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 10OpenSSL Releases 11---------------- 12 13 - [OpenSSL 3.0](#openssl-30) 14 - [OpenSSL 1.1.1](#openssl-111) 15 - [OpenSSL 1.1.0](#openssl-110) 16 - [OpenSSL 1.0.2](#openssl-102) 17 - [OpenSSL 1.0.1](#openssl-101) 18 - [OpenSSL 1.0.0](#openssl-100) 19 - [OpenSSL 0.9.x](#openssl-09x) 20 21OpenSSL 3.0 22----------- 23 24For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries 25listed here are only a brief description. 26The migration guide contains more detailed information related to new features, 27breaking changes, and mappings for the large list of deprecated functions. 28 29[Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod 30 31### Changes between 3.0.13 and 3.0.14 [4 Jun 2024] 32 33 * Fixed potential use after free after SSL_free_buffers() is called. 34 35 The SSL_free_buffers function is used to free the internal OpenSSL 36 buffer used when processing an incoming record from the network. 37 The call is only expected to succeed if the buffer is not currently 38 in use. However, two scenarios have been identified where the buffer 39 is freed even when still in use. 40 41 The first scenario occurs where a record header has been received 42 from the network and processed by OpenSSL, but the full record body 43 has not yet arrived. In this case calling SSL_free_buffers will succeed 44 even though a record has only been partially processed and the buffer 45 is still in use. 46 47 The second scenario occurs where a full record containing application 48 data has been received and processed by OpenSSL but the application has 49 only read part of this data. Again a call to SSL_free_buffers will 50 succeed even though the buffer is still in use. 51 52 ([CVE-2024-4741]) 53 54 *Matt Caswell* 55 56 * Fixed an issue where checking excessively long DSA keys or parameters may 57 be very slow. 58 59 Applications that use the functions EVP_PKEY_param_check() or 60 EVP_PKEY_public_check() to check a DSA public key or DSA parameters may 61 experience long delays. Where the key or parameters that are being checked 62 have been obtained from an untrusted source this may lead to a Denial of 63 Service. 64 65 To resolve this issue DSA keys larger than OPENSSL_DSA_MAX_MODULUS_BITS 66 will now fail the check immediately with a DSA_R_MODULUS_TOO_LARGE error 67 reason. 68 69 ([CVE-2024-4603]) 70 71 *Tomáš Mráz* 72 73 * Fixed an issue where some non-default TLS server configurations can cause 74 unbounded memory growth when processing TLSv1.3 sessions. An attacker may 75 exploit certain server configurations to trigger unbounded memory growth that 76 would lead to a Denial of Service 77 78 This problem can occur in TLSv1.3 if the non-default SSL_OP_NO_TICKET option 79 is being used (but not if early_data is also configured and the default 80 anti-replay protection is in use). In this case, under certain conditions, 81 the session cache can get into an incorrect state and it will fail to flush 82 properly as it fills. The session cache will continue to grow in an unbounded 83 manner. A malicious client could deliberately create the scenario for this 84 failure to force a Denial of Service. It may also happen by accident in 85 normal operation. 86 87 ([CVE-2024-2511]) 88 89 *Matt Caswell* 90 91 * New atexit configuration switch, which controls whether the OPENSSL_cleanup 92 is registered when libcrypto is unloaded. This can be used on platforms 93 where using atexit() from shared libraries causes crashes on exit. 94 95 *Randall S. Becker* 96 97### Changes between 3.0.12 and 3.0.13 [30 Jan 2024] 98 99 * A file in PKCS12 format can contain certificates and keys and may come from 100 an untrusted source. The PKCS12 specification allows certain fields to be 101 NULL, but OpenSSL did not correctly check for this case. A fix has been 102 applied to prevent a NULL pointer dereference that results in OpenSSL 103 crashing. If an application processes PKCS12 files from an untrusted source 104 using the OpenSSL APIs then that application will be vulnerable to this 105 issue prior to this fix. 106 107 OpenSSL APIs that were vulnerable to this are: PKCS12_parse(), 108 PKCS12_unpack_p7data(), PKCS12_unpack_p7encdata(), PKCS12_unpack_authsafes() 109 and PKCS12_newpass(). 110 111 We have also fixed a similar issue in SMIME_write_PKCS7(). However since this 112 function is related to writing data we do not consider it security 113 significant. 114 115 ([CVE-2024-0727]) 116 117 *Matt Caswell* 118 119 * When function EVP_PKEY_public_check() is called on RSA public keys, 120 a computation is done to confirm that the RSA modulus, n, is composite. 121 For valid RSA keys, n is a product of two or more large primes and this 122 computation completes quickly. However, if n is an overly large prime, 123 then this computation would take a long time. 124 125 An application that calls EVP_PKEY_public_check() and supplies an RSA key 126 obtained from an untrusted source could be vulnerable to a Denial of Service 127 attack. 128 129 The function EVP_PKEY_public_check() is not called from other OpenSSL 130 functions however it is called from the OpenSSL pkey command line 131 application. For that reason that application is also vulnerable if used 132 with the "-pubin" and "-check" options on untrusted data. 133 134 To resolve this issue RSA keys larger than OPENSSL_RSA_MAX_MODULUS_BITS will 135 now fail the check immediately with an RSA_R_MODULUS_TOO_LARGE error reason. 136 137 ([CVE-2023-6237]) 138 139 *Tomáš Mráz* 140 141 * Restore the encoding of SM2 PrivateKeyInfo and SubjectPublicKeyInfo to 142 have the contained AlgorithmIdentifier.algorithm set to id-ecPublicKey 143 rather than SM2. 144 145 *Richard Levitte* 146 147 * The POLY1305 MAC (message authentication code) implementation in OpenSSL 148 for PowerPC CPUs saves the contents of vector registers in different 149 order than they are restored. Thus the contents of some of these vector 150 registers is corrupted when returning to the caller. The vulnerable code is 151 used only on newer PowerPC processors supporting the PowerISA 2.07 152 instructions. 153 154 The consequences of this kind of internal application state corruption can 155 be various - from no consequences, if the calling application does not 156 depend on the contents of non-volatile XMM registers at all, to the worst 157 consequences, where the attacker could get complete control of the 158 application process. However unless the compiler uses the vector registers 159 for storing pointers, the most likely consequence, if any, would be an 160 incorrect result of some application dependent calculations or a crash 161 leading to a denial of service. 162 163 ([CVE-2023-6129]) 164 165 *Rohan McLure* 166 167 * Fix excessive time spent in DH check / generation with large Q parameter 168 value. 169 170 Applications that use the functions DH_generate_key() to generate an 171 X9.42 DH key may experience long delays. Likewise, applications that use 172 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check() 173 to check an X9.42 DH key or X9.42 DH parameters may experience long delays. 174 Where the key or parameters that are being checked have been obtained from 175 an untrusted source this may lead to a Denial of Service. 176 177 ([CVE-2023-5678]) 178 179 *Richard Levitte* 180 181### Changes between 3.0.11 and 3.0.12 [24 Oct 2023] 182 183 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(), 184 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters 185 that alter the key or IV length ([CVE-2023-5363]). 186 187 *Paul Dale* 188 189### Changes between 3.0.10 and 3.0.11 [19 Sep 2023] 190 191 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows. 192 193 The POLY1305 MAC (message authentication code) implementation in OpenSSL 194 does not save the contents of non-volatile XMM registers on Windows 64 195 platform when calculating the MAC of data larger than 64 bytes. Before 196 returning to the caller all the XMM registers are set to zero rather than 197 restoring their previous content. The vulnerable code is used only on newer 198 x86_64 processors supporting the AVX512-IFMA instructions. 199 200 The consequences of this kind of internal application state corruption can 201 be various - from no consequences, if the calling application does not 202 depend on the contents of non-volatile XMM registers at all, to the worst 203 consequences, where the attacker could get complete control of the 204 application process. However given the contents of the registers are just 205 zeroized so the attacker cannot put arbitrary values inside, the most likely 206 consequence, if any, would be an incorrect result of some application 207 dependent calculations or a crash leading to a denial of service. 208 209 ([CVE-2023-4807]) 210 211 *Bernd Edlinger* 212 213### Changes between 3.0.9 and 3.0.10 [1 Aug 2023] 214 215 * Fix excessive time spent checking DH q parameter value. 216 217 The function DH_check() performs various checks on DH parameters. After 218 fixing CVE-2023-3446 it was discovered that a large q parameter value can 219 also trigger an overly long computation during some of these checks. 220 A correct q value, if present, cannot be larger than the modulus p 221 parameter, thus it is unnecessary to perform these checks if q is larger 222 than p. 223 224 If DH_check() is called with such q parameter value, 225 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally 226 intensive checks are skipped. 227 228 ([CVE-2023-3817]) 229 230 *Tomáš Mráz* 231 232 * Fix DH_check() excessive time with over sized modulus. 233 234 The function DH_check() performs various checks on DH parameters. One of 235 those checks confirms that the modulus ("p" parameter) is not too large. 236 Trying to use a very large modulus is slow and OpenSSL will not normally use 237 a modulus which is over 10,000 bits in length. 238 239 However the DH_check() function checks numerous aspects of the key or 240 parameters that have been supplied. Some of those checks use the supplied 241 modulus value even if it has already been found to be too large. 242 243 A new limit has been added to DH_check of 32,768 bits. Supplying a 244 key/parameters with a modulus over this size will simply cause DH_check() to 245 fail. 246 247 ([CVE-2023-3446]) 248 249 *Matt Caswell* 250 251 * Do not ignore empty associated data entries with AES-SIV. 252 253 The AES-SIV algorithm allows for authentication of multiple associated 254 data entries along with the encryption. To authenticate empty data the 255 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`) 256 with NULL pointer as the output buffer and 0 as the input buffer length. 257 The AES-SIV implementation in OpenSSL just returns success for such call 258 instead of performing the associated data authentication operation. 259 The empty data thus will not be authenticated. ([CVE-2023-2975]) 260 261 Thanks to Juerg Wullschleger (Google) for discovering the issue. 262 263 The fix changes the authentication tag value and the ciphertext for 264 applications that use empty associated data entries with AES-SIV. 265 To decrypt data encrypted with previous versions of OpenSSL the application 266 has to skip calls to `EVP_DecryptUpdate()` for empty associated data 267 entries. 268 269 *Tomáš Mráz* 270 271### Changes between 3.0.8 and 3.0.9 [30 May 2023] 272 273 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic 274 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form. 275 276 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical 277 numeric text form. For gigantic sub-identifiers, this would take a very 278 long time, the time complexity being O(n^2) where n is the size of that 279 sub-identifier. ([CVE-2023-2650]) 280 281 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT 282 IDENTIFIER to canonical numeric text form if the size of that OBJECT 283 IDENTIFIER is 586 bytes or less, and fail otherwise. 284 285 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT 286 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at 287 most 128 sub-identifiers, and that the maximum value that each sub- 288 identifier may have is 2^32-1 (4294967295 decimal). 289 290 For each byte of every sub-identifier, only the 7 lower bits are part of 291 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with 292 these restrictions may occupy is 32 * 128 / 7, which is approximately 586 293 bytes. 294 295 *Richard Levitte* 296 297 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which 298 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can 299 trigger a crash of an application using AES-XTS decryption if the memory 300 just after the buffer being decrypted is not mapped. 301 Thanks to Anton Romanov (Amazon) for discovering the issue. 302 ([CVE-2023-1255]) 303 304 *Nevine Ebeid* 305 306 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]). 307 The previous fix for this timing side channel turned out to cause 308 a severe 2-3x performance regression in the typical use case 309 compared to 3.0.7. The new fix uses existing constant time 310 code paths, and restores the previous performance level while 311 fully eliminating all existing timing side channels. 312 The fix was developed by Bernd Edlinger with testing support 313 by Hubert Kario. 314 315 *Bernd Edlinger* 316 317 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention 318 that it does not enable policy checking. Thanks to David Benjamin for 319 discovering this issue. 320 ([CVE-2023-0466]) 321 322 *Tomáš Mráz* 323 324 * Fixed an issue where invalid certificate policies in leaf certificates are 325 silently ignored by OpenSSL and other certificate policy checks are skipped 326 for that certificate. A malicious CA could use this to deliberately assert 327 invalid certificate policies in order to circumvent policy checking on the 328 certificate altogether. 329 ([CVE-2023-0465]) 330 331 *Matt Caswell* 332 333 * Limited the number of nodes created in a policy tree to mitigate 334 against CVE-2023-0464. The default limit is set to 1000 nodes, which 335 should be sufficient for most installations. If required, the limit 336 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build 337 time define to a desired maximum number of nodes or zero to allow 338 unlimited growth. 339 ([CVE-2023-0464]) 340 341 *Paul Dale* 342 343### Changes between 3.0.7 and 3.0.8 [7 Feb 2023] 344 345 * Fixed NULL dereference during PKCS7 data verification. 346 347 A NULL pointer can be dereferenced when signatures are being 348 verified on PKCS7 signed or signedAndEnveloped data. In case the hash 349 algorithm used for the signature is known to the OpenSSL library but 350 the implementation of the hash algorithm is not available the digest 351 initialization will fail. There is a missing check for the return 352 value from the initialization function which later leads to invalid 353 usage of the digest API most likely leading to a crash. 354 ([CVE-2023-0401]) 355 356 PKCS7 data is processed by the SMIME library calls and also by the 357 time stamp (TS) library calls. The TLS implementation in OpenSSL does 358 not call these functions however third party applications would be 359 affected if they call these functions to verify signatures on untrusted 360 data. 361 362 *Tomáš Mráz* 363 364 * Fixed X.400 address type confusion in X.509 GeneralName. 365 366 There is a type confusion vulnerability relating to X.400 address processing 367 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING 368 but the public structure definition for GENERAL_NAME incorrectly specified 369 the type of the x400Address field as ASN1_TYPE. This field is subsequently 370 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather 371 than an ASN1_STRING. 372 373 When CRL checking is enabled (i.e. the application sets the 374 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to 375 pass arbitrary pointers to a memcmp call, enabling them to read memory 376 contents or enact a denial of service. 377 ([CVE-2023-0286]) 378 379 *Hugo Landau* 380 381 * Fixed NULL dereference validating DSA public key. 382 383 An invalid pointer dereference on read can be triggered when an 384 application tries to check a malformed DSA public key by the 385 EVP_PKEY_public_check() function. This will most likely lead 386 to an application crash. This function can be called on public 387 keys supplied from untrusted sources which could allow an attacker 388 to cause a denial of service attack. 389 390 The TLS implementation in OpenSSL does not call this function 391 but applications might call the function if there are additional 392 security requirements imposed by standards such as FIPS 140-3. 393 ([CVE-2023-0217]) 394 395 *Shane Lontis, Tomáš Mráz* 396 397 * Fixed Invalid pointer dereference in d2i_PKCS7 functions. 398 399 An invalid pointer dereference on read can be triggered when an 400 application tries to load malformed PKCS7 data with the 401 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions. 402 403 The result of the dereference is an application crash which could 404 lead to a denial of service attack. The TLS implementation in OpenSSL 405 does not call this function however third party applications might 406 call these functions on untrusted data. 407 ([CVE-2023-0216]) 408 409 *Tomáš Mráz* 410 411 * Fixed Use-after-free following BIO_new_NDEF. 412 413 The public API function BIO_new_NDEF is a helper function used for 414 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL 415 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also 416 be called directly by end user applications. 417 418 The function receives a BIO from the caller, prepends a new BIO_f_asn1 419 filter BIO onto the front of it to form a BIO chain, and then returns 420 the new head of the BIO chain to the caller. Under certain conditions, 421 for example if a CMS recipient public key is invalid, the new filter BIO 422 is freed and the function returns a NULL result indicating a failure. 423 However, in this case, the BIO chain is not properly cleaned up and the 424 BIO passed by the caller still retains internal pointers to the previously 425 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO 426 then a use-after-free will occur. This will most likely result in a crash. 427 ([CVE-2023-0215]) 428 429 *Viktor Dukhovni, Matt Caswell* 430 431 * Fixed Double free after calling PEM_read_bio_ex. 432 433 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and 434 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload 435 data. If the function succeeds then the "name_out", "header" and "data" 436 arguments are populated with pointers to buffers containing the relevant 437 decoded data. The caller is responsible for freeing those buffers. It is 438 possible to construct a PEM file that results in 0 bytes of payload data. 439 In this case PEM_read_bio_ex() will return a failure code but will populate 440 the header argument with a pointer to a buffer that has already been freed. 441 If the caller also frees this buffer then a double free will occur. This 442 will most likely lead to a crash. 443 444 The functions PEM_read_bio() and PEM_read() are simple wrappers around 445 PEM_read_bio_ex() and therefore these functions are also directly affected. 446 447 These functions are also called indirectly by a number of other OpenSSL 448 functions including PEM_X509_INFO_read_bio_ex() and 449 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL 450 internal uses of these functions are not vulnerable because the caller does 451 not free the header argument if PEM_read_bio_ex() returns a failure code. 452 ([CVE-2022-4450]) 453 454 *Kurt Roeckx, Matt Caswell* 455 456 * Fixed Timing Oracle in RSA Decryption. 457 458 A timing based side channel exists in the OpenSSL RSA Decryption 459 implementation which could be sufficient to recover a plaintext across 460 a network in a Bleichenbacher style attack. To achieve a successful 461 decryption an attacker would have to be able to send a very large number 462 of trial messages for decryption. The vulnerability affects all RSA padding 463 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE. 464 ([CVE-2022-4304]) 465 466 *Dmitry Belyavsky, Hubert Kario* 467 468 * Fixed X.509 Name Constraints Read Buffer Overflow. 469 470 A read buffer overrun can be triggered in X.509 certificate verification, 471 specifically in name constraint checking. The read buffer overrun might 472 result in a crash which could lead to a denial of service attack. 473 In a TLS client, this can be triggered by connecting to a malicious 474 server. In a TLS server, this can be triggered if the server requests 475 client authentication and a malicious client connects. 476 ([CVE-2022-4203]) 477 478 *Viktor Dukhovni* 479 480 * Fixed X.509 Policy Constraints Double Locking security issue. 481 482 If an X.509 certificate contains a malformed policy constraint and 483 policy processing is enabled, then a write lock will be taken twice 484 recursively. On some operating systems (most widely: Windows) this 485 results in a denial of service when the affected process hangs. Policy 486 processing being enabled on a publicly facing server is not considered 487 to be a common setup. 488 ([CVE-2022-3996]) 489 490 *Paul Dale* 491 492 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and 493 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor 494 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and 495 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting 496 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using 497 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases. 498 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to` 499 for legacy EC and SM2 keys is also changed similarly to honor the 500 equivalent conversion format flag as specified in the underlying 501 `EC_KEY` object being exported to a provider, when this function is 502 called through `EVP_PKEY_export()`. 503 504 *Nicola Tuveri* 505 506### Changes between 3.0.6 and 3.0.7 [1 Nov 2022] 507 508 * Fixed two buffer overflows in punycode decoding functions. 509 510 A buffer overrun can be triggered in X.509 certificate verification, 511 specifically in name constraint checking. Note that this occurs after 512 certificate chain signature verification and requires either a CA to 513 have signed the malicious certificate or for the application to continue 514 certificate verification despite failure to construct a path to a trusted 515 issuer. 516 517 In a TLS client, this can be triggered by connecting to a malicious 518 server. In a TLS server, this can be triggered if the server requests 519 client authentication and a malicious client connects. 520 521 An attacker can craft a malicious email address to overflow 522 an arbitrary number of bytes containing the `.` character (decimal 46) 523 on the stack. This buffer overflow could result in a crash (causing a 524 denial of service). 525 ([CVE-2022-3786]) 526 527 An attacker can craft a malicious email address to overflow four 528 attacker-controlled bytes on the stack. This buffer overflow could 529 result in a crash (causing a denial of service) or potentially remote code 530 execution depending on stack layout for any given platform/compiler. 531 ([CVE-2022-3602]) 532 533 *Paul Dale* 534 535 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT 536 parameters in OpenSSL code. 537 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR, 538 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT. 539 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead. 540 Using these invalid names may cause algorithms to use slower methods 541 that ignore the CRT parameters. 542 543 *Shane Lontis* 544 545 * Fixed a regression introduced in 3.0.6 version raising errors on some stack 546 operations. 547 548 *Tomáš Mráz* 549 550 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate 551 data to be signed before signing the certificate. 552 553 *Gibeom Gwon* 554 555 * Added RIPEMD160 to the default provider. 556 557 *Paul Dale* 558 559 * Ensured that the key share group sent or accepted for the key exchange 560 is allowed for the protocol version. 561 562 *Matt Caswell* 563 564### Changes between 3.0.5 and 3.0.6 [11 Oct 2022] 565 566 * OpenSSL supports creating a custom cipher via the legacy 567 EVP_CIPHER_meth_new() function and associated function calls. This function 568 was deprecated in OpenSSL 3.0 and application authors are instead encouraged 569 to use the new provider mechanism in order to implement custom ciphers. 570 571 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers 572 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and 573 EVP_CipherInit_ex2() functions (as well as other similarly named encryption 574 and decryption initialisation functions). Instead of using the custom cipher 575 directly it incorrectly tries to fetch an equivalent cipher from the 576 available providers. An equivalent cipher is found based on the NID passed to 577 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a 578 given cipher. However it is possible for an application to incorrectly pass 579 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef 580 is used in this way the OpenSSL encryption/decryption initialisation function 581 will match the NULL cipher as being equivalent and will fetch this from the 582 available providers. This will succeed if the default provider has been 583 loaded (or if a third party provider has been loaded that offers this 584 cipher). Using the NULL cipher means that the plaintext is emitted as the 585 ciphertext. 586 587 Applications are only affected by this issue if they call 588 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an 589 encryption/decryption initialisation function. Applications that only use 590 SSL/TLS are not impacted by this issue. 591 ([CVE-2022-3358]) 592 593 *Matt Caswell* 594 595 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures 596 on MacOS 10.11 597 598 *Richard Levitte* 599 600 * Fixed the linux-mips64 Configure target which was missing the 601 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that 602 platform. 603 604 *Adam Joseph* 605 606 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a 607 ticket 608 609 *Matt Caswell* 610 611 * Correctly handle a retransmitted ClientHello in DTLS 612 613 *Matt Caswell* 614 615 * Fixed detection of ktls support in cross-compile environment on Linux 616 617 *Tomas Mraz* 618 619 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider 620 against 3.0.x 621 622 *Paul Dale* 623 624 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to 625 report correct results in some cases 626 627 *Matt Caswell* 628 629 * Fix UWP builds by defining VirtualLock 630 631 *Charles Milette* 632 633 * For known safe primes use the minimum key length according to RFC 7919. 634 Longer private key sizes unnecessarily raise the cycles needed to compute the 635 shared secret without any increase of the real security. This fixes a 636 regression from 1.1.1 where these shorter keys were generated for the known 637 safe primes. 638 639 *Tomas Mraz* 640 641 * Added the loongarch64 target 642 643 *Shi Pujin* 644 645 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were 646 only passed to the FIPS provider and not to the default or legacy provider. 647 648 *Juergen Christ* 649 650 * Fixed reported performance degradation on aarch64. Restored the 651 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid 652 32-bit lane assignment in CTR mode") for 64bit targets only, since it is 653 reportedly 2-17% slower and the silicon errata only affects 32bit targets. 654 The new algorithm is still used for 32 bit targets. 655 656 *Bernd Edlinger* 657 658 * Added a missing header for memcmp that caused compilation failure on some 659 platforms 660 661 *Gregor Jasny* 662 663### Changes between 3.0.4 and 3.0.5 [5 Jul 2022] 664 665 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA 666 implementation for X86_64 CPUs supporting the AVX512IFMA instructions. 667 This issue makes the RSA implementation with 2048 bit private keys 668 incorrect on such machines and memory corruption will happen during 669 the computation. As a consequence of the memory corruption an attacker 670 may be able to trigger a remote code execution on the machine performing 671 the computation. 672 673 SSL/TLS servers or other servers using 2048 bit RSA private keys running 674 on machines supporting AVX512IFMA instructions of the X86_64 architecture 675 are affected by this issue. 676 ([CVE-2022-2274]) 677 678 *Xi Ruoyao* 679 680 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised 681 implementation would not encrypt the entirety of the data under some 682 circumstances. This could reveal sixteen bytes of data that was 683 preexisting in the memory that wasn't written. In the special case of 684 "in place" encryption, sixteen bytes of the plaintext would be revealed. 685 686 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS, 687 they are both unaffected. 688 ([CVE-2022-2097]) 689 690 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño* 691 692### Changes between 3.0.3 and 3.0.4 [21 Jun 2022] 693 694 * In addition to the c_rehash shell command injection identified in 695 CVE-2022-1292, further bugs where the c_rehash script does not 696 properly sanitise shell metacharacters to prevent command injection have been 697 fixed. 698 699 When the CVE-2022-1292 was fixed it was not discovered that there 700 are other places in the script where the file names of certificates 701 being hashed were possibly passed to a command executed through the shell. 702 703 This script is distributed by some operating systems in a manner where 704 it is automatically executed. On such operating systems, an attacker 705 could execute arbitrary commands with the privileges of the script. 706 707 Use of the c_rehash script is considered obsolete and should be replaced 708 by the OpenSSL rehash command line tool. 709 (CVE-2022-2068) 710 711 *Daniel Fiala, Tomáš Mráz* 712 713 * Case insensitive string comparison no longer uses locales. It has instead 714 been directly implemented. 715 716 *Paul Dale* 717 718### Changes between 3.0.2 and 3.0.3 [3 May 2022] 719 720 * Case insensitive string comparison is reimplemented via new locale-agnostic 721 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for 722 comparison. The previous implementation had problems when the Turkish locale 723 was used. 724 725 *Dmitry Belyavskiy* 726 727 * Fixed a bug in the c_rehash script which was not properly sanitising shell 728 metacharacters to prevent command injection. This script is distributed by 729 some operating systems in a manner where it is automatically executed. On 730 such operating systems, an attacker could execute arbitrary commands with the 731 privileges of the script. 732 733 Use of the c_rehash script is considered obsolete and should be replaced 734 by the OpenSSL rehash command line tool. 735 (CVE-2022-1292) 736 737 *Tomáš Mráz* 738 739 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer 740 certificate on an OCSP response. The bug caused the function in the case 741 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie 742 response (meaning a successful verification) even in the case where the 743 response signing certificate fails to verify. 744 745 It is anticipated that most users of `OCSP_basic_verify` will not use the 746 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return 747 a negative value (indicating a fatal error) in the case of a certificate 748 verification failure. The normal expected return value in this case would be 749 0. 750 751 This issue also impacts the command line OpenSSL "ocsp" application. When 752 verifying an ocsp response with the "-no_cert_checks" option the command line 753 application will report that the verification is successful even though it 754 has in fact failed. In this case the incorrect successful response will also 755 be accompanied by error messages showing the failure and contradicting the 756 apparently successful result. 757 ([CVE-2022-1343]) 758 759 *Matt Caswell* 760 761 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the 762 AAD data as the MAC key. This made the MAC key trivially predictable. 763 764 An attacker could exploit this issue by performing a man-in-the-middle attack 765 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such 766 that the modified data would still pass the MAC integrity check. 767 768 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0 769 endpoint will always be rejected by the recipient and the connection will 770 fail at that point. Many application protocols require data to be sent from 771 the client to the server first. Therefore, in such a case, only an OpenSSL 772 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client. 773 774 If both endpoints are OpenSSL 3.0 then the attacker could modify data being 775 sent in both directions. In this case both clients and servers could be 776 affected, regardless of the application protocol. 777 778 Note that in the absence of an attacker this bug means that an OpenSSL 3.0 779 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete 780 the handshake when using this ciphersuite. 781 782 The confidentiality of data is not impacted by this issue, i.e. an attacker 783 cannot decrypt data that has been encrypted using this ciphersuite - they can 784 only modify it. 785 786 In order for this attack to work both endpoints must legitimately negotiate 787 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in 788 OpenSSL 3.0, and is not available within the default provider or the default 789 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been 790 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the 791 following must have occurred: 792 793 1) OpenSSL must have been compiled with the (non-default) compile time option 794 enable-weak-ssl-ciphers 795 796 2) OpenSSL must have had the legacy provider explicitly loaded (either 797 through application code or via configuration) 798 799 3) The ciphersuite must have been explicitly added to the ciphersuite list 800 801 4) The libssl security level must have been set to 0 (default is 1) 802 803 5) A version of SSL/TLS below TLSv1.3 must have been negotiated 804 805 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any 806 others that both endpoints have in common 807 (CVE-2022-1434) 808 809 *Matt Caswell* 810 811 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory 812 occuppied by the removed hash table entries. 813 814 This function is used when decoding certificates or keys. If a long lived 815 process periodically decodes certificates or keys its memory usage will 816 expand without bounds and the process might be terminated by the operating 817 system causing a denial of service. Also traversing the empty hash table 818 entries will take increasingly more time. 819 820 Typically such long lived processes might be TLS clients or TLS servers 821 configured to accept client certificate authentication. 822 (CVE-2022-1473) 823 824 *Hugo Landau, Aliaksei Levin* 825 826 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report 827 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other 828 statistics are no longer supported. For compatibility, these statistics are 829 still listed in the output but are now always reported as zero. 830 831 *Hugo Landau* 832 833### Changes between 3.0.1 and 3.0.2 [15 Mar 2022] 834 835 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever 836 for non-prime moduli. 837 838 Internally this function is used when parsing certificates that contain 839 elliptic curve public keys in compressed form or explicit elliptic curve 840 parameters with a base point encoded in compressed form. 841 842 It is possible to trigger the infinite loop by crafting a certificate that 843 has invalid explicit curve parameters. 844 845 Since certificate parsing happens prior to verification of the certificate 846 signature, any process that parses an externally supplied certificate may thus 847 be subject to a denial of service attack. The infinite loop can also be 848 reached when parsing crafted private keys as they can contain explicit 849 elliptic curve parameters. 850 851 Thus vulnerable situations include: 852 853 - TLS clients consuming server certificates 854 - TLS servers consuming client certificates 855 - Hosting providers taking certificates or private keys from customers 856 - Certificate authorities parsing certification requests from subscribers 857 - Anything else which parses ASN.1 elliptic curve parameters 858 859 Also any other applications that use the BN_mod_sqrt() where the attacker 860 can control the parameter values are vulnerable to this DoS issue. 861 ([CVE-2022-0778]) 862 863 *Tomáš Mráz* 864 865 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489) 866 to the list of ciphersuites providing Perfect Forward Secrecy as 867 required by SECLEVEL >= 3. 868 869 *Dmitry Belyavskiy, Nicola Tuveri* 870 871 * Made the AES constant time code for no-asm configurations 872 optional due to the resulting 95% performance degradation. 873 The AES constant time code can be enabled, for no assembly 874 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME 875 876 *Paul Dale* 877 878 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty 879 passphrase strings. 880 881 *Darshan Sen* 882 883 * The negative return value handling of the certificate verification callback 884 was reverted. The replacement is to set the verification retry state with 885 the SSL_set_retry_verify() function. 886 887 *Tomáš Mráz* 888 889### Changes between 3.0.0 and 3.0.1 [14 Dec 2021] 890 891 * Fixed invalid handling of X509_verify_cert() internal errors in libssl 892 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to 893 verify a certificate supplied by a server. That function may return a 894 negative return value to indicate an internal error (for example out of 895 memory). Such a negative return value is mishandled by OpenSSL and will cause 896 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate 897 success and a subsequent call to SSL_get_error() to return the value 898 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be 899 returned by OpenSSL if the application has previously called 900 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this 901 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be 902 totally unexpected and applications may not behave correctly as a result. The 903 exact behaviour will depend on the application but it could result in 904 crashes, infinite loops or other similar incorrect responses. 905 906 This issue is made more serious in combination with a separate bug in OpenSSL 907 3.0 that will cause X509_verify_cert() to indicate an internal error when 908 processing a certificate chain. This will occur where a certificate does not 909 include the Subject Alternative Name extension but where a Certificate 910 Authority has enforced name constraints. This issue can occur even with valid 911 chains. 912 ([CVE-2021-4044]) 913 914 *Matt Caswell* 915 916 * Corrected a few file name and file reference bugs in the build, 917 installation and setup scripts, which lead to installation verification 918 failures. Slightly enhanced the installation verification script. 919 920 *Richard Levitte* 921 922 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private 923 keys. 924 925 *Richard Levitte* 926 927 * Fixed PVK encoder to properly query for the passphrase. 928 929 *Tomáš Mráz* 930 931 * Multiple fixes in the OSSL_HTTP API functions. 932 933 *David von Oheimb* 934 935 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the 936 OSSL_PARAM_INTEGER data type and return error on negative numbers 937 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make 938 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers. 939 940 *Richard Levitte* 941 942 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex. 943 944 *Tomáš Mráz* 945 946 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD. 947 948 *Allan Jude* 949 950 * Multiple threading fixes. 951 952 *Matt Caswell* 953 954 * Added NULL digest implementation to keep compatibility with 1.1.1 version. 955 956 *Tomáš Mráz* 957 958 * Allow fetching an operation from the provider that owns an unexportable key 959 as a fallback if that is still allowed by the property query. 960 961 *Richard Levitte* 962 963### Changes between 1.1.1 and 3.0.0 [7 sep 2021] 964 965 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now 966 deprecated. 967 968 *Matt Caswell* 969 970 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the 971 S390X capability vector to zero. This simplifies testing of different code 972 paths on S390X architecture. 973 974 *Patrick Steuer* 975 976 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed 977 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from 978 SP 800-38D". The communication will fail at this point. 979 980 *Paul Dale* 981 982 * The EC_GROUP_clear_free() function is deprecated as there is nothing 983 confidential in EC_GROUP data. 984 985 *Nicola Tuveri* 986 987 * The byte order mark (BOM) character is ignored if encountered at the 988 beginning of a PEM-formatted file. 989 990 *Dmitry Belyavskiy* 991 992 * Added CMS support for the Russian GOST algorithms. 993 994 *Dmitry Belyavskiy* 995 996 * Due to move of the implementation of cryptographic operations 997 to the providers, validation of various operation parameters can 998 be postponed until the actual operation is executed where previously 999 it happened immediately when an operation parameter was set. 1000 1001 For example when setting an unsupported curve with 1002 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not 1003 fail but later keygen operations with the EVP_PKEY_CTX will fail. 1004 1005 *OpenSSL team members and many third party contributors* 1006 1007 * The EVP_get_cipherbyname() function will return NULL for algorithms such as 1008 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were 1009 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch() 1010 instead to retrieve these algorithms from a provider. 1011 1012 *Shane Lontis* 1013 1014 * On build targets where the multilib postfix is set in the build 1015 configuration the libdir directory was changing based on whether 1016 the lib directory with the multilib postfix exists on the system 1017 or not. This unpredictable behavior was removed and eventual 1018 multilib postfix is now always added to the default libdir. Use 1019 `--libdir=lib` to override the libdir if adding the postfix is 1020 undesirable. 1021 1022 *Jan Lána* 1023 1024 * The triple DES key wrap functionality now conforms to RFC 3217 but is 1025 no longer interoperable with OpenSSL 1.1.1. 1026 1027 *Paul Dale* 1028 1029 * The ERR_GET_FUNC() function was removed. With the loss of meaningful 1030 function codes, this function can only cause problems for calling 1031 applications. 1032 1033 *Paul Dale* 1034 1035 * Add a configurable flag to output date formats as ISO 8601. Does not 1036 change the default date format. 1037 1038 *William Edmisten* 1039 1040 * Version of MSVC earlier than 1300 could get link warnings, which could 1041 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set. 1042 Support for this flag has been removed. 1043 1044 *Rich Salz* 1045 1046 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG, 1047 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for 1048 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG 1049 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set 1050 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency. 1051 1052 *Rich Salz* 1053 1054 * The signatures of the functions to get and set options on SSL and 1055 SSL_CTX objects changed from "unsigned long" to "uint64_t" type. 1056 Some source code changes may be required. 1057 1058 *Rich Salz* 1059 1060 * The public definitions of conf_method_st and conf_st have been 1061 deprecated. They will be made opaque in a future release. 1062 1063 *Rich Salz and Tomáš Mráz* 1064 1065 * Client-initiated renegotiation is disabled by default. To allow it, use 1066 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION 1067 flag, or the "ClientRenegotiation" config parameter as appropriate. 1068 1069 *Rich Salz* 1070 1071 * Add "abspath" and "includedir" pragma's to config files, to prevent, 1072 or modify relative pathname inclusion. 1073 1074 *Rich Salz* 1075 1076 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2 1077 validated. Please consult the README-FIPS and 1078 README-PROVIDERS files, as well as the migration guide. 1079 1080 *OpenSSL team members and many third party contributors* 1081 1082 * For the key types DH and DHX the allowed settable parameters are now different. 1083 1084 *Shane Lontis* 1085 1086 * The openssl commands that read keys, certificates, and CRLs now 1087 automatically detect the PEM or DER format of the input files. 1088 1089 *David von Oheimb, Richard Levitte, and Tomáš Mráz* 1090 1091 * Added enhanced PKCS#12 APIs which accept a library context. 1092 1093 *Jon Spillett* 1094 1095 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl" 1096 1097 *Matt Caswell* 1098 1099 * Added support for Kernel TLS (KTLS). 1100 1101 *Boris Pismenny, John Baldwin and Andrew Gallatin* 1102 1103 * Support for RFC 5746 secure renegotiation is now required by default for 1104 SSL or TLS connections to succeed. 1105 1106 *Benjamin Kaduk* 1107 1108 * The signature of the `copy` functional parameter of the 1109 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is 1110 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly 1111 the signature of the `pub_decode` functional parameter of the 1112 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is 1113 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`. 1114 1115 *David von Oheimb* 1116 1117 * The error return values from some control calls (ctrl) have changed. 1118 1119 *Paul Dale* 1120 1121 * A public key check is now performed during EVP_PKEY_derive_set_peer(). 1122 1123 *Shane Lontis* 1124 1125 * Many functions in the EVP_ namespace that are getters of values from 1126 implementations or contexts were renamed to include get or get0 in their 1127 names. Old names are provided as macro aliases for compatibility and 1128 are not deprecated. 1129 1130 *Tomáš Mráz* 1131 1132 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT, 1133 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT, 1134 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations 1135 are deprecated. 1136 1137 *Tomáš Mráz* 1138 1139 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for 1140 more key types. 1141 1142 * The output from the command line applications may have minor 1143 changes. 1144 1145 *Paul Dale* 1146 1147 * The output from numerous "printing" may have minor changes. 1148 1149 *David von Oheimb* 1150 1151 * Windows thread synchronization uses read/write primitives (SRWLock) when 1152 supported by the OS, otherwise CriticalSection continues to be used. 1153 1154 *Vincent Drake* 1155 1156 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to 1157 work on read only BIO source/sinks that do not support these functions. 1158 This allows piping or redirection of a file BIO using stdin to be buffered 1159 into memory. This is used internally in OSSL_DECODER_from_bio(). 1160 1161 *Shane Lontis* 1162 1163 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1 1164 this function would return one of the values OSSL_STORE_INFO_NAME, 1165 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or 1166 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported 1167 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now 1168 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications 1169 using this function should be amended to handle the changed return value. 1170 1171 *Richard Levitte* 1172 1173 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035) 1174 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations. 1175 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present. 1176 Correct the semantics of checking the validation chain in case ESSCertID{,v2} 1177 contains more than one certificate identifier: This means that all 1178 certificates referenced there MUST be part of the validation chain. 1179 1180 *David von Oheimb* 1181 1182 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4, 1183 RC5, DESX and DES have been moved to the legacy provider. 1184 1185 *Matt Caswell* 1186 1187 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and 1188 RIPEMD-160 have been moved to the legacy provider. 1189 1190 *Matt Caswell* 1191 1192 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a 1193 provided key. 1194 1195 *Dmitry Belyavskiy* 1196 1197 * The deprecated functions EVP_PKEY_get0_RSA(), 1198 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(), 1199 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as 1200 well as the similarly named "get1" functions behave differently in 1201 OpenSSL 3.0. 1202 1203 *Matt Caswell* 1204 1205 * A number of functions handling low-level keys or engines were deprecated 1206 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(), 1207 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and 1208 EVP_PKEY_get0_siphash(). 1209 1210 *Matt Caswell* 1211 1212 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into 1213 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF 1214 will need to load the legacy crypto provider. This includes these PBE 1215 algorithms which use this KDF: 1216 - NID_pbeWithMD2AndDES_CBC 1217 - NID_pbeWithMD5AndDES_CBC 1218 - NID_pbeWithSHA1AndRC2_CBC 1219 - NID_pbeWithMD2AndRC2_CBC 1220 - NID_pbeWithMD5AndRC2_CBC 1221 - NID_pbeWithSHA1AndDES_CBC 1222 1223 *Jon Spillett* 1224 1225 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and 1226 BIO_debug_callback() functions. 1227 1228 *Tomáš Mráz* 1229 1230 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and 1231 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. 1232 1233 *Tomáš Mráz* 1234 1235 * The RAND_METHOD APIs have been deprecated. 1236 1237 *Paul Dale* 1238 1239 * The SRP APIs have been deprecated. 1240 1241 *Matt Caswell* 1242 1243 * Add a compile time option to prevent the caching of provider fetched 1244 algorithms. This is enabled by including the no-cached-fetch option 1245 at configuration time. 1246 1247 *Paul Dale* 1248 1249 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration 1250 count of PKCS12_DEFAULT_ITER. 1251 1252 *Tomáš Mráz and Sahana Prasad* 1253 1254 * The openssl speed command does not use low-level API calls anymore. 1255 1256 *Tomáš Mráz* 1257 1258 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA 1259 capable processors. 1260 1261 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)* 1262 1263 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3. 1264 1265 *Matt Caswell* 1266 1267 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that 1268 providers may supply their own group implementations (using either the "key 1269 exchange" or the "key encapsulation" methods) which will automatically be 1270 detected and used by libssl. 1271 1272 *Matt Caswell, Nicola Tuveri* 1273 1274 * The undocumented function X509_certificate_type() has been deprecated; 1275 1276 *Rich Salz* 1277 1278 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range(). 1279 1280 *Tomáš Mráz* 1281 1282 * Removed RSA padding mode for SSLv23 (which was only used for 1283 SSLv2). This includes the functions RSA_padding_check_SSLv23() and 1284 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated 1285 `rsautl` command. 1286 1287 *Rich Salz* 1288 1289 * Deprecated the obsolete X9.31 RSA key generation related functions. 1290 1291 * While a callback function set via `SSL_CTX_set_cert_verify_callback()` 1292 is not allowed to return a value > 1, this is no more taken as failure. 1293 1294 *Viktor Dukhovni and David von Oheimb* 1295 1296 * Deprecated the obsolete X9.31 RSA key generation related functions 1297 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and 1298 BN_X931_generate_prime_ex(). 1299 1300 *Tomáš Mráz* 1301 1302 * The default key generation method for the regular 2-prime RSA keys was 1303 changed to the FIPS 186-4 B.3.6 method. 1304 1305 *Shane Lontis* 1306 1307 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions. 1308 1309 *Kurt Roeckx* 1310 1311 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn(). 1312 1313 *Rich Salz* 1314 1315 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and 1316 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*(). 1317 1318 *Rich Salz, Richard Levitte, and David von Oheimb* 1319 1320 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`. 1321 1322 *David von Oheimb* 1323 1324 * Deprecated `OCSP_parse_url()`. 1325 1326 *David von Oheimb* 1327 1328 * Validation of SM2 keys has been separated from the validation of regular EC 1329 keys. 1330 1331 *Nicola Tuveri* 1332 1333 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck` 1334 switches: a validation failure triggers an early exit, returning a failure 1335 exit status to the parent process. 1336 1337 *Nicola Tuveri* 1338 1339 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites() 1340 to ignore unknown ciphers. 1341 1342 *Otto Hollmann* 1343 1344 * The `-cipher-commands` and `-digest-commands` options 1345 of the command line utility `list` have been deprecated. 1346 Instead use the `-cipher-algorithms` and `-digest-algorithms` options. 1347 1348 *Dmitry Belyavskiy* 1349 1350 * Added convenience functions for generating asymmetric key pairs: 1351 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)> 1352 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>. 1353 1354 *David von Oheimb* 1355 1356 * All of the low level EC_KEY functions have been deprecated. 1357 1358 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz* 1359 1360 * Deprecated all the libcrypto and libssl error string loading 1361 functions. 1362 1363 *Richard Levitte* 1364 1365 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as 1366 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been 1367 deprecated. 1368 1369 *Matt Caswell* 1370 1371 * The `-crypt` option to the `passwd` command line tool has been removed. 1372 1373 *Paul Dale* 1374 1375 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands 1376 were removed. 1377 1378 *Rich Salz* 1379 1380 * Add support for AES Key Wrap inverse ciphers to the EVP layer. 1381 1382 *Shane Lontis* 1383 1384 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and 1385 EVP_PKEY_get1_tls_encodedpoint(). 1386 1387 *Matt Caswell* 1388 1389 * The security callback, which can be customised by application code, supports 1390 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter 1391 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases. 1392 1393 *Matt Caswell* 1394 1395 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public 1396 interface. Their functionality remains unchanged. 1397 1398 *Jordan Montgomery* 1399 1400 * Added new option for 'openssl list', '-providers', which will display the 1401 list of loaded providers, their names, version and status. It optionally 1402 displays their gettable parameters. 1403 1404 *Paul Dale* 1405 1406 * Removed EVP_PKEY_set_alias_type(). 1407 1408 *Richard Levitte* 1409 1410 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced 1411 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred. 1412 1413 *Jeremy Walch* 1414 1415 * Changed all "STACK" functions to be macros instead of inline functions. Macro 1416 parameters are still checked for type safety at compile time via helper 1417 inline functions. 1418 1419 *Matt Caswell* 1420 1421 * Remove the RAND_DRBG API 1422 1423 *Paul Dale and Matthias St. Pierre* 1424 1425 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses 1426 as well as actual hostnames. 1427 1428 *David Woodhouse* 1429 1430 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently 1431 ignore TLS protocol version bounds when configuring DTLS-based contexts, and 1432 conversely, silently ignore DTLS protocol version bounds when configuring 1433 TLS-based contexts. The commands can be repeated to set bounds of both 1434 types. The same applies with the corresponding "min_protocol" and 1435 "max_protocol" command-line switches, in case some application uses both TLS 1436 and DTLS. 1437 1438 SSL_CTX instances that are created for a fixed protocol version (e.g. 1439 `TLSv1_server_method()`) also silently ignore version bounds. Previously 1440 attempts to apply bounds to these protocol versions would result in an 1441 error. Now only the "version-flexible" SSL_CTX instances are subject to 1442 limits in configuration files in command-line options. 1443 1444 *Viktor Dukhovni* 1445 1446 * Deprecated the `ENGINE` API. Engines should be replaced with providers 1447 going forward. 1448 1449 *Paul Dale* 1450 1451 * Reworked the recorded ERR codes to make better space for system errors. 1452 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the 1453 given code is a system error (true) or an OpenSSL error (false). 1454 1455 *Richard Levitte* 1456 1457 * Reworked the test perl framework to better allow parallel testing. 1458 1459 *Nicola Tuveri and David von Oheimb* 1460 1461 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and 1462 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported. 1463 1464 *Shane Lontis* 1465 1466 * 'Configure' has been changed to figure out the configuration target if 1467 none is given on the command line. Consequently, the 'config' script is 1468 now only a mere wrapper. All documentation is changed to only mention 1469 'Configure'. 1470 1471 *Rich Salz and Richard Levitte* 1472 1473 * Added a library context `OSSL_LIB_CTX` that applications as well as 1474 other libraries can use to form a separate context within which 1475 libcrypto operations are performed. 1476 1477 *Richard Levitte* 1478 1479 * Added various `_ex` functions to the OpenSSL API that support using 1480 a non-default `OSSL_LIB_CTX`. 1481 1482 *OpenSSL team* 1483 1484 * Handshake now fails if Extended Master Secret extension is dropped 1485 on renegotiation. 1486 1487 *Tomáš Mráz* 1488 1489 * Dropped interactive mode from the `openssl` program. 1490 1491 *Richard Levitte* 1492 1493 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`. 1494 1495 *David von Oheimb and Shane Lontis* 1496 1497 * Deprecated `EC_METHOD_get_field_type()`. 1498 1499 *Billy Bob Brumley* 1500 1501 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(), 1502 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method() 1503 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method(). 1504 1505 *Billy Bob Brumley* 1506 1507 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of(). 1508 1509 *Billy Bob Brumley* 1510 1511 * Add CAdES-BES signature verification support, mostly derived 1512 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein. 1513 1514 *Filipe Raimundo da Silva* 1515 1516 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API. 1517 1518 *Antonio Iacono* 1519 1520 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM 1521 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS). 1522 1523 *Jakub Zelenka* 1524 1525 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). 1526 1527 *Billy Bob Brumley* 1528 1529 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and 1530 EC_KEY_precompute_mult(). 1531 1532 *Billy Bob Brumley* 1533 1534 * Deprecated EC_POINTs_mul(). 1535 1536 *Billy Bob Brumley* 1537 1538 * Removed FIPS_mode() and FIPS_mode_set(). 1539 1540 *Shane Lontis* 1541 1542 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. 1543 1544 *Dmitry Belyavskiy* 1545 1546 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and 1547 EC_POINT_get_Jprojective_coordinates_GFp(). 1548 1549 *Billy Bob Brumley* 1550 1551 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM 1552 arrays to be more easily constructed via a series of utility functions. 1553 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using 1554 the various push functions and finally convert to a passable OSSL_PARAM 1555 array using OSSL_PARAM_BLD_to_param(). 1556 1557 *Paul Dale* 1558 1559 * The security strength of SHA1 and MD5 based signatures in TLS has been 1560 reduced. 1561 1562 *Kurt Roeckx* 1563 1564 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to 1565 contain a provider side internal key. 1566 1567 *Richard Levitte* 1568 1569 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated. 1570 1571 *Richard Levitte* 1572 1573 * Project text documents not yet having a proper file name extension 1574 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to 1575 `*.md` as far as reasonable, else `*.txt`, for better use with file managers. 1576 1577 *David von Oheimb* 1578 1579 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT) 1580 have been converted to Markdown with the goal to produce documents 1581 which not only look pretty when viewed online in the browser, but 1582 remain well readable inside a plain text editor. 1583 1584 To achieve this goal, a 'minimalistic' Markdown style has been applied 1585 which avoids formatting elements that interfere too much with the 1586 reading flow in the text file. For example, it 1587 1588 * avoids [ATX headings][] and uses [setext headings][] instead 1589 (which works for `<h1>` and `<h2>` headings only). 1590 * avoids [inline links][] and uses [reference links][] instead. 1591 * avoids [fenced code blocks][] and uses [indented code blocks][] instead. 1592 1593 [ATX headings]: https://github.github.com/gfm/#atx-headings 1594 [setext headings]: https://github.github.com/gfm/#setext-headings 1595 [inline links]: https://github.github.com/gfm/#inline-link 1596 [reference links]: https://github.github.com/gfm/#reference-link 1597 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks 1598 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks 1599 1600 *Matthias St. Pierre* 1601 1602 * The test suite is changed to preserve results of each test recipe. 1603 A new directory test-runs/ with subdirectories named like the 1604 test recipes are created in the build tree for this purpose. 1605 1606 *Richard Levitte* 1607 1608 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712). 1609 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`. 1610 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points. 1611 1612 *David von Oheimb, Martin Peylo* 1613 1614 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`. 1615 It supports arbitrary request and response content types, GET redirection, 1616 TLS, connections via HTTP(S) proxies, connections and exchange via 1617 user-defined BIOs (allowing implicit connections), persistent connections, 1618 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details. 1619 The legacy OCSP-focused (and only partly documented) API 1620 is retained for backward compatibility, while most of it is deprecated. 1621 1622 *David von Oheimb* 1623 1624 * Added `util/check-format.pl`, a tool for checking adherence to the 1625 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>. 1626 The checks performed are incomplete and yield some false positives. 1627 Still the tool should be useful for detecting most typical glitches. 1628 1629 *David von Oheimb* 1630 1631 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended: 1632 If domain name resolution yields multiple IP addresses all of them are tried 1633 after `connect()` failures. 1634 1635 *David von Oheimb* 1636 1637 * All of the low level RSA functions have been deprecated. 1638 1639 *Paul Dale* 1640 1641 * X509 certificates signed using SHA1 are no longer allowed at security 1642 level 1 and above. 1643 1644 *Kurt Roeckx* 1645 1646 * The command line utilities dhparam, dsa, gendsa and dsaparam have been 1647 modified to use PKEY APIs. These commands are now in maintenance mode 1648 and no new features will be added to them. 1649 1650 *Paul Dale* 1651 1652 * The command line utility rsautl has been deprecated. 1653 1654 *Paul Dale* 1655 1656 * The command line utilities genrsa and rsa have been modified to use PKEY 1657 APIs. They now write PKCS#8 keys by default. These commands are now in 1658 maintenance mode and no new features will be added to them. 1659 1660 *Paul Dale* 1661 1662 * All of the low level DH functions have been deprecated. 1663 1664 *Paul Dale and Matt Caswell* 1665 1666 * All of the low level DSA functions have been deprecated. 1667 1668 *Paul Dale* 1669 1670 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to 1671 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC. 1672 1673 *Richard Levitte* 1674 1675 * Deprecated low level ECDH and ECDSA functions. 1676 1677 *Paul Dale* 1678 1679 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old(). 1680 1681 *Richard Levitte* 1682 1683 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits() 1684 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed 1685 a new formulation to include all the things it can be used for, 1686 as well as words of caution. 1687 1688 *Richard Levitte* 1689 1690 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated. 1691 1692 *Paul Dale* 1693 1694 * All of the low level HMAC functions have been deprecated. 1695 1696 *Paul Dale and David von Oheimb* 1697 1698 * Over two thousand fixes were made to the documentation, including: 1699 - Common options (such as -rand/-writerand, TLS version control, etc) 1700 were refactored and point to newly-enhanced descriptions in openssl.pod. 1701 - Added style conformance for all options (with help from Richard Levitte), 1702 documented all reported missing options, added a CI build to check 1703 that all options are documented and that no unimplemented options 1704 are documented. 1705 - Documented some internals, such as all use of environment variables. 1706 - Addressed all internal broken L<> references. 1707 1708 *Rich Salz* 1709 1710 * All of the low level CMAC functions have been deprecated. 1711 1712 *Paul Dale* 1713 1714 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest 1715 functions have been deprecated. 1716 1717 *Paul Dale and David von Oheimb* 1718 1719 * Corrected the documentation of the return values from the `EVP_DigestSign*` 1720 set of functions. The documentation mentioned negative values for some 1721 errors, but this was never the case, so the mention of negative values 1722 was removed. 1723 1724 Code that followed the documentation and thereby check with something 1725 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed. 1726 1727 *Richard Levitte* 1728 1729 * All of the low level cipher functions have been deprecated. 1730 1731 *Matt Caswell and Paul Dale* 1732 1733 * Removed include/openssl/opensslconf.h.in and replaced it with 1734 include/openssl/configuration.h.in, which differs in not including 1735 <openssl/macros.h>. A short header include/openssl/opensslconf.h 1736 was added to include both. 1737 1738 This allows internal hacks where one might need to modify the set 1739 of configured macros, for example this if deprecated symbols are 1740 still supposed to be available internally: 1741 1742 #include <openssl/configuration.h> 1743 1744 #undef OPENSSL_NO_DEPRECATED 1745 #define OPENSSL_SUPPRESS_DEPRECATED 1746 1747 #include <openssl/macros.h> 1748 1749 This should not be used by applications that use the exported 1750 symbols, as that will lead to linking errors. 1751 1752 *Richard Levitte* 1753 1754 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure 1755 used in exponentiation with 512-bit moduli. No EC algorithms are 1756 affected. Analysis suggests that attacks against 2-prime RSA1024, 1757 3-prime RSA1536, and DSA1024 as a result of this defect would be very 1758 difficult to perform and are not believed likely. Attacks against DH512 1759 are considered just feasible. However, for an attack the target would 1760 have to re-use the DH512 private key, which is not recommended anyway. 1761 Also applications directly using the low-level API BN_mod_exp may be 1762 affected if they use BN_FLG_CONSTTIME. 1763 ([CVE-2019-1551]) 1764 1765 *Andy Polyakov* 1766 1767 * Most memory-debug features have been deprecated, and the functionality 1768 replaced with no-ops. 1769 1770 *Rich Salz* 1771 1772 * Added documentation for the STACK API. 1773 1774 *Rich Salz* 1775 1776 * Introduced a new method type and API, OSSL_ENCODER, to represent 1777 generic encoders. These do the same sort of job that PEM writers 1778 and d2i functions do, but with support for methods supplied by 1779 providers, and the possibility for providers to support other 1780 formats as well. 1781 1782 *Richard Levitte* 1783 1784 * Introduced a new method type and API, OSSL_DECODER, to represent 1785 generic decoders. These do the same sort of job that PEM readers 1786 and i2d functions do, but with support for methods supplied by 1787 providers, and the possibility for providers to support other 1788 formats as well. 1789 1790 *Richard Levitte* 1791 1792 * Added a .pragma directive to the syntax of configuration files, to 1793 allow varying behavior in a supported and predictable manner. 1794 Currently added pragma: 1795 1796 .pragma dollarid:on 1797 1798 This allows dollar signs to be a keyword character unless it's 1799 followed by a opening brace or parenthesis. This is useful for 1800 platforms where dollar signs are commonly used in names, such as 1801 volume names and system directory names on VMS. 1802 1803 *Richard Levitte* 1804 1805 * Added functionality to create an EVP_PKEY from user data. 1806 1807 *Richard Levitte* 1808 1809 * Change the interpretation of the '--api' configuration option to 1810 mean that this is a desired API compatibility level with no 1811 further meaning. The previous interpretation, that this would 1812 also mean to remove all deprecated symbols up to and including 1813 the given version, no requires that 'no-deprecated' is also used 1814 in the configuration. 1815 1816 When building applications, the desired API compatibility level 1817 can be set with the OPENSSL_API_COMPAT macro like before. For 1818 API compatibility version below 3.0, the old style numerical 1819 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L. 1820 For version 3.0 and on, the value is expected to be the decimal 1821 value calculated from the major and minor version like this: 1822 1823 MAJOR * 10000 + MINOR * 100 1824 1825 Examples: 1826 1827 -DOPENSSL_API_COMPAT=30000 For 3.0 1828 -DOPENSSL_API_COMPAT=30200 For 3.2 1829 1830 To hide declarations that are deprecated up to and including the 1831 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be 1832 given when building the application as well. 1833 1834 *Richard Levitte* 1835 1836 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow 1837 access to certificate and CRL stores via URIs and OSSL_STORE 1838 loaders. 1839 1840 This adds the following functions: 1841 1842 - X509_LOOKUP_store() 1843 - X509_STORE_load_file() 1844 - X509_STORE_load_path() 1845 - X509_STORE_load_store() 1846 - SSL_add_store_cert_subjects_to_stack() 1847 - SSL_CTX_set_default_verify_store() 1848 - SSL_CTX_load_verify_file() 1849 - SSL_CTX_load_verify_dir() 1850 - SSL_CTX_load_verify_store() 1851 1852 *Richard Levitte* 1853 1854 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY. 1855 The presence of this system service is determined at run-time. 1856 1857 *Richard Levitte* 1858 1859 * Added functionality to create an EVP_PKEY context based on data 1860 for methods from providers. This takes an algorithm name and a 1861 property query string and simply stores them, with the intent 1862 that any operation that uses this context will use those strings 1863 to fetch the needed methods implicitly, thereby making the port 1864 of application written for pre-3.0 OpenSSL easier. 1865 1866 *Richard Levitte* 1867 1868 * The undocumented function NCONF_WIN32() has been deprecated; for 1869 conversion details see the HISTORY section of doc/man5/config.pod 1870 1871 *Rich Salz* 1872 1873 * Introduced the new functions EVP_DigestSignInit_ex() and 1874 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and 1875 EVP_DigestVerifyUpdate() have been converted to functions. See the man 1876 pages for further details. 1877 1878 *Matt Caswell* 1879 1880 * Over two thousand fixes were made to the documentation, including: 1881 adding missing command flags, better style conformance, documentation 1882 of internals, etc. 1883 1884 *Rich Salz, Richard Levitte* 1885 1886 * s390x assembly pack: add hardware-support for P-256, P-384, P-521, 1887 X25519, X448, Ed25519 and Ed448. 1888 1889 *Patrick Steuer* 1890 1891 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just 1892 the first value. 1893 1894 *Jon Spillett* 1895 1896 * Deprecated the public definition of `ERR_STATE` as well as the function 1897 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an 1898 opaque type. 1899 1900 *Richard Levitte* 1901 1902 * Added ERR functionality to give callers access to the stored function 1903 names that have replaced the older function code based functions. 1904 1905 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(), 1906 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(), 1907 ERR_peek_error_all() and ERR_peek_last_error_all(). 1908 1909 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(), 1910 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and 1911 ERR_func_error_string(). 1912 1913 *Richard Levitte* 1914 1915 * Extended testing to be verbose for failing tests only. The make variables 1916 VERBOSE_FAILURE or VF can be used to enable this: 1917 1918 $ make VF=1 test # Unix 1919 $ mms /macro=(VF=1) test ! OpenVMS 1920 $ nmake VF=1 test # Windows 1921 1922 *Richard Levitte* 1923 1924 * Added the `-copy_extensions` option to the `x509` command for use with 1925 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument, 1926 all extensions in the request are copied to the certificate or vice versa. 1927 1928 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>* 1929 1930 * Added the `-copy_extensions` option to the `req` command for use with 1931 `-x509`. When given with the `copy` or `copyall` argument, 1932 all extensions in the certification request are copied to the certificate. 1933 1934 *David von Oheimb* 1935 1936 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates 1937 they generate are by default RFC 5280 compliant in the following sense: 1938 There is a subjectKeyIdentifier extension with a hash value of the public key 1939 and for not self-signed certs there is an authorityKeyIdentifier extension 1940 with a keyIdentifier field or issuer information identifying the signing key. 1941 This is done unless some configuration overrides the new default behavior, 1942 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`. 1943 1944 *David von Oheimb* 1945 1946 * Added several checks to `X509_verify_cert()` according to requirements in 1947 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set 1948 (which may be done by using the CLI option `-x509_strict`): 1949 * The basicConstraints of CA certificates must be marked critical. 1950 * CA certificates must explicitly include the keyUsage extension. 1951 * If a pathlenConstraint is given the key usage keyCertSign must be allowed. 1952 * The issuer name of any certificate must not be empty. 1953 * The subject name of CA certs, certs with keyUsage crlSign, 1954 and certs without subjectAlternativeName must not be empty. 1955 * If a subjectAlternativeName extension is given it must not be empty. 1956 * The signatureAlgorithm field and the cert signature must be consistent. 1957 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier 1958 must not be marked critical. 1959 * The authorityKeyIdentifier must be given for X.509v3 certs 1960 unless they are self-signed. 1961 * The subjectKeyIdentifier must be given for all X.509v3 CA certs. 1962 1963 *David von Oheimb* 1964 1965 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys 1966 with explicit curve parameters (specifiedCurve) as required by RFC 5480. 1967 1968 *Tomáš Mráz* 1969 1970 * For built-in EC curves, ensure an EC_GROUP built from the curve name is 1971 used even when parsing explicit parameters, when loading a encoded key 1972 or calling `EC_GROUP_new_from_ecpkparameters()`/ 1973 `EC_GROUP_new_from_ecparameters()`. 1974 This prevents bypass of security hardening and performance gains, 1975 especially for curves with specialized EC_METHODs. 1976 By default, if a key encoded with explicit parameters is loaded and later 1977 encoded, the output is still encoded with explicit parameters, even if 1978 internally a "named" EC_GROUP is used for computation. 1979 1980 *Nicola Tuveri* 1981 1982 * Compute ECC cofactors if not provided during EC_GROUP construction. Before 1983 this change, EC_GROUP_set_generator would accept order and/or cofactor as 1984 NULL. After this change, only the cofactor parameter can be NULL. It also 1985 does some minimal sanity checks on the passed order. 1986 ([CVE-2019-1547]) 1987 1988 *Billy Bob Brumley* 1989 1990 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey. 1991 An attack is simple, if the first CMS_recipientInfo is valid but the 1992 second CMS_recipientInfo is chosen ciphertext. If the second 1993 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct 1994 encryption key will be replaced by garbage, and the message cannot be 1995 decoded, but if the RSA decryption fails, the correct encryption key is 1996 used and the recipient will not notice the attack. 1997 As a work around for this potential attack the length of the decrypted 1998 key must be equal to the cipher default key length, in case the 1999 certifiate is not given and all recipientInfo are tried out. 2000 The old behaviour can be re-enabled in the CMS code by setting the 2001 CMS_DEBUG_DECRYPT flag. 2002 2003 *Bernd Edlinger* 2004 2005 * Early start up entropy quality from the DEVRANDOM seed source has been 2006 improved for older Linux systems. The RAND subsystem will wait for 2007 /dev/random to be producing output before seeding from /dev/urandom. 2008 The seeded state is stored for future library initialisations using 2009 a system global shared memory segment. The shared memory identifier 2010 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to 2011 the desired value. The default identifier is 114. 2012 2013 *Paul Dale* 2014 2015 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1 2016 when primes for RSA keys are computed. 2017 Since we previously always generated primes == 2 (mod 3) for RSA keys, 2018 the 2-prime and 3-prime RSA modules were easy to distinguish, since 2019 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting 2020 2-prime vs. 3-prime RSA keys was possible by computing N mod 3. 2021 This avoids possible fingerprinting of newly generated RSA modules. 2022 2023 *Bernd Edlinger* 2024 2025 * Correct the extended master secret constant on EBCDIC systems. Without this 2026 fix TLS connections between an EBCDIC system and a non-EBCDIC system that 2027 negotiate EMS will fail. Unfortunately this also means that TLS connections 2028 between EBCDIC systems with this fix, and EBCDIC systems without this 2029 fix will fail if they negotiate EMS. 2030 2031 *Matt Caswell* 2032 2033 * Changed the library initialisation so that the config file is now loaded 2034 by default. This was already the case for libssl. It now occurs for both 2035 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to 2036 `OPENSSL_init_crypto()` to suppress automatic loading of a config file. 2037 2038 *Matt Caswell* 2039 2040 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`, 2041 where the former acts as a replacement for `ERR_put_error()`, and the 2042 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`. 2043 `ERR_raise_data()` adds more flexibility by taking a format string and 2044 an arbitrary number of arguments following it, to be processed with 2045 `BIO_snprintf()`. 2046 2047 *Richard Levitte* 2048 2049 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used 2050 to check if a named provider is loaded and available. When called, it 2051 will also activate all fallback providers if such are still present. 2052 2053 *Richard Levitte* 2054 2055 * Enforce a minimum DH modulus size of 512 bits. 2056 2057 *Bernd Edlinger* 2058 2059 * Changed DH parameters to generate the order q subgroup instead of 2q. 2060 Previously generated DH parameters are still accepted by DH_check 2061 but DH_generate_key works around that by clearing bit 0 of the 2062 private key for those. This avoids leaking bit 0 of the private key. 2063 2064 *Bernd Edlinger* 2065 2066 * Significantly reduce secure memory usage by the randomness pools. 2067 2068 *Paul Dale* 2069 2070 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been 2071 deprecated. 2072 2073 *Rich Salz* 2074 2075 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange 2076 algorithms. An implementation of a key exchange algorithm can be obtained 2077 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be 2078 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to 2079 the older EVP_PKEY_derive_init() function. See the man pages for the new 2080 functions for further details. 2081 2082 *Matt Caswell* 2083 2084 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function. 2085 2086 *Matt Caswell* 2087 2088 * Removed the function names from error messages and deprecated the 2089 xxx_F_xxx define's. 2090 2091 *Richard Levitte* 2092 2093 * Removed NextStep support and the macro OPENSSL_UNISTD 2094 2095 *Rich Salz* 2096 2097 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL, 2098 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL. 2099 Also removed "export var as function" capability; we do not export 2100 variables, only functions. 2101 2102 *Rich Salz* 2103 2104 * RC5_32_set_key has been changed to return an int type, with 0 indicating 2105 an error and 1 indicating success. In previous versions of OpenSSL this 2106 was a void type. If a key was set longer than the maximum possible this 2107 would crash. 2108 2109 *Matt Caswell* 2110 2111 * Support SM2 signing and verification schemes with X509 certificate. 2112 2113 *Paul Yang* 2114 2115 * Use SHA256 as the default digest for TS query in the `ts` app. 2116 2117 *Tomáš Mráz* 2118 2119 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898. 2120 2121 *Shane Lontis* 2122 2123 * Default cipher lists/suites are now available via a function, the 2124 #defines are deprecated. 2125 2126 *Todd Short* 2127 2128 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and 2129 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries 2130 for Windows Store apps easier. Also, the "no-uplink" option has been added. 2131 2132 *Kenji Mouri* 2133 2134 * Join the directories crypto/x509 and crypto/x509v3 2135 2136 *Richard Levitte* 2137 2138 * Added command 'openssl kdf' that uses the EVP_KDF API. 2139 2140 *Shane Lontis* 2141 2142 * Added command 'openssl mac' that uses the EVP_MAC API. 2143 2144 *Shane Lontis* 2145 2146 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such 2147 as default directories. Also added the command 'openssl info' 2148 for scripting purposes. 2149 2150 *Richard Levitte* 2151 2152 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been 2153 deprecated. 2154 2155 *Matt Caswell* 2156 2157 * Add prediction resistance to the DRBG reseeding process. 2158 2159 *Paul Dale* 2160 2161 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as 2162 mandated by IEEE Std 1619-2018. 2163 2164 *Paul Dale* 2165 2166 * Added newline escaping functionality to a filename when using openssl dgst. 2167 This output format is to replicate the output format found in the `*sum` 2168 checksum programs. This aims to preserve backward compatibility. 2169 2170 *Matt Eaton, Richard Levitte, and Paul Dale* 2171 2172 * Removed the heartbeat message in DTLS feature, as it has very 2173 little usage and doesn't seem to fulfill a valuable purpose. 2174 The configuration option is now deprecated. 2175 2176 *Richard Levitte* 2177 2178 * Changed the output of 'openssl {digestname} < file' to display the 2179 digest name in its output. 2180 2181 *Richard Levitte* 2182 2183 * Added a new generic trace API which provides support for enabling 2184 instrumentation through trace output. 2185 2186 *Richard Levitte & Matthias St. Pierre* 2187 2188 * Added build tests for C++. These are generated files that only do one 2189 thing, to include one public OpenSSL head file each. This tests that 2190 the public header files can be usefully included in a C++ application. 2191 2192 This test isn't enabled by default. It can be enabled with the option 2193 'enable-buildtest-c++'. 2194 2195 *Richard Levitte* 2196 2197 * Added KB KDF (EVP_KDF_KB) to EVP_KDF. 2198 2199 *Robbie Harwood* 2200 2201 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF. 2202 2203 *Simo Sorce* 2204 2205 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF. 2206 2207 *Shane Lontis* 2208 2209 * Added KMAC to EVP_MAC. 2210 2211 *Shane Lontis* 2212 2213 * Added property based algorithm implementation selection framework to 2214 the core. 2215 2216 *Paul Dale* 2217 2218 * Added SCA hardening for modular field inversion in EC_GROUP through 2219 a new dedicated field_inv() pointer in EC_METHOD. 2220 This also addresses a leakage affecting conversions from projective 2221 to affine coordinates. 2222 2223 *Billy Bob Brumley, Nicola Tuveri* 2224 2225 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF 2226 implementations. This includes an EVP_PKEY to EVP_KDF bridge for 2227 those algorithms that were already supported through the EVP_PKEY API 2228 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2 2229 and scrypt are now wrappers that call EVP_KDF. 2230 2231 *David Makepeace* 2232 2233 * Build devcrypto engine as a dynamic engine. 2234 2235 *Eneas U de Queiroz* 2236 2237 * Add keyed BLAKE2 to EVP_MAC. 2238 2239 *Antoine Salon* 2240 2241 * Fix a bug in the computation of the endpoint-pair shared secret used 2242 by DTLS over SCTP. This breaks interoperability with older versions 2243 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime 2244 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling 2245 interoperability with such broken implementations. However, enabling 2246 this switch breaks interoperability with correct implementations. 2247 2248 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a 2249 re-used X509_PUBKEY object if the second PUBKEY is malformed. 2250 2251 *Bernd Edlinger* 2252 2253 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0(). 2254 2255 *Richard Levitte* 2256 2257 * Changed the license to the Apache License v2.0. 2258 2259 *Richard Levitte* 2260 2261 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH. 2262 2263 - Major releases (indicated by incrementing the MAJOR release number) 2264 may introduce incompatible API/ABI changes. 2265 - Minor releases (indicated by incrementing the MINOR release number) 2266 may introduce new features but retain API/ABI compatibility. 2267 - Patch releases (indicated by incrementing the PATCH number) 2268 are intended for bug fixes and other improvements of existing 2269 features only (like improving performance or adding documentation) 2270 and retain API/ABI compatibility. 2271 2272 *Richard Levitte* 2273 2274 * Add support for RFC5297 SIV mode (siv128), including AES-SIV. 2275 2276 *Todd Short* 2277 2278 * Remove the 'dist' target and add a tarball building script. The 2279 'dist' target has fallen out of use, and it shouldn't be 2280 necessary to configure just to create a source distribution. 2281 2282 *Richard Levitte* 2283 2284 * Recreate the OS390-Unix config target. It no longer relies on a 2285 special script like it did for OpenSSL pre-1.1.0. 2286 2287 *Richard Levitte* 2288 2289 * Instead of having the source directories listed in Configure, add 2290 a 'build.info' keyword SUBDIRS to indicate what sub-directories to 2291 look into. 2292 2293 *Richard Levitte* 2294 2295 * Add GMAC to EVP_MAC. 2296 2297 *Paul Dale* 2298 2299 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC. 2300 2301 *Richard Levitte* 2302 2303 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC 2304 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge, 2305 to facilitate the continued use of MACs through raw private keys in 2306 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`. 2307 2308 *Richard Levitte* 2309 2310 * Deprecate ECDH_KDF_X9_62(). 2311 2312 *Antoine Salon* 2313 2314 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for 2315 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names 2316 are retained for backwards compatibility. 2317 2318 *Antoine Salon* 2319 2320 * AES-XTS mode now enforces that its two keys are different to mitigate 2321 the attacked described in "Efficient Instantiations of Tweakable 2322 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway. 2323 Details of this attack can be obtained from: 2324 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf> 2325 2326 *Paul Dale* 2327 2328 * Rename the object files, i.e. give them other names than in previous 2329 versions. Their names now include the name of the final product, as 2330 well as its type mnemonic (bin, lib, shlib). 2331 2332 *Richard Levitte* 2333 2334 * Added new option for 'openssl list', '-objects', which will display the 2335 list of built in objects, i.e. OIDs with names. 2336 2337 *Richard Levitte* 2338 2339 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`, 2340 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to 2341 be set explicitly. 2342 2343 *Chris Novakovic* 2344 2345 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path 2346 improves application performance by removing data copies and providing 2347 applications with zero-copy system calls such as sendfile and splice. 2348 2349 *Boris Pismenny* 2350 2351 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. 2352 2353 *Martin Elshuber* 2354 2355 * `PKCS12_parse` now maintains the order of the parsed certificates 2356 when outputting them via `*ca` (rather than reversing it). 2357 2358 *David von Oheimb* 2359 2360 * Deprecated pthread fork support methods. 2361 2362 *Randall S. Becker* 2363 2364 * Added support for FFDHE key exchange in TLS 1.3. 2365 2366 *Raja Ashok* 2367 2368 * Added a new concept for OpenSSL plugability: providers. This 2369 functionality is designed to replace the ENGINE API and ENGINE 2370 implementations, and to be much more dynamic, allowing provider 2371 authors to introduce new algorithms among other things, as long as 2372 there's an API that supports the algorithm type. 2373 2374 With this concept comes a new core API for interaction between 2375 libcrypto and provider implementations. Public libcrypto functions 2376 that want to use providers do so through this core API. 2377 2378 The main documentation for this core API is found in 2379 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn 2380 refer to other manuals describing the API specific for supported 2381 algorithm types (also called operations). 2382 2383 *The OpenSSL team* 2384 2385OpenSSL 1.1.1 2386------------- 2387 2388### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx] 2389 2390 * Avoid loading of a dynamic engine twice. 2391 2392 *Bernd Edlinger* 2393 2394 * Prioritise DANE TLSA issuer certs over peer certs 2395 2396 *Viktor Dukhovni* 2397 2398 * Fixed random API for MacOS prior to 10.12 2399 2400 These MacOS versions don't support the CommonCrypto APIs 2401 2402 *Lenny Primak* 2403 2404### Changes between 1.1.1k and 1.1.1l [24 Aug 2021] 2405 2406 * Fixed an SM2 Decryption Buffer Overflow. 2407 2408 In order to decrypt SM2 encrypted data an application is expected to 2409 call the API function EVP_PKEY_decrypt(). Typically an application will 2410 call this function twice. The first time, on entry, the "out" parameter 2411 can be NULL and, on exit, the "outlen" parameter is populated with the 2412 buffer size required to hold the decrypted plaintext. The application 2413 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt() 2414 again, but this time passing a non-NULL value for the "out" parameter. 2415 2416 A bug in the implementation of the SM2 decryption code means that the 2417 calculation of the buffer size required to hold the plaintext returned 2418 by the first call to EVP_PKEY_decrypt() can be smaller than the actual 2419 size required by the second call. This can lead to a buffer overflow 2420 when EVP_PKEY_decrypt() is called by the application a second time with 2421 a buffer that is too small. 2422 2423 A malicious attacker who is able present SM2 content for decryption to 2424 an application could cause attacker chosen data to overflow the buffer 2425 by up to a maximum of 62 bytes altering the contents of other data held 2426 after the buffer, possibly changing application behaviour or causing 2427 the application to crash. The location of the buffer is application 2428 dependent but is typically heap allocated. 2429 ([CVE-2021-3711]) 2430 2431 *Matt Caswell* 2432 2433 * Fixed various read buffer overruns processing ASN.1 strings 2434 2435 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING 2436 structure which contains a buffer holding the string data and a field 2437 holding the buffer length. This contrasts with normal C strings which 2438 are repesented as a buffer for the string data which is terminated 2439 with a NUL (0) byte. 2440 2441 Although not a strict requirement, ASN.1 strings that are parsed using 2442 OpenSSL's own "d2i" functions (and other similar parsing functions) as 2443 well as any string whose value has been set with the ASN1_STRING_set() 2444 function will additionally NUL terminate the byte array in the 2445 ASN1_STRING structure. 2446 2447 However, it is possible for applications to directly construct valid 2448 ASN1_STRING structures which do not NUL terminate the byte array by 2449 directly setting the "data" and "length" fields in the ASN1_STRING 2450 array. This can also happen by using the ASN1_STRING_set0() function. 2451 2452 Numerous OpenSSL functions that print ASN.1 data have been found to 2453 assume that the ASN1_STRING byte array will be NUL terminated, even 2454 though this is not guaranteed for strings that have been directly 2455 constructed. Where an application requests an ASN.1 structure to be 2456 printed, and where that ASN.1 structure contains ASN1_STRINGs that have 2457 been directly constructed by the application without NUL terminating 2458 the "data" field, then a read buffer overrun can occur. 2459 2460 The same thing can also occur during name constraints processing 2461 of certificates (for example if a certificate has been directly 2462 constructed by the application instead of loading it via the OpenSSL 2463 parsing functions, and the certificate contains non NUL terminated 2464 ASN1_STRING structures). It can also occur in the X509_get1_email(), 2465 X509_REQ_get1_email() and X509_get1_ocsp() functions. 2466 2467 If a malicious actor can cause an application to directly construct an 2468 ASN1_STRING and then process it through one of the affected OpenSSL 2469 functions then this issue could be hit. This might result in a crash 2470 (causing a Denial of Service attack). It could also result in the 2471 disclosure of private memory contents (such as private keys, or 2472 sensitive plaintext). 2473 ([CVE-2021-3712]) 2474 2475 *Matt Caswell* 2476 2477### Changes between 1.1.1j and 1.1.1k [25 Mar 2021] 2478 2479 * Fixed a problem with verifying a certificate chain when using the 2480 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of 2481 the certificates present in a certificate chain. It is not set by default. 2482 2483 Starting from OpenSSL version 1.1.1h a check to disallow certificates in 2484 the chain that have explicitly encoded elliptic curve parameters was added 2485 as an additional strict check. 2486 2487 An error in the implementation of this check meant that the result of a 2488 previous check to confirm that certificates in the chain are valid CA 2489 certificates was overwritten. This effectively bypasses the check 2490 that non-CA certificates must not be able to issue other certificates. 2491 2492 If a "purpose" has been configured then there is a subsequent opportunity 2493 for checks that the certificate is a valid CA. All of the named "purpose" 2494 values implemented in libcrypto perform this check. Therefore, where 2495 a purpose is set the certificate chain will still be rejected even when the 2496 strict flag has been used. A purpose is set by default in libssl client and 2497 server certificate verification routines, but it can be overridden or 2498 removed by an application. 2499 2500 In order to be affected, an application must explicitly set the 2501 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose 2502 for the certificate verification or, in the case of TLS client or server 2503 applications, override the default purpose. 2504 ([CVE-2021-3450]) 2505 2506 *Tomáš Mráz* 2507 2508 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously 2509 crafted renegotiation ClientHello message from a client. If a TLSv1.2 2510 renegotiation ClientHello omits the signature_algorithms extension (where it 2511 was present in the initial ClientHello), but includes a 2512 signature_algorithms_cert extension then a NULL pointer dereference will 2513 result, leading to a crash and a denial of service attack. 2514 2515 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled 2516 (which is the default configuration). OpenSSL TLS clients are not impacted by 2517 this issue. 2518 ([CVE-2021-3449]) 2519 2520 *Peter Kästle and Samuel Sapalski* 2521 2522### Changes between 1.1.1i and 1.1.1j [16 Feb 2021] 2523 2524 * Fixed the X509_issuer_and_serial_hash() function. It attempts to 2525 create a unique hash value based on the issuer and serial number data 2526 contained within an X509 certificate. However it was failing to correctly 2527 handle any errors that may occur while parsing the issuer field (which might 2528 occur if the issuer field is maliciously constructed). This may subsequently 2529 result in a NULL pointer deref and a crash leading to a potential denial of 2530 service attack. 2531 ([CVE-2021-23841]) 2532 2533 *Matt Caswell* 2534 2535 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING 2536 padding mode to correctly check for rollback attacks. This is considered a 2537 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is 2538 CVE-2021-23839. 2539 2540 *Matt Caswell* 2541 2542 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate 2543 functions. Previously they could overflow the output length argument in some 2544 cases where the input length is close to the maximum permissable length for 2545 an integer on the platform. In such cases the return value from the function 2546 call would be 1 (indicating success), but the output length value would be 2547 negative. This could cause applications to behave incorrectly or crash. 2548 ([CVE-2021-23840]) 2549 2550 *Matt Caswell* 2551 2552 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous 2553 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This 2554 could be exploited in a side channel attack to recover the password. Since 2555 the attack is local host only this is outside of the current OpenSSL 2556 threat model and therefore no CVE is assigned. 2557 2558 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this 2559 issue. 2560 2561 *Matt Caswell* 2562 2563### Changes between 1.1.1h and 1.1.1i [8 Dec 2020] 2564 2565 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function 2566 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME. 2567 If an attacker can control both items being compared then this could lead 2568 to a possible denial of service attack. OpenSSL itself uses the 2569 GENERAL_NAME_cmp function for two purposes: 2570 1) Comparing CRL distribution point names between an available CRL and a 2571 CRL distribution point embedded in an X509 certificate 2572 2) When verifying that a timestamp response token signer matches the 2573 timestamp authority name (exposed via the API functions 2574 TS_RESP_verify_response and TS_RESP_verify_token) 2575 ([CVE-2020-1971]) 2576 2577 *Matt Caswell* 2578 2579### Changes between 1.1.1g and 1.1.1h [22 Sep 2020] 2580 2581 * Certificates with explicit curve parameters are now disallowed in 2582 verification chains if the X509_V_FLAG_X509_STRICT flag is used. 2583 2584 *Tomáš Mráz* 2585 2586 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently 2587 ignore TLS protocol version bounds when configuring DTLS-based contexts, and 2588 conversely, silently ignore DTLS protocol version bounds when configuring 2589 TLS-based contexts. The commands can be repeated to set bounds of both 2590 types. The same applies with the corresponding "min_protocol" and 2591 "max_protocol" command-line switches, in case some application uses both TLS 2592 and DTLS. 2593 2594 SSL_CTX instances that are created for a fixed protocol version (e.g. 2595 TLSv1_server_method()) also silently ignore version bounds. Previously 2596 attempts to apply bounds to these protocol versions would result in an 2597 error. Now only the "version-flexible" SSL_CTX instances are subject to 2598 limits in configuration files in command-line options. 2599 2600 *Viktor Dukhovni* 2601 2602 * Handshake now fails if Extended Master Secret extension is dropped 2603 on renegotiation. 2604 2605 *Tomáš Mráz* 2606 2607 * The Oracle Developer Studio compiler will start reporting deprecated APIs 2608 2609### Changes between 1.1.1f and 1.1.1g [21 Apr 2020] 2610 2611 * Fixed segmentation fault in SSL_check_chain() 2612 Server or client applications that call the SSL_check_chain() function 2613 during or after a TLS 1.3 handshake may crash due to a NULL pointer 2614 dereference as a result of incorrect handling of the 2615 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid 2616 or unrecognised signature algorithm is received from the peer. This could 2617 be exploited by a malicious peer in a Denial of Service attack. 2618 ([CVE-2020-1967]) 2619 2620 *Benjamin Kaduk* 2621 2622 * Added AES consttime code for no-asm configurations 2623 an optional constant time support for AES was added 2624 when building openssl for no-asm. 2625 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME 2626 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME 2627 At this time this feature is by default disabled. 2628 It will be enabled by default in 3.0. 2629 2630 *Bernd Edlinger* 2631 2632### Changes between 1.1.1e and 1.1.1f [31 Mar 2020] 2633 2634 * Revert the change of EOF detection while reading in libssl to avoid 2635 regressions in applications depending on the current way of reporting 2636 the EOF. As the existing method is not fully accurate the change to 2637 reporting the EOF via SSL_ERROR_SSL is kept on the current development 2638 branch and will be present in the 3.0 release. 2639 2640 *Tomáš Mráz* 2641 2642 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1 2643 when primes for RSA keys are computed. 2644 Since we previously always generated primes == 2 (mod 3) for RSA keys, 2645 the 2-prime and 3-prime RSA modules were easy to distinguish, since 2646 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting 2647 2-prime vs. 3-prime RSA keys was possible by computing N mod 3. 2648 This avoids possible fingerprinting of newly generated RSA modules. 2649 2650 *Bernd Edlinger* 2651 2652### Changes between 1.1.1d and 1.1.1e [17 Mar 2020] 2653 2654 * Properly detect EOF while reading in libssl. Previously if we hit an EOF 2655 while reading in libssl then we would report an error back to the 2656 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add 2657 an error to the stack (which means we instead return SSL_ERROR_SSL) and 2658 therefore give a hint as to what went wrong. 2659 2660 *Matt Caswell* 2661 2662 * Check that ed25519 and ed448 are allowed by the security level. Previously 2663 signature algorithms not using an MD were not being checked that they were 2664 allowed by the security level. 2665 2666 *Kurt Roeckx* 2667 2668 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername() 2669 was not quite right. The behaviour was not consistent between resumption 2670 and normal handshakes, and also not quite consistent with historical 2671 behaviour. The behaviour in various scenarios has been clarified and 2672 it has been updated to make it match historical behaviour as closely as 2673 possible. 2674 2675 *Matt Caswell* 2676 2677 * *[VMS only]* The header files that the VMS compilers include automatically, 2678 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas 2679 that the C++ compiler doesn't understand. This is a shortcoming in the 2680 compiler, but can be worked around with `__cplusplus` guards. 2681 2682 C++ applications that use OpenSSL libraries must be compiled using the 2683 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL 2684 functions. Otherwise, only functions with symbols of less than 31 2685 characters can be used, as the linker will not be able to successfully 2686 resolve symbols with longer names. 2687 2688 *Richard Levitte* 2689 2690 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY. 2691 The presence of this system service is determined at run-time. 2692 2693 *Richard Levitte* 2694 2695 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just 2696 the first value. 2697 2698 *Jon Spillett* 2699 2700### Changes between 1.1.1c and 1.1.1d [10 Sep 2019] 2701 2702 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random 2703 number generator (RNG). This was intended to include protection in the 2704 event of a fork() system call in order to ensure that the parent and child 2705 processes did not share the same RNG state. However this protection was not 2706 being used in the default case. 2707 2708 A partial mitigation for this issue is that the output from a high 2709 precision timer is mixed into the RNG state so the likelihood of a parent 2710 and child process sharing state is significantly reduced. 2711 2712 If an application already calls OPENSSL_init_crypto() explicitly using 2713 OPENSSL_INIT_ATFORK then this problem does not occur at all. 2714 ([CVE-2019-1549]) 2715 2716 *Matthias St. Pierre* 2717 2718 * For built-in EC curves, ensure an EC_GROUP built from the curve name is 2719 used even when parsing explicit parameters, when loading a encoded key 2720 or calling `EC_GROUP_new_from_ecpkparameters()`/ 2721 `EC_GROUP_new_from_ecparameters()`. 2722 This prevents bypass of security hardening and performance gains, 2723 especially for curves with specialized EC_METHODs. 2724 By default, if a key encoded with explicit parameters is loaded and later 2725 encoded, the output is still encoded with explicit parameters, even if 2726 internally a "named" EC_GROUP is used for computation. 2727 2728 *Nicola Tuveri* 2729 2730 * Compute ECC cofactors if not provided during EC_GROUP construction. Before 2731 this change, EC_GROUP_set_generator would accept order and/or cofactor as 2732 NULL. After this change, only the cofactor parameter can be NULL. It also 2733 does some minimal sanity checks on the passed order. 2734 ([CVE-2019-1547]) 2735 2736 *Billy Bob Brumley* 2737 2738 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey. 2739 An attack is simple, if the first CMS_recipientInfo is valid but the 2740 second CMS_recipientInfo is chosen ciphertext. If the second 2741 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct 2742 encryption key will be replaced by garbage, and the message cannot be 2743 decoded, but if the RSA decryption fails, the correct encryption key is 2744 used and the recipient will not notice the attack. 2745 As a work around for this potential attack the length of the decrypted 2746 key must be equal to the cipher default key length, in case the 2747 certifiate is not given and all recipientInfo are tried out. 2748 The old behaviour can be re-enabled in the CMS code by setting the 2749 CMS_DEBUG_DECRYPT flag. 2750 ([CVE-2019-1563]) 2751 2752 *Bernd Edlinger* 2753 2754 * Early start up entropy quality from the DEVRANDOM seed source has been 2755 improved for older Linux systems. The RAND subsystem will wait for 2756 /dev/random to be producing output before seeding from /dev/urandom. 2757 The seeded state is stored for future library initialisations using 2758 a system global shared memory segment. The shared memory identifier 2759 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to 2760 the desired value. The default identifier is 114. 2761 2762 *Paul Dale* 2763 2764 * Correct the extended master secret constant on EBCDIC systems. Without this 2765 fix TLS connections between an EBCDIC system and a non-EBCDIC system that 2766 negotiate EMS will fail. Unfortunately this also means that TLS connections 2767 between EBCDIC systems with this fix, and EBCDIC systems without this 2768 fix will fail if they negotiate EMS. 2769 2770 *Matt Caswell* 2771 2772 * Use Windows installation paths in the mingw builds 2773 2774 Mingw isn't a POSIX environment per se, which means that Windows 2775 paths should be used for installation. 2776 ([CVE-2019-1552]) 2777 2778 *Richard Levitte* 2779 2780 * Changed DH_check to accept parameters with order q and 2q subgroups. 2781 With order 2q subgroups the bit 0 of the private key is not secret 2782 but DH_generate_key works around that by clearing bit 0 of the 2783 private key for those. This avoids leaking bit 0 of the private key. 2784 2785 *Bernd Edlinger* 2786 2787 * Significantly reduce secure memory usage by the randomness pools. 2788 2789 *Paul Dale* 2790 2791 * Revert the DEVRANDOM_WAIT feature for Linux systems 2792 2793 The DEVRANDOM_WAIT feature added a select() call to wait for the 2794 /dev/random device to become readable before reading from the 2795 /dev/urandom device. 2796 2797 It turned out that this change had negative side effects on 2798 performance which were not acceptable. After some discussion it 2799 was decided to revert this feature and leave it up to the OS 2800 resp. the platform maintainer to ensure a proper initialization 2801 during early boot time. 2802 2803 *Matthias St. Pierre* 2804 2805### Changes between 1.1.1b and 1.1.1c [28 May 2019] 2806 2807 * Add build tests for C++. These are generated files that only do one 2808 thing, to include one public OpenSSL head file each. This tests that 2809 the public header files can be usefully included in a C++ application. 2810 2811 This test isn't enabled by default. It can be enabled with the option 2812 'enable-buildtest-c++'. 2813 2814 *Richard Levitte* 2815 2816 * Enable SHA3 pre-hashing for ECDSA and DSA. 2817 2818 *Patrick Steuer* 2819 2820 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024. 2821 This changes the size when using the `genpkey` command when no size is given. 2822 It fixes an omission in earlier changes that changed all RSA, DSA and DH 2823 generation commands to use 2048 bits by default. 2824 2825 *Kurt Roeckx* 2826 2827 * Reorganize the manual pages to consistently have RETURN VALUES, 2828 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust 2829 util/fix-doc-nits accordingly. 2830 2831 *Paul Yang, Joshua Lock* 2832 2833 * Add the missing accessor EVP_PKEY_get0_engine() 2834 2835 *Matt Caswell* 2836 2837 * Have commands like `s_client` and `s_server` output the signature scheme 2838 along with other cipher suite parameters when debugging. 2839 2840 *Lorinczy Zsigmond* 2841 2842 * Make OPENSSL_config() error agnostic again. 2843 2844 *Richard Levitte* 2845 2846 * Do the error handling in RSA decryption constant time. 2847 2848 *Bernd Edlinger* 2849 2850 * Prevent over long nonces in ChaCha20-Poly1305. 2851 2852 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input 2853 for every encryption operation. RFC 7539 specifies that the nonce value 2854 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length 2855 and front pads the nonce with 0 bytes if it is less than 12 2856 bytes. However it also incorrectly allows a nonce to be set of up to 16 2857 bytes. In this case only the last 12 bytes are significant and any 2858 additional leading bytes are ignored. 2859 2860 It is a requirement of using this cipher that nonce values are 2861 unique. Messages encrypted using a reused nonce value are susceptible to 2862 serious confidentiality and integrity attacks. If an application changes 2863 the default nonce length to be longer than 12 bytes and then makes a 2864 change to the leading bytes of the nonce expecting the new value to be a 2865 new unique nonce then such an application could inadvertently encrypt 2866 messages with a reused nonce. 2867 2868 Additionally the ignored bytes in a long nonce are not covered by the 2869 integrity guarantee of this cipher. Any application that relies on the 2870 integrity of these ignored leading bytes of a long nonce may be further 2871 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS, 2872 is safe because no such use sets such a long nonce value. However user 2873 applications that use this cipher directly and set a non-default nonce 2874 length to be longer than 12 bytes may be vulnerable. 2875 2876 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk 2877 Greef of Ronomon. 2878 ([CVE-2019-1543]) 2879 2880 *Matt Caswell* 2881 2882 * Add DEVRANDOM_WAIT feature for Linux systems 2883 2884 On older Linux systems where the getrandom() system call is not available, 2885 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG. 2886 Contrary to getrandom(), the /dev/urandom device will not block during 2887 early boot when the kernel CSPRNG has not been seeded yet. 2888 2889 To mitigate this known weakness, use select() to wait for /dev/random to 2890 become readable before reading from /dev/urandom. 2891 2892 * Ensure that SM2 only uses SM3 as digest algorithm 2893 2894 *Paul Yang* 2895 2896### Changes between 1.1.1a and 1.1.1b [26 Feb 2019] 2897 2898 * Change the info callback signals for the start and end of a post-handshake 2899 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START 2900 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get 2901 confused by this and assume that a TLSv1.2 renegotiation has started. This 2902 can break KeyUpdate handling. Instead we no longer signal the start and end 2903 of a post handshake message exchange (although the messages themselves are 2904 still signalled). This could break some applications that were expecting 2905 the old signals. However without this KeyUpdate is not usable for many 2906 applications. 2907 2908 *Matt Caswell* 2909 2910### Changes between 1.1.1 and 1.1.1a [20 Nov 2018] 2911 2912 * Timing vulnerability in DSA signature generation 2913 2914 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a 2915 timing side channel attack. An attacker could use variations in the signing 2916 algorithm to recover the private key. 2917 2918 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser. 2919 ([CVE-2018-0734]) 2920 2921 *Paul Dale* 2922 2923 * Timing vulnerability in ECDSA signature generation 2924 2925 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a 2926 timing side channel attack. An attacker could use variations in the signing 2927 algorithm to recover the private key. 2928 2929 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser. 2930 ([CVE-2018-0735]) 2931 2932 *Paul Dale* 2933 2934 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input 2935 if its length exceeds 4096 bytes. The limit has been raised to a buffer size 2936 of two gigabytes and the error handling improved. 2937 2938 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been 2939 categorized as a normal bug, not a security issue, because the DRBG reseeds 2940 automatically and is fully functional even without additional randomness 2941 provided by the application. 2942 2943### Changes between 1.1.0i and 1.1.1 [11 Sep 2018] 2944 2945 * Add a new ClientHello callback. Provides a callback interface that gives 2946 the application the ability to adjust the nascent SSL object at the 2947 earliest stage of ClientHello processing, immediately after extensions have 2948 been collected but before they have been processed. In particular, this 2949 callback can adjust the supported TLS versions in response to the contents 2950 of the ClientHello 2951 2952 *Benjamin Kaduk* 2953 2954 * Add SM2 base algorithm support. 2955 2956 *Jack Lloyd* 2957 2958 * s390x assembly pack: add (improved) hardware-support for the following 2959 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb, 2960 aes-cfb/cfb8, aes-ecb. 2961 2962 *Patrick Steuer* 2963 2964 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str 2965 parameter is no longer accepted, as it leads to a corrupt table. NULL 2966 pem_str is reserved for alias entries only. 2967 2968 *Richard Levitte* 2969 2970 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder 2971 step for prime curves. The new implementation is based on formulae from 2972 differential addition-and-doubling in homogeneous projective coordinates 2973 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant 2974 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves 2975 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified 2976 to work in projective coordinates. 2977 2978 *Billy Bob Brumley, Nicola Tuveri* 2979 2980 * Change generating and checking of primes so that the error rate of not 2981 being prime depends on the intended use based on the size of the input. 2982 For larger primes this will result in more rounds of Miller-Rabin. 2983 The maximal error rate for primes with more than 1080 bits is lowered 2984 to 2^-128. 2985 2986 *Kurt Roeckx, Annie Yousar* 2987 2988 * Increase the number of Miller-Rabin rounds for DSA key generating to 64. 2989 2990 *Kurt Roeckx* 2991 2992 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when 2993 moving between systems, and to avoid confusion when a Windows build is 2994 done with mingw vs with MSVC. For POSIX installs, there's still a 2995 symlink or copy named 'tsget' to avoid that confusion as well. 2996 2997 *Richard Levitte* 2998 2999 * Revert blinding in ECDSA sign and instead make problematic addition 3000 length-invariant. Switch even to fixed-length Montgomery multiplication. 3001 3002 *Andy Polyakov* 3003 3004 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder 3005 step for binary curves. The new implementation is based on formulae from 3006 differential addition-and-doubling in mixed Lopez-Dahab projective 3007 coordinates, modified to independently blind the operands. 3008 3009 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri* 3010 3011 * Add a scaffold to optionally enhance the Montgomery ladder implementation 3012 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing 3013 EC_METHODs to implement their own specialized "ladder step", to take 3014 advantage of more favorable coordinate systems or more efficient 3015 differential addition-and-doubling algorithms. 3016 3017 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri* 3018 3019 * Modified the random device based seed sources to keep the relevant 3020 file descriptors open rather than reopening them on each access. 3021 This allows such sources to operate in a chroot() jail without 3022 the associated device nodes being available. This behaviour can be 3023 controlled using RAND_keep_random_devices_open(). 3024 3025 *Paul Dale* 3026 3027 * Numerous side-channel attack mitigations have been applied. This may have 3028 performance impacts for some algorithms for the benefit of improved 3029 security. Specific changes are noted in this change log by their respective 3030 authors. 3031 3032 *Matt Caswell* 3033 3034 * AIX shared library support overhaul. Switch to AIX "natural" way of 3035 handling shared libraries, which means collecting shared objects of 3036 different versions and bitnesses in one common archive. This allows to 3037 mitigate conflict between 1.0 and 1.1 side-by-side installations. It 3038 doesn't affect the way 3rd party applications are linked, only how 3039 multi-version installation is managed. 3040 3041 *Andy Polyakov* 3042 3043 * Make ec_group_do_inverse_ord() more robust and available to other 3044 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA 3045 mitigations are applied to the fallback BN_mod_inverse(). 3046 When using this function rather than BN_mod_inverse() directly, new 3047 EC cryptosystem implementations are then safer-by-default. 3048 3049 *Billy Bob Brumley* 3050 3051 * Add coordinate blinding for EC_POINT and implement projective 3052 coordinate blinding for generic prime curves as a countermeasure to 3053 chosen point SCA attacks. 3054 3055 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley* 3056 3057 * Add blinding to ECDSA and DSA signatures to protect against side channel 3058 attacks discovered by Keegan Ryan (NCC Group). 3059 3060 *Matt Caswell* 3061 3062 * Enforce checking in the `pkeyutl` command to ensure that the input 3063 length does not exceed the maximum supported digest length when performing 3064 a sign, verify or verifyrecover operation. 3065 3066 *Matt Caswell* 3067 3068 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking 3069 I/O in combination with something like select() or poll() will hang. This 3070 can be turned off again using SSL_CTX_clear_mode(). 3071 Many applications do not properly handle non-application data records, and 3072 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works 3073 around the problems in those applications, but can also break some. 3074 It's recommended to read the manpages about SSL_read(), SSL_write(), 3075 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and 3076 SSL_CTX_set_read_ahead() again. 3077 3078 *Kurt Roeckx* 3079 3080 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we 3081 now allow empty (zero character) pass phrases. 3082 3083 *Richard Levitte* 3084 3085 * Apply blinding to binary field modular inversion and remove patent 3086 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation. 3087 3088 *Billy Bob Brumley* 3089 3090 * Deprecate ec2_mult.c and unify scalar multiplication code paths for 3091 binary and prime elliptic curves. 3092 3093 *Billy Bob Brumley* 3094 3095 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for 3096 constant time fixed point multiplication. 3097 3098 *Billy Bob Brumley* 3099 3100 * Revise elliptic curve scalar multiplication with timing attack 3101 defenses: ec_wNAF_mul redirects to a constant time implementation 3102 when computing fixed point and variable point multiplication (which 3103 in OpenSSL are mostly used with secret scalars in keygen, sign, 3104 ECDH derive operations). 3105 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García, 3106 Sohaib ul Hassan* 3107 3108 * Updated CONTRIBUTING 3109 3110 *Rich Salz* 3111 3112 * Updated DRBG / RAND to request nonce and additional low entropy 3113 randomness from the system. 3114 3115 *Matthias St. Pierre* 3116 3117 * Updated 'openssl rehash' to use OpenSSL consistent default. 3118 3119 *Richard Levitte* 3120 3121 * Moved the load of the ssl_conf module to libcrypto, which helps 3122 loading engines that libssl uses before libssl is initialised. 3123 3124 *Matt Caswell* 3125 3126 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA 3127 3128 *Matt Caswell* 3129 3130 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases. 3131 3132 *Ingo Schwarze, Rich Salz* 3133 3134 * Added output of accepting IP address and port for 'openssl s_server' 3135 3136 *Richard Levitte* 3137 3138 * Added a new API for TLSv1.3 ciphersuites: 3139 SSL_CTX_set_ciphersuites() 3140 SSL_set_ciphersuites() 3141 3142 *Matt Caswell* 3143 3144 * Memory allocation failures consistently add an error to the error 3145 stack. 3146 3147 *Rich Salz* 3148 3149 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values 3150 in libcrypto when run as setuid/setgid. 3151 3152 *Bernd Edlinger* 3153 3154 * Load any config file by default when libssl is used. 3155 3156 *Matt Caswell* 3157 3158 * Added new public header file <openssl/rand_drbg.h> and documentation 3159 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview. 3160 3161 *Matthias St. Pierre* 3162 3163 * QNX support removed (cannot find contributors to get their approval 3164 for the license change). 3165 3166 *Rich Salz* 3167 3168 * TLSv1.3 replay protection for early data has been implemented. See the 3169 SSL_read_early_data() man page for further details. 3170 3171 *Matt Caswell* 3172 3173 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite 3174 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and 3175 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3. 3176 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration 3177 would otherwise inadvertently disable all TLSv1.3 ciphersuites the 3178 configuration has been separated out. See the ciphers man page or the 3179 SSL_CTX_set_ciphersuites() man page for more information. 3180 3181 *Matt Caswell* 3182 3183 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running 3184 in responder mode now supports the new "-multi" option, which 3185 spawns the specified number of child processes to handle OCSP 3186 requests. The "-timeout" option now also limits the OCSP 3187 responder's patience to wait to receive the full client request 3188 on a newly accepted connection. Child processes are respawned 3189 as needed, and the CA index file is automatically reloaded 3190 when changed. This makes it possible to run the "ocsp" responder 3191 as a long-running service, making the OpenSSL CA somewhat more 3192 feature-complete. In this mode, most diagnostic messages logged 3193 after entering the event loop are logged via syslog(3) rather than 3194 written to stderr. 3195 3196 *Viktor Dukhovni* 3197 3198 * Added support for X448 and Ed448. Heavily based on original work by 3199 Mike Hamburg. 3200 3201 *Matt Caswell* 3202 3203 * Extend OSSL_STORE with capabilities to search and to narrow the set of 3204 objects loaded. This adds the functions OSSL_STORE_expect() and 3205 OSSL_STORE_find() as well as needed tools to construct searches and 3206 get the search data out of them. 3207 3208 *Richard Levitte* 3209 3210 * Support for TLSv1.3 added. Note that users upgrading from an earlier 3211 version of OpenSSL should review their configuration settings to ensure 3212 that they are still appropriate for TLSv1.3. For further information see: 3213 <https://wiki.openssl.org/index.php/TLS1.3> 3214 3215 *Matt Caswell* 3216 3217 * Grand redesign of the OpenSSL random generator 3218 3219 The default RAND method now utilizes an AES-CTR DRBG according to 3220 NIST standard SP 800-90Ar1. The new random generator is essentially 3221 a port of the default random generator from the OpenSSL FIPS 2.0 3222 object module. It is a hybrid deterministic random bit generator 3223 using an AES-CTR bit stream and which seeds and reseeds itself 3224 automatically using trusted system entropy sources. 3225 3226 Some of its new features are: 3227 - Support for multiple DRBG instances with seed chaining. 3228 - The default RAND method makes use of a DRBG. 3229 - There is a public and private DRBG instance. 3230 - The DRBG instances are fork-safe. 3231 - Keep all global DRBG instances on the secure heap if it is enabled. 3232 - The public and private DRBG instance are per thread for lock free 3233 operation 3234 3235 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre* 3236 3237 * Changed Configure so it only says what it does and doesn't dump 3238 so much data. Instead, ./configdata.pm should be used as a script 3239 to display all sorts of configuration data. 3240 3241 *Richard Levitte* 3242 3243 * Added processing of "make variables" to Configure. 3244 3245 *Richard Levitte* 3246 3247 * Added SHA512/224 and SHA512/256 algorithm support. 3248 3249 *Paul Dale* 3250 3251 * The last traces of Netware support, first removed in 1.1.0, have 3252 now been removed. 3253 3254 *Rich Salz* 3255 3256 * Get rid of Makefile.shared, and in the process, make the processing 3257 of certain files (rc.obj, or the .def/.map/.opt files produced from 3258 the ordinal files) more visible and hopefully easier to trace and 3259 debug (or make silent). 3260 3261 *Richard Levitte* 3262 3263 * Make it possible to have environment variable assignments as 3264 arguments to config / Configure. 3265 3266 *Richard Levitte* 3267 3268 * Add multi-prime RSA (RFC 8017) support. 3269 3270 *Paul Yang* 3271 3272 * Add SM3 implemented according to GB/T 32905-2016 3273 *Jack Lloyd <jack.lloyd@ribose.com>,* 3274 *Ronald Tse <ronald.tse@ribose.com>,* 3275 *Erick Borsboom <erick.borsboom@ribose.com>* 3276 3277 * Add 'Maximum Fragment Length' TLS extension negotiation and support 3278 as documented in RFC6066. 3279 Based on a patch from Tomasz Moń 3280 3281 *Filipe Raimundo da Silva* 3282 3283 * Add SM4 implemented according to GB/T 32907-2016. 3284 *Jack Lloyd <jack.lloyd@ribose.com>,* 3285 *Ronald Tse <ronald.tse@ribose.com>,* 3286 *Erick Borsboom <erick.borsboom@ribose.com>* 3287 3288 * Reimplement -newreq-nodes and ERR_error_string_n; the 3289 original author does not agree with the license change. 3290 3291 *Rich Salz* 3292 3293 * Add ARIA AEAD TLS support. 3294 3295 *Jon Spillett* 3296 3297 * Some macro definitions to support VS6 have been removed. Visual 3298 Studio 6 has not worked since 1.1.0 3299 3300 *Rich Salz* 3301 3302 * Add ERR_clear_last_mark(), to allow callers to clear the last mark 3303 without clearing the errors. 3304 3305 *Richard Levitte* 3306 3307 * Add "atfork" functions. If building on a system that without 3308 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application 3309 requirements. The RAND facility now uses/requires this. 3310 3311 *Rich Salz* 3312 3313 * Add SHA3. 3314 3315 *Andy Polyakov* 3316 3317 * The UI API becomes a permanent and integral part of libcrypto, i.e. 3318 not possible to disable entirely. However, it's still possible to 3319 disable the console reading UI method, UI_OpenSSL() (use UI_null() 3320 as a fallback). 3321 3322 To disable, configure with 'no-ui-console'. 'no-ui' is still 3323 possible to use as an alias. Check at compile time with the 3324 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still 3325 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE. 3326 3327 *Richard Levitte* 3328 3329 * Add a STORE module, which implements a uniform and URI based reader of 3330 stores that can contain keys, certificates, CRLs and numerous other 3331 objects. The main API is loosely based on a few stdio functions, 3332 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof, 3333 OSSL_STORE_error and OSSL_STORE_close. 3334 The implementation uses backends called "loaders" to implement arbitrary 3335 URI schemes. There is one built in "loader" for the 'file' scheme. 3336 3337 *Richard Levitte* 3338 3339 * Add devcrypto engine. This has been implemented against cryptodev-linux, 3340 then adjusted to work on FreeBSD 8.4 as well. 3341 Enable by configuring with 'enable-devcryptoeng'. This is done by default 3342 on BSD implementations, as cryptodev.h is assumed to exist on all of them. 3343 3344 *Richard Levitte* 3345 3346 * Module names can prefixed with OSSL_ or OPENSSL_. This affects 3347 util/mkerr.pl, which is adapted to allow those prefixes, leading to 3348 error code calls like this: 3349 3350 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER); 3351 3352 With this change, we claim the namespaces OSSL and OPENSSL in a manner 3353 that can be encoded in C. For the foreseeable future, this will only 3354 affect new modules. 3355 3356 *Richard Levitte and Tim Hudson* 3357 3358 * Removed BSD cryptodev engine. 3359 3360 *Rich Salz* 3361 3362 * Add a build target 'build_all_generated', to build all generated files 3363 and only that. This can be used to prepare everything that requires 3364 things like perl for a system that lacks perl and then move everything 3365 to that system and do the rest of the build there. 3366 3367 *Richard Levitte* 3368 3369 * In the UI interface, make it possible to duplicate the user data. This 3370 can be used by engines that need to retain the data for a longer time 3371 than just the call where this user data is passed. 3372 3373 *Richard Levitte* 3374 3375 * Ignore the '-named_curve auto' value for compatibility of applications 3376 with OpenSSL 1.0.2. 3377 3378 *Tomáš Mráz <tmraz@fedoraproject.org>* 3379 3380 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2 3381 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such 3382 alerts across multiple records (some of which could be empty). In practice 3383 it make no sense to send an empty alert record, or to fragment one. TLSv1.3 3384 prohibits this altogether and other libraries (BoringSSL, NSS) do not 3385 support this at all. Supporting it adds significant complexity to the 3386 record layer, and its removal is unlikely to cause interoperability 3387 issues. 3388 3389 *Matt Caswell* 3390 3391 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed 3392 with Z. These are meant to replace LONG and ZLONG and to be size safe. 3393 The use of LONG and ZLONG is discouraged and scheduled for deprecation 3394 in OpenSSL 1.2.0. 3395 3396 *Richard Levitte* 3397 3398 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string, 3399 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t. 3400 3401 *Richard Levitte, Andy Polyakov* 3402 3403 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine() 3404 does for RSA, etc. 3405 3406 *Richard Levitte* 3407 3408 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target 3409 platform rather than 'mingw'. 3410 3411 *Richard Levitte* 3412 3413 * The functions X509_STORE_add_cert and X509_STORE_add_crl return 3414 success if they are asked to add an object which already exists 3415 in the store. This change cascades to other functions which load 3416 certificates and CRLs. 3417 3418 *Paul Dale* 3419 3420 * x86_64 assembly pack: annotate code with DWARF CFI directives to 3421 facilitate stack unwinding even from assembly subroutines. 3422 3423 *Andy Polyakov* 3424 3425 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN. 3426 Also remove OPENSSL_GLOBAL entirely, as it became a no-op. 3427 3428 *Richard Levitte* 3429 3430 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c. 3431 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1, 3432 which is the minimum version we support. 3433 3434 *Richard Levitte* 3435 3436 * Certificate time validation (X509_cmp_time) enforces stricter 3437 compliance with RFC 5280. Fractional seconds and timezone offsets 3438 are no longer allowed. 3439 3440 *Emilia Käsper* 3441 3442 * Add support for ARIA 3443 3444 *Paul Dale* 3445 3446 * s_client will now send the Server Name Indication (SNI) extension by 3447 default unless the new "-noservername" option is used. The server name is 3448 based on the host provided to the "-connect" option unless overridden by 3449 using "-servername". 3450 3451 *Matt Caswell* 3452 3453 * Add support for SipHash 3454 3455 *Todd Short* 3456 3457 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0 3458 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to 3459 prevent issues where no progress is being made and the peer continually 3460 sends unrecognised record types, using up resources processing them. 3461 3462 *Matt Caswell* 3463 3464 * 'openssl passwd' can now produce SHA256 and SHA512 based output, 3465 using the algorithm defined in 3466 <https://www.akkadia.org/drepper/SHA-crypt.txt> 3467 3468 *Richard Levitte* 3469 3470 * Heartbeat support has been removed; the ABI is changed for now. 3471 3472 *Richard Levitte, Rich Salz* 3473 3474 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd. 3475 3476 *Emilia Käsper* 3477 3478 * The RSA "null" method, which was partially supported to avoid patent 3479 issues, has been replaced to always returns NULL. 3480 3481 *Rich Salz* 3482 3483OpenSSL 1.1.0 3484------------- 3485 3486### Changes between 1.1.0k and 1.1.0l [10 Sep 2019] 3487 3488 * For built-in EC curves, ensure an EC_GROUP built from the curve name is 3489 used even when parsing explicit parameters, when loading a encoded key 3490 or calling `EC_GROUP_new_from_ecpkparameters()`/ 3491 `EC_GROUP_new_from_ecparameters()`. 3492 This prevents bypass of security hardening and performance gains, 3493 especially for curves with specialized EC_METHODs. 3494 By default, if a key encoded with explicit parameters is loaded and later 3495 encoded, the output is still encoded with explicit parameters, even if 3496 internally a "named" EC_GROUP is used for computation. 3497 3498 *Nicola Tuveri* 3499 3500 * Compute ECC cofactors if not provided during EC_GROUP construction. Before 3501 this change, EC_GROUP_set_generator would accept order and/or cofactor as 3502 NULL. After this change, only the cofactor parameter can be NULL. It also 3503 does some minimal sanity checks on the passed order. 3504 ([CVE-2019-1547]) 3505 3506 *Billy Bob Brumley* 3507 3508 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey. 3509 An attack is simple, if the first CMS_recipientInfo is valid but the 3510 second CMS_recipientInfo is chosen ciphertext. If the second 3511 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct 3512 encryption key will be replaced by garbage, and the message cannot be 3513 decoded, but if the RSA decryption fails, the correct encryption key is 3514 used and the recipient will not notice the attack. 3515 As a work around for this potential attack the length of the decrypted 3516 key must be equal to the cipher default key length, in case the 3517 certifiate is not given and all recipientInfo are tried out. 3518 The old behaviour can be re-enabled in the CMS code by setting the 3519 CMS_DEBUG_DECRYPT flag. 3520 ([CVE-2019-1563]) 3521 3522 *Bernd Edlinger* 3523 3524 * Use Windows installation paths in the mingw builds 3525 3526 Mingw isn't a POSIX environment per se, which means that Windows 3527 paths should be used for installation. 3528 ([CVE-2019-1552]) 3529 3530 *Richard Levitte* 3531 3532### Changes between 1.1.0j and 1.1.0k [28 May 2019] 3533 3534 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024. 3535 This changes the size when using the `genpkey` command when no size is given. 3536 It fixes an omission in earlier changes that changed all RSA, DSA and DH 3537 generation commands to use 2048 bits by default. 3538 3539 *Kurt Roeckx* 3540 3541 * Prevent over long nonces in ChaCha20-Poly1305. 3542 3543 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input 3544 for every encryption operation. RFC 7539 specifies that the nonce value 3545 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length 3546 and front pads the nonce with 0 bytes if it is less than 12 3547 bytes. However it also incorrectly allows a nonce to be set of up to 16 3548 bytes. In this case only the last 12 bytes are significant and any 3549 additional leading bytes are ignored. 3550 3551 It is a requirement of using this cipher that nonce values are 3552 unique. Messages encrypted using a reused nonce value are susceptible to 3553 serious confidentiality and integrity attacks. If an application changes 3554 the default nonce length to be longer than 12 bytes and then makes a 3555 change to the leading bytes of the nonce expecting the new value to be a 3556 new unique nonce then such an application could inadvertently encrypt 3557 messages with a reused nonce. 3558 3559 Additionally the ignored bytes in a long nonce are not covered by the 3560 integrity guarantee of this cipher. Any application that relies on the 3561 integrity of these ignored leading bytes of a long nonce may be further 3562 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS, 3563 is safe because no such use sets such a long nonce value. However user 3564 applications that use this cipher directly and set a non-default nonce 3565 length to be longer than 12 bytes may be vulnerable. 3566 3567 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk 3568 Greef of Ronomon. 3569 ([CVE-2019-1543]) 3570 3571 *Matt Caswell* 3572 3573 * Added SCA hardening for modular field inversion in EC_GROUP through 3574 a new dedicated field_inv() pointer in EC_METHOD. 3575 This also addresses a leakage affecting conversions from projective 3576 to affine coordinates. 3577 3578 *Billy Bob Brumley, Nicola Tuveri* 3579 3580 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a 3581 re-used X509_PUBKEY object if the second PUBKEY is malformed. 3582 3583 *Bernd Edlinger* 3584 3585 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0(). 3586 3587 *Richard Levitte* 3588 3589 * Remove the 'dist' target and add a tarball building script. The 3590 'dist' target has fallen out of use, and it shouldn't be 3591 necessary to configure just to create a source distribution. 3592 3593 *Richard Levitte* 3594 3595### Changes between 1.1.0i and 1.1.0j [20 Nov 2018] 3596 3597 * Timing vulnerability in DSA signature generation 3598 3599 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a 3600 timing side channel attack. An attacker could use variations in the signing 3601 algorithm to recover the private key. 3602 3603 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser. 3604 ([CVE-2018-0734]) 3605 3606 *Paul Dale* 3607 3608 * Timing vulnerability in ECDSA signature generation 3609 3610 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a 3611 timing side channel attack. An attacker could use variations in the signing 3612 algorithm to recover the private key. 3613 3614 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser. 3615 ([CVE-2018-0735]) 3616 3617 *Paul Dale* 3618 3619 * Add coordinate blinding for EC_POINT and implement projective 3620 coordinate blinding for generic prime curves as a countermeasure to 3621 chosen point SCA attacks. 3622 3623 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley* 3624 3625### Changes between 1.1.0h and 1.1.0i [14 Aug 2018] 3626 3627 * Client DoS due to large DH parameter 3628 3629 During key agreement in a TLS handshake using a DH(E) based ciphersuite a 3630 malicious server can send a very large prime value to the client. This will 3631 cause the client to spend an unreasonably long period of time generating a 3632 key for this prime resulting in a hang until the client has finished. This 3633 could be exploited in a Denial Of Service attack. 3634 3635 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken 3636 ([CVE-2018-0732]) 3637 3638 *Guido Vranken* 3639 3640 * Cache timing vulnerability in RSA Key Generation 3641 3642 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to 3643 a cache timing side channel attack. An attacker with sufficient access to 3644 mount cache timing attacks during the RSA key generation process could 3645 recover the private key. 3646 3647 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera 3648 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia. 3649 ([CVE-2018-0737]) 3650 3651 *Billy Brumley* 3652 3653 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str 3654 parameter is no longer accepted, as it leads to a corrupt table. NULL 3655 pem_str is reserved for alias entries only. 3656 3657 *Richard Levitte* 3658 3659 * Revert blinding in ECDSA sign and instead make problematic addition 3660 length-invariant. Switch even to fixed-length Montgomery multiplication. 3661 3662 *Andy Polyakov* 3663 3664 * Change generating and checking of primes so that the error rate of not 3665 being prime depends on the intended use based on the size of the input. 3666 For larger primes this will result in more rounds of Miller-Rabin. 3667 The maximal error rate for primes with more than 1080 bits is lowered 3668 to 2^-128. 3669 3670 *Kurt Roeckx, Annie Yousar* 3671 3672 * Increase the number of Miller-Rabin rounds for DSA key generating to 64. 3673 3674 *Kurt Roeckx* 3675 3676 * Add blinding to ECDSA and DSA signatures to protect against side channel 3677 attacks discovered by Keegan Ryan (NCC Group). 3678 3679 *Matt Caswell* 3680 3681 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we 3682 now allow empty (zero character) pass phrases. 3683 3684 *Richard Levitte* 3685 3686 * Certificate time validation (X509_cmp_time) enforces stricter 3687 compliance with RFC 5280. Fractional seconds and timezone offsets 3688 are no longer allowed. 3689 3690 *Emilia Käsper* 3691 3692 * Fixed a text canonicalisation bug in CMS 3693 3694 Where a CMS detached signature is used with text content the text goes 3695 through a canonicalisation process first prior to signing or verifying a 3696 signature. This process strips trailing space at the end of lines, converts 3697 line terminators to CRLF and removes additional trailing line terminators 3698 at the end of a file. A bug in the canonicalisation process meant that 3699 some characters, such as form-feed, were incorrectly treated as whitespace 3700 and removed. This is contrary to the specification (RFC5485). This fix 3701 could mean that detached text data signed with an earlier version of 3702 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data 3703 signed with a fixed OpenSSL may fail to verify with an earlier version of 3704 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data 3705 and use the "-binary" flag (for the "cms" command line application) or set 3706 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()). 3707 3708 *Matt Caswell* 3709 3710### Changes between 1.1.0g and 1.1.0h [27 Mar 2018] 3711 3712 * Constructed ASN.1 types with a recursive definition could exceed the stack 3713 3714 Constructed ASN.1 types with a recursive definition (such as can be found 3715 in PKCS7) could eventually exceed the stack given malicious input with 3716 excessive recursion. This could result in a Denial Of Service attack. There 3717 are no such structures used within SSL/TLS that come from untrusted sources 3718 so this is considered safe. 3719 3720 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz 3721 project. 3722 ([CVE-2018-0739]) 3723 3724 *Matt Caswell* 3725 3726 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC 3727 3728 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is 3729 effectively reduced to only comparing the least significant bit of each 3730 byte. This allows an attacker to forge messages that would be considered as 3731 authenticated in an amount of tries lower than that guaranteed by the 3732 security claims of the scheme. The module can only be compiled by the 3733 HP-UX assembler, so that only HP-UX PA-RISC targets are affected. 3734 3735 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg 3736 (IBM). 3737 ([CVE-2018-0733]) 3738 3739 *Andy Polyakov* 3740 3741 * Add a build target 'build_all_generated', to build all generated files 3742 and only that. This can be used to prepare everything that requires 3743 things like perl for a system that lacks perl and then move everything 3744 to that system and do the rest of the build there. 3745 3746 *Richard Levitte* 3747 3748 * Backport SSL_OP_NO_RENGOTIATION 3749 3750 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the 3751 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity 3752 changes this is no longer possible in 1.1.0. Therefore the new 3753 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to 3754 1.1.0 to provide equivalent functionality. 3755 3756 Note that if an application built against 1.1.0h headers (or above) is run 3757 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be 3758 accepted but nothing will happen, i.e. renegotiation will not be prevented. 3759 3760 *Matt Caswell* 3761 3762 * Removed the OS390-Unix config target. It relied on a script that doesn't 3763 exist. 3764 3765 *Rich Salz* 3766 3767 * rsaz_1024_mul_avx2 overflow bug on x86_64 3768 3769 There is an overflow bug in the AVX2 Montgomery multiplication procedure 3770 used in exponentiation with 1024-bit moduli. No EC algorithms are affected. 3771 Analysis suggests that attacks against RSA and DSA as a result of this 3772 defect would be very difficult to perform and are not believed likely. 3773 Attacks against DH1024 are considered just feasible, because most of the 3774 work necessary to deduce information about a private key may be performed 3775 offline. The amount of resources required for such an attack would be 3776 significant. However, for an attack on TLS to be meaningful, the server 3777 would have to share the DH1024 private key among multiple clients, which is 3778 no longer an option since CVE-2016-0701. 3779 3780 This only affects processors that support the AVX2 but not ADX extensions 3781 like Intel Haswell (4th generation). 3782 3783 This issue was reported to OpenSSL by David Benjamin (Google). The issue 3784 was originally found via the OSS-Fuzz project. 3785 ([CVE-2017-3738]) 3786 3787 *Andy Polyakov* 3788 3789### Changes between 1.1.0f and 1.1.0g [2 Nov 2017] 3790 3791 * bn_sqrx8x_internal carry bug on x86_64 3792 3793 There is a carry propagating bug in the x86_64 Montgomery squaring 3794 procedure. No EC algorithms are affected. Analysis suggests that attacks 3795 against RSA and DSA as a result of this defect would be very difficult to 3796 perform and are not believed likely. Attacks against DH are considered just 3797 feasible (although very difficult) because most of the work necessary to 3798 deduce information about a private key may be performed offline. The amount 3799 of resources required for such an attack would be very significant and 3800 likely only accessible to a limited number of attackers. An attacker would 3801 additionally need online access to an unpatched system using the target 3802 private key in a scenario with persistent DH parameters and a private 3803 key that is shared between multiple clients. 3804 3805 This only affects processors that support the BMI1, BMI2 and ADX extensions 3806 like Intel Broadwell (5th generation) and later or AMD Ryzen. 3807 3808 This issue was reported to OpenSSL by the OSS-Fuzz project. 3809 ([CVE-2017-3736]) 3810 3811 *Andy Polyakov* 3812 3813 * Malformed X.509 IPAddressFamily could cause OOB read 3814 3815 If an X.509 certificate has a malformed IPAddressFamily extension, 3816 OpenSSL could do a one-byte buffer overread. The most likely result 3817 would be an erroneous display of the certificate in text format. 3818 3819 This issue was reported to OpenSSL by the OSS-Fuzz project. 3820 ([CVE-2017-3735]) 3821 3822 *Rich Salz* 3823 3824### Changes between 1.1.0e and 1.1.0f [25 May 2017] 3825 3826 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target 3827 platform rather than 'mingw'. 3828 3829 *Richard Levitte* 3830 3831 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c. 3832 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1, 3833 which is the minimum version we support. 3834 3835 *Richard Levitte* 3836 3837### Changes between 1.1.0d and 1.1.0e [16 Feb 2017] 3838 3839 * Encrypt-Then-Mac renegotiation crash 3840 3841 During a renegotiation handshake if the Encrypt-Then-Mac extension is 3842 negotiated where it was not in the original handshake (or vice-versa) then 3843 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients 3844 and servers are affected. 3845 3846 This issue was reported to OpenSSL by Joe Orton (Red Hat). 3847 ([CVE-2017-3733]) 3848 3849 *Matt Caswell* 3850 3851### Changes between 1.1.0c and 1.1.0d [26 Jan 2017] 3852 3853 * Truncated packet could crash via OOB read 3854 3855 If one side of an SSL/TLS path is running on a 32-bit host and a specific 3856 cipher is being used, then a truncated packet can cause that host to 3857 perform an out-of-bounds read, usually resulting in a crash. 3858 3859 This issue was reported to OpenSSL by Robert Święcki of Google. 3860 ([CVE-2017-3731]) 3861 3862 *Andy Polyakov* 3863 3864 * Bad (EC)DHE parameters cause a client crash 3865 3866 If a malicious server supplies bad parameters for a DHE or ECDHE key 3867 exchange then this can result in the client attempting to dereference a 3868 NULL pointer leading to a client crash. This could be exploited in a Denial 3869 of Service attack. 3870 3871 This issue was reported to OpenSSL by Guido Vranken. 3872 ([CVE-2017-3730]) 3873 3874 *Matt Caswell* 3875 3876 * BN_mod_exp may produce incorrect results on x86_64 3877 3878 There is a carry propagating bug in the x86_64 Montgomery squaring 3879 procedure. No EC algorithms are affected. Analysis suggests that attacks 3880 against RSA and DSA as a result of this defect would be very difficult to 3881 perform and are not believed likely. Attacks against DH are considered just 3882 feasible (although very difficult) because most of the work necessary to 3883 deduce information about a private key may be performed offline. The amount 3884 of resources required for such an attack would be very significant and 3885 likely only accessible to a limited number of attackers. An attacker would 3886 additionally need online access to an unpatched system using the target 3887 private key in a scenario with persistent DH parameters and a private 3888 key that is shared between multiple clients. For example this can occur by 3889 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very 3890 similar to CVE-2015-3193 but must be treated as a separate problem. 3891 3892 This issue was reported to OpenSSL by the OSS-Fuzz project. 3893 ([CVE-2017-3732]) 3894 3895 *Andy Polyakov* 3896 3897### Changes between 1.1.0b and 1.1.0c [10 Nov 2016] 3898 3899 * ChaCha20/Poly1305 heap-buffer-overflow 3900 3901 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to 3902 a DoS attack by corrupting larger payloads. This can result in an OpenSSL 3903 crash. This issue is not considered to be exploitable beyond a DoS. 3904 3905 This issue was reported to OpenSSL by Robert Święcki (Google Security Team) 3906 ([CVE-2016-7054]) 3907 3908 *Richard Levitte* 3909 3910 * CMS Null dereference 3911 3912 Applications parsing invalid CMS structures can crash with a NULL pointer 3913 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE 3914 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the 3915 structure callback if an attempt is made to free certain invalid encodings. 3916 Only CHOICE structures using a callback which do not handle NULL value are 3917 affected. 3918 3919 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure. 3920 ([CVE-2016-7053]) 3921 3922 *Stephen Henson* 3923 3924 * Montgomery multiplication may produce incorrect results 3925 3926 There is a carry propagating bug in the Broadwell-specific Montgomery 3927 multiplication procedure that handles input lengths divisible by, but 3928 longer than 256 bits. Analysis suggests that attacks against RSA, DSA 3929 and DH private keys are impossible. This is because the subroutine in 3930 question is not used in operations with the private key itself and an input 3931 of the attacker's direct choice. Otherwise the bug can manifest itself as 3932 transient authentication and key negotiation failures or reproducible 3933 erroneous outcome of public-key operations with specially crafted input. 3934 Among EC algorithms only Brainpool P-512 curves are affected and one 3935 presumably can attack ECDH key negotiation. Impact was not analyzed in 3936 detail, because pre-requisites for attack are considered unlikely. Namely 3937 multiple clients have to choose the curve in question and the server has to 3938 share the private key among them, neither of which is default behaviour. 3939 Even then only clients that chose the curve will be affected. 3940 3941 This issue was publicly reported as transient failures and was not 3942 initially recognized as a security issue. Thanks to Richard Morgan for 3943 providing reproducible case. 3944 ([CVE-2016-7055]) 3945 3946 *Andy Polyakov* 3947 3948 * Removed automatic addition of RPATH in shared libraries and executables, 3949 as this was a remainder from OpenSSL 1.0.x and isn't needed any more. 3950 3951 *Richard Levitte* 3952 3953### Changes between 1.1.0a and 1.1.0b [26 Sep 2016] 3954 3955 * Fix Use After Free for large message sizes 3956 3957 The patch applied to address CVE-2016-6307 resulted in an issue where if a 3958 message larger than approx 16k is received then the underlying buffer to 3959 store the incoming message is reallocated and moved. Unfortunately a 3960 dangling pointer to the old location is left which results in an attempt to 3961 write to the previously freed location. This is likely to result in a 3962 crash, however it could potentially lead to execution of arbitrary code. 3963 3964 This issue only affects OpenSSL 1.1.0a. 3965 3966 This issue was reported to OpenSSL by Robert Święcki. 3967 ([CVE-2016-6309]) 3968 3969 *Matt Caswell* 3970 3971### Changes between 1.1.0 and 1.1.0a [22 Sep 2016] 3972 3973 * OCSP Status Request extension unbounded memory growth 3974 3975 A malicious client can send an excessively large OCSP Status Request 3976 extension. If that client continually requests renegotiation, sending a 3977 large OCSP Status Request extension each time, then there will be unbounded 3978 memory growth on the server. This will eventually lead to a Denial Of 3979 Service attack through memory exhaustion. Servers with a default 3980 configuration are vulnerable even if they do not support OCSP. Builds using 3981 the "no-ocsp" build time option are not affected. 3982 3983 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 3984 ([CVE-2016-6304]) 3985 3986 *Matt Caswell* 3987 3988 * SSL_peek() hang on empty record 3989 3990 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer 3991 sends an empty record. This could be exploited by a malicious peer in a 3992 Denial Of Service attack. 3993 3994 This issue was reported to OpenSSL by Alex Gaynor. 3995 ([CVE-2016-6305]) 3996 3997 *Matt Caswell* 3998 3999 * Excessive allocation of memory in tls_get_message_header() and 4000 dtls1_preprocess_fragment() 4001 4002 A (D)TLS message includes 3 bytes for its length in the header for the 4003 message. This would allow for messages up to 16Mb in length. Messages of 4004 this length are excessive and OpenSSL includes a check to ensure that a 4005 peer is sending reasonably sized messages in order to avoid too much memory 4006 being consumed to service a connection. A flaw in the logic of version 4007 1.1.0 means that memory for the message is allocated too early, prior to 4008 the excessive message length check. Due to way memory is allocated in 4009 OpenSSL this could mean an attacker could force up to 21Mb to be allocated 4010 to service a connection. This could lead to a Denial of Service through 4011 memory exhaustion. However, the excessive message length check still takes 4012 place, and this would cause the connection to immediately fail. Assuming 4013 that the application calls SSL_free() on the failed connection in a timely 4014 manner then the 21Mb of allocated memory will then be immediately freed 4015 again. Therefore the excessive memory allocation will be transitory in 4016 nature. This then means that there is only a security impact if: 4017 4018 1) The application does not call SSL_free() in a timely manner in the event 4019 that the connection fails 4020 or 4021 2) The application is working in a constrained environment where there is 4022 very little free memory 4023 or 4024 3) The attacker initiates multiple connection attempts such that there are 4025 multiple connections in a state where memory has been allocated for the 4026 connection; SSL_free() has not yet been called; and there is insufficient 4027 memory to service the multiple requests. 4028 4029 Except in the instance of (1) above any Denial Of Service is likely to be 4030 transitory because as soon as the connection fails the memory is 4031 subsequently freed again in the SSL_free() call. However there is an 4032 increased risk during this period of application crashes due to the lack of 4033 memory - which would then mean a more serious Denial of Service. 4034 4035 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 4036 (CVE-2016-6307 and CVE-2016-6308) 4037 4038 *Matt Caswell* 4039 4040 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler, 4041 had to be removed. Primary reason is that vendor assembler can't 4042 assemble our modules with -KPIC flag. As result it, assembly 4043 support, was not even available as option. But its lack means 4044 lack of side-channel resistant code, which is incompatible with 4045 security by todays standards. Fortunately gcc is readily available 4046 prepackaged option, which we firmly point at... 4047 4048 *Andy Polyakov* 4049 4050### Changes between 1.0.2h and 1.1.0 [25 Aug 2016] 4051 4052 * Windows command-line tool supports UTF-8 opt-in option for arguments 4053 and console input. Setting OPENSSL_WIN32_UTF8 environment variable 4054 (to any value) allows Windows user to access PKCS#12 file generated 4055 with Windows CryptoAPI and protected with non-ASCII password, as well 4056 as files generated under UTF-8 locale on Linux also protected with 4057 non-ASCII password. 4058 4059 *Andy Polyakov* 4060 4061 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites 4062 have been disabled by default and removed from DEFAULT, just like RC4. 4063 See the RC4 item below to re-enable both. 4064 4065 *Rich Salz* 4066 4067 * The method for finding the storage location for the Windows RAND seed file 4068 has changed. First we check %RANDFILE%. If that is not set then we check 4069 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If 4070 all else fails we fall back to C:\. 4071 4072 *Matt Caswell* 4073 4074 * The EVP_EncryptUpdate() function has had its return type changed from void 4075 to int. A return of 0 indicates and error while a return of 1 indicates 4076 success. 4077 4078 *Matt Caswell* 4079 4080 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and 4081 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch 4082 off the constant time implementation for RSA, DSA and DH have been made 4083 no-ops and deprecated. 4084 4085 *Matt Caswell* 4086 4087 * Windows RAND implementation was simplified to only get entropy by 4088 calling CryptGenRandom(). Various other RAND-related tickets 4089 were also closed. 4090 4091 *Joseph Wylie Yandle, Rich Salz* 4092 4093 * The stack and lhash API's were renamed to start with `OPENSSL_SK_` 4094 and `OPENSSL_LH_`, respectively. The old names are available 4095 with API compatibility. They new names are now completely documented. 4096 4097 *Rich Salz* 4098 4099 * Unify TYPE_up_ref(obj) methods signature. 4100 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(), 4101 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an 4102 int (instead of void) like all others TYPE_up_ref() methods. 4103 So now these methods also check the return value of CRYPTO_atomic_add(), 4104 and the validity of object reference counter. 4105 4106 *fdasilvayy@gmail.com* 4107 4108 * With Windows Visual Studio builds, the .pdb files are installed 4109 alongside the installed libraries and executables. For a static 4110 library installation, ossl_static.pdb is the associate compiler 4111 generated .pdb file to be used when linking programs. 4112 4113 *Richard Levitte* 4114 4115 * Remove openssl.spec. Packaging files belong with the packagers. 4116 4117 *Richard Levitte* 4118 4119 * Automatic Darwin/OSX configuration has had a refresh, it will now 4120 recognise x86_64 architectures automatically. You can still decide 4121 to build for a different bitness with the environment variable 4122 KERNEL_BITS (can be 32 or 64), for example: 4123 4124 KERNEL_BITS=32 ./config 4125 4126 *Richard Levitte* 4127 4128 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0, 4129 256 bit AES and HMAC with SHA256. 4130 4131 *Steve Henson* 4132 4133 * Remove support for MIPS o32 ABI on IRIX (and IRIX only). 4134 4135 *Andy Polyakov* 4136 4137 * Triple-DES ciphers have been moved from HIGH to MEDIUM. 4138 4139 *Rich Salz* 4140 4141 * To enable users to have their own config files and build file templates, 4142 Configure looks in the directory indicated by the environment variable 4143 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/ 4144 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical 4145 name and is used as is. 4146 4147 *Richard Levitte* 4148 4149 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX, 4150 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type 4151 X509_CERT_FILE_CTX was removed. 4152 4153 *Rich Salz* 4154 4155 * "shared" builds are now the default. To create only static libraries use 4156 the "no-shared" Configure option. 4157 4158 *Matt Caswell* 4159 4160 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options. 4161 All of these option have not worked for some while and are fundamental 4162 algorithms. 4163 4164 *Matt Caswell* 4165 4166 * Make various cleanup routines no-ops and mark them as deprecated. Most 4167 global cleanup functions are no longer required because they are handled 4168 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages). 4169 Explicitly de-initing can cause problems (e.g. where a library that uses 4170 OpenSSL de-inits, but an application is still using it). The affected 4171 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(), 4172 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(), 4173 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and 4174 COMP_zlib_cleanup(). 4175 4176 *Matt Caswell* 4177 4178 * --strict-warnings no longer enables runtime debugging options 4179 such as REF_DEBUG. Instead, debug options are automatically 4180 enabled with '--debug' builds. 4181 4182 *Andy Polyakov, Emilia Käsper* 4183 4184 * Made DH and DH_METHOD opaque. The structures for managing DH objects 4185 have been moved out of the public header files. New functions for managing 4186 these have been added. 4187 4188 *Matt Caswell* 4189 4190 * Made RSA and RSA_METHOD opaque. The structures for managing RSA 4191 objects have been moved out of the public header files. New 4192 functions for managing these have been added. 4193 4194 *Richard Levitte* 4195 4196 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects 4197 have been moved out of the public header files. New functions for managing 4198 these have been added. 4199 4200 *Matt Caswell* 4201 4202 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been 4203 moved out of the public header files. New functions for managing these 4204 have been added. 4205 4206 *Matt Caswell* 4207 4208 * Removed no-rijndael as a config option. Rijndael is an old name for AES. 4209 4210 *Matt Caswell* 4211 4212 * Removed the mk1mf build scripts. 4213 4214 *Richard Levitte* 4215 4216 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so 4217 it is always safe to #include a header now. 4218 4219 *Rich Salz* 4220 4221 * Removed the aged BC-32 config and all its supporting scripts 4222 4223 *Richard Levitte* 4224 4225 * Removed support for Ultrix, Netware, and OS/2. 4226 4227 *Rich Salz* 4228 4229 * Add support for HKDF. 4230 4231 *Alessandro Ghedini* 4232 4233 * Add support for blake2b and blake2s 4234 4235 *Bill Cox* 4236 4237 * Added support for "pipelining". Ciphers that have the 4238 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple 4239 encryptions/decryptions simultaneously. There are currently no built-in 4240 ciphers with this property but the expectation is that engines will be able 4241 to offer it to significantly improve throughput. Support has been extended 4242 into libssl so that multiple records for a single connection can be 4243 processed in one go (for >=TLS 1.1). 4244 4245 *Matt Caswell* 4246 4247 * Added the AFALG engine. This is an async capable engine which is able to 4248 offload work to the Linux kernel. In this initial version it only supports 4249 AES128-CBC. The kernel must be version 4.1.0 or greater. 4250 4251 *Catriona Lucey* 4252 4253 * OpenSSL now uses a new threading API. It is no longer necessary to 4254 set locking callbacks to use OpenSSL in a multi-threaded environment. There 4255 are two supported threading models: pthreads and windows threads. It is 4256 also possible to configure OpenSSL at compile time for "no-threads". The 4257 old threading API should no longer be used. The functions have been 4258 replaced with "no-op" compatibility macros. 4259 4260 *Alessandro Ghedini, Matt Caswell* 4261 4262 * Modify behavior of ALPN to invoke callback after SNI/servername 4263 callback, such that updates to the SSL_CTX affect ALPN. 4264 4265 *Todd Short* 4266 4267 * Add SSL_CIPHER queries for authentication and key-exchange. 4268 4269 *Todd Short* 4270 4271 * Changes to the DEFAULT cipherlist: 4272 - Prefer (EC)DHE handshakes over plain RSA. 4273 - Prefer AEAD ciphers over legacy ciphers. 4274 - Prefer ECDSA over RSA when both certificates are available. 4275 - Prefer TLSv1.2 ciphers/PRF. 4276 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the 4277 default cipherlist. 4278 4279 *Emilia Käsper* 4280 4281 * Change the ECC default curve list to be this, in order: x25519, 4282 secp256r1, secp521r1, secp384r1. 4283 4284 *Rich Salz* 4285 4286 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are 4287 disabled by default. They can be re-enabled using the 4288 enable-weak-ssl-ciphers option to Configure. 4289 4290 *Matt Caswell* 4291 4292 * If the server has ALPN configured, but supports no protocols that the 4293 client advertises, send a fatal "no_application_protocol" alert. 4294 This behaviour is SHALL in RFC 7301, though it isn't universally 4295 implemented by other servers. 4296 4297 *Emilia Käsper* 4298 4299 * Add X25519 support. 4300 Add ASN.1 and EVP_PKEY methods for X25519. This includes support 4301 for public and private key encoding using the format documented in 4302 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports 4303 key generation and key derivation. 4304 4305 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses 4306 X25519(29). 4307 4308 *Steve Henson* 4309 4310 * Deprecate SRP_VBASE_get_by_user. 4311 SRP_VBASE_get_by_user had inconsistent memory management behaviour. 4312 In order to fix an unavoidable memory leak ([CVE-2016-0798]), 4313 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP 4314 seed, even if the seed is configured. 4315 4316 Users should use SRP_VBASE_get1_by_user instead. Note that in 4317 SRP_VBASE_get1_by_user, caller must free the returned value. Note 4318 also that even though configuring the SRP seed attempts to hide 4319 invalid usernames by continuing the handshake with fake 4320 credentials, this behaviour is not constant time and no strong 4321 guarantees are made that the handshake is indistinguishable from 4322 that of a valid user. 4323 4324 *Emilia Käsper* 4325 4326 * Configuration change; it's now possible to build dynamic engines 4327 without having to build shared libraries and vice versa. This 4328 only applies to the engines in `engines/`, those in `crypto/engine/` 4329 will always be built into libcrypto (i.e. "static"). 4330 4331 Building dynamic engines is enabled by default; to disable, use 4332 the configuration option "disable-dynamic-engine". 4333 4334 The only requirements for building dynamic engines are the 4335 presence of the DSO module and building with position independent 4336 code, so they will also automatically be disabled if configuring 4337 with "disable-dso" or "disable-pic". 4338 4339 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE 4340 are also taken away from openssl/opensslconf.h, as they are 4341 irrelevant. 4342 4343 *Richard Levitte* 4344 4345 * Configuration change; if there is a known flag to compile 4346 position independent code, it will always be applied on the 4347 libcrypto and libssl object files, and never on the application 4348 object files. This means other libraries that use routines from 4349 libcrypto / libssl can be made into shared libraries regardless 4350 of how OpenSSL was configured. 4351 4352 If this isn't desirable, the configuration options "disable-pic" 4353 or "no-pic" can be used to disable the use of PIC. This will 4354 also disable building shared libraries and dynamic engines. 4355 4356 *Richard Levitte* 4357 4358 * Removed JPAKE code. It was experimental and has no wide use. 4359 4360 *Rich Salz* 4361 4362 * The INSTALL_PREFIX Makefile variable has been renamed to 4363 DESTDIR. That makes for less confusion on what this variable 4364 is for. Also, the configuration option --install_prefix is 4365 removed. 4366 4367 *Richard Levitte* 4368 4369 * Heartbeat for TLS has been removed and is disabled by default 4370 for DTLS; configure with enable-heartbeats. Code that uses the 4371 old #define's might need to be updated. 4372 4373 *Emilia Käsper, Rich Salz* 4374 4375 * Rename REF_CHECK to REF_DEBUG. 4376 4377 *Rich Salz* 4378 4379 * New "unified" build system 4380 4381 The "unified" build system is aimed to be a common system for all 4382 platforms we support. With it comes new support for VMS. 4383 4384 This system builds supports building in a different directory tree 4385 than the source tree. It produces one Makefile (for unix family 4386 or lookalikes), or one descrip.mms (for VMS). 4387 4388 The source of information to make the Makefile / descrip.mms is 4389 small files called 'build.info', holding the necessary 4390 information for each directory with source to compile, and a 4391 template in Configurations, like unix-Makefile.tmpl or 4392 descrip.mms.tmpl. 4393 4394 With this change, the library names were also renamed on Windows 4395 and on VMS. They now have names that are closer to the standard 4396 on Unix, and include the major version number, and in certain 4397 cases, the architecture they are built for. See "Notes on shared 4398 libraries" in INSTALL. 4399 4400 We rely heavily on the perl module Text::Template. 4401 4402 *Richard Levitte* 4403 4404 * Added support for auto-initialisation and de-initialisation of the library. 4405 OpenSSL no longer requires explicit init or deinit routines to be called, 4406 except in certain circumstances. See the OPENSSL_init_crypto() and 4407 OPENSSL_init_ssl() man pages for further information. 4408 4409 *Matt Caswell* 4410 4411 * The arguments to the DTLSv1_listen function have changed. Specifically the 4412 "peer" argument is now expected to be a BIO_ADDR object. 4413 4414 * Rewrite of BIO networking library. The BIO library lacked consistent 4415 support of IPv6, and adding it required some more extensive 4416 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types, 4417 which hold all types of addresses and chains of address information. 4418 It also introduces a new API, with functions like BIO_socket, 4419 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept. 4420 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram 4421 have been adapted accordingly. 4422 4423 *Richard Levitte* 4424 4425 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without 4426 the leading 0-byte. 4427 4428 *Emilia Käsper* 4429 4430 * CRIME protection: disable compression by default, even if OpenSSL is 4431 compiled with zlib enabled. Applications can still enable compression 4432 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by 4433 using the SSL_CONF library to configure compression. 4434 4435 *Emilia Käsper* 4436 4437 * The signature of the session callback configured with 4438 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer 4439 was explicitly marked as `const unsigned char*` instead of 4440 `unsigned char*`. 4441 4442 *Emilia Käsper* 4443 4444 * Always DPURIFY. Remove the use of uninitialized memory in the 4445 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op. 4446 4447 *Emilia Käsper* 4448 4449 * Removed many obsolete configuration items, including 4450 DES_PTR, DES_RISC1, DES_RISC2, DES_INT 4451 MD2_CHAR, MD2_INT, MD2_LONG 4452 BF_PTR, BF_PTR2 4453 IDEA_SHORT, IDEA_LONG 4454 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX 4455 4456 *Rich Salz, with advice from Andy Polyakov* 4457 4458 * Many BN internals have been moved to an internal header file. 4459 4460 *Rich Salz with help from Andy Polyakov* 4461 4462 * Configuration and writing out the results from it has changed. 4463 Files such as Makefile include/openssl/opensslconf.h and are now 4464 produced through general templates, such as Makefile.in and 4465 crypto/opensslconf.h.in and some help from the perl module 4466 Text::Template. 4467 4468 Also, the center of configuration information is no longer 4469 Makefile. Instead, Configure produces a perl module in 4470 configdata.pm which holds most of the config data (in the hash 4471 table %config), the target data that comes from the target 4472 configuration in one of the `Configurations/*.conf` files (in 4473 %target). 4474 4475 *Richard Levitte* 4476 4477 * To clarify their intended purposes, the Configure options 4478 --prefix and --openssldir change their semantics, and become more 4479 straightforward and less interdependent. 4480 4481 --prefix shall be used exclusively to give the location INSTALLTOP 4482 where programs, scripts, libraries, include files and manuals are 4483 going to be installed. The default is now /usr/local. 4484 4485 --openssldir shall be used exclusively to give the default 4486 location OPENSSLDIR where certificates, private keys, CRLs are 4487 managed. This is also where the default openssl.cnf gets 4488 installed. 4489 If the directory given with this option is a relative path, the 4490 values of both the --prefix value and the --openssldir value will 4491 be combined to become OPENSSLDIR. 4492 The default for --openssldir is INSTALLTOP/ssl. 4493 4494 Anyone who uses --openssldir to specify where OpenSSL is to be 4495 installed MUST change to use --prefix instead. 4496 4497 *Richard Levitte* 4498 4499 * The GOST engine was out of date and therefore it has been removed. An up 4500 to date GOST engine is now being maintained in an external repository. 4501 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains 4502 support for GOST ciphersuites (these are only activated if a GOST engine 4503 is present). 4504 4505 *Matt Caswell* 4506 4507 * EGD is no longer supported by default; use enable-egd when 4508 configuring. 4509 4510 *Ben Kaduk and Rich Salz* 4511 4512 * The distribution now has Makefile.in files, which are used to 4513 create Makefile's when Configure is run. *Configure must be run 4514 before trying to build now.* 4515 4516 *Rich Salz* 4517 4518 * The return value for SSL_CIPHER_description() for error conditions 4519 has changed. 4520 4521 *Rich Salz* 4522 4523 * Support for RFC6698/RFC7671 DANE TLSA peer authentication. 4524 4525 Obtaining and performing DNSSEC validation of TLSA records is 4526 the application's responsibility. The application provides 4527 the TLSA records of its choice to OpenSSL, and these are then 4528 used to authenticate the peer. 4529 4530 The TLSA records need not even come from DNS. They can, for 4531 example, be used to implement local end-entity certificate or 4532 trust-anchor "pinning", where the "pin" data takes the form 4533 of TLSA records, which can augment or replace verification 4534 based on the usual WebPKI public certification authorities. 4535 4536 *Viktor Dukhovni* 4537 4538 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL 4539 continues to support deprecated interfaces in default builds. 4540 However, applications are strongly advised to compile their 4541 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides 4542 the declarations of all interfaces deprecated in 0.9.8, 1.0.0 4543 or the 1.1.0 releases. 4544 4545 In environments in which all applications have been ported to 4546 not use any deprecated interfaces OpenSSL's Configure script 4547 should be used with the --api=1.1.0 option to entirely remove 4548 support for the deprecated features from the library and 4549 unconditionally disable them in the installed headers. 4550 Essentially the same effect can be achieved with the "no-deprecated" 4551 argument to Configure, except that this will always restrict 4552 the build to just the latest API, rather than a fixed API 4553 version. 4554 4555 As applications are ported to future revisions of the API, 4556 they should update their compile-time OPENSSL_API_COMPAT define 4557 accordingly, but in most cases should be able to continue to 4558 compile with later releases. 4559 4560 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are 4561 0x10000000L and 0x00908000L, respectively. However those 4562 versions did not support the OPENSSL_API_COMPAT feature, and 4563 so applications are not typically tested for explicit support 4564 of just the undeprecated features of either release. 4565 4566 *Viktor Dukhovni* 4567 4568 * Add support for setting the minimum and maximum supported protocol. 4569 It can bet set via the SSL_set_min_proto_version() and 4570 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and 4571 MaxProtocol. It's recommended to use the new APIs to disable 4572 protocols instead of disabling individual protocols using 4573 SSL_set_options() or SSL_CONF's Protocol. This change also 4574 removes support for disabling TLS 1.2 in the OpenSSL TLS 4575 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT. 4576 4577 *Kurt Roeckx* 4578 4579 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl. 4580 4581 *Andy Polyakov* 4582 4583 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD 4584 and integrates ECDSA and ECDH functionality into EC. Implementations can 4585 now redirect key generation and no longer need to convert to or from 4586 ECDSA_SIG format. 4587 4588 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just 4589 include the ec.h header file instead. 4590 4591 *Steve Henson* 4592 4593 * Remove support for all 40 and 56 bit ciphers. This includes all the export 4594 ciphers who are no longer supported and drops support the ephemeral RSA key 4595 exchange. The LOW ciphers currently doesn't have any ciphers in it. 4596 4597 *Kurt Roeckx* 4598 4599 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX 4600 opaque. For HMAC_CTX, the following constructors and destructors 4601 were added: 4602 4603 HMAC_CTX *HMAC_CTX_new(void); 4604 void HMAC_CTX_free(HMAC_CTX *ctx); 4605 4606 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and 4607 destroy such methods has been added. See EVP_MD_meth_new(3) and 4608 EVP_CIPHER_meth_new(3) for documentation. 4609 4610 Additional changes: 4611 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and 4612 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and 4613 `EVP_MD_CTX_reset()` should be called instead to reinitialise 4614 an already created structure. 4615 2) For consistency with the majority of our object creators and 4616 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to 4617 `EVP_MD_CTX_(new|free)`. The old names are retained as macros 4618 for deprecated builds. 4619 4620 *Richard Levitte* 4621 4622 * Added ASYNC support. Libcrypto now includes the async sub-library to enable 4623 cryptographic operations to be performed asynchronously as long as an 4624 asynchronous capable engine is used. See the ASYNC_start_job() man page for 4625 further details. Libssl has also had this capability integrated with the 4626 introduction of the new mode SSL_MODE_ASYNC and associated error 4627 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man 4628 pages. This work was developed in partnership with Intel Corp. 4629 4630 *Matt Caswell* 4631 4632 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is 4633 always enabled now. If you want to disable the support you should 4634 exclude it using the list of supported ciphers. This also means that the 4635 "-no_ecdhe" option has been removed from s_server. 4636 4637 *Kurt Roeckx* 4638 4639 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls 4640 SSL_{CTX_}set1_curves() which can set a list. 4641 4642 *Kurt Roeckx* 4643 4644 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the 4645 curve you want to support using SSL_{CTX_}set1_curves(). 4646 4647 *Kurt Roeckx* 4648 4649 * State machine rewrite. The state machine code has been significantly 4650 refactored in order to remove much duplication of code and solve issues 4651 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for 4652 further details). This change does have some associated API changes. 4653 Notably the SSL_state() function has been removed and replaced by 4654 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int. 4655 SSL_set_state() has been removed altogether. The previous handshake states 4656 defined in ssl.h and ssl3.h have also been removed. 4657 4658 *Matt Caswell* 4659 4660 * All instances of the string "ssleay" in the public API were replaced 4661 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's) 4662 Some error codes related to internal RSA_eay API's were renamed. 4663 4664 *Rich Salz* 4665 4666 * The demo files in crypto/threads were moved to demo/threads. 4667 4668 *Rich Salz* 4669 4670 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp, 4671 sureware and ubsec. 4672 4673 *Matt Caswell, Rich Salz* 4674 4675 * New ASN.1 embed macro. 4676 4677 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the 4678 structure is not allocated: it is part of the parent. That is instead of 4679 4680 FOO *x; 4681 4682 it must be: 4683 4684 FOO x; 4685 4686 This reduces memory fragmentation and make it impossible to accidentally 4687 set a mandatory field to NULL. 4688 4689 This currently only works for some fields specifically a SEQUENCE, CHOICE, 4690 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is 4691 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or 4692 SEQUENCE OF. 4693 4694 *Steve Henson* 4695 4696 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled. 4697 4698 *Emilia Käsper* 4699 4700 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although 4701 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also 4702 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add 4703 DES and RC4 ciphersuites. 4704 4705 *Matt Caswell* 4706 4707 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs. 4708 This changes the decoding behaviour for some invalid messages, 4709 though the change is mostly in the more lenient direction, and 4710 legacy behaviour is preserved as much as possible. 4711 4712 *Emilia Käsper* 4713 4714 * Fix no-stdio build. 4715 *David Woodhouse <David.Woodhouse@intel.com> and also* 4716 *Ivan Nestlerode <ivan.nestlerode@sonos.com>* 4717 4718 * New testing framework 4719 The testing framework has been largely rewritten and is now using 4720 perl and the perl modules Test::Harness and an extended variant of 4721 Test::More called OpenSSL::Test to do its work. All test scripts in 4722 test/ have been rewritten into test recipes, and all direct calls to 4723 executables in test/Makefile have become individual recipes using the 4724 simplified testing OpenSSL::Test::Simple. 4725 4726 For documentation on our testing modules, do: 4727 4728 perldoc test/testlib/OpenSSL/Test/Simple.pm 4729 perldoc test/testlib/OpenSSL/Test.pm 4730 4731 *Richard Levitte* 4732 4733 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT 4734 are used; the latter aborts on memory leaks (usually checked on exit). 4735 Some undocumented "set malloc, etc., hooks" functions were removed 4736 and others were changed. All are now documented. 4737 4738 *Rich Salz* 4739 4740 * In DSA_generate_parameters_ex, if the provided seed is too short, 4741 return an error 4742 4743 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>* 4744 4745 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites 4746 from RFC4279, RFC4785, RFC5487, RFC5489. 4747 4748 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the 4749 original RSA_PSK patch. 4750 4751 *Steve Henson* 4752 4753 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay 4754 era flag was never set throughout the codebase (only read). Also removed 4755 SSL3_FLAGS_POP_BUFFER which was only used if 4756 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set. 4757 4758 *Matt Caswell* 4759 4760 * Changed the default name options in the "ca", "crl", "req" and "x509" 4761 to be "oneline" instead of "compat". 4762 4763 *Richard Levitte* 4764 4765 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're 4766 not aware of clients that still exhibit this bug, and the workaround 4767 hasn't been working properly for a while. 4768 4769 *Emilia Käsper* 4770 4771 * The return type of BIO_number_read() and BIO_number_written() as well as 4772 the corresponding num_read and num_write members in the BIO structure has 4773 changed from unsigned long to uint64_t. On platforms where an unsigned 4774 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is 4775 transferred. 4776 4777 *Matt Caswell* 4778 4779 * Given the pervasive nature of TLS extensions it is inadvisable to run 4780 OpenSSL without support for them. It also means that maintaining 4781 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably 4782 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed. 4783 4784 *Matt Caswell* 4785 4786 * Removed support for the two export grade static DH ciphersuites 4787 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites 4788 were newly added (along with a number of other static DH ciphersuites) to 4789 1.0.2. However the two export ones have *never* worked since they were 4790 introduced. It seems strange in any case to be adding new export 4791 ciphersuites, and given "logjam" it also does not seem correct to fix them. 4792 4793 *Matt Caswell* 4794 4795 * Version negotiation has been rewritten. In particular SSLv23_method(), 4796 SSLv23_client_method() and SSLv23_server_method() have been deprecated, 4797 and turned into macros which simply call the new preferred function names 4798 TLS_method(), TLS_client_method() and TLS_server_method(). All new code 4799 should use the new names instead. Also as part of this change the ssl23.h 4800 header file has been removed. 4801 4802 *Matt Caswell* 4803 4804 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This 4805 code and the associated standard is no longer considered fit-for-purpose. 4806 4807 *Matt Caswell* 4808 4809 * RT2547 was closed. When generating a private key, try to make the 4810 output file readable only by the owner. This behavior change might 4811 be noticeable when interacting with other software. 4812 4813 * Documented all exdata functions. Added CRYPTO_free_ex_index. 4814 Added a test. 4815 4816 *Rich Salz* 4817 4818 * Added HTTP GET support to the ocsp command. 4819 4820 *Rich Salz* 4821 4822 * Changed default digest for the dgst and enc commands from MD5 to 4823 sha256 4824 4825 *Rich Salz* 4826 4827 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead. 4828 4829 *Matt Caswell* 4830 4831 * Added support for TLS extended master secret from 4832 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an 4833 initial patch which was a great help during development. 4834 4835 *Steve Henson* 4836 4837 * All libssl internal structures have been removed from the public header 4838 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is 4839 now redundant). Users should not attempt to access internal structures 4840 directly. Instead they should use the provided API functions. 4841 4842 *Matt Caswell* 4843 4844 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used. 4845 Access to deprecated functions can be re-enabled by running config with 4846 "enable-deprecated". In addition applications wishing to use deprecated 4847 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour 4848 will, by default, disable some transitive includes that previously existed 4849 in the header files (e.g. ec.h will no longer, by default, include bn.h) 4850 4851 *Matt Caswell* 4852 4853 * Added support for OCB mode. OpenSSL has been granted a patent license 4854 compatible with the OpenSSL license for use of OCB. Details are available 4855 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support 4856 for OCB can be removed by calling config with no-ocb. 4857 4858 *Matt Caswell* 4859 4860 * SSLv2 support has been removed. It still supports receiving a SSLv2 4861 compatible client hello. 4862 4863 *Kurt Roeckx* 4864 4865 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz], 4866 done while fixing the error code for the key-too-small case. 4867 4868 *Annie Yousar <a.yousar@informatik.hu-berlin.de>* 4869 4870 * CA.sh has been removed; use CA.pl instead. 4871 4872 *Rich Salz* 4873 4874 * Removed old DES API. 4875 4876 *Rich Salz* 4877 4878 * Remove various unsupported platforms: 4879 Sony NEWS4 4880 BEOS and BEOS_R5 4881 NeXT 4882 SUNOS 4883 MPE/iX 4884 Sinix/ReliantUNIX RM400 4885 DGUX 4886 NCR 4887 Tandem 4888 Cray 4889 16-bit platforms such as WIN16 4890 4891 *Rich Salz* 4892 4893 * Clean up OPENSSL_NO_xxx #define's 4894 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF 4895 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx 4896 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC 4897 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160 4898 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO 4899 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY 4900 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP 4901 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK 4902 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY 4903 - Remove MS_STATIC; it's a relic from platforms <32 bits. 4904 4905 *Rich Salz* 4906 4907 * Cleaned up dead code 4908 Remove all but one '#ifdef undef' which is to be looked at. 4909 4910 *Rich Salz* 4911 4912 * Clean up calling of xxx_free routines. 4913 Just like free(), fix most of the xxx_free routines to accept 4914 NULL. Remove the non-null checks from callers. Save much code. 4915 4916 *Rich Salz* 4917 4918 * Add secure heap for storage of private keys (when possible). 4919 Add BIO_s_secmem(), CBIGNUM, etc. 4920 Contributed by Akamai Technologies under our Corporate CLA. 4921 4922 *Rich Salz* 4923 4924 * Experimental support for a new, fast, unbiased prime candidate generator, 4925 bn_probable_prime_dh_coprime(). Not currently used by any prime generator. 4926 4927 *Felix Laurie von Massenbach <felix@erbridge.co.uk>* 4928 4929 * New output format NSS in the sess_id command line tool. This allows 4930 exporting the session id and the master key in NSS keylog format. 4931 4932 *Martin Kaiser <martin@kaiser.cx>* 4933 4934 * Harmonize version and its documentation. -f flag is used to display 4935 compilation flags. 4936 4937 *mancha <mancha1@zoho.com>* 4938 4939 * Fix eckey_priv_encode so it immediately returns an error upon a failure 4940 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue. 4941 4942 *mancha <mancha1@zoho.com>* 4943 4944 * Fix some double frees. These are not thought to be exploitable. 4945 4946 *mancha <mancha1@zoho.com>* 4947 4948 * A missing bounds check in the handling of the TLS heartbeat extension 4949 can be used to reveal up to 64k of memory to a connected client or 4950 server. 4951 4952 Thanks for Neel Mehta of Google Security for discovering this bug and to 4953 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 4954 preparing the fix ([CVE-2014-0160]) 4955 4956 *Adam Langley, Bodo Moeller* 4957 4958 * Fix for the attack described in the paper "Recovering OpenSSL 4959 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 4960 by Yuval Yarom and Naomi Benger. Details can be obtained from: 4961 <http://eprint.iacr.org/2014/140> 4962 4963 Thanks to Yuval Yarom and Naomi Benger for discovering this 4964 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076]) 4965 4966 *Yuval Yarom and Naomi Benger* 4967 4968 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file(): 4969 this fixes a limitation in previous versions of OpenSSL. 4970 4971 *Steve Henson* 4972 4973 * Experimental encrypt-then-mac support. 4974 4975 Experimental support for encrypt then mac from 4976 draft-gutmann-tls-encrypt-then-mac-02.txt 4977 4978 To enable it set the appropriate extension number (0x42 for the test 4979 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42 4980 4981 For non-compliant peers (i.e. just about everything) this should have no 4982 effect. 4983 4984 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE. 4985 4986 *Steve Henson* 4987 4988 * Add EVP support for key wrapping algorithms, to avoid problems with 4989 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in 4990 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap 4991 algorithms and include tests cases. 4992 4993 *Steve Henson* 4994 4995 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for 4996 enveloped data. 4997 4998 *Steve Henson* 4999 5000 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest, 5001 MGF1 digest and OAEP label. 5002 5003 *Steve Henson* 5004 5005 * Make openssl verify return errors. 5006 5007 *Chris Palmer <palmer@google.com> and Ben Laurie* 5008 5009 * New function ASN1_TIME_diff to calculate the difference between two 5010 ASN1_TIME structures or one structure and the current time. 5011 5012 *Steve Henson* 5013 5014 * Update fips_test_suite to support multiple command line options. New 5015 test to induce all self test errors in sequence and check expected 5016 failures. 5017 5018 *Steve Henson* 5019 5020 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and 5021 sign or verify all in one operation. 5022 5023 *Steve Henson* 5024 5025 * Add fips_algvs: a multicall fips utility incorporating all the algorithm 5026 test programs and fips_test_suite. Includes functionality to parse 5027 the minimal script output of fipsalgest.pl directly. 5028 5029 *Steve Henson* 5030 5031 * Add authorisation parameter to FIPS_module_mode_set(). 5032 5033 *Steve Henson* 5034 5035 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves. 5036 5037 *Steve Henson* 5038 5039 * Use separate DRBG fields for internal and external flags. New function 5040 FIPS_drbg_health_check() to perform on demand health checking. Add 5041 generation tests to fips_test_suite with reduced health check interval to 5042 demonstrate periodic health checking. Add "nodh" option to 5043 fips_test_suite to skip very slow DH test. 5044 5045 *Steve Henson* 5046 5047 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers 5048 based on NID. 5049 5050 *Steve Henson* 5051 5052 * More extensive health check for DRBG checking many more failure modes. 5053 New function FIPS_selftest_drbg_all() to handle every possible DRBG 5054 combination: call this in fips_test_suite. 5055 5056 *Steve Henson* 5057 5058 * Add support for canonical generation of DSA parameter 'g'. See 5059 FIPS 186-3 A.2.3. 5060 5061 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and 5062 POST to handle HMAC cases. 5063 5064 *Steve Henson* 5065 5066 * Add functions FIPS_module_version() and FIPS_module_version_text() 5067 to return numerical and string versions of the FIPS module number. 5068 5069 *Steve Henson* 5070 5071 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and 5072 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented 5073 outside the validated module in the FIPS capable OpenSSL. 5074 5075 *Steve Henson* 5076 5077 * Minor change to DRBG entropy callback semantics. In some cases 5078 there is no multiple of the block length between min_len and 5079 max_len. Allow the callback to return more than max_len bytes 5080 of entropy but discard any extra: it is the callback's responsibility 5081 to ensure that the extra data discarded does not impact the 5082 requested amount of entropy. 5083 5084 *Steve Henson* 5085 5086 * Add PRNG security strength checks to RSA, DSA and ECDSA using 5087 information in FIPS186-3, SP800-57 and SP800-131A. 5088 5089 *Steve Henson* 5090 5091 * CCM support via EVP. Interface is very similar to GCM case except we 5092 must supply all data in one chunk (i.e. no update, final) and the 5093 message length must be supplied if AAD is used. Add algorithm test 5094 support. 5095 5096 *Steve Henson* 5097 5098 * Initial version of POST overhaul. Add POST callback to allow the status 5099 of POST to be monitored and/or failures induced. Modify fips_test_suite 5100 to use callback. Always run all selftests even if one fails. 5101 5102 *Steve Henson* 5103 5104 * XTS support including algorithm test driver in the fips_gcmtest program. 5105 Note: this does increase the maximum key length from 32 to 64 bytes but 5106 there should be no binary compatibility issues as existing applications 5107 will never use XTS mode. 5108 5109 *Steve Henson* 5110 5111 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies 5112 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also 5113 performs algorithm blocking for unapproved PRNG types. Also do not 5114 set PRNG type in FIPS_mode_set(): leave this to the application. 5115 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with 5116 the standard OpenSSL PRNG: set additional data to a date time vector. 5117 5118 *Steve Henson* 5119 5120 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`. 5121 This shouldn't present any incompatibility problems because applications 5122 shouldn't be using these directly and any that are will need to rethink 5123 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2 5124 5125 *Steve Henson* 5126 5127 * Extensive self tests and health checking required by SP800-90 DRBG. 5128 Remove strength parameter from FIPS_drbg_instantiate and always 5129 instantiate at maximum supported strength. 5130 5131 *Steve Henson* 5132 5133 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing. 5134 5135 *Steve Henson* 5136 5137 * New algorithm test program fips_dhvs to handle DH primitives only testing. 5138 5139 *Steve Henson* 5140 5141 * New function DH_compute_key_padded() to compute a DH key and pad with 5142 leading zeroes if needed: this complies with SP800-56A et al. 5143 5144 *Steve Henson* 5145 5146 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by 5147 anything, incomplete, subject to change and largely untested at present. 5148 5149 *Steve Henson* 5150 5151 * Modify fipscanisteronly build option to only build the necessary object 5152 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile. 5153 5154 *Steve Henson* 5155 5156 * Add experimental option FIPSSYMS to give all symbols in 5157 fipscanister.o and FIPS or fips prefix. This will avoid 5158 conflicts with future versions of OpenSSL. Add perl script 5159 util/fipsas.pl to preprocess assembly language source files 5160 and rename any affected symbols. 5161 5162 *Steve Henson* 5163 5164 * Add selftest checks and algorithm block of non-fips algorithms in 5165 FIPS mode. Remove DES2 from selftests. 5166 5167 *Steve Henson* 5168 5169 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just 5170 return internal method without any ENGINE dependencies. Add new 5171 tiny fips sign and verify functions. 5172 5173 *Steve Henson* 5174 5175 * New build option no-ec2m to disable characteristic 2 code. 5176 5177 *Steve Henson* 5178 5179 * New build option "fipscanisteronly". This only builds fipscanister.o 5180 and (currently) associated fips utilities. Uses the file Makefile.fips 5181 instead of Makefile.org as the prototype. 5182 5183 *Steve Henson* 5184 5185 * Add some FIPS mode restrictions to GCM. Add internal IV generator. 5186 Update fips_gcmtest to use IV generator. 5187 5188 *Steve Henson* 5189 5190 * Initial, experimental EVP support for AES-GCM. AAD can be input by 5191 setting output buffer to NULL. The `*Final` function must be 5192 called although it will not retrieve any additional data. The tag 5193 can be set or retrieved with a ctrl. The IV length is by default 12 5194 bytes (96 bits) but can be set to an alternative value. If the IV 5195 length exceeds the maximum IV length (currently 16 bytes) it cannot be 5196 set before the key. 5197 5198 *Steve Henson* 5199 5200 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the 5201 underlying do_cipher function handles all cipher semantics itself 5202 including padding and finalisation. This is useful if (for example) 5203 an ENGINE cipher handles block padding itself. The behaviour of 5204 do_cipher is subtly changed if this flag is set: the return value 5205 is the number of characters written to the output buffer (zero is 5206 no longer an error code) or a negative error code. Also if the 5207 input buffer is NULL and length 0 finalisation should be performed. 5208 5209 *Steve Henson* 5210 5211 * If a candidate issuer certificate is already part of the constructed 5212 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case. 5213 5214 *Steve Henson* 5215 5216 * Improve forward-security support: add functions 5217 5218 void SSL_CTX_set_not_resumable_session_callback( 5219 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure)) 5220 void SSL_set_not_resumable_session_callback( 5221 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure)) 5222 5223 for use by SSL/TLS servers; the callback function will be called whenever a 5224 new session is created, and gets to decide whether the session may be 5225 cached to make it resumable (return 0) or not (return 1). (As by the 5226 SSL/TLS protocol specifications, the session_id sent by the server will be 5227 empty to indicate that the session is not resumable; also, the server will 5228 not generate RFC 4507 (RFC 5077) session tickets.) 5229 5230 A simple reasonable callback implementation is to return is_forward_secure. 5231 This parameter will be set to 1 or 0 depending on the ciphersuite selected 5232 by the SSL/TLS server library, indicating whether it can provide forward 5233 security. 5234 5235 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)* 5236 5237 * New -verify_name option in command line utilities to set verification 5238 parameters by name. 5239 5240 *Steve Henson* 5241 5242 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE. 5243 Add CMAC pkey methods. 5244 5245 *Steve Henson* 5246 5247 * Experimental renegotiation in s_server -www mode. If the client 5248 browses /reneg connection is renegotiated. If /renegcert it is 5249 renegotiated requesting a certificate. 5250 5251 *Steve Henson* 5252 5253 * Add an "external" session cache for debugging purposes to s_server. This 5254 should help trace issues which normally are only apparent in deployed 5255 multi-process servers. 5256 5257 *Steve Henson* 5258 5259 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where 5260 return value is ignored. NB. The functions RAND_add(), RAND_seed(), 5261 BIO_set_cipher() and some obscure PEM functions were changed so they 5262 can now return an error. The RAND changes required a change to the 5263 RAND_METHOD structure. 5264 5265 *Steve Henson* 5266 5267 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of 5268 a gcc attribute to warn if the result of a function is ignored. This 5269 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h 5270 whose return value is often ignored. 5271 5272 *Steve Henson* 5273 5274 * New -noct, -requestct, -requirect and -ctlogfile options for s_client. 5275 These allow SCTs (signed certificate timestamps) to be requested and 5276 validated when establishing a connection. 5277 5278 *Rob Percival <robpercival@google.com>* 5279 5280OpenSSL 1.0.2 5281------------- 5282 5283### Changes between 1.0.2s and 1.0.2t [10 Sep 2019] 5284 5285 * For built-in EC curves, ensure an EC_GROUP built from the curve name is 5286 used even when parsing explicit parameters, when loading a encoded key 5287 or calling `EC_GROUP_new_from_ecpkparameters()`/ 5288 `EC_GROUP_new_from_ecparameters()`. 5289 This prevents bypass of security hardening and performance gains, 5290 especially for curves with specialized EC_METHODs. 5291 By default, if a key encoded with explicit parameters is loaded and later 5292 encoded, the output is still encoded with explicit parameters, even if 5293 internally a "named" EC_GROUP is used for computation. 5294 5295 *Nicola Tuveri* 5296 5297 * Compute ECC cofactors if not provided during EC_GROUP construction. Before 5298 this change, EC_GROUP_set_generator would accept order and/or cofactor as 5299 NULL. After this change, only the cofactor parameter can be NULL. It also 5300 does some minimal sanity checks on the passed order. 5301 ([CVE-2019-1547]) 5302 5303 *Billy Bob Brumley* 5304 5305 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey. 5306 An attack is simple, if the first CMS_recipientInfo is valid but the 5307 second CMS_recipientInfo is chosen ciphertext. If the second 5308 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct 5309 encryption key will be replaced by garbage, and the message cannot be 5310 decoded, but if the RSA decryption fails, the correct encryption key is 5311 used and the recipient will not notice the attack. 5312 As a work around for this potential attack the length of the decrypted 5313 key must be equal to the cipher default key length, in case the 5314 certifiate is not given and all recipientInfo are tried out. 5315 The old behaviour can be re-enabled in the CMS code by setting the 5316 CMS_DEBUG_DECRYPT flag. 5317 ([CVE-2019-1563]) 5318 5319 *Bernd Edlinger* 5320 5321 * Document issue with installation paths in diverse Windows builds 5322 5323 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL 5324 binaries and run-time config file. 5325 ([CVE-2019-1552]) 5326 5327 *Richard Levitte* 5328 5329### Changes between 1.0.2r and 1.0.2s [28 May 2019] 5330 5331 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024. 5332 This changes the size when using the `genpkey` command when no size is given. 5333 It fixes an omission in earlier changes that changed all RSA, DSA and DH 5334 generation commands to use 2048 bits by default. 5335 5336 *Kurt Roeckx* 5337 5338 * Add FIPS support for Android Arm 64-bit 5339 5340 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object 5341 Module in Version 2.0.10. For some reason, the corresponding target 5342 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be 5343 built with FIPS support on Android Arm 64-bit. This omission has been 5344 fixed. 5345 5346 *Matthias St. Pierre* 5347 5348### Changes between 1.0.2q and 1.0.2r [26 Feb 2019] 5349 5350 * 0-byte record padding oracle 5351 5352 If an application encounters a fatal protocol error and then calls 5353 SSL_shutdown() twice (once to send a close_notify, and once to receive one) 5354 then OpenSSL can respond differently to the calling application if a 0 byte 5355 record is received with invalid padding compared to if a 0 byte record is 5356 received with an invalid MAC. If the application then behaves differently 5357 based on that in a way that is detectable to the remote peer, then this 5358 amounts to a padding oracle that could be used to decrypt data. 5359 5360 In order for this to be exploitable "non-stitched" ciphersuites must be in 5361 use. Stitched ciphersuites are optimised implementations of certain 5362 commonly used ciphersuites. Also the application must call SSL_shutdown() 5363 twice even if a protocol error has occurred (applications should not do 5364 this but some do anyway). 5365 5366 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod 5367 Aviram, with additional investigation by Steven Collison and Andrew 5368 Hourselt. It was reported to OpenSSL on 10th December 2018. 5369 ([CVE-2019-1559]) 5370 5371 *Matt Caswell* 5372 5373 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0(). 5374 5375 *Richard Levitte* 5376 5377### Changes between 1.0.2p and 1.0.2q [20 Nov 2018] 5378 5379 * Microarchitecture timing vulnerability in ECC scalar multiplication 5380 5381 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been 5382 shown to be vulnerable to a microarchitecture timing side channel attack. 5383 An attacker with sufficient access to mount local timing attacks during 5384 ECDSA signature generation could recover the private key. 5385 5386 This issue was reported to OpenSSL on 26th October 2018 by Alejandro 5387 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and 5388 Nicola Tuveri. 5389 ([CVE-2018-5407]) 5390 5391 *Billy Brumley* 5392 5393 * Timing vulnerability in DSA signature generation 5394 5395 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a 5396 timing side channel attack. An attacker could use variations in the signing 5397 algorithm to recover the private key. 5398 5399 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser. 5400 ([CVE-2018-0734]) 5401 5402 *Paul Dale* 5403 5404 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object 5405 Module, accidentally introduced while backporting security fixes from the 5406 development branch and hindering the use of ECC in FIPS mode. 5407 5408 *Nicola Tuveri* 5409 5410### Changes between 1.0.2o and 1.0.2p [14 Aug 2018] 5411 5412 * Client DoS due to large DH parameter 5413 5414 During key agreement in a TLS handshake using a DH(E) based ciphersuite a 5415 malicious server can send a very large prime value to the client. This will 5416 cause the client to spend an unreasonably long period of time generating a 5417 key for this prime resulting in a hang until the client has finished. This 5418 could be exploited in a Denial Of Service attack. 5419 5420 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken 5421 ([CVE-2018-0732]) 5422 5423 *Guido Vranken* 5424 5425 * Cache timing vulnerability in RSA Key Generation 5426 5427 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to 5428 a cache timing side channel attack. An attacker with sufficient access to 5429 mount cache timing attacks during the RSA key generation process could 5430 recover the private key. 5431 5432 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera 5433 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia. 5434 ([CVE-2018-0737]) 5435 5436 *Billy Brumley* 5437 5438 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str 5439 parameter is no longer accepted, as it leads to a corrupt table. NULL 5440 pem_str is reserved for alias entries only. 5441 5442 *Richard Levitte* 5443 5444 * Revert blinding in ECDSA sign and instead make problematic addition 5445 length-invariant. Switch even to fixed-length Montgomery multiplication. 5446 5447 *Andy Polyakov* 5448 5449 * Change generating and checking of primes so that the error rate of not 5450 being prime depends on the intended use based on the size of the input. 5451 For larger primes this will result in more rounds of Miller-Rabin. 5452 The maximal error rate for primes with more than 1080 bits is lowered 5453 to 2^-128. 5454 5455 *Kurt Roeckx, Annie Yousar* 5456 5457 * Increase the number of Miller-Rabin rounds for DSA key generating to 64. 5458 5459 *Kurt Roeckx* 5460 5461 * Add blinding to ECDSA and DSA signatures to protect against side channel 5462 attacks discovered by Keegan Ryan (NCC Group). 5463 5464 *Matt Caswell* 5465 5466 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we 5467 now allow empty (zero character) pass phrases. 5468 5469 *Richard Levitte* 5470 5471 * Certificate time validation (X509_cmp_time) enforces stricter 5472 compliance with RFC 5280. Fractional seconds and timezone offsets 5473 are no longer allowed. 5474 5475 *Emilia Käsper* 5476 5477### Changes between 1.0.2n and 1.0.2o [27 Mar 2018] 5478 5479 * Constructed ASN.1 types with a recursive definition could exceed the stack 5480 5481 Constructed ASN.1 types with a recursive definition (such as can be found 5482 in PKCS7) could eventually exceed the stack given malicious input with 5483 excessive recursion. This could result in a Denial Of Service attack. There 5484 are no such structures used within SSL/TLS that come from untrusted sources 5485 so this is considered safe. 5486 5487 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz 5488 project. 5489 ([CVE-2018-0739]) 5490 5491 *Matt Caswell* 5492 5493### Changes between 1.0.2m and 1.0.2n [7 Dec 2017] 5494 5495 * Read/write after SSL object in error state 5496 5497 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state" 5498 mechanism. The intent was that if a fatal error occurred during a handshake 5499 then OpenSSL would move into the error state and would immediately fail if 5500 you attempted to continue the handshake. This works as designed for the 5501 explicit handshake functions (SSL_do_handshake(), SSL_accept() and 5502 SSL_connect()), however due to a bug it does not work correctly if 5503 SSL_read() or SSL_write() is called directly. In that scenario, if the 5504 handshake fails then a fatal error will be returned in the initial function 5505 call. If SSL_read()/SSL_write() is subsequently called by the application 5506 for the same SSL object then it will succeed and the data is passed without 5507 being decrypted/encrypted directly from the SSL/TLS record layer. 5508 5509 In order to exploit this issue an application bug would have to be present 5510 that resulted in a call to SSL_read()/SSL_write() being issued after having 5511 already received a fatal error. 5512 5513 This issue was reported to OpenSSL by David Benjamin (Google). 5514 ([CVE-2017-3737]) 5515 5516 *Matt Caswell* 5517 5518 * rsaz_1024_mul_avx2 overflow bug on x86_64 5519 5520 There is an overflow bug in the AVX2 Montgomery multiplication procedure 5521 used in exponentiation with 1024-bit moduli. No EC algorithms are affected. 5522 Analysis suggests that attacks against RSA and DSA as a result of this 5523 defect would be very difficult to perform and are not believed likely. 5524 Attacks against DH1024 are considered just feasible, because most of the 5525 work necessary to deduce information about a private key may be performed 5526 offline. The amount of resources required for such an attack would be 5527 significant. However, for an attack on TLS to be meaningful, the server 5528 would have to share the DH1024 private key among multiple clients, which is 5529 no longer an option since CVE-2016-0701. 5530 5531 This only affects processors that support the AVX2 but not ADX extensions 5532 like Intel Haswell (4th generation). 5533 5534 This issue was reported to OpenSSL by David Benjamin (Google). The issue 5535 was originally found via the OSS-Fuzz project. 5536 ([CVE-2017-3738]) 5537 5538 *Andy Polyakov* 5539 5540### Changes between 1.0.2l and 1.0.2m [2 Nov 2017] 5541 5542 * bn_sqrx8x_internal carry bug on x86_64 5543 5544 There is a carry propagating bug in the x86_64 Montgomery squaring 5545 procedure. No EC algorithms are affected. Analysis suggests that attacks 5546 against RSA and DSA as a result of this defect would be very difficult to 5547 perform and are not believed likely. Attacks against DH are considered just 5548 feasible (although very difficult) because most of the work necessary to 5549 deduce information about a private key may be performed offline. The amount 5550 of resources required for such an attack would be very significant and 5551 likely only accessible to a limited number of attackers. An attacker would 5552 additionally need online access to an unpatched system using the target 5553 private key in a scenario with persistent DH parameters and a private 5554 key that is shared between multiple clients. 5555 5556 This only affects processors that support the BMI1, BMI2 and ADX extensions 5557 like Intel Broadwell (5th generation) and later or AMD Ryzen. 5558 5559 This issue was reported to OpenSSL by the OSS-Fuzz project. 5560 ([CVE-2017-3736]) 5561 5562 *Andy Polyakov* 5563 5564 * Malformed X.509 IPAddressFamily could cause OOB read 5565 5566 If an X.509 certificate has a malformed IPAddressFamily extension, 5567 OpenSSL could do a one-byte buffer overread. The most likely result 5568 would be an erroneous display of the certificate in text format. 5569 5570 This issue was reported to OpenSSL by the OSS-Fuzz project. 5571 5572 *Rich Salz* 5573 5574### Changes between 1.0.2k and 1.0.2l [25 May 2017] 5575 5576 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target 5577 platform rather than 'mingw'. 5578 5579 *Richard Levitte* 5580 5581### Changes between 1.0.2j and 1.0.2k [26 Jan 2017] 5582 5583 * Truncated packet could crash via OOB read 5584 5585 If one side of an SSL/TLS path is running on a 32-bit host and a specific 5586 cipher is being used, then a truncated packet can cause that host to 5587 perform an out-of-bounds read, usually resulting in a crash. 5588 5589 This issue was reported to OpenSSL by Robert Święcki of Google. 5590 ([CVE-2017-3731]) 5591 5592 *Andy Polyakov* 5593 5594 * BN_mod_exp may produce incorrect results on x86_64 5595 5596 There is a carry propagating bug in the x86_64 Montgomery squaring 5597 procedure. No EC algorithms are affected. Analysis suggests that attacks 5598 against RSA and DSA as a result of this defect would be very difficult to 5599 perform and are not believed likely. Attacks against DH are considered just 5600 feasible (although very difficult) because most of the work necessary to 5601 deduce information about a private key may be performed offline. The amount 5602 of resources required for such an attack would be very significant and 5603 likely only accessible to a limited number of attackers. An attacker would 5604 additionally need online access to an unpatched system using the target 5605 private key in a scenario with persistent DH parameters and a private 5606 key that is shared between multiple clients. For example this can occur by 5607 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very 5608 similar to CVE-2015-3193 but must be treated as a separate problem. 5609 5610 This issue was reported to OpenSSL by the OSS-Fuzz project. 5611 ([CVE-2017-3732]) 5612 5613 *Andy Polyakov* 5614 5615 * Montgomery multiplication may produce incorrect results 5616 5617 There is a carry propagating bug in the Broadwell-specific Montgomery 5618 multiplication procedure that handles input lengths divisible by, but 5619 longer than 256 bits. Analysis suggests that attacks against RSA, DSA 5620 and DH private keys are impossible. This is because the subroutine in 5621 question is not used in operations with the private key itself and an input 5622 of the attacker's direct choice. Otherwise the bug can manifest itself as 5623 transient authentication and key negotiation failures or reproducible 5624 erroneous outcome of public-key operations with specially crafted input. 5625 Among EC algorithms only Brainpool P-512 curves are affected and one 5626 presumably can attack ECDH key negotiation. Impact was not analyzed in 5627 detail, because pre-requisites for attack are considered unlikely. Namely 5628 multiple clients have to choose the curve in question and the server has to 5629 share the private key among them, neither of which is default behaviour. 5630 Even then only clients that chose the curve will be affected. 5631 5632 This issue was publicly reported as transient failures and was not 5633 initially recognized as a security issue. Thanks to Richard Morgan for 5634 providing reproducible case. 5635 ([CVE-2016-7055]) 5636 5637 *Andy Polyakov* 5638 5639 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0 5640 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to 5641 prevent issues where no progress is being made and the peer continually 5642 sends unrecognised record types, using up resources processing them. 5643 5644 *Matt Caswell* 5645 5646### Changes between 1.0.2i and 1.0.2j [26 Sep 2016] 5647 5648 * Missing CRL sanity check 5649 5650 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0 5651 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use 5652 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception. 5653 5654 This issue only affects the OpenSSL 1.0.2i 5655 ([CVE-2016-7052]) 5656 5657 *Matt Caswell* 5658 5659### Changes between 1.0.2h and 1.0.2i [22 Sep 2016] 5660 5661 * OCSP Status Request extension unbounded memory growth 5662 5663 A malicious client can send an excessively large OCSP Status Request 5664 extension. If that client continually requests renegotiation, sending a 5665 large OCSP Status Request extension each time, then there will be unbounded 5666 memory growth on the server. This will eventually lead to a Denial Of 5667 Service attack through memory exhaustion. Servers with a default 5668 configuration are vulnerable even if they do not support OCSP. Builds using 5669 the "no-ocsp" build time option are not affected. 5670 5671 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 5672 ([CVE-2016-6304]) 5673 5674 *Matt Caswell* 5675 5676 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from 5677 HIGH to MEDIUM. 5678 5679 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan 5680 Leurent (INRIA) 5681 ([CVE-2016-2183]) 5682 5683 *Rich Salz* 5684 5685 * OOB write in MDC2_Update() 5686 5687 An overflow can occur in MDC2_Update() either if called directly or 5688 through the EVP_DigestUpdate() function using MDC2. If an attacker 5689 is able to supply very large amounts of input data after a previous 5690 call to EVP_EncryptUpdate() with a partial block then a length check 5691 can overflow resulting in a heap corruption. 5692 5693 The amount of data needed is comparable to SIZE_MAX which is impractical 5694 on most platforms. 5695 5696 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 5697 ([CVE-2016-6303]) 5698 5699 *Stephen Henson* 5700 5701 * Malformed SHA512 ticket DoS 5702 5703 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a 5704 DoS attack where a malformed ticket will result in an OOB read which will 5705 ultimately crash. 5706 5707 The use of SHA512 in TLS session tickets is comparatively rare as it requires 5708 a custom server callback and ticket lookup mechanism. 5709 5710 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 5711 ([CVE-2016-6302]) 5712 5713 *Stephen Henson* 5714 5715 * OOB write in BN_bn2dec() 5716 5717 The function BN_bn2dec() does not check the return value of BN_div_word(). 5718 This can cause an OOB write if an application uses this function with an 5719 overly large BIGNUM. This could be a problem if an overly large certificate 5720 or CRL is printed out from an untrusted source. TLS is not affected because 5721 record limits will reject an oversized certificate before it is parsed. 5722 5723 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 5724 ([CVE-2016-2182]) 5725 5726 *Stephen Henson* 5727 5728 * OOB read in TS_OBJ_print_bio() 5729 5730 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is 5731 the total length the OID text representation would use and not the amount 5732 of data written. This will result in OOB reads when large OIDs are 5733 presented. 5734 5735 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 5736 ([CVE-2016-2180]) 5737 5738 *Stephen Henson* 5739 5740 * Pointer arithmetic undefined behaviour 5741 5742 Avoid some undefined pointer arithmetic 5743 5744 A common idiom in the codebase is to check limits in the following manner: 5745 "p + len > limit" 5746 5747 Where "p" points to some malloc'd data of SIZE bytes and 5748 limit == p + SIZE 5749 5750 "len" here could be from some externally supplied data (e.g. from a TLS 5751 message). 5752 5753 The rules of C pointer arithmetic are such that "p + len" is only well 5754 defined where len <= SIZE. Therefore the above idiom is actually 5755 undefined behaviour. 5756 5757 For example this could cause problems if some malloc implementation 5758 provides an address for "p" such that "p + len" actually overflows for 5759 values of len that are too big and therefore p + len < limit. 5760 5761 This issue was reported to OpenSSL by Guido Vranken 5762 ([CVE-2016-2177]) 5763 5764 *Matt Caswell* 5765 5766 * Constant time flag not preserved in DSA signing 5767 5768 Operations in the DSA signing algorithm should run in constant time in 5769 order to avoid side channel attacks. A flaw in the OpenSSL DSA 5770 implementation means that a non-constant time codepath is followed for 5771 certain operations. This has been demonstrated through a cache-timing 5772 attack to be sufficient for an attacker to recover the private DSA key. 5773 5774 This issue was reported by César Pereida (Aalto University), Billy Brumley 5775 (Tampere University of Technology), and Yuval Yarom (The University of 5776 Adelaide and NICTA). 5777 ([CVE-2016-2178]) 5778 5779 *César Pereida* 5780 5781 * DTLS buffered message DoS 5782 5783 In a DTLS connection where handshake messages are delivered out-of-order 5784 those messages that OpenSSL is not yet ready to process will be buffered 5785 for later use. Under certain circumstances, a flaw in the logic means that 5786 those messages do not get removed from the buffer even though the handshake 5787 has been completed. An attacker could force up to approx. 15 messages to 5788 remain in the buffer when they are no longer required. These messages will 5789 be cleared when the DTLS connection is closed. The default maximum size for 5790 a message is 100k. Therefore the attacker could force an additional 1500k 5791 to be consumed per connection. By opening many simulataneous connections an 5792 attacker could cause a DoS attack through memory exhaustion. 5793 5794 This issue was reported to OpenSSL by Quan Luo. 5795 ([CVE-2016-2179]) 5796 5797 *Matt Caswell* 5798 5799 * DTLS replay protection DoS 5800 5801 A flaw in the DTLS replay attack protection mechanism means that records 5802 that arrive for future epochs update the replay protection "window" before 5803 the MAC for the record has been validated. This could be exploited by an 5804 attacker by sending a record for the next epoch (which does not have to 5805 decrypt or have a valid MAC), with a very large sequence number. This means 5806 that all subsequent legitimate packets are dropped causing a denial of 5807 service for a specific DTLS connection. 5808 5809 This issue was reported to OpenSSL by the OCAP audit team. 5810 ([CVE-2016-2181]) 5811 5812 *Matt Caswell* 5813 5814 * Certificate message OOB reads 5815 5816 In OpenSSL 1.0.2 and earlier some missing message length checks can result 5817 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a 5818 theoretical DoS risk but this has not been observed in practice on common 5819 platforms. 5820 5821 The messages affected are client certificate, client certificate request 5822 and server certificate. As a result the attack can only be performed 5823 against a client or a server which enables client authentication. 5824 5825 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 5826 ([CVE-2016-6306]) 5827 5828 *Stephen Henson* 5829 5830### Changes between 1.0.2g and 1.0.2h [3 May 2016] 5831 5832 * Prevent padding oracle in AES-NI CBC MAC check 5833 5834 A MITM attacker can use a padding oracle attack to decrypt traffic 5835 when the connection uses an AES CBC cipher and the server support 5836 AES-NI. 5837 5838 This issue was introduced as part of the fix for Lucky 13 padding 5839 attack ([CVE-2013-0169]). The padding check was rewritten to be in 5840 constant time by making sure that always the same bytes are read and 5841 compared against either the MAC or padding bytes. But it no longer 5842 checked that there was enough data to have both the MAC and padding 5843 bytes. 5844 5845 This issue was reported by Juraj Somorovsky using TLS-Attacker. 5846 5847 *Kurt Roeckx* 5848 5849 * Fix EVP_EncodeUpdate overflow 5850 5851 An overflow can occur in the EVP_EncodeUpdate() function which is used for 5852 Base64 encoding of binary data. If an attacker is able to supply very large 5853 amounts of input data then a length check can overflow resulting in a heap 5854 corruption. 5855 5856 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by 5857 the `PEM_write_bio*` family of functions. These are mainly used within the 5858 OpenSSL command line applications, so any application which processes data 5859 from an untrusted source and outputs it as a PEM file should be considered 5860 vulnerable to this issue. User applications that call these APIs directly 5861 with large amounts of untrusted data may also be vulnerable. 5862 5863 This issue was reported by Guido Vranken. 5864 ([CVE-2016-2105]) 5865 5866 *Matt Caswell* 5867 5868 * Fix EVP_EncryptUpdate overflow 5869 5870 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker 5871 is able to supply very large amounts of input data after a previous call to 5872 EVP_EncryptUpdate() with a partial block then a length check can overflow 5873 resulting in a heap corruption. Following an analysis of all OpenSSL 5874 internal usage of the EVP_EncryptUpdate() function all usage is one of two 5875 forms. The first form is where the EVP_EncryptUpdate() call is known to be 5876 the first called function after an EVP_EncryptInit(), and therefore that 5877 specific call must be safe. The second form is where the length passed to 5878 EVP_EncryptUpdate() can be seen from the code to be some small value and 5879 therefore there is no possibility of an overflow. Since all instances are 5880 one of these two forms, it is believed that there can be no overflows in 5881 internal code due to this problem. It should be noted that 5882 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths. 5883 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances 5884 of these calls have also been analysed too and it is believed there are no 5885 instances in internal usage where an overflow could occur. 5886 5887 This issue was reported by Guido Vranken. 5888 ([CVE-2016-2106]) 5889 5890 *Matt Caswell* 5891 5892 * Prevent ASN.1 BIO excessive memory allocation 5893 5894 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio() 5895 a short invalid encoding can cause allocation of large amounts of memory 5896 potentially consuming excessive resources or exhausting memory. 5897 5898 Any application parsing untrusted data through d2i BIO functions is 5899 affected. The memory based functions such as d2i_X509() are *not* affected. 5900 Since the memory based functions are used by the TLS library, TLS 5901 applications are not affected. 5902 5903 This issue was reported by Brian Carpenter. 5904 ([CVE-2016-2109]) 5905 5906 *Stephen Henson* 5907 5908 * EBCDIC overread 5909 5910 ASN1 Strings that are over 1024 bytes can cause an overread in applications 5911 using the X509_NAME_oneline() function on EBCDIC systems. This could result 5912 in arbitrary stack data being returned in the buffer. 5913 5914 This issue was reported by Guido Vranken. 5915 ([CVE-2016-2176]) 5916 5917 *Matt Caswell* 5918 5919 * Modify behavior of ALPN to invoke callback after SNI/servername 5920 callback, such that updates to the SSL_CTX affect ALPN. 5921 5922 *Todd Short* 5923 5924 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the 5925 default. 5926 5927 *Kurt Roeckx* 5928 5929 * Only remove the SSLv2 methods with the no-ssl2-method option. When the 5930 methods are enabled and ssl2 is disabled the methods return NULL. 5931 5932 *Kurt Roeckx* 5933 5934### Changes between 1.0.2f and 1.0.2g [1 Mar 2016] 5935 5936* Disable weak ciphers in SSLv3 and up in default builds of OpenSSL. 5937 Builds that are not configured with "enable-weak-ssl-ciphers" will not 5938 provide any "EXPORT" or "LOW" strength ciphers. 5939 5940 *Viktor Dukhovni* 5941 5942* Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2 5943 is by default disabled at build-time. Builds that are not configured with 5944 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used, 5945 users who want to negotiate SSLv2 via the version-flexible SSLv23_method() 5946 will need to explicitly call either of: 5947 5948 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2); 5949 or 5950 SSL_clear_options(ssl, SSL_OP_NO_SSLv2); 5951 5952 as appropriate. Even if either of those is used, or the application 5953 explicitly uses the version-specific SSLv2_method() or its client and 5954 server variants, SSLv2 ciphers vulnerable to exhaustive search key 5955 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT 5956 ciphers, and SSLv2 56-bit DES are no longer available. 5957 ([CVE-2016-0800]) 5958 5959 *Viktor Dukhovni* 5960 5961 * Fix a double-free in DSA code 5962 5963 A double free bug was discovered when OpenSSL parses malformed DSA private 5964 keys and could lead to a DoS attack or memory corruption for applications 5965 that receive DSA private keys from untrusted sources. This scenario is 5966 considered rare. 5967 5968 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using 5969 libFuzzer. 5970 ([CVE-2016-0705]) 5971 5972 *Stephen Henson* 5973 5974 * Disable SRP fake user seed to address a server memory leak. 5975 5976 Add a new method SRP_VBASE_get1_by_user that handles the seed properly. 5977 5978 SRP_VBASE_get_by_user had inconsistent memory management behaviour. 5979 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user 5980 was changed to ignore the "fake user" SRP seed, even if the seed 5981 is configured. 5982 5983 Users should use SRP_VBASE_get1_by_user instead. Note that in 5984 SRP_VBASE_get1_by_user, caller must free the returned value. Note 5985 also that even though configuring the SRP seed attempts to hide 5986 invalid usernames by continuing the handshake with fake 5987 credentials, this behaviour is not constant time and no strong 5988 guarantees are made that the handshake is indistinguishable from 5989 that of a valid user. 5990 ([CVE-2016-0798]) 5991 5992 *Emilia Käsper* 5993 5994 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption 5995 5996 In the BN_hex2bn function the number of hex digits is calculated using an 5997 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For 5998 large values of `i` this can result in `bn_expand` not allocating any 5999 memory because `i * 4` is negative. This can leave the internal BIGNUM data 6000 field as NULL leading to a subsequent NULL ptr deref. For very large values 6001 of `i`, the calculation `i * 4` could be a positive value smaller than `i`. 6002 In this case memory is allocated to the internal BIGNUM data field, but it 6003 is insufficiently sized leading to heap corruption. A similar issue exists 6004 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn 6005 is ever called by user applications with very large untrusted hex/dec data. 6006 This is anticipated to be a rare occurrence. 6007 6008 All OpenSSL internal usage of these functions use data that is not expected 6009 to be untrusted, e.g. config file data or application command line 6010 arguments. If user developed applications generate config file data based 6011 on untrusted data then it is possible that this could also lead to security 6012 consequences. This is also anticipated to be rare. 6013 6014 This issue was reported to OpenSSL by Guido Vranken. 6015 ([CVE-2016-0797]) 6016 6017 *Matt Caswell* 6018 6019 * Fix memory issues in `BIO_*printf` functions 6020 6021 The internal `fmtstr` function used in processing a "%s" format string in 6022 the `BIO_*printf` functions could overflow while calculating the length of a 6023 string and cause an OOB read when printing very long strings. 6024 6025 Additionally the internal `doapr_outch` function can attempt to write to an 6026 OOB memory location (at an offset from the NULL pointer) in the event of a 6027 memory allocation failure. In 1.0.2 and below this could be caused where 6028 the size of a buffer to be allocated is greater than INT_MAX. E.g. this 6029 could be in processing a very long "%s" format string. Memory leaks can 6030 also occur. 6031 6032 The first issue may mask the second issue dependent on compiler behaviour. 6033 These problems could enable attacks where large amounts of untrusted data 6034 is passed to the `BIO_*printf` functions. If applications use these functions 6035 in this way then they could be vulnerable. OpenSSL itself uses these 6036 functions when printing out human-readable dumps of ASN.1 data. Therefore 6037 applications that print this data could be vulnerable if the data is from 6038 untrusted sources. OpenSSL command line applications could also be 6039 vulnerable where they print out ASN.1 data, or if untrusted data is passed 6040 as command line arguments. 6041 6042 Libssl is not considered directly vulnerable. Additionally certificates etc 6043 received via remote connections via libssl are also unlikely to be able to 6044 trigger these issues because of message size limits enforced within libssl. 6045 6046 This issue was reported to OpenSSL Guido Vranken. 6047 ([CVE-2016-0799]) 6048 6049 *Matt Caswell* 6050 6051 * Side channel attack on modular exponentiation 6052 6053 A side-channel attack was found which makes use of cache-bank conflicts on 6054 the Intel Sandy-Bridge microarchitecture which could lead to the recovery 6055 of RSA keys. The ability to exploit this issue is limited as it relies on 6056 an attacker who has control of code in a thread running on the same 6057 hyper-threaded core as the victim thread which is performing decryptions. 6058 6059 This issue was reported to OpenSSL by Yuval Yarom, The University of 6060 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and 6061 Nadia Heninger, University of Pennsylvania with more information at 6062 <http://cachebleed.info>. 6063 ([CVE-2016-0702]) 6064 6065 *Andy Polyakov* 6066 6067 * Change the `req` command to generate a 2048-bit RSA/DSA key by default, 6068 if no keysize is specified with default_bits. This fixes an 6069 omission in an earlier change that changed all RSA/DSA key generation 6070 commands to use 2048 bits by default. 6071 6072 *Emilia Käsper* 6073 6074### Changes between 1.0.2e and 1.0.2f [28 Jan 2016] 6075 6076 * DH small subgroups 6077 6078 Historically OpenSSL only ever generated DH parameters based on "safe" 6079 primes. More recently (in version 1.0.2) support was provided for 6080 generating X9.42 style parameter files such as those required for RFC 5114 6081 support. The primes used in such files may not be "safe". Where an 6082 application is using DH configured with parameters based on primes that are 6083 not "safe" then an attacker could use this fact to find a peer's private 6084 DH exponent. This attack requires that the attacker complete multiple 6085 handshakes in which the peer uses the same private DH exponent. For example 6086 this could be used to discover a TLS server's private DH exponent if it's 6087 reusing the private DH exponent or it's using a static DH ciphersuite. 6088 6089 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in 6090 TLS. It is not on by default. If the option is not set then the server 6091 reuses the same private DH exponent for the life of the server process and 6092 would be vulnerable to this attack. It is believed that many popular 6093 applications do set this option and would therefore not be at risk. 6094 6095 The fix for this issue adds an additional check where a "q" parameter is 6096 available (as is the case in X9.42 based parameters). This detects the 6097 only known attack, and is the only possible defense for static DH 6098 ciphersuites. This could have some performance impact. 6099 6100 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by 6101 default and cannot be disabled. This could have some performance impact. 6102 6103 This issue was reported to OpenSSL by Antonio Sanso (Adobe). 6104 ([CVE-2016-0701]) 6105 6106 *Matt Caswell* 6107 6108 * SSLv2 doesn't block disabled ciphers 6109 6110 A malicious client can negotiate SSLv2 ciphers that have been disabled on 6111 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have 6112 been disabled, provided that the SSLv2 protocol was not also disabled via 6113 SSL_OP_NO_SSLv2. 6114 6115 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram 6116 and Sebastian Schinzel. 6117 ([CVE-2015-3197]) 6118 6119 *Viktor Dukhovni* 6120 6121### Changes between 1.0.2d and 1.0.2e [3 Dec 2015] 6122 6123 * BN_mod_exp may produce incorrect results on x86_64 6124 6125 There is a carry propagating bug in the x86_64 Montgomery squaring 6126 procedure. No EC algorithms are affected. Analysis suggests that attacks 6127 against RSA and DSA as a result of this defect would be very difficult to 6128 perform and are not believed likely. Attacks against DH are considered just 6129 feasible (although very difficult) because most of the work necessary to 6130 deduce information about a private key may be performed offline. The amount 6131 of resources required for such an attack would be very significant and 6132 likely only accessible to a limited number of attackers. An attacker would 6133 additionally need online access to an unpatched system using the target 6134 private key in a scenario with persistent DH parameters and a private 6135 key that is shared between multiple clients. For example this can occur by 6136 default in OpenSSL DHE based SSL/TLS ciphersuites. 6137 6138 This issue was reported to OpenSSL by Hanno Böck. 6139 ([CVE-2015-3193]) 6140 6141 *Andy Polyakov* 6142 6143 * Certificate verify crash with missing PSS parameter 6144 6145 The signature verification routines will crash with a NULL pointer 6146 dereference if presented with an ASN.1 signature using the RSA PSS 6147 algorithm and absent mask generation function parameter. Since these 6148 routines are used to verify certificate signature algorithms this can be 6149 used to crash any certificate verification operation and exploited in a 6150 DoS attack. Any application which performs certificate verification is 6151 vulnerable including OpenSSL clients and servers which enable client 6152 authentication. 6153 6154 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG). 6155 ([CVE-2015-3194]) 6156 6157 *Stephen Henson* 6158 6159 * X509_ATTRIBUTE memory leak 6160 6161 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak 6162 memory. This structure is used by the PKCS#7 and CMS routines so any 6163 application which reads PKCS#7 or CMS data from untrusted sources is 6164 affected. SSL/TLS is not affected. 6165 6166 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using 6167 libFuzzer. 6168 ([CVE-2015-3195]) 6169 6170 *Stephen Henson* 6171 6172 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs. 6173 This changes the decoding behaviour for some invalid messages, 6174 though the change is mostly in the more lenient direction, and 6175 legacy behaviour is preserved as much as possible. 6176 6177 *Emilia Käsper* 6178 6179 * In DSA_generate_parameters_ex, if the provided seed is too short, 6180 return an error 6181 6182 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>* 6183 6184### Changes between 1.0.2c and 1.0.2d [9 Jul 2015] 6185 6186 * Alternate chains certificate forgery 6187 6188 During certificate verification, OpenSSL will attempt to find an 6189 alternative certificate chain if the first attempt to build such a chain 6190 fails. An error in the implementation of this logic can mean that an 6191 attacker could cause certain checks on untrusted certificates to be 6192 bypassed, such as the CA flag, enabling them to use a valid leaf 6193 certificate to act as a CA and "issue" an invalid certificate. 6194 6195 This issue was reported to OpenSSL by Adam Langley/David Benjamin 6196 (Google/BoringSSL). 6197 6198 *Matt Caswell* 6199 6200### Changes between 1.0.2b and 1.0.2c [12 Jun 2015] 6201 6202 * Fix HMAC ABI incompatibility. The previous version introduced an ABI 6203 incompatibility in the handling of HMAC. The previous ABI has now been 6204 restored. 6205 6206 *Matt Caswell* 6207 6208### Changes between 1.0.2a and 1.0.2b [11 Jun 2015] 6209 6210 * Malformed ECParameters causes infinite loop 6211 6212 When processing an ECParameters structure OpenSSL enters an infinite loop 6213 if the curve specified is over a specially malformed binary polynomial 6214 field. 6215 6216 This can be used to perform denial of service against any 6217 system which processes public keys, certificate requests or 6218 certificates. This includes TLS clients and TLS servers with 6219 client authentication enabled. 6220 6221 This issue was reported to OpenSSL by Joseph Barr-Pixton. 6222 ([CVE-2015-1788]) 6223 6224 *Andy Polyakov* 6225 6226 * Exploitable out-of-bounds read in X509_cmp_time 6227 6228 X509_cmp_time does not properly check the length of the ASN1_TIME 6229 string and can read a few bytes out of bounds. In addition, 6230 X509_cmp_time accepts an arbitrary number of fractional seconds in the 6231 time string. 6232 6233 An attacker can use this to craft malformed certificates and CRLs of 6234 various sizes and potentially cause a segmentation fault, resulting in 6235 a DoS on applications that verify certificates or CRLs. TLS clients 6236 that verify CRLs are affected. TLS clients and servers with client 6237 authentication enabled may be affected if they use custom verification 6238 callbacks. 6239 6240 This issue was reported to OpenSSL by Robert Swiecki (Google), and 6241 independently by Hanno Böck. 6242 ([CVE-2015-1789]) 6243 6244 *Emilia Käsper* 6245 6246 * PKCS7 crash with missing EnvelopedContent 6247 6248 The PKCS#7 parsing code does not handle missing inner EncryptedContent 6249 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs 6250 with missing content and trigger a NULL pointer dereference on parsing. 6251 6252 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 6253 structures from untrusted sources are affected. OpenSSL clients and 6254 servers are not affected. 6255 6256 This issue was reported to OpenSSL by Michal Zalewski (Google). 6257 ([CVE-2015-1790]) 6258 6259 *Emilia Käsper* 6260 6261 * CMS verify infinite loop with unknown hash function 6262 6263 When verifying a signedData message the CMS code can enter an infinite loop 6264 if presented with an unknown hash function OID. This can be used to perform 6265 denial of service against any system which verifies signedData messages using 6266 the CMS code. 6267 This issue was reported to OpenSSL by Johannes Bauer. 6268 ([CVE-2015-1792]) 6269 6270 *Stephen Henson* 6271 6272 * Race condition handling NewSessionTicket 6273 6274 If a NewSessionTicket is received by a multi-threaded client when attempting to 6275 reuse a previous ticket then a race condition can occur potentially leading to 6276 a double free of the ticket data. 6277 ([CVE-2015-1791]) 6278 6279 *Matt Caswell* 6280 6281 * Only support 256-bit or stronger elliptic curves with the 6282 'ecdh_auto' setting (server) or by default (client). Of supported 6283 curves, prefer P-256 (both). 6284 6285 *Emilia Kasper* 6286 6287### Changes between 1.0.2 and 1.0.2a [19 Mar 2015] 6288 6289 * ClientHello sigalgs DoS fix 6290 6291 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an 6292 invalid signature algorithms extension a NULL pointer dereference will 6293 occur. This can be exploited in a DoS attack against the server. 6294 6295 This issue was was reported to OpenSSL by David Ramos of Stanford 6296 University. 6297 ([CVE-2015-0291]) 6298 6299 *Stephen Henson and Matt Caswell* 6300 6301 * Multiblock corrupted pointer fix 6302 6303 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This 6304 feature only applies on 64 bit x86 architecture platforms that support AES 6305 NI instructions. A defect in the implementation of "multiblock" can cause 6306 OpenSSL's internal write buffer to become incorrectly set to NULL when 6307 using non-blocking IO. Typically, when the user application is using a 6308 socket BIO for writing, this will only result in a failed connection. 6309 However if some other BIO is used then it is likely that a segmentation 6310 fault will be triggered, thus enabling a potential DoS attack. 6311 6312 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller. 6313 ([CVE-2015-0290]) 6314 6315 *Matt Caswell* 6316 6317 * Segmentation fault in DTLSv1_listen fix 6318 6319 The DTLSv1_listen function is intended to be stateless and processes the 6320 initial ClientHello from many peers. It is common for user code to loop 6321 over the call to DTLSv1_listen until a valid ClientHello is received with 6322 an associated cookie. A defect in the implementation of DTLSv1_listen means 6323 that state is preserved in the SSL object from one invocation to the next 6324 that can lead to a segmentation fault. Errors processing the initial 6325 ClientHello can trigger this scenario. An example of such an error could be 6326 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only 6327 server. 6328 6329 This issue was reported to OpenSSL by Per Allansson. 6330 ([CVE-2015-0207]) 6331 6332 *Matt Caswell* 6333 6334 * Segmentation fault in ASN1_TYPE_cmp fix 6335 6336 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is 6337 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check 6338 certificate signature algorithm consistency this can be used to crash any 6339 certificate verification operation and exploited in a DoS attack. Any 6340 application which performs certificate verification is vulnerable including 6341 OpenSSL clients and servers which enable client authentication. 6342 ([CVE-2015-0286]) 6343 6344 *Stephen Henson* 6345 6346 * Segmentation fault for invalid PSS parameters fix 6347 6348 The signature verification routines will crash with a NULL pointer 6349 dereference if presented with an ASN.1 signature using the RSA PSS 6350 algorithm and invalid parameters. Since these routines are used to verify 6351 certificate signature algorithms this can be used to crash any 6352 certificate verification operation and exploited in a DoS attack. Any 6353 application which performs certificate verification is vulnerable including 6354 OpenSSL clients and servers which enable client authentication. 6355 6356 This issue was was reported to OpenSSL by Brian Carpenter. 6357 ([CVE-2015-0208]) 6358 6359 *Stephen Henson* 6360 6361 * ASN.1 structure reuse memory corruption fix 6362 6363 Reusing a structure in ASN.1 parsing may allow an attacker to cause 6364 memory corruption via an invalid write. Such reuse is and has been 6365 strongly discouraged and is believed to be rare. 6366 6367 Applications that parse structures containing CHOICE or ANY DEFINED BY 6368 components may be affected. Certificate parsing (d2i_X509 and related 6369 functions) are however not affected. OpenSSL clients and servers are 6370 not affected. 6371 ([CVE-2015-0287]) 6372 6373 *Stephen Henson* 6374 6375 * PKCS7 NULL pointer dereferences fix 6376 6377 The PKCS#7 parsing code does not handle missing outer ContentInfo 6378 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with 6379 missing content and trigger a NULL pointer dereference on parsing. 6380 6381 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or 6382 otherwise parse PKCS#7 structures from untrusted sources are 6383 affected. OpenSSL clients and servers are not affected. 6384 6385 This issue was reported to OpenSSL by Michal Zalewski (Google). 6386 ([CVE-2015-0289]) 6387 6388 *Emilia Käsper* 6389 6390 * DoS via reachable assert in SSLv2 servers fix 6391 6392 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in 6393 servers that both support SSLv2 and enable export cipher suites by sending 6394 a specially crafted SSLv2 CLIENT-MASTER-KEY message. 6395 6396 This issue was discovered by Sean Burford (Google) and Emilia Käsper 6397 (OpenSSL development team). 6398 ([CVE-2015-0293]) 6399 6400 *Emilia Käsper* 6401 6402 * Empty CKE with client auth and DHE fix 6403 6404 If client auth is used then a server can seg fault in the event of a DHE 6405 ciphersuite being selected and a zero length ClientKeyExchange message 6406 being sent by the client. This could be exploited in a DoS attack. 6407 ([CVE-2015-1787]) 6408 6409 *Matt Caswell* 6410 6411 * Handshake with unseeded PRNG fix 6412 6413 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake 6414 with an unseeded PRNG. The conditions are: 6415 - The client is on a platform where the PRNG has not been seeded 6416 automatically, and the user has not seeded manually 6417 - A protocol specific client method version has been used (i.e. not 6418 SSL_client_methodv23) 6419 - A ciphersuite is used that does not require additional random data from 6420 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA). 6421 6422 If the handshake succeeds then the client random that has been used will 6423 have been generated from a PRNG with insufficient entropy and therefore the 6424 output may be predictable. 6425 6426 For example using the following command with an unseeded openssl will 6427 succeed on an unpatched platform: 6428 6429 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA 6430 ([CVE-2015-0285]) 6431 6432 *Matt Caswell* 6433 6434 * Use After Free following d2i_ECPrivatekey error fix 6435 6436 A malformed EC private key file consumed via the d2i_ECPrivateKey function 6437 could cause a use after free condition. This, in turn, could cause a double 6438 free in several private key parsing functions (such as d2i_PrivateKey 6439 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption 6440 for applications that receive EC private keys from untrusted 6441 sources. This scenario is considered rare. 6442 6443 This issue was discovered by the BoringSSL project and fixed in their 6444 commit 517073cd4b. 6445 ([CVE-2015-0209]) 6446 6447 *Matt Caswell* 6448 6449 * X509_to_X509_REQ NULL pointer deref fix 6450 6451 The function X509_to_X509_REQ will crash with a NULL pointer dereference if 6452 the certificate key is invalid. This function is rarely used in practice. 6453 6454 This issue was discovered by Brian Carpenter. 6455 ([CVE-2015-0288]) 6456 6457 *Stephen Henson* 6458 6459 * Removed the export ciphers from the DEFAULT ciphers 6460 6461 *Kurt Roeckx* 6462 6463### Changes between 1.0.1l and 1.0.2 [22 Jan 2015] 6464 6465 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g. 6466 ARMv5 through ARMv8, as opposite to "locking" it to single one. 6467 So far those who have to target multiple platforms would compromise 6468 and argue that binary targeting say ARMv5 would still execute on 6469 ARMv8. "Universal" build resolves this compromise by providing 6470 near-optimal performance even on newer platforms. 6471 6472 *Andy Polyakov* 6473 6474 * Accelerated NIST P-256 elliptic curve implementation for x86_64 6475 (other platforms pending). 6476 6477 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov* 6478 6479 * Add support for the SignedCertificateTimestampList certificate and 6480 OCSP response extensions from RFC6962. 6481 6482 *Rob Stradling* 6483 6484 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 6485 for corner cases. (Certain input points at infinity could lead to 6486 bogus results, with non-infinity inputs mapped to infinity too.) 6487 6488 *Bodo Moeller* 6489 6490 * Initial support for PowerISA 2.0.7, first implemented in POWER8. 6491 This covers AES, SHA256/512 and GHASH. "Initial" means that most 6492 common cases are optimized and there still is room for further 6493 improvements. Vector Permutation AES for Altivec is also added. 6494 6495 *Andy Polyakov* 6496 6497 * Add support for little-endian ppc64 Linux target. 6498 6499 *Marcelo Cerri (IBM)* 6500 6501 * Initial support for AMRv8 ISA crypto extensions. This covers AES, 6502 SHA1, SHA256 and GHASH. "Initial" means that most common cases 6503 are optimized and there still is room for further improvements. 6504 Both 32- and 64-bit modes are supported. 6505 6506 *Andy Polyakov, Ard Biesheuvel (Linaro)* 6507 6508 * Improved ARMv7 NEON support. 6509 6510 *Andy Polyakov* 6511 6512 * Support for SPARC Architecture 2011 crypto extensions, first 6513 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1, 6514 SHA256/512, MD5, GHASH and modular exponentiation. 6515 6516 *Andy Polyakov, David Miller* 6517 6518 * Accelerated modular exponentiation for Intel processors, a.k.a. 6519 RSAZ. 6520 6521 *Shay Gueron & Vlad Krasnov (Intel Corp)* 6522 6523 * Support for new and upcoming Intel processors, including AVX2, 6524 BMI and SHA ISA extensions. This includes additional "stitched" 6525 implementations, AESNI-SHA256 and GCM, and multi-buffer support 6526 for TLS encrypt. 6527 6528 This work was sponsored by Intel Corp. 6529 6530 *Andy Polyakov* 6531 6532 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method() 6533 supports both DTLS 1.2 and 1.0 and should use whatever version the peer 6534 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only. 6535 6536 *Steve Henson* 6537 6538 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file(): 6539 this fixes a limitation in previous versions of OpenSSL. 6540 6541 *Steve Henson* 6542 6543 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest, 6544 MGF1 digest and OAEP label. 6545 6546 *Steve Henson* 6547 6548 * Add EVP support for key wrapping algorithms, to avoid problems with 6549 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in 6550 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap 6551 algorithms and include tests cases. 6552 6553 *Steve Henson* 6554 6555 * Add functions to allocate and set the fields of an ECDSA_METHOD 6556 structure. 6557 6558 *Douglas E. Engert, Steve Henson* 6559 6560 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the 6561 difference in days and seconds between two tm or ASN1_TIME structures. 6562 6563 *Steve Henson* 6564 6565 * Add -rev test option to s_server to just reverse order of characters 6566 received by client and send back to server. Also prints an abbreviated 6567 summary of the connection parameters. 6568 6569 *Steve Henson* 6570 6571 * New option -brief for s_client and s_server to print out a brief summary 6572 of connection parameters. 6573 6574 *Steve Henson* 6575 6576 * Add callbacks for arbitrary TLS extensions. 6577 6578 *Trevor Perrin <trevp@trevp.net> and Ben Laurie* 6579 6580 * New option -crl_download in several openssl utilities to download CRLs 6581 from CRLDP extension in certificates. 6582 6583 *Steve Henson* 6584 6585 * New options -CRL and -CRLform for s_client and s_server for CRLs. 6586 6587 *Steve Henson* 6588 6589 * New function X509_CRL_diff to generate a delta CRL from the difference 6590 of two full CRLs. Add support to "crl" utility. 6591 6592 *Steve Henson* 6593 6594 * New functions to set lookup_crls function and to retrieve 6595 X509_STORE from X509_STORE_CTX. 6596 6597 *Steve Henson* 6598 6599 * Print out deprecated issuer and subject unique ID fields in 6600 certificates. 6601 6602 *Steve Henson* 6603 6604 * Extend OCSP I/O functions so they can be used for simple general purpose 6605 HTTP as well as OCSP. New wrapper function which can be used to download 6606 CRLs using the OCSP API. 6607 6608 *Steve Henson* 6609 6610 * Delegate command line handling in s_client/s_server to SSL_CONF APIs. 6611 6612 *Steve Henson* 6613 6614 * `SSL_CONF*` functions. These provide a common framework for application 6615 configuration using configuration files or command lines. 6616 6617 *Steve Henson* 6618 6619 * SSL/TLS tracing code. This parses out SSL/TLS records using the 6620 message callback and prints the results. Needs compile time option 6621 "enable-ssl-trace". New options to s_client and s_server to enable 6622 tracing. 6623 6624 *Steve Henson* 6625 6626 * New ctrl and macro to retrieve supported points extensions. 6627 Print out extension in s_server and s_client. 6628 6629 *Steve Henson* 6630 6631 * New functions to retrieve certificate signature and signature 6632 OID NID. 6633 6634 *Steve Henson* 6635 6636 * Add functions to retrieve and manipulate the raw cipherlist sent by a 6637 client to OpenSSL. 6638 6639 *Steve Henson* 6640 6641 * New Suite B modes for TLS code. These use and enforce the requirements 6642 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and 6643 only use Suite B curves. The Suite B modes can be set by using the 6644 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring. 6645 6646 *Steve Henson* 6647 6648 * New chain verification flags for Suite B levels of security. Check 6649 algorithms are acceptable when flags are set in X509_verify_cert. 6650 6651 *Steve Henson* 6652 6653 * Make tls1_check_chain return a set of flags indicating checks passed 6654 by a certificate chain. Add additional tests to handle client 6655 certificates: checks for matching certificate type and issuer name 6656 comparison. 6657 6658 *Steve Henson* 6659 6660 * If an attempt is made to use a signature algorithm not in the peer 6661 preference list abort the handshake. If client has no suitable 6662 signature algorithms in response to a certificate request do not 6663 use the certificate. 6664 6665 *Steve Henson* 6666 6667 * If server EC tmp key is not in client preference list abort handshake. 6668 6669 *Steve Henson* 6670 6671 * Add support for certificate stores in CERT structure. This makes it 6672 possible to have different stores per SSL structure or one store in 6673 the parent SSL_CTX. Include distinct stores for certificate chain 6674 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN 6675 to build and store a certificate chain in CERT structure: returning 6676 an error if the chain cannot be built: this will allow applications 6677 to test if a chain is correctly configured. 6678 6679 Note: if the CERT based stores are not set then the parent SSL_CTX 6680 store is used to retain compatibility with existing behaviour. 6681 6682 *Steve Henson* 6683 6684 * New function ssl_set_client_disabled to set a ciphersuite disabled 6685 mask based on the current session, check mask when sending client 6686 hello and checking the requested ciphersuite. 6687 6688 *Steve Henson* 6689 6690 * New ctrls to retrieve and set certificate types in a certificate 6691 request message. Print out received values in s_client. If certificate 6692 types is not set with custom values set sensible values based on 6693 supported signature algorithms. 6694 6695 *Steve Henson* 6696 6697 * Support for distinct client and server supported signature algorithms. 6698 6699 *Steve Henson* 6700 6701 * Add certificate callback. If set this is called whenever a certificate 6702 is required by client or server. An application can decide which 6703 certificate chain to present based on arbitrary criteria: for example 6704 supported signature algorithms. Add very simple example to s_server. 6705 This fixes many of the problems and restrictions of the existing client 6706 certificate callback: for example you can now clear an existing 6707 certificate and specify the whole chain. 6708 6709 *Steve Henson* 6710 6711 * Add new "valid_flags" field to CERT_PKEY structure which determines what 6712 the certificate can be used for (if anything). Set valid_flags field 6713 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used 6714 to have similar checks in it. 6715 6716 Add new "cert_flags" field to CERT structure and include a "strict mode". 6717 This enforces some TLS certificate requirements (such as only permitting 6718 certificate signature algorithms contained in the supported algorithms 6719 extension) which some implementations ignore: this option should be used 6720 with caution as it could cause interoperability issues. 6721 6722 *Steve Henson* 6723 6724 * Update and tidy signature algorithm extension processing. Work out 6725 shared signature algorithms based on preferences and peer algorithms 6726 and print them out in s_client and s_server. Abort handshake if no 6727 shared signature algorithms. 6728 6729 *Steve Henson* 6730 6731 * Add new functions to allow customised supported signature algorithms 6732 for SSL and SSL_CTX structures. Add options to s_client and s_server 6733 to support them. 6734 6735 *Steve Henson* 6736 6737 * New function SSL_certs_clear() to delete all references to certificates 6738 from an SSL structure. Before this once a certificate had been added 6739 it couldn't be removed. 6740 6741 *Steve Henson* 6742 6743 * Integrate hostname, email address and IP address checking with certificate 6744 verification. New verify options supporting checking in openssl utility. 6745 6746 *Steve Henson* 6747 6748 * Fixes and wildcard matching support to hostname and email checking 6749 functions. Add manual page. 6750 6751 *Florian Weimer (Red Hat Product Security Team)* 6752 6753 * New functions to check a hostname email or IP address against a 6754 certificate. Add options x509 utility to print results of checks against 6755 a certificate. 6756 6757 *Steve Henson* 6758 6759 * Fix OCSP checking. 6760 6761 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie* 6762 6763 * Initial experimental support for explicitly trusted non-root CAs. 6764 OpenSSL still tries to build a complete chain to a root but if an 6765 intermediate CA has a trust setting included that is used. The first 6766 setting is used: whether to trust (e.g., -addtrust option to the x509 6767 utility) or reject. 6768 6769 *Steve Henson* 6770 6771 * Add -trusted_first option which attempts to find certificates in the 6772 trusted store even if an untrusted chain is also supplied. 6773 6774 *Steve Henson* 6775 6776 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE, 6777 platform support for Linux and Android. 6778 6779 *Andy Polyakov* 6780 6781 * Support for linux-x32, ILP32 environment in x86_64 framework. 6782 6783 *Andy Polyakov* 6784 6785 * Experimental multi-implementation support for FIPS capable OpenSSL. 6786 When in FIPS mode the approved implementations are used as normal, 6787 when not in FIPS mode the internal unapproved versions are used instead. 6788 This means that the FIPS capable OpenSSL isn't forced to use the 6789 (often lower performance) FIPS implementations outside FIPS mode. 6790 6791 *Steve Henson* 6792 6793 * Transparently support X9.42 DH parameters when calling 6794 PEM_read_bio_DHparameters. This means existing applications can handle 6795 the new parameter format automatically. 6796 6797 *Steve Henson* 6798 6799 * Initial experimental support for X9.42 DH parameter format: mainly 6800 to support use of 'q' parameter for RFC5114 parameters. 6801 6802 *Steve Henson* 6803 6804 * Add DH parameters from RFC5114 including test data to dhtest. 6805 6806 *Steve Henson* 6807 6808 * Support for automatic EC temporary key parameter selection. If enabled 6809 the most preferred EC parameters are automatically used instead of 6810 hardcoded fixed parameters. Now a server just has to call: 6811 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically 6812 support ECDH and use the most appropriate parameters. 6813 6814 *Steve Henson* 6815 6816 * Enhance and tidy EC curve and point format TLS extension code. Use 6817 static structures instead of allocation if default values are used. 6818 New ctrls to set curves we wish to support and to retrieve shared curves. 6819 Print out shared curves in s_server. New options to s_server and s_client 6820 to set list of supported curves. 6821 6822 *Steve Henson* 6823 6824 * New ctrls to retrieve supported signature algorithms and 6825 supported curve values as an array of NIDs. Extend openssl utility 6826 to print out received values. 6827 6828 *Steve Henson* 6829 6830 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert 6831 between NIDs and the more common NIST names such as "P-256". Enhance 6832 ecparam utility and ECC method to recognise the NIST names for curves. 6833 6834 *Steve Henson* 6835 6836 * Enhance SSL/TLS certificate chain handling to support different 6837 chains for each certificate instead of one chain in the parent SSL_CTX. 6838 6839 *Steve Henson* 6840 6841 * Support for fixed DH ciphersuite client authentication: where both 6842 server and client use DH certificates with common parameters. 6843 6844 *Steve Henson* 6845 6846 * Support for fixed DH ciphersuites: those requiring DH server 6847 certificates. 6848 6849 *Steve Henson* 6850 6851 * New function i2d_re_X509_tbs for re-encoding the TBS portion of 6852 the certificate. 6853 Note: Related 1.0.2-beta specific macros X509_get_cert_info, 6854 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and 6855 X509_CINF_get_signature were reverted post internal team review. 6856 6857OpenSSL 1.0.1 6858------------- 6859 6860### Changes between 1.0.1t and 1.0.1u [22 Sep 2016] 6861 6862 * OCSP Status Request extension unbounded memory growth 6863 6864 A malicious client can send an excessively large OCSP Status Request 6865 extension. If that client continually requests renegotiation, sending a 6866 large OCSP Status Request extension each time, then there will be unbounded 6867 memory growth on the server. This will eventually lead to a Denial Of 6868 Service attack through memory exhaustion. Servers with a default 6869 configuration are vulnerable even if they do not support OCSP. Builds using 6870 the "no-ocsp" build time option are not affected. 6871 6872 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 6873 ([CVE-2016-6304]) 6874 6875 *Matt Caswell* 6876 6877 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from 6878 HIGH to MEDIUM. 6879 6880 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan 6881 Leurent (INRIA) 6882 ([CVE-2016-2183]) 6883 6884 *Rich Salz* 6885 6886 * OOB write in MDC2_Update() 6887 6888 An overflow can occur in MDC2_Update() either if called directly or 6889 through the EVP_DigestUpdate() function using MDC2. If an attacker 6890 is able to supply very large amounts of input data after a previous 6891 call to EVP_EncryptUpdate() with a partial block then a length check 6892 can overflow resulting in a heap corruption. 6893 6894 The amount of data needed is comparable to SIZE_MAX which is impractical 6895 on most platforms. 6896 6897 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 6898 ([CVE-2016-6303]) 6899 6900 *Stephen Henson* 6901 6902 * Malformed SHA512 ticket DoS 6903 6904 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a 6905 DoS attack where a malformed ticket will result in an OOB read which will 6906 ultimately crash. 6907 6908 The use of SHA512 in TLS session tickets is comparatively rare as it requires 6909 a custom server callback and ticket lookup mechanism. 6910 6911 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 6912 ([CVE-2016-6302]) 6913 6914 *Stephen Henson* 6915 6916 * OOB write in BN_bn2dec() 6917 6918 The function BN_bn2dec() does not check the return value of BN_div_word(). 6919 This can cause an OOB write if an application uses this function with an 6920 overly large BIGNUM. This could be a problem if an overly large certificate 6921 or CRL is printed out from an untrusted source. TLS is not affected because 6922 record limits will reject an oversized certificate before it is parsed. 6923 6924 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 6925 ([CVE-2016-2182]) 6926 6927 *Stephen Henson* 6928 6929 * OOB read in TS_OBJ_print_bio() 6930 6931 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is 6932 the total length the OID text representation would use and not the amount 6933 of data written. This will result in OOB reads when large OIDs are 6934 presented. 6935 6936 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 6937 ([CVE-2016-2180]) 6938 6939 *Stephen Henson* 6940 6941 * Pointer arithmetic undefined behaviour 6942 6943 Avoid some undefined pointer arithmetic 6944 6945 A common idiom in the codebase is to check limits in the following manner: 6946 "p + len > limit" 6947 6948 Where "p" points to some malloc'd data of SIZE bytes and 6949 limit == p + SIZE 6950 6951 "len" here could be from some externally supplied data (e.g. from a TLS 6952 message). 6953 6954 The rules of C pointer arithmetic are such that "p + len" is only well 6955 defined where len <= SIZE. Therefore the above idiom is actually 6956 undefined behaviour. 6957 6958 For example this could cause problems if some malloc implementation 6959 provides an address for "p" such that "p + len" actually overflows for 6960 values of len that are too big and therefore p + len < limit. 6961 6962 This issue was reported to OpenSSL by Guido Vranken 6963 ([CVE-2016-2177]) 6964 6965 *Matt Caswell* 6966 6967 * Constant time flag not preserved in DSA signing 6968 6969 Operations in the DSA signing algorithm should run in constant time in 6970 order to avoid side channel attacks. A flaw in the OpenSSL DSA 6971 implementation means that a non-constant time codepath is followed for 6972 certain operations. This has been demonstrated through a cache-timing 6973 attack to be sufficient for an attacker to recover the private DSA key. 6974 6975 This issue was reported by César Pereida (Aalto University), Billy Brumley 6976 (Tampere University of Technology), and Yuval Yarom (The University of 6977 Adelaide and NICTA). 6978 ([CVE-2016-2178]) 6979 6980 *César Pereida* 6981 6982 * DTLS buffered message DoS 6983 6984 In a DTLS connection where handshake messages are delivered out-of-order 6985 those messages that OpenSSL is not yet ready to process will be buffered 6986 for later use. Under certain circumstances, a flaw in the logic means that 6987 those messages do not get removed from the buffer even though the handshake 6988 has been completed. An attacker could force up to approx. 15 messages to 6989 remain in the buffer when they are no longer required. These messages will 6990 be cleared when the DTLS connection is closed. The default maximum size for 6991 a message is 100k. Therefore the attacker could force an additional 1500k 6992 to be consumed per connection. By opening many simulataneous connections an 6993 attacker could cause a DoS attack through memory exhaustion. 6994 6995 This issue was reported to OpenSSL by Quan Luo. 6996 ([CVE-2016-2179]) 6997 6998 *Matt Caswell* 6999 7000 * DTLS replay protection DoS 7001 7002 A flaw in the DTLS replay attack protection mechanism means that records 7003 that arrive for future epochs update the replay protection "window" before 7004 the MAC for the record has been validated. This could be exploited by an 7005 attacker by sending a record for the next epoch (which does not have to 7006 decrypt or have a valid MAC), with a very large sequence number. This means 7007 that all subsequent legitimate packets are dropped causing a denial of 7008 service for a specific DTLS connection. 7009 7010 This issue was reported to OpenSSL by the OCAP audit team. 7011 ([CVE-2016-2181]) 7012 7013 *Matt Caswell* 7014 7015 * Certificate message OOB reads 7016 7017 In OpenSSL 1.0.2 and earlier some missing message length checks can result 7018 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a 7019 theoretical DoS risk but this has not been observed in practice on common 7020 platforms. 7021 7022 The messages affected are client certificate, client certificate request 7023 and server certificate. As a result the attack can only be performed 7024 against a client or a server which enables client authentication. 7025 7026 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 7027 ([CVE-2016-6306]) 7028 7029 *Stephen Henson* 7030 7031### Changes between 1.0.1s and 1.0.1t [3 May 2016] 7032 7033 * Prevent padding oracle in AES-NI CBC MAC check 7034 7035 A MITM attacker can use a padding oracle attack to decrypt traffic 7036 when the connection uses an AES CBC cipher and the server support 7037 AES-NI. 7038 7039 This issue was introduced as part of the fix for Lucky 13 padding 7040 attack ([CVE-2013-0169]). The padding check was rewritten to be in 7041 constant time by making sure that always the same bytes are read and 7042 compared against either the MAC or padding bytes. But it no longer 7043 checked that there was enough data to have both the MAC and padding 7044 bytes. 7045 7046 This issue was reported by Juraj Somorovsky using TLS-Attacker. 7047 ([CVE-2016-2107]) 7048 7049 *Kurt Roeckx* 7050 7051 * Fix EVP_EncodeUpdate overflow 7052 7053 An overflow can occur in the EVP_EncodeUpdate() function which is used for 7054 Base64 encoding of binary data. If an attacker is able to supply very large 7055 amounts of input data then a length check can overflow resulting in a heap 7056 corruption. 7057 7058 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by 7059 the `PEM_write_bio*` family of functions. These are mainly used within the 7060 OpenSSL command line applications, so any application which processes data 7061 from an untrusted source and outputs it as a PEM file should be considered 7062 vulnerable to this issue. User applications that call these APIs directly 7063 with large amounts of untrusted data may also be vulnerable. 7064 7065 This issue was reported by Guido Vranken. 7066 ([CVE-2016-2105]) 7067 7068 *Matt Caswell* 7069 7070 * Fix EVP_EncryptUpdate overflow 7071 7072 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker 7073 is able to supply very large amounts of input data after a previous call to 7074 EVP_EncryptUpdate() with a partial block then a length check can overflow 7075 resulting in a heap corruption. Following an analysis of all OpenSSL 7076 internal usage of the EVP_EncryptUpdate() function all usage is one of two 7077 forms. The first form is where the EVP_EncryptUpdate() call is known to be 7078 the first called function after an EVP_EncryptInit(), and therefore that 7079 specific call must be safe. The second form is where the length passed to 7080 EVP_EncryptUpdate() can be seen from the code to be some small value and 7081 therefore there is no possibility of an overflow. Since all instances are 7082 one of these two forms, it is believed that there can be no overflows in 7083 internal code due to this problem. It should be noted that 7084 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths. 7085 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances 7086 of these calls have also been analysed too and it is believed there are no 7087 instances in internal usage where an overflow could occur. 7088 7089 This issue was reported by Guido Vranken. 7090 ([CVE-2016-2106]) 7091 7092 *Matt Caswell* 7093 7094 * Prevent ASN.1 BIO excessive memory allocation 7095 7096 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio() 7097 a short invalid encoding can casuse allocation of large amounts of memory 7098 potentially consuming excessive resources or exhausting memory. 7099 7100 Any application parsing untrusted data through d2i BIO functions is 7101 affected. The memory based functions such as d2i_X509() are *not* affected. 7102 Since the memory based functions are used by the TLS library, TLS 7103 applications are not affected. 7104 7105 This issue was reported by Brian Carpenter. 7106 ([CVE-2016-2109]) 7107 7108 *Stephen Henson* 7109 7110 * EBCDIC overread 7111 7112 ASN1 Strings that are over 1024 bytes can cause an overread in applications 7113 using the X509_NAME_oneline() function on EBCDIC systems. This could result 7114 in arbitrary stack data being returned in the buffer. 7115 7116 This issue was reported by Guido Vranken. 7117 ([CVE-2016-2176]) 7118 7119 *Matt Caswell* 7120 7121 * Modify behavior of ALPN to invoke callback after SNI/servername 7122 callback, such that updates to the SSL_CTX affect ALPN. 7123 7124 *Todd Short* 7125 7126 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the 7127 default. 7128 7129 *Kurt Roeckx* 7130 7131 * Only remove the SSLv2 methods with the no-ssl2-method option. When the 7132 methods are enabled and ssl2 is disabled the methods return NULL. 7133 7134 *Kurt Roeckx* 7135 7136### Changes between 1.0.1r and 1.0.1s [1 Mar 2016] 7137 7138* Disable weak ciphers in SSLv3 and up in default builds of OpenSSL. 7139 Builds that are not configured with "enable-weak-ssl-ciphers" will not 7140 provide any "EXPORT" or "LOW" strength ciphers. 7141 7142 *Viktor Dukhovni* 7143 7144* Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2 7145 is by default disabled at build-time. Builds that are not configured with 7146 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used, 7147 users who want to negotiate SSLv2 via the version-flexible SSLv23_method() 7148 will need to explicitly call either of: 7149 7150 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2); 7151 or 7152 SSL_clear_options(ssl, SSL_OP_NO_SSLv2); 7153 7154 as appropriate. Even if either of those is used, or the application 7155 explicitly uses the version-specific SSLv2_method() or its client and 7156 server variants, SSLv2 ciphers vulnerable to exhaustive search key 7157 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT 7158 ciphers, and SSLv2 56-bit DES are no longer available. 7159 ([CVE-2016-0800]) 7160 7161 *Viktor Dukhovni* 7162 7163 * Fix a double-free in DSA code 7164 7165 A double free bug was discovered when OpenSSL parses malformed DSA private 7166 keys and could lead to a DoS attack or memory corruption for applications 7167 that receive DSA private keys from untrusted sources. This scenario is 7168 considered rare. 7169 7170 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using 7171 libFuzzer. 7172 ([CVE-2016-0705]) 7173 7174 *Stephen Henson* 7175 7176 * Disable SRP fake user seed to address a server memory leak. 7177 7178 Add a new method SRP_VBASE_get1_by_user that handles the seed properly. 7179 7180 SRP_VBASE_get_by_user had inconsistent memory management behaviour. 7181 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user 7182 was changed to ignore the "fake user" SRP seed, even if the seed 7183 is configured. 7184 7185 Users should use SRP_VBASE_get1_by_user instead. Note that in 7186 SRP_VBASE_get1_by_user, caller must free the returned value. Note 7187 also that even though configuring the SRP seed attempts to hide 7188 invalid usernames by continuing the handshake with fake 7189 credentials, this behaviour is not constant time and no strong 7190 guarantees are made that the handshake is indistinguishable from 7191 that of a valid user. 7192 ([CVE-2016-0798]) 7193 7194 *Emilia Käsper* 7195 7196 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption 7197 7198 In the BN_hex2bn function the number of hex digits is calculated using an 7199 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For 7200 large values of `i` this can result in `bn_expand` not allocating any 7201 memory because `i * 4` is negative. This can leave the internal BIGNUM data 7202 field as NULL leading to a subsequent NULL ptr deref. For very large values 7203 of `i`, the calculation `i * 4` could be a positive value smaller than `i`. 7204 In this case memory is allocated to the internal BIGNUM data field, but it 7205 is insufficiently sized leading to heap corruption. A similar issue exists 7206 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn 7207 is ever called by user applications with very large untrusted hex/dec data. 7208 This is anticipated to be a rare occurrence. 7209 7210 All OpenSSL internal usage of these functions use data that is not expected 7211 to be untrusted, e.g. config file data or application command line 7212 arguments. If user developed applications generate config file data based 7213 on untrusted data then it is possible that this could also lead to security 7214 consequences. This is also anticipated to be rare. 7215 7216 This issue was reported to OpenSSL by Guido Vranken. 7217 ([CVE-2016-0797]) 7218 7219 *Matt Caswell* 7220 7221 * Fix memory issues in `BIO_*printf` functions 7222 7223 The internal `fmtstr` function used in processing a "%s" format string in 7224 the `BIO_*printf` functions could overflow while calculating the length of a 7225 string and cause an OOB read when printing very long strings. 7226 7227 Additionally the internal `doapr_outch` function can attempt to write to an 7228 OOB memory location (at an offset from the NULL pointer) in the event of a 7229 memory allocation failure. In 1.0.2 and below this could be caused where 7230 the size of a buffer to be allocated is greater than INT_MAX. E.g. this 7231 could be in processing a very long "%s" format string. Memory leaks can 7232 also occur. 7233 7234 The first issue may mask the second issue dependent on compiler behaviour. 7235 These problems could enable attacks where large amounts of untrusted data 7236 is passed to the `BIO_*printf` functions. If applications use these functions 7237 in this way then they could be vulnerable. OpenSSL itself uses these 7238 functions when printing out human-readable dumps of ASN.1 data. Therefore 7239 applications that print this data could be vulnerable if the data is from 7240 untrusted sources. OpenSSL command line applications could also be 7241 vulnerable where they print out ASN.1 data, or if untrusted data is passed 7242 as command line arguments. 7243 7244 Libssl is not considered directly vulnerable. Additionally certificates etc 7245 received via remote connections via libssl are also unlikely to be able to 7246 trigger these issues because of message size limits enforced within libssl. 7247 7248 This issue was reported to OpenSSL Guido Vranken. 7249 ([CVE-2016-0799]) 7250 7251 *Matt Caswell* 7252 7253 * Side channel attack on modular exponentiation 7254 7255 A side-channel attack was found which makes use of cache-bank conflicts on 7256 the Intel Sandy-Bridge microarchitecture which could lead to the recovery 7257 of RSA keys. The ability to exploit this issue is limited as it relies on 7258 an attacker who has control of code in a thread running on the same 7259 hyper-threaded core as the victim thread which is performing decryptions. 7260 7261 This issue was reported to OpenSSL by Yuval Yarom, The University of 7262 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and 7263 Nadia Heninger, University of Pennsylvania with more information at 7264 <http://cachebleed.info>. 7265 ([CVE-2016-0702]) 7266 7267 *Andy Polyakov* 7268 7269 * Change the req command to generate a 2048-bit RSA/DSA key by default, 7270 if no keysize is specified with default_bits. This fixes an 7271 omission in an earlier change that changed all RSA/DSA key generation 7272 commands to use 2048 bits by default. 7273 7274 *Emilia Käsper* 7275 7276### Changes between 1.0.1q and 1.0.1r [28 Jan 2016] 7277 7278 * Protection for DH small subgroup attacks 7279 7280 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been 7281 switched on by default and cannot be disabled. This could have some 7282 performance impact. 7283 7284 *Matt Caswell* 7285 7286 * SSLv2 doesn't block disabled ciphers 7287 7288 A malicious client can negotiate SSLv2 ciphers that have been disabled on 7289 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have 7290 been disabled, provided that the SSLv2 protocol was not also disabled via 7291 SSL_OP_NO_SSLv2. 7292 7293 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram 7294 and Sebastian Schinzel. 7295 ([CVE-2015-3197]) 7296 7297 *Viktor Dukhovni* 7298 7299 * Reject DH handshakes with parameters shorter than 1024 bits. 7300 7301 *Kurt Roeckx* 7302 7303### Changes between 1.0.1p and 1.0.1q [3 Dec 2015] 7304 7305 * Certificate verify crash with missing PSS parameter 7306 7307 The signature verification routines will crash with a NULL pointer 7308 dereference if presented with an ASN.1 signature using the RSA PSS 7309 algorithm and absent mask generation function parameter. Since these 7310 routines are used to verify certificate signature algorithms this can be 7311 used to crash any certificate verification operation and exploited in a 7312 DoS attack. Any application which performs certificate verification is 7313 vulnerable including OpenSSL clients and servers which enable client 7314 authentication. 7315 7316 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG). 7317 ([CVE-2015-3194]) 7318 7319 *Stephen Henson* 7320 7321 * X509_ATTRIBUTE memory leak 7322 7323 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak 7324 memory. This structure is used by the PKCS#7 and CMS routines so any 7325 application which reads PKCS#7 or CMS data from untrusted sources is 7326 affected. SSL/TLS is not affected. 7327 7328 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using 7329 libFuzzer. 7330 ([CVE-2015-3195]) 7331 7332 *Stephen Henson* 7333 7334 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs. 7335 This changes the decoding behaviour for some invalid messages, 7336 though the change is mostly in the more lenient direction, and 7337 legacy behaviour is preserved as much as possible. 7338 7339 *Emilia Käsper* 7340 7341 * In DSA_generate_parameters_ex, if the provided seed is too short, 7342 use a random seed, as already documented. 7343 7344 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>* 7345 7346### Changes between 1.0.1o and 1.0.1p [9 Jul 2015] 7347 7348 * Alternate chains certificate forgery 7349 7350 During certificate verfification, OpenSSL will attempt to find an 7351 alternative certificate chain if the first attempt to build such a chain 7352 fails. An error in the implementation of this logic can mean that an 7353 attacker could cause certain checks on untrusted certificates to be 7354 bypassed, such as the CA flag, enabling them to use a valid leaf 7355 certificate to act as a CA and "issue" an invalid certificate. 7356 7357 This issue was reported to OpenSSL by Adam Langley/David Benjamin 7358 (Google/BoringSSL). 7359 ([CVE-2015-1793]) 7360 7361 *Matt Caswell* 7362 7363 * Race condition handling PSK identify hint 7364 7365 If PSK identity hints are received by a multi-threaded client then 7366 the values are wrongly updated in the parent SSL_CTX structure. This can 7367 result in a race condition potentially leading to a double free of the 7368 identify hint data. 7369 ([CVE-2015-3196]) 7370 7371 *Stephen Henson* 7372 7373### Changes between 1.0.1n and 1.0.1o [12 Jun 2015] 7374 7375 * Fix HMAC ABI incompatibility. The previous version introduced an ABI 7376 incompatibility in the handling of HMAC. The previous ABI has now been 7377 restored. 7378 7379### Changes between 1.0.1m and 1.0.1n [11 Jun 2015] 7380 7381 * Malformed ECParameters causes infinite loop 7382 7383 When processing an ECParameters structure OpenSSL enters an infinite loop 7384 if the curve specified is over a specially malformed binary polynomial 7385 field. 7386 7387 This can be used to perform denial of service against any 7388 system which processes public keys, certificate requests or 7389 certificates. This includes TLS clients and TLS servers with 7390 client authentication enabled. 7391 7392 This issue was reported to OpenSSL by Joseph Barr-Pixton. 7393 ([CVE-2015-1788]) 7394 7395 *Andy Polyakov* 7396 7397 * Exploitable out-of-bounds read in X509_cmp_time 7398 7399 X509_cmp_time does not properly check the length of the ASN1_TIME 7400 string and can read a few bytes out of bounds. In addition, 7401 X509_cmp_time accepts an arbitrary number of fractional seconds in the 7402 time string. 7403 7404 An attacker can use this to craft malformed certificates and CRLs of 7405 various sizes and potentially cause a segmentation fault, resulting in 7406 a DoS on applications that verify certificates or CRLs. TLS clients 7407 that verify CRLs are affected. TLS clients and servers with client 7408 authentication enabled may be affected if they use custom verification 7409 callbacks. 7410 7411 This issue was reported to OpenSSL by Robert Swiecki (Google), and 7412 independently by Hanno Böck. 7413 ([CVE-2015-1789]) 7414 7415 *Emilia Käsper* 7416 7417 * PKCS7 crash with missing EnvelopedContent 7418 7419 The PKCS#7 parsing code does not handle missing inner EncryptedContent 7420 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs 7421 with missing content and trigger a NULL pointer dereference on parsing. 7422 7423 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 7424 structures from untrusted sources are affected. OpenSSL clients and 7425 servers are not affected. 7426 7427 This issue was reported to OpenSSL by Michal Zalewski (Google). 7428 ([CVE-2015-1790]) 7429 7430 *Emilia Käsper* 7431 7432 * CMS verify infinite loop with unknown hash function 7433 7434 When verifying a signedData message the CMS code can enter an infinite loop 7435 if presented with an unknown hash function OID. This can be used to perform 7436 denial of service against any system which verifies signedData messages using 7437 the CMS code. 7438 This issue was reported to OpenSSL by Johannes Bauer. 7439 ([CVE-2015-1792]) 7440 7441 *Stephen Henson* 7442 7443 * Race condition handling NewSessionTicket 7444 7445 If a NewSessionTicket is received by a multi-threaded client when attempting to 7446 reuse a previous ticket then a race condition can occur potentially leading to 7447 a double free of the ticket data. 7448 ([CVE-2015-1791]) 7449 7450 *Matt Caswell* 7451 7452 * Reject DH handshakes with parameters shorter than 768 bits. 7453 7454 *Kurt Roeckx and Emilia Kasper* 7455 7456 * dhparam: generate 2048-bit parameters by default. 7457 7458 *Kurt Roeckx and Emilia Kasper* 7459 7460### Changes between 1.0.1l and 1.0.1m [19 Mar 2015] 7461 7462 * Segmentation fault in ASN1_TYPE_cmp fix 7463 7464 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is 7465 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check 7466 certificate signature algorithm consistency this can be used to crash any 7467 certificate verification operation and exploited in a DoS attack. Any 7468 application which performs certificate verification is vulnerable including 7469 OpenSSL clients and servers which enable client authentication. 7470 ([CVE-2015-0286]) 7471 7472 *Stephen Henson* 7473 7474 * ASN.1 structure reuse memory corruption fix 7475 7476 Reusing a structure in ASN.1 parsing may allow an attacker to cause 7477 memory corruption via an invalid write. Such reuse is and has been 7478 strongly discouraged and is believed to be rare. 7479 7480 Applications that parse structures containing CHOICE or ANY DEFINED BY 7481 components may be affected. Certificate parsing (d2i_X509 and related 7482 functions) are however not affected. OpenSSL clients and servers are 7483 not affected. 7484 ([CVE-2015-0287]) 7485 7486 *Stephen Henson* 7487 7488 * PKCS7 NULL pointer dereferences fix 7489 7490 The PKCS#7 parsing code does not handle missing outer ContentInfo 7491 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with 7492 missing content and trigger a NULL pointer dereference on parsing. 7493 7494 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or 7495 otherwise parse PKCS#7 structures from untrusted sources are 7496 affected. OpenSSL clients and servers are not affected. 7497 7498 This issue was reported to OpenSSL by Michal Zalewski (Google). 7499 ([CVE-2015-0289]) 7500 7501 *Emilia Käsper* 7502 7503 * DoS via reachable assert in SSLv2 servers fix 7504 7505 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in 7506 servers that both support SSLv2 and enable export cipher suites by sending 7507 a specially crafted SSLv2 CLIENT-MASTER-KEY message. 7508 7509 This issue was discovered by Sean Burford (Google) and Emilia Käsper 7510 (OpenSSL development team). 7511 ([CVE-2015-0293]) 7512 7513 *Emilia Käsper* 7514 7515 * Use After Free following d2i_ECPrivatekey error fix 7516 7517 A malformed EC private key file consumed via the d2i_ECPrivateKey function 7518 could cause a use after free condition. This, in turn, could cause a double 7519 free in several private key parsing functions (such as d2i_PrivateKey 7520 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption 7521 for applications that receive EC private keys from untrusted 7522 sources. This scenario is considered rare. 7523 7524 This issue was discovered by the BoringSSL project and fixed in their 7525 commit 517073cd4b. 7526 ([CVE-2015-0209]) 7527 7528 *Matt Caswell* 7529 7530 * X509_to_X509_REQ NULL pointer deref fix 7531 7532 The function X509_to_X509_REQ will crash with a NULL pointer dereference if 7533 the certificate key is invalid. This function is rarely used in practice. 7534 7535 This issue was discovered by Brian Carpenter. 7536 ([CVE-2015-0288]) 7537 7538 *Stephen Henson* 7539 7540 * Removed the export ciphers from the DEFAULT ciphers 7541 7542 *Kurt Roeckx* 7543 7544### Changes between 1.0.1k and 1.0.1l [15 Jan 2015] 7545 7546 * Build fixes for the Windows and OpenVMS platforms 7547 7548 *Matt Caswell and Richard Levitte* 7549 7550### Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 7551 7552 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 7553 message can cause a segmentation fault in OpenSSL due to a NULL pointer 7554 dereference. This could lead to a Denial Of Service attack. Thanks to 7555 Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 7556 ([CVE-2014-3571]) 7557 7558 *Steve Henson* 7559 7560 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 7561 dtls1_buffer_record function under certain conditions. In particular this 7562 could occur if an attacker sent repeated DTLS records with the same 7563 sequence number but for the next epoch. The memory leak could be exploited 7564 by an attacker in a Denial of Service attack through memory exhaustion. 7565 Thanks to Chris Mueller for reporting this issue. 7566 ([CVE-2015-0206]) 7567 7568 *Matt Caswell* 7569 7570 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 7571 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 7572 method would be set to NULL which could later result in a NULL pointer 7573 dereference. Thanks to Frank Schmirler for reporting this issue. 7574 ([CVE-2014-3569]) 7575 7576 *Kurt Roeckx* 7577 7578 * Abort handshake if server key exchange message is omitted for ephemeral 7579 ECDH ciphersuites. 7580 7581 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 7582 reporting this issue. 7583 ([CVE-2014-3572]) 7584 7585 *Steve Henson* 7586 7587 * Remove non-export ephemeral RSA code on client and server. This code 7588 violated the TLS standard by allowing the use of temporary RSA keys in 7589 non-export ciphersuites and could be used by a server to effectively 7590 downgrade the RSA key length used to a value smaller than the server 7591 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 7592 INRIA or reporting this issue. 7593 ([CVE-2015-0204]) 7594 7595 *Steve Henson* 7596 7597 * Fixed issue where DH client certificates are accepted without verification. 7598 An OpenSSL server will accept a DH certificate for client authentication 7599 without the certificate verify message. This effectively allows a client to 7600 authenticate without the use of a private key. This only affects servers 7601 which trust a client certificate authority which issues certificates 7602 containing DH keys: these are extremely rare and hardly ever encountered. 7603 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 7604 this issue. 7605 ([CVE-2015-0205]) 7606 7607 *Steve Henson* 7608 7609 * Ensure that the session ID context of an SSL is updated when its 7610 SSL_CTX is updated via SSL_set_SSL_CTX. 7611 7612 The session ID context is typically set from the parent SSL_CTX, 7613 and can vary with the CTX. 7614 7615 *Adam Langley* 7616 7617 * Fix various certificate fingerprint issues. 7618 7619 By using non-DER or invalid encodings outside the signed portion of a 7620 certificate the fingerprint can be changed without breaking the signature. 7621 Although no details of the signed portion of the certificate can be changed 7622 this can cause problems with some applications: e.g. those using the 7623 certificate fingerprint for blacklists. 7624 7625 1. Reject signatures with non zero unused bits. 7626 7627 If the BIT STRING containing the signature has non zero unused bits reject 7628 the signature. All current signature algorithms require zero unused bits. 7629 7630 2. Check certificate algorithm consistency. 7631 7632 Check the AlgorithmIdentifier inside TBS matches the one in the 7633 certificate signature. NB: this will result in signature failure 7634 errors for some broken certificates. 7635 7636 Thanks to Konrad Kraszewski from Google for reporting this issue. 7637 7638 3. Check DSA/ECDSA signatures use DER. 7639 7640 Re-encode DSA/ECDSA signatures and compare with the original received 7641 signature. Return an error if there is a mismatch. 7642 7643 This will reject various cases including garbage after signature 7644 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 7645 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 7646 (negative or with leading zeroes). 7647 7648 Further analysis was conducted and fixes were developed by Stephen Henson 7649 of the OpenSSL core team. 7650 7651 ([CVE-2014-8275]) 7652 7653 *Steve Henson* 7654 7655 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 7656 results on some platforms, including x86_64. This bug occurs at random 7657 with a very low probability, and is not known to be exploitable in any 7658 way, though its exact impact is difficult to determine. Thanks to Pieter 7659 Wuille (Blockstream) who reported this issue and also suggested an initial 7660 fix. Further analysis was conducted by the OpenSSL development team and 7661 Adam Langley of Google. The final fix was developed by Andy Polyakov of 7662 the OpenSSL core team. 7663 ([CVE-2014-3570]) 7664 7665 *Andy Polyakov* 7666 7667 * Do not resume sessions on the server if the negotiated protocol 7668 version does not match the session's version. Resuming with a different 7669 version, while not strictly forbidden by the RFC, is of questionable 7670 sanity and breaks all known clients. 7671 7672 *David Benjamin, Emilia Käsper* 7673 7674 * Tighten handling of the ChangeCipherSpec (CCS) message: reject 7675 early CCS messages during renegotiation. (Note that because 7676 renegotiation is encrypted, this early CCS was not exploitable.) 7677 7678 *Emilia Käsper* 7679 7680 * Tighten client-side session ticket handling during renegotiation: 7681 ensure that the client only accepts a session ticket if the server sends 7682 the extension anew in the ServerHello. Previously, a TLS client would 7683 reuse the old extension state and thus accept a session ticket if one was 7684 announced in the initial ServerHello. 7685 7686 Similarly, ensure that the client requires a session ticket if one 7687 was advertised in the ServerHello. Previously, a TLS client would 7688 ignore a missing NewSessionTicket message. 7689 7690 *Emilia Käsper* 7691 7692### Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 7693 7694 * SRTP Memory Leak. 7695 7696 A flaw in the DTLS SRTP extension parsing code allows an attacker, who 7697 sends a carefully crafted handshake message, to cause OpenSSL to fail 7698 to free up to 64k of memory causing a memory leak. This could be 7699 exploited in a Denial Of Service attack. This issue affects OpenSSL 7700 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 7701 whether SRTP is used or configured. Implementations of OpenSSL that 7702 have been compiled with OPENSSL_NO_SRTP defined are not affected. 7703 7704 The fix was developed by the OpenSSL team. 7705 ([CVE-2014-3513]) 7706 7707 *OpenSSL team* 7708 7709 * Session Ticket Memory Leak. 7710 7711 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 7712 integrity of that ticket is first verified. In the event of a session 7713 ticket integrity check failing, OpenSSL will fail to free memory 7714 causing a memory leak. By sending a large number of invalid session 7715 tickets an attacker could exploit this issue in a Denial Of Service 7716 attack. 7717 ([CVE-2014-3567]) 7718 7719 *Steve Henson* 7720 7721 * Build option no-ssl3 is incomplete. 7722 7723 When OpenSSL is configured with "no-ssl3" as a build option, servers 7724 could accept and complete a SSL 3.0 handshake, and clients could be 7725 configured to send them. 7726 ([CVE-2014-3568]) 7727 7728 *Akamai and the OpenSSL team* 7729 7730 * Add support for TLS_FALLBACK_SCSV. 7731 Client applications doing fallback retries should call 7732 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 7733 ([CVE-2014-3566]) 7734 7735 *Adam Langley, Bodo Moeller* 7736 7737 * Add additional DigestInfo checks. 7738 7739 Re-encode DigestInto in DER and check against the original when 7740 verifying RSA signature: this will reject any improperly encoded 7741 DigestInfo structures. 7742 7743 Note: this is a precautionary measure and no attacks are currently known. 7744 7745 *Steve Henson* 7746 7747### Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 7748 7749 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 7750 SRP code can be overrun an internal buffer. Add sanity check that 7751 g, A, B < N to SRP code. 7752 7753 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 7754 Group for discovering this issue. 7755 ([CVE-2014-3512]) 7756 7757 *Steve Henson* 7758 7759 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 7760 TLS 1.0 instead of higher protocol versions when the ClientHello message 7761 is badly fragmented. This allows a man-in-the-middle attacker to force a 7762 downgrade to TLS 1.0 even if both the server and the client support a 7763 higher protocol version, by modifying the client's TLS records. 7764 7765 Thanks to David Benjamin and Adam Langley (Google) for discovering and 7766 researching this issue. 7767 ([CVE-2014-3511]) 7768 7769 *David Benjamin* 7770 7771 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 7772 to a denial of service attack. A malicious server can crash the client 7773 with a null pointer dereference (read) by specifying an anonymous (EC)DH 7774 ciphersuite and sending carefully crafted handshake messages. 7775 7776 Thanks to Felix Gröbert (Google) for discovering and researching this 7777 issue. 7778 ([CVE-2014-3510]) 7779 7780 *Emilia Käsper* 7781 7782 * By sending carefully crafted DTLS packets an attacker could cause openssl 7783 to leak memory. This can be exploited through a Denial of Service attack. 7784 Thanks to Adam Langley for discovering and researching this issue. 7785 ([CVE-2014-3507]) 7786 7787 *Adam Langley* 7788 7789 * An attacker can force openssl to consume large amounts of memory whilst 7790 processing DTLS handshake messages. This can be exploited through a 7791 Denial of Service attack. 7792 Thanks to Adam Langley for discovering and researching this issue. 7793 ([CVE-2014-3506]) 7794 7795 *Adam Langley* 7796 7797 * An attacker can force an error condition which causes openssl to crash 7798 whilst processing DTLS packets due to memory being freed twice. This 7799 can be exploited through a Denial of Service attack. 7800 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 7801 this issue. 7802 ([CVE-2014-3505]) 7803 7804 *Adam Langley* 7805 7806 * If a multithreaded client connects to a malicious server using a resumed 7807 session and the server sends an ec point format extension it could write 7808 up to 255 bytes to freed memory. 7809 7810 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 7811 issue. 7812 ([CVE-2014-3509]) 7813 7814 *Gabor Tyukasz* 7815 7816 * A malicious server can crash an OpenSSL client with a null pointer 7817 dereference (read) by specifying an SRP ciphersuite even though it was not 7818 properly negotiated with the client. This can be exploited through a 7819 Denial of Service attack. 7820 7821 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for 7822 discovering and researching this issue. 7823 ([CVE-2014-5139]) 7824 7825 *Steve Henson* 7826 7827 * A flaw in OBJ_obj2txt may cause pretty printing functions such as 7828 X509_name_oneline, X509_name_print_ex et al. to leak some information 7829 from the stack. Applications may be affected if they echo pretty printing 7830 output to the attacker. 7831 7832 Thanks to Ivan Fratric (Google) for discovering this issue. 7833 ([CVE-2014-3508]) 7834 7835 *Emilia Käsper, and Steve Henson* 7836 7837 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 7838 for corner cases. (Certain input points at infinity could lead to 7839 bogus results, with non-infinity inputs mapped to infinity too.) 7840 7841 *Bodo Moeller* 7842 7843### Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 7844 7845 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 7846 handshake can force the use of weak keying material in OpenSSL 7847 SSL/TLS clients and servers. 7848 7849 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 7850 researching this issue. ([CVE-2014-0224]) 7851 7852 *KIKUCHI Masashi, Steve Henson* 7853 7854 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 7855 OpenSSL DTLS client the code can be made to recurse eventually crashing 7856 in a DoS attack. 7857 7858 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 7859 ([CVE-2014-0221]) 7860 7861 *Imre Rad, Steve Henson* 7862 7863 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 7864 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 7865 client or server. This is potentially exploitable to run arbitrary 7866 code on a vulnerable client or server. 7867 7868 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195]) 7869 7870 *Jüri Aedla, Steve Henson* 7871 7872 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 7873 are subject to a denial of service attack. 7874 7875 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering 7876 this issue. ([CVE-2014-3470]) 7877 7878 *Felix Gröbert, Ivan Fratric, Steve Henson* 7879 7880 * Harmonize version and its documentation. -f flag is used to display 7881 compilation flags. 7882 7883 *mancha <mancha1@zoho.com>* 7884 7885 * Fix eckey_priv_encode so it immediately returns an error upon a failure 7886 in i2d_ECPrivateKey. 7887 7888 *mancha <mancha1@zoho.com>* 7889 7890 * Fix some double frees. These are not thought to be exploitable. 7891 7892 *mancha <mancha1@zoho.com>* 7893 7894### Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 7895 7896 * A missing bounds check in the handling of the TLS heartbeat extension 7897 can be used to reveal up to 64k of memory to a connected client or 7898 server. 7899 7900 Thanks for Neel Mehta of Google Security for discovering this bug and to 7901 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 7902 preparing the fix ([CVE-2014-0160]) 7903 7904 *Adam Langley, Bodo Moeller* 7905 7906 * Fix for the attack described in the paper "Recovering OpenSSL 7907 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 7908 by Yuval Yarom and Naomi Benger. Details can be obtained from: 7909 <http://eprint.iacr.org/2014/140> 7910 7911 Thanks to Yuval Yarom and Naomi Benger for discovering this 7912 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076]) 7913 7914 *Yuval Yarom and Naomi Benger* 7915 7916 * TLS pad extension: draft-agl-tls-padding-03 7917 7918 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 7919 TLS client Hello record length value would otherwise be > 255 and 7920 less that 512 pad with a dummy extension containing zeroes so it 7921 is at least 512 bytes long. 7922 7923 *Adam Langley, Steve Henson* 7924 7925### Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 7926 7927 * Fix for TLS record tampering bug. A carefully crafted invalid 7928 handshake could crash OpenSSL with a NULL pointer exception. 7929 Thanks to Anton Johansson for reporting this issues. 7930 ([CVE-2013-4353]) 7931 7932 * Keep original DTLS digest and encryption contexts in retransmission 7933 structures so we can use the previous session parameters if they need 7934 to be resent. ([CVE-2013-6450]) 7935 7936 *Steve Henson* 7937 7938 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 7939 avoids preferring ECDHE-ECDSA ciphers when the client appears to be 7940 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 7941 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 7942 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 7943 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 7944 7945 *Rob Stradling, Adam Langley* 7946 7947### Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 7948 7949 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI 7950 supporting platforms or when small records were transferred. 7951 7952 *Andy Polyakov, Steve Henson* 7953 7954### Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 7955 7956 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 7957 7958 This addresses the flaw in CBC record processing discovered by 7959 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 7960 at: <http://www.isg.rhul.ac.uk/tls/> 7961 7962 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 7963 Security Group at Royal Holloway, University of London 7964 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 7965 Emilia Käsper for the initial patch. 7966 ([CVE-2013-0169]) 7967 7968 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson* 7969 7970 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 7971 ciphersuites which can be exploited in a denial of service attack. 7972 Thanks go to and to Adam Langley <agl@chromium.org> for discovering 7973 and detecting this bug and to Wolfgang Ettlinger 7974 <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 7975 ([CVE-2012-2686]) 7976 7977 *Adam Langley* 7978 7979 * Return an error when checking OCSP signatures when key is NULL. 7980 This fixes a DoS attack. ([CVE-2013-0166]) 7981 7982 *Steve Henson* 7983 7984 * Make openssl verify return errors. 7985 7986 *Chris Palmer <palmer@google.com> and Ben Laurie* 7987 7988 * Call OCSP Stapling callback after ciphersuite has been chosen, so 7989 the right response is stapled. Also change SSL_get_certificate() 7990 so it returns the certificate actually sent. 7991 See <http://rt.openssl.org/Ticket/Display.html?id=2836>. 7992 7993 *Rob Stradling <rob.stradling@comodo.com>* 7994 7995 * Fix possible deadlock when decoding public keys. 7996 7997 *Steve Henson* 7998 7999 * Don't use TLS 1.0 record version number in initial client hello 8000 if renegotiating. 8001 8002 *Steve Henson* 8003 8004### Changes between 1.0.1b and 1.0.1c [10 May 2012] 8005 8006 * Sanity check record length before skipping explicit IV in TLS 8007 1.2, 1.1 and DTLS to fix DoS attack. 8008 8009 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 8010 fuzzing as a service testing platform. 8011 ([CVE-2012-2333]) 8012 8013 *Steve Henson* 8014 8015 * Initialise tkeylen properly when encrypting CMS messages. 8016 Thanks to Solar Designer of Openwall for reporting this issue. 8017 8018 *Steve Henson* 8019 8020 * In FIPS mode don't try to use composite ciphers as they are not 8021 approved. 8022 8023 *Steve Henson* 8024 8025### Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 8026 8027 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 8028 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 8029 mean any application compiled against OpenSSL 1.0.0 headers setting 8030 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling 8031 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 8032 0x10000000L Any application which was previously compiled against 8033 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 8034 will need to be recompiled as a result. Letting be results in 8035 inability to disable specifically TLS 1.1 and in client context, 8036 in unlike event, limit maximum offered version to TLS 1.0 [see below]. 8037 8038 *Steve Henson* 8039 8040 * In order to ensure interoperability SSL_OP_NO_protocolX does not 8041 disable just protocol X, but all protocols above X *if* there are 8042 protocols *below* X still enabled. In more practical terms it means 8043 that if application wants to disable TLS1.0 in favor of TLS1.1 and 8044 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass 8045 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to 8046 client side. 8047 8048 *Andy Polyakov* 8049 8050### Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 8051 8052 * Check for potentially exploitable overflows in asn1_d2i_read_bio 8053 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 8054 in CRYPTO_realloc_clean. 8055 8056 Thanks to Tavis Ormandy, Google Security Team, for discovering this 8057 issue and to Adam Langley <agl@chromium.org> for fixing it. 8058 ([CVE-2012-2110]) 8059 8060 *Adam Langley (Google), Tavis Ormandy, Google Security Team* 8061 8062 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 8063 8064 *Adam Langley* 8065 8066 * Workarounds for some broken servers that "hang" if a client hello 8067 record length exceeds 255 bytes. 8068 8069 1. Do not use record version number > TLS 1.0 in initial client 8070 hello: some (but not all) hanging servers will now work. 8071 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 8072 the number of ciphers sent in the client hello. This should be 8073 set to an even number, such as 50, for example by passing: 8074 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 8075 Most broken servers should now work. 8076 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 8077 TLS 1.2 client support entirely. 8078 8079 *Steve Henson* 8080 8081 * Fix SEGV in Vector Permutation AES module observed in OpenSSH. 8082 8083 *Andy Polyakov* 8084 8085### Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 8086 8087 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET 8088 STRING form instead of a DigestInfo. 8089 8090 *Steve Henson* 8091 8092 * The format used for MDC2 RSA signatures is inconsistent between EVP 8093 and the RSA_sign/RSA_verify functions. This was made more apparent when 8094 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 8095 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 8096 the correct format in RSA_verify so both forms transparently work. 8097 8098 *Steve Henson* 8099 8100 * Some servers which support TLS 1.0 can choke if we initially indicate 8101 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 8102 encrypted premaster secret. As a workaround use the maximum permitted 8103 client version in client hello, this should keep such servers happy 8104 and still work with previous versions of OpenSSL. 8105 8106 *Steve Henson* 8107 8108 * Add support for TLS/DTLS heartbeats. 8109 8110 *Robin Seggelmann <seggelmann@fh-muenster.de>* 8111 8112 * Add support for SCTP. 8113 8114 *Robin Seggelmann <seggelmann@fh-muenster.de>* 8115 8116 * Improved PRNG seeding for VOS. 8117 8118 *Paul Green <Paul.Green@stratus.com>* 8119 8120 * Extensive assembler packs updates, most notably: 8121 8122 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 8123 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 8124 - x86_64: bit-sliced AES implementation; 8125 - ARM: NEON support, contemporary platforms optimizations; 8126 - s390x: z196 support; 8127 - `*`: GHASH and GF(2^m) multiplication implementations; 8128 8129 *Andy Polyakov* 8130 8131 * Make TLS-SRP code conformant with RFC 5054 API cleanup 8132 (removal of unnecessary code) 8133 8134 *Peter Sylvester <peter.sylvester@edelweb.fr>* 8135 8136 * Add TLS key material exporter from RFC 5705. 8137 8138 *Eric Rescorla* 8139 8140 * Add DTLS-SRTP negotiation from RFC 5764. 8141 8142 *Eric Rescorla* 8143 8144 * Add Next Protocol Negotiation, 8145 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be 8146 disabled with a no-npn flag to config or Configure. Code donated 8147 by Google. 8148 8149 *Adam Langley <agl@google.com> and Ben Laurie* 8150 8151 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 8152 NIST-P256, NIST-P521, with constant-time single point multiplication on 8153 typical inputs. Compiler support for the nonstandard type `__uint128_t` is 8154 required to use this (present in gcc 4.4 and later, for 64-bit builds). 8155 Code made available under Apache License version 2.0. 8156 8157 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 8158 line to include this in your build of OpenSSL, and run "make depend" (or 8159 "make update"). This enables the following EC_METHODs: 8160 8161 EC_GFp_nistp224_method() 8162 EC_GFp_nistp256_method() 8163 EC_GFp_nistp521_method() 8164 8165 EC_GROUP_new_by_curve_name() will automatically use these (while 8166 EC_GROUP_new_curve_GFp() currently prefers the more flexible 8167 implementations). 8168 8169 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)* 8170 8171 * Use type ossl_ssize_t instead of ssize_t which isn't available on 8172 all platforms. Move ssize_t definition from e_os.h to the public 8173 header file e_os2.h as it now appears in public header file cms.h 8174 8175 *Steve Henson* 8176 8177 * New -sigopt option to the ca, req and x509 utilities. Additional 8178 signature parameters can be passed using this option and in 8179 particular PSS. 8180 8181 *Steve Henson* 8182 8183 * Add RSA PSS signing function. This will generate and set the 8184 appropriate AlgorithmIdentifiers for PSS based on those in the 8185 corresponding EVP_MD_CTX structure. No application support yet. 8186 8187 *Steve Henson* 8188 8189 * Support for companion algorithm specific ASN1 signing routines. 8190 New function ASN1_item_sign_ctx() signs a pre-initialised 8191 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 8192 the appropriate parameters. 8193 8194 *Steve Henson* 8195 8196 * Add new algorithm specific ASN1 verification initialisation function 8197 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 8198 handling will be the same no matter what EVP_PKEY_METHOD is used. 8199 Add a PSS handler to support verification of PSS signatures: checked 8200 against a number of sample certificates. 8201 8202 *Steve Henson* 8203 8204 * Add signature printing for PSS. Add PSS OIDs. 8205 8206 *Steve Henson, Martin Kaiser <lists@kaiser.cx>* 8207 8208 * Add algorithm specific signature printing. An individual ASN1 method 8209 can now print out signatures instead of the standard hex dump. 8210 8211 More complex signatures (e.g. PSS) can print out more meaningful 8212 information. Include DSA version that prints out the signature 8213 parameters r, s. 8214 8215 *Steve Henson* 8216 8217 * Password based recipient info support for CMS library: implementing 8218 RFC3211. 8219 8220 *Steve Henson* 8221 8222 * Split password based encryption into PBES2 and PBKDF2 functions. This 8223 neatly separates the code into cipher and PBE sections and is required 8224 for some algorithms that split PBES2 into separate pieces (such as 8225 password based CMS). 8226 8227 *Steve Henson* 8228 8229 * Session-handling fixes: 8230 - Fix handling of connections that are resuming with a session ID, 8231 but also support Session Tickets. 8232 - Fix a bug that suppressed issuing of a new ticket if the client 8233 presented a ticket with an expired session. 8234 - Try to set the ticket lifetime hint to something reasonable. 8235 - Make tickets shorter by excluding irrelevant information. 8236 - On the client side, don't ignore renewed tickets. 8237 8238 *Adam Langley, Bodo Moeller (Google)* 8239 8240 * Fix PSK session representation. 8241 8242 *Bodo Moeller* 8243 8244 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 8245 8246 This work was sponsored by Intel. 8247 8248 *Andy Polyakov* 8249 8250 * Add GCM support to TLS library. Some custom code is needed to split 8251 the IV between the fixed (from PRF) and explicit (from TLS record) 8252 portions. This adds all GCM ciphersuites supported by RFC5288 and 8253 RFC5289. Generalise some `AES*` cipherstrings to include GCM and 8254 add a special AESGCM string for GCM only. 8255 8256 *Steve Henson* 8257 8258 * Expand range of ctrls for AES GCM. Permit setting invocation 8259 field on decrypt and retrieval of invocation field only on encrypt. 8260 8261 *Steve Henson* 8262 8263 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 8264 As required by RFC5289 these ciphersuites cannot be used if for 8265 versions of TLS earlier than 1.2. 8266 8267 *Steve Henson* 8268 8269 * For FIPS capable OpenSSL interpret a NULL default public key method 8270 as unset and return the appropriate default but do *not* set the default. 8271 This means we can return the appropriate method in applications that 8272 switch between FIPS and non-FIPS modes. 8273 8274 *Steve Henson* 8275 8276 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 8277 ENGINE is used then we cannot handle that in the FIPS module so we 8278 keep original code iff non-FIPS operations are allowed. 8279 8280 *Steve Henson* 8281 8282 * Add -attime option to openssl utilities. 8283 8284 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson* 8285 8286 * Redirect DSA and DH operations to FIPS module in FIPS mode. 8287 8288 *Steve Henson* 8289 8290 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 8291 FIPS EC methods unconditionally for now. 8292 8293 *Steve Henson* 8294 8295 * New build option no-ec2m to disable characteristic 2 code. 8296 8297 *Steve Henson* 8298 8299 * Backport libcrypto audit of return value checking from 1.1.0-dev; not 8300 all cases can be covered as some introduce binary incompatibilities. 8301 8302 *Steve Henson* 8303 8304 * Redirect RSA operations to FIPS module including keygen, 8305 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 8306 8307 *Steve Henson* 8308 8309 * Add similar low-level API blocking to ciphers. 8310 8311 *Steve Henson* 8312 8313 * low-level digest APIs are not approved in FIPS mode: any attempt 8314 to use these will cause a fatal error. Applications that *really* want 8315 to use them can use the `private_*` version instead. 8316 8317 *Steve Henson* 8318 8319 * Redirect cipher operations to FIPS module for FIPS builds. 8320 8321 *Steve Henson* 8322 8323 * Redirect digest operations to FIPS module for FIPS builds. 8324 8325 *Steve Henson* 8326 8327 * Update build system to add "fips" flag which will link in fipscanister.o 8328 for static and shared library builds embedding a signature if needed. 8329 8330 *Steve Henson* 8331 8332 * Output TLS supported curves in preference order instead of numerical 8333 order. This is currently hardcoded for the highest order curves first. 8334 This should be configurable so applications can judge speed vs strength. 8335 8336 *Steve Henson* 8337 8338 * Add TLS v1.2 server support for client authentication. 8339 8340 *Steve Henson* 8341 8342 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 8343 and enable MD5. 8344 8345 *Steve Henson* 8346 8347 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying 8348 FIPS modules versions. 8349 8350 *Steve Henson* 8351 8352 * Add TLS v1.2 client side support for client authentication. Keep cache 8353 of handshake records longer as we don't know the hash algorithm to use 8354 until after the certificate request message is received. 8355 8356 *Steve Henson* 8357 8358 * Initial TLS v1.2 client support. Add a default signature algorithms 8359 extension including all the algorithms we support. Parse new signature 8360 format in client key exchange. Relax some ECC signing restrictions for 8361 TLS v1.2 as indicated in RFC5246. 8362 8363 *Steve Henson* 8364 8365 * Add server support for TLS v1.2 signature algorithms extension. Switch 8366 to new signature format when needed using client digest preference. 8367 All server ciphersuites should now work correctly in TLS v1.2. No client 8368 support yet and no support for client certificates. 8369 8370 *Steve Henson* 8371 8372 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 8373 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 8374 ciphersuites. At present only RSA key exchange ciphersuites work with 8375 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 8376 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 8377 and version checking. 8378 8379 *Steve Henson* 8380 8381 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 8382 with this defined it will not be affected by any changes to ssl internal 8383 structures. Add several utility functions to allow openssl application 8384 to work with OPENSSL_NO_SSL_INTERN defined. 8385 8386 *Steve Henson* 8387 8388 * A long standing patch to add support for SRP from EdelWeb (Peter 8389 Sylvester and Christophe Renou) was integrated. 8390 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester 8391 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and 8392 Ben Laurie* 8393 8394 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 8395 8396 *Steve Henson* 8397 8398 * Permit abbreviated handshakes when renegotiating using the function 8399 SSL_renegotiate_abbreviated(). 8400 8401 *Robin Seggelmann <seggelmann@fh-muenster.de>* 8402 8403 * Add call to ENGINE_register_all_complete() to 8404 ENGINE_load_builtin_engines(), so some implementations get used 8405 automatically instead of needing explicit application support. 8406 8407 *Steve Henson* 8408 8409 * Add support for TLS key exporter as described in RFC5705. 8410 8411 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson* 8412 8413 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 8414 a few changes are required: 8415 8416 Add SSL_OP_NO_TLSv1_1 flag. 8417 Add TLSv1_1 methods. 8418 Update version checking logic to handle version 1.1. 8419 Add explicit IV handling (ported from DTLS code). 8420 Add command line options to s_client/s_server. 8421 8422 *Steve Henson* 8423 8424OpenSSL 1.0.0 8425------------- 8426 8427### Changes between 1.0.0s and 1.0.0t [3 Dec 2015] 8428 8429 * X509_ATTRIBUTE memory leak 8430 8431 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak 8432 memory. This structure is used by the PKCS#7 and CMS routines so any 8433 application which reads PKCS#7 or CMS data from untrusted sources is 8434 affected. SSL/TLS is not affected. 8435 8436 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using 8437 libFuzzer. 8438 ([CVE-2015-3195]) 8439 8440 *Stephen Henson* 8441 8442 * Race condition handling PSK identify hint 8443 8444 If PSK identity hints are received by a multi-threaded client then 8445 the values are wrongly updated in the parent SSL_CTX structure. This can 8446 result in a race condition potentially leading to a double free of the 8447 identify hint data. 8448 ([CVE-2015-3196]) 8449 8450 *Stephen Henson* 8451 8452### Changes between 1.0.0r and 1.0.0s [11 Jun 2015] 8453 8454 * Malformed ECParameters causes infinite loop 8455 8456 When processing an ECParameters structure OpenSSL enters an infinite loop 8457 if the curve specified is over a specially malformed binary polynomial 8458 field. 8459 8460 This can be used to perform denial of service against any 8461 system which processes public keys, certificate requests or 8462 certificates. This includes TLS clients and TLS servers with 8463 client authentication enabled. 8464 8465 This issue was reported to OpenSSL by Joseph Barr-Pixton. 8466 ([CVE-2015-1788]) 8467 8468 *Andy Polyakov* 8469 8470 * Exploitable out-of-bounds read in X509_cmp_time 8471 8472 X509_cmp_time does not properly check the length of the ASN1_TIME 8473 string and can read a few bytes out of bounds. In addition, 8474 X509_cmp_time accepts an arbitrary number of fractional seconds in the 8475 time string. 8476 8477 An attacker can use this to craft malformed certificates and CRLs of 8478 various sizes and potentially cause a segmentation fault, resulting in 8479 a DoS on applications that verify certificates or CRLs. TLS clients 8480 that verify CRLs are affected. TLS clients and servers with client 8481 authentication enabled may be affected if they use custom verification 8482 callbacks. 8483 8484 This issue was reported to OpenSSL by Robert Swiecki (Google), and 8485 independently by Hanno Böck. 8486 ([CVE-2015-1789]) 8487 8488 *Emilia Käsper* 8489 8490 * PKCS7 crash with missing EnvelopedContent 8491 8492 The PKCS#7 parsing code does not handle missing inner EncryptedContent 8493 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs 8494 with missing content and trigger a NULL pointer dereference on parsing. 8495 8496 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 8497 structures from untrusted sources are affected. OpenSSL clients and 8498 servers are not affected. 8499 8500 This issue was reported to OpenSSL by Michal Zalewski (Google). 8501 ([CVE-2015-1790]) 8502 8503 *Emilia Käsper* 8504 8505 * CMS verify infinite loop with unknown hash function 8506 8507 When verifying a signedData message the CMS code can enter an infinite loop 8508 if presented with an unknown hash function OID. This can be used to perform 8509 denial of service against any system which verifies signedData messages using 8510 the CMS code. 8511 This issue was reported to OpenSSL by Johannes Bauer. 8512 ([CVE-2015-1792]) 8513 8514 *Stephen Henson* 8515 8516 * Race condition handling NewSessionTicket 8517 8518 If a NewSessionTicket is received by a multi-threaded client when attempting to 8519 reuse a previous ticket then a race condition can occur potentially leading to 8520 a double free of the ticket data. 8521 ([CVE-2015-1791]) 8522 8523 *Matt Caswell* 8524 8525### Changes between 1.0.0q and 1.0.0r [19 Mar 2015] 8526 8527 * Segmentation fault in ASN1_TYPE_cmp fix 8528 8529 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is 8530 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check 8531 certificate signature algorithm consistency this can be used to crash any 8532 certificate verification operation and exploited in a DoS attack. Any 8533 application which performs certificate verification is vulnerable including 8534 OpenSSL clients and servers which enable client authentication. 8535 ([CVE-2015-0286]) 8536 8537 *Stephen Henson* 8538 8539 * ASN.1 structure reuse memory corruption fix 8540 8541 Reusing a structure in ASN.1 parsing may allow an attacker to cause 8542 memory corruption via an invalid write. Such reuse is and has been 8543 strongly discouraged and is believed to be rare. 8544 8545 Applications that parse structures containing CHOICE or ANY DEFINED BY 8546 components may be affected. Certificate parsing (d2i_X509 and related 8547 functions) are however not affected. OpenSSL clients and servers are 8548 not affected. 8549 ([CVE-2015-0287]) 8550 8551 *Stephen Henson* 8552 8553 * PKCS7 NULL pointer dereferences fix 8554 8555 The PKCS#7 parsing code does not handle missing outer ContentInfo 8556 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with 8557 missing content and trigger a NULL pointer dereference on parsing. 8558 8559 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or 8560 otherwise parse PKCS#7 structures from untrusted sources are 8561 affected. OpenSSL clients and servers are not affected. 8562 8563 This issue was reported to OpenSSL by Michal Zalewski (Google). 8564 ([CVE-2015-0289]) 8565 8566 *Emilia Käsper* 8567 8568 * DoS via reachable assert in SSLv2 servers fix 8569 8570 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in 8571 servers that both support SSLv2 and enable export cipher suites by sending 8572 a specially crafted SSLv2 CLIENT-MASTER-KEY message. 8573 8574 This issue was discovered by Sean Burford (Google) and Emilia Käsper 8575 (OpenSSL development team). 8576 ([CVE-2015-0293]) 8577 8578 *Emilia Käsper* 8579 8580 * Use After Free following d2i_ECPrivatekey error fix 8581 8582 A malformed EC private key file consumed via the d2i_ECPrivateKey function 8583 could cause a use after free condition. This, in turn, could cause a double 8584 free in several private key parsing functions (such as d2i_PrivateKey 8585 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption 8586 for applications that receive EC private keys from untrusted 8587 sources. This scenario is considered rare. 8588 8589 This issue was discovered by the BoringSSL project and fixed in their 8590 commit 517073cd4b. 8591 ([CVE-2015-0209]) 8592 8593 *Matt Caswell* 8594 8595 * X509_to_X509_REQ NULL pointer deref fix 8596 8597 The function X509_to_X509_REQ will crash with a NULL pointer dereference if 8598 the certificate key is invalid. This function is rarely used in practice. 8599 8600 This issue was discovered by Brian Carpenter. 8601 ([CVE-2015-0288]) 8602 8603 *Stephen Henson* 8604 8605 * Removed the export ciphers from the DEFAULT ciphers 8606 8607 *Kurt Roeckx* 8608 8609### Changes between 1.0.0p and 1.0.0q [15 Jan 2015] 8610 8611 * Build fixes for the Windows and OpenVMS platforms 8612 8613 *Matt Caswell and Richard Levitte* 8614 8615### Changes between 1.0.0o and 1.0.0p [8 Jan 2015] 8616 8617 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 8618 message can cause a segmentation fault in OpenSSL due to a NULL pointer 8619 dereference. This could lead to a Denial Of Service attack. Thanks to 8620 Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 8621 ([CVE-2014-3571]) 8622 8623 *Steve Henson* 8624 8625 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 8626 dtls1_buffer_record function under certain conditions. In particular this 8627 could occur if an attacker sent repeated DTLS records with the same 8628 sequence number but for the next epoch. The memory leak could be exploited 8629 by an attacker in a Denial of Service attack through memory exhaustion. 8630 Thanks to Chris Mueller for reporting this issue. 8631 ([CVE-2015-0206]) 8632 8633 *Matt Caswell* 8634 8635 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 8636 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 8637 method would be set to NULL which could later result in a NULL pointer 8638 dereference. Thanks to Frank Schmirler for reporting this issue. 8639 ([CVE-2014-3569]) 8640 8641 *Kurt Roeckx* 8642 8643 * Abort handshake if server key exchange message is omitted for ephemeral 8644 ECDH ciphersuites. 8645 8646 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 8647 reporting this issue. 8648 ([CVE-2014-3572]) 8649 8650 *Steve Henson* 8651 8652 * Remove non-export ephemeral RSA code on client and server. This code 8653 violated the TLS standard by allowing the use of temporary RSA keys in 8654 non-export ciphersuites and could be used by a server to effectively 8655 downgrade the RSA key length used to a value smaller than the server 8656 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 8657 INRIA or reporting this issue. 8658 ([CVE-2015-0204]) 8659 8660 *Steve Henson* 8661 8662 * Fixed issue where DH client certificates are accepted without verification. 8663 An OpenSSL server will accept a DH certificate for client authentication 8664 without the certificate verify message. This effectively allows a client to 8665 authenticate without the use of a private key. This only affects servers 8666 which trust a client certificate authority which issues certificates 8667 containing DH keys: these are extremely rare and hardly ever encountered. 8668 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 8669 this issue. 8670 ([CVE-2015-0205]) 8671 8672 *Steve Henson* 8673 8674 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 8675 results on some platforms, including x86_64. This bug occurs at random 8676 with a very low probability, and is not known to be exploitable in any 8677 way, though its exact impact is difficult to determine. Thanks to Pieter 8678 Wuille (Blockstream) who reported this issue and also suggested an initial 8679 fix. Further analysis was conducted by the OpenSSL development team and 8680 Adam Langley of Google. The final fix was developed by Andy Polyakov of 8681 the OpenSSL core team. 8682 ([CVE-2014-3570]) 8683 8684 *Andy Polyakov* 8685 8686 * Fix various certificate fingerprint issues. 8687 8688 By using non-DER or invalid encodings outside the signed portion of a 8689 certificate the fingerprint can be changed without breaking the signature. 8690 Although no details of the signed portion of the certificate can be changed 8691 this can cause problems with some applications: e.g. those using the 8692 certificate fingerprint for blacklists. 8693 8694 1. Reject signatures with non zero unused bits. 8695 8696 If the BIT STRING containing the signature has non zero unused bits reject 8697 the signature. All current signature algorithms require zero unused bits. 8698 8699 2. Check certificate algorithm consistency. 8700 8701 Check the AlgorithmIdentifier inside TBS matches the one in the 8702 certificate signature. NB: this will result in signature failure 8703 errors for some broken certificates. 8704 8705 Thanks to Konrad Kraszewski from Google for reporting this issue. 8706 8707 3. Check DSA/ECDSA signatures use DER. 8708 8709 Reencode DSA/ECDSA signatures and compare with the original received 8710 signature. Return an error if there is a mismatch. 8711 8712 This will reject various cases including garbage after signature 8713 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 8714 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 8715 (negative or with leading zeroes). 8716 8717 Further analysis was conducted and fixes were developed by Stephen Henson 8718 of the OpenSSL core team. 8719 8720 ([CVE-2014-8275]) 8721 8722 *Steve Henson* 8723 8724### Changes between 1.0.0n and 1.0.0o [15 Oct 2014] 8725 8726 * Session Ticket Memory Leak. 8727 8728 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 8729 integrity of that ticket is first verified. In the event of a session 8730 ticket integrity check failing, OpenSSL will fail to free memory 8731 causing a memory leak. By sending a large number of invalid session 8732 tickets an attacker could exploit this issue in a Denial Of Service 8733 attack. 8734 ([CVE-2014-3567]) 8735 8736 *Steve Henson* 8737 8738 * Build option no-ssl3 is incomplete. 8739 8740 When OpenSSL is configured with "no-ssl3" as a build option, servers 8741 could accept and complete a SSL 3.0 handshake, and clients could be 8742 configured to send them. 8743 ([CVE-2014-3568]) 8744 8745 *Akamai and the OpenSSL team* 8746 8747 * Add support for TLS_FALLBACK_SCSV. 8748 Client applications doing fallback retries should call 8749 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 8750 ([CVE-2014-3566]) 8751 8752 *Adam Langley, Bodo Moeller* 8753 8754 * Add additional DigestInfo checks. 8755 8756 Reencode DigestInto in DER and check against the original when 8757 verifying RSA signature: this will reject any improperly encoded 8758 DigestInfo structures. 8759 8760 Note: this is a precautionary measure and no attacks are currently known. 8761 8762 *Steve Henson* 8763 8764### Changes between 1.0.0m and 1.0.0n [6 Aug 2014] 8765 8766 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 8767 to a denial of service attack. A malicious server can crash the client 8768 with a null pointer dereference (read) by specifying an anonymous (EC)DH 8769 ciphersuite and sending carefully crafted handshake messages. 8770 8771 Thanks to Felix Gröbert (Google) for discovering and researching this 8772 issue. 8773 ([CVE-2014-3510]) 8774 8775 *Emilia Käsper* 8776 8777 * By sending carefully crafted DTLS packets an attacker could cause openssl 8778 to leak memory. This can be exploited through a Denial of Service attack. 8779 Thanks to Adam Langley for discovering and researching this issue. 8780 ([CVE-2014-3507]) 8781 8782 *Adam Langley* 8783 8784 * An attacker can force openssl to consume large amounts of memory whilst 8785 processing DTLS handshake messages. This can be exploited through a 8786 Denial of Service attack. 8787 Thanks to Adam Langley for discovering and researching this issue. 8788 ([CVE-2014-3506]) 8789 8790 *Adam Langley* 8791 8792 * An attacker can force an error condition which causes openssl to crash 8793 whilst processing DTLS packets due to memory being freed twice. This 8794 can be exploited through a Denial of Service attack. 8795 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 8796 this issue. 8797 ([CVE-2014-3505]) 8798 8799 *Adam Langley* 8800 8801 * If a multithreaded client connects to a malicious server using a resumed 8802 session and the server sends an ec point format extension it could write 8803 up to 255 bytes to freed memory. 8804 8805 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 8806 issue. 8807 ([CVE-2014-3509]) 8808 8809 *Gabor Tyukasz* 8810 8811 * A flaw in OBJ_obj2txt may cause pretty printing functions such as 8812 X509_name_oneline, X509_name_print_ex et al. to leak some information 8813 from the stack. Applications may be affected if they echo pretty printing 8814 output to the attacker. 8815 8816 Thanks to Ivan Fratric (Google) for discovering this issue. 8817 ([CVE-2014-3508]) 8818 8819 *Emilia Käsper, and Steve Henson* 8820 8821 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 8822 for corner cases. (Certain input points at infinity could lead to 8823 bogus results, with non-infinity inputs mapped to infinity too.) 8824 8825 *Bodo Moeller* 8826 8827### Changes between 1.0.0l and 1.0.0m [5 Jun 2014] 8828 8829 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 8830 handshake can force the use of weak keying material in OpenSSL 8831 SSL/TLS clients and servers. 8832 8833 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 8834 researching this issue. ([CVE-2014-0224]) 8835 8836 *KIKUCHI Masashi, Steve Henson* 8837 8838 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 8839 OpenSSL DTLS client the code can be made to recurse eventually crashing 8840 in a DoS attack. 8841 8842 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 8843 ([CVE-2014-0221]) 8844 8845 *Imre Rad, Steve Henson* 8846 8847 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 8848 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 8849 client or server. This is potentially exploitable to run arbitrary 8850 code on a vulnerable client or server. 8851 8852 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195]) 8853 8854 *Jüri Aedla, Steve Henson* 8855 8856 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 8857 are subject to a denial of service attack. 8858 8859 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering 8860 this issue. ([CVE-2014-3470]) 8861 8862 *Felix Gröbert, Ivan Fratric, Steve Henson* 8863 8864 * Harmonize version and its documentation. -f flag is used to display 8865 compilation flags. 8866 8867 *mancha <mancha1@zoho.com>* 8868 8869 * Fix eckey_priv_encode so it immediately returns an error upon a failure 8870 in i2d_ECPrivateKey. 8871 8872 *mancha <mancha1@zoho.com>* 8873 8874 * Fix some double frees. These are not thought to be exploitable. 8875 8876 *mancha <mancha1@zoho.com>* 8877 8878 * Fix for the attack described in the paper "Recovering OpenSSL 8879 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 8880 by Yuval Yarom and Naomi Benger. Details can be obtained from: 8881 <http://eprint.iacr.org/2014/140> 8882 8883 Thanks to Yuval Yarom and Naomi Benger for discovering this 8884 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076]) 8885 8886 *Yuval Yarom and Naomi Benger* 8887 8888### Changes between 1.0.0k and 1.0.0l [6 Jan 2014] 8889 8890 * Keep original DTLS digest and encryption contexts in retransmission 8891 structures so we can use the previous session parameters if they need 8892 to be resent. ([CVE-2013-6450]) 8893 8894 *Steve Henson* 8895 8896 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 8897 avoids preferring ECDHE-ECDSA ciphers when the client appears to be 8898 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 8899 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 8900 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 8901 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 8902 8903 *Rob Stradling, Adam Langley* 8904 8905### Changes between 1.0.0j and 1.0.0k [5 Feb 2013] 8906 8907 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 8908 8909 This addresses the flaw in CBC record processing discovered by 8910 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 8911 at: <http://www.isg.rhul.ac.uk/tls/> 8912 8913 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 8914 Security Group at Royal Holloway, University of London 8915 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 8916 Emilia Käsper for the initial patch. 8917 ([CVE-2013-0169]) 8918 8919 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson* 8920 8921 * Return an error when checking OCSP signatures when key is NULL. 8922 This fixes a DoS attack. ([CVE-2013-0166]) 8923 8924 *Steve Henson* 8925 8926 * Call OCSP Stapling callback after ciphersuite has been chosen, so 8927 the right response is stapled. Also change SSL_get_certificate() 8928 so it returns the certificate actually sent. 8929 See <http://rt.openssl.org/Ticket/Display.html?id=2836>. 8930 (This is a backport) 8931 8932 *Rob Stradling <rob.stradling@comodo.com>* 8933 8934 * Fix possible deadlock when decoding public keys. 8935 8936 *Steve Henson* 8937 8938### Changes between 1.0.0i and 1.0.0j [10 May 2012] 8939 8940[NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after 8941OpenSSL 1.0.1.] 8942 8943 * Sanity check record length before skipping explicit IV in DTLS 8944 to fix DoS attack. 8945 8946 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 8947 fuzzing as a service testing platform. 8948 ([CVE-2012-2333]) 8949 8950 *Steve Henson* 8951 8952 * Initialise tkeylen properly when encrypting CMS messages. 8953 Thanks to Solar Designer of Openwall for reporting this issue. 8954 8955 *Steve Henson* 8956 8957### Changes between 1.0.0h and 1.0.0i [19 Apr 2012] 8958 8959 * Check for potentially exploitable overflows in asn1_d2i_read_bio 8960 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 8961 in CRYPTO_realloc_clean. 8962 8963 Thanks to Tavis Ormandy, Google Security Team, for discovering this 8964 issue and to Adam Langley <agl@chromium.org> for fixing it. 8965 ([CVE-2012-2110]) 8966 8967 *Adam Langley (Google), Tavis Ormandy, Google Security Team* 8968 8969### Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 8970 8971 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 8972 in CMS and PKCS7 code. When RSA decryption fails use a random key for 8973 content decryption and always return the same error. Note: this attack 8974 needs on average 2^20 messages so it only affects automated senders. The 8975 old behaviour can be re-enabled in the CMS code by setting the 8976 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 8977 an MMA defence is not necessary. 8978 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 8979 this issue. ([CVE-2012-0884]) 8980 8981 *Steve Henson* 8982 8983 * Fix CVE-2011-4619: make sure we really are receiving a 8984 client hello before rejecting multiple SGC restarts. Thanks to 8985 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 8986 8987 *Steve Henson* 8988 8989### Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 8990 8991 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 8992 Thanks to Antonio Martin, Enterprise Secure Access Research and 8993 Development, Cisco Systems, Inc. for discovering this bug and 8994 preparing a fix. ([CVE-2012-0050]) 8995 8996 *Antonio Martin* 8997 8998### Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 8999 9000 * Nadhem Alfardan and Kenny Paterson have discovered an extension 9001 of the Vaudenay padding oracle attack on CBC mode encryption 9002 which enables an efficient plaintext recovery attack against 9003 the OpenSSL implementation of DTLS. Their attack exploits timing 9004 differences arising during decryption processing. A research 9005 paper describing this attack can be found at: 9006 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf> 9007 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 9008 Security Group at Royal Holloway, University of London 9009 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 9010 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 9011 for preparing the fix. ([CVE-2011-4108]) 9012 9013 *Robin Seggelmann, Michael Tuexen* 9014 9015 * Clear bytes used for block padding of SSL 3.0 records. 9016 ([CVE-2011-4576]) 9017 9018 *Adam Langley (Google)* 9019 9020 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George 9021 Kadianakis <desnacked@gmail.com> for discovering this issue and 9022 Adam Langley for preparing the fix. ([CVE-2011-4619]) 9023 9024 *Adam Langley (Google)* 9025 9026 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027]) 9027 9028 *Andrey Kulikov <amdeich@gmail.com>* 9029 9030 * Prevent malformed RFC3779 data triggering an assertion failure. 9031 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 9032 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577]) 9033 9034 *Rob Austein <sra@hactrn.net>* 9035 9036 * Improved PRNG seeding for VOS. 9037 9038 *Paul Green <Paul.Green@stratus.com>* 9039 9040 * Fix ssl_ciph.c set-up race. 9041 9042 *Adam Langley (Google)* 9043 9044 * Fix spurious failures in ecdsatest.c. 9045 9046 *Emilia Käsper (Google)* 9047 9048 * Fix the BIO_f_buffer() implementation (which was mixing different 9049 interpretations of the `..._len` fields). 9050 9051 *Adam Langley (Google)* 9052 9053 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 9054 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 9055 threads won't reuse the same blinding coefficients. 9056 9057 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 9058 lock to call BN_BLINDING_invert_ex, and avoids one use of 9059 BN_BLINDING_update for each BN_BLINDING structure (previously, 9060 the last update always remained unused). 9061 9062 *Emilia Käsper (Google)* 9063 9064 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 9065 9066 *Bob Buckholz (Google)* 9067 9068### Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 9069 9070 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted 9071 by initialising X509_STORE_CTX properly. ([CVE-2011-3207]) 9072 9073 *Kaspar Brand <ossl@velox.ch>* 9074 9075 * Fix SSL memory handling for (EC)DH ciphersuites, in particular 9076 for multi-threaded use of ECDH. ([CVE-2011-3210]) 9077 9078 *Adam Langley (Google)* 9079 9080 * Fix x509_name_ex_d2i memory leak on bad inputs. 9081 9082 *Bodo Moeller* 9083 9084 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 9085 signature public key algorithm by using OID xref utilities instead. 9086 Before this you could only use some ECC ciphersuites with SHA1 only. 9087 9088 *Steve Henson* 9089 9090 * Add protection against ECDSA timing attacks as mentioned in the paper 9091 by Billy Bob Brumley and Nicola Tuveri, see: 9092 <http://eprint.iacr.org/2011/232.pdf> 9093 9094 *Billy Bob Brumley and Nicola Tuveri* 9095 9096### Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 9097 9098 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 9099 9100 *Neel Mehta, Adam Langley, Bodo Moeller (Google)* 9101 9102 * Fix bug in string printing code: if *any* escaping is enabled we must 9103 escape the escape character (backslash) or the resulting string is 9104 ambiguous. 9105 9106 *Steve Henson* 9107 9108### Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 9109 9110 * Disable code workaround for ancient and obsolete Netscape browsers 9111 and servers: an attacker can use it in a ciphersuite downgrade attack. 9112 Thanks to Martin Rex for discovering this bug. CVE-2010-4180 9113 9114 *Steve Henson* 9115 9116 * Fixed J-PAKE implementation error, originally discovered by 9117 Sebastien Martini, further info and confirmation from Stefan 9118 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 9119 9120 *Ben Laurie* 9121 9122### Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 9123 9124 * Fix extension code to avoid race conditions which can result in a buffer 9125 overrun vulnerability: resumed sessions must not be modified as they can 9126 be shared by multiple threads. CVE-2010-3864 9127 9128 *Steve Henson* 9129 9130 * Fix WIN32 build system to correctly link an ENGINE directory into 9131 a DLL. 9132 9133 *Steve Henson* 9134 9135### Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 9136 9137 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover 9138 ([CVE-2010-1633]) 9139 9140 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>* 9141 9142### Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 9143 9144 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 9145 context. The operation can be customised via the ctrl mechanism in 9146 case ENGINEs want to include additional functionality. 9147 9148 *Steve Henson* 9149 9150 * Tolerate yet another broken PKCS#8 key format: private key value negative. 9151 9152 *Steve Henson* 9153 9154 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 9155 output hashes compatible with older versions of OpenSSL. 9156 9157 *Willy Weisz <weisz@vcpc.univie.ac.at>* 9158 9159 * Fix compression algorithm handling: if resuming a session use the 9160 compression algorithm of the resumed session instead of determining 9161 it from client hello again. Don't allow server to change algorithm. 9162 9163 *Steve Henson* 9164 9165 * Add load_crls() function to commands tidying load_certs() too. Add option 9166 to verify utility to allow additional CRLs to be included. 9167 9168 *Steve Henson* 9169 9170 * Update OCSP request code to permit adding custom headers to the request: 9171 some responders need this. 9172 9173 *Steve Henson* 9174 9175 * The function EVP_PKEY_sign() returns <=0 on error: check return code 9176 correctly. 9177 9178 *Julia Lawall <julia@diku.dk>* 9179 9180 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it 9181 needlessly dereferenced structures, used obsolete functions and 9182 didn't handle all updated verify codes correctly. 9183 9184 *Steve Henson* 9185 9186 * Disable MD2 in the default configuration. 9187 9188 *Steve Henson* 9189 9190 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 9191 indicate the initial BIO being pushed or popped. This makes it possible 9192 to determine whether the BIO is the one explicitly called or as a result 9193 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 9194 it handles reference counts correctly and doesn't zero out the I/O bio 9195 when it is not being explicitly popped. WARNING: applications which 9196 included workarounds for the old buggy behaviour will need to be modified 9197 or they could free up already freed BIOs. 9198 9199 *Steve Henson* 9200 9201 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 9202 renaming to all platforms (within the 0.9.8 branch, this was 9203 done conditionally on Netware platforms to avoid a name clash). 9204 9205 *Guenter <lists@gknw.net>* 9206 9207 * Add ECDHE and PSK support to DTLS. 9208 9209 *Michael Tuexen <tuexen@fh-muenster.de>* 9210 9211 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 9212 be used on C++. 9213 9214 *Steve Henson* 9215 9216 * Add "missing" function EVP_MD_flags() (without this the only way to 9217 retrieve a digest flags is by accessing the structure directly. Update 9218 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest 9219 or cipher is registered as in the "from" argument. Print out all 9220 registered digests in the dgst usage message instead of manually 9221 attempting to work them out. 9222 9223 *Steve Henson* 9224 9225 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 9226 this allows the use of compression and extensions. Change default cipher 9227 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 9228 by default unless an application cipher string requests it. 9229 9230 *Steve Henson* 9231 9232 * Alter match criteria in PKCS12_parse(). It used to try to use local 9233 key ids to find matching certificates and keys but some PKCS#12 files 9234 don't follow the (somewhat unwritten) rules and this strategy fails. 9235 Now just gather all certificates together and the first private key 9236 then look for the first certificate that matches the key. 9237 9238 *Steve Henson* 9239 9240 * Support use of registered digest and cipher names for dgst and cipher 9241 commands instead of having to add each one as a special case. So now 9242 you can do: 9243 9244 openssl sha256 foo 9245 9246 as well as: 9247 9248 openssl dgst -sha256 foo 9249 9250 and this works for ENGINE based algorithms too. 9251 9252 *Steve Henson* 9253 9254 * Update Gost ENGINE to support parameter files. 9255 9256 *Victor B. Wagner <vitus@cryptocom.ru>* 9257 9258 * Support GeneralizedTime in ca utility. 9259 9260 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson* 9261 9262 * Enhance the hash format used for certificate directory links. The new 9263 form uses the canonical encoding (meaning equivalent names will work 9264 even if they aren't identical) and uses SHA1 instead of MD5. This form 9265 is incompatible with the older format and as a result c_rehash should 9266 be used to rebuild symbolic links. 9267 9268 *Steve Henson* 9269 9270 * Make PKCS#8 the default write format for private keys, replacing the 9271 traditional format. This form is standardised, more secure and doesn't 9272 include an implicit MD5 dependency. 9273 9274 *Steve Henson* 9275 9276 * Add a $gcc_devteam_warn option to Configure. The idea is that any code 9277 committed to OpenSSL should pass this lot as a minimum. 9278 9279 *Steve Henson* 9280 9281 * Add session ticket override functionality for use by EAP-FAST. 9282 9283 *Jouni Malinen <j@w1.fi>* 9284 9285 * Modify HMAC functions to return a value. Since these can be implemented 9286 in an ENGINE errors can occur. 9287 9288 *Steve Henson* 9289 9290 * Type-checked OBJ_bsearch_ex. 9291 9292 *Ben Laurie* 9293 9294 * Type-checked OBJ_bsearch. Also some constification necessitated 9295 by type-checking. Still to come: TXT_DB, bsearch(?), 9296 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 9297 CONF_VALUE. 9298 9299 *Ben Laurie* 9300 9301 * New function OPENSSL_gmtime_adj() to add a specific number of days and 9302 seconds to a tm structure directly, instead of going through OS 9303 specific date routines. This avoids any issues with OS routines such 9304 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures 9305 and X509_time_adj_ex() to cover the extended range. The existing 9306 X509_time_adj() is still usable and will no longer have any date issues. 9307 9308 *Steve Henson* 9309 9310 * Delta CRL support. New use deltas option which will attempt to locate 9311 and search any appropriate delta CRLs available. 9312 9313 This work was sponsored by Google. 9314 9315 *Steve Henson* 9316 9317 * Support for CRLs partitioned by reason code. Reorganise CRL processing 9318 code and add additional score elements. Validate alternate CRL paths 9319 as part of the CRL checking and indicate a new error "CRL path validation 9320 error" in this case. Applications wanting additional details can use 9321 the verify callback and check the new "parent" field. If this is not 9322 NULL CRL path validation is taking place. Existing applications won't 9323 see this because it requires extended CRL support which is off by 9324 default. 9325 9326 This work was sponsored by Google. 9327 9328 *Steve Henson* 9329 9330 * Support for freshest CRL extension. 9331 9332 This work was sponsored by Google. 9333 9334 *Steve Henson* 9335 9336 * Initial indirect CRL support. Currently only supported in the CRLs 9337 passed directly and not via lookup. Process certificate issuer 9338 CRL entry extension and lookup CRL entries by bother issuer name 9339 and serial number. Check and process CRL issuer entry in IDP extension. 9340 9341 This work was sponsored by Google. 9342 9343 *Steve Henson* 9344 9345 * Add support for distinct certificate and CRL paths. The CRL issuer 9346 certificate is validated separately in this case. Only enabled if 9347 an extended CRL support flag is set: this flag will enable additional 9348 CRL functionality in future. 9349 9350 This work was sponsored by Google. 9351 9352 *Steve Henson* 9353 9354 * Add support for policy mappings extension. 9355 9356 This work was sponsored by Google. 9357 9358 *Steve Henson* 9359 9360 * Fixes to pathlength constraint, self issued certificate handling, 9361 policy processing to align with RFC3280 and PKITS tests. 9362 9363 This work was sponsored by Google. 9364 9365 *Steve Henson* 9366 9367 * Support for name constraints certificate extension. DN, email, DNS 9368 and URI types are currently supported. 9369 9370 This work was sponsored by Google. 9371 9372 *Steve Henson* 9373 9374 * To cater for systems that provide a pointer-based thread ID rather 9375 than numeric, deprecate the current numeric thread ID mechanism and 9376 replace it with a structure and associated callback type. This 9377 mechanism allows a numeric "hash" to be extracted from a thread ID in 9378 either case, and on platforms where pointers are larger than 'long', 9379 mixing is done to help ensure the numeric 'hash' is usable even if it 9380 can't be guaranteed unique. The default mechanism is to use "&errno" 9381 as a pointer-based thread ID to distinguish between threads. 9382 9383 Applications that want to provide their own thread IDs should now use 9384 CRYPTO_THREADID_set_callback() to register a callback that will call 9385 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 9386 9387 Note that ERR_remove_state() is now deprecated, because it is tied 9388 to the assumption that thread IDs are numeric. ERR_remove_state(0) 9389 to free the current thread's error state should be replaced by 9390 ERR_remove_thread_state(NULL). 9391 9392 (This new approach replaces the functions CRYPTO_set_idptr_callback(), 9393 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 9394 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 9395 application was previously providing a numeric thread callback that 9396 was inappropriate for distinguishing threads, then uniqueness might 9397 have been obtained with &errno that happened immediately in the 9398 intermediate development versions of OpenSSL; this is no longer the 9399 case, the numeric thread callback will now override the automatic use 9400 of &errno.) 9401 9402 *Geoff Thorpe, with help from Bodo Moeller* 9403 9404 * Initial support for different CRL issuing certificates. This covers a 9405 simple case where the self issued certificates in the chain exist and 9406 the real CRL issuer is higher in the existing chain. 9407 9408 This work was sponsored by Google. 9409 9410 *Steve Henson* 9411 9412 * Removed effectively defunct crypto/store from the build. 9413 9414 *Ben Laurie* 9415 9416 * Revamp of STACK to provide stronger type-checking. Still to come: 9417 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 9418 ASN1_STRING, CONF_VALUE. 9419 9420 *Ben Laurie* 9421 9422 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 9423 RAM on SSL connections. This option can save about 34k per idle SSL. 9424 9425 *Nick Mathewson* 9426 9427 * Revamp of LHASH to provide stronger type-checking. Still to come: 9428 STACK, TXT_DB, bsearch, qsort. 9429 9430 *Ben Laurie* 9431 9432 * Initial support for Cryptographic Message Syntax (aka CMS) based 9433 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 9434 support for data, signedData, compressedData, digestedData and 9435 encryptedData, envelopedData types included. Scripts to check against 9436 RFC4134 examples draft and interop and consistency checks of many 9437 content types and variants. 9438 9439 *Steve Henson* 9440 9441 * Add options to enc utility to support use of zlib compression BIO. 9442 9443 *Steve Henson* 9444 9445 * Extend mk1mf to support importing of options and assembly language 9446 files from Configure script, currently only included in VC-WIN32. 9447 The assembly language rules can now optionally generate the source 9448 files from the associated perl scripts. 9449 9450 *Steve Henson* 9451 9452 * Implement remaining functionality needed to support GOST ciphersuites. 9453 Interop testing has been performed using CryptoPro implementations. 9454 9455 *Victor B. Wagner <vitus@cryptocom.ru>* 9456 9457 * s390x assembler pack. 9458 9459 *Andy Polyakov* 9460 9461 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 9462 "family." 9463 9464 *Andy Polyakov* 9465 9466 * Implement Opaque PRF Input TLS extension as specified in 9467 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 9468 official specification yet and no extension type assignment by 9469 IANA exists, this extension (for now) will have to be explicitly 9470 enabled when building OpenSSL by providing the extension number 9471 to use. For example, specify an option 9472 9473 -DTLSEXT_TYPE_opaque_prf_input=0x9527 9474 9475 to the "config" or "Configure" script to enable the extension, 9476 assuming extension number 0x9527 (which is a completely arbitrary 9477 and unofficial assignment based on the MD5 hash of the Internet 9478 Draft). Note that by doing so, you potentially lose 9479 interoperability with other TLS implementations since these might 9480 be using the same extension number for other purposes. 9481 9482 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 9483 opaque PRF input value to use in the handshake. This will create 9484 an internal copy of the length-'len' string at 'src', and will 9485 return non-zero for success. 9486 9487 To get more control and flexibility, provide a callback function 9488 by using 9489 9490 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 9491 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 9492 9493 where 9494 9495 int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 9496 void *arg; 9497 9498 Callback function 'cb' will be called in handshakes, and is 9499 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 9500 Argument 'arg' is for application purposes (the value as given to 9501 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 9502 be provided to the callback function). The callback function 9503 has to return non-zero to report success: usually 1 to use opaque 9504 PRF input just if possible, or 2 to enforce use of the opaque PRF 9505 input. In the latter case, the library will abort the handshake 9506 if opaque PRF input is not successfully negotiated. 9507 9508 Arguments 'peerinput' and 'len' given to the callback function 9509 will always be NULL and 0 in the case of a client. A server will 9510 see the client's opaque PRF input through these variables if 9511 available (NULL and 0 otherwise). Note that if the server 9512 provides an opaque PRF input, the length must be the same as the 9513 length of the client's opaque PRF input. 9514 9515 Note that the callback function will only be called when creating 9516 a new session (session resumption can resume whatever was 9517 previously negotiated), and will not be called in SSL 2.0 9518 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 9519 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 9520 for applications that need to enforce opaque PRF input. 9521 9522 *Bodo Moeller* 9523 9524 * Update ssl code to support digests other than SHA1+MD5 for handshake 9525 MAC. 9526 9527 *Victor B. Wagner <vitus@cryptocom.ru>* 9528 9529 * Add RFC4507 support to OpenSSL. This includes the corrections in 9530 RFC4507bis. The encrypted ticket format is an encrypted encoded 9531 SSL_SESSION structure, that way new session features are automatically 9532 supported. 9533 9534 If a client application caches session in an SSL_SESSION structure 9535 support is transparent because tickets are now stored in the encoded 9536 SSL_SESSION. 9537 9538 The SSL_CTX structure automatically generates keys for ticket 9539 protection in servers so again support should be possible 9540 with no application modification. 9541 9542 If a client or server wishes to disable RFC4507 support then the option 9543 SSL_OP_NO_TICKET can be set. 9544 9545 Add a TLS extension debugging callback to allow the contents of any client 9546 or server extensions to be examined. 9547 9548 This work was sponsored by Google. 9549 9550 *Steve Henson* 9551 9552 * Final changes to avoid use of pointer pointer casts in OpenSSL. 9553 OpenSSL should now compile cleanly on gcc 4.2 9554 9555 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson* 9556 9557 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 9558 support including streaming MAC support: this is required for GOST 9559 ciphersuite support. 9560 9561 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson* 9562 9563 * Add option -stream to use PKCS#7 streaming in smime utility. New 9564 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 9565 to output in BER and PEM format. 9566 9567 *Steve Henson* 9568 9569 * Experimental support for use of HMAC via EVP_PKEY interface. This 9570 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The 9571 EVP_PKEY "key" in this case is the HMAC key, potentially allowing 9572 ENGINE support for HMAC keys which are unextractable. New -mac and 9573 -macopt options to dgst utility. 9574 9575 *Steve Henson* 9576 9577 * New option -sigopt to dgst utility. Update dgst to use 9578 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use 9579 alternative signing parameters such as X9.31 or PSS in the dgst 9580 utility. 9581 9582 *Steve Henson* 9583 9584 * Change ssl_cipher_apply_rule(), the internal function that does 9585 the work each time a ciphersuite string requests enabling 9586 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 9587 removing ("!foo+bar") a class of ciphersuites: Now it maintains 9588 the order of disabled ciphersuites such that those ciphersuites 9589 that most recently went from enabled to disabled not only stay 9590 in order with respect to each other, but also have higher priority 9591 than other disabled ciphersuites the next time ciphersuites are 9592 enabled again. 9593 9594 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 9595 the same ciphersuites as with "HIGH" alone, but in a specific 9596 order where the PSK ciphersuites come first (since they are the 9597 most recently disabled ciphersuites when "HIGH" is parsed). 9598 9599 Also, change ssl_create_cipher_list() (using this new 9600 functionality) such that between otherwise identical 9601 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in 9602 the default order. 9603 9604 *Bodo Moeller* 9605 9606 * Change ssl_create_cipher_list() so that it automatically 9607 arranges the ciphersuites in reasonable order before starting 9608 to process the rule string. Thus, the definition for "DEFAULT" 9609 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 9610 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`. 9611 This makes it much easier to arrive at a reasonable default order 9612 in applications for which anonymous ciphers are OK (meaning 9613 that you can't actually use DEFAULT). 9614 9615 *Bodo Moeller; suggested by Victor Duchovni* 9616 9617 * Split the SSL/TLS algorithm mask (as used for ciphersuite string 9618 processing) into multiple integers instead of setting 9619 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 9620 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 9621 (These masks as well as the individual bit definitions are hidden 9622 away into the non-exported interface ssl/ssl_locl.h, so this 9623 change to the definition of the SSL_CIPHER structure shouldn't 9624 affect applications.) This give us more bits for each of these 9625 categories, so there is no longer a need to coagulate AES128 and 9626 AES256 into a single algorithm bit, and to coagulate Camellia128 9627 and Camellia256 into a single algorithm bit, which has led to all 9628 kinds of kludges. 9629 9630 Thus, among other things, the kludge introduced in 0.9.7m and 9631 0.9.8e for masking out AES256 independently of AES128 or masking 9632 out Camellia256 independently of AES256 is not needed here in 0.9.9. 9633 9634 With the change, we also introduce new ciphersuite aliases that 9635 so far were missing: "AES128", "AES256", "CAMELLIA128", and 9636 "CAMELLIA256". 9637 9638 *Bodo Moeller* 9639 9640 * Add support for dsa-with-SHA224 and dsa-with-SHA256. 9641 Use the leftmost N bytes of the signature input if the input is 9642 larger than the prime q (with N being the size in bytes of q). 9643 9644 *Nils Larsch* 9645 9646 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 9647 it yet and it is largely untested. 9648 9649 *Steve Henson* 9650 9651 * Add support for the ecdsa-with-SHA224/256/384/512 signature types. 9652 9653 *Nils Larsch* 9654 9655 * Initial incomplete changes to avoid need for function casts in OpenSSL 9656 some compilers (gcc 4.2 and later) reject their use. Safestack is 9657 reimplemented. Update ASN1 to avoid use of legacy functions. 9658 9659 *Steve Henson* 9660 9661 * Win32/64 targets are linked with Winsock2. 9662 9663 *Andy Polyakov* 9664 9665 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 9666 to external functions. This can be used to increase CRL handling 9667 efficiency especially when CRLs are very large by (for example) storing 9668 the CRL revoked certificates in a database. 9669 9670 *Steve Henson* 9671 9672 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so 9673 new CRLs added to a directory can be used. New command line option 9674 -verify_return_error to s_client and s_server. This causes real errors 9675 to be returned by the verify callback instead of carrying on no matter 9676 what. This reflects the way a "real world" verify callback would behave. 9677 9678 *Steve Henson* 9679 9680 * GOST engine, supporting several GOST algorithms and public key formats. 9681 Kindly donated by Cryptocom. 9682 9683 *Cryptocom* 9684 9685 * Partial support for Issuing Distribution Point CRL extension. CRLs 9686 partitioned by DP are handled but no indirect CRL or reason partitioning 9687 (yet). Complete overhaul of CRL handling: now the most suitable CRL is 9688 selected via a scoring technique which handles IDP and AKID in CRLs. 9689 9690 *Steve Henson* 9691 9692 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 9693 will ultimately be used for all verify operations: this will remove the 9694 X509_STORE dependency on certificate verification and allow alternative 9695 lookup methods. X509_STORE based implementations of these two callbacks. 9696 9697 *Steve Henson* 9698 9699 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 9700 Modify get_crl() to find a valid (unexpired) CRL if possible. 9701 9702 *Steve Henson* 9703 9704 * New function X509_CRL_match() to check if two CRLs are identical. Normally 9705 this would be called X509_CRL_cmp() but that name is already used by 9706 a function that just compares CRL issuer names. Cache several CRL 9707 extensions in X509_CRL structure and cache CRLDP in X509. 9708 9709 *Steve Henson* 9710 9711 * Store a "canonical" representation of X509_NAME structure (ASN1 Name) 9712 this maps equivalent X509_NAME structures into a consistent structure. 9713 Name comparison can then be performed rapidly using memcmp(). 9714 9715 *Steve Henson* 9716 9717 * Non-blocking OCSP request processing. Add -timeout option to ocsp 9718 utility. 9719 9720 *Steve Henson* 9721 9722 * Allow digests to supply their own micalg string for S/MIME type using 9723 the ctrl EVP_MD_CTRL_MICALG. 9724 9725 *Steve Henson* 9726 9727 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the 9728 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 9729 ctrl. It can then customise the structure before and/or after signing 9730 if necessary. 9731 9732 *Steve Henson* 9733 9734 * New function OBJ_add_sigid() to allow application defined signature OIDs 9735 to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 9736 to free up any added signature OIDs. 9737 9738 *Steve Henson* 9739 9740 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 9741 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 9742 digest and cipher tables. New options added to openssl utility: 9743 list-message-digest-algorithms and list-cipher-algorithms. 9744 9745 *Steve Henson* 9746 9747 * Change the array representation of binary polynomials: the list 9748 of degrees of non-zero coefficients is now terminated with -1. 9749 Previously it was terminated with 0, which was also part of the 9750 value; thus, the array representation was not applicable to 9751 polynomials where t^0 has coefficient zero. This change makes 9752 the array representation useful in a more general context. 9753 9754 *Douglas Stebila* 9755 9756 * Various modifications and fixes to SSL/TLS cipher string 9757 handling. For ECC, the code now distinguishes between fixed ECDH 9758 with RSA certificates on the one hand and with ECDSA certificates 9759 on the other hand, since these are separate ciphersuites. The 9760 unused code for Fortezza ciphersuites has been removed. 9761 9762 For consistency with EDH, ephemeral ECDH is now called "EECDH" 9763 (not "ECDHE"). For consistency with the code for DH 9764 certificates, use of ECDH certificates is now considered ECDH 9765 authentication, not RSA or ECDSA authentication (the latter is 9766 merely the CA's signing algorithm and not actively used in the 9767 protocol). 9768 9769 The temporary ciphersuite alias "ECCdraft" is no longer 9770 available, and ECC ciphersuites are no longer excluded from "ALL" 9771 and "DEFAULT". The following aliases now exist for RFC 4492 9772 ciphersuites, most of these by analogy with the DH case: 9773 9774 kECDHr - ECDH cert, signed with RSA 9775 kECDHe - ECDH cert, signed with ECDSA 9776 kECDH - ECDH cert (signed with either RSA or ECDSA) 9777 kEECDH - ephemeral ECDH 9778 ECDH - ECDH cert or ephemeral ECDH 9779 9780 aECDH - ECDH cert 9781 aECDSA - ECDSA cert 9782 ECDSA - ECDSA cert 9783 9784 AECDH - anonymous ECDH 9785 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 9786 9787 *Bodo Moeller* 9788 9789 * Add additional S/MIME capabilities for AES and GOST ciphers if supported. 9790 Use correct micalg parameters depending on digest(s) in signed message. 9791 9792 *Steve Henson* 9793 9794 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 9795 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 9796 9797 *Steve Henson* 9798 9799 * Initial engine support for EVP_PKEY_METHOD. New functions to permit 9800 an engine to register a method. Add ENGINE lookups for methods and 9801 functional reference processing. 9802 9803 *Steve Henson* 9804 9805 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of 9806 `EVP_{Sign,Verify}*` which allow an application to customise the signature 9807 process. 9808 9809 *Steve Henson* 9810 9811 * New -resign option to smime utility. This adds one or more signers 9812 to an existing PKCS#7 signedData structure. Also -md option to use an 9813 alternative message digest algorithm for signing. 9814 9815 *Steve Henson* 9816 9817 * Tidy up PKCS#7 routines and add new functions to make it easier to 9818 create PKCS7 structures containing multiple signers. Update smime 9819 application to support multiple signers. 9820 9821 *Steve Henson* 9822 9823 * New -macalg option to pkcs12 utility to allow setting of an alternative 9824 digest MAC. 9825 9826 *Steve Henson* 9827 9828 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 9829 Reorganize PBE internals to lookup from a static table using NIDs, 9830 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 9831 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 9832 PRF which will be automatically used with PBES2. 9833 9834 *Steve Henson* 9835 9836 * Replace the algorithm specific calls to generate keys in "req" with the 9837 new API. 9838 9839 *Steve Henson* 9840 9841 * Update PKCS#7 enveloped data routines to use new API. This is now 9842 supported by any public key method supporting the encrypt operation. A 9843 ctrl is added to allow the public key algorithm to examine or modify 9844 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 9845 a no op. 9846 9847 *Steve Henson* 9848 9849 * Add a ctrl to asn1 method to allow a public key algorithm to express 9850 a default digest type to use. In most cases this will be SHA1 but some 9851 algorithms (such as GOST) need to specify an alternative digest. The 9852 return value indicates how strong the preference is 1 means optional and 9853 2 is mandatory (that is it is the only supported type). Modify 9854 ASN1_item_sign() to accept a NULL digest argument to indicate it should 9855 use the default md. Update openssl utilities to use the default digest 9856 type for signing if it is not explicitly indicated. 9857 9858 *Steve Henson* 9859 9860 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 9861 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 9862 signing method from the key type. This effectively removes the link 9863 between digests and public key types. 9864 9865 *Steve Henson* 9866 9867 * Add an OID cross reference table and utility functions. Its purpose is to 9868 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 9869 rsaEncryption. This will allow some of the algorithm specific hackery 9870 needed to use the correct OID to be removed. 9871 9872 *Steve Henson* 9873 9874 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 9875 structures for PKCS7_sign(). They are now set up by the relevant public 9876 key ASN1 method. 9877 9878 *Steve Henson* 9879 9880 * Add provisional EC pkey method with support for ECDSA and ECDH. 9881 9882 *Steve Henson* 9883 9884 * Add support for key derivation (agreement) in the API, DH method and 9885 pkeyutl. 9886 9887 *Steve Henson* 9888 9889 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 9890 public and private key formats. As a side effect these add additional 9891 command line functionality not previously available: DSA signatures can be 9892 generated and verified using pkeyutl and DH key support and generation in 9893 pkey, genpkey. 9894 9895 *Steve Henson* 9896 9897 * BeOS support. 9898 9899 *Oliver Tappe <zooey@hirschkaefer.de>* 9900 9901 * New make target "install_html_docs" installs HTML renditions of the 9902 manual pages. 9903 9904 *Oliver Tappe <zooey@hirschkaefer.de>* 9905 9906 * New utility "genpkey" this is analogous to "genrsa" etc except it can 9907 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 9908 support key and parameter generation and add initial key generation 9909 functionality for RSA. 9910 9911 *Steve Henson* 9912 9913 * Add functions for main EVP_PKEY_method operations. The undocumented 9914 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to 9915 `EVP_PKEY_{encrypt,decrypt}_old`. 9916 9917 *Steve Henson* 9918 9919 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public 9920 key API, doesn't do much yet. 9921 9922 *Steve Henson* 9923 9924 * New function EVP_PKEY_asn1_get0_info() to retrieve information about 9925 public key algorithms. New option to openssl utility: 9926 "list-public-key-algorithms" to print out info. 9927 9928 *Steve Henson* 9929 9930 * Implement the Supported Elliptic Curves Extension for 9931 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 9932 9933 *Douglas Stebila* 9934 9935 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 9936 EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 9937 9938 *Steve Henson* 9939 9940 * New utilities pkey and pkeyparam. These are similar to algorithm specific 9941 utilities such as rsa, dsa, dsaparam etc except they process any key 9942 type. 9943 9944 *Steve Henson* 9945 9946 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 9947 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 9948 EVP_PKEY_print_param() to print public key data from an EVP_PKEY 9949 structure. 9950 9951 *Steve Henson* 9952 9953 * Initial support for pluggable public key ASN1. 9954 De-spaghettify the public key ASN1 handling. Move public and private 9955 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 9956 algorithm specific handling to a single module within the relevant 9957 algorithm directory. Add functions to allow (near) opaque processing 9958 of public and private key structures. 9959 9960 *Steve Henson* 9961 9962 * Implement the Supported Point Formats Extension for 9963 ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 9964 9965 *Douglas Stebila* 9966 9967 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 9968 for the psk identity [hint] and the psk callback functions to the 9969 SSL_SESSION, SSL and SSL_CTX structure. 9970 9971 New ciphersuites: 9972 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 9973 PSK-AES256-CBC-SHA 9974 9975 New functions: 9976 SSL_CTX_use_psk_identity_hint 9977 SSL_get_psk_identity_hint 9978 SSL_get_psk_identity 9979 SSL_use_psk_identity_hint 9980 9981 *Mika Kousa and Pasi Eronen of Nokia Corporation* 9982 9983 * Add RFC 3161 compliant time stamp request creation, response generation 9984 and response verification functionality. 9985 9986 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project* 9987 9988 * Add initial support for TLS extensions, specifically for the server_name 9989 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 9990 have new members for a host name. The SSL data structure has an 9991 additional member `SSL_CTX *initial_ctx` so that new sessions can be 9992 stored in that context to allow for session resumption, even after the 9993 SSL has been switched to a new SSL_CTX in reaction to a client's 9994 server_name extension. 9995 9996 New functions (subject to change): 9997 9998 SSL_get_servername() 9999 SSL_get_servername_type() 10000 SSL_set_SSL_CTX() 10001 10002 New CTRL codes and macros (subject to change): 10003 10004 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 10005 - SSL_CTX_set_tlsext_servername_callback() 10006 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 10007 - SSL_CTX_set_tlsext_servername_arg() 10008 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 10009 10010 openssl s_client has a new '-servername ...' option. 10011 10012 openssl s_server has new options '-servername_host ...', '-cert2 ...', 10013 '-key2 ...', '-servername_fatal' (subject to change). This allows 10014 testing the HostName extension for a specific single host name ('-cert' 10015 and '-key' remain fallbacks for handshakes without HostName 10016 negotiation). If the unrecognized_name alert has to be sent, this by 10017 default is a warning; it becomes fatal with the '-servername_fatal' 10018 option. 10019 10020 *Peter Sylvester, Remy Allais, Christophe Renou* 10021 10022 * Whirlpool hash implementation is added. 10023 10024 *Andy Polyakov* 10025 10026 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 10027 bn(64,32). Because of instruction set limitations it doesn't have 10028 any negative impact on performance. This was done mostly in order 10029 to make it possible to share assembler modules, such as bn_mul_mont 10030 implementations, between 32- and 64-bit builds without hassle. 10031 10032 *Andy Polyakov* 10033 10034 * Move code previously exiled into file crypto/ec/ec2_smpt.c 10035 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 10036 macro. 10037 10038 *Bodo Moeller* 10039 10040 * New candidate for BIGNUM assembler implementation, bn_mul_mont, 10041 dedicated Montgomery multiplication procedure, is introduced. 10042 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 10043 "64-bit" performance on certain 32-bit targets. 10044 10045 *Andy Polyakov* 10046 10047 * New option SSL_OP_NO_COMP to disable use of compression selectively 10048 in SSL structures. New SSL ctrl to set maximum send fragment size. 10049 Save memory by setting the I/O buffer sizes dynamically instead of 10050 using the maximum available value. 10051 10052 *Steve Henson* 10053 10054 * New option -V for 'openssl ciphers'. This prints the ciphersuite code 10055 in addition to the text details. 10056 10057 *Bodo Moeller* 10058 10059 * Very, very preliminary EXPERIMENTAL support for printing of general 10060 ASN1 structures. This currently produces rather ugly output and doesn't 10061 handle several customised structures at all. 10062 10063 *Steve Henson* 10064 10065 * Integrated support for PVK file format and some related formats such 10066 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 10067 these in the 'rsa' and 'dsa' utilities. 10068 10069 *Steve Henson* 10070 10071 * Support for PKCS#1 RSAPublicKey format on rsa utility command line. 10072 10073 *Steve Henson* 10074 10075 * Remove the ancient ASN1_METHOD code. This was only ever used in one 10076 place for the (very old) "NETSCAPE" format certificates which are now 10077 handled using new ASN1 code equivalents. 10078 10079 *Steve Henson* 10080 10081 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 10082 pointer and make the SSL_METHOD parameter in SSL_CTX_new, 10083 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 10084 10085 *Nils Larsch* 10086 10087 * Modify CRL distribution points extension code to print out previously 10088 unsupported fields. Enhance extension setting code to allow setting of 10089 all fields. 10090 10091 *Steve Henson* 10092 10093 * Add print and set support for Issuing Distribution Point CRL extension. 10094 10095 *Steve Henson* 10096 10097 * Change 'Configure' script to enable Camellia by default. 10098 10099 *NTT* 10100 10101OpenSSL 0.9.x 10102------------- 10103 10104### Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 10105 10106 * When rejecting SSL/TLS records due to an incorrect version number, never 10107 update s->server with a new major version number. As of 10108 - OpenSSL 0.9.8m if 'short' is a 16-bit type, 10109 - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 10110 the previous behavior could result in a read attempt at NULL when 10111 receiving specific incorrect SSL/TLS records once record payload 10112 protection is active. ([CVE-2010-0740]) 10113 10114 *Bodo Moeller, Adam Langley <agl@chromium.org>* 10115 10116 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 10117 could be crashed if the relevant tables were not present (e.g. chrooted). 10118 10119 *Tomas Hoger <thoger@redhat.com>* 10120 10121### Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 10122 10123 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245]) 10124 10125 *Martin Olsson, Neel Mehta* 10126 10127 * Fix X509_STORE locking: Every 'objs' access requires a lock (to 10128 accommodate for stack sorting, always a write lock!). 10129 10130 *Bodo Moeller* 10131 10132 * On some versions of WIN32 Heap32Next is very slow. This can cause 10133 excessive delays in the RAND_poll(): over a minute. As a workaround 10134 include a time check in the inner Heap32Next loop too. 10135 10136 *Steve Henson* 10137 10138 * The code that handled flushing of data in SSL/TLS originally used the 10139 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 10140 the problem outlined in PR#1949. The fix suggested there however can 10141 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 10142 of Apache). So instead simplify the code to flush unconditionally. 10143 This should be fine since flushing with no data to flush is a no op. 10144 10145 *Steve Henson* 10146 10147 * Handle TLS versions 2.0 and later properly and correctly use the 10148 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 10149 off ancient servers have a habit of sticking around for a while... 10150 10151 *Steve Henson* 10152 10153 * Modify compression code so it frees up structures without using the 10154 ex_data callbacks. This works around a problem where some applications 10155 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 10156 restarting) then use compression (e.g. SSL with compression) later. 10157 This results in significant per-connection memory leaks and 10158 has caused some security issues including CVE-2008-1678 and 10159 CVE-2009-4355. 10160 10161 *Steve Henson* 10162 10163 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 10164 change when encrypting or decrypting. 10165 10166 *Bodo Moeller* 10167 10168 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 10169 connect and renegotiate with servers which do not support RI. 10170 Until RI is more widely deployed this option is enabled by default. 10171 10172 *Steve Henson* 10173 10174 * Add "missing" ssl ctrls to clear options and mode. 10175 10176 *Steve Henson* 10177 10178 * If client attempts to renegotiate and doesn't support RI respond with 10179 a no_renegotiation alert as required by RFC5746. Some renegotiating 10180 TLS clients will continue a connection gracefully when they receive 10181 the alert. Unfortunately OpenSSL mishandled this alert and would hang 10182 waiting for a server hello which it will never receive. Now we treat a 10183 received no_renegotiation alert as a fatal error. This is because 10184 applications requesting a renegotiation might well expect it to succeed 10185 and would have no code in place to handle the server denying it so the 10186 only safe thing to do is to terminate the connection. 10187 10188 *Steve Henson* 10189 10190 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 10191 peer supports secure renegotiation and 0 otherwise. Print out peer 10192 renegotiation support in s_client/s_server. 10193 10194 *Steve Henson* 10195 10196 * Replace the highly broken and deprecated SPKAC certification method with 10197 the updated NID creation version. This should correctly handle UTF8. 10198 10199 *Steve Henson* 10200 10201 * Implement RFC5746. Re-enable renegotiation but require the extension 10202 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 10203 turns out to be a bad idea. It has been replaced by 10204 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 10205 SSL_CTX_set_options(). This is really not recommended unless you 10206 know what you are doing. 10207 10208 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson* 10209 10210 * Fixes to stateless session resumption handling. Use initial_ctx when 10211 issuing and attempting to decrypt tickets in case it has changed during 10212 servername handling. Use a non-zero length session ID when attempting 10213 stateless session resumption: this makes it possible to determine if 10214 a resumption has occurred immediately after receiving server hello 10215 (several places in OpenSSL subtly assume this) instead of later in 10216 the handshake. 10217 10218 *Steve Henson* 10219 10220 * The functions ENGINE_ctrl(), OPENSSL_isservice(), 10221 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 10222 fixes for a few places where the return code is not checked 10223 correctly. 10224 10225 *Julia Lawall <julia@diku.dk>* 10226 10227 * Add --strict-warnings option to Configure script to include devteam 10228 warnings in other configurations. 10229 10230 *Steve Henson* 10231 10232 * Add support for --libdir option and LIBDIR variable in makefiles. This 10233 makes it possible to install openssl libraries in locations which 10234 have names other than "lib", for example "/usr/lib64" which some 10235 systems need. 10236 10237 *Steve Henson, based on patch from Jeremy Utley* 10238 10239 * Don't allow the use of leading 0x80 in OIDs. This is a violation of 10240 X690 8.9.12 and can produce some misleading textual output of OIDs. 10241 10242 *Steve Henson, reported by Dan Kaminsky* 10243 10244 * Delete MD2 from algorithm tables. This follows the recommendation in 10245 several standards that it is not used in new applications due to 10246 several cryptographic weaknesses. For binary compatibility reasons 10247 the MD2 API is still compiled in by default. 10248 10249 *Steve Henson* 10250 10251 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 10252 and restored. 10253 10254 *Steve Henson* 10255 10256 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 10257 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 10258 clash. 10259 10260 *Guenter <lists@gknw.net>* 10261 10262 * Fix the server certificate chain building code to use X509_verify_cert(), 10263 it used to have an ad-hoc builder which was unable to cope with anything 10264 other than a simple chain. 10265 10266 *David Woodhouse <dwmw2@infradead.org>, Steve Henson* 10267 10268 * Don't check self signed certificate signatures in X509_verify_cert() 10269 by default (a flag can override this): it just wastes time without 10270 adding any security. As a useful side effect self signed root CAs 10271 with non-FIPS digests are now usable in FIPS mode. 10272 10273 *Steve Henson* 10274 10275 * In dtls1_process_out_of_seq_message() the check if the current message 10276 is already buffered was missing. For every new message was memory 10277 allocated, allowing an attacker to perform an denial of service attack 10278 with sending out of seq handshake messages until there is no memory 10279 left. Additionally every future message was buffered, even if the 10280 sequence number made no sense and would be part of another handshake. 10281 So only messages with sequence numbers less than 10 in advance will be 10282 buffered. ([CVE-2009-1378]) 10283 10284 *Robin Seggelmann, discovered by Daniel Mentz* 10285 10286 * Records are buffered if they arrive with a future epoch to be 10287 processed after finishing the corresponding handshake. There is 10288 currently no limitation to this buffer allowing an attacker to perform 10289 a DOS attack with sending records with future epochs until there is no 10290 memory left. This patch adds the pqueue_size() function to determine 10291 the size of a buffer and limits the record buffer to 100 entries. 10292 ([CVE-2009-1377]) 10293 10294 *Robin Seggelmann, discovered by Daniel Mentz* 10295 10296 * Keep a copy of frag->msg_header.frag_len so it can be used after the 10297 parent structure is freed. ([CVE-2009-1379]) 10298 10299 *Daniel Mentz* 10300 10301 * Handle non-blocking I/O properly in SSL_shutdown() call. 10302 10303 *Darryl Miles <darryl-mailinglists@netbauds.net>* 10304 10305 * Add `2.5.4.*` OIDs 10306 10307 *Ilya O. <vrghost@gmail.com>* 10308 10309### Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 10310 10311 * Disable renegotiation completely - this fixes a severe security 10312 problem ([CVE-2009-3555]) at the cost of breaking all 10313 renegotiation. Renegotiation can be re-enabled by setting 10314 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 10315 run-time. This is really not recommended unless you know what 10316 you're doing. 10317 10318 *Ben Laurie* 10319 10320### Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 10321 10322 * Don't set val to NULL when freeing up structures, it is freed up by 10323 underlying code. If `sizeof(void *) > sizeof(long)` this can result in 10324 zeroing past the valid field. ([CVE-2009-0789]) 10325 10326 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>* 10327 10328 * Fix bug where return value of CMS_SignerInfo_verify_content() was not 10329 checked correctly. This would allow some invalid signed attributes to 10330 appear to verify correctly. ([CVE-2009-0591]) 10331 10332 *Ivan Nestlerode <inestlerode@us.ibm.com>* 10333 10334 * Reject UniversalString and BMPString types with invalid lengths. This 10335 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 10336 a legal length. ([CVE-2009-0590]) 10337 10338 *Steve Henson* 10339 10340 * Set S/MIME signing as the default purpose rather than setting it 10341 unconditionally. This allows applications to override it at the store 10342 level. 10343 10344 *Steve Henson* 10345 10346 * Permit restricted recursion of ASN1 strings. This is needed in practice 10347 to handle some structures. 10348 10349 *Steve Henson* 10350 10351 * Improve efficiency of mem_gets: don't search whole buffer each time 10352 for a '\n' 10353 10354 *Jeremy Shapiro <jnshapir@us.ibm.com>* 10355 10356 * New -hex option for openssl rand. 10357 10358 *Matthieu Herrb* 10359 10360 * Print out UTF8String and NumericString when parsing ASN1. 10361 10362 *Steve Henson* 10363 10364 * Support NumericString type for name components. 10365 10366 *Steve Henson* 10367 10368 * Allow CC in the environment to override the automatically chosen 10369 compiler. Note that nothing is done to ensure flags work with the 10370 chosen compiler. 10371 10372 *Ben Laurie* 10373 10374### Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 10375 10376 * Properly check EVP_VerifyFinal() and similar return values 10377 ([CVE-2008-5077]). 10378 10379 *Ben Laurie, Bodo Moeller, Google Security Team* 10380 10381 * Enable TLS extensions by default. 10382 10383 *Ben Laurie* 10384 10385 * Allow the CHIL engine to be loaded, whether the application is 10386 multithreaded or not. (This does not release the developer from the 10387 obligation to set up the dynamic locking callbacks.) 10388 10389 *Sander Temme <sander@temme.net>* 10390 10391 * Use correct exit code if there is an error in dgst command. 10392 10393 *Steve Henson; problem pointed out by Roland Dirlewanger* 10394 10395 * Tweak Configure so that you need to say "experimental-jpake" to enable 10396 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 10397 10398 *Bodo Moeller* 10399 10400 * Add experimental JPAKE support, including demo authentication in 10401 s_client and s_server. 10402 10403 *Ben Laurie* 10404 10405 * Set the comparison function in v3_addr_canonize(). 10406 10407 *Rob Austein <sra@hactrn.net>* 10408 10409 * Add support for XMPP STARTTLS in s_client. 10410 10411 *Philip Paeps <philip@freebsd.org>* 10412 10413 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 10414 to ensure that even with this option, only ciphersuites in the 10415 server's preference list will be accepted. (Note that the option 10416 applies only when resuming a session, so the earlier behavior was 10417 just about the algorithm choice for symmetric cryptography.) 10418 10419 *Bodo Moeller* 10420 10421### Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 10422 10423 * Fix NULL pointer dereference if a DTLS server received 10424 ChangeCipherSpec as first record ([CVE-2009-1386]). 10425 10426 *PR #1679* 10427 10428 * Fix a state transition in s3_srvr.c and d1_srvr.c 10429 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`). 10430 10431 *Nagendra Modadugu* 10432 10433 * The fix in 0.9.8c that supposedly got rid of unsafe 10434 double-checked locking was incomplete for RSA blinding, 10435 addressing just one layer of what turns out to have been 10436 doubly unsafe triple-checked locking. 10437 10438 So now fix this for real by retiring the MONT_HELPER macro 10439 in crypto/rsa/rsa_eay.c. 10440 10441 *Bodo Moeller; problem pointed out by Marius Schilder* 10442 10443 * Various precautionary measures: 10444 10445 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 10446 10447 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 10448 (NB: This would require knowledge of the secret session ticket key 10449 to exploit, in which case you'd be SOL either way.) 10450 10451 - Change bn_nist.c so that it will properly handle input BIGNUMs 10452 outside the expected range. 10453 10454 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 10455 builds. 10456 10457 *Neel Mehta, Bodo Moeller* 10458 10459 * Allow engines to be "soft loaded" - i.e. optionally don't die if 10460 the load fails. Useful for distros. 10461 10462 *Ben Laurie and the FreeBSD team* 10463 10464 * Add support for Local Machine Keyset attribute in PKCS#12 files. 10465 10466 *Steve Henson* 10467 10468 * Fix BN_GF2m_mod_arr() top-bit cleanup code. 10469 10470 *Huang Ying* 10471 10472 * Expand ENGINE to support engine supplied SSL client certificate functions. 10473 10474 This work was sponsored by Logica. 10475 10476 *Steve Henson* 10477 10478 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 10479 keystores. Support for SSL/TLS client authentication too. 10480 Not compiled unless enable-capieng specified to Configure. 10481 10482 This work was sponsored by Logica. 10483 10484 *Steve Henson* 10485 10486 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using 10487 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 10488 attribute creation routines such as certificate requests and PKCS#12 10489 files. 10490 10491 *Steve Henson* 10492 10493### Changes between 0.9.8g and 0.9.8h [28 May 2008] 10494 10495 * Fix flaw if 'Server Key exchange message' is omitted from a TLS 10496 handshake which could lead to a client crash as found using the 10497 Codenomicon TLS test suite ([CVE-2008-1672]) 10498 10499 *Steve Henson, Mark Cox* 10500 10501 * Fix double free in TLS server name extensions which could lead to 10502 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891]) 10503 10504 *Joe Orton* 10505 10506 * Clear error queue in SSL_CTX_use_certificate_chain_file() 10507 10508 Clear the error queue to ensure that error entries left from 10509 older function calls do not interfere with the correct operation. 10510 10511 *Lutz Jaenicke, Erik de Castro Lopo* 10512 10513 * Remove root CA certificates of commercial CAs: 10514 10515 The OpenSSL project does not recommend any specific CA and does not 10516 have any policy with respect to including or excluding any CA. 10517 Therefore it does not make any sense to ship an arbitrary selection 10518 of root CA certificates with the OpenSSL software. 10519 10520 *Lutz Jaenicke* 10521 10522 * RSA OAEP patches to fix two separate invalid memory reads. 10523 The first one involves inputs when 'lzero' is greater than 10524 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 10525 before the beginning of from). The second one involves inputs where 10526 the 'db' section contains nothing but zeroes (there is a one-byte 10527 invalid read after the end of 'db'). 10528 10529 *Ivan Nestlerode <inestlerode@us.ibm.com>* 10530 10531 * Partial backport from 0.9.9-dev: 10532 10533 Introduce bn_mul_mont (dedicated Montgomery multiplication 10534 procedure) as a candidate for BIGNUM assembler implementation. 10535 While 0.9.9-dev uses assembler for various architectures, only 10536 x86_64 is available by default here in the 0.9.8 branch, and 10537 32-bit x86 is available through a compile-time setting. 10538 10539 To try the 32-bit x86 assembler implementation, use Configure 10540 option "enable-montasm" (which exists only for this backport). 10541 10542 As "enable-montasm" for 32-bit x86 disclaims code stability 10543 anyway, in this constellation we activate additional code 10544 backported from 0.9.9-dev for further performance improvements, 10545 namely BN_from_montgomery_word. (To enable this otherwise, 10546 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.) 10547 10548 *Andy Polyakov (backport partially by Bodo Moeller)* 10549 10550 * Add TLS session ticket callback. This allows an application to set 10551 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 10552 values. This is useful for key rollover for example where several key 10553 sets may exist with different names. 10554 10555 *Steve Henson* 10556 10557 * Reverse ENGINE-internal logic for caching default ENGINE handles. 10558 This was broken until now in 0.9.8 releases, such that the only way 10559 a registered ENGINE could be used (assuming it initialises 10560 successfully on the host) was to explicitly set it as the default 10561 for the relevant algorithms. This is in contradiction with 0.9.7 10562 behaviour and the documentation. With this fix, when an ENGINE is 10563 registered into a given algorithm's table of implementations, the 10564 'uptodate' flag is reset so that auto-discovery will be used next 10565 time a new context for that algorithm attempts to select an 10566 implementation. 10567 10568 *Ian Lister (tweaked by Geoff Thorpe)* 10569 10570 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 10571 implementation in the following ways: 10572 10573 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 10574 hard coded. 10575 10576 Lack of BER streaming support means one pass streaming processing is 10577 only supported if data is detached: setting the streaming flag is 10578 ignored for embedded content. 10579 10580 CMS support is disabled by default and must be explicitly enabled 10581 with the enable-cms configuration option. 10582 10583 *Steve Henson* 10584 10585 * Update the GMP engine glue to do direct copies between BIGNUM and 10586 mpz_t when openssl and GMP use the same limb size. Otherwise the 10587 existing "conversion via a text string export" trick is still used. 10588 10589 *Paul Sheer <paulsheer@gmail.com>* 10590 10591 * Zlib compression BIO. This is a filter BIO which compressed and 10592 uncompresses any data passed through it. 10593 10594 *Steve Henson* 10595 10596 * Add AES_wrap_key() and AES_unwrap_key() functions to implement 10597 RFC3394 compatible AES key wrapping. 10598 10599 *Steve Henson* 10600 10601 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 10602 sets string data without copying. X509_ALGOR_set0() and 10603 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 10604 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 10605 from an X509_ATTRIBUTE structure optionally checking it occurs only 10606 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 10607 data. 10608 10609 *Steve Henson* 10610 10611 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 10612 to get the expected BN_FLG_CONSTTIME behavior. 10613 10614 *Bodo Moeller (Google)* 10615 10616 * Netware support: 10617 10618 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 10619 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 10620 - added some more tests to do_tests.pl 10621 - fixed RunningProcess usage so that it works with newer LIBC NDKs too 10622 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 10623 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 10624 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 10625 - various changes to netware.pl to enable gcc-cross builds on Win32 10626 platform 10627 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 10628 - various changes to fix missing prototype warnings 10629 - fixed x86nasm.pl to create correct asm files for NASM COFF output 10630 - added AES, WHIRLPOOL and CPUID assembler code to build files 10631 - added missing AES assembler make rules to mk1mf.pl 10632 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply 10633 10634 *Guenter Knauf <eflash@gmx.net>* 10635 10636 * Implement certificate status request TLS extension defined in RFC3546. 10637 A client can set the appropriate parameters and receive the encoded 10638 OCSP response via a callback. A server can query the supplied parameters 10639 and set the encoded OCSP response in the callback. Add simplified examples 10640 to s_client and s_server. 10641 10642 *Steve Henson* 10643 10644### Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 10645 10646 * Fix various bugs: 10647 + Binary incompatibility of ssl_ctx_st structure 10648 + DTLS interoperation with non-compliant servers 10649 + Don't call get_session_cb() without proposed session 10650 + Fix ia64 assembler code 10651 10652 *Andy Polyakov, Steve Henson* 10653 10654### Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 10655 10656 * DTLS Handshake overhaul. There were longstanding issues with 10657 OpenSSL DTLS implementation, which were making it impossible for 10658 RFC 4347 compliant client to communicate with OpenSSL server. 10659 Unfortunately just fixing these incompatibilities would "cut off" 10660 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 10661 server keeps tolerating non RFC compliant syntax. The opposite is 10662 not true, 0.9.8f client can not communicate with earlier server. 10663 This update even addresses CVE-2007-4995. 10664 10665 *Andy Polyakov* 10666 10667 * Changes to avoid need for function casts in OpenSSL: some compilers 10668 (gcc 4.2 and later) reject their use. 10669 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 10670 Steve Henson* 10671 10672 * Add RFC4507 support to OpenSSL. This includes the corrections in 10673 RFC4507bis. The encrypted ticket format is an encrypted encoded 10674 SSL_SESSION structure, that way new session features are automatically 10675 supported. 10676 10677 If a client application caches session in an SSL_SESSION structure 10678 support is transparent because tickets are now stored in the encoded 10679 SSL_SESSION. 10680 10681 The SSL_CTX structure automatically generates keys for ticket 10682 protection in servers so again support should be possible 10683 with no application modification. 10684 10685 If a client or server wishes to disable RFC4507 support then the option 10686 SSL_OP_NO_TICKET can be set. 10687 10688 Add a TLS extension debugging callback to allow the contents of any client 10689 or server extensions to be examined. 10690 10691 This work was sponsored by Google. 10692 10693 *Steve Henson* 10694 10695 * Add initial support for TLS extensions, specifically for the server_name 10696 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 10697 have new members for a host name. The SSL data structure has an 10698 additional member `SSL_CTX *initial_ctx` so that new sessions can be 10699 stored in that context to allow for session resumption, even after the 10700 SSL has been switched to a new SSL_CTX in reaction to a client's 10701 server_name extension. 10702 10703 New functions (subject to change): 10704 10705 SSL_get_servername() 10706 SSL_get_servername_type() 10707 SSL_set_SSL_CTX() 10708 10709 New CTRL codes and macros (subject to change): 10710 10711 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 10712 - SSL_CTX_set_tlsext_servername_callback() 10713 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 10714 - SSL_CTX_set_tlsext_servername_arg() 10715 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 10716 10717 openssl s_client has a new '-servername ...' option. 10718 10719 openssl s_server has new options '-servername_host ...', '-cert2 ...', 10720 '-key2 ...', '-servername_fatal' (subject to change). This allows 10721 testing the HostName extension for a specific single host name ('-cert' 10722 and '-key' remain fallbacks for handshakes without HostName 10723 negotiation). If the unrecognized_name alert has to be sent, this by 10724 default is a warning; it becomes fatal with the '-servername_fatal' 10725 option. 10726 10727 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson* 10728 10729 * Add AES and SSE2 assembly language support to VC++ build. 10730 10731 *Steve Henson* 10732 10733 * Mitigate attack on final subtraction in Montgomery reduction. 10734 10735 *Andy Polyakov* 10736 10737 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 10738 (which previously caused an internal error). 10739 10740 *Bodo Moeller* 10741 10742 * Squeeze another 10% out of IGE mode when in != out. 10743 10744 *Ben Laurie* 10745 10746 * AES IGE mode speedup. 10747 10748 *Dean Gaudet (Google)* 10749 10750 * Add the Korean symmetric 128-bit cipher SEED (see 10751 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and 10752 add SEED ciphersuites from RFC 4162: 10753 10754 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 10755 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 10756 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 10757 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 10758 10759 To minimize changes between patchlevels in the OpenSSL 0.9.8 10760 series, SEED remains excluded from compilation unless OpenSSL 10761 is configured with 'enable-seed'. 10762 10763 *KISA, Bodo Moeller* 10764 10765 * Mitigate branch prediction attacks, which can be practical if a 10766 single processor is shared, allowing a spy process to extract 10767 information. For detailed background information, see 10768 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron, 10769 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 10770 and Necessary Software Countermeasures"). The core of the change 10771 are new versions BN_div_no_branch() and 10772 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 10773 respectively, which are slower, but avoid the security-relevant 10774 conditional branches. These are automatically called by BN_div() 10775 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 10776 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 10777 remove a conditional branch. 10778 10779 BN_FLG_CONSTTIME is the new name for the previous 10780 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 10781 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 10782 in the exponent causes BN_mod_exp_mont() to use the alternative 10783 implementation in BN_mod_exp_mont_consttime().) The old name 10784 remains as a deprecated alias. 10785 10786 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 10787 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 10788 constant-time implementations for more than just exponentiation. 10789 Here too the old name is kept as a deprecated alias. 10790 10791 BN_BLINDING_new() will now use BN_dup() for the modulus so that 10792 the BN_BLINDING structure gets an independent copy of the 10793 modulus. This means that the previous `BIGNUM *m` argument to 10794 BN_BLINDING_new() and to BN_BLINDING_create_param() now 10795 essentially becomes `const BIGNUM *m`, although we can't actually 10796 change this in the header file before 0.9.9. It allows 10797 RSA_setup_blinding() to use BN_with_flags() on the modulus to 10798 enable BN_FLG_CONSTTIME. 10799 10800 *Matthew D Wood (Intel Corp)* 10801 10802 * In the SSL/TLS server implementation, be strict about session ID 10803 context matching (which matters if an application uses a single 10804 external cache for different purposes). Previously, 10805 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 10806 set. This did ensure strict client verification, but meant that, 10807 with applications using a single external cache for quite 10808 different requirements, clients could circumvent ciphersuite 10809 restrictions for a given session ID context by starting a session 10810 in a different context. 10811 10812 *Bodo Moeller* 10813 10814 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 10815 a ciphersuite string such as "DEFAULT:RSA" cannot enable 10816 authentication-only ciphersuites. 10817 10818 *Bodo Moeller* 10819 10820 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 10821 not complete and could lead to a possible single byte overflow 10822 ([CVE-2007-5135]) [Ben Laurie] 10823 10824### Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 10825 10826 * Since AES128 and AES256 (and similarly Camellia128 and 10827 Camellia256) share a single mask bit in the logic of 10828 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 10829 kludge to work properly if AES128 is available and AES256 isn't 10830 (or if Camellia128 is available and Camellia256 isn't). 10831 10832 *Victor Duchovni* 10833 10834 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 10835 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 10836 When a point or a seed is encoded in a BIT STRING, we need to 10837 prevent the removal of trailing zero bits to get the proper DER 10838 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 10839 of a NamedBitList, for which trailing 0 bits need to be removed.) 10840 10841 *Bodo Moeller* 10842 10843 * Have SSL/TLS server implementation tolerate "mismatched" record 10844 protocol version while receiving ClientHello even if the 10845 ClientHello is fragmented. (The server can't insist on the 10846 particular protocol version it has chosen before the ServerHello 10847 message has informed the client about his choice.) 10848 10849 *Bodo Moeller* 10850 10851 * Add RFC 3779 support. 10852 10853 *Rob Austein for ARIN, Ben Laurie* 10854 10855 * Load error codes if they are not already present instead of using a 10856 static variable. This allows them to be cleanly unloaded and reloaded. 10857 Improve header file function name parsing. 10858 10859 *Steve Henson* 10860 10861 * extend SMTP and IMAP protocol emulation in s_client to use EHLO 10862 or CAPABILITY handshake as required by RFCs. 10863 10864 *Goetz Babin-Ebell* 10865 10866### Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 10867 10868 * Introduce limits to prevent malicious keys being able to 10869 cause a denial of service. ([CVE-2006-2940]) 10870 10871 *Steve Henson, Bodo Moeller* 10872 10873 * Fix ASN.1 parsing of certain invalid structures that can result 10874 in a denial of service. ([CVE-2006-2937]) [Steve Henson] 10875 10876 * Fix buffer overflow in SSL_get_shared_ciphers() function. 10877 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team] 10878 10879 * Fix SSL client code which could crash if connecting to a 10880 malicious SSLv2 server. ([CVE-2006-4343]) 10881 10882 *Tavis Ormandy and Will Drewry, Google Security Team* 10883 10884 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 10885 match only those. Before that, "AES256-SHA" would be interpreted 10886 as a pattern and match "AES128-SHA" too (since AES128-SHA got 10887 the same strength classification in 0.9.7h) as we currently only 10888 have a single AES bit in the ciphersuite description bitmap. 10889 That change, however, also applied to ciphersuite strings such as 10890 "RC4-MD5" that intentionally matched multiple ciphersuites -- 10891 namely, SSL 2.0 ciphersuites in addition to the more common ones 10892 from SSL 3.0/TLS 1.0. 10893 10894 So we change the selection algorithm again: Naming an explicit 10895 ciphersuite selects this one ciphersuite, and any other similar 10896 ciphersuite (same bitmap) from *other* protocol versions. 10897 Thus, "RC4-MD5" again will properly select both the SSL 2.0 10898 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 10899 10900 Since SSL 2.0 does not have any ciphersuites for which the 10901 128/256 bit distinction would be relevant, this works for now. 10902 The proper fix will be to use different bits for AES128 and 10903 AES256, which would have avoided the problems from the beginning; 10904 however, bits are scarce, so we can only do this in a new release 10905 (not just a patchlevel) when we can change the SSL_CIPHER 10906 definition to split the single 'unsigned long mask' bitmap into 10907 multiple values to extend the available space. 10908 10909 *Bodo Moeller* 10910 10911### Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 10912 10913 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 10914 ([CVE-2006-4339]) [Ben Laurie and Google Security Team] 10915 10916 * Add AES IGE and biIGE modes. 10917 10918 *Ben Laurie* 10919 10920 * Change the Unix randomness entropy gathering to use poll() when 10921 possible instead of select(), since the latter has some 10922 undesirable limitations. 10923 10924 *Darryl Miles via Richard Levitte and Bodo Moeller* 10925 10926 * Disable "ECCdraft" ciphersuites more thoroughly. Now special 10927 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 10928 cannot be implicitly activated as part of, e.g., the "AES" alias. 10929 However, please upgrade to OpenSSL 0.9.9[-dev] for 10930 non-experimental use of the ECC ciphersuites to get TLS extension 10931 support, which is required for curve and point format negotiation 10932 to avoid potential handshake problems. 10933 10934 *Bodo Moeller* 10935 10936 * Disable rogue ciphersuites: 10937 10938 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 10939 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 10940 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 10941 10942 The latter two were purportedly from 10943 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 10944 appear there. 10945 10946 Also deactivate the remaining ciphersuites from 10947 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 10948 unofficial, and the ID has long expired. 10949 10950 *Bodo Moeller* 10951 10952 * Fix RSA blinding Heisenbug (problems sometimes occurred on 10953 dual-core machines) and other potential thread-safety issues. 10954 10955 *Bodo Moeller* 10956 10957 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 10958 versions), which is now available for royalty-free use 10959 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>). 10960 Also, add Camellia TLS ciphersuites from RFC 4132. 10961 10962 To minimize changes between patchlevels in the OpenSSL 0.9.8 10963 series, Camellia remains excluded from compilation unless OpenSSL 10964 is configured with 'enable-camellia'. 10965 10966 *NTT* 10967 10968 * Disable the padding bug check when compression is in use. The padding 10969 bug check assumes the first packet is of even length, this is not 10970 necessarily true if compression is enabled and can result in false 10971 positives causing handshake failure. The actual bug test is ancient 10972 code so it is hoped that implementations will either have fixed it by 10973 now or any which still have the bug do not support compression. 10974 10975 *Steve Henson* 10976 10977### Changes between 0.9.8a and 0.9.8b [04 May 2006] 10978 10979 * When applying a cipher rule check to see if string match is an explicit 10980 cipher suite and only match that one cipher suite if it is. 10981 10982 *Steve Henson* 10983 10984 * Link in manifests for VC++ if needed. 10985 10986 *Austin Ziegler <halostatue@gmail.com>* 10987 10988 * Update support for ECC-based TLS ciphersuites according to 10989 draft-ietf-tls-ecc-12.txt with proposed changes (but without 10990 TLS extensions, which are supported starting with the 0.9.9 10991 branch, not in the OpenSSL 0.9.8 branch). 10992 10993 *Douglas Stebila* 10994 10995 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 10996 opaque EVP_CIPHER_CTX handling. 10997 10998 *Steve Henson* 10999 11000 * Fixes and enhancements to zlib compression code. We now only use 11001 "zlib1.dll" and use the default `__cdecl` calling convention on Win32 11002 to conform with the standards mentioned here: 11003 <http://www.zlib.net/DLL_FAQ.txt> 11004 Static zlib linking now works on Windows and the new --with-zlib-include 11005 --with-zlib-lib options to Configure can be used to supply the location 11006 of the headers and library. Gracefully handle case where zlib library 11007 can't be loaded. 11008 11009 *Steve Henson* 11010 11011 * Several fixes and enhancements to the OID generation code. The old code 11012 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 11013 handle numbers larger than ULONG_MAX, truncated printing and had a 11014 non standard OBJ_obj2txt() behaviour. 11015 11016 *Steve Henson* 11017 11018 * Add support for building of engines under engine/ as shared libraries 11019 under VC++ build system. 11020 11021 *Steve Henson* 11022 11023 * Corrected the numerous bugs in the Win32 path splitter in DSO. 11024 Hopefully, we will not see any false combination of paths any more. 11025 11026 *Richard Levitte* 11027 11028### Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 11029 11030 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 11031 (part of SSL_OP_ALL). This option used to disable the 11032 countermeasure against man-in-the-middle protocol-version 11033 rollback in the SSL 2.0 server implementation, which is a bad 11034 idea. ([CVE-2005-2969]) 11035 11036 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 11037 for Information Security, National Institute of Advanced Industrial 11038 Science and Technology [AIST], Japan)* 11039 11040 * Add two function to clear and return the verify parameter flags. 11041 11042 *Steve Henson* 11043 11044 * Keep cipherlists sorted in the source instead of sorting them at 11045 runtime, thus removing the need for a lock. 11046 11047 *Nils Larsch* 11048 11049 * Avoid some small subgroup attacks in Diffie-Hellman. 11050 11051 *Nick Mathewson and Ben Laurie* 11052 11053 * Add functions for well-known primes. 11054 11055 *Nick Mathewson* 11056 11057 * Extended Windows CE support. 11058 11059 *Satoshi Nakamura and Andy Polyakov* 11060 11061 * Initialize SSL_METHOD structures at compile time instead of during 11062 runtime, thus removing the need for a lock. 11063 11064 *Steve Henson* 11065 11066 * Make PKCS7_decrypt() work even if no certificate is supplied by 11067 attempting to decrypt each encrypted key in turn. Add support to 11068 smime utility. 11069 11070 *Steve Henson* 11071 11072### Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 11073 11074[NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 11075OpenSSL 0.9.8.] 11076 11077 * Add libcrypto.pc and libssl.pc for those who feel they need them. 11078 11079 *Richard Levitte* 11080 11081 * Change CA.sh and CA.pl so they don't bundle the CSR and the private 11082 key into the same file any more. 11083 11084 *Richard Levitte* 11085 11086 * Add initial support for Win64, both IA64 and AMD64/x64 flavors. 11087 11088 *Andy Polyakov* 11089 11090 * Add -utf8 command line and config file option to 'ca'. 11091 11092 *Stefan <stf@udoma.org* 11093 11094 * Removed the macro des_crypt(), as it seems to conflict with some 11095 libraries. Use DES_crypt(). 11096 11097 *Richard Levitte* 11098 11099 * Correct naming of the 'chil' and '4758cca' ENGINEs. This 11100 involves renaming the source and generated shared-libs for 11101 both. The engines will accept the corrected or legacy ids 11102 ('ncipher' and '4758_cca' respectively) when binding. NB, 11103 this only applies when building 'shared'. 11104 11105 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe* 11106 11107 * Add attribute functions to EVP_PKEY structure. Modify 11108 PKCS12_create() to recognize a CSP name attribute and 11109 use it. Make -CSP option work again in pkcs12 utility. 11110 11111 *Steve Henson* 11112 11113 * Add new functionality to the bn blinding code: 11114 - automatic re-creation of the BN_BLINDING parameters after 11115 a fixed number of uses (currently 32) 11116 - add new function for parameter creation 11117 - introduce flags to control the update behaviour of the 11118 BN_BLINDING parameters 11119 - hide BN_BLINDING structure 11120 Add a second BN_BLINDING slot to the RSA structure to improve 11121 performance when a single RSA object is shared among several 11122 threads. 11123 11124 *Nils Larsch* 11125 11126 * Add support for DTLS. 11127 11128 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie* 11129 11130 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 11131 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 11132 11133 *Walter Goulet* 11134 11135 * Remove buggy and incomplete DH cert support from 11136 ssl/ssl_rsa.c and ssl/s3_both.c 11137 11138 *Nils Larsch* 11139 11140 * Use SHA-1 instead of MD5 as the default digest algorithm for 11141 the `apps/openssl` commands. 11142 11143 *Nils Larsch* 11144 11145 * Compile clean with "-Wall -Wmissing-prototypes 11146 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 11147 DEBUG_SAFESTACK must also be set. 11148 11149 *Ben Laurie* 11150 11151 * Change ./Configure so that certain algorithms can be disabled by default. 11152 The new counterpiece to "no-xxx" is "enable-xxx". 11153 11154 The patented RC5 and MDC2 algorithms will now be disabled unless 11155 "enable-rc5" and "enable-mdc2", respectively, are specified. 11156 11157 (IDEA remains enabled despite being patented. This is because IDEA 11158 is frequently required for interoperability, and there is no license 11159 fee for non-commercial use. As before, "no-idea" can be used to 11160 avoid this algorithm.) 11161 11162 *Bodo Moeller* 11163 11164 * Add processing of proxy certificates (see RFC 3820). This work was 11165 sponsored by KTH (The Royal Institute of Technology in Stockholm) and 11166 EGEE (Enabling Grids for E-science in Europe). 11167 11168 *Richard Levitte* 11169 11170 * RC4 performance overhaul on modern architectures/implementations, such 11171 as Intel P4, IA-64 and AMD64. 11172 11173 *Andy Polyakov* 11174 11175 * New utility extract-section.pl. This can be used specify an alternative 11176 section number in a pod file instead of having to treat each file as 11177 a separate case in Makefile. This can be done by adding two lines to the 11178 pod file: 11179 11180 =for comment openssl_section:XXX 11181 11182 The blank line is mandatory. 11183 11184 *Steve Henson* 11185 11186 * New arguments -certform, -keyform and -pass for s_client and s_server 11187 to allow alternative format key and certificate files and passphrase 11188 sources. 11189 11190 *Steve Henson* 11191 11192 * New structure X509_VERIFY_PARAM which combines current verify parameters, 11193 update associated structures and add various utility functions. 11194 11195 Add new policy related verify parameters, include policy checking in 11196 standard verify code. Enhance 'smime' application with extra parameters 11197 to support policy checking and print out. 11198 11199 *Steve Henson* 11200 11201 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3 11202 Nehemiah processors. These extensions support AES encryption in hardware 11203 as well as RNG (though RNG support is currently disabled). 11204 11205 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov* 11206 11207 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally). 11208 11209 *Geoff Thorpe* 11210 11211 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 11212 11213 *Andy Polyakov and a number of other people* 11214 11215 * Improved PowerPC platform support. Most notably BIGNUM assembler 11216 implementation contributed by IBM. 11217 11218 *Suresh Chari, Peter Waltenberg, Andy Polyakov* 11219 11220 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 11221 exponent rather than 'unsigned long'. There is a corresponding change to 11222 the new 'rsa_keygen' element of the RSA_METHOD structure. 11223 11224 *Jelte Jansen, Geoff Thorpe* 11225 11226 * Functionality for creating the initial serial number file is now 11227 moved from CA.pl to the 'ca' utility with a new option -create_serial. 11228 11229 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 11230 number file to 1, which is bound to cause problems. To avoid 11231 the problems while respecting compatibility between different 0.9.7 11232 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 11233 CA.pl for serial number initialization. With the new release 0.9.8, 11234 we can fix the problem directly in the 'ca' utility.) 11235 11236 *Steve Henson* 11237 11238 * Reduced header interdependencies by declaring more opaque objects in 11239 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 11240 give fewer recursive includes, which could break lazy source code - so 11241 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 11242 developers should define this symbol when building and using openssl to 11243 ensure they track the recommended behaviour, interfaces, [etc], but 11244 backwards-compatible behaviour prevails when this isn't defined. 11245 11246 *Geoff Thorpe* 11247 11248 * New function X509_POLICY_NODE_print() which prints out policy nodes. 11249 11250 *Steve Henson* 11251 11252 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 11253 This will generate a random key of the appropriate length based on the 11254 cipher context. The EVP_CIPHER can provide its own random key generation 11255 routine to support keys of a specific form. This is used in the des and 11256 3des routines to generate a key of the correct parity. Update S/MIME 11257 code to use new functions and hence generate correct parity DES keys. 11258 Add EVP_CHECK_DES_KEY #define to return an error if the key is not 11259 valid (weak or incorrect parity). 11260 11261 *Steve Henson* 11262 11263 * Add a local set of CRLs that can be used by X509_verify_cert() as well 11264 as looking them up. This is useful when the verified structure may contain 11265 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 11266 present unless the new PKCS7_NO_CRL flag is asserted. 11267 11268 *Steve Henson* 11269 11270 * Extend ASN1 oid configuration module. It now additionally accepts the 11271 syntax: 11272 11273 shortName = some long name, 1.2.3.4 11274 11275 *Steve Henson* 11276 11277 * Reimplemented the BN_CTX implementation. There is now no more static 11278 limitation on the number of variables it can handle nor the depth of the 11279 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 11280 information can now expand as required, and rather than having a single 11281 static array of bignums, BN_CTX now uses a linked-list of such arrays 11282 allowing it to expand on demand whilst maintaining the usefulness of 11283 BN_CTX's "bundling". 11284 11285 *Geoff Thorpe* 11286 11287 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 11288 to allow all RSA operations to function using a single BN_CTX. 11289 11290 *Geoff Thorpe* 11291 11292 * Preliminary support for certificate policy evaluation and checking. This 11293 is initially intended to pass the tests outlined in "Conformance Testing 11294 of Relying Party Client Certificate Path Processing Logic" v1.07. 11295 11296 *Steve Henson* 11297 11298 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 11299 remained unused and not that useful. A variety of other little bignum 11300 tweaks and fixes have also been made continuing on from the audit (see 11301 below). 11302 11303 *Geoff Thorpe* 11304 11305 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with 11306 associated ASN1, EVP and SSL functions and old ASN1 macros. 11307 11308 *Richard Levitte* 11309 11310 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 11311 and this should never fail. So the return value from the use of 11312 BN_set_word() (which can fail due to needless expansion) is now deprecated; 11313 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 11314 11315 *Geoff Thorpe* 11316 11317 * BN_CTX_get() should return zero-valued bignums, providing the same 11318 initialised value as BN_new(). 11319 11320 *Geoff Thorpe, suggested by Ulf Möller* 11321 11322 * Support for inhibitAnyPolicy certificate extension. 11323 11324 *Steve Henson* 11325 11326 * An audit of the BIGNUM code is underway, for which debugging code is 11327 enabled when BN_DEBUG is defined. This makes stricter enforcements on what 11328 is considered valid when processing BIGNUMs, and causes execution to 11329 assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 11330 further steps are taken to deliberately pollute unused data in BIGNUM 11331 structures to try and expose faulty code further on. For now, openssl will 11332 (in its default mode of operation) continue to tolerate the inconsistent 11333 forms that it has tolerated in the past, but authors and packagers should 11334 consider trying openssl and their own applications when compiled with 11335 these debugging symbols defined. It will help highlight potential bugs in 11336 their own code, and will improve the test coverage for OpenSSL itself. At 11337 some point, these tighter rules will become openssl's default to improve 11338 maintainability, though the assert()s and other overheads will remain only 11339 in debugging configurations. See bn.h for more details. 11340 11341 *Geoff Thorpe, Nils Larsch, Ulf Möller* 11342 11343 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 11344 that can only be obtained through BN_CTX_new() (which implicitly 11345 initialises it). The presence of this function only made it possible 11346 to overwrite an existing structure (and cause memory leaks). 11347 11348 *Geoff Thorpe* 11349 11350 * Because of the callback-based approach for implementing LHASH as a 11351 template type, lh_insert() adds opaque objects to hash-tables and 11352 lh_doall() or lh_doall_arg() are typically used with a destructor callback 11353 to clean up those corresponding objects before destroying the hash table 11354 (and losing the object pointers). So some over-zealous constifications in 11355 LHASH have been relaxed so that lh_insert() does not take (nor store) the 11356 objects as "const" and the `lh_doall[_arg]` callback wrappers are not 11357 prototyped to have "const" restrictions on the object pointers they are 11358 given (and so aren't required to cast them away any more). 11359 11360 *Geoff Thorpe* 11361 11362 * The tmdiff.h API was so ugly and minimal that our own timing utility 11363 (speed) prefers to use its own implementation. The two implementations 11364 haven't been consolidated as yet (volunteers?) but the tmdiff API has had 11365 its object type properly exposed (MS_TM) instead of casting to/from 11366 `char *`. This may still change yet if someone realises MS_TM and 11367 `ms_time_***` 11368 aren't necessarily the greatest nomenclatures - but this is what was used 11369 internally to the implementation so I've used that for now. 11370 11371 *Geoff Thorpe* 11372 11373 * Ensure that deprecated functions do not get compiled when 11374 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 11375 the self-tests were still using deprecated key-generation functions so 11376 these have been updated also. 11377 11378 *Geoff Thorpe* 11379 11380 * Reorganise PKCS#7 code to separate the digest location functionality 11381 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest(). 11382 New function PKCS7_set_digest() to set the digest type for PKCS#7 11383 digestedData type. Add additional code to correctly generate the 11384 digestedData type and add support for this type in PKCS7 initialization 11385 functions. 11386 11387 *Steve Henson* 11388 11389 * New function PKCS7_set0_type_other() this initializes a PKCS7 11390 structure of type "other". 11391 11392 *Steve Henson* 11393 11394 * Fix prime generation loop in crypto/bn/bn_prime.pl by making 11395 sure the loop does correctly stop and breaking ("division by zero") 11396 modulus operations are not performed. The (pre-generated) prime 11397 table crypto/bn/bn_prime.h was already correct, but it could not be 11398 re-generated on some platforms because of the "division by zero" 11399 situation in the script. 11400 11401 *Ralf S. Engelschall* 11402 11403 * Update support for ECC-based TLS ciphersuites according to 11404 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 11405 SHA-1 now is only used for "small" curves (where the 11406 representation of a field element takes up to 24 bytes); for 11407 larger curves, the field element resulting from ECDH is directly 11408 used as premaster secret. 11409 11410 *Douglas Stebila (Sun Microsystems Laboratories)* 11411 11412 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 11413 curve secp160r1 to the tests. 11414 11415 *Douglas Stebila (Sun Microsystems Laboratories)* 11416 11417 * Add the possibility to load symbols globally with DSO. 11418 11419 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte* 11420 11421 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 11422 control of the error stack. 11423 11424 *Richard Levitte* 11425 11426 * Add support for STORE in ENGINE. 11427 11428 *Richard Levitte* 11429 11430 * Add the STORE type. The intention is to provide a common interface 11431 to certificate and key stores, be they simple file-based stores, or 11432 HSM-type store, or LDAP stores, or... 11433 NOTE: The code is currently UNTESTED and isn't really used anywhere. 11434 11435 *Richard Levitte* 11436 11437 * Add a generic structure called OPENSSL_ITEM. This can be used to 11438 pass a list of arguments to any function as well as provide a way 11439 for a function to pass data back to the caller. 11440 11441 *Richard Levitte* 11442 11443 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 11444 works like BUF_strdup() but can be used to duplicate a portion of 11445 a string. The copy gets NUL-terminated. BUF_memdup() duplicates 11446 a memory area. 11447 11448 *Richard Levitte* 11449 11450 * Add the function sk_find_ex() which works like sk_find(), but will 11451 return an index to an element even if an exact match couldn't be 11452 found. The index is guaranteed to point at the element where the 11453 searched-for key would be inserted to preserve sorting order. 11454 11455 *Richard Levitte* 11456 11457 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 11458 takes an extra flags argument for optional functionality. Currently, 11459 the following flags are defined: 11460 11461 OBJ_BSEARCH_VALUE_ON_NOMATCH 11462 This one gets OBJ_bsearch_ex() to return a pointer to the first 11463 element where the comparing function returns a negative or zero 11464 number. 11465 11466 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 11467 This one gets OBJ_bsearch_ex() to return a pointer to the first 11468 element where the comparing function returns zero. This is useful 11469 if there are more than one element where the comparing function 11470 returns zero. 11471 11472 *Richard Levitte* 11473 11474 * Make it possible to create self-signed certificates with 'openssl ca' 11475 in such a way that the self-signed certificate becomes part of the 11476 CA database and uses the same mechanisms for serial number generation 11477 as all other certificate signing. The new flag '-selfsign' enables 11478 this functionality. Adapt CA.sh and CA.pl.in. 11479 11480 *Richard Levitte* 11481 11482 * Add functionality to check the public key of a certificate request 11483 against a given private. This is useful to check that a certificate 11484 request can be signed by that key (self-signing). 11485 11486 *Richard Levitte* 11487 11488 * Make it possible to have multiple active certificates with the same 11489 subject in the CA index file. This is done only if the keyword 11490 'unique_subject' is set to 'no' in the main CA section (default 11491 if 'CA_default') of the configuration file. The value is saved 11492 with the database itself in a separate index attribute file, 11493 named like the index file with '.attr' appended to the name. 11494 11495 *Richard Levitte* 11496 11497 * Generate multi-valued AVAs using '+' notation in config files for 11498 req and dirName. 11499 11500 *Steve Henson* 11501 11502 * Support for nameConstraints certificate extension. 11503 11504 *Steve Henson* 11505 11506 * Support for policyConstraints certificate extension. 11507 11508 *Steve Henson* 11509 11510 * Support for policyMappings certificate extension. 11511 11512 *Steve Henson* 11513 11514 * Make sure the default DSA_METHOD implementation only uses its 11515 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 11516 and change its own handlers to be NULL so as to remove unnecessary 11517 indirection. This lets alternative implementations fallback to the 11518 default implementation more easily. 11519 11520 *Geoff Thorpe* 11521 11522 * Support for directoryName in GeneralName related extensions 11523 in config files. 11524 11525 *Steve Henson* 11526 11527 * Make it possible to link applications using Makefile.shared. 11528 Make that possible even when linking against static libraries! 11529 11530 *Richard Levitte* 11531 11532 * Support for single pass processing for S/MIME signing. This now 11533 means that S/MIME signing can be done from a pipe, in addition 11534 cleartext signing (multipart/signed type) is effectively streaming 11535 and the signed data does not need to be all held in memory. 11536 11537 This is done with a new flag PKCS7_STREAM. When this flag is set 11538 PKCS7_sign() only initializes the PKCS7 structure and the actual signing 11539 is done after the data is output (and digests calculated) in 11540 SMIME_write_PKCS7(). 11541 11542 *Steve Henson* 11543 11544 * Add full support for -rpath/-R, both in shared libraries and 11545 applications, at least on the platforms where it's known how 11546 to do it. 11547 11548 *Richard Levitte* 11549 11550 * In crypto/ec/ec_mult.c, implement fast point multiplication with 11551 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 11552 will now compute a table of multiples of the generator that 11553 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 11554 faster (notably in the case of a single point multiplication, 11555 scalar * generator). 11556 11557 *Nils Larsch, Bodo Moeller* 11558 11559 * IPv6 support for certificate extensions. The various extensions 11560 which use the IP:a.b.c.d can now take IPv6 addresses using the 11561 formats of RFC1884 2.2 . IPv6 addresses are now also displayed 11562 correctly. 11563 11564 *Steve Henson* 11565 11566 * Added an ENGINE that implements RSA by performing private key 11567 exponentiations with the GMP library. The conversions to and from 11568 GMP's mpz_t format aren't optimised nor are any montgomery forms 11569 cached, and on x86 it appears OpenSSL's own performance has caught up. 11570 However there are likely to be other architectures where GMP could 11571 provide a boost. This ENGINE is not built in by default, but it can be 11572 specified at Configure time and should be accompanied by the necessary 11573 linker additions, eg; 11574 ./config -DOPENSSL_USE_GMP -lgmp 11575 11576 *Geoff Thorpe* 11577 11578 * "openssl engine" will not display ENGINE/DSO load failure errors when 11579 testing availability of engines with "-t" - the old behaviour is 11580 produced by increasing the feature's verbosity with "-tt". 11581 11582 *Geoff Thorpe* 11583 11584 * ECDSA routines: under certain error conditions uninitialized BN objects 11585 could be freed. Solution: make sure initialization is performed early 11586 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 11587 via PR#459) 11588 11589 *Lutz Jaenicke* 11590 11591 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 11592 and DH_METHOD (eg. by ENGINE implementations) to override the normal 11593 software implementations. For DSA and DH, parameter generation can 11594 also be overridden by providing the appropriate method callbacks. 11595 11596 *Geoff Thorpe* 11597 11598 * Change the "progress" mechanism used in key-generation and 11599 primality testing to functions that take a new BN_GENCB pointer in 11600 place of callback/argument pairs. The new API functions have `_ex` 11601 postfixes and the older functions are reimplemented as wrappers for 11602 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 11603 declarations of the old functions to help (graceful) attempts to 11604 migrate to the new functions. Also, the new key-generation API 11605 functions operate on a caller-supplied key-structure and return 11606 success/failure rather than returning a key or NULL - this is to 11607 help make "keygen" another member function of RSA_METHOD etc. 11608 11609 Example for using the new callback interface: 11610 11611 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 11612 void *my_arg = ...; 11613 BN_GENCB my_cb; 11614 11615 BN_GENCB_set(&my_cb, my_callback, my_arg); 11616 11617 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 11618 /* For the meaning of a, b in calls to my_callback(), see the 11619 * documentation of the function that calls the callback. 11620 * cb will point to my_cb; my_arg can be retrieved as cb->arg. 11621 * my_callback should return 1 if it wants BN_is_prime_ex() 11622 * to continue, or 0 to stop. 11623 */ 11624 11625 *Geoff Thorpe* 11626 11627 * Change the ZLIB compression method to be stateful, and make it 11628 available to TLS with the number defined in 11629 draft-ietf-tls-compression-04.txt. 11630 11631 *Richard Levitte* 11632 11633 * Add the ASN.1 structures and functions for CertificatePair, which 11634 is defined as follows (according to X.509_4thEditionDraftV6.pdf): 11635 11636 CertificatePair ::= SEQUENCE { 11637 forward [0] Certificate OPTIONAL, 11638 reverse [1] Certificate OPTIONAL, 11639 -- at least one of the pair shall be present -- } 11640 11641 Also implement the PEM functions to read and write certificate 11642 pairs, and defined the PEM tag as "CERTIFICATE PAIR". 11643 11644 This needed to be defined, mostly for the sake of the LDAP 11645 attribute crossCertificatePair, but may prove useful elsewhere as 11646 well. 11647 11648 *Richard Levitte* 11649 11650 * Make it possible to inhibit symlinking of shared libraries in 11651 Makefile.shared, for Cygwin's sake. 11652 11653 *Richard Levitte* 11654 11655 * Extend the BIGNUM API by creating a function 11656 void BN_set_negative(BIGNUM *a, int neg); 11657 and a macro that behave like 11658 int BN_is_negative(const BIGNUM *a); 11659 11660 to avoid the need to access 'a->neg' directly in applications. 11661 11662 *Nils Larsch* 11663 11664 * Implement fast modular reduction for pseudo-Mersenne primes 11665 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 11666 EC_GROUP_new_curve_GFp() will now automatically use this 11667 if applicable. 11668 11669 *Nils Larsch <nla@trustcenter.de>* 11670 11671 * Add new lock type (CRYPTO_LOCK_BN). 11672 11673 *Bodo Moeller* 11674 11675 * Change the ENGINE framework to automatically load engines 11676 dynamically from specific directories unless they could be 11677 found to already be built in or loaded. Move all the 11678 current engines except for the cryptodev one to a new 11679 directory engines/. 11680 The engines in engines/ are built as shared libraries if 11681 the "shared" options was given to ./Configure or ./config. 11682 Otherwise, they are inserted in libcrypto.a. 11683 /usr/local/ssl/engines is the default directory for dynamic 11684 engines, but that can be overridden at configure time through 11685 the usual use of --prefix and/or --openssldir, and at run 11686 time with the environment variable OPENSSL_ENGINES. 11687 11688 *Geoff Thorpe and Richard Levitte* 11689 11690 * Add Makefile.shared, a helper makefile to build shared 11691 libraries. Adapt Makefile.org. 11692 11693 *Richard Levitte* 11694 11695 * Add version info to Win32 DLLs. 11696 11697 *Peter 'Luna' Runestig" <peter@runestig.com>* 11698 11699 * Add new 'medium level' PKCS#12 API. Certificates and keys 11700 can be added using this API to created arbitrary PKCS#12 11701 files while avoiding the low-level API. 11702 11703 New options to PKCS12_create(), key or cert can be NULL and 11704 will then be omitted from the output file. The encryption 11705 algorithm NIDs can be set to -1 for no encryption, the mac 11706 iteration count can be set to 0 to omit the mac. 11707 11708 Enhance pkcs12 utility by making the -nokeys and -nocerts 11709 options work when creating a PKCS#12 file. New option -nomac 11710 to omit the mac, NONE can be set for an encryption algorithm. 11711 New code is modified to use the enhanced PKCS12_create() 11712 instead of the low-level API. 11713 11714 *Steve Henson* 11715 11716 * Extend ASN1 encoder to support indefinite length constructed 11717 encoding. This can output sequences tags and octet strings in 11718 this form. Modify pk7_asn1.c to support indefinite length 11719 encoding. This is experimental and needs additional code to 11720 be useful, such as an ASN1 bio and some enhanced streaming 11721 PKCS#7 code. 11722 11723 Extend template encode functionality so that tagging is passed 11724 down to the template encoder. 11725 11726 *Steve Henson* 11727 11728 * Let 'openssl req' fail if an argument to '-newkey' is not 11729 recognized instead of using RSA as a default. 11730 11731 *Bodo Moeller* 11732 11733 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 11734 As these are not official, they are not included in "ALL"; 11735 the "ECCdraft" ciphersuite group alias can be used to select them. 11736 11737 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)* 11738 11739 * Add ECDH engine support. 11740 11741 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)* 11742 11743 * Add ECDH in new directory crypto/ecdh/. 11744 11745 *Douglas Stebila (Sun Microsystems Laboratories)* 11746 11747 * Let BN_rand_range() abort with an error after 100 iterations 11748 without success (which indicates a broken PRNG). 11749 11750 *Bodo Moeller* 11751 11752 * Change BN_mod_sqrt() so that it verifies that the input value 11753 is really the square of the return value. (Previously, 11754 BN_mod_sqrt would show GIGO behaviour.) 11755 11756 *Bodo Moeller* 11757 11758 * Add named elliptic curves over binary fields from X9.62, SECG, 11759 and WAP/WTLS; add OIDs that were still missing. 11760 11761 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)* 11762 11763 * Extend the EC library for elliptic curves over binary fields 11764 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 11765 New EC_METHOD: 11766 11767 EC_GF2m_simple_method 11768 11769 New API functions: 11770 11771 EC_GROUP_new_curve_GF2m 11772 EC_GROUP_set_curve_GF2m 11773 EC_GROUP_get_curve_GF2m 11774 EC_POINT_set_affine_coordinates_GF2m 11775 EC_POINT_get_affine_coordinates_GF2m 11776 EC_POINT_set_compressed_coordinates_GF2m 11777 11778 Point compression for binary fields is disabled by default for 11779 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 11780 enable it). 11781 11782 As binary polynomials are represented as BIGNUMs, various members 11783 of the EC_GROUP and EC_POINT data structures can be shared 11784 between the implementations for prime fields and binary fields; 11785 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m) 11786 are essentially identical to their `..._GFp` counterparts. 11787 (For simplicity, the `..._GFp` prefix has been dropped from 11788 various internal method names.) 11789 11790 An internal 'field_div' method (similar to 'field_mul' and 11791 'field_sqr') has been added; this is used only for binary fields. 11792 11793 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)* 11794 11795 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 11796 through methods ('mul', 'precompute_mult'). 11797 11798 The generic implementations (now internally called 'ec_wNAF_mul' 11799 and 'ec_wNAF_precomputed_mult') remain the default if these 11800 methods are undefined. 11801 11802 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)* 11803 11804 * New function EC_GROUP_get_degree, which is defined through 11805 EC_METHOD. For curves over prime fields, this returns the bit 11806 length of the modulus. 11807 11808 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)* 11809 11810 * New functions EC_GROUP_dup, EC_POINT_dup. 11811 (These simply call ..._new and ..._copy). 11812 11813 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)* 11814 11815 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 11816 Polynomials are represented as BIGNUMs (where the sign bit is not 11817 used) in the following functions [macros]: 11818 11819 BN_GF2m_add 11820 BN_GF2m_sub [= BN_GF2m_add] 11821 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 11822 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 11823 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 11824 BN_GF2m_mod_inv 11825 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 11826 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 11827 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 11828 BN_GF2m_cmp [= BN_ucmp] 11829 11830 (Note that only the 'mod' functions are actually for fields GF(2^m). 11831 BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 11832 11833 For some functions, an the irreducible polynomial defining a 11834 field can be given as an 'unsigned int[]' with strictly 11835 decreasing elements giving the indices of those bits that are set; 11836 i.e., p[] represents the polynomial 11837 f(t) = t^p[0] + t^p[1] + ... + t^p[k] 11838 where 11839 p[0] > p[1] > ... > p[k] = 0. 11840 This applies to the following functions: 11841 11842 BN_GF2m_mod_arr 11843 BN_GF2m_mod_mul_arr 11844 BN_GF2m_mod_sqr_arr 11845 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 11846 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 11847 BN_GF2m_mod_exp_arr 11848 BN_GF2m_mod_sqrt_arr 11849 BN_GF2m_mod_solve_quad_arr 11850 BN_GF2m_poly2arr 11851 BN_GF2m_arr2poly 11852 11853 Conversion can be performed by the following functions: 11854 11855 BN_GF2m_poly2arr 11856 BN_GF2m_arr2poly 11857 11858 bntest.c has additional tests for binary polynomial arithmetic. 11859 11860 Two implementations for BN_GF2m_mod_div() are available. 11861 The default algorithm simply uses BN_GF2m_mod_inv() and 11862 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 11863 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 11864 copyright notice in crypto/bn/bn_gf2m.c before enabling it). 11865 11866 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)* 11867 11868 * Add new error code 'ERR_R_DISABLED' that can be used when some 11869 functionality is disabled at compile-time. 11870 11871 *Douglas Stebila <douglas.stebila@sun.com>* 11872 11873 * Change default behaviour of 'openssl asn1parse' so that more 11874 information is visible when viewing, e.g., a certificate: 11875 11876 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 11877 mode the content of non-printable OCTET STRINGs is output in a 11878 style similar to INTEGERs, but with '[HEX DUMP]' prepended to 11879 avoid the appearance of a printable string. 11880 11881 *Nils Larsch <nla@trustcenter.de>* 11882 11883 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 11884 functions 11885 EC_GROUP_set_asn1_flag() 11886 EC_GROUP_get_asn1_flag() 11887 EC_GROUP_set_point_conversion_form() 11888 EC_GROUP_get_point_conversion_form() 11889 These control ASN1 encoding details: 11890 - Curves (i.e., groups) are encoded explicitly unless asn1_flag 11891 has been set to OPENSSL_EC_NAMED_CURVE. 11892 - Points are encoded in uncompressed form by default; options for 11893 asn1_for are as for point2oct, namely 11894 POINT_CONVERSION_COMPRESSED 11895 POINT_CONVERSION_UNCOMPRESSED 11896 POINT_CONVERSION_HYBRID 11897 11898 Also add 'seed' and 'seed_len' members to EC_GROUP with access 11899 functions 11900 EC_GROUP_set_seed() 11901 EC_GROUP_get0_seed() 11902 EC_GROUP_get_seed_len() 11903 This is used only for ASN1 purposes (so far). 11904 11905 *Nils Larsch <nla@trustcenter.de>* 11906 11907 * Add 'field_type' member to EC_METHOD, which holds the NID 11908 of the appropriate field type OID. The new function 11909 EC_METHOD_get_field_type() returns this value. 11910 11911 *Nils Larsch <nla@trustcenter.de>* 11912 11913 * Add functions 11914 EC_POINT_point2bn() 11915 EC_POINT_bn2point() 11916 EC_POINT_point2hex() 11917 EC_POINT_hex2point() 11918 providing useful interfaces to EC_POINT_point2oct() and 11919 EC_POINT_oct2point(). 11920 11921 *Nils Larsch <nla@trustcenter.de>* 11922 11923 * Change internals of the EC library so that the functions 11924 EC_GROUP_set_generator() 11925 EC_GROUP_get_generator() 11926 EC_GROUP_get_order() 11927 EC_GROUP_get_cofactor() 11928 are implemented directly in crypto/ec/ec_lib.c and not dispatched 11929 to methods, which would lead to unnecessary code duplication when 11930 adding different types of curves. 11931 11932 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller* 11933 11934 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 11935 arithmetic, and such that modified wNAFs are generated 11936 (which avoid length expansion in many cases). 11937 11938 *Bodo Moeller* 11939 11940 * Add a function EC_GROUP_check_discriminant() (defined via 11941 EC_METHOD) that verifies that the curve discriminant is non-zero. 11942 11943 Add a function EC_GROUP_check() that makes some sanity tests 11944 on a EC_GROUP, its generator and order. This includes 11945 EC_GROUP_check_discriminant(). 11946 11947 *Nils Larsch <nla@trustcenter.de>* 11948 11949 * Add ECDSA in new directory crypto/ecdsa/. 11950 11951 Add applications 'openssl ecparam' and 'openssl ecdsa' 11952 (these are based on 'openssl dsaparam' and 'openssl dsa'). 11953 11954 ECDSA support is also included in various other files across the 11955 library. Most notably, 11956 - 'openssl req' now has a '-newkey ecdsa:file' option; 11957 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 11958 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 11959 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 11960 them suitable for ECDSA where domain parameters must be 11961 extracted before the specific public key; 11962 - ECDSA engine support has been added. 11963 11964 *Nils Larsch <nla@trustcenter.de>* 11965 11966 * Include some named elliptic curves, and add OIDs from X9.62, 11967 SECG, and WAP/WTLS. Each curve can be obtained from the new 11968 function 11969 EC_GROUP_new_by_curve_name(), 11970 and the list of available named curves can be obtained with 11971 EC_get_builtin_curves(). 11972 Also add a 'curve_name' member to EC_GROUP objects, which can be 11973 accessed via 11974 EC_GROUP_set_curve_name() 11975 EC_GROUP_get_curve_name() 11976 11977 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller* 11978 11979 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 11980 was actually never needed) and in BN_mul(). The removal in BN_mul() 11981 required a small change in bn_mul_part_recursive() and the addition 11982 of the functions bn_cmp_part_words(), bn_sub_part_words() and 11983 bn_add_part_words(), which do the same thing as bn_cmp_words(), 11984 bn_sub_words() and bn_add_words() except they take arrays with 11985 differing sizes. 11986 11987 *Richard Levitte* 11988 11989### Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 11990 11991 * Cleanse PEM buffers before freeing them since they may contain 11992 sensitive data. 11993 11994 *Benjamin Bennett <ben@psc.edu>* 11995 11996 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 11997 a ciphersuite string such as "DEFAULT:RSA" cannot enable 11998 authentication-only ciphersuites. 11999 12000 *Bodo Moeller* 12001 12002 * Since AES128 and AES256 share a single mask bit in the logic of 12003 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 12004 kludge to work properly if AES128 is available and AES256 isn't. 12005 12006 *Victor Duchovni* 12007 12008 * Expand security boundary to match 1.1.1 module. 12009 12010 *Steve Henson* 12011 12012 * Remove redundant features: hash file source, editing of test vectors 12013 modify fipsld to use external fips_premain.c signature. 12014 12015 *Steve Henson* 12016 12017 * New perl script mkfipsscr.pl to create shell scripts or batch files to 12018 run algorithm test programs. 12019 12020 *Steve Henson* 12021 12022 * Make algorithm test programs more tolerant of whitespace. 12023 12024 *Steve Henson* 12025 12026 * Have SSL/TLS server implementation tolerate "mismatched" record 12027 protocol version while receiving ClientHello even if the 12028 ClientHello is fragmented. (The server can't insist on the 12029 particular protocol version it has chosen before the ServerHello 12030 message has informed the client about his choice.) 12031 12032 *Bodo Moeller* 12033 12034 * Load error codes if they are not already present instead of using a 12035 static variable. This allows them to be cleanly unloaded and reloaded. 12036 12037 *Steve Henson* 12038 12039### Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 12040 12041 * Introduce limits to prevent malicious keys being able to 12042 cause a denial of service. ([CVE-2006-2940]) 12043 12044 *Steve Henson, Bodo Moeller* 12045 12046 * Fix ASN.1 parsing of certain invalid structures that can result 12047 in a denial of service. ([CVE-2006-2937]) [Steve Henson] 12048 12049 * Fix buffer overflow in SSL_get_shared_ciphers() function. 12050 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team] 12051 12052 * Fix SSL client code which could crash if connecting to a 12053 malicious SSLv2 server. ([CVE-2006-4343]) 12054 12055 *Tavis Ormandy and Will Drewry, Google Security Team* 12056 12057 * Change ciphersuite string processing so that an explicit 12058 ciphersuite selects this one ciphersuite (so that "AES256-SHA" 12059 will no longer include "AES128-SHA"), and any other similar 12060 ciphersuite (same bitmap) from *other* protocol versions (so that 12061 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 12062 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 12063 changes from 0.9.8b and 0.9.8d. 12064 12065 *Bodo Moeller* 12066 12067### Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 12068 12069 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 12070 ([CVE-2006-4339]) [Ben Laurie and Google Security Team] 12071 12072 * Change the Unix randomness entropy gathering to use poll() when 12073 possible instead of select(), since the latter has some 12074 undesirable limitations. 12075 12076 *Darryl Miles via Richard Levitte and Bodo Moeller* 12077 12078 * Disable rogue ciphersuites: 12079 12080 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 12081 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 12082 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 12083 12084 The latter two were purportedly from 12085 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 12086 appear there. 12087 12088 Also deactivate the remaining ciphersuites from 12089 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 12090 unofficial, and the ID has long expired. 12091 12092 *Bodo Moeller* 12093 12094 * Fix RSA blinding Heisenbug (problems sometimes occurred on 12095 dual-core machines) and other potential thread-safety issues. 12096 12097 *Bodo Moeller* 12098 12099### Changes between 0.9.7i and 0.9.7j [04 May 2006] 12100 12101 * Adapt fipsld and the build system to link against the validated FIPS 12102 module in FIPS mode. 12103 12104 *Steve Henson* 12105 12106 * Fixes for VC++ 2005 build under Windows. 12107 12108 *Steve Henson* 12109 12110 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 12111 from a Windows bash shell such as MSYS. It is autodetected from the 12112 "config" script when run from a VC++ environment. Modify standard VC++ 12113 build to use fipscanister.o from the GNU make build. 12114 12115 *Steve Henson* 12116 12117### Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 12118 12119 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 12120 The value now differs depending on if you build for FIPS or not. 12121 BEWARE! A program linked with a shared FIPSed libcrypto can't be 12122 safely run with a non-FIPSed libcrypto, as it may crash because of 12123 the difference induced by this change. 12124 12125 *Andy Polyakov* 12126 12127### Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 12128 12129 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 12130 (part of SSL_OP_ALL). This option used to disable the 12131 countermeasure against man-in-the-middle protocol-version 12132 rollback in the SSL 2.0 server implementation, which is a bad 12133 idea. ([CVE-2005-2969]) 12134 12135 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 12136 for Information Security, National Institute of Advanced Industrial 12137 Science and Technology [AIST, Japan)]* 12138 12139 * Minimal support for X9.31 signatures and PSS padding modes. This is 12140 mainly for FIPS compliance and not fully integrated at this stage. 12141 12142 *Steve Henson* 12143 12144 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 12145 the exponentiation using a fixed-length exponent. (Otherwise, 12146 the information leaked through timing could expose the secret key 12147 after many signatures; cf. Bleichenbacher's attack on DSA with 12148 biased k.) 12149 12150 *Bodo Moeller* 12151 12152 * Make a new fixed-window mod_exp implementation the default for 12153 RSA, DSA, and DH private-key operations so that the sequence of 12154 squares and multiplies and the memory access pattern are 12155 independent of the particular secret key. This will mitigate 12156 cache-timing and potential related attacks. 12157 12158 BN_mod_exp_mont_consttime() is the new exponentiation implementation, 12159 and this is automatically used by BN_mod_exp_mont() if the new flag 12160 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 12161 will use this BN flag for private exponents unless the flag 12162 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 12163 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 12164 12165 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller* 12166 12167 * Change the client implementation for SSLv23_method() and 12168 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 12169 Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 12170 (Previously, the SSL 2.0 backwards compatible Client Hello 12171 message format would be used even with SSL_OP_NO_SSLv2.) 12172 12173 *Bodo Moeller* 12174 12175 * Add support for smime-type MIME parameter in S/MIME messages which some 12176 clients need. 12177 12178 *Steve Henson* 12179 12180 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in 12181 a threadsafe manner. Modify rsa code to use new function and add calls 12182 to dsa and dh code (which had race conditions before). 12183 12184 *Steve Henson* 12185 12186 * Include the fixed error library code in the C error file definitions 12187 instead of fixing them up at runtime. This keeps the error code 12188 structures constant. 12189 12190 *Steve Henson* 12191 12192### Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 12193 12194[NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 12195OpenSSL 0.9.8.] 12196 12197 * Fixes for newer kerberos headers. NB: the casts are needed because 12198 the 'length' field is signed on one version and unsigned on another 12199 with no (?) obvious way to tell the difference, without these VC++ 12200 complains. Also the "definition" of FAR (blank) is no longer included 12201 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 12202 some needed definitions. 12203 12204 *Steve Henson* 12205 12206 * Undo Cygwin change. 12207 12208 *Ulf Möller* 12209 12210 * Added support for proxy certificates according to RFC 3820. 12211 Because they may be a security thread to unaware applications, 12212 they must be explicitly allowed in run-time. See 12213 docs/HOWTO/proxy_certificates.txt for further information. 12214 12215 *Richard Levitte* 12216 12217### Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 12218 12219 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 12220 server and client random values. Previously 12221 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 12222 less random data when sizeof(time_t) > 4 (some 64 bit platforms). 12223 12224 This change has negligible security impact because: 12225 12226 1. Server and client random values still have 24 bytes of pseudo random 12227 data. 12228 12229 2. Server and client random values are sent in the clear in the initial 12230 handshake. 12231 12232 3. The master secret is derived using the premaster secret (48 bytes in 12233 size for static RSA ciphersuites) as well as client server and random 12234 values. 12235 12236 The OpenSSL team would like to thank the UK NISCC for bringing this issue 12237 to our attention. 12238 12239 *Stephen Henson, reported by UK NISCC* 12240 12241 * Use Windows randomness collection on Cygwin. 12242 12243 *Ulf Möller* 12244 12245 * Fix hang in EGD/PRNGD query when communication socket is closed 12246 prematurely by EGD/PRNGD. 12247 12248 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014* 12249 12250 * Prompt for pass phrases when appropriate for PKCS12 input format. 12251 12252 *Steve Henson* 12253 12254 * Back-port of selected performance improvements from development 12255 branch, as well as improved support for PowerPC platforms. 12256 12257 *Andy Polyakov* 12258 12259 * Add lots of checks for memory allocation failure, error codes to indicate 12260 failure and freeing up memory if a failure occurs. 12261 12262 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson* 12263 12264 * Add new -passin argument to dgst. 12265 12266 *Steve Henson* 12267 12268 * Perform some character comparisons of different types in X509_NAME_cmp: 12269 this is needed for some certificates that re-encode DNs into UTF8Strings 12270 (in violation of RFC3280) and can't or won't issue name rollover 12271 certificates. 12272 12273 *Steve Henson* 12274 12275 * Make an explicit check during certificate validation to see that 12276 the CA setting in each certificate on the chain is correct. As a 12277 side effect always do the following basic checks on extensions, 12278 not just when there's an associated purpose to the check: 12279 12280 - if there is an unhandled critical extension (unless the user 12281 has chosen to ignore this fault) 12282 - if the path length has been exceeded (if one is set at all) 12283 - that certain extensions fit the associated purpose (if one has 12284 been given) 12285 12286 *Richard Levitte* 12287 12288### Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 12289 12290 * Avoid a race condition when CRLs are checked in a multi threaded 12291 environment. This would happen due to the reordering of the revoked 12292 entries during signature checking and serial number lookup. Now the 12293 encoding is cached and the serial number sort performed under a lock. 12294 Add new STACK function sk_is_sorted(). 12295 12296 *Steve Henson* 12297 12298 * Add Delta CRL to the extension code. 12299 12300 *Steve Henson* 12301 12302 * Various fixes to s3_pkt.c so alerts are sent properly. 12303 12304 *David Holmes <d.holmes@f5.com>* 12305 12306 * Reduce the chances of duplicate issuer name and serial numbers (in 12307 violation of RFC3280) using the OpenSSL certificate creation utilities. 12308 This is done by creating a random 64 bit value for the initial serial 12309 number when a serial number file is created or when a self signed 12310 certificate is created using 'openssl req -x509'. The initial serial 12311 number file is created using 'openssl x509 -next_serial' in CA.pl 12312 rather than being initialized to 1. 12313 12314 *Steve Henson* 12315 12316### Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 12317 12318 * Fix null-pointer assignment in do_change_cipher_spec() revealed 12319 by using the Codenomicon TLS Test Tool ([CVE-2004-0079]) 12320 12321 *Joe Orton, Steve Henson* 12322 12323 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 12324 ([CVE-2004-0112]) 12325 12326 *Joe Orton, Steve Henson* 12327 12328 * Make it possible to have multiple active certificates with the same 12329 subject in the CA index file. This is done only if the keyword 12330 'unique_subject' is set to 'no' in the main CA section (default 12331 if 'CA_default') of the configuration file. The value is saved 12332 with the database itself in a separate index attribute file, 12333 named like the index file with '.attr' appended to the name. 12334 12335 *Richard Levitte* 12336 12337 * X509 verify fixes. Disable broken certificate workarounds when 12338 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 12339 keyUsage extension present. Don't accept CRLs with unhandled critical 12340 extensions: since verify currently doesn't process CRL extensions this 12341 rejects a CRL with *any* critical extensions. Add new verify error codes 12342 for these cases. 12343 12344 *Steve Henson* 12345 12346 * When creating an OCSP nonce use an OCTET STRING inside the extnValue. 12347 A clarification of RFC2560 will require the use of OCTET STRINGs and 12348 some implementations cannot handle the current raw format. Since OpenSSL 12349 copies and compares OCSP nonces as opaque blobs without any attempt at 12350 parsing them this should not create any compatibility issues. 12351 12352 *Steve Henson* 12353 12354 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 12355 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 12356 this HMAC (and other) operations are several times slower than OpenSSL 12357 < 0.9.7. 12358 12359 *Steve Henson* 12360 12361 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 12362 12363 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>* 12364 12365 * Use the correct content when signing type "other". 12366 12367 *Steve Henson* 12368 12369### Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 12370 12371 * Fix various bugs revealed by running the NISCC test suite: 12372 12373 Stop out of bounds reads in the ASN1 code when presented with 12374 invalid tags (CVE-2003-0543 and CVE-2003-0544). 12375 12376 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]). 12377 12378 If verify callback ignores invalid public key errors don't try to check 12379 certificate signature with the NULL public key. 12380 12381 *Steve Henson* 12382 12383 * New -ignore_err option in ocsp application to stop the server 12384 exiting on the first error in a request. 12385 12386 *Steve Henson* 12387 12388 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 12389 if the server requested one: as stated in TLS 1.0 and SSL 3.0 12390 specifications. 12391 12392 *Steve Henson* 12393 12394 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 12395 extra data after the compression methods not only for TLS 1.0 12396 but also for SSL 3.0 (as required by the specification). 12397 12398 *Bodo Moeller; problem pointed out by Matthias Loepfe* 12399 12400 * Change X509_certificate_type() to mark the key as exported/exportable 12401 when it's 512 *bits* long, not 512 bytes. 12402 12403 *Richard Levitte* 12404 12405 * Change AES_cbc_encrypt() so it outputs exact multiple of 12406 blocks during encryption. 12407 12408 *Richard Levitte* 12409 12410 * Various fixes to base64 BIO and non blocking I/O. On write 12411 flushes were not handled properly if the BIO retried. On read 12412 data was not being buffered properly and had various logic bugs. 12413 This also affects blocking I/O when the data being decoded is a 12414 certain size. 12415 12416 *Steve Henson* 12417 12418 * Various S/MIME bugfixes and compatibility changes: 12419 output correct application/pkcs7 MIME type if 12420 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 12421 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 12422 of files as .eml work). Correctly handle very long lines in MIME 12423 parser. 12424 12425 *Steve Henson* 12426 12427### Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 12428 12429 * Countermeasure against the Klima-Pokorny-Rosa extension of 12430 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 12431 a protocol version number mismatch like a decryption error 12432 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 12433 12434 *Bodo Moeller* 12435 12436 * Turn on RSA blinding by default in the default implementation 12437 to avoid a timing attack. Applications that don't want it can call 12438 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 12439 They would be ill-advised to do so in most cases. 12440 12441 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller* 12442 12443 * Change RSA blinding code so that it works when the PRNG is not 12444 seeded (in this case, the secret RSA exponent is abused as 12445 an unpredictable seed -- if it is not unpredictable, there 12446 is no point in blinding anyway). Make RSA blinding thread-safe 12447 by remembering the creator's thread ID in rsa->blinding and 12448 having all other threads use local one-time blinding factors 12449 (this requires more computation than sharing rsa->blinding, but 12450 avoids excessive locking; and if an RSA object is not shared 12451 between threads, blinding will still be very fast). 12452 12453 *Bodo Moeller* 12454 12455 * Fixed a typo bug that would cause ENGINE_set_default() to set an 12456 ENGINE as defaults for all supported algorithms irrespective of 12457 the 'flags' parameter. 'flags' is now honoured, so applications 12458 should make sure they are passing it correctly. 12459 12460 *Geoff Thorpe* 12461 12462 * Target "mingw" now allows native Windows code to be generated in 12463 the Cygwin environment as well as with the MinGW compiler. 12464 12465 *Ulf Moeller* 12466 12467### Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 12468 12469 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 12470 via timing by performing a MAC computation even if incorrect 12471 block cipher padding has been found. This is a countermeasure 12472 against active attacks where the attacker has to distinguish 12473 between bad padding and a MAC verification error. ([CVE-2003-0078]) 12474 12475 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 12476 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 12477 Martin Vuagnoux (EPFL, Ilion)* 12478 12479 * Make the no-err option work as intended. The intention with no-err 12480 is not to have the whole error stack handling routines removed from 12481 libcrypto, it's only intended to remove all the function name and 12482 reason texts, thereby removing some of the footprint that may not 12483 be interesting if those errors aren't displayed anyway. 12484 12485 NOTE: it's still possible for any application or module to have its 12486 own set of error texts inserted. The routines are there, just not 12487 used by default when no-err is given. 12488 12489 *Richard Levitte* 12490 12491 * Add support for FreeBSD on IA64. 12492 12493 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454* 12494 12495 * Adjust DES_cbc_cksum() so it returns the same value as the MIT 12496 Kerberos function mit_des_cbc_cksum(). Before this change, 12497 the value returned by DES_cbc_cksum() was like the one from 12498 mit_des_cbc_cksum(), except the bytes were swapped. 12499 12500 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte* 12501 12502 * Allow an application to disable the automatic SSL chain building. 12503 Before this a rather primitive chain build was always performed in 12504 ssl3_output_cert_chain(): an application had no way to send the 12505 correct chain if the automatic operation produced an incorrect result. 12506 12507 Now the chain builder is disabled if either: 12508 12509 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 12510 12511 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 12512 12513 The reasoning behind this is that an application would not want the 12514 auto chain building to take place if extra chain certificates are 12515 present and it might also want a means of sending no additional 12516 certificates (for example the chain has two certificates and the 12517 root is omitted). 12518 12519 *Steve Henson* 12520 12521 * Add the possibility to build without the ENGINE framework. 12522 12523 *Steven Reddie <smr@essemer.com.au> via Richard Levitte* 12524 12525 * Under Win32 gmtime() can return NULL: check return value in 12526 OPENSSL_gmtime(). Add error code for case where gmtime() fails. 12527 12528 *Steve Henson* 12529 12530 * DSA routines: under certain error conditions uninitialized BN objects 12531 could be freed. Solution: make sure initialization is performed early 12532 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 12533 Nils Larsch <nla@trustcenter.de> via PR#459) 12534 12535 *Lutz Jaenicke* 12536 12537 * Another fix for SSLv2 session ID handling: the session ID was incorrectly 12538 checked on reconnect on the client side, therefore session resumption 12539 could still fail with a "ssl session id is different" error. This 12540 behaviour is masked when SSL_OP_ALL is used due to 12541 SSL_OP_MICROSOFT_SESS_ID_BUG being set. 12542 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 12543 followup to PR #377. 12544 12545 *Lutz Jaenicke* 12546 12547 * IA-32 assembler support enhancements: unified ELF targets, support 12548 for SCO/Caldera platforms, fix for Cygwin shared build. 12549 12550 *Andy Polyakov* 12551 12552 * Add support for FreeBSD on sparc64. As a consequence, support for 12553 FreeBSD on non-x86 processors is separate from x86 processors on 12554 the config script, much like the NetBSD support. 12555 12556 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>* 12557 12558### Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 12559 12560[NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 12561OpenSSL 0.9.7.] 12562 12563 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED 12564 code (06) was taken as the first octet of the session ID and the last 12565 octet was ignored consequently. As a result SSLv2 client side session 12566 caching could not have worked due to the session ID mismatch between 12567 client and server. 12568 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 12569 PR #377. 12570 12571 *Lutz Jaenicke* 12572 12573 * Change the declaration of needed Kerberos libraries to use EX_LIBS 12574 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 12575 removed entirely. 12576 12577 *Richard Levitte* 12578 12579 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 12580 seems that in spite of existing for more than a year, many application 12581 author have done nothing to provide the necessary callbacks, which 12582 means that this particular engine will not work properly anywhere. 12583 This is a very unfortunate situation which forces us, in the name 12584 of usability, to give the hw_ncipher.c a static lock, which is part 12585 of libcrypto. 12586 NOTE: This is for the 0.9.7 series ONLY. This hack will never 12587 appear in 0.9.8 or later. We EXPECT application authors to have 12588 dealt properly with this when 0.9.8 is released (unless we actually 12589 make such changes in the libcrypto locking code that changes will 12590 have to be made anyway). 12591 12592 *Richard Levitte* 12593 12594 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 12595 octets have been read, EOF or an error occurs. Without this change 12596 some truncated ASN1 structures will not produce an error. 12597 12598 *Steve Henson* 12599 12600 * Disable Heimdal support, since it hasn't been fully implemented. 12601 Still give the possibility to force the use of Heimdal, but with 12602 warnings and a request that patches get sent to openssl-dev. 12603 12604 *Richard Levitte* 12605 12606 * Add the VC-CE target, introduce the WINCE sysname, and add 12607 INSTALL.WCE and appropriate conditionals to make it build. 12608 12609 *Steven Reddie <smr@essemer.com.au> via Richard Levitte* 12610 12611 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 12612 cygssl-x.y.z.dll, where x, y and z are the major, minor and 12613 edit numbers of the version. 12614 12615 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte* 12616 12617 * Introduce safe string copy and catenation functions 12618 (BUF_strlcpy() and BUF_strlcat()). 12619 12620 *Ben Laurie (CHATS) and Richard Levitte* 12621 12622 * Avoid using fixed-size buffers for one-line DNs. 12623 12624 *Ben Laurie (CHATS)* 12625 12626 * Add BUF_MEM_grow_clean() to avoid information leakage when 12627 resizing buffers containing secrets, and use where appropriate. 12628 12629 *Ben Laurie (CHATS)* 12630 12631 * Avoid using fixed size buffers for configuration file location. 12632 12633 *Ben Laurie (CHATS)* 12634 12635 * Avoid filename truncation for various CA files. 12636 12637 *Ben Laurie (CHATS)* 12638 12639 * Use sizeof in preference to magic numbers. 12640 12641 *Ben Laurie (CHATS)* 12642 12643 * Avoid filename truncation in cert requests. 12644 12645 *Ben Laurie (CHATS)* 12646 12647 * Add assertions to check for (supposedly impossible) buffer 12648 overflows. 12649 12650 *Ben Laurie (CHATS)* 12651 12652 * Don't cache truncated DNS entries in the local cache (this could 12653 potentially lead to a spoofing attack). 12654 12655 *Ben Laurie (CHATS)* 12656 12657 * Fix various buffers to be large enough for hex/decimal 12658 representations in a platform independent manner. 12659 12660 *Ben Laurie (CHATS)* 12661 12662 * Add CRYPTO_realloc_clean() to avoid information leakage when 12663 resizing buffers containing secrets, and use where appropriate. 12664 12665 *Ben Laurie (CHATS)* 12666 12667 * Add BIO_indent() to avoid much slightly worrying code to do 12668 indents. 12669 12670 *Ben Laurie (CHATS)* 12671 12672 * Convert sprintf()/BIO_puts() to BIO_printf(). 12673 12674 *Ben Laurie (CHATS)* 12675 12676 * buffer_gets() could terminate with the buffer only half 12677 full. Fixed. 12678 12679 *Ben Laurie (CHATS)* 12680 12681 * Add assertions to prevent user-supplied crypto functions from 12682 overflowing internal buffers by having large block sizes, etc. 12683 12684 *Ben Laurie (CHATS)* 12685 12686 * New OPENSSL_assert() macro (similar to assert(), but enabled 12687 unconditionally). 12688 12689 *Ben Laurie (CHATS)* 12690 12691 * Eliminate unused copy of key in RC4. 12692 12693 *Ben Laurie (CHATS)* 12694 12695 * Eliminate unused and incorrectly sized buffers for IV in pem.h. 12696 12697 *Ben Laurie (CHATS)* 12698 12699 * Fix off-by-one error in EGD path. 12700 12701 *Ben Laurie (CHATS)* 12702 12703 * If RANDFILE path is too long, ignore instead of truncating. 12704 12705 *Ben Laurie (CHATS)* 12706 12707 * Eliminate unused and incorrectly sized X.509 structure 12708 CBCParameter. 12709 12710 *Ben Laurie (CHATS)* 12711 12712 * Eliminate unused and dangerous function knumber(). 12713 12714 *Ben Laurie (CHATS)* 12715 12716 * Eliminate unused and dangerous structure, KSSL_ERR. 12717 12718 *Ben Laurie (CHATS)* 12719 12720 * Protect against overlong session ID context length in an encoded 12721 session object. Since these are local, this does not appear to be 12722 exploitable. 12723 12724 *Ben Laurie (CHATS)* 12725 12726 * Change from security patch (see 0.9.6e below) that did not affect 12727 the 0.9.6 release series: 12728 12729 Remote buffer overflow in SSL3 protocol - an attacker could 12730 supply an oversized master key in Kerberos-enabled versions. 12731 ([CVE-2002-0657]) 12732 12733 *Ben Laurie (CHATS)* 12734 12735 * Change the SSL kerb5 codes to match RFC 2712. 12736 12737 *Richard Levitte* 12738 12739 * Make -nameopt work fully for req and add -reqopt switch. 12740 12741 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson* 12742 12743 * The "block size" for block ciphers in CFB and OFB mode should be 1. 12744 12745 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>* 12746 12747 * Make sure tests can be performed even if the corresponding algorithms 12748 have been removed entirely. This was also the last step to make 12749 OpenSSL compilable with DJGPP under all reasonable conditions. 12750 12751 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>* 12752 12753 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 12754 to allow version independent disabling of normally unselected ciphers, 12755 which may be activated as a side-effect of selecting a single cipher. 12756 12757 (E.g., cipher list string "RSA" enables ciphersuites that are left 12758 out of "ALL" because they do not provide symmetric encryption. 12759 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 12760 12761 *Lutz Jaenicke, Bodo Moeller* 12762 12763 * Add appropriate support for separate platform-dependent build 12764 directories. The recommended way to make a platform-dependent 12765 build directory is the following (tested on Linux), maybe with 12766 some local tweaks: 12767 12768 # Place yourself outside of the OpenSSL source tree. In 12769 # this example, the environment variable OPENSSL_SOURCE 12770 # is assumed to contain the absolute OpenSSL source directory. 12771 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 12772 cd objtree/"`uname -s`-`uname -r`-`uname -m`" 12773 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 12774 mkdir -p `dirname $F` 12775 ln -s $OPENSSL_SOURCE/$F $F 12776 done 12777 12778 To be absolutely sure not to disturb the source tree, a "make clean" 12779 is a good thing. If it isn't successful, don't worry about it, 12780 it probably means the source directory is very clean. 12781 12782 *Richard Levitte* 12783 12784 * Make sure any ENGINE control commands make local copies of string 12785 pointers passed to them whenever necessary. Otherwise it is possible 12786 the caller may have overwritten (or deallocated) the original string 12787 data when a later ENGINE operation tries to use the stored values. 12788 12789 *Götz Babin-Ebell <babinebell@trustcenter.de>* 12790 12791 * Improve diagnostics in file reading and command-line digests. 12792 12793 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>* 12794 12795 * Add AES modes CFB and OFB to the object database. Correct an 12796 error in AES-CFB decryption. 12797 12798 *Richard Levitte* 12799 12800 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 12801 allows existing EVP_CIPHER_CTX structures to be reused after 12802 calling `EVP_*Final()`. This behaviour is used by encryption 12803 BIOs and some applications. This has the side effect that 12804 applications must explicitly clean up cipher contexts with 12805 EVP_CIPHER_CTX_cleanup() or they will leak memory. 12806 12807 *Steve Henson* 12808 12809 * Check the values of dna and dnb in bn_mul_recursive before calling 12810 bn_mul_comba (a non zero value means the a or b arrays do not contain 12811 n2 elements) and fallback to bn_mul_normal if either is not zero. 12812 12813 *Steve Henson* 12814 12815 * Fix escaping of non-ASCII characters when using the -subj option 12816 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 12817 12818 *Lutz Jaenicke* 12819 12820 * Make object definitions compliant to LDAP (RFC2256): SN is the short 12821 form for "surname", serialNumber has no short form. 12822 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 12823 therefore remove "mail" short name for "internet 7". 12824 The OID for unique identifiers in X509 certificates is 12825 x500UniqueIdentifier, not uniqueIdentifier. 12826 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 12827 12828 *Lutz Jaenicke* 12829 12830 * Add an "init" command to the ENGINE config module and auto initialize 12831 ENGINEs. Without any "init" command the ENGINE will be initialized 12832 after all ctrl commands have been executed on it. If init=1 the 12833 ENGINE is initialized at that point (ctrls before that point are run 12834 on the uninitialized ENGINE and after on the initialized one). If 12835 init=0 then the ENGINE will not be initialized at all. 12836 12837 *Steve Henson* 12838 12839 * Fix the 'app_verify_callback' interface so that the user-defined 12840 argument is actually passed to the callback: In the 12841 SSL_CTX_set_cert_verify_callback() prototype, the callback 12842 declaration has been changed from 12843 int (*cb)() 12844 into 12845 int (*cb)(X509_STORE_CTX *,void *); 12846 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 12847 i=s->ctx->app_verify_callback(&ctx) 12848 has been changed into 12849 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 12850 12851 To update applications using SSL_CTX_set_cert_verify_callback(), 12852 a dummy argument can be added to their callback functions. 12853 12854 *D. K. Smetters <smetters@parc.xerox.com>* 12855 12856 * Added the '4758cca' ENGINE to support IBM 4758 cards. 12857 12858 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe* 12859 12860 * Add and OPENSSL_LOAD_CONF define which will cause 12861 OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 12862 This allows older applications to transparently support certain 12863 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 12864 Two new functions OPENSSL_add_all_algorithms_noconf() which will never 12865 load the config file and OPENSSL_add_all_algorithms_conf() which will 12866 always load it have also been added. 12867 12868 *Steve Henson* 12869 12870 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 12871 Adjust NIDs and EVP layer. 12872 12873 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte* 12874 12875 * Config modules support in openssl utility. 12876 12877 Most commands now load modules from the config file, 12878 though in a few (such as version) this isn't done 12879 because it couldn't be used for anything. 12880 12881 In the case of ca and req the config file used is 12882 the same as the utility itself: that is the -config 12883 command line option can be used to specify an 12884 alternative file. 12885 12886 *Steve Henson* 12887 12888 * Move default behaviour from OPENSSL_config(). If appname is NULL 12889 use "openssl_conf" if filename is NULL use default openssl config file. 12890 12891 *Steve Henson* 12892 12893 * Add an argument to OPENSSL_config() to allow the use of an alternative 12894 config section name. Add a new flag to tolerate a missing config file 12895 and move code to CONF_modules_load_file(). 12896 12897 *Steve Henson* 12898 12899 * Support for crypto accelerator cards from Accelerated Encryption 12900 Processing, www.aep.ie. (Use engine 'aep') 12901 The support was copied from 0.9.6c [engine] and adapted/corrected 12902 to work with the new engine framework. 12903 12904 *AEP Inc. and Richard Levitte* 12905 12906 * Support for SureWare crypto accelerator cards from Baltimore 12907 Technologies. (Use engine 'sureware') 12908 The support was copied from 0.9.6c [engine] and adapted 12909 to work with the new engine framework. 12910 12911 *Richard Levitte* 12912 12913 * Have the CHIL engine fork-safe (as defined by nCipher) and actually 12914 make the newer ENGINE framework commands for the CHIL engine work. 12915 12916 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte* 12917 12918 * Make it possible to produce shared libraries on ReliantUNIX. 12919 12920 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte* 12921 12922 * Add the configuration target debug-linux-ppro. 12923 Make 'openssl rsa' use the general key loading routines 12924 implemented in `apps.c`, and make those routines able to 12925 handle the key format FORMAT_NETSCAPE and the variant 12926 FORMAT_IISSGC. 12927 12928 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte* 12929 12930 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 12931 12932 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte* 12933 12934 * Add -keyform to rsautl, and document -engine. 12935 12936 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>* 12937 12938 * Change BIO_new_file (crypto/bio/bss_file.c) to use new 12939 BIO_R_NO_SUCH_FILE error code rather than the generic 12940 ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 12941 12942 *Ben Laurie* 12943 12944 * Add new functions 12945 ERR_peek_last_error 12946 ERR_peek_last_error_line 12947 ERR_peek_last_error_line_data. 12948 These are similar to 12949 ERR_peek_error 12950 ERR_peek_error_line 12951 ERR_peek_error_line_data, 12952 but report on the latest error recorded rather than the first one 12953 still in the error queue. 12954 12955 *Ben Laurie, Bodo Moeller* 12956 12957 * default_algorithms option in ENGINE config module. This allows things 12958 like: 12959 default_algorithms = ALL 12960 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 12961 12962 *Steve Henson* 12963 12964 * Preliminary ENGINE config module. 12965 12966 *Steve Henson* 12967 12968 * New experimental application configuration code. 12969 12970 *Steve Henson* 12971 12972 * Change the AES code to follow the same name structure as all other 12973 symmetric ciphers, and behave the same way. Move everything to 12974 the directory crypto/aes, thereby obsoleting crypto/rijndael. 12975 12976 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte* 12977 12978 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 12979 12980 *Ben Laurie and Theo de Raadt* 12981 12982 * Add option to output public keys in req command. 12983 12984 *Massimiliano Pala madwolf@openca.org* 12985 12986 * Use wNAFs in EC_POINTs_mul() for improved efficiency 12987 (up to about 10% better than before for P-192 and P-224). 12988 12989 *Bodo Moeller* 12990 12991 * New functions/macros 12992 12993 SSL_CTX_set_msg_callback(ctx, cb) 12994 SSL_CTX_set_msg_callback_arg(ctx, arg) 12995 SSL_set_msg_callback(ssl, cb) 12996 SSL_set_msg_callback_arg(ssl, arg) 12997 12998 to request calling a callback function 12999 13000 void cb(int write_p, int version, int content_type, 13001 const void *buf, size_t len, SSL *ssl, void *arg) 13002 13003 whenever a protocol message has been completely received 13004 (write_p == 0) or sent (write_p == 1). Here 'version' is the 13005 protocol version according to which the SSL library interprets 13006 the current protocol message (SSL2_VERSION, SSL3_VERSION, or 13007 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 13008 the content type as defined in the SSL 3.0/TLS 1.0 protocol 13009 specification (change_cipher_spec(20), alert(21), handshake(22)). 13010 'buf' and 'len' point to the actual message, 'ssl' to the 13011 SSL object, and 'arg' is the application-defined value set by 13012 SSL[_CTX]_set_msg_callback_arg(). 13013 13014 'openssl s_client' and 'openssl s_server' have new '-msg' options 13015 to enable a callback that displays all protocol messages. 13016 13017 *Bodo Moeller* 13018 13019 * Change the shared library support so shared libraries are built as 13020 soon as the corresponding static library is finished, and thereby get 13021 openssl and the test programs linked against the shared library. 13022 This still only happens when the keyword "shard" has been given to 13023 the configuration scripts. 13024 13025 NOTE: shared library support is still an experimental thing, and 13026 backward binary compatibility is still not guaranteed. 13027 13028 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte* 13029 13030 * Add support for Subject Information Access extension. 13031 13032 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>* 13033 13034 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 13035 additional bytes when new memory had to be allocated, not just 13036 when reusing an existing buffer. 13037 13038 *Bodo Moeller* 13039 13040 * New command line and configuration option 'utf8' for the req command. 13041 This allows field values to be specified as UTF8 strings. 13042 13043 *Steve Henson* 13044 13045 * Add -multi and -mr options to "openssl speed" - giving multiple parallel 13046 runs for the former and machine-readable output for the latter. 13047 13048 *Ben Laurie* 13049 13050 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 13051 of the e-mail address in the DN (i.e., it will go into a certificate 13052 extension only). The new configuration file option 'email_in_dn = no' 13053 has the same effect. 13054 13055 *Massimiliano Pala madwolf@openca.org* 13056 13057 * Change all functions with names starting with `des_` to be starting 13058 with `DES_` instead. Add wrappers that are compatible with libdes, 13059 but are named `_ossl_old_des_*`. Finally, add macros that map the 13060 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes 13061 compatibility is desired. If OpenSSL 0.9.6c compatibility is 13062 desired, the `des_*` symbols will be mapped to `DES_*`, with one 13063 exception. 13064 13065 Since we provide two compatibility mappings, the user needs to 13066 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 13067 compatibility is desired. The default (i.e., when that macro 13068 isn't defined) is OpenSSL 0.9.6c compatibility. 13069 13070 There are also macros that enable and disable the support of old 13071 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 13072 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 13073 are defined, the default will apply: to support the old des routines. 13074 13075 In either case, one must include openssl/des.h to get the correct 13076 definitions. Do not try to just include openssl/des_old.h, that 13077 won't work. 13078 13079 NOTE: This is a major break of an old API into a new one. Software 13080 authors are encouraged to switch to the `DES_` style functions. Some 13081 time in the future, des_old.h and the libdes compatibility functions 13082 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 13083 default), and then completely removed. 13084 13085 *Richard Levitte* 13086 13087 * Test for certificates which contain unsupported critical extensions. 13088 If such a certificate is found during a verify operation it is 13089 rejected by default: this behaviour can be overridden by either 13090 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 13091 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 13092 X509_supported_extension() has also been added which returns 1 if a 13093 particular extension is supported. 13094 13095 *Steve Henson* 13096 13097 * Modify the behaviour of EVP cipher functions in similar way to digests 13098 to retain compatibility with existing code. 13099 13100 *Steve Henson* 13101 13102 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 13103 compatibility with existing code. In particular the 'ctx' parameter does 13104 not have to be to be initialized before the call to EVP_DigestInit() and 13105 it is tidied up after a call to EVP_DigestFinal(). New function 13106 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 13107 EVP_MD_CTX_copy() changed to not require the destination to be 13108 initialized valid and new function EVP_MD_CTX_copy_ex() added which 13109 requires the destination to be valid. 13110 13111 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 13112 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 13113 13114 *Steve Henson* 13115 13116 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it 13117 so that complete 'Handshake' protocol structures are kept in memory 13118 instead of overwriting 'msg_type' and 'length' with 'body' data. 13119 13120 *Bodo Moeller* 13121 13122 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 13123 13124 *Massimo Santin via Richard Levitte* 13125 13126 * Major restructuring to the underlying ENGINE code. This includes 13127 reduction of linker bloat, separation of pure "ENGINE" manipulation 13128 (initialisation, etc) from functionality dealing with implementations 13129 of specific crypto interfaces. This change also introduces integrated 13130 support for symmetric ciphers and digest implementations - so ENGINEs 13131 can now accelerate these by providing EVP_CIPHER and EVP_MD 13132 implementations of their own. This is detailed in 13133 [crypto/engine/README.md](crypto/engine/README.md) 13134 as it couldn't be adequately described here. However, there are a few 13135 API changes worth noting - some RSA, DSA, DH, and RAND functions that 13136 were changed in the original introduction of ENGINE code have now 13137 reverted back - the hooking from this code to ENGINE is now a good 13138 deal more passive and at run-time, operations deal directly with 13139 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 13140 dereferencing through an ENGINE pointer any more. Also, the ENGINE 13141 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed - 13142 they were not being used by the framework as there is no concept of a 13143 BIGNUM_METHOD and they could not be generalised to the new 13144 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 13145 ENGINE_cpy() has been removed as it cannot be consistently defined in 13146 the new code. 13147 13148 *Geoff Thorpe* 13149 13150 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 13151 13152 *Steve Henson* 13153 13154 * Change mkdef.pl to sort symbols that get the same entry number, 13155 and make sure the automatically generated functions `ERR_load_*` 13156 become part of libeay.num as well. 13157 13158 *Richard Levitte* 13159 13160 * New function SSL_renegotiate_pending(). This returns true once 13161 renegotiation has been requested (either SSL_renegotiate() call 13162 or HelloRequest/ClientHello received from the peer) and becomes 13163 false once a handshake has been completed. 13164 (For servers, SSL_renegotiate() followed by SSL_do_handshake() 13165 sends a HelloRequest, but does not ensure that a handshake takes 13166 place. SSL_renegotiate_pending() is useful for checking if the 13167 client has followed the request.) 13168 13169 *Bodo Moeller* 13170 13171 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 13172 By default, clients may request session resumption even during 13173 renegotiation (if session ID contexts permit); with this option, 13174 session resumption is possible only in the first handshake. 13175 13176 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 13177 more bits available for options that should not be part of 13178 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 13179 13180 *Bodo Moeller* 13181 13182 * Add some demos for certificate and certificate request creation. 13183 13184 *Steve Henson* 13185 13186 * Make maximum certificate chain size accepted from the peer application 13187 settable (`SSL*_get/set_max_cert_list()`), as proposed by 13188 "Douglas E. Engert" <deengert@anl.gov>. 13189 13190 *Lutz Jaenicke* 13191 13192 * Add support for shared libraries for Unixware-7 13193 (Boyd Lynn Gerber <gerberb@zenez.com>). 13194 13195 *Lutz Jaenicke* 13196 13197 * Add a "destroy" handler to ENGINEs that allows structural cleanup to 13198 be done prior to destruction. Use this to unload error strings from 13199 ENGINEs that load their own error strings. NB: This adds two new API 13200 functions to "get" and "set" this destroy handler in an ENGINE. 13201 13202 *Geoff Thorpe* 13203 13204 * Alter all existing ENGINE implementations (except "openssl" and 13205 "openbsd") to dynamically instantiate their own error strings. This 13206 makes them more flexible to be built both as statically-linked ENGINEs 13207 and self-contained shared-libraries loadable via the "dynamic" ENGINE. 13208 Also, add stub code to each that makes building them as self-contained 13209 shared-libraries easier (see [README-Engine.md](README-Engine.md)). 13210 13211 *Geoff Thorpe* 13212 13213 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 13214 implementations into applications that are completely implemented in 13215 self-contained shared-libraries. The "dynamic" ENGINE exposes control 13216 commands that can be used to configure what shared-library to load and 13217 to control aspects of the way it is handled. Also, made an update to 13218 the [README-Engine.md](README-Engine.md) file 13219 that brings its information up-to-date and 13220 provides some information and instructions on the "dynamic" ENGINE 13221 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 13222 13223 *Geoff Thorpe* 13224 13225 * Make it possible to unload ranges of ERR strings with a new 13226 "ERR_unload_strings" function. 13227 13228 *Geoff Thorpe* 13229 13230 * Add a copy() function to EVP_MD. 13231 13232 *Ben Laurie* 13233 13234 * Make EVP_MD routines take a context pointer instead of just the 13235 md_data void pointer. 13236 13237 *Ben Laurie* 13238 13239 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 13240 that the digest can only process a single chunk of data 13241 (typically because it is provided by a piece of 13242 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 13243 is only going to provide a single chunk of data, and hence the 13244 framework needn't accumulate the data for oneshot drivers. 13245 13246 *Ben Laurie* 13247 13248 * As with "ERR", make it possible to replace the underlying "ex_data" 13249 functions. This change also alters the storage and management of global 13250 ex_data state - it's now all inside ex_data.c and all "class" code (eg. 13251 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 13252 index counters. The API functions that use this state have been changed 13253 to take a "class_index" rather than pointers to the class's local STACK 13254 and counter, and there is now an API function to dynamically create new 13255 classes. This centralisation allows us to (a) plug a lot of the 13256 thread-safety problems that existed, and (b) makes it possible to clean 13257 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 13258 such data would previously have always leaked in application code and 13259 workarounds were in place to make the memory debugging turn a blind eye 13260 to it. Application code that doesn't use this new function will still 13261 leak as before, but their memory debugging output will announce it now 13262 rather than letting it slide. 13263 13264 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 13265 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 13266 has a return value to indicate success or failure. 13267 13268 *Geoff Thorpe* 13269 13270 * Make it possible to replace the underlying "ERR" functions such that the 13271 global state (2 LHASH tables and 2 locks) is only used by the "default" 13272 implementation. This change also adds two functions to "get" and "set" 13273 the implementation prior to it being automatically set the first time 13274 any other ERR function takes place. Ie. an application can call "get", 13275 pass the return value to a module it has just loaded, and that module 13276 can call its own "set" function using that value. This means the 13277 module's "ERR" operations will use (and modify) the error state in the 13278 application and not in its own statically linked copy of OpenSSL code. 13279 13280 *Geoff Thorpe* 13281 13282 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment 13283 reference counts. This performs normal REF_PRINT/REF_CHECK macros on 13284 the operation, and provides a more encapsulated way for external code 13285 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 13286 to use these functions rather than manually incrementing the counts. 13287 13288 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 13289 13290 *Geoff Thorpe* 13291 13292 * Add EVP test program. 13293 13294 *Ben Laurie* 13295 13296 * Add symmetric cipher support to ENGINE. Expect the API to change! 13297 13298 *Ben Laurie* 13299 13300 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 13301 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 13302 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 13303 These allow a CRL to be built without having to access X509_CRL fields 13304 directly. Modify 'ca' application to use new functions. 13305 13306 *Steve Henson* 13307 13308 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 13309 bug workarounds. Rollback attack detection is a security feature. 13310 The problem will only arise on OpenSSL servers when TLSv1 is not 13311 available (sslv3_server_method() or SSL_OP_NO_TLSv1). 13312 Software authors not wanting to support TLSv1 will have special reasons 13313 for their choice and can explicitly enable this option. 13314 13315 *Bodo Moeller, Lutz Jaenicke* 13316 13317 * Rationalise EVP so it can be extended: don't include a union of 13318 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 13319 (similar to those existing for EVP_CIPHER_CTX). 13320 Usage example: 13321 13322 EVP_MD_CTX md; 13323 13324 EVP_MD_CTX_init(&md); /* new function call */ 13325 EVP_DigestInit(&md, EVP_sha1()); 13326 EVP_DigestUpdate(&md, in, len); 13327 EVP_DigestFinal(&md, out, NULL); 13328 EVP_MD_CTX_cleanup(&md); /* new function call */ 13329 13330 *Ben Laurie* 13331 13332 * Make DES key schedule conform to the usual scheme, as well as 13333 correcting its structure. This means that calls to DES functions 13334 now have to pass a pointer to a des_key_schedule instead of a 13335 plain des_key_schedule (which was actually always a pointer 13336 anyway): E.g., 13337 13338 des_key_schedule ks; 13339 13340 des_set_key_checked(..., &ks); 13341 des_ncbc_encrypt(..., &ks, ...); 13342 13343 (Note that a later change renames 'des_...' into 'DES_...'.) 13344 13345 *Ben Laurie* 13346 13347 * Initial reduction of linker bloat: the use of some functions, such as 13348 PEM causes large amounts of unused functions to be linked in due to 13349 poor organisation. For example pem_all.c contains every PEM function 13350 which has a knock on effect of linking in large amounts of (unused) 13351 ASN1 code. Grouping together similar functions and splitting unrelated 13352 functions prevents this. 13353 13354 *Steve Henson* 13355 13356 * Cleanup of EVP macros. 13357 13358 *Ben Laurie* 13359 13360 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the 13361 correct `_ecb suffix`. 13362 13363 *Ben Laurie* 13364 13365 * Add initial OCSP responder support to ocsp application. The 13366 revocation information is handled using the text based index 13367 use by the ca application. The responder can either handle 13368 requests generated internally, supplied in files (for example 13369 via a CGI script) or using an internal minimal server. 13370 13371 *Steve Henson* 13372 13373 * Add configuration choices to get zlib compression for TLS. 13374 13375 *Richard Levitte* 13376 13377 * Changes to Kerberos SSL for RFC 2712 compliance: 13378 1. Implemented real KerberosWrapper, instead of just using 13379 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 13380 2. Implemented optional authenticator field of KerberosWrapper. 13381 13382 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 13383 and authenticator structs; see crypto/krb5/. 13384 13385 Generalized Kerberos calls to support multiple Kerberos libraries. 13386 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu> 13387 via Richard Levitte* 13388 13389 * Cause 'openssl speed' to use fully hard-coded DSA keys as it 13390 already does with RSA. testdsa.h now has 'priv_key/pub_key' 13391 values for each of the key sizes rather than having just 13392 parameters (and 'speed' generating keys each time). 13393 13394 *Geoff Thorpe* 13395 13396 * Speed up EVP routines. 13397 Before: 13398crypt 13399pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 13400s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 13401s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 13402s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 13403crypt 13404s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 13405s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 13406s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 13407 After: 13408crypt 13409s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 13410crypt 13411s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 13412 13413 *Ben Laurie* 13414 13415 * Added the OS2-EMX target. 13416 13417 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte* 13418 13419 * Rewrite commands to use `NCONF` routines instead of the old `CONF`. 13420 New functions to support `NCONF` routines in extension code. 13421 New function `CONF_set_nconf()` 13422 to allow functions which take an `NCONF` to also handle the old `LHASH` 13423 structure: this means that the old `CONF` compatible routines can be 13424 retained (in particular w.rt. extensions) without having to duplicate the 13425 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack. 13426 13427 *Steve Henson* 13428 13429 * Enhance the general user interface with mechanisms for inner control 13430 and with possibilities to have yes/no kind of prompts. 13431 13432 *Richard Levitte* 13433 13434 * Change all calls to low-level digest routines in the library and 13435 applications to use EVP. Add missing calls to HMAC_cleanup() and 13436 don't assume HMAC_CTX can be copied using memcpy(). 13437 13438 *Verdon Walker <VWalker@novell.com>, Steve Henson* 13439 13440 * Add the possibility to control engines through control names but with 13441 arbitrary arguments instead of just a string. 13442 Change the key loaders to take a UI_METHOD instead of a callback 13443 function pointer. NOTE: this breaks binary compatibility with earlier 13444 versions of OpenSSL [engine]. 13445 Adapt the nCipher code for these new conditions and add a card insertion 13446 callback. 13447 13448 *Richard Levitte* 13449 13450 * Enhance the general user interface with mechanisms to better support 13451 dialog box interfaces, application-defined prompts, the possibility 13452 to use defaults (for example default passwords from somewhere else) 13453 and interrupts/cancellations. 13454 13455 *Richard Levitte* 13456 13457 * Tidy up PKCS#12 attribute handling. Add support for the CSP name 13458 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 13459 13460 *Steve Henson* 13461 13462 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 13463 tidy up some unnecessarily weird code in 'sk_new()'). 13464 13465 *Geoff, reported by Diego Tartara <dtartara@novamens.com>* 13466 13467 * Change the key loading routines for ENGINEs to use the same kind 13468 callback (pem_password_cb) as all other routines that need this 13469 kind of callback. 13470 13471 *Richard Levitte* 13472 13473 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 13474 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 13475 than this minimum value is recommended. 13476 13477 *Lutz Jaenicke* 13478 13479 * New random seeder for OpenVMS, using the system process statistics 13480 that are easily reachable. 13481 13482 *Richard Levitte* 13483 13484 * Windows apparently can't transparently handle global 13485 variables defined in DLLs. Initialisations such as: 13486 13487 const ASN1_ITEM *it = &ASN1_INTEGER_it; 13488 13489 won't compile. This is used by the any applications that need to 13490 declare their own ASN1 modules. This was fixed by adding the option 13491 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 13492 needed for static libraries under Win32. 13493 13494 *Steve Henson* 13495 13496 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 13497 setting of purpose and trust fields. New X509_STORE trust and 13498 purpose functions and tidy up setting in other SSL functions. 13499 13500 *Steve Henson* 13501 13502 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 13503 structure. These are inherited by X509_STORE_CTX when it is 13504 initialised. This allows various defaults to be set in the 13505 X509_STORE structure (such as flags for CRL checking and custom 13506 purpose or trust settings) for functions which only use X509_STORE_CTX 13507 internally such as S/MIME. 13508 13509 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 13510 trust settings if they are not set in X509_STORE. This allows X509_STORE 13511 purposes and trust (in S/MIME for example) to override any set by default. 13512 13513 Add command line options for CRL checking to smime, s_client and s_server 13514 applications. 13515 13516 *Steve Henson* 13517 13518 * Initial CRL based revocation checking. If the CRL checking flag(s) 13519 are set then the CRL is looked up in the X509_STORE structure and 13520 its validity and signature checked, then if the certificate is found 13521 in the CRL the verify fails with a revoked error. 13522 13523 Various new CRL related callbacks added to X509_STORE_CTX structure. 13524 13525 Command line options added to 'verify' application to support this. 13526 13527 This needs some additional work, such as being able to handle multiple 13528 CRLs with different times, extension based lookup (rather than just 13529 by subject name) and ultimately more complete V2 CRL extension 13530 handling. 13531 13532 *Steve Henson* 13533 13534 * Add a general user interface API (crypto/ui/). This is designed 13535 to replace things like des_read_password and friends (backward 13536 compatibility functions using this new API are provided). 13537 The purpose is to remove prompting functions from the DES code 13538 section as well as provide for prompting through dialog boxes in 13539 a window system and the like. 13540 13541 *Richard Levitte* 13542 13543 * Add "ex_data" support to ENGINE so implementations can add state at a 13544 per-structure level rather than having to store it globally. 13545 13546 *Geoff* 13547 13548 * Make it possible for ENGINE structures to be copied when retrieved by 13549 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 13550 This causes the "original" ENGINE structure to act like a template, 13551 analogous to the RSA vs. RSA_METHOD type of separation. Because of this 13552 operational state can be localised to each ENGINE structure, despite the 13553 fact they all share the same "methods". New ENGINE structures returned in 13554 this case have no functional references and the return value is the single 13555 structural reference. This matches the single structural reference returned 13556 by ENGINE_by_id() normally, when it is incremented on the pre-existing 13557 ENGINE structure. 13558 13559 *Geoff* 13560 13561 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 13562 needs to match any other type at all we need to manually clear the 13563 tag cache. 13564 13565 *Steve Henson* 13566 13567 * Changes to the "openssl engine" utility to include; 13568 - verbosity levels ('-v', '-vv', and '-vvv') that provide information 13569 about an ENGINE's available control commands. 13570 - executing control commands from command line arguments using the 13571 '-pre' and '-post' switches. '-post' is only used if '-t' is 13572 specified and the ENGINE is successfully initialised. The syntax for 13573 the individual commands are colon-separated, for example; 13574 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 13575 13576 *Geoff* 13577 13578 * New dynamic control command support for ENGINEs. ENGINEs can now 13579 declare their own commands (numbers), names (strings), descriptions, 13580 and input types for run-time discovery by calling applications. A 13581 subset of these commands are implicitly classed as "executable" 13582 depending on their input type, and only these can be invoked through 13583 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 13584 can be based on user input, config files, etc). The distinction is 13585 that "executable" commands cannot return anything other than a boolean 13586 result and can only support numeric or string input, whereas some 13587 discoverable commands may only be for direct use through 13588 ENGINE_ctrl(), eg. supporting the exchange of binary data, function 13589 pointers, or other custom uses. The "executable" commands are to 13590 support parameterisations of ENGINE behaviour that can be 13591 unambiguously defined by ENGINEs and used consistently across any 13592 OpenSSL-based application. Commands have been added to all the 13593 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 13594 control over shared-library paths without source code alterations. 13595 13596 *Geoff* 13597 13598 * Changed all ENGINE implementations to dynamically allocate their 13599 ENGINEs rather than declaring them statically. Apart from this being 13600 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 13601 this also allows the implementations to compile without using the 13602 internal engine_int.h header. 13603 13604 *Geoff* 13605 13606 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 13607 'const' value. Any code that should be able to modify a RAND_METHOD 13608 should already have non-const pointers to it (ie. they should only 13609 modify their own ones). 13610 13611 *Geoff* 13612 13613 * Made a variety of little tweaks to the ENGINE code. 13614 - "atalla" and "ubsec" string definitions were moved from header files 13615 to C code. "nuron" string definitions were placed in variables 13616 rather than hard-coded - allowing parameterisation of these values 13617 later on via ctrl() commands. 13618 - Removed unused "#if 0"'d code. 13619 - Fixed engine list iteration code so it uses ENGINE_free() to release 13620 structural references. 13621 - Constified the RAND_METHOD element of ENGINE structures. 13622 - Constified various get/set functions as appropriate and added 13623 missing functions (including a catch-all ENGINE_cpy that duplicates 13624 all ENGINE values onto a new ENGINE except reference counts/state). 13625 - Removed NULL parameter checks in get/set functions. Setting a method 13626 or function to NULL is a way of cancelling out a previously set 13627 value. Passing a NULL ENGINE parameter is just plain stupid anyway 13628 and doesn't justify the extra error symbols and code. 13629 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 13630 flags from engine_int.h to engine.h. 13631 - Changed prototypes for ENGINE handler functions (init(), finish(), 13632 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 13633 13634 *Geoff* 13635 13636 * Implement binary inversion algorithm for BN_mod_inverse in addition 13637 to the algorithm using long division. The binary algorithm can be 13638 used only if the modulus is odd. On 32-bit systems, it is faster 13639 only for relatively small moduli (roughly 20-30% for 128-bit moduli, 13640 roughly 5-15% for 256-bit moduli), so we use it only for moduli 13641 up to 450 bits. In 64-bit environments, the binary algorithm 13642 appears to be advantageous for much longer moduli; here we use it 13643 for moduli up to 2048 bits. 13644 13645 *Bodo Moeller* 13646 13647 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 13648 could not support the combine flag in choice fields. 13649 13650 *Steve Henson* 13651 13652 * Add a 'copy_extensions' option to the 'ca' utility. This copies 13653 extensions from a certificate request to the certificate. 13654 13655 *Steve Henson* 13656 13657 * Allow multiple 'certopt' and 'nameopt' options to be separated 13658 by commas. Add 'namopt' and 'certopt' options to the 'ca' config 13659 file: this allows the display of the certificate about to be 13660 signed to be customised, to allow certain fields to be included 13661 or excluded and extension details. The old system didn't display 13662 multicharacter strings properly, omitted fields not in the policy 13663 and couldn't display additional details such as extensions. 13664 13665 *Steve Henson* 13666 13667 * Function EC_POINTs_mul for multiple scalar multiplication 13668 of an arbitrary number of elliptic curve points 13669 \sum scalars[i]*points[i], 13670 optionally including the generator defined for the EC_GROUP: 13671 scalar*generator + \sum scalars[i]*points[i]. 13672 13673 EC_POINT_mul is a simple wrapper function for the typical case 13674 that the point list has just one item (besides the optional 13675 generator). 13676 13677 *Bodo Moeller* 13678 13679 * First EC_METHODs for curves over GF(p): 13680 13681 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 13682 operations and provides various method functions that can also 13683 operate with faster implementations of modular arithmetic. 13684 13685 EC_GFp_mont_method() reuses most functions that are part of 13686 EC_GFp_simple_method, but uses Montgomery arithmetic. 13687 13688 *Bodo Moeller; point addition and point doubling 13689 implementation directly derived from source code provided by 13690 Lenka Fibikova <fibikova@exp-math.uni-essen.de>* 13691 13692 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 13693 crypto/ec/ec_lib.c): 13694 13695 Curves are EC_GROUP objects (with an optional group generator) 13696 based on EC_METHODs that are built into the library. 13697 13698 Points are EC_POINT objects based on EC_GROUP objects. 13699 13700 Most of the framework would be able to handle curves over arbitrary 13701 finite fields, but as there are no obvious types for fields other 13702 than GF(p), some functions are limited to that for now. 13703 13704 *Bodo Moeller* 13705 13706 * Add the -HTTP option to s_server. It is similar to -WWW, but requires 13707 that the file contains a complete HTTP response. 13708 13709 *Richard Levitte* 13710 13711 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 13712 change the def and num file printf format specifier from "%-40sXXX" 13713 to "%-39s XXX". The latter will always guarantee a space after the 13714 field while the former will cause them to run together if the field 13715 is 40 of more characters long. 13716 13717 *Steve Henson* 13718 13719 * Constify the cipher and digest 'method' functions and structures 13720 and modify related functions to take constant EVP_MD and EVP_CIPHER 13721 pointers. 13722 13723 *Steve Henson* 13724 13725 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them 13726 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 13727 13728 *Bodo Moeller* 13729 13730 * Modify `EVP_Digest*()` routines so they now return values. Although the 13731 internal software routines can never fail additional hardware versions 13732 might. 13733 13734 *Steve Henson* 13735 13736 * Clean up crypto/err/err.h and change some error codes to avoid conflicts: 13737 13738 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 13739 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 13740 13741 ASN1 error codes 13742 ERR_R_NESTED_ASN1_ERROR 13743 ... 13744 ERR_R_MISSING_ASN1_EOS 13745 were 4 .. 9, conflicting with 13746 ERR_LIB_RSA (= ERR_R_RSA_LIB) 13747 ... 13748 ERR_LIB_PEM (= ERR_R_PEM_LIB). 13749 They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 13750 13751 Add new error code 'ERR_R_INTERNAL_ERROR'. 13752 13753 *Bodo Moeller* 13754 13755 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 13756 suffices. 13757 13758 *Bodo Moeller* 13759 13760 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This 13761 sets the subject name for a new request or supersedes the 13762 subject name in a given request. Formats that can be parsed are 13763 'CN=Some Name, OU=myOU, C=IT' 13764 and 13765 'CN=Some Name/OU=myOU/C=IT'. 13766 13767 Add options '-batch' and '-verbose' to 'openssl req'. 13768 13769 *Massimiliano Pala <madwolf@hackmasters.net>* 13770 13771 * Introduce the possibility to access global variables through 13772 functions on platform were that's the best way to handle exporting 13773 global variables in shared libraries. To enable this functionality, 13774 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 13775 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 13776 is normally done by Configure or something similar). 13777 13778 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 13779 in the source file (foo.c) like this: 13780 13781 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 13782 OPENSSL_IMPLEMENT_GLOBAL(double,bar); 13783 13784 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 13785 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 13786 13787 OPENSSL_DECLARE_GLOBAL(int,foo); 13788 #define foo OPENSSL_GLOBAL_REF(foo) 13789 OPENSSL_DECLARE_GLOBAL(double,bar); 13790 #define bar OPENSSL_GLOBAL_REF(bar) 13791 13792 The #defines are very important, and therefore so is including the 13793 header file everywhere where the defined globals are used. 13794 13795 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 13796 of ASN.1 items, but that structure is a bit different. 13797 13798 The largest change is in util/mkdef.pl which has been enhanced with 13799 better and easier to understand logic to choose which symbols should 13800 go into the Windows .def files as well as a number of fixes and code 13801 cleanup (among others, algorithm keywords are now sorted 13802 lexicographically to avoid constant rewrites). 13803 13804 *Richard Levitte* 13805 13806 * In BN_div() keep a copy of the sign of 'num' before writing the 13807 result to 'rm' because if rm==num the value will be overwritten 13808 and produce the wrong result if 'num' is negative: this caused 13809 problems with BN_mod() and BN_nnmod(). 13810 13811 *Steve Henson* 13812 13813 * Function OCSP_request_verify(). This checks the signature on an 13814 OCSP request and verifies the signer certificate. The signer 13815 certificate is just checked for a generic purpose and OCSP request 13816 trust settings. 13817 13818 *Steve Henson* 13819 13820 * Add OCSP_check_validity() function to check the validity of OCSP 13821 responses. OCSP responses are prepared in real time and may only 13822 be a few seconds old. Simply checking that the current time lies 13823 between thisUpdate and nextUpdate max reject otherwise valid responses 13824 caused by either OCSP responder or client clock inaccuracy. Instead 13825 we allow thisUpdate and nextUpdate to fall within a certain period of 13826 the current time. The age of the response can also optionally be 13827 checked. Two new options -validity_period and -status_age added to 13828 ocsp utility. 13829 13830 *Steve Henson* 13831 13832 * If signature or public key algorithm is unrecognized print out its 13833 OID rather that just UNKNOWN. 13834 13835 *Steve Henson* 13836 13837 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 13838 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 13839 ID to be generated from the issuer certificate alone which can then be 13840 passed to OCSP_id_issuer_cmp(). 13841 13842 *Steve Henson* 13843 13844 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 13845 ASN1 modules to export functions returning ASN1_ITEM pointers 13846 instead of the ASN1_ITEM structures themselves. This adds several 13847 new macros which allow the underlying ASN1 function/structure to 13848 be accessed transparently. As a result code should not use ASN1_ITEM 13849 references directly (such as &X509_it) but instead use the relevant 13850 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 13851 use of the new ASN1 code on platforms where exporting structures 13852 is problematical (for example in shared libraries) but exporting 13853 functions returning pointers to structures is not. 13854 13855 *Steve Henson* 13856 13857 * Add support for overriding the generation of SSL/TLS session IDs. 13858 These callbacks can be registered either in an SSL_CTX or per SSL. 13859 The purpose of this is to allow applications to control, if they wish, 13860 the arbitrary values chosen for use as session IDs, particularly as it 13861 can be useful for session caching in multiple-server environments. A 13862 command-line switch for testing this (and any client code that wishes 13863 to use such a feature) has been added to "s_server". 13864 13865 *Geoff Thorpe, Lutz Jaenicke* 13866 13867 * Modify mkdef.pl to recognise and parse preprocessor conditionals 13868 of the form `#if defined(...) || defined(...) || ...` and 13869 `#if !defined(...) && !defined(...) && ...`. This also avoids 13870 the growing number of special cases it was previously handling. 13871 13872 *Richard Levitte* 13873 13874 * Make all configuration macros available for application by making 13875 sure they are available in opensslconf.h, by giving them names starting 13876 with `OPENSSL_` to avoid conflicts with other packages and by making 13877 sure e_os2.h will cover all platform-specific cases together with 13878 opensslconf.h. 13879 Additionally, it is now possible to define configuration/platform- 13880 specific names (called "system identities"). In the C code, these 13881 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another 13882 macro with the name beginning with `OPENSSL_SYS_`, which is determined 13883 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on 13884 what is available. 13885 13886 *Richard Levitte* 13887 13888 * New option -set_serial to 'req' and 'x509' this allows the serial 13889 number to use to be specified on the command line. Previously self 13890 signed certificates were hard coded with serial number 0 and the 13891 CA options of 'x509' had to use a serial number in a file which was 13892 auto incremented. 13893 13894 *Steve Henson* 13895 13896 * New options to 'ca' utility to support V2 CRL entry extensions. 13897 Currently CRL reason, invalidity date and hold instruction are 13898 supported. Add new CRL extensions to V3 code and some new objects. 13899 13900 *Steve Henson* 13901 13902 * New function EVP_CIPHER_CTX_set_padding() this is used to 13903 disable standard block padding (aka PKCS#5 padding) in the EVP 13904 API, which was previously mandatory. This means that the data is 13905 not padded in any way and so the total length much be a multiple 13906 of the block size, otherwise an error occurs. 13907 13908 *Steve Henson* 13909 13910 * Initial (incomplete) OCSP SSL support. 13911 13912 *Steve Henson* 13913 13914 * New function OCSP_parse_url(). This splits up a URL into its host, 13915 port and path components: primarily to parse OCSP URLs. New -url 13916 option to ocsp utility. 13917 13918 *Steve Henson* 13919 13920 * New nonce behavior. The return value of OCSP_check_nonce() now 13921 reflects the various checks performed. Applications can decide 13922 whether to tolerate certain situations such as an absent nonce 13923 in a response when one was present in a request: the ocsp application 13924 just prints out a warning. New function OCSP_add1_basic_nonce() 13925 this is to allow responders to include a nonce in a response even if 13926 the request is nonce-less. 13927 13928 *Steve Henson* 13929 13930 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are 13931 skipped when using openssl x509 multiple times on a single input file, 13932 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`. 13933 13934 *Bodo Moeller* 13935 13936 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 13937 set string type: to handle setting ASN1_TIME structures. Fix ca 13938 utility to correctly initialize revocation date of CRLs. 13939 13940 *Steve Henson* 13941 13942 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 13943 the clients preferred ciphersuites and rather use its own preferences. 13944 Should help to work around M$ SGC (Server Gated Cryptography) bug in 13945 Internet Explorer by ensuring unchanged hash method during stepup. 13946 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 13947 13948 *Lutz Jaenicke* 13949 13950 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 13951 to aes and add a new 'exist' option to print out symbols that don't 13952 appear to exist. 13953 13954 *Steve Henson* 13955 13956 * Additional options to ocsp utility to allow flags to be set and 13957 additional certificates supplied. 13958 13959 *Steve Henson* 13960 13961 * Add the option -VAfile to 'openssl ocsp', so the user can give the 13962 OCSP client a number of certificate to only verify the response 13963 signature against. 13964 13965 *Richard Levitte* 13966 13967 * Update Rijndael code to version 3.0 and change EVP AES ciphers to 13968 handle the new API. Currently only ECB, CBC modes supported. Add new 13969 AES OIDs. 13970 13971 Add TLS AES ciphersuites as described in RFC3268, "Advanced 13972 Encryption Standard (AES) Ciphersuites for Transport Layer 13973 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 13974 not enabled by default and were not part of the "ALL" ciphersuite 13975 alias because they were not yet official; they could be 13976 explicitly requested by specifying the "AESdraft" ciphersuite 13977 group alias. In the final release of OpenSSL 0.9.7, the group 13978 alias is called "AES" and is part of "ALL".) 13979 13980 *Ben Laurie, Steve Henson, Bodo Moeller* 13981 13982 * New function OCSP_copy_nonce() to copy nonce value (if present) from 13983 request to response. 13984 13985 *Steve Henson* 13986 13987 * Functions for OCSP responders. OCSP_request_onereq_count(), 13988 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 13989 extract information from a certificate request. OCSP_response_create() 13990 creates a response and optionally adds a basic response structure. 13991 OCSP_basic_add1_status() adds a complete single response to a basic 13992 response and returns the OCSP_SINGLERESP structure just added (to allow 13993 extensions to be included for example). OCSP_basic_add1_cert() adds a 13994 certificate to a basic response and OCSP_basic_sign() signs a basic 13995 response with various flags. New helper functions ASN1_TIME_check() 13996 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 13997 (converts ASN1_TIME to GeneralizedTime). 13998 13999 *Steve Henson* 14000 14001 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 14002 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 14003 structure from a certificate. X509_pubkey_digest() digests the public_key 14004 contents: this is used in various key identifiers. 14005 14006 *Steve Henson* 14007 14008 * Make sk_sort() tolerate a NULL argument. 14009 14010 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>* 14011 14012 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 14013 passed by the function are trusted implicitly. If any of them signed the 14014 response then it is assumed to be valid and is not verified. 14015 14016 *Steve Henson* 14017 14018 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 14019 to data. This was previously part of the PKCS7 ASN1 code. This 14020 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 14021 *Steve Henson, reported by Kenneth R. Robinette 14022 <support@securenetterm.com>* 14023 14024 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 14025 routines: without these tracing memory leaks is very painful. 14026 Fix leaks in PKCS12 and PKCS7 routines. 14027 14028 *Steve Henson* 14029 14030 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 14031 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 14032 effectively meant GeneralizedTime would never be used. Now it 14033 is initialised to -1 but X509_time_adj() now has to check the value 14034 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 14035 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 14036 *Steve Henson, reported by Kenneth R. Robinette 14037 <support@securenetterm.com>* 14038 14039 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 14040 result in a zero length in the ASN1_INTEGER structure which was 14041 not consistent with the structure when d2i_ASN1_INTEGER() was used 14042 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 14043 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 14044 where it did not print out a minus for negative ASN1_INTEGER. 14045 14046 *Steve Henson* 14047 14048 * Add summary printout to ocsp utility. The various functions which 14049 convert status values to strings have been renamed to: 14050 OCSP_response_status_str(), OCSP_cert_status_str() and 14051 OCSP_crl_reason_str() and are no longer static. New options 14052 to verify nonce values and to disable verification. OCSP response 14053 printout format cleaned up. 14054 14055 *Steve Henson* 14056 14057 * Add additional OCSP certificate checks. These are those specified 14058 in RFC2560. This consists of two separate checks: the CA of the 14059 certificate being checked must either be the OCSP signer certificate 14060 or the issuer of the OCSP signer certificate. In the latter case the 14061 OCSP signer certificate must contain the OCSP signing extended key 14062 usage. This check is performed by attempting to match the OCSP 14063 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 14064 in the OCSP_CERTID structures of the response. 14065 14066 *Steve Henson* 14067 14068 * Initial OCSP certificate verification added to OCSP_basic_verify() 14069 and related routines. This uses the standard OpenSSL certificate 14070 verify routines to perform initial checks (just CA validity) and 14071 to obtain the certificate chain. Then additional checks will be 14072 performed on the chain. Currently the root CA is checked to see 14073 if it is explicitly trusted for OCSP signing. This is used to set 14074 a root CA as a global signing root: that is any certificate that 14075 chains to that CA is an acceptable OCSP signing certificate. 14076 14077 *Steve Henson* 14078 14079 * New '-extfile ...' option to 'openssl ca' for reading X.509v3 14080 extensions from a separate configuration file. 14081 As when reading extensions from the main configuration file, 14082 the '-extensions ...' option may be used for specifying the 14083 section to use. 14084 14085 *Massimiliano Pala <madwolf@comune.modena.it>* 14086 14087 * New OCSP utility. Allows OCSP requests to be generated or 14088 read. The request can be sent to a responder and the output 14089 parsed, outputted or printed in text form. Not complete yet: 14090 still needs to check the OCSP response validity. 14091 14092 *Steve Henson* 14093 14094 * New subcommands for 'openssl ca': 14095 `openssl ca -status <serial>` prints the status of the cert with 14096 the given serial number (according to the index file). 14097 `openssl ca -updatedb` updates the expiry status of certificates 14098 in the index file. 14099 14100 *Massimiliano Pala <madwolf@comune.modena.it>* 14101 14102 * New '-newreq-nodes' command option to CA.pl. This is like 14103 '-newreq', but calls 'openssl req' with the '-nodes' option 14104 so that the resulting key is not encrypted. 14105 14106 *Damien Miller <djm@mindrot.org>* 14107 14108 * New configuration for the GNU Hurd. 14109 14110 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte* 14111 14112 * Initial code to implement OCSP basic response verify. This 14113 is currently incomplete. Currently just finds the signer's 14114 certificate and verifies the signature on the response. 14115 14116 *Steve Henson* 14117 14118 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in 14119 value of OPENSSLDIR. This is available via the new '-d' option 14120 to 'openssl version', and is also included in 'openssl version -a'. 14121 14122 *Bodo Moeller* 14123 14124 * Allowing defining memory allocation callbacks that will be given 14125 file name and line number information in additional arguments 14126 (a `const char*` and an int). The basic functionality remains, as 14127 well as the original possibility to just replace malloc(), 14128 realloc() and free() by functions that do not know about these 14129 additional arguments. To register and find out the current 14130 settings for extended allocation functions, the following 14131 functions are provided: 14132 14133 CRYPTO_set_mem_ex_functions 14134 CRYPTO_set_locked_mem_ex_functions 14135 CRYPTO_get_mem_ex_functions 14136 CRYPTO_get_locked_mem_ex_functions 14137 14138 These work the same way as CRYPTO_set_mem_functions and friends. 14139 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an 14140 extended allocation function is enabled. 14141 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where 14142 a conventional allocation function is enabled. 14143 14144 *Richard Levitte, Bodo Moeller* 14145 14146 * Finish off removing the remaining LHASH function pointer casts. 14147 There should no longer be any prototype-casting required when using 14148 the LHASH abstraction, and any casts that remain are "bugs". See 14149 the callback types and macros at the head of lhash.h for details 14150 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 14151 14152 *Geoff Thorpe* 14153 14154 * Add automatic query of EGD sockets in RAND_poll() for the unix variant. 14155 If /dev/[u]random devices are not available or do not return enough 14156 entropy, EGD style sockets (served by EGD or PRNGD) will automatically 14157 be queried. 14158 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 14159 /etc/entropy will be queried once each in this sequence, querying stops 14160 when enough entropy was collected without querying more sockets. 14161 14162 *Lutz Jaenicke* 14163 14164 * Change the Unix RAND_poll() variant to be able to poll several 14165 random devices, as specified by DEVRANDOM, until a sufficient amount 14166 of data has been collected. We spend at most 10 ms on each file 14167 (select timeout) and read in non-blocking mode. DEVRANDOM now 14168 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 14169 (previously it was just the string "/dev/urandom"), so on typical 14170 platforms the 10 ms delay will never occur. 14171 Also separate out the Unix variant to its own file, rand_unix.c. 14172 For VMS, there's a currently-empty rand_vms.c. 14173 14174 *Richard Levitte* 14175 14176 * Move OCSP client related routines to ocsp_cl.c. These 14177 provide utility functions which an application needing 14178 to issue a request to an OCSP responder and analyse the 14179 response will typically need: as opposed to those which an 14180 OCSP responder itself would need which will be added later. 14181 14182 OCSP_request_sign() signs an OCSP request with an API similar 14183 to PKCS7_sign(). OCSP_response_status() returns status of OCSP 14184 response. OCSP_response_get1_basic() extracts basic response 14185 from response. OCSP_resp_find_status(): finds and extracts status 14186 information from an OCSP_CERTID structure (which will be created 14187 when the request structure is built). These are built from lower 14188 level functions which work on OCSP_SINGLERESP structures but 14189 won't normally be used unless the application wishes to examine 14190 extensions in the OCSP response for example. 14191 14192 Replace nonce routines with a pair of functions. 14193 OCSP_request_add1_nonce() adds a nonce value and optionally 14194 generates a random value. OCSP_check_nonce() checks the 14195 validity of the nonce in an OCSP response. 14196 14197 *Steve Henson* 14198 14199 * Change function OCSP_request_add() to OCSP_request_add0_id(). 14200 This doesn't copy the supplied OCSP_CERTID and avoids the 14201 need to free up the newly created id. Change return type 14202 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 14203 This can then be used to add extensions to the request. 14204 Deleted OCSP_request_new(), since most of its functionality 14205 is now in OCSP_REQUEST_new() (and the case insensitive name 14206 clash) apart from the ability to set the request name which 14207 will be added elsewhere. 14208 14209 *Steve Henson* 14210 14211 * Update OCSP API. Remove obsolete extensions argument from 14212 various functions. Extensions are now handled using the new 14213 OCSP extension code. New simple OCSP HTTP function which 14214 can be used to send requests and parse the response. 14215 14216 *Steve Henson* 14217 14218 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 14219 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 14220 uses the special reorder version of SET OF to sort the attributes 14221 and reorder them to match the encoded order. This resolves a long 14222 standing problem: a verify on a PKCS7 structure just after signing 14223 it used to fail because the attribute order did not match the 14224 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 14225 it uses the received order. This is necessary to tolerate some broken 14226 software that does not order SET OF. This is handled by encoding 14227 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 14228 to produce the required SET OF. 14229 14230 *Steve Henson* 14231 14232 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 14233 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 14234 files to get correct declarations of the ASN.1 item variables. 14235 14236 *Richard Levitte* 14237 14238 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 14239 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 14240 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 14241 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 14242 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 14243 ASN1_ITEM and no wrapper functions. 14244 14245 *Steve Henson* 14246 14247 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 14248 replace the old function pointer based I/O routines. Change most of 14249 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these. 14250 14251 *Steve Henson* 14252 14253 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor 14254 lines, recognize more "algorithms" that can be deselected, and make 14255 it complain about algorithm deselection that isn't recognised. 14256 14257 *Richard Levitte* 14258 14259 * New ASN1 functions to handle dup, sign, verify, digest, pack and 14260 unpack operations in terms of ASN1_ITEM. Modify existing wrappers 14261 to use new functions. Add NO_ASN1_OLD which can be set to remove 14262 some old style ASN1 functions: this can be used to determine if old 14263 code will still work when these eventually go away. 14264 14265 *Steve Henson* 14266 14267 * New extension functions for OCSP structures, these follow the 14268 same conventions as certificates and CRLs. 14269 14270 *Steve Henson* 14271 14272 * New function X509V3_add1_i2d(). This automatically encodes and 14273 adds an extension. Its behaviour can be customised with various 14274 flags to append, replace or delete. Various wrappers added for 14275 certificates and CRLs. 14276 14277 *Steve Henson* 14278 14279 * Fix to avoid calling the underlying ASN1 print routine when 14280 an extension cannot be parsed. Correct a typo in the 14281 OCSP_SERVICELOC extension. Tidy up print OCSP format. 14282 14283 *Steve Henson* 14284 14285 * Make mkdef.pl parse some of the ASN1 macros and add appropriate 14286 entries for variables. 14287 14288 *Steve Henson* 14289 14290 * Add functionality to `apps/openssl.c` for detecting locking 14291 problems: As the program is single-threaded, all we have 14292 to do is register a locking callback using an array for 14293 storing which locks are currently held by the program. 14294 14295 *Bodo Moeller* 14296 14297 * Use a lock around the call to CRYPTO_get_ex_new_index() in 14298 SSL_get_ex_data_X509_STORE_idx(), which is used in 14299 ssl_verify_cert_chain() and thus can be called at any time 14300 during TLS/SSL handshakes so that thread-safety is essential. 14301 Unfortunately, the ex_data design is not at all suited 14302 for multi-threaded use, so it probably should be abolished. 14303 14304 *Bodo Moeller* 14305 14306 * Added Broadcom "ubsec" ENGINE to OpenSSL. 14307 14308 *Broadcom, tweaked and integrated by Geoff Thorpe* 14309 14310 * Move common extension printing code to new function 14311 X509V3_print_extensions(). Reorganise OCSP print routines and 14312 implement some needed OCSP ASN1 functions. Add OCSP extensions. 14313 14314 *Steve Henson* 14315 14316 * New function X509_signature_print() to remove duplication in some 14317 print routines. 14318 14319 *Steve Henson* 14320 14321 * Add a special meaning when SET OF and SEQUENCE OF flags are both 14322 set (this was treated exactly the same as SET OF previously). This 14323 is used to reorder the STACK representing the structure to match the 14324 encoding. This will be used to get round a problem where a PKCS7 14325 structure which was signed could not be verified because the STACK 14326 order did not reflect the encoded order. 14327 14328 *Steve Henson* 14329 14330 * Reimplement the OCSP ASN1 module using the new code. 14331 14332 *Steve Henson* 14333 14334 * Update the X509V3 code to permit the use of an ASN1_ITEM structure 14335 for its ASN1 operations. The old style function pointers still exist 14336 for now but they will eventually go away. 14337 14338 *Steve Henson* 14339 14340 * Merge in replacement ASN1 code from the ASN1 branch. This almost 14341 completely replaces the old ASN1 functionality with a table driven 14342 encoder and decoder which interprets an ASN1_ITEM structure describing 14343 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 14344 largely maintained. Almost all of the old asn1_mac.h macro based ASN1 14345 has also been converted to the new form. 14346 14347 *Steve Henson* 14348 14349 * Change BN_mod_exp_recp so that negative moduli are tolerated 14350 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 14351 so that BN_mod_exp_mont and BN_mod_exp_mont_word work 14352 for negative moduli. 14353 14354 *Bodo Moeller* 14355 14356 * Fix BN_uadd and BN_usub: Always return non-negative results instead 14357 of not touching the result's sign bit. 14358 14359 *Bodo Moeller* 14360 14361 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be 14362 set. 14363 14364 *Bodo Moeller* 14365 14366 * Changed the LHASH code to use prototypes for callbacks, and created 14367 macros to declare and implement thin (optionally static) functions 14368 that provide type-safety and avoid function pointer casting for the 14369 type-specific callbacks. 14370 14371 *Geoff Thorpe* 14372 14373 * Added Kerberos Cipher Suites to be used with TLS, as written in 14374 RFC 2712. 14375 *Veers Staats <staatsvr@asc.hpc.mil>, 14376 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte* 14377 14378 * Reformat the FAQ so the different questions and answers can be divided 14379 in sections depending on the subject. 14380 14381 *Richard Levitte* 14382 14383 * Have the zlib compression code load ZLIB.DLL dynamically under 14384 Windows. 14385 14386 *Richard Levitte* 14387 14388 * New function BN_mod_sqrt for computing square roots modulo a prime 14389 (using the probabilistic Tonelli-Shanks algorithm unless 14390 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 14391 be handled deterministically). 14392 14393 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller* 14394 14395 * Make BN_mod_inverse faster by explicitly handling small quotients 14396 in the Euclid loop. (Speed gain about 20% for small moduli [256 or 14397 512 bits], about 30% for larger ones [1024 or 2048 bits].) 14398 14399 *Bodo Moeller* 14400 14401 * New function BN_kronecker. 14402 14403 *Bodo Moeller* 14404 14405 * Fix BN_gcd so that it works on negative inputs; the result is 14406 positive unless both parameters are zero. 14407 Previously something reasonably close to an infinite loop was 14408 possible because numbers could be growing instead of shrinking 14409 in the implementation of Euclid's algorithm. 14410 14411 *Bodo Moeller* 14412 14413 * Fix BN_is_word() and BN_is_one() macros to take into account the 14414 sign of the number in question. 14415 14416 Fix BN_is_word(a,w) to work correctly for w == 0. 14417 14418 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 14419 because its test if the absolute value of 'a' equals 'w'. 14420 Note that BN_abs_is_word does *not* handle w == 0 reliably; 14421 it exists mostly for use in the implementations of BN_is_zero(), 14422 BN_is_one(), and BN_is_word(). 14423 14424 *Bodo Moeller* 14425 14426 * New function BN_swap. 14427 14428 *Bodo Moeller* 14429 14430 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 14431 the exponentiation functions are more likely to produce reasonable 14432 results on negative inputs. 14433 14434 *Bodo Moeller* 14435 14436 * Change BN_mod_mul so that the result is always non-negative. 14437 Previously, it could be negative if one of the factors was negative; 14438 I don't think anyone really wanted that behaviour. 14439 14440 *Bodo Moeller* 14441 14442 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c` 14443 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`, 14444 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`) 14445 and add new functions: 14446 14447 BN_nnmod 14448 BN_mod_sqr 14449 BN_mod_add 14450 BN_mod_add_quick 14451 BN_mod_sub 14452 BN_mod_sub_quick 14453 BN_mod_lshift1 14454 BN_mod_lshift1_quick 14455 BN_mod_lshift 14456 BN_mod_lshift_quick 14457 14458 These functions always generate non-negative results. 14459 14460 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r` 14461 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead). 14462 14463 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as 14464 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`] 14465 be reduced modulo `m`. 14466 14467 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller* 14468 14469<!-- 14470 The following entry accidentally appeared in the CHANGES file 14471 distributed with OpenSSL 0.9.7. The modifications described in 14472 it do *not* apply to OpenSSL 0.9.7. 14473 14474 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 14475 was actually never needed) and in BN_mul(). The removal in BN_mul() 14476 required a small change in bn_mul_part_recursive() and the addition 14477 of the functions bn_cmp_part_words(), bn_sub_part_words() and 14478 bn_add_part_words(), which do the same thing as bn_cmp_words(), 14479 bn_sub_words() and bn_add_words() except they take arrays with 14480 differing sizes. 14481 14482 *Richard Levitte* 14483--> 14484 14485 * In 'openssl passwd', verify passwords read from the terminal 14486 unless the '-salt' option is used (which usually means that 14487 verification would just waste user's time since the resulting 14488 hash is going to be compared with some given password hash) 14489 or the new '-noverify' option is used. 14490 14491 This is an incompatible change, but it does not affect 14492 non-interactive use of 'openssl passwd' (passwords on the command 14493 line, '-stdin' option, '-in ...' option) and thus should not 14494 cause any problems. 14495 14496 *Bodo Moeller* 14497 14498 * Remove all references to RSAref, since there's no more need for it. 14499 14500 *Richard Levitte* 14501 14502 * Make DSO load along a path given through an environment variable 14503 (SHLIB_PATH) with shl_load(). 14504 14505 *Richard Levitte* 14506 14507 * Constify the ENGINE code as a result of BIGNUM constification. 14508 Also constify the RSA code and most things related to it. In a 14509 few places, most notable in the depth of the ASN.1 code, ugly 14510 casts back to non-const were required (to be solved at a later 14511 time) 14512 14513 *Richard Levitte* 14514 14515 * Make it so the openssl application has all engines loaded by default. 14516 14517 *Richard Levitte* 14518 14519 * Constify the BIGNUM routines a little more. 14520 14521 *Richard Levitte* 14522 14523 * Add the following functions: 14524 14525 ENGINE_load_cswift() 14526 ENGINE_load_chil() 14527 ENGINE_load_atalla() 14528 ENGINE_load_nuron() 14529 ENGINE_load_builtin_engines() 14530 14531 That way, an application can itself choose if external engines that 14532 are built-in in OpenSSL shall ever be used or not. The benefit is 14533 that applications won't have to be linked with libdl or other dso 14534 libraries unless it's really needed. 14535 14536 Changed 'openssl engine' to load all engines on demand. 14537 Changed the engine header files to avoid the duplication of some 14538 declarations (they differed!). 14539 14540 *Richard Levitte* 14541 14542 * 'openssl engine' can now list capabilities. 14543 14544 *Richard Levitte* 14545 14546 * Better error reporting in 'openssl engine'. 14547 14548 *Richard Levitte* 14549 14550 * Never call load_dh_param(NULL) in s_server. 14551 14552 *Bodo Moeller* 14553 14554 * Add engine application. It can currently list engines by name and 14555 identity, and test if they are actually available. 14556 14557 *Richard Levitte* 14558 14559 * Improve RPM specification file by forcing symbolic linking and making 14560 sure the installed documentation is also owned by root.root. 14561 14562 *Damien Miller <djm@mindrot.org>* 14563 14564 * Give the OpenSSL applications more possibilities to make use of 14565 keys (public as well as private) handled by engines. 14566 14567 *Richard Levitte* 14568 14569 * Add OCSP code that comes from CertCo. 14570 14571 *Richard Levitte* 14572 14573 * Add VMS support for the Rijndael code. 14574 14575 *Richard Levitte* 14576 14577 * Added untested support for Nuron crypto accelerator. 14578 14579 *Ben Laurie* 14580 14581 * Add support for external cryptographic devices. This code was 14582 previously distributed separately as the "engine" branch. 14583 14584 *Geoff Thorpe, Richard Levitte* 14585 14586 * Rework the filename-translation in the DSO code. It is now possible to 14587 have far greater control over how a "name" is turned into a filename 14588 depending on the operating environment and any oddities about the 14589 different shared library filenames on each system. 14590 14591 *Geoff Thorpe* 14592 14593 * Support threads on FreeBSD-elf in Configure. 14594 14595 *Richard Levitte* 14596 14597 * Fix for SHA1 assembly problem with MASM: it produces 14598 warnings about corrupt line number information when assembling 14599 with debugging information. This is caused by the overlapping 14600 of two sections. 14601 14602 *Bernd Matthes <mainbug@celocom.de>, Steve Henson* 14603 14604 * NCONF changes. 14605 NCONF_get_number() has no error checking at all. As a replacement, 14606 NCONF_get_number_e() is defined (`_e` for "error checking") and is 14607 promoted strongly. The old NCONF_get_number is kept around for 14608 binary backward compatibility. 14609 Make it possible for methods to load from something other than a BIO, 14610 by providing a function pointer that is given a name instead of a BIO. 14611 For example, this could be used to load configuration data from an 14612 LDAP server. 14613 14614 *Richard Levitte* 14615 14616 * Fix for non blocking accept BIOs. Added new I/O special reason 14617 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 14618 with non blocking I/O was not possible because no retry code was 14619 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 14620 this case. 14621 14622 *Steve Henson* 14623 14624 * Added the beginnings of Rijndael support. 14625 14626 *Ben Laurie* 14627 14628 * Fix for bug in DirectoryString mask setting. Add support for 14629 X509_NAME_print_ex() in 'req' and X509_print_ex() function 14630 to allow certificate printing to more controllable, additional 14631 'certopt' option to 'x509' to allow new printing options to be 14632 set. 14633 14634 *Steve Henson* 14635 14636 * Clean old EAY MD5 hack from e_os.h. 14637 14638 *Richard Levitte* 14639 14640### Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 14641 14642 * Fix null-pointer assignment in do_change_cipher_spec() revealed 14643 by using the Codenomicon TLS Test Tool ([CVE-2004-0079]) 14644 14645 *Joe Orton, Steve Henson* 14646 14647### Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 14648 14649 * Fix additional bug revealed by the NISCC test suite: 14650 14651 Stop bug triggering large recursion when presented with 14652 certain ASN.1 tags ([CVE-2003-0851]) 14653 14654 *Steve Henson* 14655 14656### Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 14657 14658 * Fix various bugs revealed by running the NISCC test suite: 14659 14660 Stop out of bounds reads in the ASN1 code when presented with 14661 invalid tags (CVE-2003-0543 and CVE-2003-0544). 14662 14663 If verify callback ignores invalid public key errors don't try to check 14664 certificate signature with the NULL public key. 14665 14666 *Steve Henson* 14667 14668 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 14669 if the server requested one: as stated in TLS 1.0 and SSL 3.0 14670 specifications. 14671 14672 *Steve Henson* 14673 14674 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 14675 extra data after the compression methods not only for TLS 1.0 14676 but also for SSL 3.0 (as required by the specification). 14677 14678 *Bodo Moeller; problem pointed out by Matthias Loepfe* 14679 14680 * Change X509_certificate_type() to mark the key as exported/exportable 14681 when it's 512 *bits* long, not 512 bytes. 14682 14683 *Richard Levitte* 14684 14685### Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 14686 14687 * Countermeasure against the Klima-Pokorny-Rosa extension of 14688 Bleichbacher's attack on PKCS #1 v1.5 padding: treat 14689 a protocol version number mismatch like a decryption error 14690 in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 14691 14692 *Bodo Moeller* 14693 14694 * Turn on RSA blinding by default in the default implementation 14695 to avoid a timing attack. Applications that don't want it can call 14696 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 14697 They would be ill-advised to do so in most cases. 14698 14699 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller* 14700 14701 * Change RSA blinding code so that it works when the PRNG is not 14702 seeded (in this case, the secret RSA exponent is abused as 14703 an unpredictable seed -- if it is not unpredictable, there 14704 is no point in blinding anyway). Make RSA blinding thread-safe 14705 by remembering the creator's thread ID in rsa->blinding and 14706 having all other threads use local one-time blinding factors 14707 (this requires more computation than sharing rsa->blinding, but 14708 avoids excessive locking; and if an RSA object is not shared 14709 between threads, blinding will still be very fast). 14710 14711 *Bodo Moeller* 14712 14713### Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 14714 14715 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 14716 via timing by performing a MAC computation even if incorrect 14717 block cipher padding has been found. This is a countermeasure 14718 against active attacks where the attacker has to distinguish 14719 between bad padding and a MAC verification error. ([CVE-2003-0078]) 14720 14721 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 14722 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 14723 Martin Vuagnoux (EPFL, Ilion)* 14724 14725### Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 14726 14727 * New function OPENSSL_cleanse(), which is used to cleanse a section of 14728 memory from its contents. This is done with a counter that will 14729 place alternating values in each byte. This can be used to solve 14730 two issues: 1) the removal of calls to memset() by highly optimizing 14731 compilers, and 2) cleansing with other values than 0, since those can 14732 be read through on certain media, for example a swap space on disk. 14733 14734 *Geoff Thorpe* 14735 14736 * Bugfix: client side session caching did not work with external caching, 14737 because the session->cipher setting was not restored when reloading 14738 from the external cache. This problem was masked, when 14739 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 14740 (Found by Steve Haslam <steve@araqnid.ddts.net>.) 14741 14742 *Lutz Jaenicke* 14743 14744 * Fix client_certificate (ssl/s2_clnt.c): The permissible total 14745 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 14746 14747 *Zeev Lieber <zeev-l@yahoo.com>* 14748 14749 * Undo an undocumented change introduced in 0.9.6e which caused 14750 repeated calls to OpenSSL_add_all_ciphers() and 14751 OpenSSL_add_all_digests() to be ignored, even after calling 14752 EVP_cleanup(). 14753 14754 *Richard Levitte* 14755 14756 * Change the default configuration reader to deal with last line not 14757 being properly terminated. 14758 14759 *Richard Levitte* 14760 14761 * Change X509_NAME_cmp() so it applies the special rules on handling 14762 DN values that are of type PrintableString, as well as RDNs of type 14763 emailAddress where the value has the type ia5String. 14764 14765 *stefank@valicert.com via Richard Levitte* 14766 14767 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 14768 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 14769 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 14770 the bitwise-OR of the two for use by the majority of applications 14771 wanting this behaviour, and update the docs. The documented 14772 behaviour and actual behaviour were inconsistent and had been 14773 changing anyway, so this is more a bug-fix than a behavioural 14774 change. 14775 14776 *Geoff Thorpe, diagnosed by Nadav Har'El* 14777 14778 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 14779 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 14780 14781 *Bodo Moeller* 14782 14783 * Fix initialization code race conditions in 14784 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 14785 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 14786 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 14787 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 14788 ssl2_get_cipher_by_char(), 14789 ssl3_get_cipher_by_char(). 14790 14791 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller* 14792 14793 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 14794 the cached sessions are flushed, as the remove_cb() might use ex_data 14795 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 14796 (see [openssl.org #212]). 14797 14798 *Geoff Thorpe, Lutz Jaenicke* 14799 14800 * Fix typo in OBJ_txt2obj which incorrectly passed the content 14801 length, instead of the encoding length to d2i_ASN1_OBJECT. 14802 14803 *Steve Henson* 14804 14805### Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 14806 14807 * [In 0.9.6g-engine release:] 14808 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`). 14809 14810 *Lynn Gazis <lgazis@rainbow.com>* 14811 14812### Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 14813 14814 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 14815 and get fix the header length calculation. 14816 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 14817 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson* 14818 14819 * Use proper error handling instead of 'assertions' in buffer 14820 overflow checks added in 0.9.6e. This prevents DoS (the 14821 assertions could call abort()). 14822 14823 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller* 14824 14825### Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 14826 14827 * Add various sanity checks to asn1_get_length() to reject 14828 the ASN1 length bytes if they exceed sizeof(long), will appear 14829 negative or the content length exceeds the length of the 14830 supplied buffer. 14831 14832 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>* 14833 14834 * Fix cipher selection routines: ciphers without encryption had no flags 14835 for the cipher strength set and where therefore not handled correctly 14836 by the selection routines (PR #130). 14837 14838 *Lutz Jaenicke* 14839 14840 * Fix EVP_dsa_sha macro. 14841 14842 *Nils Larsch* 14843 14844 * New option 14845 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 14846 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 14847 that was added in OpenSSL 0.9.6d. 14848 14849 As the countermeasure turned out to be incompatible with some 14850 broken SSL implementations, the new option is part of SSL_OP_ALL. 14851 SSL_OP_ALL is usually employed when compatibility with weird SSL 14852 implementations is desired (e.g. '-bugs' option to 's_client' and 14853 's_server'), so the new option is automatically set in many 14854 applications. 14855 14856 *Bodo Moeller* 14857 14858 * Changes in security patch: 14859 14860 Changes marked "(CHATS)" were sponsored by the Defense Advanced 14861 Research Projects Agency (DARPA) and Air Force Research Laboratory, 14862 Air Force Materiel Command, USAF, under agreement number 14863 F30602-01-2-0537. 14864 14865 * Add various sanity checks to asn1_get_length() to reject 14866 the ASN1 length bytes if they exceed sizeof(long), will appear 14867 negative or the content length exceeds the length of the 14868 supplied buffer. ([CVE-2002-0659]) 14869 14870 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>* 14871 14872 * Assertions for various potential buffer overflows, not known to 14873 happen in practice. 14874 14875 *Ben Laurie (CHATS)* 14876 14877 * Various temporary buffers to hold ASCII versions of integers were 14878 too small for 64 bit platforms. ([CVE-2002-0655]) 14879 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>* 14880 14881 * Remote buffer overflow in SSL3 protocol - an attacker could 14882 supply an oversized session ID to a client. ([CVE-2002-0656]) 14883 14884 *Ben Laurie (CHATS)* 14885 14886 * Remote buffer overflow in SSL2 protocol - an attacker could 14887 supply an oversized client master key. ([CVE-2002-0656]) 14888 14889 *Ben Laurie (CHATS)* 14890 14891### Changes between 0.9.6c and 0.9.6d [9 May 2002] 14892 14893 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 14894 encoded as NULL) with id-dsa-with-sha1. 14895 14896 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller* 14897 14898 * Check various `X509_...()` return values in `apps/req.c`. 14899 14900 *Nils Larsch <nla@trustcenter.de>* 14901 14902 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 14903 an end-of-file condition would erroneously be flagged, when the CRLF 14904 was just at the end of a processed block. The bug was discovered when 14905 processing data through a buffering memory BIO handing the data to a 14906 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 14907 <ptsekov@syntrex.com> and Nedelcho Stanev. 14908 14909 *Lutz Jaenicke* 14910 14911 * Implement a countermeasure against a vulnerability recently found 14912 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 14913 before application data chunks to avoid the use of known IVs 14914 with data potentially chosen by the attacker. 14915 14916 *Bodo Moeller* 14917 14918 * Fix length checks in ssl3_get_client_hello(). 14919 14920 *Bodo Moeller* 14921 14922 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently 14923 to prevent ssl3_read_internal() from incorrectly assuming that 14924 ssl3_read_bytes() found application data while handshake 14925 processing was enabled when in fact s->s3->in_read_app_data was 14926 merely automatically cleared during the initial handshake. 14927 14928 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>* 14929 14930 * Fix object definitions for Private and Enterprise: they were not 14931 recognized in their shortname (=lowercase) representation. Extend 14932 obj_dat.pl to issue an error when using undefined keywords instead 14933 of silently ignoring the problem (Svenning Sorensen 14934 <sss@sss.dnsalias.net>). 14935 14936 *Lutz Jaenicke* 14937 14938 * Fix DH_generate_parameters() so that it works for 'non-standard' 14939 generators, i.e. generators other than 2 and 5. (Previously, the 14940 code did not properly initialise the 'add' and 'rem' values to 14941 BN_generate_prime().) 14942 14943 In the new general case, we do not insist that 'generator' is 14944 actually a primitive root: This requirement is rather pointless; 14945 a generator of the order-q subgroup is just as good, if not 14946 better. 14947 14948 *Bodo Moeller* 14949 14950 * Map new X509 verification errors to alerts. Discovered and submitted by 14951 Tom Wu <tom@arcot.com>. 14952 14953 *Lutz Jaenicke* 14954 14955 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 14956 returning non-zero before the data has been completely received 14957 when using non-blocking I/O. 14958 14959 *Bodo Moeller; problem pointed out by John Hughes* 14960 14961 * Some of the ciphers missed the strength entry (SSL_LOW etc). 14962 14963 *Ben Laurie, Lutz Jaenicke* 14964 14965 * Fix bug in SSL_clear(): bad sessions were not removed (found by 14966 Yoram Zahavi <YoramZ@gilian.com>). 14967 14968 *Lutz Jaenicke* 14969 14970 * Add information about CygWin 1.3 and on, and preserve proper 14971 configuration for the versions before that. 14972 14973 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte* 14974 14975 * Make removal from session cache (SSL_CTX_remove_session()) more robust: 14976 check whether we deal with a copy of a session and do not delete from 14977 the cache in this case. Problem reported by "Izhar Shoshani Levi" 14978 <izhar@checkpoint.com>. 14979 14980 *Lutz Jaenicke* 14981 14982 * Do not store session data into the internal session cache, if it 14983 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 14984 flag is set). Proposed by Aslam <aslam@funk.com>. 14985 14986 *Lutz Jaenicke* 14987 14988 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 14989 value is 0. 14990 14991 *Richard Levitte* 14992 14993 * [In 0.9.6d-engine release:] 14994 Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 14995 14996 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte* 14997 14998 * Add the configuration target linux-s390x. 14999 15000 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte* 15001 15002 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 15003 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 15004 variable as an indication that a ClientHello message has been 15005 received. As the flag value will be lost between multiple 15006 invocations of ssl3_accept when using non-blocking I/O, the 15007 function may not be aware that a handshake has actually taken 15008 place, thus preventing a new session from being added to the 15009 session cache. 15010 15011 To avoid this problem, we now set s->new_session to 2 instead of 15012 using a local variable. 15013 15014 *Lutz Jaenicke, Bodo Moeller* 15015 15016 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 15017 if the SSL_R_LENGTH_MISMATCH error is detected. 15018 15019 *Geoff Thorpe, Bodo Moeller* 15020 15021 * New 'shared_ldflag' column in Configure platform table. 15022 15023 *Richard Levitte* 15024 15025 * Fix EVP_CIPHER_mode macro. 15026 15027 *"Dan S. Camper" <dan@bti.net>* 15028 15029 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 15030 type, we must throw them away by setting rr->length to 0. 15031 15032 *D P Chang <dpc@qualys.com>* 15033 15034### Changes between 0.9.6b and 0.9.6c [21 dec 2001] 15035 15036 * Fix BN_rand_range bug pointed out by Dominikus Scherkl 15037 <Dominikus.Scherkl@biodata.com>. (The previous implementation 15038 worked incorrectly for those cases where range = `10..._2` and 15039 `3*range` is two bits longer than range.) 15040 15041 *Bodo Moeller* 15042 15043 * Only add signing time to PKCS7 structures if it is not already 15044 present. 15045 15046 *Steve Henson* 15047 15048 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 15049 OBJ_ld_ce should be OBJ_id_ce. 15050 Also some ip-pda OIDs in crypto/objects/objects.txt were 15051 incorrect (cf. RFC 3039). 15052 15053 *Matt Cooper, Frederic Giudicelli, Bodo Moeller* 15054 15055 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 15056 returns early because it has nothing to do. 15057 15058 *Andy Schneider <andy.schneider@bjss.co.uk>* 15059 15060 * [In 0.9.6c-engine release:] 15061 Fix mutex callback return values in crypto/engine/hw_ncipher.c. 15062 15063 *Andy Schneider <andy.schneider@bjss.co.uk>* 15064 15065 * [In 0.9.6c-engine release:] 15066 Add support for Cryptographic Appliance's keyserver technology. 15067 (Use engine 'keyclient') 15068 15069 *Cryptographic Appliances and Geoff Thorpe* 15070 15071 * Add a configuration entry for OS/390 Unix. The C compiler 'c89' 15072 is called via tools/c89.sh because arguments have to be 15073 rearranged (all '-L' options must appear before the first object 15074 modules). 15075 15076 *Richard Shapiro <rshapiro@abinitio.com>* 15077 15078 * [In 0.9.6c-engine release:] 15079 Add support for Broadcom crypto accelerator cards, backported 15080 from 0.9.7. 15081 15082 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox* 15083 15084 * [In 0.9.6c-engine release:] 15085 Add support for SureWare crypto accelerator cards from 15086 Baltimore Technologies. (Use engine 'sureware') 15087 15088 *Baltimore Technologies and Mark Cox* 15089 15090 * [In 0.9.6c-engine release:] 15091 Add support for crypto accelerator cards from Accelerated 15092 Encryption Processing, www.aep.ie. (Use engine 'aep') 15093 15094 *AEP Inc. and Mark Cox* 15095 15096 * Add a configuration entry for gcc on UnixWare. 15097 15098 *Gary Benson <gbenson@redhat.com>* 15099 15100 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 15101 messages are stored in a single piece (fixed-length part and 15102 variable-length part combined) and fix various bugs found on the way. 15103 15104 *Bodo Moeller* 15105 15106 * Disable caching in BIO_gethostbyname(), directly use gethostbyname() 15107 instead. BIO_gethostbyname() does not know what timeouts are 15108 appropriate, so entries would stay in cache even when they have 15109 become invalid. 15110 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>* 15111 15112 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 15113 faced with a pathologically small ClientHello fragment that does 15114 not contain client_version: Instead of aborting with an error, 15115 simply choose the highest available protocol version (i.e., 15116 TLS 1.0 unless it is disabled). In practice, ClientHello 15117 messages are never sent like this, but this change gives us 15118 strictly correct behaviour at least for TLS. 15119 15120 *Bodo Moeller* 15121 15122 * Fix SSL handshake functions and SSL_clear() such that SSL_clear() 15123 never resets s->method to s->ctx->method when called from within 15124 one of the SSL handshake functions. 15125 15126 *Bodo Moeller; problem pointed out by Niko Baric* 15127 15128 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 15129 (sent using the client's version number) if client_version is 15130 smaller than the protocol version in use. Also change 15131 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 15132 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 15133 the client will at least see that alert. 15134 15135 *Bodo Moeller* 15136 15137 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 15138 correctly. 15139 15140 *Bodo Moeller* 15141 15142 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 15143 client receives HelloRequest while in a handshake. 15144 15145 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>* 15146 15147 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 15148 should end in 'break', not 'goto end' which circumvents various 15149 cleanups done in state SSL_ST_OK. But session related stuff 15150 must be disabled for SSL_ST_OK in the case that we just sent a 15151 HelloRequest. 15152 15153 Also avoid some overhead by not calling ssl_init_wbio_buffer() 15154 before just sending a HelloRequest. 15155 15156 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>* 15157 15158 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 15159 reveal whether illegal block cipher padding was found or a MAC 15160 verification error occurred. (Neither SSLerr() codes nor alerts 15161 are directly visible to potential attackers, but the information 15162 may leak via logfiles.) 15163 15164 Similar changes are not required for the SSL 2.0 implementation 15165 because the number of padding bytes is sent in clear for SSL 2.0, 15166 and the extra bytes are just ignored. However ssl/s2_pkt.c 15167 failed to verify that the purported number of padding bytes is in 15168 the legal range. 15169 15170 *Bodo Moeller* 15171 15172 * Add OpenUNIX-8 support including shared libraries 15173 (Boyd Lynn Gerber <gerberb@zenez.com>). 15174 15175 *Lutz Jaenicke* 15176 15177 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 15178 'wristwatch attack' using huge encoding parameters (cf. 15179 James H. Manger's CRYPTO 2001 paper). Note that the 15180 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 15181 encoding parameters and hence was not vulnerable. 15182 15183 *Bodo Moeller* 15184 15185 * BN_sqr() bug fix. 15186 15187 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>* 15188 15189 * Rabin-Miller test analyses assume uniformly distributed witnesses, 15190 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 15191 followed by modular reduction. 15192 15193 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>* 15194 15195 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 15196 equivalent based on BN_pseudo_rand() instead of BN_rand(). 15197 15198 *Bodo Moeller* 15199 15200 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 15201 This function was broken, as the check for a new client hello message 15202 to handle SGC did not allow these large messages. 15203 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 15204 15205 *Lutz Jaenicke* 15206 15207 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`. 15208 15209 *Lutz Jaenicke* 15210 15211 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 15212 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 15213 15214 *Lutz Jaenicke* 15215 15216 * Rework the configuration and shared library support for Tru64 Unix. 15217 The configuration part makes use of modern compiler features and 15218 still retains old compiler behavior for those that run older versions 15219 of the OS. The shared library support part includes a variant that 15220 uses the RPATH feature, and is available through the special 15221 configuration target "alpha-cc-rpath", which will never be selected 15222 automatically. 15223 15224 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte* 15225 15226 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 15227 with the same message size as in ssl3_get_certificate_request(). 15228 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 15229 messages might inadvertently be reject as too long. 15230 15231 *Petr Lampa <lampa@fee.vutbr.cz>* 15232 15233 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 15234 15235 *Andy Polyakov* 15236 15237 * Modified SSL library such that the verify_callback that has been set 15238 specifically for an SSL object with SSL_set_verify() is actually being 15239 used. Before the change, a verify_callback set with this function was 15240 ignored and the verify_callback() set in the SSL_CTX at the time of 15241 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 15242 to allow the necessary settings. 15243 15244 *Lutz Jaenicke* 15245 15246 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 15247 explicitly to NULL, as at least on Solaris 8 this seems not always to be 15248 done automatically (in contradiction to the requirements of the C 15249 standard). This made problems when used from OpenSSH. 15250 15251 *Lutz Jaenicke* 15252 15253 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 15254 dh->length and always used 15255 15256 BN_rand_range(priv_key, dh->p). 15257 15258 BN_rand_range() is not necessary for Diffie-Hellman, and this 15259 specific range makes Diffie-Hellman unnecessarily inefficient if 15260 dh->length (recommended exponent length) is much smaller than the 15261 length of dh->p. We could use BN_rand_range() if the order of 15262 the subgroup was stored in the DH structure, but we only have 15263 dh->length. 15264 15265 So switch back to 15266 15267 BN_rand(priv_key, l, ...) 15268 15269 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 15270 otherwise. 15271 15272 *Bodo Moeller* 15273 15274 * In 15275 15276 RSA_eay_public_encrypt 15277 RSA_eay_private_decrypt 15278 RSA_eay_private_encrypt (signing) 15279 RSA_eay_public_decrypt (signature verification) 15280 15281 (default implementations for RSA_public_encrypt, 15282 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 15283 always reject numbers >= n. 15284 15285 *Bodo Moeller* 15286 15287 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 15288 to synchronize access to 'locking_thread'. This is necessary on 15289 systems where access to 'locking_thread' (an 'unsigned long' 15290 variable) is not atomic. 15291 15292 *Bodo Moeller* 15293 15294 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 15295 *before* setting the 'crypto_lock_rand' flag. The previous code had 15296 a race condition if 0 is a valid thread ID. 15297 15298 *Travis Vitek <vitek@roguewave.com>* 15299 15300 * Add support for shared libraries under Irix. 15301 15302 *Albert Chin-A-Young <china@thewrittenword.com>* 15303 15304 * Add configuration option to build on Linux on both big-endian and 15305 little-endian MIPS. 15306 15307 *Ralf Baechle <ralf@uni-koblenz.de>* 15308 15309 * Add the possibility to create shared libraries on HP-UX. 15310 15311 *Richard Levitte* 15312 15313### Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 15314 15315 * Change ssleay_rand_bytes (crypto/rand/md_rand.c) 15316 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 15317 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 15318 PRNG state recovery was possible based on the output of 15319 one PRNG request appropriately sized to gain knowledge on 15320 'md' followed by enough consecutive 1-byte PRNG requests 15321 to traverse all of 'state'. 15322 15323 1. When updating 'md_local' (the current thread's copy of 'md') 15324 during PRNG output generation, hash all of the previous 15325 'md_local' value, not just the half used for PRNG output. 15326 15327 2. Make the number of bytes from 'state' included into the hash 15328 independent from the number of PRNG bytes requested. 15329 15330 The first measure alone would be sufficient to avoid 15331 Markku-Juhani's attack. (Actually it had never occurred 15332 to me that the half of 'md_local' used for chaining was the 15333 half from which PRNG output bytes were taken -- I had always 15334 assumed that the secret half would be used.) The second 15335 measure makes sure that additional data from 'state' is never 15336 mixed into 'md_local' in small portions; this heuristically 15337 further strengthens the PRNG. 15338 15339 *Bodo Moeller* 15340 15341 * Fix crypto/bn/asm/mips3.s. 15342 15343 *Andy Polyakov* 15344 15345 * When only the key is given to "enc", the IV is undefined. Print out 15346 an error message in this case. 15347 15348 *Lutz Jaenicke* 15349 15350 * Handle special case when X509_NAME is empty in X509 printing routines. 15351 15352 *Steve Henson* 15353 15354 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 15355 positive and less than q. 15356 15357 *Bodo Moeller* 15358 15359 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is 15360 used: it isn't thread safe and the add_lock_callback should handle 15361 that itself. 15362 15363 *Paul Rose <Paul.Rose@bridge.com>* 15364 15365 * Verify that incoming data obeys the block size in 15366 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 15367 15368 *Bodo Moeller* 15369 15370 * Fix OAEP check. 15371 15372 *Ulf Möller, Bodo Möller* 15373 15374 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 15375 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 15376 when fixing the server behaviour for backwards-compatible 'client 15377 hello' messages. (Note that the attack is impractical against 15378 SSL 3.0 and TLS 1.0 anyway because length and version checking 15379 means that the probability of guessing a valid ciphertext is 15380 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 15381 paper.) 15382 15383 Before 0.9.5, the countermeasure (hide the error by generating a 15384 random 'decryption result') did not work properly because 15385 ERR_clear_error() was missing, meaning that SSL_get_error() would 15386 detect the supposedly ignored error. 15387 15388 Both problems are now fixed. 15389 15390 *Bodo Moeller* 15391 15392 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 15393 (previously it was 1024). 15394 15395 *Bodo Moeller* 15396 15397 * Fix for compatibility mode trust settings: ignore trust settings 15398 unless some valid trust or reject settings are present. 15399 15400 *Steve Henson* 15401 15402 * Fix for blowfish EVP: its a variable length cipher. 15403 15404 *Steve Henson* 15405 15406 * Fix various bugs related to DSA S/MIME verification. Handle missing 15407 parameters in DSA public key structures and return an error in the 15408 DSA routines if parameters are absent. 15409 15410 *Steve Henson* 15411 15412 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 15413 in the current directory if neither $RANDFILE nor $HOME was set. 15414 RAND_file_name() in 0.9.6a returned NULL in this case. This has 15415 caused some confusion to Windows users who haven't defined $HOME. 15416 Thus RAND_file_name() is changed again: e_os.h can define a 15417 DEFAULT_HOME, which will be used if $HOME is not set. 15418 For Windows, we use "C:"; on other platforms, we still require 15419 environment variables. 15420 15421 * Move 'if (!initialized) RAND_poll()' into regions protected by 15422 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 15423 having multiple threads call RAND_poll() concurrently. 15424 15425 *Bodo Moeller* 15426 15427 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 15428 combination of a flag and a thread ID variable. 15429 Otherwise while one thread is in ssleay_rand_bytes (which sets the 15430 flag), *other* threads can enter ssleay_add_bytes without obeying 15431 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 15432 that they do not hold after the first thread unsets add_do_not_lock). 15433 15434 *Bodo Moeller* 15435 15436 * Change bctest again: '-x' expressions are not available in all 15437 versions of 'test'. 15438 15439 *Bodo Moeller* 15440 15441### Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 15442 15443 * Fix a couple of memory leaks in PKCS7_dataDecode() 15444 15445 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>* 15446 15447 * Change Configure and Makefiles to provide EXE_EXT, which will contain 15448 the default extension for executables, if any. Also, make the perl 15449 scripts that use symlink() to test if it really exists and use "cp" 15450 if it doesn't. All this made OpenSSL compilable and installable in 15451 CygWin. 15452 15453 *Richard Levitte* 15454 15455 * Fix for asn1_GetSequence() for indefinite length constructed data. 15456 If SEQUENCE is length is indefinite just set c->slen to the total 15457 amount of data available. 15458 15459 *Steve Henson, reported by shige@FreeBSD.org* 15460 15461 *This change does not apply to 0.9.7.* 15462 15463 * Change bctest to avoid here-documents inside command substitution 15464 (workaround for FreeBSD /bin/sh bug). 15465 For compatibility with Ultrix, avoid shell functions (introduced 15466 in the bctest version that searches along $PATH). 15467 15468 *Bodo Moeller* 15469 15470 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 15471 with des_encrypt() defined on some operating systems, like Solaris 15472 and UnixWare. 15473 15474 *Richard Levitte* 15475 15476 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 15477 On the Importance of Eliminating Errors in Cryptographic 15478 Computations, J. Cryptology 14 (2001) 2, 101-119, 15479 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>). 15480 15481 *Ulf Moeller* 15482 15483 * MIPS assembler BIGNUM division bug fix. 15484 15485 *Andy Polyakov* 15486 15487 * Disabled incorrect Alpha assembler code. 15488 15489 *Richard Levitte* 15490 15491 * Fix PKCS#7 decode routines so they correctly update the length 15492 after reading an EOC for the EXPLICIT tag. 15493 15494 *Steve Henson* 15495 15496 *This change does not apply to 0.9.7.* 15497 15498 * Fix bug in PKCS#12 key generation routines. This was triggered 15499 if a 3DES key was generated with a 0 initial byte. Include 15500 PKCS12_BROKEN_KEYGEN compilation option to retain the old 15501 (but broken) behaviour. 15502 15503 *Steve Henson* 15504 15505 * Enhance bctest to search for a working bc along $PATH and print 15506 it when found. 15507 15508 *Tim Rice <tim@multitalents.net> via Richard Levitte* 15509 15510 * Fix memory leaks in err.c: free err_data string if necessary; 15511 don't write to the wrong index in ERR_set_error_data. 15512 15513 *Bodo Moeller* 15514 15515 * Implement ssl23_peek (analogous to ssl23_read), which previously 15516 did not exist. 15517 15518 *Bodo Moeller* 15519 15520 * Replace rdtsc with `_emit` statements for VC++ version 5. 15521 15522 *Jeremy Cooper <jeremy@baymoo.org>* 15523 15524 * Make it possible to reuse SSLv2 sessions. 15525 15526 *Richard Levitte* 15527 15528 * In copy_email() check for >= 0 as a return value for 15529 X509_NAME_get_index_by_NID() since 0 is a valid index. 15530 15531 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>* 15532 15533 * Avoid coredump with unsupported or invalid public keys by checking if 15534 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 15535 PKCS7_verify() fails with non detached data. 15536 15537 *Steve Henson* 15538 15539 * Don't use getenv in library functions when run as setuid/setgid. 15540 New function OPENSSL_issetugid(). 15541 15542 *Ulf Moeller* 15543 15544 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 15545 due to incorrect handling of multi-threading: 15546 15547 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 15548 15549 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 15550 15551 3. Count how many times MemCheck_off() has been called so that 15552 nested use can be treated correctly. This also avoids 15553 inband-signalling in the previous code (which relied on the 15554 assumption that thread ID 0 is impossible). 15555 15556 *Bodo Moeller* 15557 15558 * Add "-rand" option also to s_client and s_server. 15559 15560 *Lutz Jaenicke* 15561 15562 * Fix CPU detection on Irix 6.x. 15563 *Kurt Hockenbury <khockenb@stevens-tech.edu> and 15564 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>* 15565 15566 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME 15567 was empty. 15568 15569 *Steve Henson* 15570 15571 *This change does not apply to 0.9.7.* 15572 15573 * Use the cached encoding of an X509_NAME structure rather than 15574 copying it. This is apparently the reason for the libsafe "errors" 15575 but the code is actually correct. 15576 15577 *Steve Henson* 15578 15579 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 15580 Bleichenbacher's DSA attack. 15581 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 15582 to be set and top=0 forces the highest bit to be set; top=-1 is new 15583 and leaves the highest bit random. 15584 15585 *Ulf Moeller, Bodo Moeller* 15586 15587 * In the `NCONF_...`-based implementations for `CONF_...` queries 15588 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 15589 a temporary CONF structure with the data component set to NULL 15590 (which gives segmentation faults in lh_retrieve). 15591 Instead, use NULL for the CONF pointer in CONF_get_string and 15592 CONF_get_number (which may use environment variables) and directly 15593 return NULL from CONF_get_section. 15594 15595 *Bodo Moeller* 15596 15597 * Fix potential buffer overrun for EBCDIC. 15598 15599 *Ulf Moeller* 15600 15601 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign 15602 keyUsage if basicConstraints absent for a CA. 15603 15604 *Steve Henson* 15605 15606 * Make SMIME_write_PKCS7() write mail header values with a format that 15607 is more generally accepted (no spaces before the semicolon), since 15608 some programs can't parse those values properly otherwise. Also make 15609 sure BIO's that break lines after each write do not create invalid 15610 headers. 15611 15612 *Richard Levitte* 15613 15614 * Make the CRL encoding routines work with empty SEQUENCE OF. The 15615 macros previously used would not encode an empty SEQUENCE OF 15616 and break the signature. 15617 15618 *Steve Henson* 15619 15620 *This change does not apply to 0.9.7.* 15621 15622 * Zero the premaster secret after deriving the master secret in 15623 DH ciphersuites. 15624 15625 *Steve Henson* 15626 15627 * Add some EVP_add_digest_alias registrations (as found in 15628 OpenSSL_add_all_digests()) to SSL_library_init() 15629 aka OpenSSL_add_ssl_algorithms(). This provides improved 15630 compatibility with peers using X.509 certificates 15631 with unconventional AlgorithmIdentifier OIDs. 15632 15633 *Bodo Moeller* 15634 15635 * Fix for Irix with NO_ASM. 15636 15637 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>* 15638 15639 * ./config script fixes. 15640 15641 *Ulf Moeller, Richard Levitte* 15642 15643 * Fix 'openssl passwd -1'. 15644 15645 *Bodo Moeller* 15646 15647 * Change PKCS12_key_gen_asc() so it can cope with non null 15648 terminated strings whose length is passed in the passlen 15649 parameter, for example from PEM callbacks. This was done 15650 by adding an extra length parameter to asc2uni(). 15651 15652 *Steve Henson, reported by <oddissey@samsung.co.kr>* 15653 15654 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 15655 call failed, free the DSA structure. 15656 15657 *Bodo Moeller* 15658 15659 * Fix to uni2asc() to cope with zero length Unicode strings. 15660 These are present in some PKCS#12 files. 15661 15662 *Steve Henson* 15663 15664 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 15665 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 15666 when writing a 32767 byte record. 15667 15668 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>* 15669 15670 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c), 15671 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`. 15672 15673 (RSA objects have a reference count access to which is protected 15674 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 15675 so they are meant to be shared between threads.) 15676 *Bodo Moeller, Geoff Thorpe; original patch submitted by 15677 "Reddie, Steven" <Steven.Reddie@ca.com>* 15678 15679 * Fix a deadlock in CRYPTO_mem_leaks(). 15680 15681 *Bodo Moeller* 15682 15683 * Use better test patterns in bntest. 15684 15685 *Ulf Möller* 15686 15687 * rand_win.c fix for Borland C. 15688 15689 *Ulf Möller* 15690 15691 * BN_rshift bugfix for n == 0. 15692 15693 *Bodo Moeller* 15694 15695 * Add a 'bctest' script that checks for some known 'bc' bugs 15696 so that 'make test' does not abort just because 'bc' is broken. 15697 15698 *Bodo Moeller* 15699 15700 * Store verify_result within SSL_SESSION also for client side to 15701 avoid potential security hole. (Re-used sessions on the client side 15702 always resulted in verify_result==X509_V_OK, not using the original 15703 result of the server certificate verification.) 15704 15705 *Lutz Jaenicke* 15706 15707 * Fix ssl3_pending: If the record in s->s3->rrec is not of type 15708 SSL3_RT_APPLICATION_DATA, return 0. 15709 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 15710 15711 *Bodo Moeller* 15712 15713 * Fix SSL_peek: 15714 Both ssl2_peek and ssl3_peek, which were totally broken in earlier 15715 releases, have been re-implemented by renaming the previous 15716 implementations of ssl2_read and ssl3_read to ssl2_read_internal 15717 and ssl3_read_internal, respectively, and adding 'peek' parameters 15718 to them. The new ssl[23]_{read,peek} functions are calls to 15719 ssl[23]_read_internal with the 'peek' flag set appropriately. 15720 A 'peek' parameter has also been added to ssl3_read_bytes, which 15721 does the actual work for ssl3_read_internal. 15722 15723 *Bodo Moeller* 15724 15725 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 15726 the method-specific "init()" handler. Also clean up ex_data after 15727 calling the method-specific "finish()" handler. Previously, this was 15728 happening the other way round. 15729 15730 *Geoff Thorpe* 15731 15732 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 15733 The previous value, 12, was not always sufficient for BN_mod_exp(). 15734 15735 *Bodo Moeller* 15736 15737 * Make sure that shared libraries get the internal name engine with 15738 the full version number and not just 0. This should mark the 15739 shared libraries as not backward compatible. Of course, this should 15740 be changed again when we can guarantee backward binary compatibility. 15741 15742 *Richard Levitte* 15743 15744 * Fix typo in get_cert_by_subject() in by_dir.c 15745 15746 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>* 15747 15748 * Rework the system to generate shared libraries: 15749 15750 - Make note of the expected extension for the shared libraries and 15751 if there is a need for symbolic links from for example libcrypto.so.0 15752 to libcrypto.so.0.9.7. There is extended info in Configure for 15753 that. 15754 15755 - Make as few rebuilds of the shared libraries as possible. 15756 15757 - Still avoid linking the OpenSSL programs with the shared libraries. 15758 15759 - When installing, install the shared libraries separately from the 15760 static ones. 15761 15762 *Richard Levitte* 15763 15764 * Fix SSL_CTX_set_read_ahead macro to actually use its argument. 15765 15766 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 15767 and not in SSL_clear because the latter is also used by the 15768 accept/connect functions; previously, the settings made by 15769 SSL_set_read_ahead would be lost during the handshake. 15770 15771 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>* 15772 15773 * Correct util/mkdef.pl to be selective about disabled algorithms. 15774 Previously, it would create entries for disabled algorithms no 15775 matter what. 15776 15777 *Richard Levitte* 15778 15779 * Added several new manual pages for SSL_* function. 15780 15781 *Lutz Jaenicke* 15782 15783### Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 15784 15785 * In ssl23_get_client_hello, generate an error message when faced 15786 with an initial SSL 3.0/TLS record that is too small to contain the 15787 first two bytes of the ClientHello message, i.e. client_version. 15788 (Note that this is a pathologic case that probably has never happened 15789 in real life.) The previous approach was to use the version number 15790 from the record header as a substitute; but our protocol choice 15791 should not depend on that one because it is not authenticated 15792 by the Finished messages. 15793 15794 *Bodo Moeller* 15795 15796 * More robust randomness gathering functions for Windows. 15797 15798 *Jeffrey Altman <jaltman@columbia.edu>* 15799 15800 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 15801 not set then we don't setup the error code for issuer check errors 15802 to avoid possibly overwriting other errors which the callback does 15803 handle. If an application does set the flag then we assume it knows 15804 what it is doing and can handle the new informational codes 15805 appropriately. 15806 15807 *Steve Henson* 15808 15809 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 15810 a general "ANY" type, as such it should be able to decode anything 15811 including tagged types. However it didn't check the class so it would 15812 wrongly interpret tagged types in the same way as their universal 15813 counterpart and unknown types were just rejected. Changed so that the 15814 tagged and unknown types are handled in the same way as a SEQUENCE: 15815 that is the encoding is stored intact. There is also a new type 15816 "V_ASN1_OTHER" which is used when the class is not universal, in this 15817 case we have no idea what the actual type is so we just lump them all 15818 together. 15819 15820 *Steve Henson* 15821 15822 * On VMS, stdout may very well lead to a file that is written to 15823 in a record-oriented fashion. That means that every write() will 15824 write a separate record, which will be read separately by the 15825 programs trying to read from it. This can be very confusing. 15826 15827 The solution is to put a BIO filter in the way that will buffer 15828 text until a linefeed is reached, and then write everything a 15829 line at a time, so every record written will be an actual line, 15830 not chunks of lines and not (usually doesn't happen, but I've 15831 seen it once) several lines in one record. BIO_f_linebuffer() is 15832 the answer. 15833 15834 Currently, it's a VMS-only method, because that's where it has 15835 been tested well enough. 15836 15837 *Richard Levitte* 15838 15839 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 15840 it can return incorrect results. 15841 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 15842 but it was in 0.9.6-beta[12].) 15843 15844 *Bodo Moeller* 15845 15846 * Disable the check for content being present when verifying detached 15847 signatures in pk7_smime.c. Some versions of Netscape (wrongly) 15848 include zero length content when signing messages. 15849 15850 *Steve Henson* 15851 15852 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 15853 BIO_ctrl (for BIO pairs). 15854 15855 *Bodo Möller* 15856 15857 * Add DSO method for VMS. 15858 15859 *Richard Levitte* 15860 15861 * Bug fix: Montgomery multiplication could produce results with the 15862 wrong sign. 15863 15864 *Ulf Möller* 15865 15866 * Add RPM specification openssl.spec and modify it to build three 15867 packages. The default package contains applications, application 15868 documentation and run-time libraries. The devel package contains 15869 include files, static libraries and function documentation. The 15870 doc package contains the contents of the doc directory. The original 15871 openssl.spec was provided by Damien Miller <djm@mindrot.org>. 15872 15873 *Richard Levitte* 15874 15875 * Add a large number of documentation files for many SSL routines. 15876 15877 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>* 15878 15879 * Add a configuration entry for Sony News 4. 15880 15881 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>* 15882 15883 * Don't set the two most significant bits to one when generating a 15884 random number < q in the DSA library. 15885 15886 *Ulf Möller* 15887 15888 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 15889 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 15890 the underlying transport is blocking) if a handshake took place. 15891 (The default behaviour is needed by applications such as s_client 15892 and s_server that use select() to determine when to use SSL_read; 15893 but for applications that know in advance when to expect data, it 15894 just makes things more complicated.) 15895 15896 *Bodo Moeller* 15897 15898 * Add RAND_egd_bytes(), which gives control over the number of bytes read 15899 from EGD. 15900 15901 *Ben Laurie* 15902 15903 * Add a few more EBCDIC conditionals that make `req` and `x509` 15904 work better on such systems. 15905 15906 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>* 15907 15908 * Add two demo programs for PKCS12_parse() and PKCS12_create(). 15909 Update PKCS12_parse() so it copies the friendlyName and the 15910 keyid to the certificates aux info. 15911 15912 *Steve Henson* 15913 15914 * Fix bug in PKCS7_verify() which caused an infinite loop 15915 if there was more than one signature. 15916 15917 *Sven Uszpelkat <su@celocom.de>* 15918 15919 * Major change in util/mkdef.pl to include extra information 15920 about each symbol, as well as presenting variables as well 15921 as functions. This change means that there's n more need 15922 to rebuild the .num files when some algorithms are excluded. 15923 15924 *Richard Levitte* 15925 15926 * Allow the verify time to be set by an application, 15927 rather than always using the current time. 15928 15929 *Steve Henson* 15930 15931 * Phase 2 verify code reorganisation. The certificate 15932 verify code now looks up an issuer certificate by a 15933 number of criteria: subject name, authority key id 15934 and key usage. It also verifies self signed certificates 15935 by the same criteria. The main comparison function is 15936 X509_check_issued() which performs these checks. 15937 15938 Lot of changes were necessary in order to support this 15939 without completely rewriting the lookup code. 15940 15941 Authority and subject key identifier are now cached. 15942 15943 The LHASH 'certs' is X509_STORE has now been replaced 15944 by a STACK_OF(X509_OBJECT). This is mainly because an 15945 LHASH can't store or retrieve multiple objects with 15946 the same hash value. 15947 15948 As a result various functions (which were all internal 15949 use only) have changed to handle the new X509_STORE 15950 structure. This will break anything that messed round 15951 with X509_STORE internally. 15952 15953 The functions X509_STORE_add_cert() now checks for an 15954 exact match, rather than just subject name. 15955 15956 The X509_STORE API doesn't directly support the retrieval 15957 of multiple certificates matching a given criteria, however 15958 this can be worked round by performing a lookup first 15959 (which will fill the cache with candidate certificates) 15960 and then examining the cache for matches. This is probably 15961 the best we can do without throwing out X509_LOOKUP 15962 entirely (maybe later...). 15963 15964 The X509_VERIFY_CTX structure has been enhanced considerably. 15965 15966 All certificate lookup operations now go via a get_issuer() 15967 callback. Although this currently uses an X509_STORE it 15968 can be replaced by custom lookups. This is a simple way 15969 to bypass the X509_STORE hackery necessary to make this 15970 work and makes it possible to use more efficient techniques 15971 in future. A very simple version which uses a simple 15972 STACK for its trusted certificate store is also provided 15973 using X509_STORE_CTX_trusted_stack(). 15974 15975 The verify_cb() and verify() callbacks now have equivalents 15976 in the X509_STORE_CTX structure. 15977 15978 X509_STORE_CTX also has a 'flags' field which can be used 15979 to customise the verify behaviour. 15980 15981 *Steve Henson* 15982 15983 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 15984 excludes S/MIME capabilities. 15985 15986 *Steve Henson* 15987 15988 * When a certificate request is read in keep a copy of the 15989 original encoding of the signed data and use it when outputting 15990 again. Signatures then use the original encoding rather than 15991 a decoded, encoded version which may cause problems if the 15992 request is improperly encoded. 15993 15994 *Steve Henson* 15995 15996 * For consistency with other BIO_puts implementations, call 15997 buffer_write(b, ...) directly in buffer_puts instead of calling 15998 BIO_write(b, ...). 15999 16000 In BIO_puts, increment b->num_write as in BIO_write. 16001 16002 *Peter.Sylvester@EdelWeb.fr* 16003 16004 * Fix BN_mul_word for the case where the word is 0. (We have to use 16005 BN_zero, we may not return a BIGNUM with an array consisting of 16006 words set to zero.) 16007 16008 *Bodo Moeller* 16009 16010 * Avoid calling abort() from within the library when problems are 16011 detected, except if preprocessor symbols have been defined 16012 (such as REF_CHECK, BN_DEBUG etc.). 16013 16014 *Bodo Moeller* 16015 16016 * New openssl application 'rsautl'. This utility can be 16017 used for low-level RSA operations. DER public key 16018 BIO/fp routines also added. 16019 16020 *Steve Henson* 16021 16022 * New Configure entry and patches for compiling on QNX 4. 16023 16024 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>* 16025 16026 * A demo state-machine implementation was sponsored by 16027 Nuron (<http://www.nuron.com/>) and is now available in 16028 demos/state_machine. 16029 16030 *Ben Laurie* 16031 16032 * New options added to the 'dgst' utility for signature 16033 generation and verification. 16034 16035 *Steve Henson* 16036 16037 * Unrecognized PKCS#7 content types are now handled via a 16038 catch all ASN1_TYPE structure. This allows unsupported 16039 types to be stored as a "blob" and an application can 16040 encode and decode it manually. 16041 16042 *Steve Henson* 16043 16044 * Fix various signed/unsigned issues to make a_strex.c 16045 compile under VC++. 16046 16047 *Oscar Jacobsson <oscar.jacobsson@celocom.com>* 16048 16049 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 16050 length if passed a buffer. ASN1_INTEGER_to_BN failed 16051 if passed a NULL BN and its argument was negative. 16052 16053 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>* 16054 16055 * Modification to PKCS#7 encoding routines to output definite 16056 length encoding. Since currently the whole structures are in 16057 memory there's not real point in using indefinite length 16058 constructed encoding. However if OpenSSL is compiled with 16059 the flag PKCS7_INDEFINITE_ENCODING the old form is used. 16060 16061 *Steve Henson* 16062 16063 * Added BIO_vprintf() and BIO_vsnprintf(). 16064 16065 *Richard Levitte* 16066 16067 * Added more prefixes to parse for in the strings written 16068 through a logging bio, to cover all the levels that are available 16069 through syslog. The prefixes are now: 16070 16071 PANIC, EMERG, EMR => LOG_EMERG 16072 ALERT, ALR => LOG_ALERT 16073 CRIT, CRI => LOG_CRIT 16074 ERROR, ERR => LOG_ERR 16075 WARNING, WARN, WAR => LOG_WARNING 16076 NOTICE, NOTE, NOT => LOG_NOTICE 16077 INFO, INF => LOG_INFO 16078 DEBUG, DBG => LOG_DEBUG 16079 16080 and as before, if none of those prefixes are present at the 16081 beginning of the string, LOG_ERR is chosen. 16082 16083 On Win32, the `LOG_*` levels are mapped according to this: 16084 16085 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 16086 LOG_WARNING => EVENTLOG_WARNING_TYPE 16087 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 16088 16089 *Richard Levitte* 16090 16091 * Made it possible to reconfigure with just the configuration 16092 argument "reconf" or "reconfigure". The command line arguments 16093 are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 16094 and are retrieved from there when reconfiguring. 16095 16096 *Richard Levitte* 16097 16098 * MD4 implemented. 16099 16100 *Assar Westerlund <assar@sics.se>, Richard Levitte* 16101 16102 * Add the arguments -CAfile and -CApath to the pkcs12 utility. 16103 16104 *Richard Levitte* 16105 16106 * The obj_dat.pl script was messing up the sorting of object 16107 names. The reason was that it compared the quoted version 16108 of strings as a result "OCSP" > "OCSP Signing" because 16109 " > SPACE. Changed script to store unquoted versions of 16110 names and add quotes on output. It was also omitting some 16111 names from the lookup table if they were given a default 16112 value (that is if SN is missing it is given the same 16113 value as LN and vice versa), these are now added on the 16114 grounds that if an object has a name we should be able to 16115 look it up. Finally added warning output when duplicate 16116 short or long names are found. 16117 16118 *Steve Henson* 16119 16120 * Changes needed for Tandem NSK. 16121 16122 *Scott Uroff <scott@xypro.com>* 16123 16124 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in 16125 RSA_padding_check_SSLv23(), special padding was never detected 16126 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 16127 version rollback attacks was not effective. 16128 16129 In s23_clnt.c, don't use special rollback-attack detection padding 16130 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 16131 client; similarly, in s23_srvr.c, don't do the rollback check if 16132 SSL 2.0 is the only protocol enabled in the server. 16133 16134 *Bodo Moeller* 16135 16136 * Make it possible to get hexdumps of unprintable data with 'openssl 16137 asn1parse'. By implication, the functions ASN1_parse_dump() and 16138 BIO_dump_indent() are added. 16139 16140 *Richard Levitte* 16141 16142 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 16143 these print out strings and name structures based on various 16144 flags including RFC2253 support and proper handling of 16145 multibyte characters. Added options to the 'x509' utility 16146 to allow the various flags to be set. 16147 16148 *Steve Henson* 16149 16150 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 16151 Also change the functions X509_cmp_current_time() and 16152 X509_gmtime_adj() work with an ASN1_TIME structure, 16153 this will enable certificates using GeneralizedTime in validity 16154 dates to be checked. 16155 16156 *Steve Henson* 16157 16158 * Make the NEG_PUBKEY_BUG code (which tolerates invalid 16159 negative public key encodings) on by default, 16160 NO_NEG_PUBKEY_BUG can be set to disable it. 16161 16162 *Steve Henson* 16163 16164 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 16165 content octets. An i2c_ASN1_OBJECT is unnecessary because 16166 the encoding can be trivially obtained from the structure. 16167 16168 *Steve Henson* 16169 16170 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`), 16171 not read locks (`CRYPTO_r_[un]lock`). 16172 16173 *Bodo Moeller* 16174 16175 * A first attempt at creating official support for shared 16176 libraries through configuration. I've kept it so the 16177 default is static libraries only, and the OpenSSL programs 16178 are always statically linked for now, but there are 16179 preparations for dynamic linking in place. 16180 This has been tested on Linux and Tru64. 16181 16182 *Richard Levitte* 16183 16184 * Randomness polling function for Win9x, as described in: 16185 Peter Gutmann, Software Generation of Practically Strong 16186 Random Numbers. 16187 16188 *Ulf Möller* 16189 16190 * Fix so PRNG is seeded in req if using an already existing 16191 DSA key. 16192 16193 *Steve Henson* 16194 16195 * New options to smime application. -inform and -outform 16196 allow alternative formats for the S/MIME message including 16197 PEM and DER. The -content option allows the content to be 16198 specified separately. This should allow things like Netscape 16199 form signing output easier to verify. 16200 16201 *Steve Henson* 16202 16203 * Fix the ASN1 encoding of tags using the 'long form'. 16204 16205 *Steve Henson* 16206 16207 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT 16208 STRING types. These convert content octets to and from the 16209 underlying type. The actual tag and length octets are 16210 already assumed to have been read in and checked. These 16211 are needed because all other string types have virtually 16212 identical handling apart from the tag. By having versions 16213 of the ASN1 functions that just operate on content octets 16214 IMPLICIT tagging can be handled properly. It also allows 16215 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 16216 and ASN1_INTEGER are identical apart from the tag. 16217 16218 *Steve Henson* 16219 16220 * Change the handling of OID objects as follows: 16221 16222 - New object identifiers are inserted in objects.txt, following 16223 the syntax given in [crypto/objects/README.md](crypto/objects/README.md). 16224 - objects.pl is used to process obj_mac.num and create a new 16225 obj_mac.h. 16226 - obj_dat.pl is used to create a new obj_dat.h, using the data in 16227 obj_mac.h. 16228 16229 This is currently kind of a hack, and the perl code in objects.pl 16230 isn't very elegant, but it works as I intended. The simplest way 16231 to check that it worked correctly is to look in obj_dat.h and 16232 check the array nid_objs and make sure the objects haven't moved 16233 around (this is important!). Additions are OK, as well as 16234 consistent name changes. 16235 16236 *Richard Levitte* 16237 16238 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 16239 16240 *Bodo Moeller* 16241 16242 * Addition of the command line parameter '-rand file' to 'openssl req'. 16243 The given file adds to whatever has already been seeded into the 16244 random pool through the RANDFILE configuration file option or 16245 environment variable, or the default random state file. 16246 16247 *Richard Levitte* 16248 16249 * mkstack.pl now sorts each macro group into lexical order. 16250 Previously the output order depended on the order the files 16251 appeared in the directory, resulting in needless rewriting 16252 of safestack.h . 16253 16254 *Steve Henson* 16255 16256 * Patches to make OpenSSL compile under Win32 again. Mostly 16257 work arounds for the VC++ problem that it treats func() as 16258 func(void). Also stripped out the parts of mkdef.pl that 16259 added extra typesafe functions: these no longer exist. 16260 16261 *Steve Henson* 16262 16263 * Reorganisation of the stack code. The macros are now all 16264 collected in safestack.h . Each macro is defined in terms of 16265 a "stack macro" of the form `SKM_<name>(type, a, b)`. The 16266 DEBUG_SAFESTACK is now handled in terms of function casts, 16267 this has the advantage of retaining type safety without the 16268 use of additional functions. If DEBUG_SAFESTACK is not defined 16269 then the non typesafe macros are used instead. Also modified the 16270 mkstack.pl script to handle the new form. Needs testing to see 16271 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 16272 the default if no major problems. Similar behaviour for ASN1_SET_OF 16273 and PKCS12_STACK_OF. 16274 16275 *Steve Henson* 16276 16277 * When some versions of IIS use the 'NET' form of private key the 16278 key derivation algorithm is different. Normally MD5(password) is 16279 used as a 128 bit RC4 key. In the modified case 16280 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some 16281 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 16282 as the old Netscape_RSA functions except they have an additional 16283 'sgckey' parameter which uses the modified algorithm. Also added 16284 an -sgckey command line option to the rsa utility. Thanks to 16285 Adrian Peck <bertie@ncipher.com> for posting details of the modified 16286 algorithm to openssl-dev. 16287 16288 *Steve Henson* 16289 16290 * The evp_local.h macros were using 'c.##kname' which resulted in 16291 invalid expansion on some systems (SCO 5.0.5 for example). 16292 Corrected to 'c.kname'. 16293 16294 *Phillip Porch <root@theporch.com>* 16295 16296 * New X509_get1_email() and X509_REQ_get1_email() functions that return 16297 a STACK of email addresses from a certificate or request, these look 16298 in the subject name and the subject alternative name extensions and 16299 omit any duplicate addresses. 16300 16301 *Steve Henson* 16302 16303 * Re-implement BN_mod_exp2_mont using independent (and larger) windows. 16304 This makes DSA verification about 2 % faster. 16305 16306 *Bodo Moeller* 16307 16308 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5 16309 (meaning that now 2^5 values will be precomputed, which is only 4 KB 16310 plus overhead for 1024 bit moduli). 16311 This makes exponentiations about 0.5 % faster for 1024 bit 16312 exponents (as measured by "openssl speed rsa2048"). 16313 16314 *Bodo Moeller* 16315 16316 * Rename memory handling macros to avoid conflicts with other 16317 software: 16318 Malloc => OPENSSL_malloc 16319 Malloc_locked => OPENSSL_malloc_locked 16320 Realloc => OPENSSL_realloc 16321 Free => OPENSSL_free 16322 16323 *Richard Levitte* 16324 16325 * New function BN_mod_exp_mont_word for small bases (roughly 15% 16326 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 16327 16328 *Bodo Moeller* 16329 16330 * CygWin32 support. 16331 16332 *John Jarvie <jjarvie@newsguy.com>* 16333 16334 * The type-safe stack code has been rejigged. It is now only compiled 16335 in when OpenSSL is configured with the DEBUG_SAFESTACK option and 16336 by default all type-specific stack functions are "#define"d back to 16337 standard stack functions. This results in more streamlined output 16338 but retains the type-safety checking possibilities of the original 16339 approach. 16340 16341 *Geoff Thorpe* 16342 16343 * The STACK code has been cleaned up, and certain type declarations 16344 that didn't make a lot of sense have been brought in line. This has 16345 also involved a cleanup of sorts in safestack.h to more correctly 16346 map type-safe stack functions onto their plain stack counterparts. 16347 This work has also resulted in a variety of "const"ifications of 16348 lots of the code, especially `_cmp` operations which should normally 16349 be prototyped with "const" parameters anyway. 16350 16351 *Geoff Thorpe* 16352 16353 * When generating bytes for the first time in md_rand.c, 'stir the pool' 16354 by seeding with STATE_SIZE dummy bytes (with zero entropy count). 16355 (The PRNG state consists of two parts, the large pool 'state' and 'md', 16356 where all of 'md' is used each time the PRNG is used, but 'state' 16357 is used only indexed by a cyclic counter. As entropy may not be 16358 well distributed from the beginning, 'md' is important as a 16359 chaining variable. However, the output function chains only half 16360 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 16361 all of 'md', and seeding with STATE_SIZE dummy bytes will result 16362 in all of 'state' being rewritten, with the new values depending 16363 on virtually all of 'md'. This overcomes the 80 bit limitation.) 16364 16365 *Bodo Moeller* 16366 16367 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 16368 the handshake is continued after ssl_verify_cert_chain(); 16369 otherwise, if SSL_VERIFY_NONE is set, remaining error codes 16370 can lead to 'unexplainable' connection aborts later. 16371 16372 *Bodo Moeller; problem tracked down by Lutz Jaenicke* 16373 16374 * Major EVP API cipher revision. 16375 Add hooks for extra EVP features. This allows various cipher 16376 parameters to be set in the EVP interface. Support added for variable 16377 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 16378 setting of RC2 and RC5 parameters. 16379 16380 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 16381 ciphers. 16382 16383 Remove lots of duplicated code from the EVP library. For example *every* 16384 cipher init() function handles the 'iv' in the same way according to the 16385 cipher mode. They also all do nothing if the 'key' parameter is NULL and 16386 for CFB and OFB modes they zero ctx->num. 16387 16388 New functionality allows removal of S/MIME code RC2 hack. 16389 16390 Most of the routines have the same form and so can be declared in terms 16391 of macros. 16392 16393 By shifting this to the top level EVP_CipherInit() it can be removed from 16394 all individual ciphers. If the cipher wants to handle IVs or keys 16395 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 16396 flags. 16397 16398 Change lots of functions like EVP_EncryptUpdate() to now return a 16399 value: although software versions of the algorithms cannot fail 16400 any installed hardware versions can. 16401 16402 *Steve Henson* 16403 16404 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 16405 this option is set, tolerate broken clients that send the negotiated 16406 protocol version number instead of the requested protocol version 16407 number. 16408 16409 *Bodo Moeller* 16410 16411 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag; 16412 i.e. non-zero for export ciphersuites, zero otherwise. 16413 Previous versions had this flag inverted, inconsistent with 16414 rsa_tmp_cb (..._TMP_RSA_CB). 16415 16416 *Bodo Moeller; problem reported by Amit Chopra* 16417 16418 * Add missing DSA library text string. Work around for some IIS 16419 key files with invalid SEQUENCE encoding. 16420 16421 *Steve Henson* 16422 16423 * Add a document (doc/standards.txt) that list all kinds of standards 16424 and so on that are implemented in OpenSSL. 16425 16426 *Richard Levitte* 16427 16428 * Enhance c_rehash script. Old version would mishandle certificates 16429 with the same subject name hash and wouldn't handle CRLs at all. 16430 Added -fingerprint option to crl utility, to support new c_rehash 16431 features. 16432 16433 *Steve Henson* 16434 16435 * Eliminate non-ANSI declarations in crypto.h and stack.h. 16436 16437 *Ulf Möller* 16438 16439 * Fix for SSL server purpose checking. Server checking was 16440 rejecting certificates which had extended key usage present 16441 but no ssl client purpose. 16442 16443 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>* 16444 16445 * Make PKCS#12 code work with no password. The PKCS#12 spec 16446 is a little unclear about how a blank password is handled. 16447 Since the password in encoded as a BMPString with terminating 16448 double NULL a zero length password would end up as just the 16449 double NULL. However no password at all is different and is 16450 handled differently in the PKCS#12 key generation code. NS 16451 treats a blank password as zero length. MSIE treats it as no 16452 password on export: but it will try both on import. We now do 16453 the same: PKCS12_parse() tries zero length and no password if 16454 the password is set to "" or NULL (NULL is now a valid password: 16455 it wasn't before) as does the pkcs12 application. 16456 16457 *Steve Henson* 16458 16459 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use 16460 perror when PEM_read_bio_X509_REQ fails, the error message must 16461 be obtained from the error queue. 16462 16463 *Bodo Moeller* 16464 16465 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 16466 it in ERR_remove_state if appropriate, and change ERR_get_state 16467 accordingly to avoid race conditions (this is necessary because 16468 thread_hash is no longer constant once set). 16469 16470 *Bodo Moeller* 16471 16472 * Bugfix for linux-elf makefile.one. 16473 16474 *Ulf Möller* 16475 16476 * RSA_get_default_method() will now cause a default 16477 RSA_METHOD to be chosen if one doesn't exist already. 16478 Previously this was only set during a call to RSA_new() 16479 or RSA_new_method(NULL) meaning it was possible for 16480 RSA_get_default_method() to return NULL. 16481 16482 *Geoff Thorpe* 16483 16484 * Added native name translation to the existing DSO code 16485 that will convert (if the flag to do so is set) filenames 16486 that are sufficiently small and have no path information 16487 into a canonical native form. Eg. "blah" converted to 16488 "libblah.so" or "blah.dll" etc. 16489 16490 *Geoff Thorpe* 16491 16492 * New function ERR_error_string_n(e, buf, len) which is like 16493 ERR_error_string(e, buf), but writes at most 'len' bytes 16494 including the 0 terminator. For ERR_error_string_n, 'buf' 16495 may not be NULL. 16496 16497 *Damien Miller <djm@mindrot.org>, Bodo Moeller* 16498 16499 * CONF library reworked to become more general. A new CONF 16500 configuration file reader "class" is implemented as well as a 16501 new functions (`NCONF_*`, for "New CONF") to handle it. The now 16502 old `CONF_*` functions are still there, but are reimplemented to 16503 work in terms of the new functions. Also, a set of functions 16504 to handle the internal storage of the configuration data is 16505 provided to make it easier to write new configuration file 16506 reader "classes" (I can definitely see something reading a 16507 configuration file in XML format, for example), called `_CONF_*`, 16508 or "the configuration storage API"... 16509 16510 The new configuration file reading functions are: 16511 16512 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 16513 NCONF_get_section, NCONF_get_string, NCONF_get_numbre 16514 16515 NCONF_default, NCONF_WIN32 16516 16517 NCONF_dump_fp, NCONF_dump_bio 16518 16519 NCONF_default and NCONF_WIN32 are method (or "class") choosers, 16520 NCONF_new creates a new CONF object. This works in the same way 16521 as other interfaces in OpenSSL, like the BIO interface. 16522 `NCONF_dump_*` dump the internal storage of the configuration file, 16523 which is useful for debugging. All other functions take the same 16524 arguments as the old `CONF_*` functions with the exception of the 16525 first that must be a `CONF *` instead of a `LHASH *`. 16526 16527 To make it easier to use the new classes with the old `CONF_*` functions, 16528 the function CONF_set_default_method is provided. 16529 16530 *Richard Levitte* 16531 16532 * Add '-tls1' option to 'openssl ciphers', which was already 16533 mentioned in the documentation but had not been implemented. 16534 (This option is not yet really useful because even the additional 16535 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 16536 16537 *Bodo Moeller* 16538 16539 * Initial DSO code added into libcrypto for letting OpenSSL (and 16540 OpenSSL-based applications) load shared libraries and bind to 16541 them in a portable way. 16542 16543 *Geoff Thorpe, with contributions from Richard Levitte* 16544 16545### Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 16546 16547 * Make sure _lrotl and _lrotr are only used with MSVC. 16548 16549 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 16550 (the default implementation of RAND_status). 16551 16552 * Rename openssl x509 option '-crlext', which was added in 0.9.5, 16553 to '-clrext' (= clear extensions), as intended and documented. 16554 *Bodo Moeller; inconsistency pointed out by Michael Attili 16555 <attili@amaxo.com>* 16556 16557 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length 16558 was larger than the MD block size. 16559 16560 *Steve Henson, pointed out by Yost William <YostW@tce.com>* 16561 16562 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 16563 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 16564 using the passed key: if the passed key was a private key the result 16565 of X509_print(), for example, would be to print out all the private key 16566 components. 16567 16568 *Steve Henson* 16569 16570 * des_quad_cksum() byte order bug fix. 16571 *Ulf Möller, using the problem description in krb4-0.9.7, where 16572 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>* 16573 16574 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 16575 discouraged. 16576 16577 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>* 16578 16579 * For easily testing in shell scripts whether some command 16580 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 16581 returns with exit code 0 iff no command of the given name is available. 16582 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 16583 the output goes to stdout and nothing is printed to stderr. 16584 Additional arguments are always ignored. 16585 16586 Since for each cipher there is a command of the same name, 16587 the 'no-cipher' compilation switches can be tested this way. 16588 16589 ('openssl no-XXX' is not able to detect pseudo-commands such 16590 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 16591 16592 *Bodo Moeller* 16593 16594 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 16595 16596 *Bodo Moeller* 16597 16598 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 16599 is set; it will be thrown away anyway because each handshake creates 16600 its own key. 16601 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 16602 to parameters -- in previous versions (since OpenSSL 0.9.3) the 16603 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning 16604 you effectively got SSL_OP_SINGLE_DH_USE when using this macro. 16605 16606 *Bodo Moeller* 16607 16608 * New s_client option -ign_eof: EOF at stdin is ignored, and 16609 'Q' and 'R' lose their special meanings (quit/renegotiate). 16610 This is part of what -quiet does; unlike -quiet, -ign_eof 16611 does not suppress any output. 16612 16613 *Richard Levitte* 16614 16615 * Add compatibility options to the purpose and trust code. The 16616 purpose X509_PURPOSE_ANY is "any purpose" which automatically 16617 accepts a certificate or CA, this was the previous behaviour, 16618 with all the associated security issues. 16619 16620 X509_TRUST_COMPAT is the old trust behaviour: only and 16621 automatically trust self signed roots in certificate store. A 16622 new trust setting X509_TRUST_DEFAULT is used to specify that 16623 a purpose has no associated trust setting and it should instead 16624 use the value in the default purpose. 16625 16626 *Steve Henson* 16627 16628 * Fix the PKCS#8 DSA private key code so it decodes keys again 16629 and fix a memory leak. 16630 16631 *Steve Henson* 16632 16633 * In util/mkerr.pl (which implements 'make errors'), preserve 16634 reason strings from the previous version of the .c file, as 16635 the default to have only downcase letters (and digits) in 16636 automatically generated reasons codes is not always appropriate. 16637 16638 *Bodo Moeller* 16639 16640 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 16641 using strerror. Previously, ERR_reason_error_string() returned 16642 library names as reason strings for SYSerr; but SYSerr is a special 16643 case where small numbers are errno values, not library numbers. 16644 16645 *Bodo Moeller* 16646 16647 * Add '-dsaparam' option to 'openssl dhparam' application. This 16648 converts DSA parameters into DH parameters. (When creating parameters, 16649 DSA_generate_parameters is used.) 16650 16651 *Bodo Moeller* 16652 16653 * Include 'length' (recommended exponent length) in C code generated 16654 by 'openssl dhparam -C'. 16655 16656 *Bodo Moeller* 16657 16658 * The second argument to set_label in perlasm was already being used 16659 so couldn't be used as a "file scope" flag. Moved to third argument 16660 which was free. 16661 16662 *Steve Henson* 16663 16664 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 16665 instead of RAND_bytes for encryption IVs and salts. 16666 16667 *Bodo Moeller* 16668 16669 * Include RAND_status() into RAND_METHOD instead of implementing 16670 it only for md_rand.c Otherwise replacing the PRNG by calling 16671 RAND_set_rand_method would be impossible. 16672 16673 *Bodo Moeller* 16674 16675 * Don't let DSA_generate_key() enter an infinite loop if the random 16676 number generation fails. 16677 16678 *Bodo Moeller* 16679 16680 * New 'rand' application for creating pseudo-random output. 16681 16682 *Bodo Moeller* 16683 16684 * Added configuration support for Linux/IA64 16685 16686 *Rolf Haberrecker <rolf@suse.de>* 16687 16688 * Assembler module support for Mingw32. 16689 16690 *Ulf Möller* 16691 16692 * Shared library support for HPUX (in shlib/). 16693 16694 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous* 16695 16696 * Shared library support for Solaris gcc. 16697 16698 *Lutz Behnke <behnke@trustcenter.de>* 16699 16700### Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 16701 16702 * PKCS7_encrypt() was adding text MIME headers twice because they 16703 were added manually and by SMIME_crlf_copy(). 16704 16705 *Steve Henson* 16706 16707 * In bntest.c don't call BN_rand with zero bits argument. 16708 16709 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>* 16710 16711 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 16712 case was implemented. This caused BN_div_recp() to fail occasionally. 16713 16714 *Ulf Möller* 16715 16716 * Add an optional second argument to the set_label() in the perl 16717 assembly language builder. If this argument exists and is set 16718 to 1 it signals that the assembler should use a symbol whose 16719 scope is the entire file, not just the current function. This 16720 is needed with MASM which uses the format label:: for this scope. 16721 16722 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>* 16723 16724 * Change the ASN1 types so they are typedefs by default. Before 16725 almost all types were #define'd to ASN1_STRING which was causing 16726 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 16727 for example. 16728 16729 *Steve Henson* 16730 16731 * Change names of new functions to the new get1/get0 naming 16732 convention: After 'get1', the caller owns a reference count 16733 and has to call `..._free`; 'get0' returns a pointer to some 16734 data structure without incrementing reference counters. 16735 (Some of the existing 'get' functions increment a reference 16736 counter, some don't.) 16737 Similarly, 'set1' and 'add1' functions increase reference 16738 counters or duplicate objects. 16739 16740 *Steve Henson* 16741 16742 * Allow for the possibility of temp RSA key generation failure: 16743 the code used to assume it always worked and crashed on failure. 16744 16745 *Steve Henson* 16746 16747 * Fix potential buffer overrun problem in BIO_printf(). 16748 *Ulf Möller, using public domain code by Patrick Powell; problem 16749 pointed out by David Sacerdote <das33@cornell.edu>* 16750 16751 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions 16752 RAND_egd() and RAND_status(). In the command line application, 16753 the EGD socket can be specified like a seed file using RANDFILE 16754 or -rand. 16755 16756 *Ulf Möller* 16757 16758 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 16759 Some CAs (e.g. Verisign) distribute certificates in this form. 16760 16761 *Steve Henson* 16762 16763 * Remove the SSL_ALLOW_ADH compile option and set the default cipher 16764 list to exclude them. This means that no special compilation option 16765 is needed to use anonymous DH: it just needs to be included in the 16766 cipher list. 16767 16768 *Steve Henson* 16769 16770 * Change the EVP_MD_CTX_type macro so its meaning consistent with 16771 EVP_MD_type. The old functionality is available in a new macro called 16772 EVP_MD_md(). Change code that uses it and update docs. 16773 16774 *Steve Henson* 16775 16776 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions 16777 where the `void *` argument is replaced by a function pointer argument. 16778 Previously `void *` was abused to point to functions, which works on 16779 many platforms, but is not correct. As these functions are usually 16780 called by macros defined in OpenSSL header files, most source code 16781 should work without changes. 16782 16783 *Richard Levitte* 16784 16785 * `<openssl/opensslconf.h>` (which is created by Configure) now contains 16786 sections with information on -D... compiler switches used for 16787 compiling the library so that applications can see them. To enable 16788 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES` 16789 must be defined. E.g., 16790 #define OPENSSL_ALGORITHM_DEFINES 16791 #include <openssl/opensslconf.h> 16792 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc. 16793 16794 *Richard Levitte, Ulf and Bodo Möller* 16795 16796 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 16797 record layer. 16798 16799 *Bodo Moeller* 16800 16801 * Change the 'other' type in certificate aux info to a STACK_OF 16802 X509_ALGOR. Although not an AlgorithmIdentifier as such it has 16803 the required ASN1 format: arbitrary types determined by an OID. 16804 16805 *Steve Henson* 16806 16807 * Add some PEM_write_X509_REQ_NEW() functions and a command line 16808 argument to 'req'. This is not because the function is newer or 16809 better than others it just uses the work 'NEW' in the certificate 16810 request header lines. Some software needs this. 16811 16812 *Steve Henson* 16813 16814 * Reorganise password command line arguments: now passwords can be 16815 obtained from various sources. Delete the PEM_cb function and make 16816 it the default behaviour: i.e. if the callback is NULL and the 16817 usrdata argument is not NULL interpret it as a null terminated pass 16818 phrase. If usrdata and the callback are NULL then the pass phrase 16819 is prompted for as usual. 16820 16821 *Steve Henson* 16822 16823 * Add support for the Compaq Atalla crypto accelerator. If it is installed, 16824 the support is automatically enabled. The resulting binaries will 16825 autodetect the card and use it if present. 16826 16827 *Ben Laurie and Compaq Inc.* 16828 16829 * Work around for Netscape hang bug. This sends certificate request 16830 and server done in one record. Since this is perfectly legal in the 16831 SSL/TLS protocol it isn't a "bug" option and is on by default. See 16832 the bugs/SSLv3 entry for more info. 16833 16834 *Steve Henson* 16835 16836 * HP-UX tune-up: new unified configs, HP C compiler bug workaround. 16837 16838 *Andy Polyakov* 16839 16840 * Add -rand argument to smime and pkcs12 applications and read/write 16841 of seed file. 16842 16843 *Steve Henson* 16844 16845 * New 'passwd' tool for crypt(3) and apr1 password hashes. 16846 16847 *Bodo Moeller* 16848 16849 * Add command line password options to the remaining applications. 16850 16851 *Steve Henson* 16852 16853 * Bug fix for BN_div_recp() for numerators with an even number of 16854 bits. 16855 16856 *Ulf Möller* 16857 16858 * More tests in bntest.c, and changed test_bn output. 16859 16860 *Ulf Möller* 16861 16862 * ./config recognizes MacOS X now. 16863 16864 *Andy Polyakov* 16865 16866 * Bug fix for BN_div() when the first words of num and divisor are 16867 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`. 16868 16869 *Ulf Möller* 16870 16871 * Add support for various broken PKCS#8 formats, and command line 16872 options to produce them. 16873 16874 *Steve Henson* 16875 16876 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 16877 get temporary BIGNUMs from a BN_CTX. 16878 16879 *Ulf Möller* 16880 16881 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 16882 for p == 0. 16883 16884 *Ulf Möller* 16885 16886 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and 16887 include a #define from the old name to the new. The original intent 16888 was that statically linked binaries could for example just call 16889 SSLeay_add_all_ciphers() to just add ciphers to the table and not 16890 link with digests. This never worked because SSLeay_add_all_digests() 16891 and SSLeay_add_all_ciphers() were in the same source file so calling 16892 one would link with the other. They are now in separate source files. 16893 16894 *Steve Henson* 16895 16896 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 16897 16898 *Steve Henson* 16899 16900 * Use a less unusual form of the Miller-Rabin primality test (it used 16901 a binary algorithm for exponentiation integrated into the Miller-Rabin 16902 loop, our standard modexp algorithms are faster). 16903 16904 *Bodo Moeller* 16905 16906 * Support for the EBCDIC character set completed. 16907 16908 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>* 16909 16910 * Source code cleanups: use const where appropriate, eliminate casts, 16911 use `void *` instead of `char *` in lhash. 16912 16913 *Ulf Möller* 16914 16915 * Bugfix: ssl3_send_server_key_exchange was not restartable 16916 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 16917 this the server could overwrite ephemeral keys that the client 16918 has already seen). 16919 16920 *Bodo Moeller* 16921 16922 * Turn DSA_is_prime into a macro that calls BN_is_prime, 16923 using 50 iterations of the Rabin-Miller test. 16924 16925 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 16926 iterations of the Rabin-Miller test as required by the appendix 16927 to FIPS PUB 186[-1]) instead of DSA_is_prime. 16928 As BN_is_prime_fasttest includes trial division, DSA parameter 16929 generation becomes much faster. 16930 16931 This implies a change for the callback functions in DSA_is_prime 16932 and DSA_generate_parameters: The callback function is called once 16933 for each positive witness in the Rabin-Miller test, not just 16934 occasionally in the inner loop; and the parameters to the 16935 callback function now provide an iteration count for the outer 16936 loop rather than for the current invocation of the inner loop. 16937 DSA_generate_parameters additionally can call the callback 16938 function with an 'iteration count' of -1, meaning that a 16939 candidate has passed the trial division test (when q is generated 16940 from an application-provided seed, trial division is skipped). 16941 16942 *Bodo Moeller* 16943 16944 * New function BN_is_prime_fasttest that optionally does trial 16945 division before starting the Rabin-Miller test and has 16946 an additional BN_CTX * argument (whereas BN_is_prime always 16947 has to allocate at least one BN_CTX). 16948 'callback(1, -1, cb_arg)' is called when a number has passed the 16949 trial division stage. 16950 16951 *Bodo Moeller* 16952 16953 * Fix for bug in CRL encoding. The validity dates weren't being handled 16954 as ASN1_TIME. 16955 16956 *Steve Henson* 16957 16958 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 16959 16960 *Steve Henson* 16961 16962 * New function BN_pseudo_rand(). 16963 16964 *Ulf Möller* 16965 16966 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 16967 bignum version of BN_from_montgomery() with the working code from 16968 SSLeay 0.9.0 (the word based version is faster anyway), and clean up 16969 the comments. 16970 16971 *Ulf Möller* 16972 16973 * Avoid a race condition in s2_clnt.c (function get_server_hello) that 16974 made it impossible to use the same SSL_SESSION data structure in 16975 SSL2 clients in multiple threads. 16976 16977 *Bodo Moeller* 16978 16979 * The return value of RAND_load_file() no longer counts bytes obtained 16980 by stat(). RAND_load_file(..., -1) is new and uses the complete file 16981 to seed the PRNG (previously an explicit byte count was required). 16982 16983 *Ulf Möller, Bodo Möller* 16984 16985 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 16986 used `char *` instead of `void *` and had casts all over the place. 16987 16988 *Steve Henson* 16989 16990 * Make BN_generate_prime() return NULL on error if ret!=NULL. 16991 16992 *Ulf Möller* 16993 16994 * Retain source code compatibility for BN_prime_checks macro: 16995 BN_is_prime(..., BN_prime_checks, ...) now uses 16996 BN_prime_checks_for_size to determine the appropriate number of 16997 Rabin-Miller iterations. 16998 16999 *Ulf Möller* 17000 17001 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 17002 DH_CHECK_P_NOT_SAFE_PRIME. 17003 (Check if this is true? OpenPGP calls them "strong".) 17004 17005 *Ulf Möller* 17006 17007 * Merge the functionality of "dh" and "gendh" programs into a new program 17008 "dhparam". The old programs are retained for now but will handle DH keys 17009 (instead of parameters) in future. 17010 17011 *Steve Henson* 17012 17013 * Make the ciphers, s_server and s_client programs check the return values 17014 when a new cipher list is set. 17015 17016 *Steve Henson* 17017 17018 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 17019 ciphers. Before when the 56bit ciphers were enabled the sorting was 17020 wrong. 17021 17022 The syntax for the cipher sorting has been extended to support sorting by 17023 cipher-strength (using the strength_bits hard coded in the tables). 17024 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`). 17025 17026 Fix a bug in the cipher-command parser: when supplying a cipher command 17027 string with an "undefined" symbol (neither command nor alphanumeric 17028 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now 17029 an error is flagged. 17030 17031 Due to the strength-sorting extension, the code of the 17032 ssl_create_cipher_list() function was completely rearranged. I hope that 17033 the readability was also increased :-) 17034 17035 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>* 17036 17037 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1 17038 for the first serial number and places 2 in the serial number file. This 17039 avoids problems when the root CA is created with serial number zero and 17040 the first user certificate has the same issuer name and serial number 17041 as the root CA. 17042 17043 *Steve Henson* 17044 17045 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 17046 the new code. Add documentation for this stuff. 17047 17048 *Steve Henson* 17049 17050 * Changes to X509_ATTRIBUTE utilities. These have been renamed from 17051 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509 17052 structures and behave in an analogous way to the X509v3 functions: 17053 they shouldn't be called directly but wrapper functions should be used 17054 instead. 17055 17056 So we also now have some wrapper functions that call the X509at functions 17057 when passed certificate requests. (TO DO: similar things can be done with 17058 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 17059 things. Some of these need some d2i or i2d and print functionality 17060 because they handle more complex structures.) 17061 17062 *Steve Henson* 17063 17064 * Add missing #ifndefs that caused missing symbols when building libssl 17065 as a shared library without RSA. Use #ifndef NO_SSL2 instead of 17066 NO_RSA in `ssl/s2*.c`. 17067 17068 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller* 17069 17070 * Precautions against using the PRNG uninitialized: RAND_bytes() now 17071 has a return value which indicates the quality of the random data 17072 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 17073 error queue. New function RAND_pseudo_bytes() generates output that is 17074 guaranteed to be unique but not unpredictable. RAND_add is like 17075 RAND_seed, but takes an extra argument for an entropy estimate 17076 (RAND_seed always assumes full entropy). 17077 17078 *Ulf Möller* 17079 17080 * Do more iterations of Rabin-Miller probable prime test (specifically, 17081 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 17082 instead of only 2 for all lengths; see BN_prime_checks_for_size definition 17083 in crypto/bn/bn_prime.c for the complete table). This guarantees a 17084 false-positive rate of at most 2^-80 for random input. 17085 17086 *Bodo Moeller* 17087 17088 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 17089 17090 *Bodo Moeller* 17091 17092 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 17093 in the 0.9.5 release), this returns the chain 17094 from an X509_CTX structure with a dup of the stack and all 17095 the X509 reference counts upped: so the stack will exist 17096 after X509_CTX_cleanup() has been called. Modify pkcs12.c 17097 to use this. 17098 17099 Also make SSL_SESSION_print() print out the verify return 17100 code. 17101 17102 *Steve Henson* 17103 17104 * Add manpage for the pkcs12 command. Also change the default 17105 behaviour so MAC iteration counts are used unless the new 17106 -nomaciter option is used. This improves file security and 17107 only older versions of MSIE (4.0 for example) need it. 17108 17109 *Steve Henson* 17110 17111 * Honor the no-xxx Configure options when creating .DEF files. 17112 17113 *Ulf Möller* 17114 17115 * Add PKCS#10 attributes to field table: challengePassword, 17116 unstructuredName and unstructuredAddress. These are taken from 17117 draft PKCS#9 v2.0 but are compatible with v1.2 provided no 17118 international characters are used. 17119 17120 More changes to X509_ATTRIBUTE code: allow the setting of types 17121 based on strings. Remove the 'loc' parameter when adding 17122 attributes because these will be a SET OF encoding which is sorted 17123 in ASN1 order. 17124 17125 *Steve Henson* 17126 17127 * Initial changes to the 'req' utility to allow request generation 17128 automation. This will allow an application to just generate a template 17129 file containing all the field values and have req construct the 17130 request. 17131 17132 Initial support for X509_ATTRIBUTE handling. Stacks of these are 17133 used all over the place including certificate requests and PKCS#7 17134 structures. They are currently handled manually where necessary with 17135 some primitive wrappers for PKCS#7. The new functions behave in a 17136 manner analogous to the X509 extension functions: they allow 17137 attributes to be looked up by NID and added. 17138 17139 Later something similar to the X509V3 code would be desirable to 17140 automatically handle the encoding, decoding and printing of the 17141 more complex types. The string types like challengePassword can 17142 be handled by the string table functions. 17143 17144 Also modified the multi byte string table handling. Now there is 17145 a 'global mask' which masks out certain types. The table itself 17146 can use the flag STABLE_NO_MASK to ignore the mask setting: this 17147 is useful when for example there is only one permissible type 17148 (as in countryName) and using the mask might result in no valid 17149 types at all. 17150 17151 *Steve Henson* 17152 17153 * Clean up 'Finished' handling, and add functions SSL_get_finished and 17154 SSL_get_peer_finished to allow applications to obtain the latest 17155 Finished messages sent to the peer or expected from the peer, 17156 respectively. (SSL_get_peer_finished is usually the Finished message 17157 actually received from the peer, otherwise the protocol will be aborted.) 17158 17159 As the Finished message are message digests of the complete handshake 17160 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 17161 be used for external authentication procedures when the authentication 17162 provided by SSL/TLS is not desired or is not enough. 17163 17164 *Bodo Moeller* 17165 17166 * Enhanced support for Alpha Linux is added. Now ./config checks if 17167 the host supports BWX extension and if Compaq C is present on the 17168 $PATH. Just exploiting of the BWX extension results in 20-30% 17169 performance kick for some algorithms, e.g. DES and RC4 to mention 17170 a couple. Compaq C in turn generates ~20% faster code for MD5 and 17171 SHA1. 17172 17173 *Andy Polyakov* 17174 17175 * Add support for MS "fast SGC". This is arguably a violation of the 17176 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 17177 weak crypto and after checking the certificate is SGC a second one 17178 with strong crypto. MS SGC stops the first handshake after receiving 17179 the server certificate message and sends a second client hello. Since 17180 a server will typically do all the time consuming operations before 17181 expecting any further messages from the client (server key exchange 17182 is the most expensive) there is little difference between the two. 17183 17184 To get OpenSSL to support MS SGC we have to permit a second client 17185 hello message after we have sent server done. In addition we have to 17186 reset the MAC if we do get this second client hello. 17187 17188 *Steve Henson* 17189 17190 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide 17191 if a DER encoded private key is RSA or DSA traditional format. Changed 17192 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 17193 format DER encoded private key. Newer code should use PKCS#8 format which 17194 has the key type encoded in the ASN1 structure. Added DER private key 17195 support to pkcs8 application. 17196 17197 *Steve Henson* 17198 17199 * SSL 3/TLS 1 servers now don't request certificates when an anonymous 17200 ciphersuites has been selected (as required by the SSL 3/TLS 1 17201 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 17202 is set, we interpret this as a request to violate the specification 17203 (the worst that can happen is a handshake failure, and 'correct' 17204 behaviour would result in a handshake failure anyway). 17205 17206 *Bodo Moeller* 17207 17208 * In SSL_CTX_add_session, take into account that there might be multiple 17209 SSL_SESSION structures with the same session ID (e.g. when two threads 17210 concurrently obtain them from an external cache). 17211 The internal cache can handle only one SSL_SESSION with a given ID, 17212 so if there's a conflict, we now throw out the old one to achieve 17213 consistency. 17214 17215 *Bodo Moeller* 17216 17217 * Add OIDs for idea and blowfish in CBC mode. This will allow both 17218 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 17219 some routines that use cipher OIDs: some ciphers do not have OIDs 17220 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 17221 example. 17222 17223 *Steve Henson* 17224 17225 * Simplify the trust setting structure and code. Now we just have 17226 two sequences of OIDs for trusted and rejected settings. These will 17227 typically have values the same as the extended key usage extension 17228 and any application specific purposes. 17229 17230 The trust checking code now has a default behaviour: it will just 17231 check for an object with the same NID as the passed id. Functions can 17232 be provided to override either the default behaviour or the behaviour 17233 for a given id. SSL client, server and email already have functions 17234 in place for compatibility: they check the NID and also return "trusted" 17235 if the certificate is self signed. 17236 17237 *Steve Henson* 17238 17239 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the 17240 traditional format into an EVP_PKEY structure. 17241 17242 *Steve Henson* 17243 17244 * Add a password callback function PEM_cb() which either prompts for 17245 a password if usr_data is NULL or otherwise assumes it is a null 17246 terminated password. Allow passwords to be passed on command line 17247 environment or config files in a few more utilities. 17248 17249 *Steve Henson* 17250 17251 * Add a bunch of DER and PEM functions to handle PKCS#8 format private 17252 keys. Add some short names for PKCS#8 PBE algorithms and allow them 17253 to be specified on the command line for the pkcs8 and pkcs12 utilities. 17254 Update documentation. 17255 17256 *Steve Henson* 17257 17258 * Support for ASN1 "NULL" type. This could be handled before by using 17259 ASN1_TYPE but there wasn't any function that would try to read a NULL 17260 and produce an error if it couldn't. For compatibility we also have 17261 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 17262 don't allocate anything because they don't need to. 17263 17264 *Steve Henson* 17265 17266 * Initial support for MacOS is now provided. Examine INSTALL.MacOS 17267 for details. 17268 17269 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>* 17270 17271 * Rebuild of the memory allocation routines used by OpenSSL code and 17272 possibly others as well. The purpose is to make an interface that 17273 provide hooks so anyone can build a separate set of allocation and 17274 deallocation routines to be used by OpenSSL, for example memory 17275 pool implementations, or something else, which was previously hard 17276 since Malloc(), Realloc() and Free() were defined as macros having 17277 the values malloc, realloc and free, respectively (except for Win32 17278 compilations). The same is provided for memory debugging code. 17279 OpenSSL already comes with functionality to find memory leaks, but 17280 this gives people a chance to debug other memory problems. 17281 17282 With these changes, a new set of functions and macros have appeared: 17283 17284 CRYPTO_set_mem_debug_functions() [F] 17285 CRYPTO_get_mem_debug_functions() [F] 17286 CRYPTO_dbg_set_options() [F] 17287 CRYPTO_dbg_get_options() [F] 17288 CRYPTO_malloc_debug_init() [M] 17289 17290 The memory debug functions are NULL by default, unless the library 17291 is compiled with CRYPTO_MDEBUG or friends is defined. If someone 17292 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 17293 gives the standard debugging functions that come with OpenSSL) or 17294 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 17295 provided by the library user) must be used. When the standard 17296 debugging functions are used, CRYPTO_dbg_set_options can be used to 17297 request additional information: 17298 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 17299 the CRYPTO_MDEBUG_xxx macro when compiling the library. 17300 17301 Also, things like CRYPTO_set_mem_functions will always give the 17302 expected result (the new set of functions is used for allocation 17303 and deallocation) at all times, regardless of platform and compiler 17304 options. 17305 17306 To finish it up, some functions that were never use in any other 17307 way than through macros have a new API and new semantic: 17308 17309 CRYPTO_dbg_malloc() 17310 CRYPTO_dbg_realloc() 17311 CRYPTO_dbg_free() 17312 17313 All macros of value have retained their old syntax. 17314 17315 *Richard Levitte and Bodo Moeller* 17316 17317 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 17318 ordering of SMIMECapabilities wasn't in "strength order" and there 17319 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 17320 algorithm. 17321 17322 *Steve Henson* 17323 17324 * Some ASN1 types with illegal zero length encoding (INTEGER, 17325 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 17326 17327 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson* 17328 17329 * Merge in my S/MIME library for OpenSSL. This provides a simple 17330 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 17331 functionality to handle multipart/signed properly) and a utility 17332 called 'smime' to call all this stuff. This is based on code I 17333 originally wrote for Celo who have kindly allowed it to be 17334 included in OpenSSL. 17335 17336 *Steve Henson* 17337 17338 * Add variants des_set_key_checked and des_set_key_unchecked of 17339 des_set_key (aka des_key_sched). Global variable des_check_key 17340 decides which of these is called by des_set_key; this way 17341 des_check_key behaves as it always did, but applications and 17342 the library itself, which was buggy for des_check_key == 1, 17343 have a cleaner way to pick the version they need. 17344 17345 *Bodo Moeller* 17346 17347 * New function PKCS12_newpass() which changes the password of a 17348 PKCS12 structure. 17349 17350 *Steve Henson* 17351 17352 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and 17353 dynamic mix. In both cases the ids can be used as an index into the 17354 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 17355 functions so they accept a list of the field values and the 17356 application doesn't need to directly manipulate the X509_TRUST 17357 structure. 17358 17359 *Steve Henson* 17360 17361 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 17362 need initialising. 17363 17364 *Steve Henson* 17365 17366 * Modify the way the V3 extension code looks up extensions. This now 17367 works in a similar way to the object code: we have some "standard" 17368 extensions in a static table which is searched with OBJ_bsearch() 17369 and the application can add dynamic ones if needed. The file 17370 crypto/x509v3/ext_dat.h now has the info: this file needs to be 17371 updated whenever a new extension is added to the core code and kept 17372 in ext_nid order. There is a simple program 'tabtest.c' which checks 17373 this. New extensions are not added too often so this file can readily 17374 be maintained manually. 17375 17376 There are two big advantages in doing things this way. The extensions 17377 can be looked up immediately and no longer need to be "added" using 17378 X509V3_add_standard_extensions(): this function now does nothing. 17379 Side note: I get *lots* of email saying the extension code doesn't 17380 work because people forget to call this function. 17381 Also no dynamic allocation is done unless new extensions are added: 17382 so if we don't add custom extensions there is no need to call 17383 X509V3_EXT_cleanup(). 17384 17385 *Steve Henson* 17386 17387 * Modify enc utility's salting as follows: make salting the default. Add a 17388 magic header, so unsalted files fail gracefully instead of just decrypting 17389 to garbage. This is because not salting is a big security hole, so people 17390 should be discouraged from doing it. 17391 17392 *Ben Laurie* 17393 17394 * Fixes and enhancements to the 'x509' utility. It allowed a message 17395 digest to be passed on the command line but it only used this 17396 parameter when signing a certificate. Modified so all relevant 17397 operations are affected by the digest parameter including the 17398 -fingerprint and -x509toreq options. Also -x509toreq choked if a 17399 DSA key was used because it didn't fix the digest. 17400 17401 *Steve Henson* 17402 17403 * Initial certificate chain verify code. Currently tests the untrusted 17404 certificates for consistency with the verify purpose (which is set 17405 when the X509_STORE_CTX structure is set up) and checks the pathlength. 17406 17407 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 17408 this is because it will reject chains with invalid extensions whereas 17409 every previous version of OpenSSL and SSLeay made no checks at all. 17410 17411 Trust code: checks the root CA for the relevant trust settings. Trust 17412 settings have an initial value consistent with the verify purpose: e.g. 17413 if the verify purpose is for SSL client use it expects the CA to be 17414 trusted for SSL client use. However the default value can be changed to 17415 permit custom trust settings: one example of this would be to only trust 17416 certificates from a specific "secure" set of CAs. 17417 17418 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 17419 which should be used for version portability: especially since the 17420 verify structure is likely to change more often now. 17421 17422 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 17423 to set them. If not set then assume SSL clients will verify SSL servers 17424 and vice versa. 17425 17426 Two new options to the verify program: -untrusted allows a set of 17427 untrusted certificates to be passed in and -purpose which sets the 17428 intended purpose of the certificate. If a purpose is set then the 17429 new chain verify code is used to check extension consistency. 17430 17431 *Steve Henson* 17432 17433 * Support for the authority information access extension. 17434 17435 *Steve Henson* 17436 17437 * Modify RSA and DSA PEM read routines to transparently handle 17438 PKCS#8 format private keys. New *_PUBKEY_* functions that handle 17439 public keys in a format compatible with certificate 17440 SubjectPublicKeyInfo structures. Unfortunately there were already 17441 functions called *_PublicKey_* which used various odd formats so 17442 these are retained for compatibility: however the DSA variants were 17443 never in a public release so they have been deleted. Changed dsa/rsa 17444 utilities to handle the new format: note no releases ever handled public 17445 keys so we should be OK. 17446 17447 The primary motivation for this change is to avoid the same fiasco 17448 that dogs private keys: there are several incompatible private key 17449 formats some of which are standard and some OpenSSL specific and 17450 require various evil hacks to allow partial transparent handling and 17451 even then it doesn't work with DER formats. Given the option anything 17452 other than PKCS#8 should be dumped: but the other formats have to 17453 stay in the name of compatibility. 17454 17455 With public keys and the benefit of hindsight one standard format 17456 is used which works with EVP_PKEY, RSA or DSA structures: though 17457 it clearly returns an error if you try to read the wrong kind of key. 17458 17459 Added a -pubkey option to the 'x509' utility to output the public key. 17460 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()` 17461 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add 17462 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`) 17463 that do the same as the `EVP_PKEY_assign_*()` except they up the 17464 reference count of the added key (they don't "swallow" the 17465 supplied key). 17466 17467 *Steve Henson* 17468 17469 * Fixes to crypto/x509/by_file.c the code to read in certificates and 17470 CRLs would fail if the file contained no certificates or no CRLs: 17471 added a new function to read in both types and return the number 17472 read: this means that if none are read it will be an error. The 17473 DER versions of the certificate and CRL reader would always fail 17474 because it isn't possible to mix certificates and CRLs in DER format 17475 without choking one or the other routine. Changed this to just read 17476 a certificate: this is the best we can do. Also modified the code 17477 in `apps/verify.c` to take notice of return codes: it was previously 17478 attempting to read in certificates from NULL pointers and ignoring 17479 any errors: this is one reason why the cert and CRL reader seemed 17480 to work. It doesn't check return codes from the default certificate 17481 routines: these may well fail if the certificates aren't installed. 17482 17483 *Steve Henson* 17484 17485 * Code to support otherName option in GeneralName. 17486 17487 *Steve Henson* 17488 17489 * First update to verify code. Change the verify utility 17490 so it warns if it is passed a self signed certificate: 17491 for consistency with the normal behaviour. X509_verify 17492 has been modified to it will now verify a self signed 17493 certificate if *exactly* the same certificate appears 17494 in the store: it was previously impossible to trust a 17495 single self signed certificate. This means that: 17496 openssl verify ss.pem 17497 now gives a warning about a self signed certificate but 17498 openssl verify -CAfile ss.pem ss.pem 17499 is OK. 17500 17501 *Steve Henson* 17502 17503 * For servers, store verify_result in SSL_SESSION data structure 17504 (and add it to external session representation). 17505 This is needed when client certificate verifications fails, 17506 but an application-provided verification callback (set by 17507 SSL_CTX_set_cert_verify_callback) allows accepting the session 17508 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 17509 but returns 1): When the session is reused, we have to set 17510 ssl->verify_result to the appropriate error code to avoid 17511 security holes. 17512 17513 *Bodo Moeller, problem pointed out by Lutz Jaenicke* 17514 17515 * Fix a bug in the new PKCS#7 code: it didn't consider the 17516 case in PKCS7_dataInit() where the signed PKCS7 structure 17517 didn't contain any existing data because it was being created. 17518 17519 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson* 17520 17521 * Add a salt to the key derivation routines in enc.c. This 17522 forms the first 8 bytes of the encrypted file. Also add a 17523 -S option to allow a salt to be input on the command line. 17524 17525 *Steve Henson* 17526 17527 * New function X509_cmp(). Oddly enough there wasn't a function 17528 to compare two certificates. We do this by working out the SHA1 17529 hash and comparing that. X509_cmp() will be needed by the trust 17530 code. 17531 17532 *Steve Henson* 17533 17534 * SSL_get1_session() is like SSL_get_session(), but increments 17535 the reference count in the SSL_SESSION returned. 17536 17537 *Geoff Thorpe <geoff@eu.c2.net>* 17538 17539 * Fix for 'req': it was adding a null to request attributes. 17540 Also change the X509_LOOKUP and X509_INFO code to handle 17541 certificate auxiliary information. 17542 17543 *Steve Henson* 17544 17545 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document 17546 the 'enc' command. 17547 17548 *Steve Henson* 17549 17550 * Add the possibility to add extra information to the memory leak 17551 detecting output, to form tracebacks, showing from where each 17552 allocation was originated: CRYPTO_push_info("constant string") adds 17553 the string plus current file name and line number to a per-thread 17554 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 17555 is like calling CYRPTO_pop_info() until the stack is empty. 17556 Also updated memory leak detection code to be multi-thread-safe. 17557 17558 *Richard Levitte* 17559 17560 * Add options -text and -noout to pkcs7 utility and delete the 17561 encryption options which never did anything. Update docs. 17562 17563 *Steve Henson* 17564 17565 * Add options to some of the utilities to allow the pass phrase 17566 to be included on either the command line (not recommended on 17567 OSes like Unix) or read from the environment. Update the 17568 manpages and fix a few bugs. 17569 17570 *Steve Henson* 17571 17572 * Add a few manpages for some of the openssl commands. 17573 17574 *Steve Henson* 17575 17576 * Fix the -revoke option in ca. It was freeing up memory twice, 17577 leaking and not finding already revoked certificates. 17578 17579 *Steve Henson* 17580 17581 * Extensive changes to support certificate auxiliary information. 17582 This involves the use of X509_CERT_AUX structure and X509_AUX 17583 functions. An X509_AUX function such as PEM_read_X509_AUX() 17584 can still read in a certificate file in the usual way but it 17585 will also read in any additional "auxiliary information". By 17586 doing things this way a fair degree of compatibility can be 17587 retained: existing certificates can have this information added 17588 using the new 'x509' options. 17589 17590 Current auxiliary information includes an "alias" and some trust 17591 settings. The trust settings will ultimately be used in enhanced 17592 certificate chain verification routines: currently a certificate 17593 can only be trusted if it is self signed and then it is trusted 17594 for all purposes. 17595 17596 *Steve Henson* 17597 17598 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`). 17599 The problem was that one of the replacement routines had not been working 17600 since SSLeay releases. For now the offending routine has been replaced 17601 with non-optimised assembler. Even so, this now gives around 95% 17602 performance improvement for 1024 bit RSA signs. 17603 17604 *Mark Cox* 17605 17606 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2 17607 handling. Most clients have the effective key size in bits equal to 17608 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 17609 A few however don't do this and instead use the size of the decrypted key 17610 to determine the RC2 key length and the AlgorithmIdentifier to determine 17611 the effective key length. In this case the effective key length can still 17612 be 40 bits but the key length can be 168 bits for example. This is fixed 17613 by manually forcing an RC2 key into the EVP_PKEY structure because the 17614 EVP code can't currently handle unusual RC2 key sizes: it always assumes 17615 the key length and effective key length are equal. 17616 17617 *Steve Henson* 17618 17619 * Add a bunch of functions that should simplify the creation of 17620 X509_NAME structures. Now you should be able to do: 17621 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 17622 and have it automatically work out the correct field type and fill in 17623 the structures. The more adventurous can try: 17624 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 17625 and it will (hopefully) work out the correct multibyte encoding. 17626 17627 *Steve Henson* 17628 17629 * Change the 'req' utility to use the new field handling and multibyte 17630 copy routines. Before the DN field creation was handled in an ad hoc 17631 way in req, ca, and x509 which was rather broken and didn't support 17632 BMPStrings or UTF8Strings. Since some software doesn't implement 17633 BMPStrings or UTF8Strings yet, they can be enabled using the config file 17634 using the dirstring_type option. See the new comment in the default 17635 openssl.cnf for more info. 17636 17637 *Steve Henson* 17638 17639 * Make crypto/rand/md_rand.c more robust: 17640 - Assure unique random numbers after fork(). 17641 - Make sure that concurrent threads access the global counter and 17642 md serializably so that we never lose entropy in them 17643 or use exactly the same state in multiple threads. 17644 Access to the large state is not always serializable because 17645 the additional locking could be a performance killer, and 17646 md should be large enough anyway. 17647 17648 *Bodo Moeller* 17649 17650 * New file `apps/app_rand.c` with commonly needed functionality 17651 for handling the random seed file. 17652 17653 Use the random seed file in some applications that previously did not: 17654 ca, 17655 dsaparam -genkey (which also ignored its '-rand' option), 17656 s_client, 17657 s_server, 17658 x509 (when signing). 17659 Except on systems with /dev/urandom, it is crucial to have a random 17660 seed file at least for key creation, DSA signing, and for DH exchanges; 17661 for RSA signatures we could do without one. 17662 17663 gendh and gendsa (unlike genrsa) used to read only the first byte 17664 of each file listed in the '-rand' option. The function as previously 17665 found in genrsa is now in app_rand.c and is used by all programs 17666 that support '-rand'. 17667 17668 *Bodo Moeller* 17669 17670 * In RAND_write_file, use mode 0600 for creating files; 17671 don't just chmod when it may be too late. 17672 17673 *Bodo Moeller* 17674 17675 * Report an error from X509_STORE_load_locations 17676 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 17677 17678 *Bill Perry* 17679 17680 * New function ASN1_mbstring_copy() this copies a string in either 17681 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 17682 into an ASN1_STRING type. A mask of permissible types is passed 17683 and it chooses the "minimal" type to use or an error if not type 17684 is suitable. 17685 17686 *Steve Henson* 17687 17688 * Add function equivalents to the various macros in asn1.h. The old 17689 macros are retained with an `M_` prefix. Code inside the library can 17690 use the `M_` macros. External code (including the openssl utility) 17691 should *NOT* in order to be "shared library friendly". 17692 17693 *Steve Henson* 17694 17695 * Add various functions that can check a certificate's extensions 17696 to see if it usable for various purposes such as SSL client, 17697 server or S/MIME and CAs of these types. This is currently 17698 VERY EXPERIMENTAL but will ultimately be used for certificate chain 17699 verification. Also added a -purpose flag to x509 utility to 17700 print out all the purposes. 17701 17702 *Steve Henson* 17703 17704 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated 17705 functions. 17706 17707 *Steve Henson* 17708 17709 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search 17710 for, obtain and decode and extension and obtain its critical flag. 17711 This allows all the necessary extension code to be handled in a 17712 single function call. 17713 17714 *Steve Henson* 17715 17716 * RC4 tune-up featuring 30-40% performance improvement on most RISC 17717 platforms. See crypto/rc4/rc4_enc.c for further details. 17718 17719 *Andy Polyakov* 17720 17721 * New -noout option to asn1parse. This causes no output to be produced 17722 its main use is when combined with -strparse and -out to extract data 17723 from a file (which may not be in ASN.1 format). 17724 17725 *Steve Henson* 17726 17727 * Fix for pkcs12 program. It was hashing an invalid certificate pointer 17728 when producing the local key id. 17729 17730 *Richard Levitte <levitte@stacken.kth.se>* 17731 17732 * New option -dhparam in s_server. This allows a DH parameter file to be 17733 stated explicitly. If it is not stated then it tries the first server 17734 certificate file. The previous behaviour hard coded the filename 17735 "server.pem". 17736 17737 *Steve Henson* 17738 17739 * Add -pubin and -pubout options to the rsa and dsa commands. These allow 17740 a public key to be input or output. For example: 17741 openssl rsa -in key.pem -pubout -out pubkey.pem 17742 Also added necessary DSA public key functions to handle this. 17743 17744 *Steve Henson* 17745 17746 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 17747 in the message. This was handled by allowing 17748 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 17749 17750 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>* 17751 17752 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 17753 to the end of the strings whereas this didn't. This would cause problems 17754 if strings read with d2i_ASN1_bytes() were later modified. 17755 17756 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>* 17757 17758 * Fix for base64 decode bug. When a base64 bio reads only one line of 17759 data and it contains EOF it will end up returning an error. This is 17760 caused by input 46 bytes long. The cause is due to the way base64 17761 BIOs find the start of base64 encoded data. They do this by trying a 17762 trial decode on each line until they find one that works. When they 17763 do a flag is set and it starts again knowing it can pass all the 17764 data directly through the decoder. Unfortunately it doesn't reset 17765 the context it uses. This means that if EOF is reached an attempt 17766 is made to pass two EOFs through the context and this causes the 17767 resulting error. This can also cause other problems as well. As is 17768 usual with these problems it takes *ages* to find and the fix is 17769 trivial: move one line. 17770 17771 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)* 17772 17773 * Ugly workaround to get s_client and s_server working under Windows. The 17774 old code wouldn't work because it needed to select() on sockets and the 17775 tty (for keypresses and to see if data could be written). Win32 only 17776 supports select() on sockets so we select() with a 1s timeout on the 17777 sockets and then see if any characters are waiting to be read, if none 17778 are present then we retry, we also assume we can always write data to 17779 the tty. This isn't nice because the code then blocks until we've 17780 received a complete line of data and it is effectively polling the 17781 keyboard at 1s intervals: however it's quite a bit better than not 17782 working at all :-) A dedicated Windows application might handle this 17783 with an event loop for example. 17784 17785 *Steve Henson* 17786 17787 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 17788 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 17789 will be called when RSA_sign() and RSA_verify() are used. This is useful 17790 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 17791 For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 17792 should *not* be used: RSA_sign() and RSA_verify() must be used instead. 17793 This necessitated the support of an extra signature type NID_md5_sha1 17794 for SSL signatures and modifications to the SSL library to use it instead 17795 of calling RSA_public_decrypt() and RSA_private_encrypt(). 17796 17797 *Steve Henson* 17798 17799 * Add new -verify -CAfile and -CApath options to the crl program, these 17800 will lookup a CRL issuers certificate and verify the signature in a 17801 similar way to the verify program. Tidy up the crl program so it 17802 no longer accesses structures directly. Make the ASN1 CRL parsing a bit 17803 less strict. It will now permit CRL extensions even if it is not 17804 a V2 CRL: this will allow it to tolerate some broken CRLs. 17805 17806 *Steve Henson* 17807 17808 * Initialize all non-automatic variables each time one of the openssl 17809 sub-programs is started (this is necessary as they may be started 17810 multiple times from the "OpenSSL>" prompt). 17811 17812 *Lennart Bang, Bodo Moeller* 17813 17814 * Preliminary compilation option RSA_NULL which disables RSA crypto without 17815 removing all other RSA functionality (this is what NO_RSA does). This 17816 is so (for example) those in the US can disable those operations covered 17817 by the RSA patent while allowing storage and parsing of RSA keys and RSA 17818 key generation. 17819 17820 *Steve Henson* 17821 17822 * Non-copying interface to BIO pairs. 17823 (still largely untested) 17824 17825 *Bodo Moeller* 17826 17827 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive 17828 ASCII string. This was handled independently in various places before. 17829 17830 *Steve Henson* 17831 17832 * New functions UTF8_getc() and UTF8_putc() that parse and generate 17833 UTF8 strings a character at a time. 17834 17835 *Steve Henson* 17836 17837 * Use client_version from client hello to select the protocol 17838 (s23_srvr.c) and for RSA client key exchange verification 17839 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 17840 17841 *Bodo Moeller* 17842 17843 * Add various utility functions to handle SPKACs, these were previously 17844 handled by poking round in the structure internals. Added new function 17845 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 17846 print, verify and generate SPKACs. Based on an original idea from 17847 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 17848 17849 *Steve Henson* 17850 17851 * RIPEMD160 is operational on all platforms and is back in 'make test'. 17852 17853 *Andy Polyakov* 17854 17855 * Allow the config file extension section to be overwritten on the 17856 command line. Based on an original idea from Massimiliano Pala 17857 <madwolf@comune.modena.it>. The new option is called -extensions 17858 and can be applied to ca, req and x509. Also -reqexts to override 17859 the request extensions in req and -crlexts to override the crl extensions 17860 in ca. 17861 17862 *Steve Henson* 17863 17864 * Add new feature to the SPKAC handling in ca. Now you can include 17865 the same field multiple times by preceding it by "XXXX." for example: 17866 1.OU="Unit name 1" 17867 2.OU="Unit name 2" 17868 this is the same syntax as used in the req config file. 17869 17870 *Steve Henson* 17871 17872 * Allow certificate extensions to be added to certificate requests. These 17873 are specified in a 'req_extensions' option of the req section of the 17874 config file. They can be printed out with the -text option to req but 17875 are otherwise ignored at present. 17876 17877 *Steve Henson* 17878 17879 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 17880 data read consists of only the final block it would not decrypted because 17881 EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 17882 A misplaced 'break' also meant the decrypted final block might not be 17883 copied until the next read. 17884 17885 *Steve Henson* 17886 17887 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 17888 a few extra parameters to the DH structure: these will be useful if 17889 for example we want the value of 'q' or implement X9.42 DH. 17890 17891 *Steve Henson* 17892 17893 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and 17894 provides hooks that allow the default DSA functions or functions on a 17895 "per key" basis to be replaced. This allows hardware acceleration and 17896 hardware key storage to be handled without major modification to the 17897 library. Also added low-level modexp hooks and CRYPTO_EX structure and 17898 associated functions. 17899 17900 *Steve Henson* 17901 17902 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 17903 as "read only": it can't be written to and the buffer it points to will 17904 not be freed. Reading from a read only BIO is much more efficient than 17905 a normal memory BIO. This was added because there are several times when 17906 an area of memory needs to be read from a BIO. The previous method was 17907 to create a memory BIO and write the data to it, this results in two 17908 copies of the data and an O(n^2) reading algorithm. There is a new 17909 function BIO_new_mem_buf() which creates a read only memory BIO from 17910 an area of memory. Also modified the PKCS#7 routines to use read only 17911 memory BIOs. 17912 17913 *Steve Henson* 17914 17915 * Bugfix: ssl23_get_client_hello did not work properly when called in 17916 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 17917 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 17918 but a retry condition occurred while trying to read the rest. 17919 17920 *Bodo Moeller* 17921 17922 * The PKCS7_ENC_CONTENT_new() function was setting the content type as 17923 NID_pkcs7_encrypted by default: this was wrong since this should almost 17924 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 17925 the encrypted data type: this is a more sensible place to put it and it 17926 allows the PKCS#12 code to be tidied up that duplicated this 17927 functionality. 17928 17929 *Steve Henson* 17930 17931 * Changed obj_dat.pl script so it takes its input and output files on 17932 the command line. This should avoid shell escape redirection problems 17933 under Win32. 17934 17935 *Steve Henson* 17936 17937 * Initial support for certificate extension requests, these are included 17938 in things like Xenroll certificate requests. Included functions to allow 17939 extensions to be obtained and added. 17940 17941 *Steve Henson* 17942 17943 * -crlf option to s_client and s_server for sending newlines as 17944 CRLF (as required by many protocols). 17945 17946 *Bodo Moeller* 17947 17948### Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 17949 17950 * Install libRSAglue.a when OpenSSL is built with RSAref. 17951 17952 *Ralf S. Engelschall* 17953 17954 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency. 17955 17956 *Andrija Antonijevic <TheAntony2@bigfoot.com>* 17957 17958 * Fix -startdate and -enddate (which was missing) arguments to 'ca' 17959 program. 17960 17961 *Steve Henson* 17962 17963 * New function DSA_dup_DH, which duplicates DSA parameters/keys as 17964 DH parameters/keys (q is lost during that conversion, but the resulting 17965 DH parameters contain its length). 17966 17967 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 17968 much faster than DH_generate_parameters (which creates parameters 17969 where `p = 2*q + 1`), and also the smaller q makes DH computations 17970 much more efficient (160-bit exponentiation instead of 1024-bit 17971 exponentiation); so this provides a convenient way to support DHE 17972 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 17973 utter importance to use 17974 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 17975 or 17976 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 17977 when such DH parameters are used, because otherwise small subgroup 17978 attacks may become possible! 17979 17980 *Bodo Moeller* 17981 17982 * Avoid memory leak in i2d_DHparams. 17983 17984 *Bodo Moeller* 17985 17986 * Allow the -k option to be used more than once in the enc program: 17987 this allows the same encrypted message to be read by multiple recipients. 17988 17989 *Steve Henson* 17990 17991 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 17992 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 17993 it will always use the numerical form of the OID, even if it has a short 17994 or long name. 17995 17996 *Steve Henson* 17997 17998 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 17999 method only got called if p,q,dmp1,dmq1,iqmp components were present, 18000 otherwise bn_mod_exp was called. In the case of hardware keys for example 18001 no private key components need be present and it might store extra data 18002 in the RSA structure, which cannot be accessed from bn_mod_exp. 18003 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 18004 private key operations. 18005 18006 *Steve Henson* 18007 18008 * Added support for SPARC Linux. 18009 18010 *Andy Polyakov* 18011 18012 * pem_password_cb function type incompatibly changed from 18013 typedef int pem_password_cb(char *buf, int size, int rwflag); 18014 to 18015 ....(char *buf, int size, int rwflag, void *userdata); 18016 so that applications can pass data to their callbacks: 18017 The `PEM[_ASN1]_{read,write}...` functions and macros now take an 18018 additional void * argument, which is just handed through whenever 18019 the password callback is called. 18020 18021 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller* 18022 18023 New function SSL_CTX_set_default_passwd_cb_userdata. 18024 18025 Compatibility note: As many C implementations push function arguments 18026 onto the stack in reverse order, the new library version is likely to 18027 interoperate with programs that have been compiled with the old 18028 pem_password_cb definition (PEM_whatever takes some data that 18029 happens to be on the stack as its last argument, and the callback 18030 just ignores this garbage); but there is no guarantee whatsoever that 18031 this will work. 18032 18033 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 18034 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 18035 problems not only on Windows, but also on some Unix platforms. 18036 To avoid problematic command lines, these definitions are now in an 18037 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 18038 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 18039 18040 *Bodo Moeller* 18041 18042 * MIPS III/IV assembler module is reimplemented. 18043 18044 *Andy Polyakov* 18045 18046 * More DES library cleanups: remove references to srand/rand and 18047 delete an unused file. 18048 18049 *Ulf Möller* 18050 18051 * Add support for the free Netwide assembler (NASM) under Win32, 18052 since not many people have MASM (ml) and it can be hard to obtain. 18053 This is currently experimental but it seems to work OK and pass all 18054 the tests. Check out INSTALL.W32 for info. 18055 18056 *Steve Henson* 18057 18058 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 18059 without temporary keys kept an extra copy of the server key, 18060 and connections with temporary keys did not free everything in case 18061 of an error. 18062 18063 *Bodo Moeller* 18064 18065 * New function RSA_check_key and new openssl rsa option -check 18066 for verifying the consistency of RSA keys. 18067 18068 *Ulf Moeller, Bodo Moeller* 18069 18070 * Various changes to make Win32 compile work: 18071 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 18072 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 18073 comparison" warnings. 18074 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update. 18075 18076 *Steve Henson* 18077 18078 * Add a debugging option to PKCS#5 v2 key generation function: when 18079 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 18080 derived keys are printed to stderr. 18081 18082 *Steve Henson* 18083 18084 * Copy the flags in ASN1_STRING_dup(). 18085 18086 *Roman E. Pavlov <pre@mo.msk.ru>* 18087 18088 * The x509 application mishandled signing requests containing DSA 18089 keys when the signing key was also DSA and the parameters didn't match. 18090 18091 It was supposed to omit the parameters when they matched the signing key: 18092 the verifying software was then supposed to automatically use the CA's 18093 parameters if they were absent from the end user certificate. 18094 18095 Omitting parameters is no longer recommended. The test was also 18096 the wrong way round! This was probably due to unusual behaviour in 18097 EVP_cmp_parameters() which returns 1 if the parameters match. 18098 This meant that parameters were omitted when they *didn't* match and 18099 the certificate was useless. Certificates signed with 'ca' didn't have 18100 this bug. 18101 18102 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>* 18103 18104 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 18105 The interface is as follows: 18106 Applications can use 18107 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 18108 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 18109 "off" is now the default. 18110 The library internally uses 18111 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 18112 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 18113 to disable memory-checking temporarily. 18114 18115 Some inconsistent states that previously were possible (and were 18116 even the default) are now avoided. 18117 18118 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 18119 with each memory chunk allocated; this is occasionally more helpful 18120 than just having a counter. 18121 18122 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 18123 18124 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 18125 extensions. 18126 18127 *Bodo Moeller* 18128 18129 * Introduce "mode" for SSL structures (with defaults in SSL_CTX), 18130 which largely parallels "options", but is for changing API behaviour, 18131 whereas "options" are about protocol behaviour. 18132 Initial "mode" flags are: 18133 18134 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 18135 a single record has been written. 18136 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 18137 retries use the same buffer location. 18138 (But all of the contents must be 18139 copied!) 18140 18141 *Bodo Moeller* 18142 18143 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 18144 worked. 18145 18146 * Fix problems with no-hmac etc. 18147 18148 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>* 18149 18150 * New functions RSA_get_default_method(), RSA_set_method() and 18151 RSA_get_method(). These allows replacement of RSA_METHODs without having 18152 to mess around with the internals of an RSA structure. 18153 18154 *Steve Henson* 18155 18156 * Fix memory leaks in DSA_do_sign and DSA_is_prime. 18157 Also really enable memory leak checks in openssl.c and in some 18158 test programs. 18159 18160 *Chad C. Mulligan, Bodo Moeller* 18161 18162 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 18163 up the length of negative integers. This has now been simplified to just 18164 store the length when it is first determined and use it later, rather 18165 than trying to keep track of where data is copied and updating it to 18166 point to the end. 18167 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>* 18168 18169 * Add a new function PKCS7_signatureVerify. This allows the verification 18170 of a PKCS#7 signature but with the signing certificate passed to the 18171 function itself. This contrasts with PKCS7_dataVerify which assumes the 18172 certificate is present in the PKCS#7 structure. This isn't always the 18173 case: certificates can be omitted from a PKCS#7 structure and be 18174 distributed by "out of band" means (such as a certificate database). 18175 18176 *Steve Henson* 18177 18178 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the 18179 function prototypes in pem.h, also change util/mkdef.pl to add the 18180 necessary function names. 18181 18182 *Steve Henson* 18183 18184 * mk1mf.pl (used by Windows builds) did not properly read the 18185 options set by Configure in the top level Makefile, and Configure 18186 was not even able to write more than one option correctly. 18187 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 18188 18189 *Bodo Moeller* 18190 18191 * New functions CONF_load_bio() and CONF_load_fp() to allow a config 18192 file to be loaded from a BIO or FILE pointer. The BIO version will 18193 for example allow memory BIOs to contain config info. 18194 18195 *Steve Henson* 18196 18197 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 18198 Whoever hopes to achieve shared-library compatibility across versions 18199 must use this, not the compile-time macro. 18200 (Exercise 0.9.4: Which is the minimum library version required by 18201 such programs?) 18202 Note: All this applies only to multi-threaded programs, others don't 18203 need locks. 18204 18205 *Bodo Moeller* 18206 18207 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests 18208 through a BIO pair triggered the default case, i.e. 18209 SSLerr(...,SSL_R_UNKNOWN_STATE). 18210 18211 *Bodo Moeller* 18212 18213 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 18214 can use the SSL library even if none of the specific BIOs is 18215 appropriate. 18216 18217 *Bodo Moeller* 18218 18219 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 18220 for the encoded length. 18221 18222 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>* 18223 18224 * Add initial documentation of the X509V3 functions. 18225 18226 *Steve Henson* 18227 18228 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and 18229 PEM_write_bio_PKCS8PrivateKey() that are equivalent to 18230 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 18231 secure PKCS#8 private key format with a high iteration count. 18232 18233 *Steve Henson* 18234 18235 * Fix determination of Perl interpreter: A perl or perl5 18236 *directory* in $PATH was also accepted as the interpreter. 18237 18238 *Ralf S. Engelschall* 18239 18240 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking 18241 wrong with it but it was very old and did things like calling 18242 PEM_ASN1_read() directly and used MD5 for the hash not to mention some 18243 unusual formatting. 18244 18245 *Steve Henson* 18246 18247 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed 18248 to use the new extension code. 18249 18250 *Steve Henson* 18251 18252 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 18253 with macros. This should make it easier to change their form, add extra 18254 arguments etc. Fix a few PEM prototypes which didn't have cipher as a 18255 constant. 18256 18257 *Steve Henson* 18258 18259 * Add to configuration table a new entry that can specify an alternative 18260 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 18261 according to Mark Crispin <MRC@Panda.COM>. 18262 18263 *Bodo Moeller* 18264 18265 * DES CBC did not update the IV. Weird. 18266 18267 *Ben Laurie* 18268lse 18269 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 18270 Changing the behaviour of the former might break existing programs -- 18271 where IV updating is needed, des_ncbc_encrypt can be used. 18272ndif 18273 18274 * When bntest is run from "make test" it drives bc to check its 18275 calculations, as well as internally checking them. If an internal check 18276 fails, it needs to cause bc to give a non-zero result or make test carries 18277 on without noticing the failure. Fixed. 18278 18279 *Ben Laurie* 18280 18281 * DES library cleanups. 18282 18283 *Ulf Möller* 18284 18285 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 18286 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 18287 ciphers. NOTE: although the key derivation function has been verified 18288 against some published test vectors it has not been extensively tested 18289 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 18290 of v2.0. 18291 18292 *Steve Henson* 18293 18294 * Instead of "mkdir -p", which is not fully portable, use new 18295 Perl script "util/mkdir-p.pl". 18296 18297 *Bodo Moeller* 18298 18299 * Rewrite the way password based encryption (PBE) is handled. It used to 18300 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 18301 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 18302 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 18303 the 'parameter' field of the AlgorithmIdentifier is passed to the 18304 underlying key generation function so it must do its own ASN1 parsing. 18305 This has also changed the EVP_PBE_CipherInit() function which now has a 18306 'parameter' argument instead of literal salt and iteration count values 18307 and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 18308 18309 *Steve Henson* 18310 18311 * Support for PKCS#5 v1.5 compatible password based encryption algorithms 18312 and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 18313 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 18314 KEY" because this clashed with PKCS#8 unencrypted string. Since this 18315 value was just used as a "magic string" and not used directly its 18316 value doesn't matter. 18317 18318 *Steve Henson* 18319 18320 * Introduce some semblance of const correctness to BN. Shame C doesn't 18321 support mutable. 18322 18323 *Ben Laurie* 18324 18325 * "linux-sparc64" configuration (ultrapenguin). 18326 18327 *Ray Miller <ray.miller@oucs.ox.ac.uk>* 18328 "linux-sparc" configuration. 18329 18330 *Christian Forster <fo@hawo.stw.uni-erlangen.de>* 18331 18332 * config now generates no-xxx options for missing ciphers. 18333 18334 *Ulf Möller* 18335 18336 * Support the EBCDIC character set (work in progress). 18337 File ebcdic.c not yet included because it has a different license. 18338 18339 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>* 18340 18341 * Support BS2000/OSD-POSIX. 18342 18343 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>* 18344 18345 * Make callbacks for key generation use `void *` instead of `char *`. 18346 18347 *Ben Laurie* 18348 18349 * Make S/MIME samples compile (not yet tested). 18350 18351 *Ben Laurie* 18352 18353 * Additional typesafe stacks. 18354 18355 *Ben Laurie* 18356 18357 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 18358 18359 *Bodo Moeller* 18360 18361### Changes between 0.9.3 and 0.9.3a [29 May 1999] 18362 18363 * New configuration variant "sco5-gcc". 18364 18365 * Updated some demos. 18366 18367 *Sean O Riordain, Wade Scholine* 18368 18369 * Add missing BIO_free at exit of pkcs12 application. 18370 18371 *Wu Zhigang* 18372 18373 * Fix memory leak in conf.c. 18374 18375 *Steve Henson* 18376 18377 * Updates for Win32 to assembler version of MD5. 18378 18379 *Steve Henson* 18380 18381 * Set #! path to perl in `apps/der_chop` to where we found it 18382 instead of using a fixed path. 18383 18384 *Bodo Moeller* 18385 18386 * SHA library changes for irix64-mips4-cc. 18387 18388 *Andy Polyakov* 18389 18390 * Improvements for VMS support. 18391 18392 *Richard Levitte* 18393 18394### Changes between 0.9.2b and 0.9.3 [24 May 1999] 18395 18396 * Bignum library bug fix. IRIX 6 passes "make test" now! 18397 This also avoids the problems with SC4.2 and unpatched SC5. 18398 18399 *Andy Polyakov <appro@fy.chalmers.se>* 18400 18401 * New functions sk_num, sk_value and sk_set to replace the previous macros. 18402 These are required because of the typesafe stack would otherwise break 18403 existing code. If old code used a structure member which used to be STACK 18404 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 18405 sk_num or sk_value it would produce an error because the num, data members 18406 are not present in STACK_OF. Now it just produces a warning. sk_set 18407 replaces the old method of assigning a value to sk_value 18408 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 18409 that does this will no longer work (and should use sk_set instead) but 18410 this could be regarded as a "questionable" behaviour anyway. 18411 18412 *Steve Henson* 18413 18414 * Fix most of the other PKCS#7 bugs. The "experimental" code can now 18415 correctly handle encrypted S/MIME data. 18416 18417 *Steve Henson* 18418 18419 * Change type of various DES function arguments from des_cblock 18420 (which means, in function argument declarations, pointer to char) 18421 to des_cblock * (meaning pointer to array with 8 char elements), 18422 which allows the compiler to do more typechecking; it was like 18423 that back in SSLeay, but with lots of ugly casts. 18424 18425 Introduce new type const_des_cblock. 18426 18427 *Bodo Moeller* 18428 18429 * Reorganise the PKCS#7 library and get rid of some of the more obvious 18430 problems: find RecipientInfo structure that matches recipient certificate 18431 and initialise the ASN1 structures properly based on passed cipher. 18432 18433 *Steve Henson* 18434 18435 * Belatedly make the BN tests actually check the results. 18436 18437 *Ben Laurie* 18438 18439 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 18440 to and from BNs: it was completely broken. New compilation option 18441 NEG_PUBKEY_BUG to allow for some broken certificates that encode public 18442 key elements as negative integers. 18443 18444 *Steve Henson* 18445 18446 * Reorganize and speed up MD5. 18447 18448 *Andy Polyakov <appro@fy.chalmers.se>* 18449 18450 * VMS support. 18451 18452 *Richard Levitte <richard@levitte.org>* 18453 18454 * New option -out to asn1parse to allow the parsed structure to be 18455 output to a file. This is most useful when combined with the -strparse 18456 option to examine the output of things like OCTET STRINGS. 18457 18458 *Steve Henson* 18459 18460 * Make SSL library a little more fool-proof by not requiring any longer 18461 that `SSL_set_{accept,connect}_state` be called before 18462 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted 18463 in many applications because usually everything *appeared* to work as 18464 intended anyway -- now it really works as intended). 18465 18466 *Bodo Moeller* 18467 18468 * Move openssl.cnf out of lib/. 18469 18470 *Ulf Möller* 18471 18472 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall 18473 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 18474 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+ 18475 18476 *Ralf S. Engelschall* 18477 18478 * Various fixes to the EVP and PKCS#7 code. It may now be able to 18479 handle PKCS#7 enveloped data properly. 18480 18481 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve* 18482 18483 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 18484 copying pointers. The cert_st handling is changed by this in 18485 various ways (and thus what used to be known as ctx->default_cert 18486 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert` 18487 any longer when s->cert does not give us what we need). 18488 ssl_cert_instantiate becomes obsolete by this change. 18489 As soon as we've got the new code right (possibly it already is?), 18490 we have solved a couple of bugs of the earlier code where s->cert 18491 was used as if it could not have been shared with other SSL structures. 18492 18493 Note that using the SSL API in certain dirty ways now will result 18494 in different behaviour than observed with earlier library versions: 18495 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx) 18496 does not influence s as it used to. 18497 18498 In order to clean up things more thoroughly, inside SSL_SESSION 18499 we don't use CERT any longer, but a new structure SESS_CERT 18500 that holds per-session data (if available); currently, this is 18501 the peer's certificate chain and, for clients, the server's certificate 18502 and temporary key. CERT holds only those values that can have 18503 meaningful defaults in an SSL_CTX. 18504 18505 *Bodo Moeller* 18506 18507 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 18508 from the internal representation. Various PKCS#7 fixes: remove some 18509 evil casts and set the enc_dig_alg field properly based on the signing 18510 key type. 18511 18512 *Steve Henson* 18513 18514 * Allow PKCS#12 password to be set from the command line or the 18515 environment. Let 'ca' get its config file name from the environment 18516 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 18517 and 'x509'). 18518 18519 *Steve Henson* 18520 18521 * Allow certificate policies extension to use an IA5STRING for the 18522 organization field. This is contrary to the PKIX definition but 18523 VeriSign uses it and IE5 only recognises this form. Document 'x509' 18524 extension option. 18525 18526 *Steve Henson* 18527 18528 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 18529 without disallowing inline assembler and the like for non-pedantic builds. 18530 18531 *Ben Laurie* 18532 18533 * Support Borland C++ builder. 18534 18535 *Janez Jere <jj@void.si>, modified by Ulf Möller* 18536 18537 * Support Mingw32. 18538 18539 *Ulf Möller* 18540 18541 * SHA-1 cleanups and performance enhancements. 18542 18543 *Andy Polyakov <appro@fy.chalmers.se>* 18544 18545 * Sparc v8plus assembler for the bignum library. 18546 18547 *Andy Polyakov <appro@fy.chalmers.se>* 18548 18549 * Accept any -xxx and +xxx compiler options in Configure. 18550 18551 *Ulf Möller* 18552 18553 * Update HPUX configuration. 18554 18555 *Anonymous* 18556 18557 * Add missing `sk_<type>_unshift()` function to safestack.h 18558 18559 *Ralf S. Engelschall* 18560 18561 * New function SSL_CTX_use_certificate_chain_file that sets the 18562 "extra_cert"s in addition to the certificate. (This makes sense 18563 only for "PEM" format files, as chains as a whole are not 18564 DER-encoded.) 18565 18566 *Bodo Moeller* 18567 18568 * Support verify_depth from the SSL API. 18569 x509_vfy.c had what can be considered an off-by-one-error: 18570 Its depth (which was not part of the external interface) 18571 was actually counting the number of certificates in a chain; 18572 now it really counts the depth. 18573 18574 *Bodo Moeller* 18575 18576 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 18577 instead of X509err, which often resulted in confusing error 18578 messages since the error codes are not globally unique 18579 (e.g. an alleged error in ssl3_accept when a certificate 18580 didn't match the private key). 18581 18582 * New function SSL_CTX_set_session_id_context that allows to set a default 18583 value (so that you don't need SSL_set_session_id_context for each 18584 connection using the SSL_CTX). 18585 18586 *Bodo Moeller* 18587 18588 * OAEP decoding bug fix. 18589 18590 *Ulf Möller* 18591 18592 * Support INSTALL_PREFIX for package builders, as proposed by 18593 David Harris. 18594 18595 *Bodo Moeller* 18596 18597 * New Configure options "threads" and "no-threads". For systems 18598 where the proper compiler options are known (currently Solaris 18599 and Linux), "threads" is the default. 18600 18601 *Bodo Moeller* 18602 18603 * New script util/mklink.pl as a faster substitute for util/mklink.sh. 18604 18605 *Bodo Moeller* 18606 18607 * Install various scripts to $(OPENSSLDIR)/misc, not to 18608 $(INSTALLTOP)/bin -- they shouldn't clutter directories 18609 such as /usr/local/bin. 18610 18611 *Bodo Moeller* 18612 18613 * "make linux-shared" to build shared libraries. 18614 18615 *Niels Poppe <niels@netbox.org>* 18616 18617 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...). 18618 18619 *Ulf Möller* 18620 18621 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 18622 extension adding in x509 utility. 18623 18624 *Steve Henson* 18625 18626 * Remove NOPROTO sections and error code comments. 18627 18628 *Ulf Möller* 18629 18630 * Partial rewrite of the DEF file generator to now parse the ANSI 18631 prototypes. 18632 18633 *Steve Henson* 18634 18635 * New Configure options --prefix=DIR and --openssldir=DIR. 18636 18637 *Ulf Möller* 18638 18639 * Complete rewrite of the error code script(s). It is all now handled 18640 by one script at the top level which handles error code gathering, 18641 header rewriting and C source file generation. It should be much better 18642 than the old method: it now uses a modified version of Ulf's parser to 18643 read the ANSI prototypes in all header files (thus the old K&R definitions 18644 aren't needed for error creation any more) and do a better job of 18645 translating function codes into names. The old 'ASN1 error code embedded 18646 in a comment' is no longer necessary and it doesn't use .err files which 18647 have now been deleted. Also the error code call doesn't have to appear all 18648 on one line (which resulted in some large lines...). 18649 18650 *Steve Henson* 18651 18652 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`. 18653 18654 *Bodo Moeller* 18655 18656 * Change behaviour of ssl2_read when facing length-0 packets: Don't return 18657 0 (which usually indicates a closed connection), but continue reading. 18658 18659 *Bodo Moeller* 18660 18661 * Fix some race conditions. 18662 18663 *Bodo Moeller* 18664 18665 * Add support for CRL distribution points extension. Add Certificate 18666 Policies and CRL distribution points documentation. 18667 18668 *Steve Henson* 18669 18670 * Move the autogenerated header file parts to crypto/opensslconf.h. 18671 18672 *Ulf Möller* 18673 18674 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 18675 8 of keying material. Merlin has also confirmed interop with this fix 18676 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 18677 18678 *Merlin Hughes <merlin@baltimore.ie>* 18679 18680 * Fix lots of warnings. 18681 18682 *Richard Levitte <levitte@stacken.kth.se>* 18683 18684 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 18685 the directory spec didn't end with a LIST_SEPARATOR_CHAR. 18686 18687 *Richard Levitte <levitte@stacken.kth.se>* 18688 18689 * Fix problems with sizeof(long) == 8. 18690 18691 *Andy Polyakov <appro@fy.chalmers.se>* 18692 18693 * Change functions to ANSI C. 18694 18695 *Ulf Möller* 18696 18697 * Fix typos in error codes. 18698 18699 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller* 18700 18701 * Remove defunct assembler files from Configure. 18702 18703 *Ulf Möller* 18704 18705 * SPARC v8 assembler BIGNUM implementation. 18706 18707 *Andy Polyakov <appro@fy.chalmers.se>* 18708 18709 * Support for Certificate Policies extension: both print and set. 18710 Various additions to support the r2i method this uses. 18711 18712 *Steve Henson* 18713 18714 * A lot of constification, and fix a bug in X509_NAME_oneline() that could 18715 return a const string when you are expecting an allocated buffer. 18716 18717 *Ben Laurie* 18718 18719 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 18720 types DirectoryString and DisplayText. 18721 18722 *Steve Henson* 18723 18724 * Add code to allow r2i extensions to access the configuration database, 18725 add an LHASH database driver and add several ctx helper functions. 18726 18727 *Steve Henson* 18728 18729 * Fix an evil bug in bn_expand2() which caused various BN functions to 18730 fail when they extended the size of a BIGNUM. 18731 18732 *Steve Henson* 18733 18734 * Various utility functions to handle SXNet extension. Modify mkdef.pl to 18735 support typesafe stack. 18736 18737 *Steve Henson* 18738 18739 * Fix typo in SSL_[gs]et_options(). 18740 18741 *Nils Frostberg <nils@medcom.se>* 18742 18743 * Delete various functions and files that belonged to the (now obsolete) 18744 old X509V3 handling code. 18745 18746 *Steve Henson* 18747 18748 * New Configure option "rsaref". 18749 18750 *Ulf Möller* 18751 18752 * Don't auto-generate pem.h. 18753 18754 *Bodo Moeller* 18755 18756 * Introduce type-safe ASN.1 SETs. 18757 18758 *Ben Laurie* 18759 18760 * Convert various additional casted stacks to type-safe STACK_OF() variants. 18761 18762 *Ben Laurie, Ralf S. Engelschall, Steve Henson* 18763 18764 * Introduce type-safe STACKs. This will almost certainly break lots of code 18765 that links with OpenSSL (well at least cause lots of warnings), but fear 18766 not: the conversion is trivial, and it eliminates loads of evil casts. A 18767 few STACKed things have been converted already. Feel free to convert more. 18768 In the fullness of time, I'll do away with the STACK type altogether. 18769 18770 *Ben Laurie* 18771 18772 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate 18773 specified in `<certfile>` by updating the entry in the index.txt file. 18774 This way one no longer has to edit the index.txt file manually for 18775 revoking a certificate. The -revoke option does the gory details now. 18776 18777 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall* 18778 18779 * Fix `openssl crl -noout -text` combination where `-noout` killed the 18780 `-text` option at all and this way the `-noout -text` combination was 18781 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`. 18782 18783 *Ralf S. Engelschall* 18784 18785 * Make sure a corresponding plain text error message exists for the 18786 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 18787 verify callback function determined that a certificate was revoked. 18788 18789 *Ralf S. Engelschall* 18790 18791 * Bugfix: In test/testenc, don't test `openssl <cipher>` for 18792 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 18793 all available ciphers including rc5, which was forgotten until now. 18794 In order to let the testing shell script know which algorithms 18795 are available, a new (up to now undocumented) command 18796 `openssl list-cipher-commands` is used. 18797 18798 *Bodo Moeller* 18799 18800 * Bugfix: s_client occasionally would sleep in select() when 18801 it should have checked SSL_pending() first. 18802 18803 *Bodo Moeller* 18804 18805 * New functions DSA_do_sign and DSA_do_verify to provide access to 18806 the raw DSA values prior to ASN.1 encoding. 18807 18808 *Ulf Möller* 18809 18810 * Tweaks to Configure 18811 18812 *Niels Poppe <niels@netbox.org>* 18813 18814 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 18815 yet... 18816 18817 *Steve Henson* 18818 18819 * New variables $(RANLIB) and $(PERL) in the Makefiles. 18820 18821 *Ulf Möller* 18822 18823 * New config option to avoid instructions that are illegal on the 80386. 18824 The default code is faster, but requires at least a 486. 18825 18826 *Ulf Möller* 18827 18828 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 18829 SSL2_SERVER_VERSION (not used at all) macros, which are now the 18830 same as SSL2_VERSION anyway. 18831 18832 *Bodo Moeller* 18833 18834 * New "-showcerts" option for s_client. 18835 18836 *Bodo Moeller* 18837 18838 * Still more PKCS#12 integration. Add pkcs12 application to openssl 18839 application. Various cleanups and fixes. 18840 18841 *Steve Henson* 18842 18843 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 18844 modify error routines to work internally. Add error codes and PBE init 18845 to library startup routines. 18846 18847 *Steve Henson* 18848 18849 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and 18850 packing functions to asn1 and evp. Changed function names and error 18851 codes along the way. 18852 18853 *Steve Henson* 18854 18855 * PKCS12 integration: and so it begins... First of several patches to 18856 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 18857 objects to objects.h 18858 18859 *Steve Henson* 18860 18861 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1 18862 and display support for Thawte strong extranet extension. 18863 18864 *Steve Henson* 18865 18866 * Add LinuxPPC support. 18867 18868 *Jeff Dubrule <igor@pobox.org>* 18869 18870 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 18871 bn_div_words in alpha.s. 18872 18873 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie* 18874 18875 * Make sure the RSA OAEP test is skipped under -DRSAref because 18876 OAEP isn't supported when OpenSSL is built with RSAref. 18877 18878 *Ulf Moeller <ulf@fitug.de>* 18879 18880 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 18881 so they no longer are missing under -DNOPROTO. 18882 18883 *Soren S. Jorvang <soren@t.dk>* 18884 18885### Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 18886 18887 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 18888 doesn't work when the session is reused. Coming soon! 18889 18890 *Ben Laurie* 18891 18892 * Fix a security hole, that allows sessions to be reused in the wrong 18893 context thus bypassing client cert protection! All software that uses 18894 client certs and session caches in multiple contexts NEEDS PATCHING to 18895 allow session reuse! A fuller solution is in the works. 18896 18897 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)* 18898 18899 * Some more source tree cleanups (removed obsolete files 18900 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 18901 permission on "config" script to be executable) and a fix for the INSTALL 18902 document. 18903 18904 *Ulf Moeller <ulf@fitug.de>* 18905 18906 * Remove some legacy and erroneous uses of malloc, free instead of 18907 Malloc, Free. 18908 18909 *Lennart Bang <lob@netstream.se>, with minor changes by Steve* 18910 18911 * Make rsa_oaep_test return non-zero on error. 18912 18913 *Ulf Moeller <ulf@fitug.de>* 18914 18915 * Add support for native Solaris shared libraries. Configure 18916 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 18917 if someone would make that last step automatic. 18918 18919 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>* 18920 18921 * ctx_size was not built with the right compiler during "make links". Fixed. 18922 18923 *Ben Laurie* 18924 18925 * Change the meaning of 'ALL' in the cipher list. It now means "everything 18926 except NULL ciphers". This means the default cipher list will no longer 18927 enable NULL ciphers. They need to be specifically enabled e.g. with 18928 the string "DEFAULT:eNULL". 18929 18930 *Steve Henson* 18931 18932 * Fix to RSA private encryption routines: if p < q then it would 18933 occasionally produce an invalid result. This will only happen with 18934 externally generated keys because OpenSSL (and SSLeay) ensure p > q. 18935 18936 *Steve Henson* 18937 18938 * Be less restrictive and allow also `perl util/perlpath.pl 18939 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`, 18940 because this way one can also use an interpreter named `perl5` (which is 18941 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 18942 installed as `perl`). 18943 18944 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>* 18945 18946 * Let util/clean-depend.pl work also with older Perl 5.00x versions. 18947 18948 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>* 18949 18950 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 18951 advapi32.lib to Win32 build and change the pem test comparison 18952 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 18953 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 18954 and crypto/des/ede_cbcm_enc.c. 18955 18956 *Steve Henson* 18957 18958 * DES quad checksum was broken on big-endian architectures. Fixed. 18959 18960 *Ben Laurie* 18961 18962 * Comment out two functions in bio.h that aren't implemented. Fix up the 18963 Win32 test batch file so it (might) work again. The Win32 test batch file 18964 is horrible: I feel ill.... 18965 18966 *Steve Henson* 18967 18968 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 18969 in e_os.h. Audit of header files to check ANSI and non ANSI 18970 sections: 10 functions were absent from non ANSI section and not exported 18971 from Windows DLLs. Fixed up libeay.num for new functions. 18972 18973 *Steve Henson* 18974 18975 * Make `openssl version` output lines consistent. 18976 18977 *Ralf S. Engelschall* 18978 18979 * Fix Win32 symbol export lists for BIO functions: Added 18980 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 18981 to ms/libeay{16,32}.def. 18982 18983 *Ralf S. Engelschall* 18984 18985 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 18986 fine under Unix and passes some trivial tests I've now added. But the 18987 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 18988 added to make sure no one expects that this stuff really works in the 18989 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 18990 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 18991 openssl_bio.xs. 18992 18993 *Ralf S. Engelschall* 18994 18995 * Fix the generation of two part addresses in perl. 18996 18997 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie* 18998 18999 * Add config entry for Linux on MIPS. 19000 19001 *John Tobey <jtobey@channel1.com>* 19002 19003 * Make links whenever Configure is run, unless we are on Windoze. 19004 19005 *Ben Laurie* 19006 19007 * Permit extensions to be added to CRLs using crl_section in openssl.cnf. 19008 Currently only issuerAltName and AuthorityKeyIdentifier make any sense 19009 in CRLs. 19010 19011 *Steve Henson* 19012 19013 * Add a useful kludge to allow package maintainers to specify compiler and 19014 other platforms details on the command line without having to patch the 19015 Configure script every time: One now can use 19016 `perl Configure <id>:<details>`, 19017 i.e. platform ids are allowed to have details appended 19018 to them (separated by colons). This is treated as there would be a static 19019 pre-configured entry in Configure's %table under key `<id>` with value 19020 `<details>` and `perl Configure <id>` is called. So, when you want to 19021 perform a quick test-compile under FreeBSD 3.1 with pgcc and without 19022 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"` 19023 now, which overrides the FreeBSD-elf entry on-the-fly. 19024 19025 *Ralf S. Engelschall* 19026 19027 * Disable new TLS1 ciphersuites by default: they aren't official yet. 19028 19029 *Ben Laurie* 19030 19031 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 19032 on the `perl Configure ...` command line. This way one can compile 19033 OpenSSL libraries with Position Independent Code (PIC) which is needed 19034 for linking it into DSOs. 19035 19036 *Ralf S. Engelschall* 19037 19038 * Remarkably, export ciphers were totally broken and no-one had noticed! 19039 Fixed. 19040 19041 *Ben Laurie* 19042 19043 * Cleaned up the LICENSE document: The official contact for any license 19044 questions now is the OpenSSL core team under openssl-core@openssl.org. 19045 And add a paragraph about the dual-license situation to make sure people 19046 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 19047 to the OpenSSL toolkit. 19048 19049 *Ralf S. Engelschall* 19050 19051 * General source tree makefile cleanups: Made `making xxx in yyy...` 19052 display consistent in the source tree and replaced `/bin/rm` by `rm`. 19053 Additionally cleaned up the `make links` target: Remove unnecessary 19054 semicolons, subsequent redundant removes, inline point.sh into mklink.sh 19055 to speed processing and no longer clutter the display with confusing 19056 stuff. Instead only the actually done links are displayed. 19057 19058 *Ralf S. Engelschall* 19059 19060 * Permit null encryption ciphersuites, used for authentication only. It used 19061 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 19062 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 19063 encryption. 19064 19065 *Ben Laurie* 19066 19067 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 19068 signed attributes when verifying signatures (this would break them), 19069 the detached data encoding was wrong and public keys obtained using 19070 X509_get_pubkey() weren't freed. 19071 19072 *Steve Henson* 19073 19074 * Add text documentation for the BUFFER functions. Also added a work around 19075 to a Win95 console bug. This was triggered by the password read stuff: the 19076 last character typed gets carried over to the next fread(). If you were 19077 generating a new cert request using 'req' for example then the last 19078 character of the passphrase would be CR which would then enter the first 19079 field as blank. 19080 19081 *Steve Henson* 19082 19083 * Added the new 'Includes OpenSSL Cryptography Software' button as 19084 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 19085 button and can be used by applications based on OpenSSL to show the 19086 relationship to the OpenSSL project. 19087 19088 *Ralf S. Engelschall* 19089 19090 * Remove confusing variables in function signatures in files 19091 ssl/ssl_lib.c and ssl/ssl.h. 19092 19093 *Lennart Bong <lob@kulthea.stacken.kth.se>* 19094 19095 * Don't install bss_file.c under PREFIX/include/ 19096 19097 *Lennart Bong <lob@kulthea.stacken.kth.se>* 19098 19099 * Get the Win32 compile working again. Modify mkdef.pl so it can handle 19100 functions that return function pointers and has support for NT specific 19101 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 19102 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 19103 unsigned to signed types: this was killing the Win32 compile. 19104 19105 *Steve Henson* 19106 19107 * Add new certificate file to stack functions, 19108 SSL_add_dir_cert_subjects_to_stack() and 19109 SSL_add_file_cert_subjects_to_stack(). These largely supplant 19110 SSL_load_client_CA_file(), and can be used to add multiple certs easily 19111 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 19112 This means that Apache-SSL and similar packages don't have to mess around 19113 to add as many CAs as they want to the preferred list. 19114 19115 *Ben Laurie* 19116 19117 * Experiment with doxygen documentation. Currently only partially applied to 19118 ssl/ssl_lib.c. 19119 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with 19120 openssl.doxy as the configuration file. 19121 19122 *Ben Laurie* 19123 19124 * Get rid of remaining C++-style comments which strict C compilers hate. 19125 19126 *Ralf S. Engelschall, pointed out by Carlos Amengual* 19127 19128 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 19129 compiled in by default: it has problems with large keys. 19130 19131 *Steve Henson* 19132 19133 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 19134 DH private keys and/or callback functions which directly correspond to 19135 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 19136 is needed for applications which have to configure certificates on a 19137 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 19138 (e.g. s_server). 19139 For the RSA certificate situation is makes no difference, but 19140 for the DSA certificate situation this fixes the "no shared cipher" 19141 problem where the OpenSSL cipher selection procedure failed because the 19142 temporary keys were not overtaken from the context and the API provided 19143 no way to reconfigure them. 19144 The new functions now let applications reconfigure the stuff and they 19145 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 19146 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 19147 non-public-API function ssl_cert_instantiate() is used as a helper 19148 function and also to reduce code redundancy inside ssl_rsa.c. 19149 19150 *Ralf S. Engelschall* 19151 19152 * Move s_server -dcert and -dkey options out of the undocumented feature 19153 area because they are useful for the DSA situation and should be 19154 recognized by the users. 19155 19156 *Ralf S. Engelschall* 19157 19158 * Fix the cipher decision scheme for export ciphers: the export bits are 19159 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 19160 SSL_EXP_MASK. So, the original variable has to be used instead of the 19161 already masked variable. 19162 19163 *Richard Levitte <levitte@stacken.kth.se>* 19164 19165 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c 19166 19167 *Richard Levitte <levitte@stacken.kth.se>* 19168 19169 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 19170 from `int` to `unsigned int` because it is a length and initialized by 19171 EVP_DigestFinal() which expects an `unsigned int *`. 19172 19173 *Richard Levitte <levitte@stacken.kth.se>* 19174 19175 * Don't hard-code path to Perl interpreter on shebang line of Configure 19176 script. Instead use the usual Shell->Perl transition trick. 19177 19178 *Ralf S. Engelschall* 19179 19180 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates 19181 (in addition to RSA certificates) to match the behaviour of `openssl dsa 19182 -noout -modulus` as it's already the case for `openssl rsa -noout 19183 -modulus`. For RSA the -modulus is the real "modulus" while for DSA 19184 currently the public key is printed (a decision which was already done by 19185 `openssl dsa -modulus` in the past) which serves a similar purpose. 19186 Additionally the NO_RSA no longer completely removes the whole -modulus 19187 option; it now only avoids using the RSA stuff. Same applies to NO_DSA 19188 now, too. 19189 19190 *Ralf S. Engelschall* 19191 19192 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 19193 BIO. See the source (crypto/evp/bio_ok.c) for more info. 19194 19195 *Arne Ansper <arne@ats.cyber.ee>* 19196 19197 * Dump the old yucky req code that tried (and failed) to allow raw OIDs 19198 to be added. Now both 'req' and 'ca' can use new objects defined in the 19199 config file. 19200 19201 *Steve Henson* 19202 19203 * Add cool BIO that does syslog (or event log on NT). 19204 19205 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie* 19206 19207 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 19208 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 19209 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 19210 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 19211 19212 *Ben Laurie* 19213 19214 * Add preliminary config info for new extension code. 19215 19216 *Steve Henson* 19217 19218 * Make RSA_NO_PADDING really use no padding. 19219 19220 *Ulf Moeller <ulf@fitug.de>* 19221 19222 * Generate errors when private/public key check is done. 19223 19224 *Ben Laurie* 19225 19226 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 19227 for some CRL extensions and new objects added. 19228 19229 *Steve Henson* 19230 19231 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private 19232 key usage extension and fuller support for authority key id. 19233 19234 *Steve Henson* 19235 19236 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 19237 padding method for RSA, which is recommended for new applications in PKCS 19238 #1 v2.0 (RFC 2437, October 1998). 19239 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 19240 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 19241 against Bleichbacher's attack on RSA. 19242 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 19243 Ben Laurie* 19244 19245 * Updates to the new SSL compression code 19246 19247 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)* 19248 19249 * Fix so that the version number in the master secret, when passed 19250 via RSA, checks that if TLS was proposed, but we roll back to SSLv3 19251 (because the server will not accept higher), that the version number 19252 is 0x03,0x01, not 0x03,0x00 19253 19254 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)* 19255 19256 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory 19257 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes 19258 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`. 19259 19260 *Steve Henson* 19261 19262 * Support for RAW extensions where an arbitrary extension can be 19263 created by including its DER encoding. See `apps/openssl.cnf` for 19264 an example. 19265 19266 *Steve Henson* 19267 19268 * Make sure latest Perl versions don't interpret some generated C array 19269 code as Perl array code in the crypto/err/err_genc.pl script. 19270 19271 *Lars Weber <3weber@informatik.uni-hamburg.de>* 19272 19273 * Modify ms/do_ms.bat to not generate assembly language makefiles since 19274 not many people have the assembler. Various Win32 compilation fixes and 19275 update to the INSTALL.W32 file with (hopefully) more accurate Win32 19276 build instructions. 19277 19278 *Steve Henson* 19279 19280 * Modify configure script 'Configure' to automatically create crypto/date.h 19281 file under Win32 and also build pem.h from pem.org. New script 19282 util/mkfiles.pl to create the MINFO file on environments that can't do a 19283 'make files': perl util/mkfiles.pl >MINFO should work. 19284 19285 *Steve Henson* 19286 19287 * Major rework of DES function declarations, in the pursuit of correctness 19288 and purity. As a result, many evil casts evaporated, and some weirdness, 19289 too. You may find this causes warnings in your code. Zapping your evil 19290 casts will probably fix them. Mostly. 19291 19292 *Ben Laurie* 19293 19294 * Fix for a typo in asn1.h. Bug fix to object creation script 19295 obj_dat.pl. It considered a zero in an object definition to mean 19296 "end of object": none of the objects in objects.h have any zeros 19297 so it wasn't spotted. 19298 19299 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>* 19300 19301 * Add support for Triple DES Cipher Block Chaining with Output Feedback 19302 Masking (CBCM). In the absence of test vectors, the best I have been able 19303 to do is check that the decrypt undoes the encrypt, so far. Send me test 19304 vectors if you have them. 19305 19306 *Ben Laurie* 19307 19308 * Correct calculation of key length for export ciphers (too much space was 19309 allocated for null ciphers). This has not been tested! 19310 19311 *Ben Laurie* 19312 19313 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage 19314 message is now correct (it understands "crypto" and "ssl" on its 19315 command line). There is also now an "update" option. This will update 19316 the util/ssleay.num and util/libeay.num files with any new functions. 19317 If you do a: 19318 perl util/mkdef.pl crypto ssl update 19319 it will update them. 19320 19321 *Steve Henson* 19322 19323 * Overhauled the Perl interface: 19324 - ported BN stuff to OpenSSL's different BN library 19325 - made the perl/ source tree CVS-aware 19326 - renamed the package from SSLeay to OpenSSL (the files still contain 19327 their history because I've copied them in the repository) 19328 - removed obsolete files (the test scripts will be replaced 19329 by better Test::Harness variants in the future) 19330 19331 *Ralf S. Engelschall* 19332 19333 * First cut for a very conservative source tree cleanup: 19334 1. merge various obsolete readme texts into doc/ssleay.txt 19335 where we collect the old documents and readme texts. 19336 2. remove the first part of files where I'm already sure that we no 19337 longer need them because of three reasons: either they are just temporary 19338 files which were left by Eric or they are preserved original files where 19339 I've verified that the diff is also available in the CVS via "cvs diff 19340 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 19341 the crypto/md/ stuff). 19342 19343 *Ralf S. Engelschall* 19344 19345 * More extension code. Incomplete support for subject and issuer alt 19346 name, issuer and authority key id. Change the i2v function parameters 19347 and add an extra 'crl' parameter in the X509V3_CTX structure: guess 19348 what that's for :-) Fix to ASN1 macro which messed up 19349 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 19350 19351 *Steve Henson* 19352 19353 * Preliminary support for ENUMERATED type. This is largely copied from the 19354 INTEGER code. 19355 19356 *Steve Henson* 19357 19358 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 19359 19360 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)* 19361 19362 * Make sure `make rehash` target really finds the `openssl` program. 19363 19364 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>* 19365 19366 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 19367 like to hear about it if this slows down other processors. 19368 19369 *Ben Laurie* 19370 19371 * Add CygWin32 platform information to Configure script. 19372 19373 *Alan Batie <batie@aahz.jf.intel.com>* 19374 19375 * Fixed ms/32all.bat script: `no_asm` -> `no-asm` 19376 19377 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>* 19378 19379 * New program nseq to manipulate netscape certificate sequences 19380 19381 *Steve Henson* 19382 19383 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 19384 few typos. 19385 19386 *Steve Henson* 19387 19388 * Fixes to BN code. Previously the default was to define BN_RECURSION 19389 but the BN code had some problems that would cause failures when 19390 doing certificate verification and some other functions. 19391 19392 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)* 19393 19394 * Add ASN1 and PEM code to support netscape certificate sequences. 19395 19396 *Steve Henson* 19397 19398 * Add ASN1 and PEM code to support netscape certificate sequences. 19399 19400 *Steve Henson* 19401 19402 * Add several PKIX and private extended key usage OIDs. 19403 19404 *Steve Henson* 19405 19406 * Modify the 'ca' program to handle the new extension code. Modify 19407 openssl.cnf for new extension format, add comments. 19408 19409 *Steve Henson* 19410 19411 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 19412 and add a sample to openssl.cnf so req -x509 now adds appropriate 19413 CA extensions. 19414 19415 *Steve Henson* 19416 19417 * Continued X509 V3 changes. Add to other makefiles, integrate with the 19418 error code, add initial support to X509_print() and x509 application. 19419 19420 *Steve Henson* 19421 19422 * Takes a deep breath and start adding X509 V3 extension support code. Add 19423 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 19424 stuff is currently isolated and isn't even compiled yet. 19425 19426 *Steve Henson* 19427 19428 * Continuing patches for GeneralizedTime. Fix up certificate and CRL 19429 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 19430 Removed the versions check from X509 routines when loading extensions: 19431 this allows certain broken certificates that don't set the version 19432 properly to be processed. 19433 19434 *Steve Henson* 19435 19436 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 19437 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 19438 can still be regenerated with "make depend". 19439 19440 *Ben Laurie* 19441 19442 * Spelling mistake in C version of CAST-128. 19443 19444 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>* 19445 19446 * Changes to the error generation code. The perl script err-code.pl 19447 now reads in the old error codes and retains the old numbers, only 19448 adding new ones if necessary. It also only changes the .err files if new 19449 codes are added. The makefiles have been modified to only insert errors 19450 when needed (to avoid needlessly modifying header files). This is done 19451 by only inserting errors if the .err file is newer than the auto generated 19452 C file. To rebuild all the error codes from scratch (the old behaviour) 19453 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 19454 or delete all the .err files. 19455 19456 *Steve Henson* 19457 19458 * CAST-128 was incorrectly implemented for short keys. The C version has 19459 been fixed, but is untested. The assembler versions are also fixed, but 19460 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 19461 to regenerate it if needed. 19462 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 19463 Hagino <itojun@kame.net>* 19464 19465 * File was opened incorrectly in randfile.c. 19466 19467 *Ulf Möller <ulf@fitug.de>* 19468 19469 * Beginning of support for GeneralizedTime. d2i, i2d, check and print 19470 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 19471 GeneralizedTime. ASN1_TIME is the proper type used in certificates et 19472 al: it's just almost always a UTCTime. Note this patch adds new error 19473 codes so do a "make errors" if there are problems. 19474 19475 *Steve Henson* 19476 19477 * Correct Linux 1 recognition in config. 19478 19479 *Ulf Möller <ulf@fitug.de>* 19480 19481 * Remove pointless MD5 hash when using DSA keys in ca. 19482 19483 *Anonymous <nobody@replay.com>* 19484 19485 * Generate an error if given an empty string as a cert directory. Also 19486 generate an error if handed NULL (previously returned 0 to indicate an 19487 error, but didn't set one). 19488 19489 *Ben Laurie, reported by Anonymous <nobody@replay.com>* 19490 19491 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 19492 19493 *Ben Laurie* 19494 19495 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 19496 parameters. This was causing a warning which killed off the Win32 compile. 19497 19498 *Steve Henson* 19499 19500 * Remove C++ style comments from crypto/bn/bn_local.h. 19501 19502 *Neil Costigan <neil.costigan@celocom.com>* 19503 19504 * The function OBJ_txt2nid was broken. It was supposed to return a nid 19505 based on a text string, looking up short and long names and finally 19506 "dot" format. The "dot" format stuff didn't work. Added new function 19507 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 19508 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 19509 OID is not part of the table. 19510 19511 *Steve Henson* 19512 19513 * Add prototypes to X509 lookup/verify methods, fixing a bug in 19514 X509_LOOKUP_by_alias(). 19515 19516 *Ben Laurie* 19517 19518 * Sort openssl functions by name. 19519 19520 *Ben Laurie* 19521 19522 * Get the `gendsa` command working and add it to the `list` command. Remove 19523 encryption from sample DSA keys (in case anyone is interested the password 19524 was "1234"). 19525 19526 *Steve Henson* 19527 19528 * Make *all* `*_free` functions accept a NULL pointer. 19529 19530 *Frans Heymans <fheymans@isaserver.be>* 19531 19532 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use 19533 NULL pointers. 19534 19535 *Anonymous <nobody@replay.com>* 19536 19537 * s_server should send the CAfile as acceptable CAs, not its own cert. 19538 19539 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>* 19540 19541 * Don't blow it for numeric `-newkey` arguments to `apps/req`. 19542 19543 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>* 19544 19545 * Temp key "for export" tests were wrong in s3_srvr.c. 19546 19547 *Anonymous <nobody@replay.com>* 19548 19549 * Add prototype for temp key callback functions 19550 SSL_CTX_set_tmp_{rsa,dh}_callback(). 19551 19552 *Ben Laurie* 19553 19554 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 19555 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 19556 19557 *Steve Henson* 19558 19559 * X509_name_add_entry() freed the wrong thing after an error. 19560 19561 *Arne Ansper <arne@ats.cyber.ee>* 19562 19563 * rsa_eay.c would attempt to free a NULL context. 19564 19565 *Arne Ansper <arne@ats.cyber.ee>* 19566 19567 * BIO_s_socket() had a broken should_retry() on Windoze. 19568 19569 *Arne Ansper <arne@ats.cyber.ee>* 19570 19571 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 19572 19573 *Arne Ansper <arne@ats.cyber.ee>* 19574 19575 * Make sure the already existing X509_STORE->depth variable is initialized 19576 in X509_STORE_new(), but document the fact that this variable is still 19577 unused in the certificate verification process. 19578 19579 *Ralf S. Engelschall* 19580 19581 * Fix the various library and `apps/` files to free up pkeys obtained from 19582 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 19583 19584 *Steve Henson* 19585 19586 * Fix reference counting in X509_PUBKEY_get(). This makes 19587 demos/maurice/example2.c work, amongst others, probably. 19588 19589 *Steve Henson and Ben Laurie* 19590 19591 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named 19592 `openssl` and second, the shortcut symlinks for the `openssl <command>` 19593 are no longer created. This way we have a single and consistent command 19594 line interface `openssl <command>`, similar to `cvs <command>`. 19595 19596 *Ralf S. Engelschall, Paul Sutton and Ben Laurie* 19597 19598 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 19599 BIT STRING wrapper always have zero unused bits. 19600 19601 *Steve Henson* 19602 19603 * Add CA.pl, perl version of CA.sh, add extended key usage OID. 19604 19605 *Steve Henson* 19606 19607 * Make the top-level INSTALL documentation easier to understand. 19608 19609 *Paul Sutton* 19610 19611 * Makefiles updated to exit if an error occurs in a sub-directory 19612 make (including if user presses ^C) [Paul Sutton] 19613 19614 * Make Montgomery context stuff explicit in RSA data structure. 19615 19616 *Ben Laurie* 19617 19618 * Fix build order of pem and err to allow for generated pem.h. 19619 19620 *Ben Laurie* 19621 19622 * Fix renumbering bug in X509_NAME_delete_entry(). 19623 19624 *Ben Laurie* 19625 19626 * Enhanced the err-ins.pl script so it makes the error library number 19627 global and can add a library name. This is needed for external ASN1 and 19628 other error libraries. 19629 19630 *Steve Henson* 19631 19632 * Fixed sk_insert which never worked properly. 19633 19634 *Steve Henson* 19635 19636 * Fix ASN1 macros so they can handle indefinite length constructed 19637 EXPLICIT tags. Some non standard certificates use these: they can now 19638 be read in. 19639 19640 *Steve Henson* 19641 19642 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 19643 into a single doc/ssleay.txt bundle. This way the information is still 19644 preserved but no longer messes up this directory. Now it's new room for 19645 the new set of documentation files. 19646 19647 *Ralf S. Engelschall* 19648 19649 * SETs were incorrectly DER encoded. This was a major pain, because they 19650 shared code with SEQUENCEs, which aren't coded the same. This means that 19651 almost everything to do with SETs or SEQUENCEs has either changed name or 19652 number of arguments. 19653 19654 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>* 19655 19656 * Fix test data to work with the above. 19657 19658 *Ben Laurie* 19659 19660 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 19661 was already fixed by Eric for 0.9.1 it seems. 19662 19663 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>* 19664 19665 * Autodetect FreeBSD3. 19666 19667 *Ben Laurie* 19668 19669 * Fix various bugs in Configure. This affects the following platforms: 19670 nextstep 19671 ncr-scde 19672 unixware-2.0 19673 unixware-2.0-pentium 19674 sco5-cc. 19675 19676 *Ben Laurie* 19677 19678 * Eliminate generated files from CVS. Reorder tests to regenerate files 19679 before they are needed. 19680 19681 *Ben Laurie* 19682 19683 * Generate Makefile.ssl from Makefile.org (to keep CVS happy). 19684 19685 *Ben Laurie* 19686 19687### Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 19688 19689 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 19690 changed SSLeay to OpenSSL in version strings. 19691 19692 *Ralf S. Engelschall* 19693 19694 * Some fixups to the top-level documents. 19695 19696 *Paul Sutton* 19697 19698 * Fixed the nasty bug where rsaref.h was not found under compile-time 19699 because the symlink to include/ was missing. 19700 19701 *Ralf S. Engelschall* 19702 19703 * Incorporated the popular no-RSA/DSA-only patches 19704 which allow to compile an RSA-free SSLeay. 19705 19706 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall* 19707 19708 * Fixed nasty rehash problem under `make -f Makefile.ssl links` 19709 when "ssleay" is still not found. 19710 19711 *Ralf S. Engelschall* 19712 19713 * Added more platforms to Configure: Cray T3E, HPUX 11, 19714 19715 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>* 19716 19717 * Updated the README file. 19718 19719 *Ralf S. Engelschall* 19720 19721 * Added various .cvsignore files in the CVS repository subdirs 19722 to make a "cvs update" really silent. 19723 19724 *Ralf S. Engelschall* 19725 19726 * Recompiled the error-definition header files and added 19727 missing symbols to the Win32 linker tables. 19728 19729 *Ralf S. Engelschall* 19730 19731 * Cleaned up the top-level documents; 19732 o new files: CHANGES and LICENSE 19733 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 19734 o merged COPYRIGHT into LICENSE 19735 o removed obsolete TODO file 19736 o renamed MICROSOFT to INSTALL.W32 19737 19738 *Ralf S. Engelschall* 19739 19740 * Removed dummy files from the 0.9.1b source tree: 19741 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 19742 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 19743 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 19744 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 19745 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 19746 19747 *Ralf S. Engelschall* 19748 19749 * Added various platform portability fixes. 19750 19751 *Mark J. Cox* 19752 19753 * The Genesis of the OpenSSL rpject: 19754 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 19755 Young and Tim J. Hudson created while they were working for C2Net until 19756 summer 1998. 19757 19758 *The OpenSSL Project* 19759 19760### Changes between 0.9.0b and 0.9.1b [not released] 19761 19762 * Updated a few CA certificates under certs/ 19763 19764 *Eric A. Young* 19765 19766 * Changed some BIGNUM api stuff. 19767 19768 *Eric A. Young* 19769 19770 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 19771 DGUX x86, Linux Alpha, etc. 19772 19773 *Eric A. Young* 19774 19775 * New COMP library [crypto/comp/] for SSL Record Layer Compression: 19776 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 19777 available). 19778 19779 *Eric A. Young* 19780 19781 * Add -strparse option to asn1pars program which parses nested 19782 binary structures 19783 19784 *Dr Stephen Henson <shenson@bigfoot.com>* 19785 19786 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 19787 19788 *Eric A. Young* 19789 19790 * DSA fix for "ca" program. 19791 19792 *Eric A. Young* 19793 19794 * Added "-genkey" option to "dsaparam" program. 19795 19796 *Eric A. Young* 19797 19798 * Added RIPE MD160 (rmd160) message digest. 19799 19800 *Eric A. Young* 19801 19802 * Added -a (all) option to "ssleay version" command. 19803 19804 *Eric A. Young* 19805 19806 * Added PLATFORM define which is the id given to Configure. 19807 19808 *Eric A. Young* 19809 19810 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 19811 19812 *Eric A. Young* 19813 19814 * Extended the ASN.1 parser routines. 19815 19816 *Eric A. Young* 19817 19818 * Extended BIO routines to support REUSEADDR, seek, tell, etc. 19819 19820 *Eric A. Young* 19821 19822 * Added a BN_CTX to the BN library. 19823 19824 *Eric A. Young* 19825 19826 * Fixed the weak key values in DES library 19827 19828 *Eric A. Young* 19829 19830 * Changed API in EVP library for cipher aliases. 19831 19832 *Eric A. Young* 19833 19834 * Added support for RC2/64bit cipher. 19835 19836 *Eric A. Young* 19837 19838 * Converted the lhash library to the crypto/mem.c functions. 19839 19840 *Eric A. Young* 19841 19842 * Added more recognized ASN.1 object ids. 19843 19844 *Eric A. Young* 19845 19846 * Added more RSA padding checks for SSL/TLS. 19847 19848 *Eric A. Young* 19849 19850 * Added BIO proxy/filter functionality. 19851 19852 *Eric A. Young* 19853 19854 * Added extra_certs to SSL_CTX which can be used 19855 send extra CA certificates to the client in the CA cert chain sending 19856 process. It can be configured with SSL_CTX_add_extra_chain_cert(). 19857 19858 *Eric A. Young* 19859 19860 * Now Fortezza is denied in the authentication phase because 19861 this is key exchange mechanism is not supported by SSLeay at all. 19862 19863 *Eric A. Young* 19864 19865 * Additional PKCS1 checks. 19866 19867 *Eric A. Young* 19868 19869 * Support the string "TLSv1" for all TLS v1 ciphers. 19870 19871 *Eric A. Young* 19872 19873 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 19874 ex_data index of the SSL context in the X509_STORE_CTX ex_data. 19875 19876 *Eric A. Young* 19877 19878 * Fixed a few memory leaks. 19879 19880 *Eric A. Young* 19881 19882 * Fixed various code and comment typos. 19883 19884 *Eric A. Young* 19885 19886 * A minor bug in ssl/s3_clnt.c where there would always be 4 0 19887 bytes sent in the client random. 19888 19889 *Edward Bishop <ebishop@spyglass.com>* 19890 19891<!-- Links --> 19892 19893[CVE-2024-4741]: https://www.openssl.org/news/vulnerabilities.html#CVE-2024-4741 19894[CVE-2024-4603]: https://www.openssl.org/news/vulnerabilities.html#CVE-2024-4603 19895[CVE-2024-2511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2024-2511 19896[CVE-2024-0727]: https://www.openssl.org/news/vulnerabilities.html#CVE-2024-0727 19897[CVE-2023-6237]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6237 19898[CVE-2023-6129]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6129 19899[CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678 19900[CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363 19901[CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807 19902[CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817 19903[CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446 19904[CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975 19905[RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5 19906[CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650 19907[CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255 19908[CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466 19909[CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465 19910[CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464 19911[CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401 19912[CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286 19913[CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217 19914[CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216 19915[CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215 19916[CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450 19917[CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304 19918[CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203 19919[CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996 19920[CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274 19921[CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097 19922[CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971 19923[CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967 19924[CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563 19925[CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559 19926[CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552 19927[CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551 19928[CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549 19929[CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547 19930[CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543 19931[CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407 19932[CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739 19933[CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737 19934[CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735 19935[CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734 19936[CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733 19937[CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732 19938[CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738 19939[CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737 19940[CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736 19941[CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735 19942[CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733 19943[CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732 19944[CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731 19945[CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730 19946[CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055 19947[CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054 19948[CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053 19949[CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052 19950[CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309 19951[CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308 19952[CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307 19953[CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306 19954[CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305 19955[CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304 19956[CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303 19957[CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302 19958[CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183 19959[CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182 19960[CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181 19961[CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180 19962[CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179 19963[CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178 19964[CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177 19965[CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176 19966[CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109 19967[CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107 19968[CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106 19969[CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105 19970[CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800 19971[CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799 19972[CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798 19973[CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797 19974[CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705 19975[CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702 19976[CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701 19977[CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197 19978[CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196 19979[CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195 19980[CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194 19981[CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193 19982[CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793 19983[CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792 19984[CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791 19985[CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790 19986[CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789 19987[CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788 19988[CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787 19989[CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293 19990[CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291 19991[CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290 19992[CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289 19993[CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288 19994[CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287 19995[CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286 19996[CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285 19997[CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209 19998[CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208 19999[CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207 20000[CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206 20001[CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205 20002[CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204 20003[CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275 20004[CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139 20005[CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572 20006[CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571 20007[CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570 20008[CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569 20009[CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568 20010[CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567 20011[CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566 20012[CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513 20013[CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512 20014[CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511 20015[CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510 20016[CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509 20017[CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508 20018[CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507 20019[CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506 20020[CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505 20021[CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470 20022[CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224 20023[CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221 20024[CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195 20025[CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160 20026[CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076 20027[CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450 20028[CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353 20029[CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169 20030[CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166 20031[CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686 20032[CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333 20033[CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110 20034[CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884 20035[CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050 20036[CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027 20037[CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619 20038[CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577 20039[CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576 20040[CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109 20041[CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108 20042[CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210 20043[CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207 20044[CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014 20045[CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252 20046[CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180 20047[CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864 20048[CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633 20049[CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740 20050[CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433 20051[CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355 20052[CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555 20053[CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245 20054[CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386 20055[CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379 20056[CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378 20057[CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377 20058[CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789 20059[CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591 20060[CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590 20061[CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077 20062[CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678 20063[CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672 20064[CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891 20065[CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135 20066[CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995 20067[CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343 20068[CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339 20069[CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738 20070[CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940 20071[CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937 20072[CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969 20073[CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112 20074[CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079 20075[CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851 20076[CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545 20077[CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544 20078[CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543 20079[CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078 20080[CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659 20081[CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657 20082[CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656 20083[CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655 20084