4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx] ###
27 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
28 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
29 used and applications should instead use the
30 L<EC_POINT_set_affine_coordinates(3)> and
31 L<EC_POINT_get_affine_coordinates(3)> functions.
35 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
36 arrays to be more easily constructed via a series of utility functions.
37 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
38 the various push functions and finally convert to a passable OSSL_PARAM
39 array using OSSL_PARAM_BLD_to_param().
43 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
44 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
45 internal keys, if they correspond to one of those built in types.
49 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
50 contain a provider side internal key.
54 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
55 They are old functions that we don't use, and that you could disable with
56 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
60 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
61 have been converted to Markdown with the goal to produce documents
62 which not only look pretty when viewed online in the browser, but
63 remain well readable inside a plain text editor.
65 To achieve this goal, a 'minimalistic' Markdown style has been applied
66 which avoids formatting elements that interfere too much with the
67 reading flow in the text file. For example, it
69 * avoids [ATX headings][] and uses [setext headings][] instead
70 (which works for `<h1>` and `<h2>` headings only).
71 * avoids [inline links][] and uses [reference links][] instead.
72 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
74 [ATX headings]: https://github.github.com/gfm/#atx-headings
75 [setext headings]: https://github.github.com/gfm/#setext-headings
76 [inline links]: https://github.github.com/gfm/#inline-link
77 [reference links]: https://github.github.com/gfm/#reference-link
78 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
79 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
84 * The test suite is changed to preserve results of each test recipe.
85 A new directory test-runs/ with subdirectories named like the
86 test recipes are created in the build tree for this purpose.
90 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
91 This adds crypto/cmp/, crpyto/crmf/, and test/cmp_*.
92 See L<OSSL_CMP_exec_IR_ses(3)> as starting point.
96 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
97 The legacy OCSP-focused and only partly documented API is retained.
98 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
102 * All of the low level RSA functions have been deprecated including:
104 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
105 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
106 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
107 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
108 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
109 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
110 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
111 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
112 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
113 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
114 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
115 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
116 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
117 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
118 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
119 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
120 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
121 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
122 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
123 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
124 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
125 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
126 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
127 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
128 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
129 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
130 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
131 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
133 Use of these low level functions has been informally discouraged for a long
134 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
135 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
136 L<EVP_PKEY_decrypt(3)>.
140 * X509 certificates signed using SHA1 are no longer allowed at security
142 In TLS/SSL the default security level is 1. It can be set either
143 using the cipher string with @SECLEVEL, or calling
144 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
145 a call to SSL_CTX_use_certificate() will fail if the security level is not
147 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
148 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
153 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
154 modified to use PKEY APIs. These commands are now in maintenance mode
155 and no new features will be added to them.
159 * The command line utility rsautl has been deprecated.
160 Instead use the pkeyutl program.
164 * The command line utilities genrsa and rsa have been modified to use PKEY
165 APIs These commands are now in maintenance mode and no new features will
170 * All of the low level DH functions have been deprecated including:
172 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
173 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
174 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
175 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
176 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
177 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
178 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
179 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
180 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
181 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
182 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
183 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
184 DH_meth_get_generate_params and DH_meth_set_generate_params.
186 Use of these low level functions has been informally discouraged for a long
187 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
188 and L<EVP_PKEY_derive(3)>.
192 * All of the low level DSA functions have been deprecated including:
194 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
195 DSA_get_default_method, DSA_set_method, DSA_get_method,
196 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
197 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
198 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
199 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
200 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
201 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
202 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
203 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
204 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
205 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
206 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
209 Use of these low level functions has been informally discouraged for a long
210 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
211 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
215 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
216 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
217 This means that applications don't have to look at the curve NID and
218 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
219 However, they still can, that EVP_PKEY_set_alias_type() call acts as
220 a no-op when the EVP_PKEY is already of the given type.
222 Parameter and key generation is also reworked to make it possible
223 to generate EVP_PKEY_SM2 parameters and keys without having to go
224 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
225 However, code that does the latter will still work as before.
229 * Deprecated low level ECDH and ECDSA functions. These include:
231 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
232 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
235 Use of these low level functions has been informally discouraged for a long
236 time. Instead applications should use the EVP_PKEY_derive(3),
237 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
241 * Deprecated the EC_KEY_METHOD functions. These include:
243 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
244 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
245 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
246 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
247 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
248 EC_KEY_METHOD_get_verify.
250 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
254 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
255 and EVP_PKEY_decrypt() instead.
256 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
257 and EVP_PKEY_encrypt() instead.
261 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
262 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
263 a new formulation to include all the things it can be used for,
264 as well as words of caution.
268 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
269 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
273 * All of the low level HMAC functions have been deprecated including:
275 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
276 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
279 Use of these low level functions has been informally discouraged for a long
280 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
281 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
282 and L<EVP_MAC_final(3)>.
286 * Over two thousand fixes were made to the documentation, including:
287 - Common options (such as -rand/-writerand, TLS version control, etc)
288 were refactored and point to newly-enhanced descriptions in openssl.pod.
289 - Added style conformance for all options (with help from Richard Levitte),
290 documented all reported missing options, added a CI build to check
291 that all options are documented and that no unimplemented options
293 - Documented some internals, such as all use of environment variables.
294 - Addressed all internal broken L<> references.
298 * All of the low level CMAC functions have been deprecated including:
300 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
301 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
303 Use of these low level functions has been informally discouraged for a long
304 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
305 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
306 and L<EVP_MAC_final(3)>.
310 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
311 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
314 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
315 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
316 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
317 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
318 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
319 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
320 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
321 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
322 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
323 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
325 Use of these low level functions has been informally discouraged
326 for a long time. Applications should use the EVP_DigestInit_ex(3),
327 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
331 * Corrected the documentation of the return values from the EVP_DigestSign*
332 set of functions. The documentation mentioned negative values for some
333 errors, but this was never the case, so the mention of negative values
336 Code that followed the documentation and thereby check with something
337 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
341 * All of the low level cipher functions have been deprecated including:
343 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
344 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
345 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
346 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
347 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
348 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
349 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
350 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
351 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
352 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
353 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
354 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
355 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
356 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
357 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
358 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
359 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
360 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
361 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
362 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
363 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
364 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
365 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
366 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
367 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
368 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
369 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
370 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
371 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
373 Use of these low level functions has been informally discouraged for
374 a long time. Applications should use the high level EVP APIs, e.g.
375 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
376 equivalently named decrypt functions instead.
378 *Matt Caswell and Paul Dale*
380 * Removed include/openssl/opensslconf.h.in and replaced it with
381 include/openssl/configuration.h.in, which differs in not including
382 <openssl/macros.h>. A short header include/openssl/opensslconf.h
383 was added to include both.
385 This allows internal hacks where one might need to modify the set
386 of configured macros, for example this if deprecated symbols are
387 still supposed to be available internally:
389 #include <openssl/configuration.h>
391 #undef OPENSSL_NO_DEPRECATED
392 #define OPENSSL_SUPPRESS_DEPRECATED
394 #include <openssl/macros.h>
396 This should not be used by applications that use the exported
397 symbols, as that will lead to linking errors.
401 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
402 used in exponentiation with 512-bit moduli. No EC algorithms are
403 affected. Analysis suggests that attacks against 2-prime RSA1024,
404 3-prime RSA1536, and DSA1024 as a result of this defect would be very
405 difficult to perform and are not believed likely. Attacks against DH512
406 are considered just feasible. However, for an attack the target would
407 have to re-use the DH512 private key, which is not recommended anyway.
408 Also applications directly using the low level API BN_mod_exp may be
409 affected if they use BN_FLG_CONSTTIME.
414 * Most memory-debug features have been deprecated, and the functionality
415 replaced with no-ops.
419 * Added documentation for the STACK API. OpenSSL only defines the STACK
420 functions where they are used.
424 * Introduced a new method type and API, OSSL_SERIALIZER, to
425 represent generic serializers. An implementation is expected to
426 be able to serialize an object associated with a given name (such
427 as an algorithm name for an asymmetric key) into forms given by
428 implementation properties.
430 Serializers are primarily used from inside libcrypto, through
431 calls to functions like EVP_PKEY_print_private(),
432 PEM_write_bio_PrivateKey() and similar.
434 Serializers are specified in such a way that they can be made to
435 directly handle the provider side portion of an object, if this
436 provider side part comes from the same provider as the serializer
437 itself, but can also be made to handle objects in parametrized
438 form (as an OSSL_PARAM array of data). This allows a provider to
439 offer generic serializers as a service for any other provider.
443 * Added a .pragma directive to the syntax of configuration files, to
444 allow varying behavior in a supported and predictable manner.
445 Currently added pragma:
449 This allows dollar signs to be a keyword character unless it's
450 followed by a opening brace or parenthesis. This is useful for
451 platforms where dollar signs are commonly used in names, such as
452 volume names and system directory names on VMS.
456 * Added functionality to create an EVP_PKEY from user data. This
457 is effectively the same as creating a RSA, DH or DSA object and
458 then assigning them to an EVP_PKEY, but directly using algorithm
459 agnostic EVP functions. A benefit is that this should be future
460 proof for public key algorithms to come.
464 * Change the interpretation of the '--api' configuration option to
465 mean that this is a desired API compatibility level with no
466 further meaning. The previous interpretation, that this would
467 also mean to remove all deprecated symbols up to and including
468 the given version, no requires that 'no-deprecated' is also used
469 in the configuration.
471 When building applications, the desired API compatibility level
472 can be set with the OPENSSL_API_COMPAT macro like before. For
473 API compatibility version below 3.0, the old style numerical
474 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
475 For version 3.0 and on, the value is expected to be the decimal
476 value calculated from the major and minor version like this:
478 MAJOR * 10000 + MINOR * 100
482 -DOPENSSL_API_COMPAT=30000 For 3.0
483 -DOPENSSL_API_COMPAT=30200 For 3.2
485 To hide declarations that are deprecated up to and including the
486 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
487 given when building the application as well.
491 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
492 access to certificate and CRL stores via URIs and OSSL_STORE
495 This adds the following functions:
497 - X509_LOOKUP_store()
498 - X509_STORE_load_file()
499 - X509_STORE_load_path()
500 - X509_STORE_load_store()
501 - SSL_add_store_cert_subjects_to_stack()
502 - SSL_CTX_set_default_verify_store()
503 - SSL_CTX_load_verify_file()
504 - SSL_CTX_load_verify_dir()
505 - SSL_CTX_load_verify_store()
507 Also, the following functions are now deprecated:
509 - X509_STORE_load_locations() (use X509_STORE_load_file(),
510 X509_STORE_load_path() or X509_STORE_load_store() instead)
511 - SSL_CTX_load_verify_locations() (use SSL_CTX_load_verify_file(),
512 SSL_CTX_load_verify_dir() or SSL_CTX_load_verify_store() instead)
516 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
517 The presence of this system service is determined at run-time.
521 * Added functionality to create an EVP_PKEY context based on data
522 for methods from providers. This takes an algorithm name and a
523 property query string and simply stores them, with the intent
524 that any operation that uses this context will use those strings
525 to fetch the needed methods implicitly, thereby making the port
526 of application written for pre-3.0 OpenSSL easier.
530 * The undocumented function NCONF_WIN32() has been deprecated; for
531 conversion details see the HISTORY section of doc/man5/config.pod
535 * Introduced the new functions EVP_DigestSignInit_ex() and
536 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
537 EVP_DigestVerifyUpdate() have been converted to functions. See the man
538 pages for further details.
542 * Over two thousand fixes were made to the documentation, including:
543 adding missing command flags, better style conformance, documentation
546 *Rich Salz, Richard Levitte*
548 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
549 X25519, X448, Ed25519 and Ed448.
553 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
558 * Deprecated the public definition of ERR_STATE as well as the function
559 ERR_get_state(). This is done in preparation of making ERR_STATE an
564 * Added ERR functionality to give callers access to the stored function
565 names that have replaced the older function code based functions.
567 New functions are ERR_get_error_func(), ERR_peek_error_func(),
568 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
569 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
570 and ERR_peek_last_error_all().
572 These functions have become deprecated: ERR_get_error_line_data(),
573 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
574 ERR_func_error_string().
578 * Extended testing to be verbose for failing tests only. The make variables
579 VERBOSE_FAILURE or VF can be used to enable this:
581 $ make VF=1 test # Unix
582 $ mms /macro=(VF=1) test ! OpenVMS
583 $ nmake VF=1 test # Windows
588 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
589 used even when parsing explicit parameters, when loading a serialized key
590 or calling `EC_GROUP_new_from_ecpkparameters()`/
591 `EC_GROUP_new_from_ecparameters()`.
592 This prevents bypass of security hardening and performance gains,
593 especially for curves with specialized EC_METHODs.
594 By default, if a key encoded with explicit parameters is loaded and later
595 serialized, the output is still encoded with explicit parameters, even if
596 internally a "named" EC_GROUP is used for computation.
600 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
601 this change, EC_GROUP_set_generator would accept order and/or cofactor as
602 NULL. After this change, only the cofactor parameter can be NULL. It also
603 does some minimal sanity checks on the passed order.
608 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
609 An attack is simple, if the first CMS_recipientInfo is valid but the
610 second CMS_recipientInfo is chosen ciphertext. If the second
611 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
612 encryption key will be replaced by garbage, and the message cannot be
613 decoded, but if the RSA decryption fails, the correct encryption key is
614 used and the recipient will not notice the attack.
615 As a work around for this potential attack the length of the decrypted
616 key must be equal to the cipher default key length, in case the
617 certifiate is not given and all recipientInfo are tried out.
618 The old behaviour can be re-enabled in the CMS code by setting the
619 CMS_DEBUG_DECRYPT flag.
623 * Early start up entropy quality from the DEVRANDOM seed source has been
624 improved for older Linux systems. The RAND subsystem will wait for
625 /dev/random to be producing output before seeding from /dev/urandom.
626 The seeded state is stored for future library initialisations using
627 a system global shared memory segment. The shared memory identifier
628 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
629 the desired value. The default identifier is 114.
633 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
634 when primes for RSA keys are computed.
635 Since we previously always generated primes == 2 (mod 3) for RSA keys,
636 the 2-prime and 3-prime RSA modules were easy to distinguish, since
637 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
638 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
639 This avoids possible fingerprinting of newly generated RSA modules.
643 * Correct the extended master secret constant on EBCDIC systems. Without this
644 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
645 negotiate EMS will fail. Unfortunately this also means that TLS connections
646 between EBCDIC systems with this fix, and EBCDIC systems without this
647 fix will fail if they negotiate EMS.
651 * Changed the library initialisation so that the config file is now loaded
652 by default. This was already the case for libssl. It now occurs for both
653 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
654 OPENSSL_init_crypto() to suppress automatic loading of a config file.
658 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
659 where the former acts as a replacement for ERR_put_error(), and the
660 latter replaces the combination ERR_put_error()+ERR_add_error_data().
661 ERR_raise_data() adds more flexibility by taking a format string and
662 an arbitrary number of arguments following it, to be processed with
667 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
668 to check if a named provider is loaded and available. When called, it
669 will also activate all fallback providers if such are still present.
673 * Enforce a minimum DH modulus size of 512 bits.
677 * Changed DH parameters to generate the order q subgroup instead of 2q.
678 Previously generated DH parameters are still accepted by DH_check
679 but DH_generate_key works around that by clearing bit 0 of the
680 private key for those. This avoids leaking bit 0 of the private key.
684 * Significantly reduce secure memory usage by the randomness pools.
688 * {CRYPTO,OPENSSL}_mem_debug_{push,pop} are now no-ops and have been
693 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
694 algorithms. An implementation of a key exchange algorithm can be obtained
695 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
696 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
697 the older EVP_PKEY_derive_init() function. See the man pages for the new
698 functions for further details.
702 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
706 * Removed the function names from error messages and deprecated the
709 * Removed NextStep support and the macro OPENSSL_UNISTD
713 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
714 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
715 Also removed "export var as function" capability; we do not export
716 variables, only functions.
720 * RC5_32_set_key has been changed to return an int type, with 0 indicating
721 an error and 1 indicating success. In previous versions of OpenSSL this
722 was a void type. If a key was set longer than the maximum possible this
727 * Support SM2 signing and verification schemes with X509 certificate.
731 * Use SHA256 as the default digest for TS query in the ts app.
735 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
736 This checks that the salt length is at least 128 bits, the derived key
737 length is at least 112 bits, and that the iteration count is at least 1000.
738 For backwards compatibility these checks are disabled by default in the
739 default provider, but are enabled by default in the fips provider.
740 To enable or disable these checks use the control
741 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
745 * Default cipher lists/suites are now available via a function, the
746 #defines are deprecated.
750 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
751 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
752 for Windows Store apps easier. Also, the "no-uplink" option has been added.
756 * Join the directories crypto/x509 and crypto/x509v3
760 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
761 This changes the size when using the genpkey app when no size is given. It
762 fixes an omission in earlier changes that changed all RSA, DSA and DH
763 generation apps to use 2048 bits by default.
767 * Added command 'openssl kdf' that uses the EVP_KDF API.
771 * Added command 'openssl mac' that uses the EVP_MAC API.
775 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
776 as default directories. Also added the command 'openssl info'
777 for scripting purposes.
781 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
782 deprecated. These undocumented functions were never integrated into the EVP
783 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
784 Bi-directional IGE mode. These modes were never formally standardised and
785 usage of these functions is believed to be very small. In particular
786 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
787 is ever used. The security implications are believed to be minimal, but
788 this issue was never fixed for backwards compatibility reasons. New code
789 should not use these modes.
793 * Add prediction resistance to the DRBG reseeding process.
797 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
798 mandated by IEEE Std 1619-2018.
802 * Added newline escaping functionality to a filename when using openssl dgst.
803 This output format is to replicate the output format found in the '*sum'
804 checksum programs. This aims to preserve backward compatibility.
806 *Matt Eaton, Richard Levitte, and Paul Dale*
808 * Removed the heartbeat message in DTLS feature, as it has very
809 little usage and doesn't seem to fulfill a valuable purpose.
810 The configuration option is now deprecated.
814 * Changed the output of 'openssl {digestname} < file' to display the
815 digest name in its output.
819 * Added a new generic trace API which provides support for enabling
820 instrumentation through trace output. This feature is mainly intended
821 as an aid for developers and is disabled by default. To utilize it,
822 OpenSSL needs to be configured with the `enable-trace` option.
824 If the tracing API is enabled, the application can activate trace output
825 by registering BIOs as trace channels for a number of tracing and debugging
828 The 'openssl' application has been expanded to enable any of the types
829 available via environment variables defined by the user, and serves as
830 one possible example on how to use this functionality.
832 *Richard Levitte & Matthias St. Pierre*
834 * Added build tests for C++. These are generated files that only do one
835 thing, to include one public OpenSSL head file each. This tests that
836 the public header files can be usefully included in a C++ application.
838 This test isn't enabled by default. It can be enabled with the option
839 'enable-buildtest-c++'.
843 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
847 * Add KMAC to EVP_MAC.
851 * Added property based algorithm implementation selection framework to
856 * Added SCA hardening for modular field inversion in EC_GROUP through
857 a new dedicated field_inv() pointer in EC_METHOD.
858 This also addresses a leakage affecting conversions from projective
859 to affine coordinates.
861 *Billy Bob Brumley, Nicola Tuveri*
863 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
864 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
865 those algorithms that were already supported through the EVP_PKEY API
866 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
867 and scrypt are now wrappers that call EVP_KDF.
871 * Build devcrypto engine as a dynamic engine.
875 * Add keyed BLAKE2 to EVP_MAC.
879 * Fix a bug in the computation of the endpoint-pair shared secret used
880 by DTLS over SCTP. This breaks interoperability with older versions
881 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
882 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
883 interoperability with such broken implementations. However, enabling
884 this switch breaks interoperability with correct implementations.
886 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
887 re-used X509_PUBKEY object if the second PUBKEY is malformed.
891 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
895 * Change the license to the Apache License v2.0.
899 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
901 - Major releases (indicated by incrementing the MAJOR release number)
902 may introduce incompatible API/ABI changes.
903 - Minor releases (indicated by incrementing the MINOR release number)
904 may introduce new features but retain API/ABI compatibility.
905 - Patch releases (indicated by incrementing the PATCH number)
906 are intended for bug fixes and other improvements of existing
907 features only (like improving performance or adding documentation)
908 and retain API/ABI compatibility.
912 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
916 * Remove the 'dist' target and add a tarball building script. The
917 'dist' target has fallen out of use, and it shouldn't be
918 necessary to configure just to create a source distribution.
922 * Recreate the OS390-Unix config target. It no longer relies on a
923 special script like it did for OpenSSL pre-1.1.0.
927 * Instead of having the source directories listed in Configure, add
928 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
933 * Add GMAC to EVP_MAC.
937 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
941 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
942 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
943 to facilitate the continued use of MACs through raw private keys in
944 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
948 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
949 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
953 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
954 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
955 are retained for backwards compatibility.
959 * AES-XTS mode now enforces that its two keys are different to mitigate
960 the attacked described in "Efficient Instantiations of Tweakable
961 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
962 Details of this attack can be obtained from:
963 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
967 * Rename the object files, i.e. give them other names than in previous
968 versions. Their names now include the name of the final product, as
969 well as its type mnemonic (bin, lib, shlib).
973 * Added new option for 'openssl list', '-objects', which will display the
974 list of built in objects, i.e. OIDs with names.
978 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
979 improves application performance by removing data copies and providing
980 applications with zero-copy system calls such as sendfile and splice.
988 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx] ###
991 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020] ###
993 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
994 while reading in libssl then we would report an error back to the
995 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
996 an error to the stack (which means we instead return SSL_ERROR_SSL) and
997 therefore give a hint as to what went wrong.
1001 * Check that ed25519 and ed448 are allowed by the security level. Previously
1002 signature algorithms not using an MD were not being checked that they were
1003 allowed by the security level.
1007 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1008 was not quite right. The behaviour was not consistent between resumption
1009 and normal handshakes, and also not quite consistent with historical
1010 behaviour. The behaviour in various scenarios has been clarified and
1011 it has been updated to make it match historical behaviour as closely as
1016 * *[VMS only]* The header files that the VMS compilers include automatically,
1017 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1018 that the C++ compiler doesn't understand. This is a shortcoming in the
1019 compiler, but can be worked around with `__cplusplus` guards.
1021 C++ applications that use OpenSSL libraries must be compiled using the
1022 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1023 functions. Otherwise, only functions with symbols of less than 31
1024 characters can be used, as the linker will not be able to successfully
1025 resolve symbols with longer names.
1029 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1030 The presence of this system service is determined at run-time.
1034 * Added newline escaping functionality to a filename when using openssl dgst.
1035 This output format is to replicate the output format found in the '*sum'
1036 checksum programs. This aims to preserve backward compatibility.
1038 *Matt Eaton, Richard Levitte, and Paul Dale*
1040 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1045 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019] ###
1047 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1048 number generator (RNG). This was intended to include protection in the
1049 event of a fork() system call in order to ensure that the parent and child
1050 processes did not share the same RNG state. However this protection was not
1051 being used in the default case.
1053 A partial mitigation for this issue is that the output from a high
1054 precision timer is mixed into the RNG state so the likelihood of a parent
1055 and child process sharing state is significantly reduced.
1057 If an application already calls OPENSSL_init_crypto() explicitly using
1058 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1061 *Matthias St. Pierre*
1063 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1064 used even when parsing explicit parameters, when loading a serialized key
1065 or calling `EC_GROUP_new_from_ecpkparameters()`/
1066 `EC_GROUP_new_from_ecparameters()`.
1067 This prevents bypass of security hardening and performance gains,
1068 especially for curves with specialized EC_METHODs.
1069 By default, if a key encoded with explicit parameters is loaded and later
1070 serialized, the output is still encoded with explicit parameters, even if
1071 internally a "named" EC_GROUP is used for computation.
1075 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1076 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1077 NULL. After this change, only the cofactor parameter can be NULL. It also
1078 does some minimal sanity checks on the passed order.
1083 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1084 An attack is simple, if the first CMS_recipientInfo is valid but the
1085 second CMS_recipientInfo is chosen ciphertext. If the second
1086 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1087 encryption key will be replaced by garbage, and the message cannot be
1088 decoded, but if the RSA decryption fails, the correct encryption key is
1089 used and the recipient will not notice the attack.
1090 As a work around for this potential attack the length of the decrypted
1091 key must be equal to the cipher default key length, in case the
1092 certifiate is not given and all recipientInfo are tried out.
1093 The old behaviour can be re-enabled in the CMS code by setting the
1094 CMS_DEBUG_DECRYPT flag.
1099 * Early start up entropy quality from the DEVRANDOM seed source has been
1100 improved for older Linux systems. The RAND subsystem will wait for
1101 /dev/random to be producing output before seeding from /dev/urandom.
1102 The seeded state is stored for future library initialisations using
1103 a system global shared memory segment. The shared memory identifier
1104 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1105 the desired value. The default identifier is 114.
1109 * Correct the extended master secret constant on EBCDIC systems. Without this
1110 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1111 negotiate EMS will fail. Unfortunately this also means that TLS connections
1112 between EBCDIC systems with this fix, and EBCDIC systems without this
1113 fix will fail if they negotiate EMS.
1117 * Use Windows installation paths in the mingw builds
1119 Mingw isn't a POSIX environment per se, which means that Windows
1120 paths should be used for installation.
1125 * Changed DH_check to accept parameters with order q and 2q subgroups.
1126 With order 2q subgroups the bit 0 of the private key is not secret
1127 but DH_generate_key works around that by clearing bit 0 of the
1128 private key for those. This avoids leaking bit 0 of the private key.
1132 * Significantly reduce secure memory usage by the randomness pools.
1136 * Revert the DEVRANDOM_WAIT feature for Linux systems
1138 The DEVRANDOM_WAIT feature added a select() call to wait for the
1139 /dev/random device to become readable before reading from the
1140 /dev/urandom device.
1142 It turned out that this change had negative side effects on
1143 performance which were not acceptable. After some discussion it
1144 was decided to revert this feature and leave it up to the OS
1145 resp. the platform maintainer to ensure a proper initialization
1146 during early boot time.
1148 *Matthias St. Pierre*
1150 ### Changes between 1.1.1b and 1.1.1c [28 May 2019] ###
1152 * Add build tests for C++. These are generated files that only do one
1153 thing, to include one public OpenSSL head file each. This tests that
1154 the public header files can be usefully included in a C++ application.
1156 This test isn't enabled by default. It can be enabled with the option
1157 'enable-buildtest-c++'.
1161 * Enable SHA3 pre-hashing for ECDSA and DSA.
1165 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1166 This changes the size when using the genpkey app when no size is given. It
1167 fixes an omission in earlier changes that changed all RSA, DSA and DH
1168 generation apps to use 2048 bits by default.
1172 * Reorganize the manual pages to consistently have RETURN VALUES,
1173 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1174 util/fix-doc-nits accordingly.
1176 *Paul Yang, Joshua Lock*
1178 * Add the missing accessor EVP_PKEY_get0_engine()
1182 * Have apps like 's_client' and 's_server' output the signature scheme
1183 along with other cipher suite parameters when debugging.
1187 * Make OPENSSL_config() error agnostic again.
1191 * Do the error handling in RSA decryption constant time.
1195 * Prevent over long nonces in ChaCha20-Poly1305.
1197 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1198 for every encryption operation. RFC 7539 specifies that the nonce value
1199 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1200 and front pads the nonce with 0 bytes if it is less than 12
1201 bytes. However it also incorrectly allows a nonce to be set of up to 16
1202 bytes. In this case only the last 12 bytes are significant and any
1203 additional leading bytes are ignored.
1205 It is a requirement of using this cipher that nonce values are
1206 unique. Messages encrypted using a reused nonce value are susceptible to
1207 serious confidentiality and integrity attacks. If an application changes
1208 the default nonce length to be longer than 12 bytes and then makes a
1209 change to the leading bytes of the nonce expecting the new value to be a
1210 new unique nonce then such an application could inadvertently encrypt
1211 messages with a reused nonce.
1213 Additionally the ignored bytes in a long nonce are not covered by the
1214 integrity guarantee of this cipher. Any application that relies on the
1215 integrity of these ignored leading bytes of a long nonce may be further
1216 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1217 is safe because no such use sets such a long nonce value. However user
1218 applications that use this cipher directly and set a non-default nonce
1219 length to be longer than 12 bytes may be vulnerable.
1221 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1227 * Add DEVRANDOM_WAIT feature for Linux systems
1229 On older Linux systems where the getrandom() system call is not available,
1230 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1231 Contrary to getrandom(), the /dev/urandom device will not block during
1232 early boot when the kernel CSPRNG has not been seeded yet.
1234 To mitigate this known weakness, use select() to wait for /dev/random to
1235 become readable before reading from /dev/urandom.
1237 * Ensure that SM2 only uses SM3 as digest algorithm
1241 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019] ###
1243 * Change the info callback signals for the start and end of a post-handshake
1244 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1245 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1246 confused by this and assume that a TLSv1.2 renegotiation has started. This
1247 can break KeyUpdate handling. Instead we no longer signal the start and end
1248 of a post handshake message exchange (although the messages themselves are
1249 still signalled). This could break some applications that were expecting
1250 the old signals. However without this KeyUpdate is not usable for many
1255 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018] ###
1257 * Timing vulnerability in DSA signature generation
1259 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1260 timing side channel attack. An attacker could use variations in the signing
1261 algorithm to recover the private key.
1263 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1268 * Timing vulnerability in ECDSA signature generation
1270 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1271 timing side channel attack. An attacker could use variations in the signing
1272 algorithm to recover the private key.
1274 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1279 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1280 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1281 of two gigabytes and the error handling improved.
1283 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1284 categorized as a normal bug, not a security issue, because the DRBG reseeds
1285 automatically and is fully functional even without additional randomness
1286 provided by the application.
1288 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018] ###
1290 * Add a new ClientHello callback. Provides a callback interface that gives
1291 the application the ability to adjust the nascent SSL object at the
1292 earliest stage of ClientHello processing, immediately after extensions have
1293 been collected but before they have been processed. In particular, this
1294 callback can adjust the supported TLS versions in response to the contents
1299 * Add SM2 base algorithm support.
1303 * s390x assembly pack: add (improved) hardware-support for the following
1304 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1305 aes-cfb/cfb8, aes-ecb.
1309 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1310 parameter is no longer accepted, as it leads to a corrupt table. NULL
1311 pem_str is reserved for alias entries only.
1315 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1316 step for prime curves. The new implementation is based on formulae from
1317 differential addition-and-doubling in homogeneous projective coordinates
1318 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1319 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1320 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1321 to work in projective coordinates.
1323 *Billy Bob Brumley, Nicola Tuveri*
1325 * Change generating and checking of primes so that the error rate of not
1326 being prime depends on the intended use based on the size of the input.
1327 For larger primes this will result in more rounds of Miller-Rabin.
1328 The maximal error rate for primes with more than 1080 bits is lowered
1331 *Kurt Roeckx, Annie Yousar*
1333 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1337 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1338 moving between systems, and to avoid confusion when a Windows build is
1339 done with mingw vs with MSVC. For POSIX installs, there's still a
1340 symlink or copy named 'tsget' to avoid that confusion as well.
1344 * Revert blinding in ECDSA sign and instead make problematic addition
1345 length-invariant. Switch even to fixed-length Montgomery multiplication.
1349 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1350 step for binary curves. The new implementation is based on formulae from
1351 differential addition-and-doubling in mixed Lopez-Dahab projective
1352 coordinates, modified to independently blind the operands.
1354 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1356 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1357 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1358 EC_METHODs to implement their own specialized "ladder step", to take
1359 advantage of more favorable coordinate systems or more efficient
1360 differential addition-and-doubling algorithms.
1362 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1364 * Modified the random device based seed sources to keep the relevant
1365 file descriptors open rather than reopening them on each access.
1366 This allows such sources to operate in a chroot() jail without
1367 the associated device nodes being available. This behaviour can be
1368 controlled using RAND_keep_random_devices_open().
1372 * Numerous side-channel attack mitigations have been applied. This may have
1373 performance impacts for some algorithms for the benefit of improved
1374 security. Specific changes are noted in this change log by their respective
1379 * AIX shared library support overhaul. Switch to AIX "natural" way of
1380 handling shared libraries, which means collecting shared objects of
1381 different versions and bitnesses in one common archive. This allows to
1382 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1383 doesn't affect the way 3rd party applications are linked, only how
1384 multi-version installation is managed.
1388 * Make ec_group_do_inverse_ord() more robust and available to other
1389 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1390 mitigations are applied to the fallback BN_mod_inverse().
1391 When using this function rather than BN_mod_inverse() directly, new
1392 EC cryptosystem implementations are then safer-by-default.
1396 * Add coordinate blinding for EC_POINT and implement projective
1397 coordinate blinding for generic prime curves as a countermeasure to
1398 chosen point SCA attacks.
1400 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1402 * Add blinding to ECDSA and DSA signatures to protect against side channel
1403 attacks discovered by Keegan Ryan (NCC Group).
1407 * Enforce checking in the pkeyutl command line app to ensure that the input
1408 length does not exceed the maximum supported digest length when performing
1409 a sign, verify or verifyrecover operation.
1413 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1414 I/O in combination with something like select() or poll() will hang. This
1415 can be turned off again using SSL_CTX_clear_mode().
1416 Many applications do not properly handle non-application data records, and
1417 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1418 around the problems in those applications, but can also break some.
1419 It's recommended to read the manpages about SSL_read(), SSL_write(),
1420 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1421 SSL_CTX_set_read_ahead() again.
1425 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1426 now allow empty (zero character) pass phrases.
1430 * Apply blinding to binary field modular inversion and remove patent
1431 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1435 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1436 binary and prime elliptic curves.
1440 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1441 constant time fixed point multiplication.
1445 * Revise elliptic curve scalar multiplication with timing attack
1446 defenses: ec_wNAF_mul redirects to a constant time implementation
1447 when computing fixed point and variable point multiplication (which
1448 in OpenSSL are mostly used with secret scalars in keygen, sign,
1449 ECDH derive operations).
1450 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1453 * Updated CONTRIBUTING
1457 * Updated DRBG / RAND to request nonce and additional low entropy
1458 randomness from the system.
1460 *Matthias St. Pierre*
1462 * Updated 'openssl rehash' to use OpenSSL consistent default.
1466 * Moved the load of the ssl_conf module to libcrypto, which helps
1467 loading engines that libssl uses before libssl is initialised.
1471 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1475 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1477 *Ingo Schwarze, Rich Salz*
1479 * Added output of accepting IP address and port for 'openssl s_server'
1483 * Added a new API for TLSv1.3 ciphersuites:
1484 SSL_CTX_set_ciphersuites()
1485 SSL_set_ciphersuites()
1489 * Memory allocation failures consistently add an error to the error
1494 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1495 in libcrypto when run as setuid/setgid.
1499 * Load any config file by default when libssl is used.
1503 * Added new public header file <openssl/rand_drbg.h> and documentation
1504 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1506 *Matthias St. Pierre*
1508 * QNX support removed (cannot find contributors to get their approval
1509 for the license change).
1513 * TLSv1.3 replay protection for early data has been implemented. See the
1514 SSL_read_early_data() man page for further details.
1518 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1519 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1520 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1521 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1522 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1523 configuration has been separated out. See the ciphers man page or the
1524 SSL_CTX_set_ciphersuites() man page for more information.
1528 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1529 in responder mode now supports the new "-multi" option, which
1530 spawns the specified number of child processes to handle OCSP
1531 requests. The "-timeout" option now also limits the OCSP
1532 responder's patience to wait to receive the full client request
1533 on a newly accepted connection. Child processes are respawned
1534 as needed, and the CA index file is automatically reloaded
1535 when changed. This makes it possible to run the "ocsp" responder
1536 as a long-running service, making the OpenSSL CA somewhat more
1537 feature-complete. In this mode, most diagnostic messages logged
1538 after entering the event loop are logged via syslog(3) rather than
1543 * Added support for X448 and Ed448. Heavily based on original work by
1548 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1549 objects loaded. This adds the functions OSSL_STORE_expect() and
1550 OSSL_STORE_find() as well as needed tools to construct searches and
1551 get the search data out of them.
1555 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1556 version of OpenSSL should review their configuration settings to ensure
1557 that they are still appropriate for TLSv1.3. For further information see:
1558 https://wiki.openssl.org/index.php/TLS1.3
1562 * Grand redesign of the OpenSSL random generator
1564 The default RAND method now utilizes an AES-CTR DRBG according to
1565 NIST standard SP 800-90Ar1. The new random generator is essentially
1566 a port of the default random generator from the OpenSSL FIPS 2.0
1567 object module. It is a hybrid deterministic random bit generator
1568 using an AES-CTR bit stream and which seeds and reseeds itself
1569 automatically using trusted system entropy sources.
1571 Some of its new features are:
1572 - Support for multiple DRBG instances with seed chaining.
1573 - The default RAND method makes use of a DRBG.
1574 - There is a public and private DRBG instance.
1575 - The DRBG instances are fork-safe.
1576 - Keep all global DRBG instances on the secure heap if it is enabled.
1577 - The public and private DRBG instance are per thread for lock free
1580 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1582 * Changed Configure so it only says what it does and doesn't dump
1583 so much data. Instead, ./configdata.pm should be used as a script
1584 to display all sorts of configuration data.
1588 * Added processing of "make variables" to Configure.
1592 * Added SHA512/224 and SHA512/256 algorithm support.
1596 * The last traces of Netware support, first removed in 1.1.0, have
1601 * Get rid of Makefile.shared, and in the process, make the processing
1602 of certain files (rc.obj, or the .def/.map/.opt files produced from
1603 the ordinal files) more visible and hopefully easier to trace and
1604 debug (or make silent).
1608 * Make it possible to have environment variable assignments as
1609 arguments to config / Configure.
1613 * Add multi-prime RSA (RFC 8017) support.
1617 * Add SM3 implemented according to GB/T 32905-2016
1618 * Jack Lloyd <jack.lloyd@ribose.com>,
1619 Ronald Tse <ronald.tse@ribose.com>,
1620 Erick Borsboom <erick.borsboom@ribose.com> *
1622 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1623 as documented in RFC6066.
1624 Based on a patch from Tomasz Moń
1626 *Filipe Raimundo da Silva*
1628 * Add SM4 implemented according to GB/T 32907-2016.
1629 * Jack Lloyd <jack.lloyd@ribose.com>,
1630 Ronald Tse <ronald.tse@ribose.com>,
1631 Erick Borsboom <erick.borsboom@ribose.com> *
1633 * Reimplement -newreq-nodes and ERR_error_string_n; the
1634 original author does not agree with the license change.
1638 * Add ARIA AEAD TLS support.
1642 * Some macro definitions to support VS6 have been removed. Visual
1643 Studio 6 has not worked since 1.1.0
1647 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1648 without clearing the errors.
1652 * Add "atfork" functions. If building on a system that without
1653 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1654 requirements. The RAND facility now uses/requires this.
1662 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1663 not possible to disable entirely. However, it's still possible to
1664 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1667 To disable, configure with 'no-ui-console'. 'no-ui' is still
1668 possible to use as an alias. Check at compile time with the
1669 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1670 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1674 * Add a STORE module, which implements a uniform and URI based reader of
1675 stores that can contain keys, certificates, CRLs and numerous other
1676 objects. The main API is loosely based on a few stdio functions,
1677 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1678 OSSL_STORE_error and OSSL_STORE_close.
1679 The implementation uses backends called "loaders" to implement arbitrary
1680 URI schemes. There is one built in "loader" for the 'file' scheme.
1684 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1685 then adjusted to work on FreeBSD 8.4 as well.
1686 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1687 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1691 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1692 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1693 error code calls like this:
1695 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1697 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1698 that can be encoded in C. For the foreseeable future, this will only
1701 *Richard Levitte and Tim Hudson*
1703 * Removed BSD cryptodev engine.
1707 * Add a build target 'build_all_generated', to build all generated files
1708 and only that. This can be used to prepare everything that requires
1709 things like perl for a system that lacks perl and then move everything
1710 to that system and do the rest of the build there.
1714 * In the UI interface, make it possible to duplicate the user data. This
1715 can be used by engines that need to retain the data for a longer time
1716 than just the call where this user data is passed.
1720 * Ignore the '-named_curve auto' value for compatibility of applications
1723 *Tomas Mraz <tmraz@fedoraproject.org>*
1725 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1726 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1727 alerts across multiple records (some of which could be empty). In practice
1728 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1729 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1730 support this at all. Supporting it adds significant complexity to the
1731 record layer, and its removal is unlikely to cause interoperability
1736 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1737 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1738 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1743 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1744 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1746 *Richard Levitte, Andy Polyakov*
1748 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1753 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1754 platform rather than 'mingw'.
1758 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1759 success if they are asked to add an object which already exists
1760 in the store. This change cascades to other functions which load
1761 certificates and CRLs.
1765 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1766 facilitate stack unwinding even from assembly subroutines.
1770 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1771 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1775 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1776 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1777 which is the minimum version we support.
1781 * Certificate time validation (X509_cmp_time) enforces stricter
1782 compliance with RFC 5280. Fractional seconds and timezone offsets
1783 are no longer allowed.
1787 * Add support for ARIA
1791 * s_client will now send the Server Name Indication (SNI) extension by
1792 default unless the new "-noservername" option is used. The server name is
1793 based on the host provided to the "-connect" option unless overridden by
1794 using "-servername".
1798 * Add support for SipHash
1802 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1803 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1804 prevent issues where no progress is being made and the peer continually
1805 sends unrecognised record types, using up resources processing them.
1809 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1810 using the algorithm defined in
1811 https://www.akkadia.org/drepper/SHA-crypt.txt
1815 * Heartbeat support has been removed; the ABI is changed for now.
1817 *Richard Levitte, Rich Salz*
1819 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1823 * The RSA "null" method, which was partially supported to avoid patent
1824 issues, has been replaced to always returns NULL.
1832 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019] ###
1834 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1835 used even when parsing explicit parameters, when loading a serialized key
1836 or calling `EC_GROUP_new_from_ecpkparameters()`/
1837 `EC_GROUP_new_from_ecparameters()`.
1838 This prevents bypass of security hardening and performance gains,
1839 especially for curves with specialized EC_METHODs.
1840 By default, if a key encoded with explicit parameters is loaded and later
1841 serialized, the output is still encoded with explicit parameters, even if
1842 internally a "named" EC_GROUP is used for computation.
1846 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1847 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1848 NULL. After this change, only the cofactor parameter can be NULL. It also
1849 does some minimal sanity checks on the passed order.
1854 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1855 An attack is simple, if the first CMS_recipientInfo is valid but the
1856 second CMS_recipientInfo is chosen ciphertext. If the second
1857 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1858 encryption key will be replaced by garbage, and the message cannot be
1859 decoded, but if the RSA decryption fails, the correct encryption key is
1860 used and the recipient will not notice the attack.
1861 As a work around for this potential attack the length of the decrypted
1862 key must be equal to the cipher default key length, in case the
1863 certifiate is not given and all recipientInfo are tried out.
1864 The old behaviour can be re-enabled in the CMS code by setting the
1865 CMS_DEBUG_DECRYPT flag.
1870 * Use Windows installation paths in the mingw builds
1872 Mingw isn't a POSIX environment per se, which means that Windows
1873 paths should be used for installation.
1878 ### Changes between 1.1.0j and 1.1.0k [28 May 2019] ###
1880 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1881 This changes the size when using the genpkey app when no size is given. It
1882 fixes an omission in earlier changes that changed all RSA, DSA and DH
1883 generation apps to use 2048 bits by default.
1887 * Prevent over long nonces in ChaCha20-Poly1305.
1889 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1890 for every encryption operation. RFC 7539 specifies that the nonce value
1891 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1892 and front pads the nonce with 0 bytes if it is less than 12
1893 bytes. However it also incorrectly allows a nonce to be set of up to 16
1894 bytes. In this case only the last 12 bytes are significant and any
1895 additional leading bytes are ignored.
1897 It is a requirement of using this cipher that nonce values are
1898 unique. Messages encrypted using a reused nonce value are susceptible to
1899 serious confidentiality and integrity attacks. If an application changes
1900 the default nonce length to be longer than 12 bytes and then makes a
1901 change to the leading bytes of the nonce expecting the new value to be a
1902 new unique nonce then such an application could inadvertently encrypt
1903 messages with a reused nonce.
1905 Additionally the ignored bytes in a long nonce are not covered by the
1906 integrity guarantee of this cipher. Any application that relies on the
1907 integrity of these ignored leading bytes of a long nonce may be further
1908 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1909 is safe because no such use sets such a long nonce value. However user
1910 applications that use this cipher directly and set a non-default nonce
1911 length to be longer than 12 bytes may be vulnerable.
1913 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1919 * Added SCA hardening for modular field inversion in EC_GROUP through
1920 a new dedicated field_inv() pointer in EC_METHOD.
1921 This also addresses a leakage affecting conversions from projective
1922 to affine coordinates.
1924 *Billy Bob Brumley, Nicola Tuveri*
1926 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1927 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1931 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1935 * Remove the 'dist' target and add a tarball building script. The
1936 'dist' target has fallen out of use, and it shouldn't be
1937 necessary to configure just to create a source distribution.
1941 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018] ###
1943 * Timing vulnerability in DSA signature generation
1945 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1946 timing side channel attack. An attacker could use variations in the signing
1947 algorithm to recover the private key.
1949 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1954 * Timing vulnerability in ECDSA signature generation
1956 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1957 timing side channel attack. An attacker could use variations in the signing
1958 algorithm to recover the private key.
1960 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1965 * Add coordinate blinding for EC_POINT and implement projective
1966 coordinate blinding for generic prime curves as a countermeasure to
1967 chosen point SCA attacks.
1969 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1971 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018] ###
1973 * Client DoS due to large DH parameter
1975 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1976 malicious server can send a very large prime value to the client. This will
1977 cause the client to spend an unreasonably long period of time generating a
1978 key for this prime resulting in a hang until the client has finished. This
1979 could be exploited in a Denial Of Service attack.
1981 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1986 * Cache timing vulnerability in RSA Key Generation
1988 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1989 a cache timing side channel attack. An attacker with sufficient access to
1990 mount cache timing attacks during the RSA key generation process could
1991 recover the private key.
1993 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1994 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1999 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2000 parameter is no longer accepted, as it leads to a corrupt table. NULL
2001 pem_str is reserved for alias entries only.
2005 * Revert blinding in ECDSA sign and instead make problematic addition
2006 length-invariant. Switch even to fixed-length Montgomery multiplication.
2010 * Change generating and checking of primes so that the error rate of not
2011 being prime depends on the intended use based on the size of the input.
2012 For larger primes this will result in more rounds of Miller-Rabin.
2013 The maximal error rate for primes with more than 1080 bits is lowered
2016 *Kurt Roeckx, Annie Yousar*
2018 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2022 * Add blinding to ECDSA and DSA signatures to protect against side channel
2023 attacks discovered by Keegan Ryan (NCC Group).
2027 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2028 now allow empty (zero character) pass phrases.
2032 * Certificate time validation (X509_cmp_time) enforces stricter
2033 compliance with RFC 5280. Fractional seconds and timezone offsets
2034 are no longer allowed.
2038 * Fixed a text canonicalisation bug in CMS
2040 Where a CMS detached signature is used with text content the text goes
2041 through a canonicalisation process first prior to signing or verifying a
2042 signature. This process strips trailing space at the end of lines, converts
2043 line terminators to CRLF and removes additional trailing line terminators
2044 at the end of a file. A bug in the canonicalisation process meant that
2045 some characters, such as form-feed, were incorrectly treated as whitespace
2046 and removed. This is contrary to the specification (RFC5485). This fix
2047 could mean that detached text data signed with an earlier version of
2048 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2049 signed with a fixed OpenSSL may fail to verify with an earlier version of
2050 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2051 and use the "-binary" flag (for the "cms" command line application) or set
2052 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2056 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018] ###
2058 * Constructed ASN.1 types with a recursive definition could exceed the stack
2060 Constructed ASN.1 types with a recursive definition (such as can be found
2061 in PKCS7) could eventually exceed the stack given malicious input with
2062 excessive recursion. This could result in a Denial Of Service attack. There
2063 are no such structures used within SSL/TLS that come from untrusted sources
2064 so this is considered safe.
2066 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2072 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2074 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2075 effectively reduced to only comparing the least significant bit of each
2076 byte. This allows an attacker to forge messages that would be considered as
2077 authenticated in an amount of tries lower than that guaranteed by the
2078 security claims of the scheme. The module can only be compiled by the
2079 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2081 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2087 * Add a build target 'build_all_generated', to build all generated files
2088 and only that. This can be used to prepare everything that requires
2089 things like perl for a system that lacks perl and then move everything
2090 to that system and do the rest of the build there.
2094 * Backport SSL_OP_NO_RENGOTIATION
2096 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2097 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2098 changes this is no longer possible in 1.1.0. Therefore the new
2099 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2100 1.1.0 to provide equivalent functionality.
2102 Note that if an application built against 1.1.0h headers (or above) is run
2103 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2104 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2108 * Removed the OS390-Unix config target. It relied on a script that doesn't
2113 * rsaz_1024_mul_avx2 overflow bug on x86_64
2115 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2116 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2117 Analysis suggests that attacks against RSA and DSA as a result of this
2118 defect would be very difficult to perform and are not believed likely.
2119 Attacks against DH1024 are considered just feasible, because most of the
2120 work necessary to deduce information about a private key may be performed
2121 offline. The amount of resources required for such an attack would be
2122 significant. However, for an attack on TLS to be meaningful, the server
2123 would have to share the DH1024 private key among multiple clients, which is
2124 no longer an option since CVE-2016-0701.
2126 This only affects processors that support the AVX2 but not ADX extensions
2127 like Intel Haswell (4th generation).
2129 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2130 was originally found via the OSS-Fuzz project.
2135 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017] ###
2137 * bn_sqrx8x_internal carry bug on x86_64
2139 There is a carry propagating bug in the x86_64 Montgomery squaring
2140 procedure. No EC algorithms are affected. Analysis suggests that attacks
2141 against RSA and DSA as a result of this defect would be very difficult to
2142 perform and are not believed likely. Attacks against DH are considered just
2143 feasible (although very difficult) because most of the work necessary to
2144 deduce information about a private key may be performed offline. The amount
2145 of resources required for such an attack would be very significant and
2146 likely only accessible to a limited number of attackers. An attacker would
2147 additionally need online access to an unpatched system using the target
2148 private key in a scenario with persistent DH parameters and a private
2149 key that is shared between multiple clients.
2151 This only affects processors that support the BMI1, BMI2 and ADX extensions
2152 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2154 This issue was reported to OpenSSL by the OSS-Fuzz project.
2159 * Malformed X.509 IPAddressFamily could cause OOB read
2161 If an X.509 certificate has a malformed IPAddressFamily extension,
2162 OpenSSL could do a one-byte buffer overread. The most likely result
2163 would be an erroneous display of the certificate in text format.
2165 This issue was reported to OpenSSL by the OSS-Fuzz project.
2170 ### Changes between 1.1.0e and 1.1.0f [25 May 2017] ###
2172 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2173 platform rather than 'mingw'.
2177 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2178 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2179 which is the minimum version we support.
2183 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017] ###
2185 * Encrypt-Then-Mac renegotiation crash
2187 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2188 negotiated where it was not in the original handshake (or vice-versa) then
2189 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2190 and servers are affected.
2192 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2197 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017] ###
2199 * Truncated packet could crash via OOB read
2201 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2202 cipher is being used, then a truncated packet can cause that host to
2203 perform an out-of-bounds read, usually resulting in a crash.
2205 This issue was reported to OpenSSL by Robert Święcki of Google.
2210 * Bad (EC)DHE parameters cause a client crash
2212 If a malicious server supplies bad parameters for a DHE or ECDHE key
2213 exchange then this can result in the client attempting to dereference a
2214 NULL pointer leading to a client crash. This could be exploited in a Denial
2217 This issue was reported to OpenSSL by Guido Vranken.
2222 * BN_mod_exp may produce incorrect results on x86_64
2224 There is a carry propagating bug in the x86_64 Montgomery squaring
2225 procedure. No EC algorithms are affected. Analysis suggests that attacks
2226 against RSA and DSA as a result of this defect would be very difficult to
2227 perform and are not believed likely. Attacks against DH are considered just
2228 feasible (although very difficult) because most of the work necessary to
2229 deduce information about a private key may be performed offline. The amount
2230 of resources required for such an attack would be very significant and
2231 likely only accessible to a limited number of attackers. An attacker would
2232 additionally need online access to an unpatched system using the target
2233 private key in a scenario with persistent DH parameters and a private
2234 key that is shared between multiple clients. For example this can occur by
2235 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2236 similar to CVE-2015-3193 but must be treated as a separate problem.
2238 This issue was reported to OpenSSL by the OSS-Fuzz project.
2243 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016] ###
2245 * ChaCha20/Poly1305 heap-buffer-overflow
2247 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
2248 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2249 crash. This issue is not considered to be exploitable beyond a DoS.
2251 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2256 * CMS Null dereference
2258 Applications parsing invalid CMS structures can crash with a NULL pointer
2259 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2260 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2261 structure callback if an attempt is made to free certain invalid encodings.
2262 Only CHOICE structures using a callback which do not handle NULL value are
2265 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2270 * Montgomery multiplication may produce incorrect results
2272 There is a carry propagating bug in the Broadwell-specific Montgomery
2273 multiplication procedure that handles input lengths divisible by, but
2274 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2275 and DH private keys are impossible. This is because the subroutine in
2276 question is not used in operations with the private key itself and an input
2277 of the attacker's direct choice. Otherwise the bug can manifest itself as
2278 transient authentication and key negotiation failures or reproducible
2279 erroneous outcome of public-key operations with specially crafted input.
2280 Among EC algorithms only Brainpool P-512 curves are affected and one
2281 presumably can attack ECDH key negotiation. Impact was not analyzed in
2282 detail, because pre-requisites for attack are considered unlikely. Namely
2283 multiple clients have to choose the curve in question and the server has to
2284 share the private key among them, neither of which is default behaviour.
2285 Even then only clients that chose the curve will be affected.
2287 This issue was publicly reported as transient failures and was not
2288 initially recognized as a security issue. Thanks to Richard Morgan for
2289 providing reproducible case.
2294 * Removed automatic addition of RPATH in shared libraries and executables,
2295 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2299 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016] ###
2301 * Fix Use After Free for large message sizes
2303 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2304 message larger than approx 16k is received then the underlying buffer to
2305 store the incoming message is reallocated and moved. Unfortunately a
2306 dangling pointer to the old location is left which results in an attempt to
2307 write to the previously freed location. This is likely to result in a
2308 crash, however it could potentially lead to execution of arbitrary code.
2310 This issue only affects OpenSSL 1.1.0a.
2312 This issue was reported to OpenSSL by Robert Święcki.
2317 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016] ###
2319 * OCSP Status Request extension unbounded memory growth
2321 A malicious client can send an excessively large OCSP Status Request
2322 extension. If that client continually requests renegotiation, sending a
2323 large OCSP Status Request extension each time, then there will be unbounded
2324 memory growth on the server. This will eventually lead to a Denial Of
2325 Service attack through memory exhaustion. Servers with a default
2326 configuration are vulnerable even if they do not support OCSP. Builds using
2327 the "no-ocsp" build time option are not affected.
2329 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2334 * SSL_peek() hang on empty record
2336 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2337 sends an empty record. This could be exploited by a malicious peer in a
2338 Denial Of Service attack.
2340 This issue was reported to OpenSSL by Alex Gaynor.
2345 * Excessive allocation of memory in tls_get_message_header() and
2346 dtls1_preprocess_fragment()
2348 A (D)TLS message includes 3 bytes for its length in the header for the
2349 message. This would allow for messages up to 16Mb in length. Messages of
2350 this length are excessive and OpenSSL includes a check to ensure that a
2351 peer is sending reasonably sized messages in order to avoid too much memory
2352 being consumed to service a connection. A flaw in the logic of version
2353 1.1.0 means that memory for the message is allocated too early, prior to
2354 the excessive message length check. Due to way memory is allocated in
2355 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2356 to service a connection. This could lead to a Denial of Service through
2357 memory exhaustion. However, the excessive message length check still takes
2358 place, and this would cause the connection to immediately fail. Assuming
2359 that the application calls SSL_free() on the failed connection in a timely
2360 manner then the 21Mb of allocated memory will then be immediately freed
2361 again. Therefore the excessive memory allocation will be transitory in
2362 nature. This then means that there is only a security impact if:
2364 1) The application does not call SSL_free() in a timely manner in the event
2365 that the connection fails
2367 2) The application is working in a constrained environment where there is
2368 very little free memory
2370 3) The attacker initiates multiple connection attempts such that there are
2371 multiple connections in a state where memory has been allocated for the
2372 connection; SSL_free() has not yet been called; and there is insufficient
2373 memory to service the multiple requests.
2375 Except in the instance of (1) above any Denial Of Service is likely to be
2376 transitory because as soon as the connection fails the memory is
2377 subsequently freed again in the SSL_free() call. However there is an
2378 increased risk during this period of application crashes due to the lack of
2379 memory - which would then mean a more serious Denial of Service.
2381 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2382 (CVE-2016-6307 and CVE-2016-6308)
2386 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2387 had to be removed. Primary reason is that vendor assembler can't
2388 assemble our modules with -KPIC flag. As result it, assembly
2389 support, was not even available as option. But its lack means
2390 lack of side-channel resistant code, which is incompatible with
2391 security by todays standards. Fortunately gcc is readily available
2392 prepackaged option, which we firmly point at...
2396 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016] ###
2398 * Windows command-line tool supports UTF-8 opt-in option for arguments
2399 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2400 (to any value) allows Windows user to access PKCS#12 file generated
2401 with Windows CryptoAPI and protected with non-ASCII password, as well
2402 as files generated under UTF-8 locale on Linux also protected with
2407 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2408 have been disabled by default and removed from DEFAULT, just like RC4.
2409 See the RC4 item below to re-enable both.
2413 * The method for finding the storage location for the Windows RAND seed file
2414 has changed. First we check %RANDFILE%. If that is not set then we check
2415 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2416 all else fails we fall back to C:\.
2420 * The EVP_EncryptUpdate() function has had its return type changed from void
2421 to int. A return of 0 indicates and error while a return of 1 indicates
2426 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2427 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2428 off the constant time implementation for RSA, DSA and DH have been made
2429 no-ops and deprecated.
2433 * Windows RAND implementation was simplified to only get entropy by
2434 calling CryptGenRandom(). Various other RAND-related tickets
2437 *Joseph Wylie Yandle, Rich Salz*
2439 * The stack and lhash API's were renamed to start with OPENSSL_SK_
2440 and OPENSSL_LH_, respectively. The old names are available
2441 with API compatibility. They new names are now completely documented.
2445 * Unify TYPE_up_ref(obj) methods signature.
2446 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2447 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2448 int (instead of void) like all others TYPE_up_ref() methods.
2449 So now these methods also check the return value of CRYPTO_atomic_add(),
2450 and the validity of object reference counter.
2452 *fdasilvayy@gmail.com*
2454 * With Windows Visual Studio builds, the .pdb files are installed
2455 alongside the installed libraries and executables. For a static
2456 library installation, ossl_static.pdb is the associate compiler
2457 generated .pdb file to be used when linking programs.
2461 * Remove openssl.spec. Packaging files belong with the packagers.
2465 * Automatic Darwin/OSX configuration has had a refresh, it will now
2466 recognise x86_64 architectures automatically. You can still decide
2467 to build for a different bitness with the environment variable
2468 KERNEL_BITS (can be 32 or 64), for example:
2470 KERNEL_BITS=32 ./config
2474 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2475 256 bit AES and HMAC with SHA256.
2479 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2483 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2487 * To enable users to have their own config files and build file templates,
2488 Configure looks in the directory indicated by the environment variable
2489 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2490 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2491 name and is used as is.
2495 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2496 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2497 X509_CERT_FILE_CTX was removed.
2501 * "shared" builds are now the default. To create only static libraries use
2502 the "no-shared" Configure option.
2506 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2507 All of these option have not worked for some while and are fundamental
2512 * Make various cleanup routines no-ops and mark them as deprecated. Most
2513 global cleanup functions are no longer required because they are handled
2514 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2515 Explicitly de-initing can cause problems (e.g. where a library that uses
2516 OpenSSL de-inits, but an application is still using it). The affected
2517 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2518 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2519 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2520 COMP_zlib_cleanup().
2524 * --strict-warnings no longer enables runtime debugging options
2525 such as REF_DEBUG. Instead, debug options are automatically
2526 enabled with '--debug' builds.
2528 *Andy Polyakov, Emilia Käsper*
2530 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2531 have been moved out of the public header files. New functions for managing
2532 these have been added.
2536 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2537 objects have been moved out of the public header files. New
2538 functions for managing these have been added.
2542 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2543 have been moved out of the public header files. New functions for managing
2544 these have been added.
2548 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2549 moved out of the public header files. New functions for managing these
2554 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2558 * Removed the mk1mf build scripts.
2562 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2563 it is always safe to #include a header now.
2567 * Removed the aged BC-32 config and all its supporting scripts
2571 * Removed support for Ultrix, Netware, and OS/2.
2575 * Add support for HKDF.
2577 *Alessandro Ghedini*
2579 * Add support for blake2b and blake2s
2583 * Added support for "pipelining". Ciphers that have the
2584 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2585 encryptions/decryptions simultaneously. There are currently no built-in
2586 ciphers with this property but the expectation is that engines will be able
2587 to offer it to significantly improve throughput. Support has been extended
2588 into libssl so that multiple records for a single connection can be
2589 processed in one go (for >=TLS 1.1).
2593 * Added the AFALG engine. This is an async capable engine which is able to
2594 offload work to the Linux kernel. In this initial version it only supports
2595 AES128-CBC. The kernel must be version 4.1.0 or greater.
2599 * OpenSSL now uses a new threading API. It is no longer necessary to
2600 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2601 are two supported threading models: pthreads and windows threads. It is
2602 also possible to configure OpenSSL at compile time for "no-threads". The
2603 old threading API should no longer be used. The functions have been
2604 replaced with "no-op" compatibility macros.
2606 *Alessandro Ghedini, Matt Caswell*
2608 * Modify behavior of ALPN to invoke callback after SNI/servername
2609 callback, such that updates to the SSL_CTX affect ALPN.
2613 * Add SSL_CIPHER queries for authentication and key-exchange.
2617 * Changes to the DEFAULT cipherlist:
2618 - Prefer (EC)DHE handshakes over plain RSA.
2619 - Prefer AEAD ciphers over legacy ciphers.
2620 - Prefer ECDSA over RSA when both certificates are available.
2621 - Prefer TLSv1.2 ciphers/PRF.
2622 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2627 * Change the ECC default curve list to be this, in order: x25519,
2628 secp256r1, secp521r1, secp384r1.
2632 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2633 disabled by default. They can be re-enabled using the
2634 enable-weak-ssl-ciphers option to Configure.
2638 * If the server has ALPN configured, but supports no protocols that the
2639 client advertises, send a fatal "no_application_protocol" alert.
2640 This behaviour is SHALL in RFC 7301, though it isn't universally
2641 implemented by other servers.
2645 * Add X25519 support.
2646 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2647 for public and private key encoding using the format documented in
2648 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2649 key generation and key derivation.
2651 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2656 * Deprecate SRP_VBASE_get_by_user.
2657 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2658 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2659 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2660 seed, even if the seed is configured.
2662 Users should use SRP_VBASE_get1_by_user instead. Note that in
2663 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2664 also that even though configuring the SRP seed attempts to hide
2665 invalid usernames by continuing the handshake with fake
2666 credentials, this behaviour is not constant time and no strong
2667 guarantees are made that the handshake is indistinguishable from
2668 that of a valid user.
2672 * Configuration change; it's now possible to build dynamic engines
2673 without having to build shared libraries and vice versa. This
2674 only applies to the engines in engines/, those in crypto/engine/
2675 will always be built into libcrypto (i.e. "static").
2677 Building dynamic engines is enabled by default; to disable, use
2678 the configuration option "disable-dynamic-engine".
2680 The only requirements for building dynamic engines are the
2681 presence of the DSO module and building with position independent
2682 code, so they will also automatically be disabled if configuring
2683 with "disable-dso" or "disable-pic".
2685 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2686 are also taken away from openssl/opensslconf.h, as they are
2691 * Configuration change; if there is a known flag to compile
2692 position independent code, it will always be applied on the
2693 libcrypto and libssl object files, and never on the application
2694 object files. This means other libraries that use routines from
2695 libcrypto / libssl can be made into shared libraries regardless
2696 of how OpenSSL was configured.
2698 If this isn't desirable, the configuration options "disable-pic"
2699 or "no-pic" can be used to disable the use of PIC. This will
2700 also disable building shared libraries and dynamic engines.
2704 * Removed JPAKE code. It was experimental and has no wide use.
2708 * The INSTALL_PREFIX Makefile variable has been renamed to
2709 DESTDIR. That makes for less confusion on what this variable
2710 is for. Also, the configuration option --install_prefix is
2715 * Heartbeat for TLS has been removed and is disabled by default
2716 for DTLS; configure with enable-heartbeats. Code that uses the
2717 old #define's might need to be updated.
2719 *Emilia Käsper, Rich Salz*
2721 * Rename REF_CHECK to REF_DEBUG.
2725 * New "unified" build system
2727 The "unified" build system is aimed to be a common system for all
2728 platforms we support. With it comes new support for VMS.
2730 This system builds supports building in a different directory tree
2731 than the source tree. It produces one Makefile (for unix family
2732 or lookalikes), or one descrip.mms (for VMS).
2734 The source of information to make the Makefile / descrip.mms is
2735 small files called 'build.info', holding the necessary
2736 information for each directory with source to compile, and a
2737 template in Configurations, like unix-Makefile.tmpl or
2740 With this change, the library names were also renamed on Windows
2741 and on VMS. They now have names that are closer to the standard
2742 on Unix, and include the major version number, and in certain
2743 cases, the architecture they are built for. See "Notes on shared
2744 libraries" in INSTALL.
2746 We rely heavily on the perl module Text::Template.
2750 * Added support for auto-initialisation and de-initialisation of the library.
2751 OpenSSL no longer requires explicit init or deinit routines to be called,
2752 except in certain circumstances. See the OPENSSL_init_crypto() and
2753 OPENSSL_init_ssl() man pages for further information.
2757 * The arguments to the DTLSv1_listen function have changed. Specifically the
2758 "peer" argument is now expected to be a BIO_ADDR object.
2760 * Rewrite of BIO networking library. The BIO library lacked consistent
2761 support of IPv6, and adding it required some more extensive
2762 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2763 which hold all types of addresses and chains of address information.
2764 It also introduces a new API, with functions like BIO_socket,
2765 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2766 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2767 have been adapted accordingly.
2771 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2776 * CRIME protection: disable compression by default, even if OpenSSL is
2777 compiled with zlib enabled. Applications can still enable compression
2778 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2779 using the SSL_CONF library to configure compression.
2783 * The signature of the session callback configured with
2784 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2785 was explicitly marked as 'const unsigned char*' instead of
2790 * Always DPURIFY. Remove the use of uninitialized memory in the
2791 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2795 * Removed many obsolete configuration items, including
2796 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2797 MD2_CHAR, MD2_INT, MD2_LONG
2799 IDEA_SHORT, IDEA_LONG
2800 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2802 *Rich Salz, with advice from Andy Polyakov*
2804 * Many BN internals have been moved to an internal header file.
2806 *Rich Salz with help from Andy Polyakov*
2808 * Configuration and writing out the results from it has changed.
2809 Files such as Makefile include/openssl/opensslconf.h and are now
2810 produced through general templates, such as Makefile.in and
2811 crypto/opensslconf.h.in and some help from the perl module
2814 Also, the center of configuration information is no longer
2815 Makefile. Instead, Configure produces a perl module in
2816 configdata.pm which holds most of the config data (in the hash
2817 table %config), the target data that comes from the target
2818 configuration in one of the Configurations/*.conf files (in
2823 * To clarify their intended purposes, the Configure options
2824 --prefix and --openssldir change their semantics, and become more
2825 straightforward and less interdependent.
2827 --prefix shall be used exclusively to give the location INSTALLTOP
2828 where programs, scripts, libraries, include files and manuals are
2829 going to be installed. The default is now /usr/local.
2831 --openssldir shall be used exclusively to give the default
2832 location OPENSSLDIR where certificates, private keys, CRLs are
2833 managed. This is also where the default openssl.cnf gets
2835 If the directory given with this option is a relative path, the
2836 values of both the --prefix value and the --openssldir value will
2837 be combined to become OPENSSLDIR.
2838 The default for --openssldir is INSTALLTOP/ssl.
2840 Anyone who uses --openssldir to specify where OpenSSL is to be
2841 installed MUST change to use --prefix instead.
2845 * The GOST engine was out of date and therefore it has been removed. An up
2846 to date GOST engine is now being maintained in an external repository.
2847 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
2848 support for GOST ciphersuites (these are only activated if a GOST engine
2853 * EGD is no longer supported by default; use enable-egd when
2856 *Ben Kaduk and Rich Salz*
2858 * The distribution now has Makefile.in files, which are used to
2859 create Makefile's when Configure is run. *Configure must be run
2860 before trying to build now.*
2864 * The return value for SSL_CIPHER_description() for error conditions
2869 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2871 Obtaining and performing DNSSEC validation of TLSA records is
2872 the application's responsibility. The application provides
2873 the TLSA records of its choice to OpenSSL, and these are then
2874 used to authenticate the peer.
2876 The TLSA records need not even come from DNS. They can, for
2877 example, be used to implement local end-entity certificate or
2878 trust-anchor "pinning", where the "pin" data takes the form
2879 of TLSA records, which can augment or replace verification
2880 based on the usual WebPKI public certification authorities.
2884 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2885 continues to support deprecated interfaces in default builds.
2886 However, applications are strongly advised to compile their
2887 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2888 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2889 or the 1.1.0 releases.
2891 In environments in which all applications have been ported to
2892 not use any deprecated interfaces OpenSSL's Configure script
2893 should be used with the --api=1.1.0 option to entirely remove
2894 support for the deprecated features from the library and
2895 unconditionally disable them in the installed headers.
2896 Essentially the same effect can be achieved with the "no-deprecated"
2897 argument to Configure, except that this will always restrict
2898 the build to just the latest API, rather than a fixed API
2901 As applications are ported to future revisions of the API,
2902 they should update their compile-time OPENSSL_API_COMPAT define
2903 accordingly, but in most cases should be able to continue to
2904 compile with later releases.
2906 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2907 0x10000000L and 0x00908000L, respectively. However those
2908 versions did not support the OPENSSL_API_COMPAT feature, and
2909 so applications are not typically tested for explicit support
2910 of just the undeprecated features of either release.
2914 * Add support for setting the minimum and maximum supported protocol.
2915 It can bet set via the SSL_set_min_proto_version() and
2916 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2917 MaxProtocol. It's recommended to use the new APIs to disable
2918 protocols instead of disabling individual protocols using
2919 SSL_set_options() or SSL_CONF's Protocol. This change also
2920 removes support for disabling TLS 1.2 in the OpenSSL TLS
2921 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2925 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2929 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2930 and integrates ECDSA and ECDH functionality into EC. Implementations can
2931 now redirect key generation and no longer need to convert to or from
2934 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2935 include the ec.h header file instead.
2939 * Remove support for all 40 and 56 bit ciphers. This includes all the export
2940 ciphers who are no longer supported and drops support the ephemeral RSA key
2941 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2945 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2946 opaque. For HMAC_CTX, the following constructors and destructors
2949 HMAC_CTX *HMAC_CTX_new(void);
2950 void HMAC_CTX_free(HMAC_CTX *ctx);
2952 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2953 destroy such methods has been added. See EVP_MD_meth_new(3) and
2954 EVP_CIPHER_meth_new(3) for documentation.
2957 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2958 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2959 EVP_MD_CTX_reset() should be called instead to reinitialise
2960 an already created structure.
2961 2) For consistency with the majority of our object creators and
2962 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2963 EVP_MD_CTX_(new|free). The old names are retained as macros
2964 for deprecated builds.
2968 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
2969 cryptographic operations to be performed asynchronously as long as an
2970 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2971 further details. Libssl has also had this capability integrated with the
2972 introduction of the new mode SSL_MODE_ASYNC and associated error
2973 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2974 pages. This work was developed in partnership with Intel Corp.
2978 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2979 always enabled now. If you want to disable the support you should
2980 exclude it using the list of supported ciphers. This also means that the
2981 "-no_ecdhe" option has been removed from s_server.
2985 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2986 SSL_{CTX_}set1_curves() which can set a list.
2990 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2991 curve you want to support using SSL_{CTX_}set1_curves().
2995 * State machine rewrite. The state machine code has been significantly
2996 refactored in order to remove much duplication of code and solve issues
2997 with the old code (see ssl/statem/README for further details). This change
2998 does have some associated API changes. Notably the SSL_state() function
2999 has been removed and replaced by SSL_get_state which now returns an
3000 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
3001 altogether. The previous handshake states defined in ssl.h and ssl3.h have
3006 * All instances of the string "ssleay" in the public API were replaced
3007 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3008 Some error codes related to internal RSA_eay API's were renamed.
3012 * The demo files in crypto/threads were moved to demo/threads.
3016 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3019 *Matt Caswell, Rich Salz*
3021 * New ASN.1 embed macro.
3023 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3024 structure is not allocated: it is part of the parent. That is instead of
3032 This reduces memory fragmentation and make it impossible to accidentally
3033 set a mandatory field to NULL.
3035 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3036 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3037 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3042 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3046 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3047 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3048 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3049 DES and RC4 ciphersuites.
3053 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3054 This changes the decoding behaviour for some invalid messages,
3055 though the change is mostly in the more lenient direction, and
3056 legacy behaviour is preserved as much as possible.
3060 * Fix no-stdio build.
3061 * David Woodhouse <David.Woodhouse@intel.com> and also
3062 Ivan Nestlerode <ivan.nestlerode@sonos.com> *
3064 * New testing framework
3065 The testing framework has been largely rewritten and is now using
3066 perl and the perl modules Test::Harness and an extended variant of
3067 Test::More called OpenSSL::Test to do its work. All test scripts in
3068 test/ have been rewritten into test recipes, and all direct calls to
3069 executables in test/Makefile have become individual recipes using the
3070 simplified testing OpenSSL::Test::Simple.
3072 For documentation on our testing modules, do:
3074 perldoc test/testlib/OpenSSL/Test/Simple.pm
3075 perldoc test/testlib/OpenSSL/Test.pm
3079 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3080 are used; the latter aborts on memory leaks (usually checked on exit).
3081 Some undocumented "set malloc, etc., hooks" functions were removed
3082 and others were changed. All are now documented.
3086 * In DSA_generate_parameters_ex, if the provided seed is too short,
3089 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3091 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3092 from RFC4279, RFC4785, RFC5487, RFC5489.
3094 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3095 original RSA_PSK patch.
3099 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3100 era flag was never set throughout the codebase (only read). Also removed
3101 SSL3_FLAGS_POP_BUFFER which was only used if
3102 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3106 * Changed the default name options in the "ca", "crl", "req" and "x509"
3107 to be "oneline" instead of "compat".
3111 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3112 not aware of clients that still exhibit this bug, and the workaround
3113 hasn't been working properly for a while.
3117 * The return type of BIO_number_read() and BIO_number_written() as well as
3118 the corresponding num_read and num_write members in the BIO structure has
3119 changed from unsigned long to uint64_t. On platforms where an unsigned
3120 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3125 * Given the pervasive nature of TLS extensions it is inadvisable to run
3126 OpenSSL without support for them. It also means that maintaining
3127 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3128 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3132 * Removed support for the two export grade static DH ciphersuites
3133 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3134 were newly added (along with a number of other static DH ciphersuites) to
3135 1.0.2. However the two export ones have *never* worked since they were
3136 introduced. It seems strange in any case to be adding new export
3137 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3141 * Version negotiation has been rewritten. In particular SSLv23_method(),
3142 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3143 and turned into macros which simply call the new preferred function names
3144 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3145 should use the new names instead. Also as part of this change the ssl23.h
3146 header file has been removed.
3150 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3151 code and the associated standard is no longer considered fit-for-purpose.
3155 * RT2547 was closed. When generating a private key, try to make the
3156 output file readable only by the owner. This behavior change might
3157 be noticeable when interacting with other software.
3159 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3164 * Added HTTP GET support to the ocsp command.
3168 * Changed default digest for the dgst and enc commands from MD5 to
3173 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3177 * Added support for TLS extended master secret from
3178 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3179 initial patch which was a great help during development.
3183 * All libssl internal structures have been removed from the public header
3184 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3185 now redundant). Users should not attempt to access internal structures
3186 directly. Instead they should use the provided API functions.
3190 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3191 Access to deprecated functions can be re-enabled by running config with
3192 "enable-deprecated". In addition applications wishing to use deprecated
3193 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3194 will, by default, disable some transitive includes that previously existed
3195 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3199 * Added support for OCB mode. OpenSSL has been granted a patent license
3200 compatible with the OpenSSL license for use of OCB. Details are available
3201 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
3202 for OCB can be removed by calling config with no-ocb.
3206 * SSLv2 support has been removed. It still supports receiving a SSLv2
3207 compatible client hello.
3211 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3212 done while fixing the error code for the key-too-small case.
3214 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3216 * CA.sh has been removed; use CA.pl instead.
3220 * Removed old DES API.
3224 * Remove various unsupported platforms:
3230 Sinix/ReliantUNIX RM400
3235 16-bit platforms such as WIN16
3239 * Clean up OPENSSL_NO_xxx #define's
3240 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3241 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3242 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3243 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3244 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3245 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3246 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3247 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3248 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3249 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3253 * Cleaned up dead code
3254 Remove all but one '#ifdef undef' which is to be looked at.
3258 * Clean up calling of xxx_free routines.
3259 Just like free(), fix most of the xxx_free routines to accept
3260 NULL. Remove the non-null checks from callers. Save much code.
3264 * Add secure heap for storage of private keys (when possible).
3265 Add BIO_s_secmem(), CBIGNUM, etc.
3266 Contributed by Akamai Technologies under our Corporate CLA.
3270 * Experimental support for a new, fast, unbiased prime candidate generator,
3271 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3273 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3275 * New output format NSS in the sess_id command line tool. This allows
3276 exporting the session id and the master key in NSS keylog format.
3278 *Martin Kaiser <martin@kaiser.cx>*
3280 * Harmonize version and its documentation. -f flag is used to display
3283 *mancha <mancha1@zoho.com>*
3285 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3286 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3288 *mancha <mancha1@zoho.com>*
3290 * Fix some double frees. These are not thought to be exploitable.
3292 *mancha <mancha1@zoho.com>*
3294 * A missing bounds check in the handling of the TLS heartbeat extension
3295 can be used to reveal up to 64k of memory to a connected client or
3298 Thanks for Neel Mehta of Google Security for discovering this bug and to
3299 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3300 preparing the fix [CVE-2014-0160][]
3302 *Adam Langley, Bodo Moeller*
3304 * Fix for the attack described in the paper "Recovering OpenSSL
3305 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3306 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3307 http://eprint.iacr.org/2014/140
3309 Thanks to Yuval Yarom and Naomi Benger for discovering this
3310 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3312 *Yuval Yarom and Naomi Benger*
3314 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3315 this fixes a limitation in previous versions of OpenSSL.
3319 * Experimental encrypt-then-mac support.
3321 Experimental support for encrypt then mac from
3322 draft-gutmann-tls-encrypt-then-mac-02.txt
3324 To enable it set the appropriate extension number (0x42 for the test
3325 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3327 For non-compliant peers (i.e. just about everything) this should have no
3330 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3335 * Add EVP support for key wrapping algorithms, to avoid problems with
3336 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3337 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3338 algorithms and include tests cases.
3342 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3347 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3348 MGF1 digest and OAEP label.
3352 * Make openssl verify return errors.
3354 *Chris Palmer <palmer@google.com> and Ben Laurie*
3356 * New function ASN1_TIME_diff to calculate the difference between two
3357 ASN1_TIME structures or one structure and the current time.
3361 * Update fips_test_suite to support multiple command line options. New
3362 test to induce all self test errors in sequence and check expected
3367 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3368 sign or verify all in one operation.
3372 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3373 test programs and fips_test_suite. Includes functionality to parse
3374 the minimal script output of fipsalgest.pl directly.
3378 * Add authorisation parameter to FIPS_module_mode_set().
3382 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3386 * Use separate DRBG fields for internal and external flags. New function
3387 FIPS_drbg_health_check() to perform on demand health checking. Add
3388 generation tests to fips_test_suite with reduced health check interval to
3389 demonstrate periodic health checking. Add "nodh" option to
3390 fips_test_suite to skip very slow DH test.
3394 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3399 * More extensive health check for DRBG checking many more failure modes.
3400 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3401 combination: call this in fips_test_suite.
3405 * Add support for canonical generation of DSA parameter 'g'. See
3408 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3409 POST to handle HMAC cases.
3413 * Add functions FIPS_module_version() and FIPS_module_version_text()
3414 to return numerical and string versions of the FIPS module number.
3418 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3419 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3420 outside the validated module in the FIPS capable OpenSSL.
3424 * Minor change to DRBG entropy callback semantics. In some cases
3425 there is no multiple of the block length between min_len and
3426 max_len. Allow the callback to return more than max_len bytes
3427 of entropy but discard any extra: it is the callback's responsibility
3428 to ensure that the extra data discarded does not impact the
3429 requested amount of entropy.
3433 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3434 information in FIPS186-3, SP800-57 and SP800-131A.
3438 * CCM support via EVP. Interface is very similar to GCM case except we
3439 must supply all data in one chunk (i.e. no update, final) and the
3440 message length must be supplied if AAD is used. Add algorithm test
3445 * Initial version of POST overhaul. Add POST callback to allow the status
3446 of POST to be monitored and/or failures induced. Modify fips_test_suite
3447 to use callback. Always run all selftests even if one fails.
3451 * XTS support including algorithm test driver in the fips_gcmtest program.
3452 Note: this does increase the maximum key length from 32 to 64 bytes but
3453 there should be no binary compatibility issues as existing applications
3454 will never use XTS mode.
3458 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3459 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3460 performs algorithm blocking for unapproved PRNG types. Also do not
3461 set PRNG type in FIPS_mode_set(): leave this to the application.
3462 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3463 the standard OpenSSL PRNG: set additional data to a date time vector.
3467 * Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
3468 This shouldn't present any incompatibility problems because applications
3469 shouldn't be using these directly and any that are will need to rethink
3470 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3474 * Extensive self tests and health checking required by SP800-90 DRBG.
3475 Remove strength parameter from FIPS_drbg_instantiate and always
3476 instantiate at maximum supported strength.
3480 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3484 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3488 * New function DH_compute_key_padded() to compute a DH key and pad with
3489 leading zeroes if needed: this complies with SP800-56A et al.
3493 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3494 anything, incomplete, subject to change and largely untested at present.
3498 * Modify fipscanisteronly build option to only build the necessary object
3499 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3503 * Add experimental option FIPSSYMS to give all symbols in
3504 fipscanister.o and FIPS or fips prefix. This will avoid
3505 conflicts with future versions of OpenSSL. Add perl script
3506 util/fipsas.pl to preprocess assembly language source files
3507 and rename any affected symbols.
3511 * Add selftest checks and algorithm block of non-fips algorithms in
3512 FIPS mode. Remove DES2 from selftests.
3516 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3517 return internal method without any ENGINE dependencies. Add new
3518 tiny fips sign and verify functions.
3522 * New build option no-ec2m to disable characteristic 2 code.
3526 * New build option "fipscanisteronly". This only builds fipscanister.o
3527 and (currently) associated fips utilities. Uses the file Makefile.fips
3528 instead of Makefile.org as the prototype.
3532 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3533 Update fips_gcmtest to use IV generator.
3537 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3538 setting output buffer to NULL. The *Final function must be
3539 called although it will not retrieve any additional data. The tag
3540 can be set or retrieved with a ctrl. The IV length is by default 12
3541 bytes (96 bits) but can be set to an alternative value. If the IV
3542 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3547 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3548 underlying do_cipher function handles all cipher semantics itself
3549 including padding and finalisation. This is useful if (for example)
3550 an ENGINE cipher handles block padding itself. The behaviour of
3551 do_cipher is subtly changed if this flag is set: the return value
3552 is the number of characters written to the output buffer (zero is
3553 no longer an error code) or a negative error code. Also if the
3554 input buffer is NULL and length 0 finalisation should be performed.
3558 * If a candidate issuer certificate is already part of the constructed
3559 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3563 * Improve forward-security support: add functions
3565 void SSL_CTX_set_not_resumable_session_callback(
3566 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3567 void SSL_set_not_resumable_session_callback(
3568 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3570 for use by SSL/TLS servers; the callback function will be called whenever a
3571 new session is created, and gets to decide whether the session may be
3572 cached to make it resumable (return 0) or not (return 1). (As by the
3573 SSL/TLS protocol specifications, the session_id sent by the server will be
3574 empty to indicate that the session is not resumable; also, the server will
3575 not generate RFC 4507 (RFC 5077) session tickets.)
3577 A simple reasonable callback implementation is to return is_forward_secure.
3578 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3579 by the SSL/TLS server library, indicating whether it can provide forward
3582 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3584 * New -verify_name option in command line utilities to set verification
3589 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3590 Add CMAC pkey methods.
3594 * Experimental renegotiation in s_server -www mode. If the client
3595 browses /reneg connection is renegotiated. If /renegcert it is
3596 renegotiated requesting a certificate.
3600 * Add an "external" session cache for debugging purposes to s_server. This
3601 should help trace issues which normally are only apparent in deployed
3602 multi-process servers.
3606 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3607 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3608 BIO_set_cipher() and some obscure PEM functions were changed so they
3609 can now return an error. The RAND changes required a change to the
3610 RAND_METHOD structure.
3614 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3615 a gcc attribute to warn if the result of a function is ignored. This
3616 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3617 whose return value is often ignored.
3621 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3622 These allow SCTs (signed certificate timestamps) to be requested and
3623 validated when establishing a connection.
3625 *Rob Percival <robpercival@google.com>*
3630 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019] ###
3632 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3633 used even when parsing explicit parameters, when loading a serialized key
3634 or calling `EC_GROUP_new_from_ecpkparameters()`/
3635 `EC_GROUP_new_from_ecparameters()`.
3636 This prevents bypass of security hardening and performance gains,
3637 especially for curves with specialized EC_METHODs.
3638 By default, if a key encoded with explicit parameters is loaded and later
3639 serialized, the output is still encoded with explicit parameters, even if
3640 internally a "named" EC_GROUP is used for computation.
3644 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3645 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3646 NULL. After this change, only the cofactor parameter can be NULL. It also
3647 does some minimal sanity checks on the passed order.
3652 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3653 An attack is simple, if the first CMS_recipientInfo is valid but the
3654 second CMS_recipientInfo is chosen ciphertext. If the second
3655 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3656 encryption key will be replaced by garbage, and the message cannot be
3657 decoded, but if the RSA decryption fails, the correct encryption key is
3658 used and the recipient will not notice the attack.
3659 As a work around for this potential attack the length of the decrypted
3660 key must be equal to the cipher default key length, in case the
3661 certifiate is not given and all recipientInfo are tried out.
3662 The old behaviour can be re-enabled in the CMS code by setting the
3663 CMS_DEBUG_DECRYPT flag.
3668 * Document issue with installation paths in diverse Windows builds
3670 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3671 binaries and run-time config file.
3676 ### Changes between 1.0.2r and 1.0.2s [28 May 2019] ###
3678 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3679 This changes the size when using the genpkey app when no size is given. It
3680 fixes an omission in earlier changes that changed all RSA, DSA and DH
3681 generation apps to use 2048 bits by default.
3685 * Add FIPS support for Android Arm 64-bit
3687 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3688 Module in Version 2.0.10. For some reason, the corresponding target
3689 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3690 built with FIPS support on Android Arm 64-bit. This omission has been
3693 *Matthias St. Pierre*
3695 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019] ###
3697 * 0-byte record padding oracle
3699 If an application encounters a fatal protocol error and then calls
3700 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3701 then OpenSSL can respond differently to the calling application if a 0 byte
3702 record is received with invalid padding compared to if a 0 byte record is
3703 received with an invalid MAC. If the application then behaves differently
3704 based on that in a way that is detectable to the remote peer, then this
3705 amounts to a padding oracle that could be used to decrypt data.
3707 In order for this to be exploitable "non-stitched" ciphersuites must be in
3708 use. Stitched ciphersuites are optimised implementations of certain
3709 commonly used ciphersuites. Also the application must call SSL_shutdown()
3710 twice even if a protocol error has occurred (applications should not do
3711 this but some do anyway).
3713 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3714 Aviram, with additional investigation by Steven Collison and Andrew
3715 Hourselt. It was reported to OpenSSL on 10th December 2018.
3720 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3724 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018] ###
3726 * Microarchitecture timing vulnerability in ECC scalar multiplication
3728 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3729 shown to be vulnerable to a microarchitecture timing side channel attack.
3730 An attacker with sufficient access to mount local timing attacks during
3731 ECDSA signature generation could recover the private key.
3733 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3734 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3740 * Timing vulnerability in DSA signature generation
3742 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3743 timing side channel attack. An attacker could use variations in the signing
3744 algorithm to recover the private key.
3746 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3751 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3752 Module, accidentally introduced while backporting security fixes from the
3753 development branch and hindering the use of ECC in FIPS mode.
3757 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018] ###
3759 * Client DoS due to large DH parameter
3761 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3762 malicious server can send a very large prime value to the client. This will
3763 cause the client to spend an unreasonably long period of time generating a
3764 key for this prime resulting in a hang until the client has finished. This
3765 could be exploited in a Denial Of Service attack.
3767 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3772 * Cache timing vulnerability in RSA Key Generation
3774 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3775 a cache timing side channel attack. An attacker with sufficient access to
3776 mount cache timing attacks during the RSA key generation process could
3777 recover the private key.
3779 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3780 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3785 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3786 parameter is no longer accepted, as it leads to a corrupt table. NULL
3787 pem_str is reserved for alias entries only.
3791 * Revert blinding in ECDSA sign and instead make problematic addition
3792 length-invariant. Switch even to fixed-length Montgomery multiplication.
3796 * Change generating and checking of primes so that the error rate of not
3797 being prime depends on the intended use based on the size of the input.
3798 For larger primes this will result in more rounds of Miller-Rabin.
3799 The maximal error rate for primes with more than 1080 bits is lowered
3802 *Kurt Roeckx, Annie Yousar*
3804 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3808 * Add blinding to ECDSA and DSA signatures to protect against side channel
3809 attacks discovered by Keegan Ryan (NCC Group).
3813 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3814 now allow empty (zero character) pass phrases.
3818 * Certificate time validation (X509_cmp_time) enforces stricter
3819 compliance with RFC 5280. Fractional seconds and timezone offsets
3820 are no longer allowed.
3824 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018] ###
3826 * Constructed ASN.1 types with a recursive definition could exceed the stack
3828 Constructed ASN.1 types with a recursive definition (such as can be found
3829 in PKCS7) could eventually exceed the stack given malicious input with
3830 excessive recursion. This could result in a Denial Of Service attack. There
3831 are no such structures used within SSL/TLS that come from untrusted sources
3832 so this is considered safe.
3834 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3840 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017] ###
3842 * Read/write after SSL object in error state
3844 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3845 mechanism. The intent was that if a fatal error occurred during a handshake
3846 then OpenSSL would move into the error state and would immediately fail if
3847 you attempted to continue the handshake. This works as designed for the
3848 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3849 SSL_connect()), however due to a bug it does not work correctly if
3850 SSL_read() or SSL_write() is called directly. In that scenario, if the
3851 handshake fails then a fatal error will be returned in the initial function
3852 call. If SSL_read()/SSL_write() is subsequently called by the application
3853 for the same SSL object then it will succeed and the data is passed without
3854 being decrypted/encrypted directly from the SSL/TLS record layer.
3856 In order to exploit this issue an application bug would have to be present
3857 that resulted in a call to SSL_read()/SSL_write() being issued after having
3858 already received a fatal error.
3860 This issue was reported to OpenSSL by David Benjamin (Google).
3865 * rsaz_1024_mul_avx2 overflow bug on x86_64
3867 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3868 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3869 Analysis suggests that attacks against RSA and DSA as a result of this
3870 defect would be very difficult to perform and are not believed likely.
3871 Attacks against DH1024 are considered just feasible, because most of the
3872 work necessary to deduce information about a private key may be performed
3873 offline. The amount of resources required for such an attack would be
3874 significant. However, for an attack on TLS to be meaningful, the server
3875 would have to share the DH1024 private key among multiple clients, which is
3876 no longer an option since CVE-2016-0701.
3878 This only affects processors that support the AVX2 but not ADX extensions
3879 like Intel Haswell (4th generation).
3881 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3882 was originally found via the OSS-Fuzz project.
3887 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017] ###
3889 * bn_sqrx8x_internal carry bug on x86_64
3891 There is a carry propagating bug in the x86_64 Montgomery squaring
3892 procedure. No EC algorithms are affected. Analysis suggests that attacks
3893 against RSA and DSA as a result of this defect would be very difficult to
3894 perform and are not believed likely. Attacks against DH are considered just
3895 feasible (although very difficult) because most of the work necessary to
3896 deduce information about a private key may be performed offline. The amount
3897 of resources required for such an attack would be very significant and
3898 likely only accessible to a limited number of attackers. An attacker would
3899 additionally need online access to an unpatched system using the target
3900 private key in a scenario with persistent DH parameters and a private
3901 key that is shared between multiple clients.
3903 This only affects processors that support the BMI1, BMI2 and ADX extensions
3904 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3906 This issue was reported to OpenSSL by the OSS-Fuzz project.
3911 * Malformed X.509 IPAddressFamily could cause OOB read
3913 If an X.509 certificate has a malformed IPAddressFamily extension,
3914 OpenSSL could do a one-byte buffer overread. The most likely result
3915 would be an erroneous display of the certificate in text format.
3917 This issue was reported to OpenSSL by the OSS-Fuzz project.
3922 ### Changes between 1.0.2k and 1.0.2l [25 May 2017] ###
3924 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3925 platform rather than 'mingw'.
3929 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017] ###
3931 * Truncated packet could crash via OOB read
3933 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3934 cipher is being used, then a truncated packet can cause that host to
3935 perform an out-of-bounds read, usually resulting in a crash.
3937 This issue was reported to OpenSSL by Robert Święcki of Google.
3942 * BN_mod_exp may produce incorrect results on x86_64
3944 There is a carry propagating bug in the x86_64 Montgomery squaring
3945 procedure. No EC algorithms are affected. Analysis suggests that attacks
3946 against RSA and DSA as a result of this defect would be very difficult to
3947 perform and are not believed likely. Attacks against DH are considered just
3948 feasible (although very difficult) because most of the work necessary to
3949 deduce information about a private key may be performed offline. The amount
3950 of resources required for such an attack would be very significant and
3951 likely only accessible to a limited number of attackers. An attacker would
3952 additionally need online access to an unpatched system using the target
3953 private key in a scenario with persistent DH parameters and a private
3954 key that is shared between multiple clients. For example this can occur by
3955 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3956 similar to CVE-2015-3193 but must be treated as a separate problem.
3958 This issue was reported to OpenSSL by the OSS-Fuzz project.
3963 * Montgomery multiplication may produce incorrect results
3965 There is a carry propagating bug in the Broadwell-specific Montgomery
3966 multiplication procedure that handles input lengths divisible by, but
3967 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3968 and DH private keys are impossible. This is because the subroutine in
3969 question is not used in operations with the private key itself and an input
3970 of the attacker's direct choice. Otherwise the bug can manifest itself as
3971 transient authentication and key negotiation failures or reproducible
3972 erroneous outcome of public-key operations with specially crafted input.
3973 Among EC algorithms only Brainpool P-512 curves are affected and one
3974 presumably can attack ECDH key negotiation. Impact was not analyzed in
3975 detail, because pre-requisites for attack are considered unlikely. Namely
3976 multiple clients have to choose the curve in question and the server has to
3977 share the private key among them, neither of which is default behaviour.
3978 Even then only clients that chose the curve will be affected.
3980 This issue was publicly reported as transient failures and was not
3981 initially recognized as a security issue. Thanks to Richard Morgan for
3982 providing reproducible case.
3987 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3988 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3989 prevent issues where no progress is being made and the peer continually
3990 sends unrecognised record types, using up resources processing them.
3994 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016] ###
3996 * Missing CRL sanity check
3998 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
3999 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4000 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4002 This issue only affects the OpenSSL 1.0.2i
4007 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016] ###
4009 * OCSP Status Request extension unbounded memory growth
4011 A malicious client can send an excessively large OCSP Status Request
4012 extension. If that client continually requests renegotiation, sending a
4013 large OCSP Status Request extension each time, then there will be unbounded
4014 memory growth on the server. This will eventually lead to a Denial Of
4015 Service attack through memory exhaustion. Servers with a default
4016 configuration are vulnerable even if they do not support OCSP. Builds using
4017 the "no-ocsp" build time option are not affected.
4019 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4024 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4027 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4033 * OOB write in MDC2_Update()
4035 An overflow can occur in MDC2_Update() either if called directly or
4036 through the EVP_DigestUpdate() function using MDC2. If an attacker
4037 is able to supply very large amounts of input data after a previous
4038 call to EVP_EncryptUpdate() with a partial block then a length check
4039 can overflow resulting in a heap corruption.
4041 The amount of data needed is comparable to SIZE_MAX which is impractical
4044 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4049 * Malformed SHA512 ticket DoS
4051 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4052 DoS attack where a malformed ticket will result in an OOB read which will
4055 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4056 a custom server callback and ticket lookup mechanism.
4058 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4063 * OOB write in BN_bn2dec()
4065 The function BN_bn2dec() does not check the return value of BN_div_word().
4066 This can cause an OOB write if an application uses this function with an
4067 overly large BIGNUM. This could be a problem if an overly large certificate
4068 or CRL is printed out from an untrusted source. TLS is not affected because
4069 record limits will reject an oversized certificate before it is parsed.
4071 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4076 * OOB read in TS_OBJ_print_bio()
4078 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4079 the total length the OID text representation would use and not the amount
4080 of data written. This will result in OOB reads when large OIDs are
4083 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4088 * Pointer arithmetic undefined behaviour
4090 Avoid some undefined pointer arithmetic
4092 A common idiom in the codebase is to check limits in the following manner:
4095 Where "p" points to some malloc'd data of SIZE bytes and
4098 "len" here could be from some externally supplied data (e.g. from a TLS
4101 The rules of C pointer arithmetic are such that "p + len" is only well
4102 defined where len <= SIZE. Therefore the above idiom is actually
4103 undefined behaviour.
4105 For example this could cause problems if some malloc implementation
4106 provides an address for "p" such that "p + len" actually overflows for
4107 values of len that are too big and therefore p + len < limit.
4109 This issue was reported to OpenSSL by Guido Vranken
4114 * Constant time flag not preserved in DSA signing
4116 Operations in the DSA signing algorithm should run in constant time in
4117 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4118 implementation means that a non-constant time codepath is followed for
4119 certain operations. This has been demonstrated through a cache-timing
4120 attack to be sufficient for an attacker to recover the private DSA key.
4122 This issue was reported by César Pereida (Aalto University), Billy Brumley
4123 (Tampere University of Technology), and Yuval Yarom (The University of
4124 Adelaide and NICTA).
4129 * DTLS buffered message DoS
4131 In a DTLS connection where handshake messages are delivered out-of-order
4132 those messages that OpenSSL is not yet ready to process will be buffered
4133 for later use. Under certain circumstances, a flaw in the logic means that
4134 those messages do not get removed from the buffer even though the handshake
4135 has been completed. An attacker could force up to approx. 15 messages to
4136 remain in the buffer when they are no longer required. These messages will
4137 be cleared when the DTLS connection is closed. The default maximum size for
4138 a message is 100k. Therefore the attacker could force an additional 1500k
4139 to be consumed per connection. By opening many simulataneous connections an
4140 attacker could cause a DoS attack through memory exhaustion.
4142 This issue was reported to OpenSSL by Quan Luo.
4147 * DTLS replay protection DoS
4149 A flaw in the DTLS replay attack protection mechanism means that records
4150 that arrive for future epochs update the replay protection "window" before
4151 the MAC for the record has been validated. This could be exploited by an
4152 attacker by sending a record for the next epoch (which does not have to
4153 decrypt or have a valid MAC), with a very large sequence number. This means
4154 that all subsequent legitimate packets are dropped causing a denial of
4155 service for a specific DTLS connection.
4157 This issue was reported to OpenSSL by the OCAP audit team.
4162 * Certificate message OOB reads
4164 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4165 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4166 theoretical DoS risk but this has not been observed in practice on common
4169 The messages affected are client certificate, client certificate request
4170 and server certificate. As a result the attack can only be performed
4171 against a client or a server which enables client authentication.
4173 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4178 ### Changes between 1.0.2g and 1.0.2h [3 May 2016] ###
4180 * Prevent padding oracle in AES-NI CBC MAC check
4182 A MITM attacker can use a padding oracle attack to decrypt traffic
4183 when the connection uses an AES CBC cipher and the server support
4186 This issue was introduced as part of the fix for Lucky 13 padding
4187 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4188 constant time by making sure that always the same bytes are read and
4189 compared against either the MAC or padding bytes. But it no longer
4190 checked that there was enough data to have both the MAC and padding
4193 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4198 * Fix EVP_EncodeUpdate overflow
4200 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4201 Base64 encoding of binary data. If an attacker is able to supply very large
4202 amounts of input data then a length check can overflow resulting in a heap
4205 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4206 the PEM_write_bio* family of functions. These are mainly used within the
4207 OpenSSL command line applications, so any application which processes data
4208 from an untrusted source and outputs it as a PEM file should be considered
4209 vulnerable to this issue. User applications that call these APIs directly
4210 with large amounts of untrusted data may also be vulnerable.
4212 This issue was reported by Guido Vranken.
4217 * Fix EVP_EncryptUpdate overflow
4219 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4220 is able to supply very large amounts of input data after a previous call to
4221 EVP_EncryptUpdate() with a partial block then a length check can overflow
4222 resulting in a heap corruption. Following an analysis of all OpenSSL
4223 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4224 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4225 the first called function after an EVP_EncryptInit(), and therefore that
4226 specific call must be safe. The second form is where the length passed to
4227 EVP_EncryptUpdate() can be seen from the code to be some small value and
4228 therefore there is no possibility of an overflow. Since all instances are
4229 one of these two forms, it is believed that there can be no overflows in
4230 internal code due to this problem. It should be noted that
4231 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4232 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4233 of these calls have also been analysed too and it is believed there are no
4234 instances in internal usage where an overflow could occur.
4236 This issue was reported by Guido Vranken.
4241 * Prevent ASN.1 BIO excessive memory allocation
4243 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4244 a short invalid encoding can cause allocation of large amounts of memory
4245 potentially consuming excessive resources or exhausting memory.
4247 Any application parsing untrusted data through d2i BIO functions is
4248 affected. The memory based functions such as d2i_X509() are *not* affected.
4249 Since the memory based functions are used by the TLS library, TLS
4250 applications are not affected.
4252 This issue was reported by Brian Carpenter.
4259 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4260 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4261 in arbitrary stack data being returned in the buffer.
4263 This issue was reported by Guido Vranken.
4268 * Modify behavior of ALPN to invoke callback after SNI/servername
4269 callback, such that updates to the SSL_CTX affect ALPN.
4273 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4278 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4279 methods are enabled and ssl2 is disabled the methods return NULL.
4283 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016] ###
4285 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4286 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4287 provide any "EXPORT" or "LOW" strength ciphers.
4291 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4292 is by default disabled at build-time. Builds that are not configured with
4293 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4294 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4295 will need to explicitly call either of:
4297 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4299 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4301 as appropriate. Even if either of those is used, or the application
4302 explicitly uses the version-specific SSLv2_method() or its client and
4303 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4304 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4305 ciphers, and SSLv2 56-bit DES are no longer available.
4310 * Fix a double-free in DSA code
4312 A double free bug was discovered when OpenSSL parses malformed DSA private
4313 keys and could lead to a DoS attack or memory corruption for applications
4314 that receive DSA private keys from untrusted sources. This scenario is
4317 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4323 * Disable SRP fake user seed to address a server memory leak.
4325 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4327 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4328 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4329 was changed to ignore the "fake user" SRP seed, even if the seed
4332 Users should use SRP_VBASE_get1_by_user instead. Note that in
4333 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4334 also that even though configuring the SRP seed attempts to hide
4335 invalid usernames by continuing the handshake with fake
4336 credentials, this behaviour is not constant time and no strong
4337 guarantees are made that the handshake is indistinguishable from
4338 that of a valid user.
4343 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4345 In the BN_hex2bn function the number of hex digits is calculated using an
4346 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
4347 large values of |i| this can result in |bn_expand| not allocating any
4348 memory because |i * 4| is negative. This can leave the internal BIGNUM data
4349 field as NULL leading to a subsequent NULL ptr deref. For very large values
4350 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
4351 In this case memory is allocated to the internal BIGNUM data field, but it
4352 is insufficiently sized leading to heap corruption. A similar issue exists
4353 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4354 is ever called by user applications with very large untrusted hex/dec data.
4355 This is anticipated to be a rare occurrence.
4357 All OpenSSL internal usage of these functions use data that is not expected
4358 to be untrusted, e.g. config file data or application command line
4359 arguments. If user developed applications generate config file data based
4360 on untrusted data then it is possible that this could also lead to security
4361 consequences. This is also anticipated to be rare.
4363 This issue was reported to OpenSSL by Guido Vranken.
4368 * Fix memory issues in BIO_*printf functions
4370 The internal |fmtstr| function used in processing a "%s" format string in
4371 the BIO_*printf functions could overflow while calculating the length of a
4372 string and cause an OOB read when printing very long strings.
4374 Additionally the internal |doapr_outch| function can attempt to write to an
4375 OOB memory location (at an offset from the NULL pointer) in the event of a
4376 memory allocation failure. In 1.0.2 and below this could be caused where
4377 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4378 could be in processing a very long "%s" format string. Memory leaks can
4381 The first issue may mask the second issue dependent on compiler behaviour.
4382 These problems could enable attacks where large amounts of untrusted data
4383 is passed to the BIO_*printf functions. If applications use these functions
4384 in this way then they could be vulnerable. OpenSSL itself uses these
4385 functions when printing out human-readable dumps of ASN.1 data. Therefore
4386 applications that print this data could be vulnerable if the data is from
4387 untrusted sources. OpenSSL command line applications could also be
4388 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4389 as command line arguments.
4391 Libssl is not considered directly vulnerable. Additionally certificates etc
4392 received via remote connections via libssl are also unlikely to be able to
4393 trigger these issues because of message size limits enforced within libssl.
4395 This issue was reported to OpenSSL Guido Vranken.
4400 * Side channel attack on modular exponentiation
4402 A side-channel attack was found which makes use of cache-bank conflicts on
4403 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4404 of RSA keys. The ability to exploit this issue is limited as it relies on
4405 an attacker who has control of code in a thread running on the same
4406 hyper-threaded core as the victim thread which is performing decryptions.
4408 This issue was reported to OpenSSL by Yuval Yarom, The University of
4409 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4410 Nadia Heninger, University of Pennsylvania with more information at
4411 http://cachebleed.info.
4416 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4417 if no keysize is specified with default_bits. This fixes an
4418 omission in an earlier change that changed all RSA/DSA key generation
4419 apps to use 2048 bits by default.
4423 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016] ###
4424 * DH small subgroups
4426 Historically OpenSSL only ever generated DH parameters based on "safe"
4427 primes. More recently (in version 1.0.2) support was provided for
4428 generating X9.42 style parameter files such as those required for RFC 5114
4429 support. The primes used in such files may not be "safe". Where an
4430 application is using DH configured with parameters based on primes that are
4431 not "safe" then an attacker could use this fact to find a peer's private
4432 DH exponent. This attack requires that the attacker complete multiple
4433 handshakes in which the peer uses the same private DH exponent. For example
4434 this could be used to discover a TLS server's private DH exponent if it's
4435 reusing the private DH exponent or it's using a static DH ciphersuite.
4437 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4438 TLS. It is not on by default. If the option is not set then the server
4439 reuses the same private DH exponent for the life of the server process and
4440 would be vulnerable to this attack. It is believed that many popular
4441 applications do set this option and would therefore not be at risk.
4443 The fix for this issue adds an additional check where a "q" parameter is
4444 available (as is the case in X9.42 based parameters). This detects the
4445 only known attack, and is the only possible defense for static DH
4446 ciphersuites. This could have some performance impact.
4448 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4449 default and cannot be disabled. This could have some performance impact.
4451 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4456 * SSLv2 doesn't block disabled ciphers
4458 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4459 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4460 been disabled, provided that the SSLv2 protocol was not also disabled via
4463 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4464 and Sebastian Schinzel.
4469 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015] ###
4471 * BN_mod_exp may produce incorrect results on x86_64
4473 There is a carry propagating bug in the x86_64 Montgomery squaring
4474 procedure. No EC algorithms are affected. Analysis suggests that attacks
4475 against RSA and DSA as a result of this defect would be very difficult to
4476 perform and are not believed likely. Attacks against DH are considered just
4477 feasible (although very difficult) because most of the work necessary to
4478 deduce information about a private key may be performed offline. The amount
4479 of resources required for such an attack would be very significant and
4480 likely only accessible to a limited number of attackers. An attacker would
4481 additionally need online access to an unpatched system using the target
4482 private key in a scenario with persistent DH parameters and a private
4483 key that is shared between multiple clients. For example this can occur by
4484 default in OpenSSL DHE based SSL/TLS ciphersuites.
4486 This issue was reported to OpenSSL by Hanno Böck.
4491 * Certificate verify crash with missing PSS parameter
4493 The signature verification routines will crash with a NULL pointer
4494 dereference if presented with an ASN.1 signature using the RSA PSS
4495 algorithm and absent mask generation function parameter. Since these
4496 routines are used to verify certificate signature algorithms this can be
4497 used to crash any certificate verification operation and exploited in a
4498 DoS attack. Any application which performs certificate verification is
4499 vulnerable including OpenSSL clients and servers which enable client
4502 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4507 * X509_ATTRIBUTE memory leak
4509 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4510 memory. This structure is used by the PKCS#7 and CMS routines so any
4511 application which reads PKCS#7 or CMS data from untrusted sources is
4512 affected. SSL/TLS is not affected.
4514 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4520 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4521 This changes the decoding behaviour for some invalid messages,
4522 though the change is mostly in the more lenient direction, and
4523 legacy behaviour is preserved as much as possible.
4527 * In DSA_generate_parameters_ex, if the provided seed is too short,
4530 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4532 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015] ###
4534 * Alternate chains certificate forgery
4536 During certificate verification, OpenSSL will attempt to find an
4537 alternative certificate chain if the first attempt to build such a chain
4538 fails. An error in the implementation of this logic can mean that an
4539 attacker could cause certain checks on untrusted certificates to be
4540 bypassed, such as the CA flag, enabling them to use a valid leaf
4541 certificate to act as a CA and "issue" an invalid certificate.
4543 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4548 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015] ###
4550 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4551 incompatibility in the handling of HMAC. The previous ABI has now been
4556 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015] ###
4558 * Malformed ECParameters causes infinite loop
4560 When processing an ECParameters structure OpenSSL enters an infinite loop
4561 if the curve specified is over a specially malformed binary polynomial
4564 This can be used to perform denial of service against any
4565 system which processes public keys, certificate requests or
4566 certificates. This includes TLS clients and TLS servers with
4567 client authentication enabled.
4569 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4574 * Exploitable out-of-bounds read in X509_cmp_time
4576 X509_cmp_time does not properly check the length of the ASN1_TIME
4577 string and can read a few bytes out of bounds. In addition,
4578 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4581 An attacker can use this to craft malformed certificates and CRLs of
4582 various sizes and potentially cause a segmentation fault, resulting in
4583 a DoS on applications that verify certificates or CRLs. TLS clients
4584 that verify CRLs are affected. TLS clients and servers with client
4585 authentication enabled may be affected if they use custom verification
4588 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4589 independently by Hanno Böck.
4594 * PKCS7 crash with missing EnvelopedContent
4596 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4597 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4598 with missing content and trigger a NULL pointer dereference on parsing.
4600 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4601 structures from untrusted sources are affected. OpenSSL clients and
4602 servers are not affected.
4604 This issue was reported to OpenSSL by Michal Zalewski (Google).
4609 * CMS verify infinite loop with unknown hash function
4611 When verifying a signedData message the CMS code can enter an infinite loop
4612 if presented with an unknown hash function OID. This can be used to perform
4613 denial of service against any system which verifies signedData messages using
4615 This issue was reported to OpenSSL by Johannes Bauer.
4620 * Race condition handling NewSessionTicket
4622 If a NewSessionTicket is received by a multi-threaded client when attempting to
4623 reuse a previous ticket then a race condition can occur potentially leading to
4624 a double free of the ticket data.
4629 * Only support 256-bit or stronger elliptic curves with the
4630 'ecdh_auto' setting (server) or by default (client). Of supported
4631 curves, prefer P-256 (both).
4635 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015] ###
4637 * ClientHello sigalgs DoS fix
4639 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4640 invalid signature algorithms extension a NULL pointer dereference will
4641 occur. This can be exploited in a DoS attack against the server.
4643 This issue was was reported to OpenSSL by David Ramos of Stanford
4647 *Stephen Henson and Matt Caswell*
4649 * Multiblock corrupted pointer fix
4651 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4652 feature only applies on 64 bit x86 architecture platforms that support AES
4653 NI instructions. A defect in the implementation of "multiblock" can cause
4654 OpenSSL's internal write buffer to become incorrectly set to NULL when
4655 using non-blocking IO. Typically, when the user application is using a
4656 socket BIO for writing, this will only result in a failed connection.
4657 However if some other BIO is used then it is likely that a segmentation
4658 fault will be triggered, thus enabling a potential DoS attack.
4660 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4665 * Segmentation fault in DTLSv1_listen fix
4667 The DTLSv1_listen function is intended to be stateless and processes the
4668 initial ClientHello from many peers. It is common for user code to loop
4669 over the call to DTLSv1_listen until a valid ClientHello is received with
4670 an associated cookie. A defect in the implementation of DTLSv1_listen means
4671 that state is preserved in the SSL object from one invocation to the next
4672 that can lead to a segmentation fault. Errors processing the initial
4673 ClientHello can trigger this scenario. An example of such an error could be
4674 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4677 This issue was reported to OpenSSL by Per Allansson.
4682 * Segmentation fault in ASN1_TYPE_cmp fix
4684 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4685 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4686 certificate signature algorithm consistency this can be used to crash any
4687 certificate verification operation and exploited in a DoS attack. Any
4688 application which performs certificate verification is vulnerable including
4689 OpenSSL clients and servers which enable client authentication.
4694 * Segmentation fault for invalid PSS parameters fix
4696 The signature verification routines will crash with a NULL pointer
4697 dereference if presented with an ASN.1 signature using the RSA PSS
4698 algorithm and invalid parameters. Since these routines are used to verify
4699 certificate signature algorithms this can be used to crash any
4700 certificate verification operation and exploited in a DoS attack. Any
4701 application which performs certificate verification is vulnerable including
4702 OpenSSL clients and servers which enable client authentication.
4704 This issue was was reported to OpenSSL by Brian Carpenter.
4709 * ASN.1 structure reuse memory corruption fix
4711 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4712 memory corruption via an invalid write. Such reuse is and has been
4713 strongly discouraged and is believed to be rare.
4715 Applications that parse structures containing CHOICE or ANY DEFINED BY
4716 components may be affected. Certificate parsing (d2i_X509 and related
4717 functions) are however not affected. OpenSSL clients and servers are
4723 * PKCS7 NULL pointer dereferences fix
4725 The PKCS#7 parsing code does not handle missing outer ContentInfo
4726 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4727 missing content and trigger a NULL pointer dereference on parsing.
4729 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4730 otherwise parse PKCS#7 structures from untrusted sources are
4731 affected. OpenSSL clients and servers are not affected.
4733 This issue was reported to OpenSSL by Michal Zalewski (Google).
4738 * DoS via reachable assert in SSLv2 servers fix
4740 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4741 servers that both support SSLv2 and enable export cipher suites by sending
4742 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4744 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4745 (OpenSSL development team).
4750 * Empty CKE with client auth and DHE fix
4752 If client auth is used then a server can seg fault in the event of a DHE
4753 ciphersuite being selected and a zero length ClientKeyExchange message
4754 being sent by the client. This could be exploited in a DoS attack.
4759 * Handshake with unseeded PRNG fix
4761 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4762 with an unseeded PRNG. The conditions are:
4763 - The client is on a platform where the PRNG has not been seeded
4764 automatically, and the user has not seeded manually
4765 - A protocol specific client method version has been used (i.e. not
4766 SSL_client_methodv23)
4767 - A ciphersuite is used that does not require additional random data from
4768 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4770 If the handshake succeeds then the client random that has been used will
4771 have been generated from a PRNG with insufficient entropy and therefore the
4772 output may be predictable.
4774 For example using the following command with an unseeded openssl will
4775 succeed on an unpatched platform:
4777 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4782 * Use After Free following d2i_ECPrivatekey error fix
4784 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4785 could cause a use after free condition. This, in turn, could cause a double
4786 free in several private key parsing functions (such as d2i_PrivateKey
4787 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4788 for applications that receive EC private keys from untrusted
4789 sources. This scenario is considered rare.
4791 This issue was discovered by the BoringSSL project and fixed in their
4797 * X509_to_X509_REQ NULL pointer deref fix
4799 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4800 the certificate key is invalid. This function is rarely used in practice.
4802 This issue was discovered by Brian Carpenter.
4807 * Removed the export ciphers from the DEFAULT ciphers
4811 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015] ###
4813 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4814 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4815 So far those who have to target multiple platforms would compromise
4816 and argue that binary targeting say ARMv5 would still execute on
4817 ARMv8. "Universal" build resolves this compromise by providing
4818 near-optimal performance even on newer platforms.
4822 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4823 (other platforms pending).
4825 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4827 * Add support for the SignedCertificateTimestampList certificate and
4828 OCSP response extensions from RFC6962.
4832 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4833 for corner cases. (Certain input points at infinity could lead to
4834 bogus results, with non-infinity inputs mapped to infinity too.)
4838 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4839 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4840 common cases are optimized and there still is room for further
4841 improvements. Vector Permutation AES for Altivec is also added.
4845 * Add support for little-endian ppc64 Linux target.
4847 *Marcelo Cerri (IBM)*
4849 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4850 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4851 are optimized and there still is room for further improvements.
4852 Both 32- and 64-bit modes are supported.
4854 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4856 * Improved ARMv7 NEON support.
4860 * Support for SPARC Architecture 2011 crypto extensions, first
4861 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4862 SHA256/512, MD5, GHASH and modular exponentiation.
4864 *Andy Polyakov, David Miller*
4866 * Accelerated modular exponentiation for Intel processors, a.k.a.
4869 *Shay Gueron & Vlad Krasnov (Intel Corp)*
4871 * Support for new and upcoming Intel processors, including AVX2,
4872 BMI and SHA ISA extensions. This includes additional "stitched"
4873 implementations, AESNI-SHA256 and GCM, and multi-buffer support
4876 This work was sponsored by Intel Corp.
4880 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
4881 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
4882 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
4886 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4887 this fixes a limitation in previous versions of OpenSSL.
4891 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4892 MGF1 digest and OAEP label.
4896 * Add EVP support for key wrapping algorithms, to avoid problems with
4897 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4898 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4899 algorithms and include tests cases.
4903 * Add functions to allocate and set the fields of an ECDSA_METHOD
4906 *Douglas E. Engert, Steve Henson*
4908 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
4909 difference in days and seconds between two tm or ASN1_TIME structures.
4913 * Add -rev test option to s_server to just reverse order of characters
4914 received by client and send back to server. Also prints an abbreviated
4915 summary of the connection parameters.
4919 * New option -brief for s_client and s_server to print out a brief summary
4920 of connection parameters.
4924 * Add callbacks for arbitrary TLS extensions.
4926 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
4928 * New option -crl_download in several openssl utilities to download CRLs
4929 from CRLDP extension in certificates.
4933 * New options -CRL and -CRLform for s_client and s_server for CRLs.
4937 * New function X509_CRL_diff to generate a delta CRL from the difference
4938 of two full CRLs. Add support to "crl" utility.
4942 * New functions to set lookup_crls function and to retrieve
4943 X509_STORE from X509_STORE_CTX.
4947 * Print out deprecated issuer and subject unique ID fields in
4952 * Extend OCSP I/O functions so they can be used for simple general purpose
4953 HTTP as well as OCSP. New wrapper function which can be used to download
4954 CRLs using the OCSP API.
4958 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
4962 * SSL_CONF* functions. These provide a common framework for application
4963 configuration using configuration files or command lines.
4967 * SSL/TLS tracing code. This parses out SSL/TLS records using the
4968 message callback and prints the results. Needs compile time option
4969 "enable-ssl-trace". New options to s_client and s_server to enable
4974 * New ctrl and macro to retrieve supported points extensions.
4975 Print out extension in s_server and s_client.
4979 * New functions to retrieve certificate signature and signature
4984 * Add functions to retrieve and manipulate the raw cipherlist sent by a
4989 * New Suite B modes for TLS code. These use and enforce the requirements
4990 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
4991 only use Suite B curves. The Suite B modes can be set by using the
4992 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
4996 * New chain verification flags for Suite B levels of security. Check
4997 algorithms are acceptable when flags are set in X509_verify_cert.
5001 * Make tls1_check_chain return a set of flags indicating checks passed
5002 by a certificate chain. Add additional tests to handle client
5003 certificates: checks for matching certificate type and issuer name
5008 * If an attempt is made to use a signature algorithm not in the peer
5009 preference list abort the handshake. If client has no suitable
5010 signature algorithms in response to a certificate request do not
5011 use the certificate.
5015 * If server EC tmp key is not in client preference list abort handshake.
5019 * Add support for certificate stores in CERT structure. This makes it
5020 possible to have different stores per SSL structure or one store in
5021 the parent SSL_CTX. Include distinct stores for certificate chain
5022 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5023 to build and store a certificate chain in CERT structure: returning
5024 an error if the chain cannot be built: this will allow applications
5025 to test if a chain is correctly configured.
5027 Note: if the CERT based stores are not set then the parent SSL_CTX
5028 store is used to retain compatibility with existing behaviour.
5033 * New function ssl_set_client_disabled to set a ciphersuite disabled
5034 mask based on the current session, check mask when sending client
5035 hello and checking the requested ciphersuite.
5039 * New ctrls to retrieve and set certificate types in a certificate
5040 request message. Print out received values in s_client. If certificate
5041 types is not set with custom values set sensible values based on
5042 supported signature algorithms.
5046 * Support for distinct client and server supported signature algorithms.
5050 * Add certificate callback. If set this is called whenever a certificate
5051 is required by client or server. An application can decide which
5052 certificate chain to present based on arbitrary criteria: for example
5053 supported signature algorithms. Add very simple example to s_server.
5054 This fixes many of the problems and restrictions of the existing client
5055 certificate callback: for example you can now clear an existing
5056 certificate and specify the whole chain.
5060 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5061 the certificate can be used for (if anything). Set valid_flags field
5062 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5063 to have similar checks in it.
5065 Add new "cert_flags" field to CERT structure and include a "strict mode".
5066 This enforces some TLS certificate requirements (such as only permitting
5067 certificate signature algorithms contained in the supported algorithms
5068 extension) which some implementations ignore: this option should be used
5069 with caution as it could cause interoperability issues.
5073 * Update and tidy signature algorithm extension processing. Work out
5074 shared signature algorithms based on preferences and peer algorithms
5075 and print them out in s_client and s_server. Abort handshake if no
5076 shared signature algorithms.
5080 * Add new functions to allow customised supported signature algorithms
5081 for SSL and SSL_CTX structures. Add options to s_client and s_server
5086 * New function SSL_certs_clear() to delete all references to certificates
5087 from an SSL structure. Before this once a certificate had been added
5088 it couldn't be removed.
5092 * Integrate hostname, email address and IP address checking with certificate
5093 verification. New verify options supporting checking in openssl utility.
5097 * Fixes and wildcard matching support to hostname and email checking
5098 functions. Add manual page.
5100 *Florian Weimer (Red Hat Product Security Team)*
5102 * New functions to check a hostname email or IP address against a
5103 certificate. Add options x509 utility to print results of checks against
5108 * Fix OCSP checking.
5110 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5112 * Initial experimental support for explicitly trusted non-root CAs.
5113 OpenSSL still tries to build a complete chain to a root but if an
5114 intermediate CA has a trust setting included that is used. The first
5115 setting is used: whether to trust (e.g., -addtrust option to the x509
5120 * Add -trusted_first option which attempts to find certificates in the
5121 trusted store even if an untrusted chain is also supplied.
5125 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5126 platform support for Linux and Android.
5130 * Support for linux-x32, ILP32 environment in x86_64 framework.
5134 * Experimental multi-implementation support for FIPS capable OpenSSL.
5135 When in FIPS mode the approved implementations are used as normal,
5136 when not in FIPS mode the internal unapproved versions are used instead.
5137 This means that the FIPS capable OpenSSL isn't forced to use the
5138 (often lower performance) FIPS implementations outside FIPS mode.
5142 * Transparently support X9.42 DH parameters when calling
5143 PEM_read_bio_DHparameters. This means existing applications can handle
5144 the new parameter format automatically.
5148 * Initial experimental support for X9.42 DH parameter format: mainly
5149 to support use of 'q' parameter for RFC5114 parameters.
5153 * Add DH parameters from RFC5114 including test data to dhtest.
5157 * Support for automatic EC temporary key parameter selection. If enabled
5158 the most preferred EC parameters are automatically used instead of
5159 hardcoded fixed parameters. Now a server just has to call:
5160 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5161 support ECDH and use the most appropriate parameters.
5165 * Enhance and tidy EC curve and point format TLS extension code. Use
5166 static structures instead of allocation if default values are used.
5167 New ctrls to set curves we wish to support and to retrieve shared curves.
5168 Print out shared curves in s_server. New options to s_server and s_client
5169 to set list of supported curves.
5173 * New ctrls to retrieve supported signature algorithms and
5174 supported curve values as an array of NIDs. Extend openssl utility
5175 to print out received values.
5179 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5180 between NIDs and the more common NIST names such as "P-256". Enhance
5181 ecparam utility and ECC method to recognise the NIST names for curves.
5185 * Enhance SSL/TLS certificate chain handling to support different
5186 chains for each certificate instead of one chain in the parent SSL_CTX.
5190 * Support for fixed DH ciphersuite client authentication: where both
5191 server and client use DH certificates with common parameters.
5195 * Support for fixed DH ciphersuites: those requiring DH server
5200 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5202 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5203 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5204 X509_CINF_get_signature were reverted post internal team review.
5210 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016] ###
5212 * OCSP Status Request extension unbounded memory growth
5214 A malicious client can send an excessively large OCSP Status Request
5215 extension. If that client continually requests renegotiation, sending a
5216 large OCSP Status Request extension each time, then there will be unbounded
5217 memory growth on the server. This will eventually lead to a Denial Of
5218 Service attack through memory exhaustion. Servers with a default
5219 configuration are vulnerable even if they do not support OCSP. Builds using
5220 the "no-ocsp" build time option are not affected.
5222 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5227 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5230 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5236 * OOB write in MDC2_Update()
5238 An overflow can occur in MDC2_Update() either if called directly or
5239 through the EVP_DigestUpdate() function using MDC2. If an attacker
5240 is able to supply very large amounts of input data after a previous
5241 call to EVP_EncryptUpdate() with a partial block then a length check
5242 can overflow resulting in a heap corruption.
5244 The amount of data needed is comparable to SIZE_MAX which is impractical
5247 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5252 * Malformed SHA512 ticket DoS
5254 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5255 DoS attack where a malformed ticket will result in an OOB read which will
5258 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5259 a custom server callback and ticket lookup mechanism.
5261 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5266 * OOB write in BN_bn2dec()
5268 The function BN_bn2dec() does not check the return value of BN_div_word().
5269 This can cause an OOB write if an application uses this function with an
5270 overly large BIGNUM. This could be a problem if an overly large certificate
5271 or CRL is printed out from an untrusted source. TLS is not affected because
5272 record limits will reject an oversized certificate before it is parsed.
5274 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5279 * OOB read in TS_OBJ_print_bio()
5281 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5282 the total length the OID text representation would use and not the amount
5283 of data written. This will result in OOB reads when large OIDs are
5286 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5291 * Pointer arithmetic undefined behaviour
5293 Avoid some undefined pointer arithmetic
5295 A common idiom in the codebase is to check limits in the following manner:
5298 Where "p" points to some malloc'd data of SIZE bytes and
5301 "len" here could be from some externally supplied data (e.g. from a TLS
5304 The rules of C pointer arithmetic are such that "p + len" is only well
5305 defined where len <= SIZE. Therefore the above idiom is actually
5306 undefined behaviour.
5308 For example this could cause problems if some malloc implementation
5309 provides an address for "p" such that "p + len" actually overflows for
5310 values of len that are too big and therefore p + len < limit.
5312 This issue was reported to OpenSSL by Guido Vranken
5317 * Constant time flag not preserved in DSA signing
5319 Operations in the DSA signing algorithm should run in constant time in
5320 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5321 implementation means that a non-constant time codepath is followed for
5322 certain operations. This has been demonstrated through a cache-timing
5323 attack to be sufficient for an attacker to recover the private DSA key.
5325 This issue was reported by César Pereida (Aalto University), Billy Brumley
5326 (Tampere University of Technology), and Yuval Yarom (The University of
5327 Adelaide and NICTA).
5332 * DTLS buffered message DoS
5334 In a DTLS connection where handshake messages are delivered out-of-order
5335 those messages that OpenSSL is not yet ready to process will be buffered
5336 for later use. Under certain circumstances, a flaw in the logic means that
5337 those messages do not get removed from the buffer even though the handshake
5338 has been completed. An attacker could force up to approx. 15 messages to
5339 remain in the buffer when they are no longer required. These messages will
5340 be cleared when the DTLS connection is closed. The default maximum size for
5341 a message is 100k. Therefore the attacker could force an additional 1500k
5342 to be consumed per connection. By opening many simulataneous connections an
5343 attacker could cause a DoS attack through memory exhaustion.
5345 This issue was reported to OpenSSL by Quan Luo.
5350 * DTLS replay protection DoS
5352 A flaw in the DTLS replay attack protection mechanism means that records
5353 that arrive for future epochs update the replay protection "window" before
5354 the MAC for the record has been validated. This could be exploited by an
5355 attacker by sending a record for the next epoch (which does not have to
5356 decrypt or have a valid MAC), with a very large sequence number. This means
5357 that all subsequent legitimate packets are dropped causing a denial of
5358 service for a specific DTLS connection.
5360 This issue was reported to OpenSSL by the OCAP audit team.
5365 * Certificate message OOB reads
5367 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5368 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5369 theoretical DoS risk but this has not been observed in practice on common
5372 The messages affected are client certificate, client certificate request
5373 and server certificate. As a result the attack can only be performed
5374 against a client or a server which enables client authentication.
5376 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5381 ### Changes between 1.0.1s and 1.0.1t [3 May 2016] ###
5383 * Prevent padding oracle in AES-NI CBC MAC check
5385 A MITM attacker can use a padding oracle attack to decrypt traffic
5386 when the connection uses an AES CBC cipher and the server support
5389 This issue was introduced as part of the fix for Lucky 13 padding
5390 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5391 constant time by making sure that always the same bytes are read and
5392 compared against either the MAC or padding bytes. But it no longer
5393 checked that there was enough data to have both the MAC and padding
5396 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5401 * Fix EVP_EncodeUpdate overflow
5403 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5404 Base64 encoding of binary data. If an attacker is able to supply very large
5405 amounts of input data then a length check can overflow resulting in a heap
5408 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5409 the PEM_write_bio* family of functions. These are mainly used within the
5410 OpenSSL command line applications, so any application which processes data
5411 from an untrusted source and outputs it as a PEM file should be considered
5412 vulnerable to this issue. User applications that call these APIs directly
5413 with large amounts of untrusted data may also be vulnerable.
5415 This issue was reported by Guido Vranken.
5420 * Fix EVP_EncryptUpdate overflow
5422 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5423 is able to supply very large amounts of input data after a previous call to
5424 EVP_EncryptUpdate() with a partial block then a length check can overflow
5425 resulting in a heap corruption. Following an analysis of all OpenSSL
5426 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5427 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5428 the first called function after an EVP_EncryptInit(), and therefore that
5429 specific call must be safe. The second form is where the length passed to
5430 EVP_EncryptUpdate() can be seen from the code to be some small value and
5431 therefore there is no possibility of an overflow. Since all instances are
5432 one of these two forms, it is believed that there can be no overflows in
5433 internal code due to this problem. It should be noted that
5434 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5435 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5436 of these calls have also been analysed too and it is believed there are no
5437 instances in internal usage where an overflow could occur.
5439 This issue was reported by Guido Vranken.
5444 * Prevent ASN.1 BIO excessive memory allocation
5446 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5447 a short invalid encoding can casuse allocation of large amounts of memory
5448 potentially consuming excessive resources or exhausting memory.
5450 Any application parsing untrusted data through d2i BIO functions is
5451 affected. The memory based functions such as d2i_X509() are *not* affected.
5452 Since the memory based functions are used by the TLS library, TLS
5453 applications are not affected.
5455 This issue was reported by Brian Carpenter.
5462 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5463 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5464 in arbitrary stack data being returned in the buffer.
5466 This issue was reported by Guido Vranken.
5471 * Modify behavior of ALPN to invoke callback after SNI/servername
5472 callback, such that updates to the SSL_CTX affect ALPN.
5476 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5481 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5482 methods are enabled and ssl2 is disabled the methods return NULL.
5486 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016] ###
5488 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5489 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5490 provide any "EXPORT" or "LOW" strength ciphers.
5494 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5495 is by default disabled at build-time. Builds that are not configured with
5496 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5497 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5498 will need to explicitly call either of:
5500 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5502 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5504 as appropriate. Even if either of those is used, or the application
5505 explicitly uses the version-specific SSLv2_method() or its client and
5506 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5507 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5508 ciphers, and SSLv2 56-bit DES are no longer available.
5513 * Fix a double-free in DSA code
5515 A double free bug was discovered when OpenSSL parses malformed DSA private
5516 keys and could lead to a DoS attack or memory corruption for applications
5517 that receive DSA private keys from untrusted sources. This scenario is
5520 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5526 * Disable SRP fake user seed to address a server memory leak.
5528 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5530 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5531 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5532 was changed to ignore the "fake user" SRP seed, even if the seed
5535 Users should use SRP_VBASE_get1_by_user instead. Note that in
5536 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5537 also that even though configuring the SRP seed attempts to hide
5538 invalid usernames by continuing the handshake with fake
5539 credentials, this behaviour is not constant time and no strong
5540 guarantees are made that the handshake is indistinguishable from
5541 that of a valid user.
5546 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5548 In the BN_hex2bn function the number of hex digits is calculated using an
5549 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
5550 large values of |i| this can result in |bn_expand| not allocating any
5551 memory because |i * 4| is negative. This can leave the internal BIGNUM data
5552 field as NULL leading to a subsequent NULL ptr deref. For very large values
5553 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
5554 In this case memory is allocated to the internal BIGNUM data field, but it
5555 is insufficiently sized leading to heap corruption. A similar issue exists
5556 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5557 is ever called by user applications with very large untrusted hex/dec data.
5558 This is anticipated to be a rare occurrence.
5560 All OpenSSL internal usage of these functions use data that is not expected
5561 to be untrusted, e.g. config file data or application command line
5562 arguments. If user developed applications generate config file data based
5563 on untrusted data then it is possible that this could also lead to security
5564 consequences. This is also anticipated to be rare.
5566 This issue was reported to OpenSSL by Guido Vranken.
5571 * Fix memory issues in BIO_*printf functions
5573 The internal |fmtstr| function used in processing a "%s" format string in
5574 the BIO_*printf functions could overflow while calculating the length of a
5575 string and cause an OOB read when printing very long strings.
5577 Additionally the internal |doapr_outch| function can attempt to write to an
5578 OOB memory location (at an offset from the NULL pointer) in the event of a
5579 memory allocation failure. In 1.0.2 and below this could be caused where
5580 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5581 could be in processing a very long "%s" format string. Memory leaks can
5584 The first issue may mask the second issue dependent on compiler behaviour.
5585 These problems could enable attacks where large amounts of untrusted data
5586 is passed to the BIO_*printf functions. If applications use these functions
5587 in this way then they could be vulnerable. OpenSSL itself uses these
5588 functions when printing out human-readable dumps of ASN.1 data. Therefore
5589 applications that print this data could be vulnerable if the data is from
5590 untrusted sources. OpenSSL command line applications could also be
5591 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5592 as command line arguments.
5594 Libssl is not considered directly vulnerable. Additionally certificates etc
5595 received via remote connections via libssl are also unlikely to be able to
5596 trigger these issues because of message size limits enforced within libssl.
5598 This issue was reported to OpenSSL Guido Vranken.
5603 * Side channel attack on modular exponentiation
5605 A side-channel attack was found which makes use of cache-bank conflicts on
5606 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5607 of RSA keys. The ability to exploit this issue is limited as it relies on
5608 an attacker who has control of code in a thread running on the same
5609 hyper-threaded core as the victim thread which is performing decryptions.
5611 This issue was reported to OpenSSL by Yuval Yarom, The University of
5612 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5613 Nadia Heninger, University of Pennsylvania with more information at
5614 http://cachebleed.info.
5619 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5620 if no keysize is specified with default_bits. This fixes an
5621 omission in an earlier change that changed all RSA/DSA key generation
5622 apps to use 2048 bits by default.
5626 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016] ###
5628 * Protection for DH small subgroup attacks
5630 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5631 switched on by default and cannot be disabled. This could have some
5636 * SSLv2 doesn't block disabled ciphers
5638 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5639 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5640 been disabled, provided that the SSLv2 protocol was not also disabled via
5643 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5644 and Sebastian Schinzel.
5649 * Reject DH handshakes with parameters shorter than 1024 bits.
5653 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015] ###
5655 * Certificate verify crash with missing PSS parameter
5657 The signature verification routines will crash with a NULL pointer
5658 dereference if presented with an ASN.1 signature using the RSA PSS
5659 algorithm and absent mask generation function parameter. Since these
5660 routines are used to verify certificate signature algorithms this can be
5661 used to crash any certificate verification operation and exploited in a
5662 DoS attack. Any application which performs certificate verification is
5663 vulnerable including OpenSSL clients and servers which enable client
5666 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5671 * X509_ATTRIBUTE memory leak
5673 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5674 memory. This structure is used by the PKCS#7 and CMS routines so any
5675 application which reads PKCS#7 or CMS data from untrusted sources is
5676 affected. SSL/TLS is not affected.
5678 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5684 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5685 This changes the decoding behaviour for some invalid messages,
5686 though the change is mostly in the more lenient direction, and
5687 legacy behaviour is preserved as much as possible.
5691 * In DSA_generate_parameters_ex, if the provided seed is too short,
5692 use a random seed, as already documented.
5694 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5696 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015] ###
5698 * Alternate chains certificate forgery
5700 During certificate verfification, OpenSSL will attempt to find an
5701 alternative certificate chain if the first attempt to build such a chain
5702 fails. An error in the implementation of this logic can mean that an
5703 attacker could cause certain checks on untrusted certificates to be
5704 bypassed, such as the CA flag, enabling them to use a valid leaf
5705 certificate to act as a CA and "issue" an invalid certificate.
5707 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5713 * Race condition handling PSK identify hint
5715 If PSK identity hints are received by a multi-threaded client then
5716 the values are wrongly updated in the parent SSL_CTX structure. This can
5717 result in a race condition potentially leading to a double free of the
5723 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015] ###
5724 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5725 incompatibility in the handling of HMAC. The previous ABI has now been
5728 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015] ###
5730 * Malformed ECParameters causes infinite loop
5732 When processing an ECParameters structure OpenSSL enters an infinite loop
5733 if the curve specified is over a specially malformed binary polynomial
5736 This can be used to perform denial of service against any
5737 system which processes public keys, certificate requests or
5738 certificates. This includes TLS clients and TLS servers with
5739 client authentication enabled.
5741 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5746 * Exploitable out-of-bounds read in X509_cmp_time
5748 X509_cmp_time does not properly check the length of the ASN1_TIME
5749 string and can read a few bytes out of bounds. In addition,
5750 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5753 An attacker can use this to craft malformed certificates and CRLs of
5754 various sizes and potentially cause a segmentation fault, resulting in
5755 a DoS on applications that verify certificates or CRLs. TLS clients
5756 that verify CRLs are affected. TLS clients and servers with client
5757 authentication enabled may be affected if they use custom verification
5760 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5761 independently by Hanno Böck.
5766 * PKCS7 crash with missing EnvelopedContent
5768 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5769 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5770 with missing content and trigger a NULL pointer dereference on parsing.
5772 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5773 structures from untrusted sources are affected. OpenSSL clients and
5774 servers are not affected.
5776 This issue was reported to OpenSSL by Michal Zalewski (Google).
5781 * CMS verify infinite loop with unknown hash function
5783 When verifying a signedData message the CMS code can enter an infinite loop
5784 if presented with an unknown hash function OID. This can be used to perform
5785 denial of service against any system which verifies signedData messages using
5787 This issue was reported to OpenSSL by Johannes Bauer.
5792 * Race condition handling NewSessionTicket
5794 If a NewSessionTicket is received by a multi-threaded client when attempting to
5795 reuse a previous ticket then a race condition can occur potentially leading to
5796 a double free of the ticket data.
5801 * Reject DH handshakes with parameters shorter than 768 bits.
5803 *Kurt Roeckx and Emilia Kasper*
5805 * dhparam: generate 2048-bit parameters by default.
5807 *Kurt Roeckx and Emilia Kasper*
5809 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015] ###
5811 * Segmentation fault in ASN1_TYPE_cmp fix
5813 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5814 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5815 certificate signature algorithm consistency this can be used to crash any
5816 certificate verification operation and exploited in a DoS attack. Any
5817 application which performs certificate verification is vulnerable including
5818 OpenSSL clients and servers which enable client authentication.
5823 * ASN.1 structure reuse memory corruption fix
5825 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5826 memory corruption via an invalid write. Such reuse is and has been
5827 strongly discouraged and is believed to be rare.
5829 Applications that parse structures containing CHOICE or ANY DEFINED BY
5830 components may be affected. Certificate parsing (d2i_X509 and related
5831 functions) are however not affected. OpenSSL clients and servers are
5837 * PKCS7 NULL pointer dereferences fix
5839 The PKCS#7 parsing code does not handle missing outer ContentInfo
5840 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5841 missing content and trigger a NULL pointer dereference on parsing.
5843 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5844 otherwise parse PKCS#7 structures from untrusted sources are
5845 affected. OpenSSL clients and servers are not affected.
5847 This issue was reported to OpenSSL by Michal Zalewski (Google).
5852 * DoS via reachable assert in SSLv2 servers fix
5854 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5855 servers that both support SSLv2 and enable export cipher suites by sending
5856 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5858 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5859 (OpenSSL development team).
5864 * Use After Free following d2i_ECPrivatekey error fix
5866 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5867 could cause a use after free condition. This, in turn, could cause a double
5868 free in several private key parsing functions (such as d2i_PrivateKey
5869 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5870 for applications that receive EC private keys from untrusted
5871 sources. This scenario is considered rare.
5873 This issue was discovered by the BoringSSL project and fixed in their
5879 * X509_to_X509_REQ NULL pointer deref fix
5881 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5882 the certificate key is invalid. This function is rarely used in practice.
5884 This issue was discovered by Brian Carpenter.
5889 * Removed the export ciphers from the DEFAULT ciphers
5893 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015] ###
5895 * Build fixes for the Windows and OpenVMS platforms
5897 *Matt Caswell and Richard Levitte*
5899 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015] ###
5901 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
5902 message can cause a segmentation fault in OpenSSL due to a NULL pointer
5903 dereference. This could lead to a Denial Of Service attack. Thanks to
5904 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
5909 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
5910 dtls1_buffer_record function under certain conditions. In particular this
5911 could occur if an attacker sent repeated DTLS records with the same
5912 sequence number but for the next epoch. The memory leak could be exploited
5913 by an attacker in a Denial of Service attack through memory exhaustion.
5914 Thanks to Chris Mueller for reporting this issue.
5919 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
5920 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
5921 method would be set to NULL which could later result in a NULL pointer
5922 dereference. Thanks to Frank Schmirler for reporting this issue.
5927 * Abort handshake if server key exchange message is omitted for ephemeral
5930 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
5931 reporting this issue.
5936 * Remove non-export ephemeral RSA code on client and server. This code
5937 violated the TLS standard by allowing the use of temporary RSA keys in
5938 non-export ciphersuites and could be used by a server to effectively
5939 downgrade the RSA key length used to a value smaller than the server
5940 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
5941 INRIA or reporting this issue.
5946 * Fixed issue where DH client certificates are accepted without verification.
5947 An OpenSSL server will accept a DH certificate for client authentication
5948 without the certificate verify message. This effectively allows a client to
5949 authenticate without the use of a private key. This only affects servers
5950 which trust a client certificate authority which issues certificates
5951 containing DH keys: these are extremely rare and hardly ever encountered.
5952 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
5958 * Ensure that the session ID context of an SSL is updated when its
5959 SSL_CTX is updated via SSL_set_SSL_CTX.
5961 The session ID context is typically set from the parent SSL_CTX,
5962 and can vary with the CTX.
5966 * Fix various certificate fingerprint issues.
5968 By using non-DER or invalid encodings outside the signed portion of a
5969 certificate the fingerprint can be changed without breaking the signature.
5970 Although no details of the signed portion of the certificate can be changed
5971 this can cause problems with some applications: e.g. those using the
5972 certificate fingerprint for blacklists.
5974 1. Reject signatures with non zero unused bits.
5976 If the BIT STRING containing the signature has non zero unused bits reject
5977 the signature. All current signature algorithms require zero unused bits.
5979 2. Check certificate algorithm consistency.
5981 Check the AlgorithmIdentifier inside TBS matches the one in the
5982 certificate signature. NB: this will result in signature failure
5983 errors for some broken certificates.
5985 Thanks to Konrad Kraszewski from Google for reporting this issue.
5987 3. Check DSA/ECDSA signatures use DER.
5989 Re-encode DSA/ECDSA signatures and compare with the original received
5990 signature. Return an error if there is a mismatch.
5992 This will reject various cases including garbage after signature
5993 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
5994 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
5995 (negative or with leading zeroes).
5997 Further analysis was conducted and fixes were developed by Stephen Henson
5998 of the OpenSSL core team.
6004 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6005 results on some platforms, including x86_64. This bug occurs at random
6006 with a very low probability, and is not known to be exploitable in any
6007 way, though its exact impact is difficult to determine. Thanks to Pieter
6008 Wuille (Blockstream) who reported this issue and also suggested an initial
6009 fix. Further analysis was conducted by the OpenSSL development team and
6010 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6011 the OpenSSL core team.
6016 *) Do not resume sessions on the server if the negotiated protocol
6017 version does not match the session's version. Resuming with a different
6018 version, while not strictly forbidden by the RFC, is of questionable
6019 sanity and breaks all known clients.
6021 *David Benjamin, Emilia Käsper*
6023 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
6024 early CCS messages during renegotiation. (Note that because
6025 renegotiation is encrypted, this early CCS was not exploitable.)
6029 *) Tighten client-side session ticket handling during renegotiation:
6030 ensure that the client only accepts a session ticket if the server sends
6031 the extension anew in the ServerHello. Previously, a TLS client would
6032 reuse the old extension state and thus accept a session ticket if one was
6033 announced in the initial ServerHello.
6035 Similarly, ensure that the client requires a session ticket if one
6036 was advertised in the ServerHello. Previously, a TLS client would
6037 ignore a missing NewSessionTicket message.
6041 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014] ###
6045 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6046 sends a carefully crafted handshake message, to cause OpenSSL to fail
6047 to free up to 64k of memory causing a memory leak. This could be
6048 exploited in a Denial Of Service attack. This issue affects OpenSSL
6049 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6050 whether SRTP is used or configured. Implementations of OpenSSL that
6051 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6053 The fix was developed by the OpenSSL team.
6058 * Session Ticket Memory Leak.
6060 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6061 integrity of that ticket is first verified. In the event of a session
6062 ticket integrity check failing, OpenSSL will fail to free memory
6063 causing a memory leak. By sending a large number of invalid session
6064 tickets an attacker could exploit this issue in a Denial Of Service
6070 * Build option no-ssl3 is incomplete.
6072 When OpenSSL is configured with "no-ssl3" as a build option, servers
6073 could accept and complete a SSL 3.0 handshake, and clients could be
6074 configured to send them.
6077 *Akamai and the OpenSSL team*
6079 * Add support for TLS_FALLBACK_SCSV.
6080 Client applications doing fallback retries should call
6081 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6084 *Adam Langley, Bodo Moeller*
6086 * Add additional DigestInfo checks.
6088 Re-encode DigestInto in DER and check against the original when
6089 verifying RSA signature: this will reject any improperly encoded
6090 DigestInfo structures.
6092 Note: this is a precautionary measure and no attacks are currently known.
6097 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014] ###
6099 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6100 SRP code can be overrun an internal buffer. Add sanity check that
6101 g, A, B < N to SRP code.
6103 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6104 Group for discovering this issue.
6109 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6110 TLS 1.0 instead of higher protocol versions when the ClientHello message
6111 is badly fragmented. This allows a man-in-the-middle attacker to force a
6112 downgrade to TLS 1.0 even if both the server and the client support a
6113 higher protocol version, by modifying the client's TLS records.
6115 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6116 researching this issue.
6121 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6122 to a denial of service attack. A malicious server can crash the client
6123 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6124 ciphersuite and sending carefully crafted handshake messages.
6126 Thanks to Felix Gröbert (Google) for discovering and researching this
6132 * By sending carefully crafted DTLS packets an attacker could cause openssl
6133 to leak memory. This can be exploited through a Denial of Service attack.
6134 Thanks to Adam Langley for discovering and researching this issue.
6139 * An attacker can force openssl to consume large amounts of memory whilst
6140 processing DTLS handshake messages. This can be exploited through a
6141 Denial of Service attack.
6142 Thanks to Adam Langley for discovering and researching this issue.
6147 * An attacker can force an error condition which causes openssl to crash
6148 whilst processing DTLS packets due to memory being freed twice. This
6149 can be exploited through a Denial of Service attack.
6150 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6156 * If a multithreaded client connects to a malicious server using a resumed
6157 session and the server sends an ec point format extension it could write
6158 up to 255 bytes to freed memory.
6160 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6166 * A malicious server can crash an OpenSSL client with a null pointer
6167 dereference (read) by specifying an SRP ciphersuite even though it was not
6168 properly negotiated with the client. This can be exploited through a
6169 Denial of Service attack.
6171 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6172 discovering and researching this issue.
6177 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6178 X509_name_oneline, X509_name_print_ex et al. to leak some information
6179 from the stack. Applications may be affected if they echo pretty printing
6180 output to the attacker.
6182 Thanks to Ivan Fratric (Google) for discovering this issue.
6185 *Emilia Käsper, and Steve Henson*
6187 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6188 for corner cases. (Certain input points at infinity could lead to
6189 bogus results, with non-infinity inputs mapped to infinity too.)
6193 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014] ###
6195 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6196 handshake can force the use of weak keying material in OpenSSL
6197 SSL/TLS clients and servers.
6199 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6200 researching this issue. [CVE-2014-0224][]
6202 *KIKUCHI Masashi, Steve Henson*
6204 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6205 OpenSSL DTLS client the code can be made to recurse eventually crashing
6208 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6211 *Imre Rad, Steve Henson*
6213 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6214 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6215 client or server. This is potentially exploitable to run arbitrary
6216 code on a vulnerable client or server.
6218 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6220 *Jüri Aedla, Steve Henson*
6222 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6223 are subject to a denial of service attack.
6225 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6226 this issue. [CVE-2014-3470][]
6228 *Felix Gröbert, Ivan Fratric, Steve Henson*
6230 * Harmonize version and its documentation. -f flag is used to display
6233 *mancha <mancha1@zoho.com>*
6235 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6236 in i2d_ECPrivateKey.
6238 *mancha <mancha1@zoho.com>*
6240 * Fix some double frees. These are not thought to be exploitable.
6242 *mancha <mancha1@zoho.com>*
6244 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014] ###
6246 * A missing bounds check in the handling of the TLS heartbeat extension
6247 can be used to reveal up to 64k of memory to a connected client or
6250 Thanks for Neel Mehta of Google Security for discovering this bug and to
6251 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6252 preparing the fix [CVE-2014-0160][]
6254 *Adam Langley, Bodo Moeller*
6256 * Fix for the attack described in the paper "Recovering OpenSSL
6257 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6258 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6259 http://eprint.iacr.org/2014/140
6261 Thanks to Yuval Yarom and Naomi Benger for discovering this
6262 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6264 *Yuval Yarom and Naomi Benger*
6266 * TLS pad extension: draft-agl-tls-padding-03
6268 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6269 TLS client Hello record length value would otherwise be > 255 and
6270 less that 512 pad with a dummy extension containing zeroes so it
6271 is at least 512 bytes long.
6274 *Adam Langley, Steve Henson*
6276 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014] ###
6278 * Fix for TLS record tampering bug. A carefully crafted invalid
6279 handshake could crash OpenSSL with a NULL pointer exception.
6280 Thanks to Anton Johansson for reporting this issues.
6283 * Keep original DTLS digest and encryption contexts in retransmission
6284 structures so we can use the previous session parameters if they need
6285 to be resent. [CVE-2013-6450][]
6289 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6290 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6291 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6292 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6293 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6294 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6296 *Rob Stradling, Adam Langley*
6298 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013] ###
6300 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6301 supporting platforms or when small records were transferred.
6303 *Andy Polyakov, Steve Henson*
6305 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013] ###
6307 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6309 This addresses the flaw in CBC record processing discovered by
6310 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6311 at: http://www.isg.rhul.ac.uk/tls/
6313 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6314 Security Group at Royal Holloway, University of London
6315 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6316 Emilia Käsper for the initial patch.
6319 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6321 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6322 ciphersuites which can be exploited in a denial of service attack.
6323 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6324 and detecting this bug and to Wolfgang Ettlinger
6325 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6330 * Return an error when checking OCSP signatures when key is NULL.
6331 This fixes a DoS attack. [CVE-2013-0166][]
6335 * Make openssl verify return errors.
6337 *Chris Palmer <palmer@google.com> and Ben Laurie*
6339 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6340 the right response is stapled. Also change SSL_get_certificate()
6341 so it returns the certificate actually sent.
6342 See http://rt.openssl.org/Ticket/Display.html?id=2836.
6344 *Rob Stradling <rob.stradling@comodo.com>*
6346 * Fix possible deadlock when decoding public keys.
6350 * Don't use TLS 1.0 record version number in initial client hello
6355 ### Changes between 1.0.1b and 1.0.1c [10 May 2012] ###
6357 * Sanity check record length before skipping explicit IV in TLS
6358 1.2, 1.1 and DTLS to fix DoS attack.
6360 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6361 fuzzing as a service testing platform.
6366 * Initialise tkeylen properly when encrypting CMS messages.
6367 Thanks to Solar Designer of Openwall for reporting this issue.
6371 * In FIPS mode don't try to use composite ciphers as they are not
6376 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012] ###
6378 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6379 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6380 mean any application compiled against OpenSSL 1.0.0 headers setting
6381 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6382 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6383 0x10000000L Any application which was previously compiled against
6384 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6385 will need to be recompiled as a result. Letting be results in
6386 inability to disable specifically TLS 1.1 and in client context,
6387 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6391 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6392 disable just protocol X, but all protocols above X *if* there are
6393 protocols *below* X still enabled. In more practical terms it means
6394 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6395 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
6396 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
6401 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012] ###
6403 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6404 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6405 in CRYPTO_realloc_clean.
6407 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6408 issue and to Adam Langley <agl@chromium.org> for fixing it.
6411 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6413 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6417 * Workarounds for some broken servers that "hang" if a client hello
6418 record length exceeds 255 bytes.
6420 1. Do not use record version number > TLS 1.0 in initial client
6421 hello: some (but not all) hanging servers will now work.
6422 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6423 the number of ciphers sent in the client hello. This should be
6424 set to an even number, such as 50, for example by passing:
6425 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6426 Most broken servers should now work.
6427 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6428 TLS 1.2 client support entirely.
6432 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6436 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012] ###
6438 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6439 STRING form instead of a DigestInfo.
6443 * The format used for MDC2 RSA signatures is inconsistent between EVP
6444 and the RSA_sign/RSA_verify functions. This was made more apparent when
6445 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6446 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6447 the correct format in RSA_verify so both forms transparently work.
6451 * Some servers which support TLS 1.0 can choke if we initially indicate
6452 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6453 encrypted premaster secret. As a workaround use the maximum permitted
6454 client version in client hello, this should keep such servers happy
6455 and still work with previous versions of OpenSSL.
6459 * Add support for TLS/DTLS heartbeats.
6461 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6463 * Add support for SCTP.
6465 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6467 * Improved PRNG seeding for VOS.
6469 *Paul Green <Paul.Green@stratus.com>*
6471 * Extensive assembler packs updates, most notably:
6473 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6474 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6475 - x86_64: bit-sliced AES implementation;
6476 - ARM: NEON support, contemporary platforms optimizations;
6477 - s390x: z196 support;
6478 - *: GHASH and GF(2^m) multiplication implementations;
6482 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6483 (removal of unnecessary code)
6485 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6487 * Add TLS key material exporter from RFC 5705.
6491 * Add DTLS-SRTP negotiation from RFC 5764.
6495 * Add Next Protocol Negotiation,
6496 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
6497 disabled with a no-npn flag to config or Configure. Code donated
6500 *Adam Langley <agl@google.com> and Ben Laurie*
6502 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6503 NIST-P256, NIST-P521, with constant-time single point multiplication on
6504 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6505 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6506 Code made available under Apache License version 2.0.
6508 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6509 line to include this in your build of OpenSSL, and run "make depend" (or
6510 "make update"). This enables the following EC_METHODs:
6512 EC_GFp_nistp224_method()
6513 EC_GFp_nistp256_method()
6514 EC_GFp_nistp521_method()
6516 EC_GROUP_new_by_curve_name() will automatically use these (while
6517 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6520 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6522 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6523 all platforms. Move ssize_t definition from e_os.h to the public
6524 header file e_os2.h as it now appears in public header file cms.h
6528 * New -sigopt option to the ca, req and x509 utilities. Additional
6529 signature parameters can be passed using this option and in
6534 * Add RSA PSS signing function. This will generate and set the
6535 appropriate AlgorithmIdentifiers for PSS based on those in the
6536 corresponding EVP_MD_CTX structure. No application support yet.
6540 * Support for companion algorithm specific ASN1 signing routines.
6541 New function ASN1_item_sign_ctx() signs a pre-initialised
6542 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6543 the appropriate parameters.
6547 * Add new algorithm specific ASN1 verification initialisation function
6548 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6549 handling will be the same no matter what EVP_PKEY_METHOD is used.
6550 Add a PSS handler to support verification of PSS signatures: checked
6551 against a number of sample certificates.
6555 * Add signature printing for PSS. Add PSS OIDs.
6557 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6559 * Add algorithm specific signature printing. An individual ASN1 method
6560 can now print out signatures instead of the standard hex dump.
6562 More complex signatures (e.g. PSS) can print out more meaningful
6563 information. Include DSA version that prints out the signature
6568 * Password based recipient info support for CMS library: implementing
6573 * Split password based encryption into PBES2 and PBKDF2 functions. This
6574 neatly separates the code into cipher and PBE sections and is required
6575 for some algorithms that split PBES2 into separate pieces (such as
6576 password based CMS).
6580 * Session-handling fixes:
6581 - Fix handling of connections that are resuming with a session ID,
6582 but also support Session Tickets.
6583 - Fix a bug that suppressed issuing of a new ticket if the client
6584 presented a ticket with an expired session.
6585 - Try to set the ticket lifetime hint to something reasonable.
6586 - Make tickets shorter by excluding irrelevant information.
6587 - On the client side, don't ignore renewed tickets.
6589 *Adam Langley, Bodo Moeller (Google)*
6591 * Fix PSK session representation.
6595 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6597 This work was sponsored by Intel.
6601 * Add GCM support to TLS library. Some custom code is needed to split
6602 the IV between the fixed (from PRF) and explicit (from TLS record)
6603 portions. This adds all GCM ciphersuites supported by RFC5288 and
6604 RFC5289. Generalise some AES* cipherstrings to include GCM and
6605 add a special AESGCM string for GCM only.
6609 * Expand range of ctrls for AES GCM. Permit setting invocation
6610 field on decrypt and retrieval of invocation field only on encrypt.
6614 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6615 As required by RFC5289 these ciphersuites cannot be used if for
6616 versions of TLS earlier than 1.2.
6620 * For FIPS capable OpenSSL interpret a NULL default public key method
6621 as unset and return the appropriate default but do *not* set the default.
6622 This means we can return the appropriate method in applications that
6623 switch between FIPS and non-FIPS modes.
6627 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6628 ENGINE is used then we cannot handle that in the FIPS module so we
6629 keep original code iff non-FIPS operations are allowed.
6633 * Add -attime option to openssl utilities.
6635 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6637 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6641 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6642 FIPS EC methods unconditionally for now.
6646 * New build option no-ec2m to disable characteristic 2 code.
6650 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6651 all cases can be covered as some introduce binary incompatibilities.
6655 * Redirect RSA operations to FIPS module including keygen,
6656 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6660 * Add similar low level API blocking to ciphers.
6664 * Low level digest APIs are not approved in FIPS mode: any attempt
6665 to use these will cause a fatal error. Applications that *really* want
6666 to use them can use the private_* version instead.
6670 * Redirect cipher operations to FIPS module for FIPS builds.
6674 * Redirect digest operations to FIPS module for FIPS builds.
6678 * Update build system to add "fips" flag which will link in fipscanister.o
6679 for static and shared library builds embedding a signature if needed.
6683 * Output TLS supported curves in preference order instead of numerical
6684 order. This is currently hardcoded for the highest order curves first.
6685 This should be configurable so applications can judge speed vs strength.
6689 * Add TLS v1.2 server support for client authentication.
6693 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6698 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6699 FIPS modules versions.
6703 * Add TLS v1.2 client side support for client authentication. Keep cache
6704 of handshake records longer as we don't know the hash algorithm to use
6705 until after the certificate request message is received.
6709 * Initial TLS v1.2 client support. Add a default signature algorithms
6710 extension including all the algorithms we support. Parse new signature
6711 format in client key exchange. Relax some ECC signing restrictions for
6712 TLS v1.2 as indicated in RFC5246.
6716 * Add server support for TLS v1.2 signature algorithms extension. Switch
6717 to new signature format when needed using client digest preference.
6718 All server ciphersuites should now work correctly in TLS v1.2. No client
6719 support yet and no support for client certificates.
6723 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6724 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6725 ciphersuites. At present only RSA key exchange ciphersuites work with
6726 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6727 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6728 and version checking.
6732 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6733 with this defined it will not be affected by any changes to ssl internal
6734 structures. Add several utility functions to allow openssl application
6735 to work with OPENSSL_NO_SSL_INTERN defined.
6739 * A long standing patch to add support for SRP from EdelWeb (Peter
6740 Sylvester and Christophe Renou) was integrated.
6741 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6742 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6745 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6749 * Permit abbreviated handshakes when renegotiating using the function
6750 SSL_renegotiate_abbreviated().
6752 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6754 * Add call to ENGINE_register_all_complete() to
6755 ENGINE_load_builtin_engines(), so some implementations get used
6756 automatically instead of needing explicit application support.
6760 * Add support for TLS key exporter as described in RFC5705.
6762 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6764 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6765 a few changes are required:
6767 Add SSL_OP_NO_TLSv1_1 flag.
6768 Add TLSv1_1 methods.
6769 Update version checking logic to handle version 1.1.
6770 Add explicit IV handling (ported from DTLS code).
6771 Add command line options to s_client/s_server.
6778 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015] ###
6780 * X509_ATTRIBUTE memory leak
6782 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6783 memory. This structure is used by the PKCS#7 and CMS routines so any
6784 application which reads PKCS#7 or CMS data from untrusted sources is
6785 affected. SSL/TLS is not affected.
6787 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6793 * Race condition handling PSK identify hint
6795 If PSK identity hints are received by a multi-threaded client then
6796 the values are wrongly updated in the parent SSL_CTX structure. This can
6797 result in a race condition potentially leading to a double free of the
6803 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015] ###
6805 * Malformed ECParameters causes infinite loop
6807 When processing an ECParameters structure OpenSSL enters an infinite loop
6808 if the curve specified is over a specially malformed binary polynomial
6811 This can be used to perform denial of service against any
6812 system which processes public keys, certificate requests or
6813 certificates. This includes TLS clients and TLS servers with
6814 client authentication enabled.
6816 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6821 * Exploitable out-of-bounds read in X509_cmp_time
6823 X509_cmp_time does not properly check the length of the ASN1_TIME
6824 string and can read a few bytes out of bounds. In addition,
6825 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6828 An attacker can use this to craft malformed certificates and CRLs of
6829 various sizes and potentially cause a segmentation fault, resulting in
6830 a DoS on applications that verify certificates or CRLs. TLS clients
6831 that verify CRLs are affected. TLS clients and servers with client
6832 authentication enabled may be affected if they use custom verification
6835 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6836 independently by Hanno Böck.
6841 * PKCS7 crash with missing EnvelopedContent
6843 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6844 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6845 with missing content and trigger a NULL pointer dereference on parsing.
6847 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6848 structures from untrusted sources are affected. OpenSSL clients and
6849 servers are not affected.
6851 This issue was reported to OpenSSL by Michal Zalewski (Google).
6856 * CMS verify infinite loop with unknown hash function
6858 When verifying a signedData message the CMS code can enter an infinite loop
6859 if presented with an unknown hash function OID. This can be used to perform
6860 denial of service against any system which verifies signedData messages using
6862 This issue was reported to OpenSSL by Johannes Bauer.
6867 * Race condition handling NewSessionTicket
6869 If a NewSessionTicket is received by a multi-threaded client when attempting to
6870 reuse a previous ticket then a race condition can occur potentially leading to
6871 a double free of the ticket data.
6876 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015] ###
6878 * Segmentation fault in ASN1_TYPE_cmp fix
6880 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6881 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6882 certificate signature algorithm consistency this can be used to crash any
6883 certificate verification operation and exploited in a DoS attack. Any
6884 application which performs certificate verification is vulnerable including
6885 OpenSSL clients and servers which enable client authentication.
6890 * ASN.1 structure reuse memory corruption fix
6892 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6893 memory corruption via an invalid write. Such reuse is and has been
6894 strongly discouraged and is believed to be rare.
6896 Applications that parse structures containing CHOICE or ANY DEFINED BY
6897 components may be affected. Certificate parsing (d2i_X509 and related
6898 functions) are however not affected. OpenSSL clients and servers are
6904 * PKCS7 NULL pointer dereferences fix
6906 The PKCS#7 parsing code does not handle missing outer ContentInfo
6907 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6908 missing content and trigger a NULL pointer dereference on parsing.
6910 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6911 otherwise parse PKCS#7 structures from untrusted sources are
6912 affected. OpenSSL clients and servers are not affected.
6914 This issue was reported to OpenSSL by Michal Zalewski (Google).
6919 * DoS via reachable assert in SSLv2 servers fix
6921 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6922 servers that both support SSLv2 and enable export cipher suites by sending
6923 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6925 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6926 (OpenSSL development team).
6931 * Use After Free following d2i_ECPrivatekey error fix
6933 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6934 could cause a use after free condition. This, in turn, could cause a double
6935 free in several private key parsing functions (such as d2i_PrivateKey
6936 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6937 for applications that receive EC private keys from untrusted
6938 sources. This scenario is considered rare.
6940 This issue was discovered by the BoringSSL project and fixed in their
6946 * X509_to_X509_REQ NULL pointer deref fix
6948 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6949 the certificate key is invalid. This function is rarely used in practice.
6951 This issue was discovered by Brian Carpenter.
6956 * Removed the export ciphers from the DEFAULT ciphers
6960 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015] ###
6962 * Build fixes for the Windows and OpenVMS platforms
6964 *Matt Caswell and Richard Levitte*
6966 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015] ###
6968 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6969 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6970 dereference. This could lead to a Denial Of Service attack. Thanks to
6971 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6976 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6977 dtls1_buffer_record function under certain conditions. In particular this
6978 could occur if an attacker sent repeated DTLS records with the same
6979 sequence number but for the next epoch. The memory leak could be exploited
6980 by an attacker in a Denial of Service attack through memory exhaustion.
6981 Thanks to Chris Mueller for reporting this issue.
6986 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6987 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6988 method would be set to NULL which could later result in a NULL pointer
6989 dereference. Thanks to Frank Schmirler for reporting this issue.
6994 * Abort handshake if server key exchange message is omitted for ephemeral
6997 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6998 reporting this issue.
7003 * Remove non-export ephemeral RSA code on client and server. This code
7004 violated the TLS standard by allowing the use of temporary RSA keys in
7005 non-export ciphersuites and could be used by a server to effectively
7006 downgrade the RSA key length used to a value smaller than the server
7007 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7008 INRIA or reporting this issue.
7013 * Fixed issue where DH client certificates are accepted without verification.
7014 An OpenSSL server will accept a DH certificate for client authentication
7015 without the certificate verify message. This effectively allows a client to
7016 authenticate without the use of a private key. This only affects servers
7017 which trust a client certificate authority which issues certificates
7018 containing DH keys: these are extremely rare and hardly ever encountered.
7019 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7025 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7026 results on some platforms, including x86_64. This bug occurs at random
7027 with a very low probability, and is not known to be exploitable in any
7028 way, though its exact impact is difficult to determine. Thanks to Pieter
7029 Wuille (Blockstream) who reported this issue and also suggested an initial
7030 fix. Further analysis was conducted by the OpenSSL development team and
7031 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7032 the OpenSSL core team.
7037 * Fix various certificate fingerprint issues.
7039 By using non-DER or invalid encodings outside the signed portion of a
7040 certificate the fingerprint can be changed without breaking the signature.
7041 Although no details of the signed portion of the certificate can be changed
7042 this can cause problems with some applications: e.g. those using the
7043 certificate fingerprint for blacklists.
7045 1. Reject signatures with non zero unused bits.
7047 If the BIT STRING containing the signature has non zero unused bits reject
7048 the signature. All current signature algorithms require zero unused bits.
7050 2. Check certificate algorithm consistency.
7052 Check the AlgorithmIdentifier inside TBS matches the one in the
7053 certificate signature. NB: this will result in signature failure
7054 errors for some broken certificates.
7056 Thanks to Konrad Kraszewski from Google for reporting this issue.
7058 3. Check DSA/ECDSA signatures use DER.
7060 Reencode DSA/ECDSA signatures and compare with the original received
7061 signature. Return an error if there is a mismatch.
7063 This will reject various cases including garbage after signature
7064 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7065 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7066 (negative or with leading zeroes).
7068 Further analysis was conducted and fixes were developed by Stephen Henson
7069 of the OpenSSL core team.
7075 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014] ###
7077 * Session Ticket Memory Leak.
7079 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7080 integrity of that ticket is first verified. In the event of a session
7081 ticket integrity check failing, OpenSSL will fail to free memory
7082 causing a memory leak. By sending a large number of invalid session
7083 tickets an attacker could exploit this issue in a Denial Of Service
7089 * Build option no-ssl3 is incomplete.
7091 When OpenSSL is configured with "no-ssl3" as a build option, servers
7092 could accept and complete a SSL 3.0 handshake, and clients could be
7093 configured to send them.
7096 *Akamai and the OpenSSL team*
7098 * Add support for TLS_FALLBACK_SCSV.
7099 Client applications doing fallback retries should call
7100 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7103 *Adam Langley, Bodo Moeller*
7105 * Add additional DigestInfo checks.
7107 Reencode DigestInto in DER and check against the original when
7108 verifying RSA signature: this will reject any improperly encoded
7109 DigestInfo structures.
7111 Note: this is a precautionary measure and no attacks are currently known.
7116 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014] ###
7118 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7119 to a denial of service attack. A malicious server can crash the client
7120 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7121 ciphersuite and sending carefully crafted handshake messages.
7123 Thanks to Felix Gröbert (Google) for discovering and researching this
7129 * By sending carefully crafted DTLS packets an attacker could cause openssl
7130 to leak memory. This can be exploited through a Denial of Service attack.
7131 Thanks to Adam Langley for discovering and researching this issue.
7136 * An attacker can force openssl to consume large amounts of memory whilst
7137 processing DTLS handshake messages. This can be exploited through a
7138 Denial of Service attack.
7139 Thanks to Adam Langley for discovering and researching this issue.
7144 * An attacker can force an error condition which causes openssl to crash
7145 whilst processing DTLS packets due to memory being freed twice. This
7146 can be exploited through a Denial of Service attack.
7147 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7153 * If a multithreaded client connects to a malicious server using a resumed
7154 session and the server sends an ec point format extension it could write
7155 up to 255 bytes to freed memory.
7157 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7163 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7164 X509_name_oneline, X509_name_print_ex et al. to leak some information
7165 from the stack. Applications may be affected if they echo pretty printing
7166 output to the attacker.
7168 Thanks to Ivan Fratric (Google) for discovering this issue.
7171 *Emilia Käsper, and Steve Henson*
7173 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7174 for corner cases. (Certain input points at infinity could lead to
7175 bogus results, with non-infinity inputs mapped to infinity too.)
7179 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014] ###
7181 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7182 handshake can force the use of weak keying material in OpenSSL
7183 SSL/TLS clients and servers.
7185 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7186 researching this issue. [CVE-2014-0224][]
7188 *KIKUCHI Masashi, Steve Henson*
7190 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7191 OpenSSL DTLS client the code can be made to recurse eventually crashing
7194 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7197 *Imre Rad, Steve Henson*
7199 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7200 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7201 client or server. This is potentially exploitable to run arbitrary
7202 code on a vulnerable client or server.
7204 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7206 *Jüri Aedla, Steve Henson*
7208 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7209 are subject to a denial of service attack.
7211 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7212 this issue. [CVE-2014-3470][]
7214 *Felix Gröbert, Ivan Fratric, Steve Henson*
7216 * Harmonize version and its documentation. -f flag is used to display
7219 *mancha <mancha1@zoho.com>*
7221 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7222 in i2d_ECPrivateKey.
7224 *mancha <mancha1@zoho.com>*
7226 * Fix some double frees. These are not thought to be exploitable.
7228 *mancha <mancha1@zoho.com>*
7230 * Fix for the attack described in the paper "Recovering OpenSSL
7231 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7232 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7233 http://eprint.iacr.org/2014/140
7235 Thanks to Yuval Yarom and Naomi Benger for discovering this
7236 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7238 *Yuval Yarom and Naomi Benger*
7240 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014] ###
7242 * Keep original DTLS digest and encryption contexts in retransmission
7243 structures so we can use the previous session parameters if they need
7244 to be resent. [CVE-2013-6450][]
7248 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7249 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7250 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7251 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7252 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7253 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7255 *Rob Stradling, Adam Langley*
7257 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013] ###
7259 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7261 This addresses the flaw in CBC record processing discovered by
7262 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7263 at: http://www.isg.rhul.ac.uk/tls/
7265 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7266 Security Group at Royal Holloway, University of London
7267 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7268 Emilia Käsper for the initial patch.
7271 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7273 * Return an error when checking OCSP signatures when key is NULL.
7274 This fixes a DoS attack. [CVE-2013-0166][]
7278 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7279 the right response is stapled. Also change SSL_get_certificate()
7280 so it returns the certificate actually sent.
7281 See http://rt.openssl.org/Ticket/Display.html?id=2836.
7282 (This is a backport)
7284 *Rob Stradling <rob.stradling@comodo.com>*
7286 * Fix possible deadlock when decoding public keys.
7290 ### Changes between 1.0.0i and 1.0.0j [10 May 2012] ###
7292 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7295 * Sanity check record length before skipping explicit IV in DTLS
7298 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7299 fuzzing as a service testing platform.
7304 * Initialise tkeylen properly when encrypting CMS messages.
7305 Thanks to Solar Designer of Openwall for reporting this issue.
7309 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012] ###
7311 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7312 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7313 in CRYPTO_realloc_clean.
7315 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7316 issue and to Adam Langley <agl@chromium.org> for fixing it.
7319 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7321 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012] ###
7323 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7324 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7325 content decryption and always return the same error. Note: this attack
7326 needs on average 2^20 messages so it only affects automated senders. The
7327 old behaviour can be re-enabled in the CMS code by setting the
7328 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7329 an MMA defence is not necessary.
7330 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7331 this issue. [CVE-2012-0884][]
7335 * Fix CVE-2011-4619: make sure we really are receiving a
7336 client hello before rejecting multiple SGC restarts. Thanks to
7337 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7341 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012] ###
7343 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7344 Thanks to Antonio Martin, Enterprise Secure Access Research and
7345 Development, Cisco Systems, Inc. for discovering this bug and
7346 preparing a fix. [CVE-2012-0050][]
7350 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012] ###
7352 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7353 of the Vaudenay padding oracle attack on CBC mode encryption
7354 which enables an efficient plaintext recovery attack against
7355 the OpenSSL implementation of DTLS. Their attack exploits timing
7356 differences arising during decryption processing. A research
7357 paper describing this attack can be found at:
7358 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
7359 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7360 Security Group at Royal Holloway, University of London
7361 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7362 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7363 for preparing the fix. [CVE-2011-4108][]
7365 *Robin Seggelmann, Michael Tuexen*
7367 * Clear bytes used for block padding of SSL 3.0 records.
7370 *Adam Langley (Google)*
7372 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7373 Kadianakis <desnacked@gmail.com> for discovering this issue and
7374 Adam Langley for preparing the fix. [CVE-2011-4619][]
7376 *Adam Langley (Google)*
7378 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7380 *Andrey Kulikov <amdeich@gmail.com>*
7382 * Prevent malformed RFC3779 data triggering an assertion failure.
7383 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7384 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7386 *Rob Austein <sra@hactrn.net>*
7388 * Improved PRNG seeding for VOS.
7390 *Paul Green <Paul.Green@stratus.com>*
7392 * Fix ssl_ciph.c set-up race.
7394 *Adam Langley (Google)*
7396 * Fix spurious failures in ecdsatest.c.
7398 *Emilia Käsper (Google)*
7400 * Fix the BIO_f_buffer() implementation (which was mixing different
7401 interpretations of the '..._len' fields).
7403 *Adam Langley (Google)*
7405 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7406 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7407 threads won't reuse the same blinding coefficients.
7409 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7410 lock to call BN_BLINDING_invert_ex, and avoids one use of
7411 BN_BLINDING_update for each BN_BLINDING structure (previously,
7412 the last update always remained unused).
7414 *Emilia Käsper (Google)*
7416 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7418 *Bob Buckholz (Google)*
7420 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011] ###
7422 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7423 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7425 *Kaspar Brand <ossl@velox.ch>*
7427 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7428 for multi-threaded use of ECDH. [CVE-2011-3210][]
7430 *Adam Langley (Google)*
7432 * Fix x509_name_ex_d2i memory leak on bad inputs.
7436 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7437 signature public key algorithm by using OID xref utilities instead.
7438 Before this you could only use some ECC ciphersuites with SHA1 only.
7442 * Add protection against ECDSA timing attacks as mentioned in the paper
7443 by Billy Bob Brumley and Nicola Tuveri, see:
7445 http://eprint.iacr.org/2011/232.pdf
7448 *Billy Bob Brumley and Nicola Tuveri*
7450 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011] ###
7452 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7454 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7456 * Fix bug in string printing code: if *any* escaping is enabled we must
7457 escape the escape character (backslash) or the resulting string is
7462 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010] ###
7464 * Disable code workaround for ancient and obsolete Netscape browsers
7465 and servers: an attacker can use it in a ciphersuite downgrade attack.
7466 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7470 * Fixed J-PAKE implementation error, originally discovered by
7471 Sebastien Martini, further info and confirmation from Stefan
7472 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7476 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010] ###
7478 * Fix extension code to avoid race conditions which can result in a buffer
7479 overrun vulnerability: resumed sessions must not be modified as they can
7480 be shared by multiple threads. CVE-2010-3864
7484 * Fix WIN32 build system to correctly link an ENGINE directory into
7489 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010] ###
7491 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7494 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7496 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010] ###
7498 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7499 context. The operation can be customised via the ctrl mechanism in
7500 case ENGINEs want to include additional functionality.
7504 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7508 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7509 output hashes compatible with older versions of OpenSSL.
7511 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7513 * Fix compression algorithm handling: if resuming a session use the
7514 compression algorithm of the resumed session instead of determining
7515 it from client hello again. Don't allow server to change algorithm.
7519 * Add load_crls() function to apps tidying load_certs() too. Add option
7520 to verify utility to allow additional CRLs to be included.
7524 * Update OCSP request code to permit adding custom headers to the request:
7525 some responders need this.
7529 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7532 *Julia Lawall <julia@diku.dk>*
7534 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7535 needlessly dereferenced structures, used obsolete functions and
7536 didn't handle all updated verify codes correctly.
7540 * Disable MD2 in the default configuration.
7544 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7545 indicate the initial BIO being pushed or popped. This makes it possible
7546 to determine whether the BIO is the one explicitly called or as a result
7547 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7548 it handles reference counts correctly and doesn't zero out the I/O bio
7549 when it is not being explicitly popped. WARNING: applications which
7550 included workarounds for the old buggy behaviour will need to be modified
7551 or they could free up already freed BIOs.
7555 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7556 renaming to all platforms (within the 0.9.8 branch, this was
7557 done conditionally on Netware platforms to avoid a name clash).
7559 *Guenter <lists@gknw.net>*
7561 * Add ECDHE and PSK support to DTLS.
7563 *Michael Tuexen <tuexen@fh-muenster.de>*
7565 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7570 * Add "missing" function EVP_MD_flags() (without this the only way to
7571 retrieve a digest flags is by accessing the structure directly. Update
7572 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
7573 or cipher is registered as in the "from" argument. Print out all
7574 registered digests in the dgst usage message instead of manually
7575 attempting to work them out.
7579 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7580 this allows the use of compression and extensions. Change default cipher
7581 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7582 by default unless an application cipher string requests it.
7586 * Alter match criteria in PKCS12_parse(). It used to try to use local
7587 key ids to find matching certificates and keys but some PKCS#12 files
7588 don't follow the (somewhat unwritten) rules and this strategy fails.
7589 Now just gather all certificates together and the first private key
7590 then look for the first certificate that matches the key.
7594 * Support use of registered digest and cipher names for dgst and cipher
7595 commands instead of having to add each one as a special case. So now
7602 openssl dgst -sha256 foo
7604 and this works for ENGINE based algorithms too.
7609 * Update Gost ENGINE to support parameter files.
7611 *Victor B. Wagner <vitus@cryptocom.ru>*
7613 * Support GeneralizedTime in ca utility.
7615 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7617 * Enhance the hash format used for certificate directory links. The new
7618 form uses the canonical encoding (meaning equivalent names will work
7619 even if they aren't identical) and uses SHA1 instead of MD5. This form
7620 is incompatible with the older format and as a result c_rehash should
7621 be used to rebuild symbolic links.
7625 * Make PKCS#8 the default write format for private keys, replacing the
7626 traditional format. This form is standardised, more secure and doesn't
7627 include an implicit MD5 dependency.
7631 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7632 committed to OpenSSL should pass this lot as a minimum.
7636 * Add session ticket override functionality for use by EAP-FAST.
7638 *Jouni Malinen <j@w1.fi>*
7640 * Modify HMAC functions to return a value. Since these can be implemented
7641 in an ENGINE errors can occur.
7645 * Type-checked OBJ_bsearch_ex.
7649 * Type-checked OBJ_bsearch. Also some constification necessitated
7650 by type-checking. Still to come: TXT_DB, bsearch(?),
7651 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7656 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7657 seconds to a tm structure directly, instead of going through OS
7658 specific date routines. This avoids any issues with OS routines such
7659 as the year 2038 bug. New *_adj() functions for ASN1 time structures
7660 and X509_time_adj_ex() to cover the extended range. The existing
7661 X509_time_adj() is still usable and will no longer have any date issues.
7665 * Delta CRL support. New use deltas option which will attempt to locate
7666 and search any appropriate delta CRLs available.
7668 This work was sponsored by Google.
7672 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7673 code and add additional score elements. Validate alternate CRL paths
7674 as part of the CRL checking and indicate a new error "CRL path validation
7675 error" in this case. Applications wanting additional details can use
7676 the verify callback and check the new "parent" field. If this is not
7677 NULL CRL path validation is taking place. Existing applications won't
7678 see this because it requires extended CRL support which is off by
7681 This work was sponsored by Google.
7685 * Support for freshest CRL extension.
7687 This work was sponsored by Google.
7691 * Initial indirect CRL support. Currently only supported in the CRLs
7692 passed directly and not via lookup. Process certificate issuer
7693 CRL entry extension and lookup CRL entries by bother issuer name
7694 and serial number. Check and process CRL issuer entry in IDP extension.
7696 This work was sponsored by Google.
7700 * Add support for distinct certificate and CRL paths. The CRL issuer
7701 certificate is validated separately in this case. Only enabled if
7702 an extended CRL support flag is set: this flag will enable additional
7703 CRL functionality in future.
7705 This work was sponsored by Google.
7709 * Add support for policy mappings extension.
7711 This work was sponsored by Google.
7715 * Fixes to pathlength constraint, self issued certificate handling,
7716 policy processing to align with RFC3280 and PKITS tests.
7718 This work was sponsored by Google.
7722 * Support for name constraints certificate extension. DN, email, DNS
7723 and URI types are currently supported.
7725 This work was sponsored by Google.
7729 * To cater for systems that provide a pointer-based thread ID rather
7730 than numeric, deprecate the current numeric thread ID mechanism and
7731 replace it with a structure and associated callback type. This
7732 mechanism allows a numeric "hash" to be extracted from a thread ID in
7733 either case, and on platforms where pointers are larger than 'long',
7734 mixing is done to help ensure the numeric 'hash' is usable even if it
7735 can't be guaranteed unique. The default mechanism is to use "&errno"
7736 as a pointer-based thread ID to distinguish between threads.
7738 Applications that want to provide their own thread IDs should now use
7739 CRYPTO_THREADID_set_callback() to register a callback that will call
7740 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7742 Note that ERR_remove_state() is now deprecated, because it is tied
7743 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7744 to free the current thread's error state should be replaced by
7745 ERR_remove_thread_state(NULL).
7747 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7748 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7749 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7750 application was previously providing a numeric thread callback that
7751 was inappropriate for distinguishing threads, then uniqueness might
7752 have been obtained with &errno that happened immediately in the
7753 intermediate development versions of OpenSSL; this is no longer the
7754 case, the numeric thread callback will now override the automatic use
7757 *Geoff Thorpe, with help from Bodo Moeller*
7759 * Initial support for different CRL issuing certificates. This covers a
7760 simple case where the self issued certificates in the chain exist and
7761 the real CRL issuer is higher in the existing chain.
7763 This work was sponsored by Google.
7767 * Removed effectively defunct crypto/store from the build.
7771 * Revamp of STACK to provide stronger type-checking. Still to come:
7772 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7773 ASN1_STRING, CONF_VALUE.
7777 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7778 RAM on SSL connections. This option can save about 34k per idle SSL.
7782 * Revamp of LHASH to provide stronger type-checking. Still to come:
7783 STACK, TXT_DB, bsearch, qsort.
7787 * Initial support for Cryptographic Message Syntax (aka CMS) based
7788 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7789 support for data, signedData, compressedData, digestedData and
7790 encryptedData, envelopedData types included. Scripts to check against
7791 RFC4134 examples draft and interop and consistency checks of many
7792 content types and variants.
7796 * Add options to enc utility to support use of zlib compression BIO.
7800 * Extend mk1mf to support importing of options and assembly language
7801 files from Configure script, currently only included in VC-WIN32.
7802 The assembly language rules can now optionally generate the source
7803 files from the associated perl scripts.
7807 * Implement remaining functionality needed to support GOST ciphersuites.
7808 Interop testing has been performed using CryptoPro implementations.
7810 *Victor B. Wagner <vitus@cryptocom.ru>*
7812 * s390x assembler pack.
7816 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7821 * Implement Opaque PRF Input TLS extension as specified in
7822 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7823 official specification yet and no extension type assignment by
7824 IANA exists, this extension (for now) will have to be explicitly
7825 enabled when building OpenSSL by providing the extension number
7826 to use. For example, specify an option
7828 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7830 to the "config" or "Configure" script to enable the extension,
7831 assuming extension number 0x9527 (which is a completely arbitrary
7832 and unofficial assignment based on the MD5 hash of the Internet
7833 Draft). Note that by doing so, you potentially lose
7834 interoperability with other TLS implementations since these might
7835 be using the same extension number for other purposes.
7837 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7838 opaque PRF input value to use in the handshake. This will create
7839 an internal copy of the length-'len' string at 'src', and will
7840 return non-zero for success.
7842 To get more control and flexibility, provide a callback function
7845 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7846 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7850 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7853 Callback function 'cb' will be called in handshakes, and is
7854 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7855 Argument 'arg' is for application purposes (the value as given to
7856 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7857 be provided to the callback function). The callback function
7858 has to return non-zero to report success: usually 1 to use opaque
7859 PRF input just if possible, or 2 to enforce use of the opaque PRF
7860 input. In the latter case, the library will abort the handshake
7861 if opaque PRF input is not successfully negotiated.
7863 Arguments 'peerinput' and 'len' given to the callback function
7864 will always be NULL and 0 in the case of a client. A server will
7865 see the client's opaque PRF input through these variables if
7866 available (NULL and 0 otherwise). Note that if the server
7867 provides an opaque PRF input, the length must be the same as the
7868 length of the client's opaque PRF input.
7870 Note that the callback function will only be called when creating
7871 a new session (session resumption can resume whatever was
7872 previously negotiated), and will not be called in SSL 2.0
7873 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7874 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7875 for applications that need to enforce opaque PRF input.
7880 * Update ssl code to support digests other than SHA1+MD5 for handshake
7884 *Victor B. Wagner <vitus@cryptocom.ru>*
7886 * Add RFC4507 support to OpenSSL. This includes the corrections in
7887 RFC4507bis. The encrypted ticket format is an encrypted encoded
7888 SSL_SESSION structure, that way new session features are automatically
7891 If a client application caches session in an SSL_SESSION structure
7892 support is transparent because tickets are now stored in the encoded
7895 The SSL_CTX structure automatically generates keys for ticket
7896 protection in servers so again support should be possible
7897 with no application modification.
7899 If a client or server wishes to disable RFC4507 support then the option
7900 SSL_OP_NO_TICKET can be set.
7902 Add a TLS extension debugging callback to allow the contents of any client
7903 or server extensions to be examined.
7905 This work was sponsored by Google.
7909 * Final changes to avoid use of pointer pointer casts in OpenSSL.
7910 OpenSSL should now compile cleanly on gcc 4.2
7912 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
7914 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
7915 support including streaming MAC support: this is required for GOST
7916 ciphersuite support.
7918 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
7920 * Add option -stream to use PKCS#7 streaming in smime utility. New
7921 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
7922 to output in BER and PEM format.
7926 * Experimental support for use of HMAC via EVP_PKEY interface. This
7927 allows HMAC to be handled via the EVP_DigestSign*() interface. The
7928 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
7929 ENGINE support for HMAC keys which are unextractable. New -mac and
7930 -macopt options to dgst utility.
7934 * New option -sigopt to dgst utility. Update dgst to use
7935 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
7936 alternative signing parameters such as X9.31 or PSS in the dgst
7941 * Change ssl_cipher_apply_rule(), the internal function that does
7942 the work each time a ciphersuite string requests enabling
7943 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
7944 removing ("!foo+bar") a class of ciphersuites: Now it maintains
7945 the order of disabled ciphersuites such that those ciphersuites
7946 that most recently went from enabled to disabled not only stay
7947 in order with respect to each other, but also have higher priority
7948 than other disabled ciphersuites the next time ciphersuites are
7951 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
7952 the same ciphersuites as with "HIGH" alone, but in a specific
7953 order where the PSK ciphersuites come first (since they are the
7954 most recently disabled ciphersuites when "HIGH" is parsed).
7956 Also, change ssl_create_cipher_list() (using this new
7957 functionality) such that between otherwise identical
7958 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
7963 * Change ssl_create_cipher_list() so that it automatically
7964 arranges the ciphersuites in reasonable order before starting
7965 to process the rule string. Thus, the definition for "DEFAULT"
7966 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
7967 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
7968 This makes it much easier to arrive at a reasonable default order
7969 in applications for which anonymous ciphers are OK (meaning
7970 that you can't actually use DEFAULT).
7972 *Bodo Moeller; suggested by Victor Duchovni*
7974 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
7975 processing) into multiple integers instead of setting
7976 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
7977 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
7978 (These masks as well as the individual bit definitions are hidden
7979 away into the non-exported interface ssl/ssl_locl.h, so this
7980 change to the definition of the SSL_CIPHER structure shouldn't
7981 affect applications.) This give us more bits for each of these
7982 categories, so there is no longer a need to coagulate AES128 and
7983 AES256 into a single algorithm bit, and to coagulate Camellia128
7984 and Camellia256 into a single algorithm bit, which has led to all
7987 Thus, among other things, the kludge introduced in 0.9.7m and
7988 0.9.8e for masking out AES256 independently of AES128 or masking
7989 out Camellia256 independently of AES256 is not needed here in 0.9.9.
7991 With the change, we also introduce new ciphersuite aliases that
7992 so far were missing: "AES128", "AES256", "CAMELLIA128", and
7997 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
7998 Use the leftmost N bytes of the signature input if the input is
7999 larger than the prime q (with N being the size in bytes of q).
8003 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8004 it yet and it is largely untested.
8008 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8012 * Initial incomplete changes to avoid need for function casts in OpenSSL
8013 some compilers (gcc 4.2 and later) reject their use. Safestack is
8014 reimplemented. Update ASN1 to avoid use of legacy functions.
8018 * Win32/64 targets are linked with Winsock2.
8022 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8023 to external functions. This can be used to increase CRL handling
8024 efficiency especially when CRLs are very large by (for example) storing
8025 the CRL revoked certificates in a database.
8029 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8030 new CRLs added to a directory can be used. New command line option
8031 -verify_return_error to s_client and s_server. This causes real errors
8032 to be returned by the verify callback instead of carrying on no matter
8033 what. This reflects the way a "real world" verify callback would behave.
8037 * GOST engine, supporting several GOST algorithms and public key formats.
8038 Kindly donated by Cryptocom.
8042 * Partial support for Issuing Distribution Point CRL extension. CRLs
8043 partitioned by DP are handled but no indirect CRL or reason partitioning
8044 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8045 selected via a scoring technique which handles IDP and AKID in CRLs.
8049 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8050 will ultimately be used for all verify operations: this will remove the
8051 X509_STORE dependency on certificate verification and allow alternative
8052 lookup methods. X509_STORE based implementations of these two callbacks.
8056 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8057 Modify get_crl() to find a valid (unexpired) CRL if possible.
8061 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8062 this would be called X509_CRL_cmp() but that name is already used by
8063 a function that just compares CRL issuer names. Cache several CRL
8064 extensions in X509_CRL structure and cache CRLDP in X509.
8068 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8069 this maps equivalent X509_NAME structures into a consistent structure.
8070 Name comparison can then be performed rapidly using memcmp().
8074 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8079 * Allow digests to supply their own micalg string for S/MIME type using
8080 the ctrl EVP_MD_CTRL_MICALG.
8084 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8085 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8086 ctrl. It can then customise the structure before and/or after signing
8091 * New function OBJ_add_sigid() to allow application defined signature OIDs
8092 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8093 to free up any added signature OIDs.
8097 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8098 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8099 digest and cipher tables. New options added to openssl utility:
8100 list-message-digest-algorithms and list-cipher-algorithms.
8104 * Change the array representation of binary polynomials: the list
8105 of degrees of non-zero coefficients is now terminated with -1.
8106 Previously it was terminated with 0, which was also part of the
8107 value; thus, the array representation was not applicable to
8108 polynomials where t^0 has coefficient zero. This change makes
8109 the array representation useful in a more general context.
8113 * Various modifications and fixes to SSL/TLS cipher string
8114 handling. For ECC, the code now distinguishes between fixed ECDH
8115 with RSA certificates on the one hand and with ECDSA certificates
8116 on the other hand, since these are separate ciphersuites. The
8117 unused code for Fortezza ciphersuites has been removed.
8119 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8120 (not "ECDHE"). For consistency with the code for DH
8121 certificates, use of ECDH certificates is now considered ECDH
8122 authentication, not RSA or ECDSA authentication (the latter is
8123 merely the CA's signing algorithm and not actively used in the
8126 The temporary ciphersuite alias "ECCdraft" is no longer
8127 available, and ECC ciphersuites are no longer excluded from "ALL"
8128 and "DEFAULT". The following aliases now exist for RFC 4492
8129 ciphersuites, most of these by analogy with the DH case:
8131 kECDHr - ECDH cert, signed with RSA
8132 kECDHe - ECDH cert, signed with ECDSA
8133 kECDH - ECDH cert (signed with either RSA or ECDSA)
8134 kEECDH - ephemeral ECDH
8135 ECDH - ECDH cert or ephemeral ECDH
8141 AECDH - anonymous ECDH
8142 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8147 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8148 Use correct micalg parameters depending on digest(s) in signed message.
8152 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8153 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8157 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8158 an engine to register a method. Add ENGINE lookups for methods and
8159 functional reference processing.
8163 * New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
8164 EVP_{Sign,Verify}* which allow an application to customise the signature
8169 * New -resign option to smime utility. This adds one or more signers
8170 to an existing PKCS#7 signedData structure. Also -md option to use an
8171 alternative message digest algorithm for signing.
8175 * Tidy up PKCS#7 routines and add new functions to make it easier to
8176 create PKCS7 structures containing multiple signers. Update smime
8177 application to support multiple signers.
8181 * New -macalg option to pkcs12 utility to allow setting of an alternative
8186 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8187 Reorganize PBE internals to lookup from a static table using NIDs,
8188 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8189 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8190 PRF which will be automatically used with PBES2.
8194 * Replace the algorithm specific calls to generate keys in "req" with the
8199 * Update PKCS#7 enveloped data routines to use new API. This is now
8200 supported by any public key method supporting the encrypt operation. A
8201 ctrl is added to allow the public key algorithm to examine or modify
8202 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8207 * Add a ctrl to asn1 method to allow a public key algorithm to express
8208 a default digest type to use. In most cases this will be SHA1 but some
8209 algorithms (such as GOST) need to specify an alternative digest. The
8210 return value indicates how strong the preference is 1 means optional and
8211 2 is mandatory (that is it is the only supported type). Modify
8212 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8213 use the default md. Update openssl utilities to use the default digest
8214 type for signing if it is not explicitly indicated.
8218 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8219 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8220 signing method from the key type. This effectively removes the link
8221 between digests and public key types.
8225 * Add an OID cross reference table and utility functions. Its purpose is to
8226 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8227 rsaEncryption. This will allow some of the algorithm specific hackery
8228 needed to use the correct OID to be removed.
8232 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8233 structures for PKCS7_sign(). They are now set up by the relevant public
8238 * Add provisional EC pkey method with support for ECDSA and ECDH.
8242 * Add support for key derivation (agreement) in the API, DH method and
8247 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8248 public and private key formats. As a side effect these add additional
8249 command line functionality not previously available: DSA signatures can be
8250 generated and verified using pkeyutl and DH key support and generation in
8257 *Oliver Tappe <zooey@hirschkaefer.de>*
8259 * New make target "install_html_docs" installs HTML renditions of the
8262 *Oliver Tappe <zooey@hirschkaefer.de>*
8264 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8265 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8266 support key and parameter generation and add initial key generation
8267 functionality for RSA.
8271 * Add functions for main EVP_PKEY_method operations. The undocumented
8272 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
8273 EVP_PKEY_{encrypt,decrypt}_old.
8277 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8278 key API, doesn't do much yet.
8282 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8283 public key algorithms. New option to openssl utility:
8284 "list-public-key-algorithms" to print out info.
8288 * Implement the Supported Elliptic Curves Extension for
8289 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8293 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8294 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8298 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8299 utilities such as rsa, dsa, dsaparam etc except they process any key
8304 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8305 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8306 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8311 * Initial support for pluggable public key ASN1.
8312 De-spaghettify the public key ASN1 handling. Move public and private
8313 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8314 algorithm specific handling to a single module within the relevant
8315 algorithm directory. Add functions to allow (near) opaque processing
8316 of public and private key structures.
8320 * Implement the Supported Point Formats Extension for
8321 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8325 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8326 for the psk identity [hint] and the psk callback functions to the
8327 SSL_SESSION, SSL and SSL_CTX structure.
8330 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8334 SSL_CTX_use_psk_identity_hint
8335 SSL_get_psk_identity_hint
8336 SSL_get_psk_identity
8337 SSL_use_psk_identity_hint
8340 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8342 * Add RFC 3161 compliant time stamp request creation, response generation
8343 and response verification functionality.
8345 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8347 * Add initial support for TLS extensions, specifically for the server_name
8348 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8349 have new members for a host name. The SSL data structure has an
8350 additional member SSL_CTX *initial_ctx so that new sessions can be
8351 stored in that context to allow for session resumption, even after the
8352 SSL has been switched to a new SSL_CTX in reaction to a client's
8353 server_name extension.
8355 New functions (subject to change):
8357 SSL_get_servername()
8358 SSL_get_servername_type()
8361 New CTRL codes and macros (subject to change):
8363 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8364 - SSL_CTX_set_tlsext_servername_callback()
8365 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8366 - SSL_CTX_set_tlsext_servername_arg()
8367 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8369 openssl s_client has a new '-servername ...' option.
8371 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8372 '-key2 ...', '-servername_fatal' (subject to change). This allows
8373 testing the HostName extension for a specific single host name ('-cert'
8374 and '-key' remain fallbacks for handshakes without HostName
8375 negotiation). If the unrecognized_name alert has to be sent, this by
8376 default is a warning; it becomes fatal with the '-servername_fatal'
8380 *Peter Sylvester, Remy Allais, Christophe Renou*
8382 * Whirlpool hash implementation is added.
8386 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8387 bn(64,32). Because of instruction set limitations it doesn't have
8388 any negative impact on performance. This was done mostly in order
8389 to make it possible to share assembler modules, such as bn_mul_mont
8390 implementations, between 32- and 64-bit builds without hassle.
8394 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8395 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8400 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8401 dedicated Montgomery multiplication procedure, is introduced.
8402 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8403 "64-bit" performance on certain 32-bit targets.
8407 * New option SSL_OP_NO_COMP to disable use of compression selectively
8408 in SSL structures. New SSL ctrl to set maximum send fragment size.
8409 Save memory by setting the I/O buffer sizes dynamically instead of
8410 using the maximum available value.
8414 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8415 in addition to the text details.
8419 * Very, very preliminary EXPERIMENTAL support for printing of general
8420 ASN1 structures. This currently produces rather ugly output and doesn't
8421 handle several customised structures at all.
8425 * Integrated support for PVK file format and some related formats such
8426 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8427 these in the 'rsa' and 'dsa' utilities.
8431 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8435 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8436 place for the (very old) "NETSCAPE" format certificates which are now
8437 handled using new ASN1 code equivalents.
8441 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8442 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8443 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8447 * Modify CRL distribution points extension code to print out previously
8448 unsupported fields. Enhance extension setting code to allow setting of
8453 * Add print and set support for Issuing Distribution Point CRL extension.
8457 * Change 'Configure' script to enable Camellia by default.
8464 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010] ###
8466 * When rejecting SSL/TLS records due to an incorrect version number, never
8467 update s->server with a new major version number. As of
8468 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8469 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8470 the previous behavior could result in a read attempt at NULL when
8471 receiving specific incorrect SSL/TLS records once record payload
8472 protection is active. [CVE-2010-0740][]
8474 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8476 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8477 could be crashed if the relevant tables were not present (e.g. chrooted).
8479 *Tomas Hoger <thoger@redhat.com>*
8481 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010] ###
8483 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8485 *Martin Olsson, Neel Mehta*
8487 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8488 accommodate for stack sorting, always a write lock!).
8492 * On some versions of WIN32 Heap32Next is very slow. This can cause
8493 excessive delays in the RAND_poll(): over a minute. As a workaround
8494 include a time check in the inner Heap32Next loop too.
8498 * The code that handled flushing of data in SSL/TLS originally used the
8499 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8500 the problem outlined in PR#1949. The fix suggested there however can
8501 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8502 of Apache). So instead simplify the code to flush unconditionally.
8503 This should be fine since flushing with no data to flush is a no op.
8507 * Handle TLS versions 2.0 and later properly and correctly use the
8508 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8509 off ancient servers have a habit of sticking around for a while...
8513 * Modify compression code so it frees up structures without using the
8514 ex_data callbacks. This works around a problem where some applications
8515 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8516 restarting) then use compression (e.g. SSL with compression) later.
8517 This results in significant per-connection memory leaks and
8518 has caused some security issues including CVE-2008-1678 and
8523 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8524 change when encrypting or decrypting.
8528 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8529 connect and renegotiate with servers which do not support RI.
8530 Until RI is more widely deployed this option is enabled by default.
8534 * Add "missing" ssl ctrls to clear options and mode.
8538 * If client attempts to renegotiate and doesn't support RI respond with
8539 a no_renegotiation alert as required by RFC5746. Some renegotiating
8540 TLS clients will continue a connection gracefully when they receive
8541 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8542 waiting for a server hello which it will never receive. Now we treat a
8543 received no_renegotiation alert as a fatal error. This is because
8544 applications requesting a renegotiation might well expect it to succeed
8545 and would have no code in place to handle the server denying it so the
8546 only safe thing to do is to terminate the connection.
8550 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8551 peer supports secure renegotiation and 0 otherwise. Print out peer
8552 renegotiation support in s_client/s_server.
8556 * Replace the highly broken and deprecated SPKAC certification method with
8557 the updated NID creation version. This should correctly handle UTF8.
8561 * Implement RFC5746. Re-enable renegotiation but require the extension
8562 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8563 turns out to be a bad idea. It has been replaced by
8564 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8565 SSL_CTX_set_options(). This is really not recommended unless you
8566 know what you are doing.
8568 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8570 * Fixes to stateless session resumption handling. Use initial_ctx when
8571 issuing and attempting to decrypt tickets in case it has changed during
8572 servername handling. Use a non-zero length session ID when attempting
8573 stateless session resumption: this makes it possible to determine if
8574 a resumption has occurred immediately after receiving server hello
8575 (several places in OpenSSL subtly assume this) instead of later in
8580 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8581 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8582 fixes for a few places where the return code is not checked
8585 *Julia Lawall <julia@diku.dk>*
8587 * Add --strict-warnings option to Configure script to include devteam
8588 warnings in other configurations.
8592 * Add support for --libdir option and LIBDIR variable in makefiles. This
8593 makes it possible to install openssl libraries in locations which
8594 have names other than "lib", for example "/usr/lib64" which some
8597 *Steve Henson, based on patch from Jeremy Utley*
8599 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8600 X690 8.9.12 and can produce some misleading textual output of OIDs.
8602 *Steve Henson, reported by Dan Kaminsky*
8604 * Delete MD2 from algorithm tables. This follows the recommendation in
8605 several standards that it is not used in new applications due to
8606 several cryptographic weaknesses. For binary compatibility reasons
8607 the MD2 API is still compiled in by default.
8611 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8616 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8617 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8620 *Guenter <lists@gknw.net>*
8622 * Fix the server certificate chain building code to use X509_verify_cert(),
8623 it used to have an ad-hoc builder which was unable to cope with anything
8624 other than a simple chain.
8626 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8628 * Don't check self signed certificate signatures in X509_verify_cert()
8629 by default (a flag can override this): it just wastes time without
8630 adding any security. As a useful side effect self signed root CAs
8631 with non-FIPS digests are now usable in FIPS mode.
8635 * In dtls1_process_out_of_seq_message() the check if the current message
8636 is already buffered was missing. For every new message was memory
8637 allocated, allowing an attacker to perform an denial of service attack
8638 with sending out of seq handshake messages until there is no memory
8639 left. Additionally every future message was buffered, even if the
8640 sequence number made no sense and would be part of another handshake.
8641 So only messages with sequence numbers less than 10 in advance will be
8642 buffered. [CVE-2009-1378][]
8644 *Robin Seggelmann, discovered by Daniel Mentz*
8646 * Records are buffered if they arrive with a future epoch to be
8647 processed after finishing the corresponding handshake. There is
8648 currently no limitation to this buffer allowing an attacker to perform
8649 a DOS attack with sending records with future epochs until there is no
8650 memory left. This patch adds the pqueue_size() function to determine
8651 the size of a buffer and limits the record buffer to 100 entries.
8654 *Robin Seggelmann, discovered by Daniel Mentz*
8656 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8657 parent structure is freed. [CVE-2009-1379][]
8661 * Handle non-blocking I/O properly in SSL_shutdown() call.
8663 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8667 *Ilya O. <vrghost@gmail.com>*
8669 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009] ###
8671 * Disable renegotiation completely - this fixes a severe security
8672 problem [CVE-2009-3555][] at the cost of breaking all
8673 renegotiation. Renegotiation can be re-enabled by setting
8674 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8675 run-time. This is really not recommended unless you know what
8680 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009] ###
8682 * Don't set val to NULL when freeing up structures, it is freed up by
8683 underlying code. If sizeof(void *) > sizeof(long) this can result in
8684 zeroing past the valid field. [CVE-2009-0789][]
8686 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8688 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8689 checked correctly. This would allow some invalid signed attributes to
8690 appear to verify correctly. [CVE-2009-0591][]
8692 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8694 * Reject UniversalString and BMPString types with invalid lengths. This
8695 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8696 a legal length. [CVE-2009-0590][]
8700 * Set S/MIME signing as the default purpose rather than setting it
8701 unconditionally. This allows applications to override it at the store
8706 * Permit restricted recursion of ASN1 strings. This is needed in practice
8707 to handle some structures.
8711 * Improve efficiency of mem_gets: don't search whole buffer each time
8714 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8716 * New -hex option for openssl rand.
8720 * Print out UTF8String and NumericString when parsing ASN1.
8724 * Support NumericString type for name components.
8728 * Allow CC in the environment to override the automatically chosen
8729 compiler. Note that nothing is done to ensure flags work with the
8734 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009] ###
8736 * Properly check EVP_VerifyFinal() and similar return values
8739 *Ben Laurie, Bodo Moeller, Google Security Team*
8741 * Enable TLS extensions by default.
8745 * Allow the CHIL engine to be loaded, whether the application is
8746 multithreaded or not. (This does not release the developer from the
8747 obligation to set up the dynamic locking callbacks.)
8749 *Sander Temme <sander@temme.net>*
8751 * Use correct exit code if there is an error in dgst command.
8753 *Steve Henson; problem pointed out by Roland Dirlewanger*
8755 * Tweak Configure so that you need to say "experimental-jpake" to enable
8756 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8760 * Add experimental JPAKE support, including demo authentication in
8761 s_client and s_server.
8765 * Set the comparison function in v3_addr_canonize().
8767 *Rob Austein <sra@hactrn.net>*
8769 * Add support for XMPP STARTTLS in s_client.
8771 *Philip Paeps <philip@freebsd.org>*
8773 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8774 to ensure that even with this option, only ciphersuites in the
8775 server's preference list will be accepted. (Note that the option
8776 applies only when resuming a session, so the earlier behavior was
8777 just about the algorithm choice for symmetric cryptography.)
8781 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008] ###
8783 * Fix NULL pointer dereference if a DTLS server received
8784 ChangeCipherSpec as first record [CVE-2009-1386][].
8788 * Fix a state transition in s3_srvr.c and d1_srvr.c
8789 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
8793 * The fix in 0.9.8c that supposedly got rid of unsafe
8794 double-checked locking was incomplete for RSA blinding,
8795 addressing just one layer of what turns out to have been
8796 doubly unsafe triple-checked locking.
8798 So now fix this for real by retiring the MONT_HELPER macro
8799 in crypto/rsa/rsa_eay.c.
8802 *Bodo Moeller; problem pointed out by Marius Schilder*
8804 * Various precautionary measures:
8806 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8808 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8809 (NB: This would require knowledge of the secret session ticket key
8810 to exploit, in which case you'd be SOL either way.)
8812 - Change bn_nist.c so that it will properly handle input BIGNUMs
8813 outside the expected range.
8815 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8819 *Neel Mehta, Bodo Moeller*
8821 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8822 the load fails. Useful for distros.
8824 *Ben Laurie and the FreeBSD team*
8826 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8830 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8834 * Expand ENGINE to support engine supplied SSL client certificate functions.
8836 This work was sponsored by Logica.
8840 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8841 keystores. Support for SSL/TLS client authentication too.
8842 Not compiled unless enable-capieng specified to Configure.
8844 This work was sponsored by Logica.
8848 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8849 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8850 attribute creation routines such as certificate requests and PKCS#12
8855 ### Changes between 0.9.8g and 0.9.8h [28 May 2008] ###
8857 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8858 handshake which could lead to a client crash as found using the
8859 Codenomicon TLS test suite [CVE-2008-1672][]
8861 *Steve Henson, Mark Cox*
8863 * Fix double free in TLS server name extensions which could lead to
8864 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8868 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8870 Clear the error queue to ensure that error entries left from
8871 older function calls do not interfere with the correct operation.
8873 *Lutz Jaenicke, Erik de Castro Lopo*
8875 * Remove root CA certificates of commercial CAs:
8877 The OpenSSL project does not recommend any specific CA and does not
8878 have any policy with respect to including or excluding any CA.
8879 Therefore it does not make any sense to ship an arbitrary selection
8880 of root CA certificates with the OpenSSL software.
8884 * RSA OAEP patches to fix two separate invalid memory reads.
8885 The first one involves inputs when 'lzero' is greater than
8886 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
8887 before the beginning of from). The second one involves inputs where
8888 the 'db' section contains nothing but zeroes (there is a one-byte
8889 invalid read after the end of 'db').
8891 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8893 * Partial backport from 0.9.9-dev:
8895 Introduce bn_mul_mont (dedicated Montgomery multiplication
8896 procedure) as a candidate for BIGNUM assembler implementation.
8897 While 0.9.9-dev uses assembler for various architectures, only
8898 x86_64 is available by default here in the 0.9.8 branch, and
8899 32-bit x86 is available through a compile-time setting.
8901 To try the 32-bit x86 assembler implementation, use Configure
8902 option "enable-montasm" (which exists only for this backport).
8904 As "enable-montasm" for 32-bit x86 disclaims code stability
8905 anyway, in this constellation we activate additional code
8906 backported from 0.9.9-dev for further performance improvements,
8907 namely BN_from_montgomery_word. (To enable this otherwise,
8908 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
8911 *Andy Polyakov (backport partially by Bodo Moeller)*
8913 * Add TLS session ticket callback. This allows an application to set
8914 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
8915 values. This is useful for key rollover for example where several key
8916 sets may exist with different names.
8920 * Reverse ENGINE-internal logic for caching default ENGINE handles.
8921 This was broken until now in 0.9.8 releases, such that the only way
8922 a registered ENGINE could be used (assuming it initialises
8923 successfully on the host) was to explicitly set it as the default
8924 for the relevant algorithms. This is in contradiction with 0.9.7
8925 behaviour and the documentation. With this fix, when an ENGINE is
8926 registered into a given algorithm's table of implementations, the
8927 'uptodate' flag is reset so that auto-discovery will be used next
8928 time a new context for that algorithm attempts to select an
8931 *Ian Lister (tweaked by Geoff Thorpe)*
8933 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
8934 implementation in the following ways:
8936 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
8939 Lack of BER streaming support means one pass streaming processing is
8940 only supported if data is detached: setting the streaming flag is
8941 ignored for embedded content.
8943 CMS support is disabled by default and must be explicitly enabled
8944 with the enable-cms configuration option.
8948 * Update the GMP engine glue to do direct copies between BIGNUM and
8949 mpz_t when openssl and GMP use the same limb size. Otherwise the
8950 existing "conversion via a text string export" trick is still used.
8952 *Paul Sheer <paulsheer@gmail.com>*
8954 * Zlib compression BIO. This is a filter BIO which compressed and
8955 uncompresses any data passed through it.
8959 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
8960 RFC3394 compatible AES key wrapping.
8964 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
8965 sets string data without copying. X509_ALGOR_set0() and
8966 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
8967 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
8968 from an X509_ATTRIBUTE structure optionally checking it occurs only
8969 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
8974 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
8975 to get the expected BN_FLG_CONSTTIME behavior.
8977 *Bodo Moeller (Google)*
8981 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
8982 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
8983 - added some more tests to do_tests.pl
8984 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
8985 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
8986 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
8987 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
8988 - various changes to netware.pl to enable gcc-cross builds on Win32
8990 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
8991 - various changes to fix missing prototype warnings
8992 - fixed x86nasm.pl to create correct asm files for NASM COFF output
8993 - added AES, WHIRLPOOL and CPUID assembler code to build files
8994 - added missing AES assembler make rules to mk1mf.pl
8995 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
8997 *Guenter Knauf <eflash@gmx.net>*
8999 * Implement certificate status request TLS extension defined in RFC3546.
9000 A client can set the appropriate parameters and receive the encoded
9001 OCSP response via a callback. A server can query the supplied parameters
9002 and set the encoded OCSP response in the callback. Add simplified examples
9003 to s_client and s_server.
9007 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007] ###
9010 + Binary incompatibility of ssl_ctx_st structure
9011 + DTLS interoperation with non-compliant servers
9012 + Don't call get_session_cb() without proposed session
9013 + Fix ia64 assembler code
9015 *Andy Polyakov, Steve Henson*
9017 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007] ###
9019 * DTLS Handshake overhaul. There were longstanding issues with
9020 OpenSSL DTLS implementation, which were making it impossible for
9021 RFC 4347 compliant client to communicate with OpenSSL server.
9022 Unfortunately just fixing these incompatibilities would "cut off"
9023 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9024 server keeps tolerating non RFC compliant syntax. The opposite is
9025 not true, 0.9.8f client can not communicate with earlier server.
9026 This update even addresses CVE-2007-4995.
9030 * Changes to avoid need for function casts in OpenSSL: some compilers
9031 (gcc 4.2 and later) reject their use.
9032 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9035 * Add RFC4507 support to OpenSSL. This includes the corrections in
9036 RFC4507bis. The encrypted ticket format is an encrypted encoded
9037 SSL_SESSION structure, that way new session features are automatically
9040 If a client application caches session in an SSL_SESSION structure
9041 support is transparent because tickets are now stored in the encoded
9044 The SSL_CTX structure automatically generates keys for ticket
9045 protection in servers so again support should be possible
9046 with no application modification.
9048 If a client or server wishes to disable RFC4507 support then the option
9049 SSL_OP_NO_TICKET can be set.
9051 Add a TLS extension debugging callback to allow the contents of any client
9052 or server extensions to be examined.
9054 This work was sponsored by Google.
9058 * Add initial support for TLS extensions, specifically for the server_name
9059 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9060 have new members for a host name. The SSL data structure has an
9061 additional member SSL_CTX *initial_ctx so that new sessions can be
9062 stored in that context to allow for session resumption, even after the
9063 SSL has been switched to a new SSL_CTX in reaction to a client's
9064 server_name extension.
9066 New functions (subject to change):
9068 SSL_get_servername()
9069 SSL_get_servername_type()
9072 New CTRL codes and macros (subject to change):
9074 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9075 - SSL_CTX_set_tlsext_servername_callback()
9076 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9077 - SSL_CTX_set_tlsext_servername_arg()
9078 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9080 openssl s_client has a new '-servername ...' option.
9082 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9083 '-key2 ...', '-servername_fatal' (subject to change). This allows
9084 testing the HostName extension for a specific single host name ('-cert'
9085 and '-key' remain fallbacks for handshakes without HostName
9086 negotiation). If the unrecognized_name alert has to be sent, this by
9087 default is a warning; it becomes fatal with the '-servername_fatal'
9091 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9093 * Add AES and SSE2 assembly language support to VC++ build.
9097 * Mitigate attack on final subtraction in Montgomery reduction.
9101 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9102 (which previously caused an internal error).
9106 * Squeeze another 10% out of IGE mode when in != out.
9110 * AES IGE mode speedup.
9112 *Dean Gaudet (Google)*
9114 * Add the Korean symmetric 128-bit cipher SEED (see
9115 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
9116 add SEED ciphersuites from RFC 4162:
9118 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9119 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9120 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9121 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9123 To minimize changes between patchlevels in the OpenSSL 0.9.8
9124 series, SEED remains excluded from compilation unless OpenSSL
9125 is configured with 'enable-seed'.
9127 *KISA, Bodo Moeller*
9129 * Mitigate branch prediction attacks, which can be practical if a
9130 single processor is shared, allowing a spy process to extract
9131 information. For detailed background information, see
9132 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
9133 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9134 and Necessary Software Countermeasures"). The core of the change
9135 are new versions BN_div_no_branch() and
9136 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9137 respectively, which are slower, but avoid the security-relevant
9138 conditional branches. These are automatically called by BN_div()
9139 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9140 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9141 remove a conditional branch.
9143 BN_FLG_CONSTTIME is the new name for the previous
9144 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9145 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9146 in the exponent causes BN_mod_exp_mont() to use the alternative
9147 implementation in BN_mod_exp_mont_consttime().) The old name
9148 remains as a deprecated alias.
9150 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9151 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9152 constant-time implementations for more than just exponentiation.
9153 Here too the old name is kept as a deprecated alias.
9155 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9156 the BN_BLINDING structure gets an independent copy of the
9157 modulus. This means that the previous "BIGNUM *m" argument to
9158 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9159 essentially becomes "const BIGNUM *m", although we can't actually
9160 change this in the header file before 0.9.9. It allows
9161 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9162 enable BN_FLG_CONSTTIME.
9165 *Matthew D Wood (Intel Corp)*
9167 * In the SSL/TLS server implementation, be strict about session ID
9168 context matching (which matters if an application uses a single
9169 external cache for different purposes). Previously,
9170 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9171 set. This did ensure strict client verification, but meant that,
9172 with applications using a single external cache for quite
9173 different requirements, clients could circumvent ciphersuite
9174 restrictions for a given session ID context by starting a session
9175 in a different context.
9179 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9180 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9181 authentication-only ciphersuites.
9185 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9186 not complete and could lead to a possible single byte overflow
9187 [CVE-2007-5135][] [Ben Laurie]
9189 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007] ###
9191 * Since AES128 and AES256 (and similarly Camellia128 and
9192 Camellia256) share a single mask bit in the logic of
9193 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9194 kludge to work properly if AES128 is available and AES256 isn't
9195 (or if Camellia128 is available and Camellia256 isn't).
9199 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9200 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9201 When a point or a seed is encoded in a BIT STRING, we need to
9202 prevent the removal of trailing zero bits to get the proper DER
9203 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9204 of a NamedBitList, for which trailing 0 bits need to be removed.)
9208 * Have SSL/TLS server implementation tolerate "mismatched" record
9209 protocol version while receiving ClientHello even if the
9210 ClientHello is fragmented. (The server can't insist on the
9211 particular protocol version it has chosen before the ServerHello
9212 message has informed the client about his choice.)
9216 * Add RFC 3779 support.
9218 *Rob Austein for ARIN, Ben Laurie*
9220 * Load error codes if they are not already present instead of using a
9221 static variable. This allows them to be cleanly unloaded and reloaded.
9222 Improve header file function name parsing.
9226 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9227 or CAPABILITY handshake as required by RFCs.
9231 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006] ###
9233 * Introduce limits to prevent malicious keys being able to
9234 cause a denial of service. [CVE-2006-2940][]
9236 *Steve Henson, Bodo Moeller*
9238 * Fix ASN.1 parsing of certain invalid structures that can result
9239 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9241 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9242 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9244 * Fix SSL client code which could crash if connecting to a
9245 malicious SSLv2 server. [CVE-2006-4343][]
9247 *Tavis Ormandy and Will Drewry, Google Security Team*
9249 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9250 match only those. Before that, "AES256-SHA" would be interpreted
9251 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9252 the same strength classification in 0.9.7h) as we currently only
9253 have a single AES bit in the ciphersuite description bitmap.
9254 That change, however, also applied to ciphersuite strings such as
9255 "RC4-MD5" that intentionally matched multiple ciphersuites --
9256 namely, SSL 2.0 ciphersuites in addition to the more common ones
9257 from SSL 3.0/TLS 1.0.
9259 So we change the selection algorithm again: Naming an explicit
9260 ciphersuite selects this one ciphersuite, and any other similar
9261 ciphersuite (same bitmap) from *other* protocol versions.
9262 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9263 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9265 Since SSL 2.0 does not have any ciphersuites for which the
9266 128/256 bit distinction would be relevant, this works for now.
9267 The proper fix will be to use different bits for AES128 and
9268 AES256, which would have avoided the problems from the beginning;
9269 however, bits are scarce, so we can only do this in a new release
9270 (not just a patchlevel) when we can change the SSL_CIPHER
9271 definition to split the single 'unsigned long mask' bitmap into
9272 multiple values to extend the available space.
9277 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006] ###
9279 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9280 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9282 * Add AES IGE and biIGE modes.
9286 * Change the Unix randomness entropy gathering to use poll() when
9287 possible instead of select(), since the latter has some
9288 undesirable limitations.
9290 *Darryl Miles via Richard Levitte and Bodo Moeller*
9292 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9293 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9294 cannot be implicitly activated as part of, e.g., the "AES" alias.
9295 However, please upgrade to OpenSSL 0.9.9[-dev] for
9296 non-experimental use of the ECC ciphersuites to get TLS extension
9297 support, which is required for curve and point format negotiation
9298 to avoid potential handshake problems.
9302 * Disable rogue ciphersuites:
9304 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9305 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9306 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9308 The latter two were purportedly from
9309 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9312 Also deactivate the remaining ciphersuites from
9313 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9314 unofficial, and the ID has long expired.
9318 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9319 dual-core machines) and other potential thread-safety issues.
9323 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9324 versions), which is now available for royalty-free use
9325 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
9326 Also, add Camellia TLS ciphersuites from RFC 4132.
9328 To minimize changes between patchlevels in the OpenSSL 0.9.8
9329 series, Camellia remains excluded from compilation unless OpenSSL
9330 is configured with 'enable-camellia'.
9334 * Disable the padding bug check when compression is in use. The padding
9335 bug check assumes the first packet is of even length, this is not
9336 necessarily true if compression is enabled and can result in false
9337 positives causing handshake failure. The actual bug test is ancient
9338 code so it is hoped that implementations will either have fixed it by
9339 now or any which still have the bug do not support compression.
9343 ### Changes between 0.9.8a and 0.9.8b [04 May 2006] ###
9345 * When applying a cipher rule check to see if string match is an explicit
9346 cipher suite and only match that one cipher suite if it is.
9350 * Link in manifests for VC++ if needed.
9352 *Austin Ziegler <halostatue@gmail.com>*
9354 * Update support for ECC-based TLS ciphersuites according to
9355 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9356 TLS extensions, which are supported starting with the 0.9.9
9357 branch, not in the OpenSSL 0.9.8 branch).
9361 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9362 opaque EVP_CIPHER_CTX handling.
9366 * Fixes and enhancements to zlib compression code. We now only use
9367 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9368 to conform with the standards mentioned here:
9369 http://www.zlib.net/DLL_FAQ.txt
9370 Static zlib linking now works on Windows and the new --with-zlib-include
9371 --with-zlib-lib options to Configure can be used to supply the location
9372 of the headers and library. Gracefully handle case where zlib library
9377 * Several fixes and enhancements to the OID generation code. The old code
9378 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9379 handle numbers larger than ULONG_MAX, truncated printing and had a
9380 non standard OBJ_obj2txt() behaviour.
9384 * Add support for building of engines under engine/ as shared libraries
9385 under VC++ build system.
9389 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9390 Hopefully, we will not see any false combination of paths any more.
9394 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005] ###
9396 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9397 (part of SSL_OP_ALL). This option used to disable the
9398 countermeasure against man-in-the-middle protocol-version
9399 rollback in the SSL 2.0 server implementation, which is a bad
9400 idea. [CVE-2005-2969][]
9402 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9403 for Information Security, National Institute of Advanced Industrial
9404 Science and Technology [AIST*, Japan)]
9406 * Add two function to clear and return the verify parameter flags.
9410 * Keep cipherlists sorted in the source instead of sorting them at
9411 runtime, thus removing the need for a lock.
9415 * Avoid some small subgroup attacks in Diffie-Hellman.
9417 *Nick Mathewson and Ben Laurie*
9419 * Add functions for well-known primes.
9423 * Extended Windows CE support.
9425 *Satoshi Nakamura and Andy Polyakov*
9427 * Initialize SSL_METHOD structures at compile time instead of during
9428 runtime, thus removing the need for a lock.
9432 * Make PKCS7_decrypt() work even if no certificate is supplied by
9433 attempting to decrypt each encrypted key in turn. Add support to
9438 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005] ###
9440 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9443 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9447 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9448 key into the same file any more.
9452 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9456 * Add -utf8 command line and config file option to 'ca'.
9458 *Stefan <stf@udoma.org*
9460 * Removed the macro des_crypt(), as it seems to conflict with some
9461 libraries. Use DES_crypt().
9465 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9466 involves renaming the source and generated shared-libs for
9467 both. The engines will accept the corrected or legacy ids
9468 ('ncipher' and '4758_cca' respectively) when binding. NB,
9469 this only applies when building 'shared'.
9471 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9473 * Add attribute functions to EVP_PKEY structure. Modify
9474 PKCS12_create() to recognize a CSP name attribute and
9475 use it. Make -CSP option work again in pkcs12 utility.
9479 * Add new functionality to the bn blinding code:
9480 - automatic re-creation of the BN_BLINDING parameters after
9481 a fixed number of uses (currently 32)
9482 - add new function for parameter creation
9483 - introduce flags to control the update behaviour of the
9484 BN_BLINDING parameters
9485 - hide BN_BLINDING structure
9486 Add a second BN_BLINDING slot to the RSA structure to improve
9487 performance when a single RSA object is shared among several
9492 * Add support for DTLS.
9494 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9496 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9497 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9501 * Remove buggy and incomplete DH cert support from
9502 ssl/ssl_rsa.c and ssl/s3_both.c
9506 * Use SHA-1 instead of MD5 as the default digest algorithm for
9507 the apps/openssl applications.
9511 * Compile clean with "-Wall -Wmissing-prototypes
9512 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9513 DEBUG_SAFESTACK must also be set.
9517 * Change ./Configure so that certain algorithms can be disabled by default.
9518 The new counterpiece to "no-xxx" is "enable-xxx".
9520 The patented RC5 and MDC2 algorithms will now be disabled unless
9521 "enable-rc5" and "enable-mdc2", respectively, are specified.
9523 (IDEA remains enabled despite being patented. This is because IDEA
9524 is frequently required for interoperability, and there is no license
9525 fee for non-commercial use. As before, "no-idea" can be used to
9526 avoid this algorithm.)
9531 * Add processing of proxy certificates (see RFC 3820). This work was
9532 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9533 EGEE (Enabling Grids for E-science in Europe).
9537 * RC4 performance overhaul on modern architectures/implementations, such
9538 as Intel P4, IA-64 and AMD64.
9542 * New utility extract-section.pl. This can be used specify an alternative
9543 section number in a pod file instead of having to treat each file as
9544 a separate case in Makefile. This can be done by adding two lines to the
9547 =for comment openssl_section:XXX
9549 The blank line is mandatory.
9554 * New arguments -certform, -keyform and -pass for s_client and s_server
9555 to allow alternative format key and certificate files and passphrase
9560 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9561 update associated structures and add various utility functions.
9563 Add new policy related verify parameters, include policy checking in
9564 standard verify code. Enhance 'smime' application with extra parameters
9565 to support policy checking and print out.
9569 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9570 Nehemiah processors. These extensions support AES encryption in hardware
9571 as well as RNG (though RNG support is currently disabled).
9573 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9575 * Deprecate BN_[get|set]_params() functions (they were ignored internally).
9579 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9581 *Andy Polyakov and a number of other people*
9583 * Improved PowerPC platform support. Most notably BIGNUM assembler
9584 implementation contributed by IBM.
9586 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9588 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9589 exponent rather than 'unsigned long'. There is a corresponding change to
9590 the new 'rsa_keygen' element of the RSA_METHOD structure.
9592 *Jelte Jansen, Geoff Thorpe*
9594 * Functionality for creating the initial serial number file is now
9595 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9597 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9598 number file to 1, which is bound to cause problems. To avoid
9599 the problems while respecting compatibility between different 0.9.7
9600 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9601 CA.pl for serial number initialization. With the new release 0.9.8,
9602 we can fix the problem directly in the 'ca' utility.)
9606 * Reduced header interdependencies by declaring more opaque objects in
9607 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9608 give fewer recursive includes, which could break lazy source code - so
9609 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9610 developers should define this symbol when building and using openssl to
9611 ensure they track the recommended behaviour, interfaces, [etc], but
9612 backwards-compatible behaviour prevails when this isn't defined.
9616 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9620 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9621 This will generate a random key of the appropriate length based on the
9622 cipher context. The EVP_CIPHER can provide its own random key generation
9623 routine to support keys of a specific form. This is used in the des and
9624 3des routines to generate a key of the correct parity. Update S/MIME
9625 code to use new functions and hence generate correct parity DES keys.
9626 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9627 valid (weak or incorrect parity).
9631 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9632 as looking them up. This is useful when the verified structure may contain
9633 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9634 present unless the new PKCS7_NO_CRL flag is asserted.
9638 * Extend ASN1 oid configuration module. It now additionally accepts the
9641 shortName = some long name, 1.2.3.4
9645 * Reimplemented the BN_CTX implementation. There is now no more static
9646 limitation on the number of variables it can handle nor the depth of the
9647 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9648 information can now expand as required, and rather than having a single
9649 static array of bignums, BN_CTX now uses a linked-list of such arrays
9650 allowing it to expand on demand whilst maintaining the usefulness of
9651 BN_CTX's "bundling".
9655 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9656 to allow all RSA operations to function using a single BN_CTX.
9660 * Preliminary support for certificate policy evaluation and checking. This
9661 is initially intended to pass the tests outlined in "Conformance Testing
9662 of Relying Party Client Certificate Path Processing Logic" v1.07.
9666 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9667 remained unused and not that useful. A variety of other little bignum
9668 tweaks and fixes have also been made continuing on from the audit (see
9673 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9674 associated ASN1, EVP and SSL functions and old ASN1 macros.
9678 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9679 and this should never fail. So the return value from the use of
9680 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9681 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9685 * BN_CTX_get() should return zero-valued bignums, providing the same
9686 initialised value as BN_new().
9688 *Geoff Thorpe, suggested by Ulf Möller*
9690 * Support for inhibitAnyPolicy certificate extension.
9694 * An audit of the BIGNUM code is underway, for which debugging code is
9695 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9696 is considered valid when processing BIGNUMs, and causes execution to
9697 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9698 further steps are taken to deliberately pollute unused data in BIGNUM
9699 structures to try and expose faulty code further on. For now, openssl will
9700 (in its default mode of operation) continue to tolerate the inconsistent
9701 forms that it has tolerated in the past, but authors and packagers should
9702 consider trying openssl and their own applications when compiled with
9703 these debugging symbols defined. It will help highlight potential bugs in
9704 their own code, and will improve the test coverage for OpenSSL itself. At
9705 some point, these tighter rules will become openssl's default to improve
9706 maintainability, though the assert()s and other overheads will remain only
9707 in debugging configurations. See bn.h for more details.
9709 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9711 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9712 that can only be obtained through BN_CTX_new() (which implicitly
9713 initialises it). The presence of this function only made it possible
9714 to overwrite an existing structure (and cause memory leaks).
9718 * Because of the callback-based approach for implementing LHASH as a
9719 template type, lh_insert() adds opaque objects to hash-tables and
9720 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9721 to clean up those corresponding objects before destroying the hash table
9722 (and losing the object pointers). So some over-zealous constifications in
9723 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9724 objects as "const" and the lh_doall[_arg] callback wrappers are not
9725 prototyped to have "const" restrictions on the object pointers they are
9726 given (and so aren't required to cast them away any more).
9730 * The tmdiff.h API was so ugly and minimal that our own timing utility
9731 (speed) prefers to use its own implementation. The two implementations
9732 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9733 its object type properly exposed (MS_TM) instead of casting to/from "char
9734 *". This may still change yet if someone realises MS_TM and "ms_time_***"
9735 aren't necessarily the greatest nomenclatures - but this is what was used
9736 internally to the implementation so I've used that for now.
9740 * Ensure that deprecated functions do not get compiled when
9741 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9742 the self-tests were still using deprecated key-generation functions so
9743 these have been updated also.
9747 * Reorganise PKCS#7 code to separate the digest location functionality
9748 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9749 New function PKCS7_set_digest() to set the digest type for PKCS#7
9750 digestedData type. Add additional code to correctly generate the
9751 digestedData type and add support for this type in PKCS7 initialization
9756 * New function PKCS7_set0_type_other() this initializes a PKCS7
9757 structure of type "other".
9761 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9762 sure the loop does correctly stop and breaking ("division by zero")
9763 modulus operations are not performed. The (pre-generated) prime
9764 table crypto/bn/bn_prime.h was already correct, but it could not be
9765 re-generated on some platforms because of the "division by zero"
9766 situation in the script.
9768 *Ralf S. Engelschall*
9770 * Update support for ECC-based TLS ciphersuites according to
9771 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9772 SHA-1 now is only used for "small" curves (where the
9773 representation of a field element takes up to 24 bytes); for
9774 larger curves, the field element resulting from ECDH is directly
9775 used as premaster secret.
9777 *Douglas Stebila (Sun Microsystems Laboratories)*
9779 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9780 curve secp160r1 to the tests.
9782 *Douglas Stebila (Sun Microsystems Laboratories)*
9784 * Add the possibility to load symbols globally with DSO.
9786 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9788 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9789 control of the error stack.
9793 * Add support for STORE in ENGINE.
9797 * Add the STORE type. The intention is to provide a common interface
9798 to certificate and key stores, be they simple file-based stores, or
9799 HSM-type store, or LDAP stores, or...
9800 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9804 * Add a generic structure called OPENSSL_ITEM. This can be used to
9805 pass a list of arguments to any function as well as provide a way
9806 for a function to pass data back to the caller.
9810 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9811 works like BUF_strdup() but can be used to duplicate a portion of
9812 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9817 * Add the function sk_find_ex() which works like sk_find(), but will
9818 return an index to an element even if an exact match couldn't be
9819 found. The index is guaranteed to point at the element where the
9820 searched-for key would be inserted to preserve sorting order.
9824 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9825 takes an extra flags argument for optional functionality. Currently,
9826 the following flags are defined:
9828 OBJ_BSEARCH_VALUE_ON_NOMATCH
9829 This one gets OBJ_bsearch_ex() to return a pointer to the first
9830 element where the comparing function returns a negative or zero
9833 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9834 This one gets OBJ_bsearch_ex() to return a pointer to the first
9835 element where the comparing function returns zero. This is useful
9836 if there are more than one element where the comparing function
9841 * Make it possible to create self-signed certificates with 'openssl ca'
9842 in such a way that the self-signed certificate becomes part of the
9843 CA database and uses the same mechanisms for serial number generation
9844 as all other certificate signing. The new flag '-selfsign' enables
9845 this functionality. Adapt CA.sh and CA.pl.in.
9849 * Add functionality to check the public key of a certificate request
9850 against a given private. This is useful to check that a certificate
9851 request can be signed by that key (self-signing).
9855 * Make it possible to have multiple active certificates with the same
9856 subject in the CA index file. This is done only if the keyword
9857 'unique_subject' is set to 'no' in the main CA section (default
9858 if 'CA_default') of the configuration file. The value is saved
9859 with the database itself in a separate index attribute file,
9860 named like the index file with '.attr' appended to the name.
9864 * Generate multi-valued AVAs using '+' notation in config files for
9869 * Support for nameConstraints certificate extension.
9873 * Support for policyConstraints certificate extension.
9877 * Support for policyMappings certificate extension.
9881 * Make sure the default DSA_METHOD implementation only uses its
9882 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9883 and change its own handlers to be NULL so as to remove unnecessary
9884 indirection. This lets alternative implementations fallback to the
9885 default implementation more easily.
9889 * Support for directoryName in GeneralName related extensions
9894 * Make it possible to link applications using Makefile.shared.
9895 Make that possible even when linking against static libraries!
9899 * Support for single pass processing for S/MIME signing. This now
9900 means that S/MIME signing can be done from a pipe, in addition
9901 cleartext signing (multipart/signed type) is effectively streaming
9902 and the signed data does not need to be all held in memory.
9904 This is done with a new flag PKCS7_STREAM. When this flag is set
9905 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
9906 is done after the data is output (and digests calculated) in
9907 SMIME_write_PKCS7().
9911 * Add full support for -rpath/-R, both in shared libraries and
9912 applications, at least on the platforms where it's known how
9917 * In crypto/ec/ec_mult.c, implement fast point multiplication with
9918 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
9919 will now compute a table of multiples of the generator that
9920 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
9921 faster (notably in the case of a single point multiplication,
9922 scalar * generator).
9924 *Nils Larsch, Bodo Moeller*
9926 * IPv6 support for certificate extensions. The various extensions
9927 which use the IP:a.b.c.d can now take IPv6 addresses using the
9928 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
9933 * Added an ENGINE that implements RSA by performing private key
9934 exponentiations with the GMP library. The conversions to and from
9935 GMP's mpz_t format aren't optimised nor are any montgomery forms
9936 cached, and on x86 it appears OpenSSL's own performance has caught up.
9937 However there are likely to be other architectures where GMP could
9938 provide a boost. This ENGINE is not built in by default, but it can be
9939 specified at Configure time and should be accompanied by the necessary
9940 linker additions, eg;
9941 ./config -DOPENSSL_USE_GMP -lgmp
9945 * "openssl engine" will not display ENGINE/DSO load failure errors when
9946 testing availability of engines with "-t" - the old behaviour is
9947 produced by increasing the feature's verbosity with "-tt".
9951 * ECDSA routines: under certain error conditions uninitialized BN objects
9952 could be freed. Solution: make sure initialization is performed early
9953 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
9958 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
9959 and DH_METHOD (eg. by ENGINE implementations) to override the normal
9960 software implementations. For DSA and DH, parameter generation can
9961 also be overridden by providing the appropriate method callbacks.
9965 * Change the "progress" mechanism used in key-generation and
9966 primality testing to functions that take a new BN_GENCB pointer in
9967 place of callback/argument pairs. The new API functions have "_ex"
9968 postfixes and the older functions are reimplemented as wrappers for
9969 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
9970 declarations of the old functions to help (graceful) attempts to
9971 migrate to the new functions. Also, the new key-generation API
9972 functions operate on a caller-supplied key-structure and return
9973 success/failure rather than returning a key or NULL - this is to
9974 help make "keygen" another member function of RSA_METHOD etc.
9976 Example for using the new callback interface:
9978 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
9982 BN_GENCB_set(&my_cb, my_callback, my_arg);
9984 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
9985 /* For the meaning of a, b in calls to my_callback(), see the
9986 * documentation of the function that calls the callback.
9987 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
9988 * my_callback should return 1 if it wants BN_is_prime_ex()
9989 * to continue, or 0 to stop.
9994 * Change the ZLIB compression method to be stateful, and make it
9995 available to TLS with the number defined in
9996 draft-ietf-tls-compression-04.txt.
10000 * Add the ASN.1 structures and functions for CertificatePair, which
10001 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10003 CertificatePair ::= SEQUENCE {
10004 forward [0] Certificate OPTIONAL,
10005 reverse [1] Certificate OPTIONAL,
10006 -- at least one of the pair shall be present -- }
10008 Also implement the PEM functions to read and write certificate
10009 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10011 This needed to be defined, mostly for the sake of the LDAP
10012 attribute crossCertificatePair, but may prove useful elsewhere as
10017 * Make it possible to inhibit symlinking of shared libraries in
10018 Makefile.shared, for Cygwin's sake.
10022 * Extend the BIGNUM API by creating a function
10023 void BN_set_negative(BIGNUM *a, int neg);
10024 and a macro that behave like
10025 int BN_is_negative(const BIGNUM *a);
10027 to avoid the need to access 'a->neg' directly in applications.
10031 * Implement fast modular reduction for pseudo-Mersenne primes
10032 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10033 EC_GROUP_new_curve_GFp() will now automatically use this
10036 *Nils Larsch <nla@trustcenter.de>*
10038 * Add new lock type (CRYPTO_LOCK_BN).
10042 * Change the ENGINE framework to automatically load engines
10043 dynamically from specific directories unless they could be
10044 found to already be built in or loaded. Move all the
10045 current engines except for the cryptodev one to a new
10046 directory engines/.
10047 The engines in engines/ are built as shared libraries if
10048 the "shared" options was given to ./Configure or ./config.
10049 Otherwise, they are inserted in libcrypto.a.
10050 /usr/local/ssl/engines is the default directory for dynamic
10051 engines, but that can be overridden at configure time through
10052 the usual use of --prefix and/or --openssldir, and at run
10053 time with the environment variable OPENSSL_ENGINES.
10055 *Geoff Thorpe and Richard Levitte*
10057 * Add Makefile.shared, a helper makefile to build shared
10058 libraries. Adapt Makefile.org.
10062 * Add version info to Win32 DLLs.
10064 *Peter 'Luna' Runestig" <peter@runestig.com>*
10066 * Add new 'medium level' PKCS#12 API. Certificates and keys
10067 can be added using this API to created arbitrary PKCS#12
10068 files while avoiding the low level API.
10070 New options to PKCS12_create(), key or cert can be NULL and
10071 will then be omitted from the output file. The encryption
10072 algorithm NIDs can be set to -1 for no encryption, the mac
10073 iteration count can be set to 0 to omit the mac.
10075 Enhance pkcs12 utility by making the -nokeys and -nocerts
10076 options work when creating a PKCS#12 file. New option -nomac
10077 to omit the mac, NONE can be set for an encryption algorithm.
10078 New code is modified to use the enhanced PKCS12_create()
10079 instead of the low level API.
10083 * Extend ASN1 encoder to support indefinite length constructed
10084 encoding. This can output sequences tags and octet strings in
10085 this form. Modify pk7_asn1.c to support indefinite length
10086 encoding. This is experimental and needs additional code to
10087 be useful, such as an ASN1 bio and some enhanced streaming
10090 Extend template encode functionality so that tagging is passed
10091 down to the template encoder.
10095 * Let 'openssl req' fail if an argument to '-newkey' is not
10096 recognized instead of using RSA as a default.
10100 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10101 As these are not official, they are not included in "ALL";
10102 the "ECCdraft" ciphersuite group alias can be used to select them.
10104 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10106 * Add ECDH engine support.
10108 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10110 * Add ECDH in new directory crypto/ecdh/.
10112 *Douglas Stebila (Sun Microsystems Laboratories)*
10114 * Let BN_rand_range() abort with an error after 100 iterations
10115 without success (which indicates a broken PRNG).
10119 * Change BN_mod_sqrt() so that it verifies that the input value
10120 is really the square of the return value. (Previously,
10121 BN_mod_sqrt would show GIGO behaviour.)
10125 * Add named elliptic curves over binary fields from X9.62, SECG,
10126 and WAP/WTLS; add OIDs that were still missing.
10128 *Sheueling Chang Shantz and Douglas Stebila
10129 (Sun Microsystems Laboratories)*
10131 * Extend the EC library for elliptic curves over binary fields
10132 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10135 EC_GF2m_simple_method
10139 EC_GROUP_new_curve_GF2m
10140 EC_GROUP_set_curve_GF2m
10141 EC_GROUP_get_curve_GF2m
10142 EC_POINT_set_affine_coordinates_GF2m
10143 EC_POINT_get_affine_coordinates_GF2m
10144 EC_POINT_set_compressed_coordinates_GF2m
10146 Point compression for binary fields is disabled by default for
10147 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10150 As binary polynomials are represented as BIGNUMs, various members
10151 of the EC_GROUP and EC_POINT data structures can be shared
10152 between the implementations for prime fields and binary fields;
10153 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
10154 are essentially identical to their ..._GFp counterparts.
10155 (For simplicity, the '..._GFp' prefix has been dropped from
10156 various internal method names.)
10158 An internal 'field_div' method (similar to 'field_mul' and
10159 'field_sqr') has been added; this is used only for binary fields.
10161 *Sheueling Chang Shantz and Douglas Stebila
10162 (Sun Microsystems Laboratories)*
10164 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10165 through methods ('mul', 'precompute_mult').
10167 The generic implementations (now internally called 'ec_wNAF_mul'
10168 and 'ec_wNAF_precomputed_mult') remain the default if these
10169 methods are undefined.
10171 *Sheueling Chang Shantz and Douglas Stebila
10172 (Sun Microsystems Laboratories)*
10174 * New function EC_GROUP_get_degree, which is defined through
10175 EC_METHOD. For curves over prime fields, this returns the bit
10176 length of the modulus.
10178 *Sheueling Chang Shantz and Douglas Stebila
10179 (Sun Microsystems Laboratories)*
10181 * New functions EC_GROUP_dup, EC_POINT_dup.
10182 (These simply call ..._new and ..._copy).
10184 *Sheueling Chang Shantz and Douglas Stebila
10185 (Sun Microsystems Laboratories)*
10187 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10188 Polynomials are represented as BIGNUMs (where the sign bit is not
10189 used) in the following functions [macros]:
10192 BN_GF2m_sub [= BN_GF2m_add]
10193 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10194 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10195 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10197 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10198 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10199 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10200 BN_GF2m_cmp [= BN_ucmp]
10202 (Note that only the 'mod' functions are actually for fields GF(2^m).
10203 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10205 For some functions, an the irreducible polynomial defining a
10206 field can be given as an 'unsigned int[]' with strictly
10207 decreasing elements giving the indices of those bits that are set;
10208 i.e., p[] represents the polynomial
10209 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10211 p[0] > p[1] > ... > p[k] = 0.
10212 This applies to the following functions:
10215 BN_GF2m_mod_mul_arr
10216 BN_GF2m_mod_sqr_arr
10217 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10218 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10219 BN_GF2m_mod_exp_arr
10220 BN_GF2m_mod_sqrt_arr
10221 BN_GF2m_mod_solve_quad_arr
10225 Conversion can be performed by the following functions:
10230 bntest.c has additional tests for binary polynomial arithmetic.
10232 Two implementations for BN_GF2m_mod_div() are available.
10233 The default algorithm simply uses BN_GF2m_mod_inv() and
10234 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10235 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10236 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10238 *Sheueling Chang Shantz and Douglas Stebila
10239 (Sun Microsystems Laboratories)*
10241 * Add new error code 'ERR_R_DISABLED' that can be used when some
10242 functionality is disabled at compile-time.
10244 *Douglas Stebila <douglas.stebila@sun.com>*
10246 * Change default behaviour of 'openssl asn1parse' so that more
10247 information is visible when viewing, e.g., a certificate:
10249 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10250 mode the content of non-printable OCTET STRINGs is output in a
10251 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10252 avoid the appearance of a printable string.
10254 *Nils Larsch <nla@trustcenter.de>*
10256 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10258 EC_GROUP_set_asn1_flag()
10259 EC_GROUP_get_asn1_flag()
10260 EC_GROUP_set_point_conversion_form()
10261 EC_GROUP_get_point_conversion_form()
10262 These control ASN1 encoding details:
10263 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10264 has been set to OPENSSL_EC_NAMED_CURVE.
10265 - Points are encoded in uncompressed form by default; options for
10266 asn1_for are as for point2oct, namely
10267 POINT_CONVERSION_COMPRESSED
10268 POINT_CONVERSION_UNCOMPRESSED
10269 POINT_CONVERSION_HYBRID
10271 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10273 EC_GROUP_set_seed()
10274 EC_GROUP_get0_seed()
10275 EC_GROUP_get_seed_len()
10276 This is used only for ASN1 purposes (so far).
10278 *Nils Larsch <nla@trustcenter.de>*
10280 * Add 'field_type' member to EC_METHOD, which holds the NID
10281 of the appropriate field type OID. The new function
10282 EC_METHOD_get_field_type() returns this value.
10284 *Nils Larsch <nla@trustcenter.de>*
10287 EC_POINT_point2bn()
10288 EC_POINT_bn2point()
10289 EC_POINT_point2hex()
10290 EC_POINT_hex2point()
10291 providing useful interfaces to EC_POINT_point2oct() and
10292 EC_POINT_oct2point().
10294 *Nils Larsch <nla@trustcenter.de>*
10296 * Change internals of the EC library so that the functions
10297 EC_GROUP_set_generator()
10298 EC_GROUP_get_generator()
10299 EC_GROUP_get_order()
10300 EC_GROUP_get_cofactor()
10301 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10302 to methods, which would lead to unnecessary code duplication when
10303 adding different types of curves.
10305 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10307 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10308 arithmetic, and such that modified wNAFs are generated
10309 (which avoid length expansion in many cases).
10313 * Add a function EC_GROUP_check_discriminant() (defined via
10314 EC_METHOD) that verifies that the curve discriminant is non-zero.
10316 Add a function EC_GROUP_check() that makes some sanity tests
10317 on a EC_GROUP, its generator and order. This includes
10318 EC_GROUP_check_discriminant().
10320 *Nils Larsch <nla@trustcenter.de>*
10322 * Add ECDSA in new directory crypto/ecdsa/.
10324 Add applications 'openssl ecparam' and 'openssl ecdsa'
10325 (these are based on 'openssl dsaparam' and 'openssl dsa').
10327 ECDSA support is also included in various other files across the
10328 library. Most notably,
10329 - 'openssl req' now has a '-newkey ecdsa:file' option;
10330 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10331 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10332 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10333 them suitable for ECDSA where domain parameters must be
10334 extracted before the specific public key;
10335 - ECDSA engine support has been added.
10337 *Nils Larsch <nla@trustcenter.de>*
10339 * Include some named elliptic curves, and add OIDs from X9.62,
10340 SECG, and WAP/WTLS. Each curve can be obtained from the new
10342 EC_GROUP_new_by_curve_name(),
10343 and the list of available named curves can be obtained with
10344 EC_get_builtin_curves().
10345 Also add a 'curve_name' member to EC_GROUP objects, which can be
10347 EC_GROUP_set_curve_name()
10348 EC_GROUP_get_curve_name()
10350 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10352 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10353 was actually never needed) and in BN_mul(). The removal in BN_mul()
10354 required a small change in bn_mul_part_recursive() and the addition
10355 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10356 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10357 bn_sub_words() and bn_add_words() except they take arrays with
10362 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007] ###
10364 * Cleanse PEM buffers before freeing them since they may contain
10367 *Benjamin Bennett <ben@psc.edu>*
10369 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10370 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10371 authentication-only ciphersuites.
10375 * Since AES128 and AES256 share a single mask bit in the logic of
10376 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10377 kludge to work properly if AES128 is available and AES256 isn't.
10381 * Expand security boundary to match 1.1.1 module.
10385 * Remove redundant features: hash file source, editing of test vectors
10386 modify fipsld to use external fips_premain.c signature.
10390 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10391 run algorithm test programs.
10395 * Make algorithm test programs more tolerant of whitespace.
10399 * Have SSL/TLS server implementation tolerate "mismatched" record
10400 protocol version while receiving ClientHello even if the
10401 ClientHello is fragmented. (The server can't insist on the
10402 particular protocol version it has chosen before the ServerHello
10403 message has informed the client about his choice.)
10407 * Load error codes if they are not already present instead of using a
10408 static variable. This allows them to be cleanly unloaded and reloaded.
10412 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006] ###
10414 * Introduce limits to prevent malicious keys being able to
10415 cause a denial of service. [CVE-2006-2940][]
10417 *Steve Henson, Bodo Moeller*
10419 * Fix ASN.1 parsing of certain invalid structures that can result
10420 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10422 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10423 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10425 * Fix SSL client code which could crash if connecting to a
10426 malicious SSLv2 server. [CVE-2006-4343][]
10428 *Tavis Ormandy and Will Drewry, Google Security Team*
10430 * Change ciphersuite string processing so that an explicit
10431 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10432 will no longer include "AES128-SHA"), and any other similar
10433 ciphersuite (same bitmap) from *other* protocol versions (so that
10434 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10435 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10436 changes from 0.9.8b and 0.9.8d.
10440 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006] ###
10442 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10443 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10445 * Change the Unix randomness entropy gathering to use poll() when
10446 possible instead of select(), since the latter has some
10447 undesirable limitations.
10449 *Darryl Miles via Richard Levitte and Bodo Moeller*
10451 * Disable rogue ciphersuites:
10453 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10454 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10455 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10457 The latter two were purportedly from
10458 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10461 Also deactivate the remaining ciphersuites from
10462 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10463 unofficial, and the ID has long expired.
10467 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10468 dual-core machines) and other potential thread-safety issues.
10472 ### Changes between 0.9.7i and 0.9.7j [04 May 2006] ###
10474 * Adapt fipsld and the build system to link against the validated FIPS
10475 module in FIPS mode.
10479 * Fixes for VC++ 2005 build under Windows.
10483 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10484 from a Windows bash shell such as MSYS. It is autodetected from the
10485 "config" script when run from a VC++ environment. Modify standard VC++
10486 build to use fipscanister.o from the GNU make build.
10490 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005] ###
10492 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10493 The value now differs depending on if you build for FIPS or not.
10494 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10495 safely run with a non-FIPSed libcrypto, as it may crash because of
10496 the difference induced by this change.
10500 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005] ###
10502 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10503 (part of SSL_OP_ALL). This option used to disable the
10504 countermeasure against man-in-the-middle protocol-version
10505 rollback in the SSL 2.0 server implementation, which is a bad
10506 idea. [CVE-2005-2969][]
10508 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10509 for Information Security, National Institute of Advanced Industrial
10510 Science and Technology [AIST*, Japan)]
10512 * Minimal support for X9.31 signatures and PSS padding modes. This is
10513 mainly for FIPS compliance and not fully integrated at this stage.
10517 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10518 the exponentiation using a fixed-length exponent. (Otherwise,
10519 the information leaked through timing could expose the secret key
10520 after many signatures; cf. Bleichenbacher's attack on DSA with
10525 * Make a new fixed-window mod_exp implementation the default for
10526 RSA, DSA, and DH private-key operations so that the sequence of
10527 squares and multiplies and the memory access pattern are
10528 independent of the particular secret key. This will mitigate
10529 cache-timing and potential related attacks.
10531 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10532 and this is automatically used by BN_mod_exp_mont() if the new flag
10533 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10534 will use this BN flag for private exponents unless the flag
10535 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10536 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10539 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10541 * Change the client implementation for SSLv23_method() and
10542 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10543 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10544 (Previously, the SSL 2.0 backwards compatible Client Hello
10545 message format would be used even with SSL_OP_NO_SSLv2.)
10549 * Add support for smime-type MIME parameter in S/MIME messages which some
10554 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10555 a threadsafe manner. Modify rsa code to use new function and add calls
10556 to dsa and dh code (which had race conditions before).
10560 * Include the fixed error library code in the C error file definitions
10561 instead of fixing them up at runtime. This keeps the error code
10562 structures constant.
10566 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005] ###
10568 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10571 * Fixes for newer kerberos headers. NB: the casts are needed because
10572 the 'length' field is signed on one version and unsigned on another
10573 with no (?) obvious way to tell the difference, without these VC++
10574 complains. Also the "definition" of FAR (blank) is no longer included
10575 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10576 some needed definitions.
10580 * Undo Cygwin change.
10584 * Added support for proxy certificates according to RFC 3820.
10585 Because they may be a security thread to unaware applications,
10586 they must be explicitly allowed in run-time. See
10587 docs/HOWTO/proxy_certificates.txt for further information.
10591 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005] ###
10593 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10594 server and client random values. Previously
10595 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10596 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10598 This change has negligible security impact because:
10600 1. Server and client random values still have 24 bytes of pseudo random
10603 2. Server and client random values are sent in the clear in the initial
10606 3. The master secret is derived using the premaster secret (48 bytes in
10607 size for static RSA ciphersuites) as well as client server and random
10610 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10613 *Stephen Henson, reported by UK NISCC*
10615 * Use Windows randomness collection on Cygwin.
10619 * Fix hang in EGD/PRNGD query when communication socket is closed
10620 prematurely by EGD/PRNGD.
10622 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10624 * Prompt for pass phrases when appropriate for PKCS12 input format.
10628 * Back-port of selected performance improvements from development
10629 branch, as well as improved support for PowerPC platforms.
10633 * Add lots of checks for memory allocation failure, error codes to indicate
10634 failure and freeing up memory if a failure occurs.
10636 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10638 * Add new -passin argument to dgst.
10642 * Perform some character comparisons of different types in X509_NAME_cmp:
10643 this is needed for some certificates that re-encode DNs into UTF8Strings
10644 (in violation of RFC3280) and can't or won't issue name rollover
10649 * Make an explicit check during certificate validation to see that
10650 the CA setting in each certificate on the chain is correct. As a
10651 side effect always do the following basic checks on extensions,
10652 not just when there's an associated purpose to the check:
10654 - if there is an unhandled critical extension (unless the user
10655 has chosen to ignore this fault)
10656 - if the path length has been exceeded (if one is set at all)
10657 - that certain extensions fit the associated purpose (if one has
10662 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004] ###
10664 * Avoid a race condition when CRLs are checked in a multi threaded
10665 environment. This would happen due to the reordering of the revoked
10666 entries during signature checking and serial number lookup. Now the
10667 encoding is cached and the serial number sort performed under a lock.
10668 Add new STACK function sk_is_sorted().
10672 * Add Delta CRL to the extension code.
10676 * Various fixes to s3_pkt.c so alerts are sent properly.
10678 *David Holmes <d.holmes@f5.com>*
10680 * Reduce the chances of duplicate issuer name and serial numbers (in
10681 violation of RFC3280) using the OpenSSL certificate creation utilities.
10682 This is done by creating a random 64 bit value for the initial serial
10683 number when a serial number file is created or when a self signed
10684 certificate is created using 'openssl req -x509'. The initial serial
10685 number file is created using 'openssl x509 -next_serial' in CA.pl
10686 rather than being initialized to 1.
10690 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004] ###
10692 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10693 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10695 *Joe Orton, Steve Henson*
10697 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10700 *Joe Orton, Steve Henson*
10702 * Make it possible to have multiple active certificates with the same
10703 subject in the CA index file. This is done only if the keyword
10704 'unique_subject' is set to 'no' in the main CA section (default
10705 if 'CA_default') of the configuration file. The value is saved
10706 with the database itself in a separate index attribute file,
10707 named like the index file with '.attr' appended to the name.
10711 * X509 verify fixes. Disable broken certificate workarounds when
10712 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10713 keyUsage extension present. Don't accept CRLs with unhandled critical
10714 extensions: since verify currently doesn't process CRL extensions this
10715 rejects a CRL with *any* critical extensions. Add new verify error codes
10720 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10721 A clarification of RFC2560 will require the use of OCTET STRINGs and
10722 some implementations cannot handle the current raw format. Since OpenSSL
10723 copies and compares OCSP nonces as opaque blobs without any attempt at
10724 parsing them this should not create any compatibility issues.
10728 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10729 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10730 this HMAC (and other) operations are several times slower than OpenSSL
10735 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10737 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10739 * Use the correct content when signing type "other".
10743 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003] ###
10745 * Fix various bugs revealed by running the NISCC test suite:
10747 Stop out of bounds reads in the ASN1 code when presented with
10748 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10750 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10752 If verify callback ignores invalid public key errors don't try to check
10753 certificate signature with the NULL public key.
10758 * New -ignore_err option in ocsp application to stop the server
10759 exiting on the first error in a request.
10763 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10764 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10769 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10770 extra data after the compression methods not only for TLS 1.0
10771 but also for SSL 3.0 (as required by the specification).
10773 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10775 * Change X509_certificate_type() to mark the key as exported/exportable
10776 when it's 512 *bits* long, not 512 bytes.
10780 * Change AES_cbc_encrypt() so it outputs exact multiple of
10781 blocks during encryption.
10785 * Various fixes to base64 BIO and non blocking I/O. On write
10786 flushes were not handled properly if the BIO retried. On read
10787 data was not being buffered properly and had various logic bugs.
10788 This also affects blocking I/O when the data being decoded is a
10793 * Various S/MIME bugfixes and compatibility changes:
10794 output correct application/pkcs7 MIME type if
10795 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10796 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10797 of files as .eml work). Correctly handle very long lines in MIME
10802 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003] ###
10804 * Countermeasure against the Klima-Pokorny-Rosa extension of
10805 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10806 a protocol version number mismatch like a decryption error
10807 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10811 * Turn on RSA blinding by default in the default implementation
10812 to avoid a timing attack. Applications that don't want it can call
10813 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10814 They would be ill-advised to do so in most cases.
10816 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10818 * Change RSA blinding code so that it works when the PRNG is not
10819 seeded (in this case, the secret RSA exponent is abused as
10820 an unpredictable seed -- if it is not unpredictable, there
10821 is no point in blinding anyway). Make RSA blinding thread-safe
10822 by remembering the creator's thread ID in rsa->blinding and
10823 having all other threads use local one-time blinding factors
10824 (this requires more computation than sharing rsa->blinding, but
10825 avoids excessive locking; and if an RSA object is not shared
10826 between threads, blinding will still be very fast).
10830 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10831 ENGINE as defaults for all supported algorithms irrespective of
10832 the 'flags' parameter. 'flags' is now honoured, so applications
10833 should make sure they are passing it correctly.
10837 * Target "mingw" now allows native Windows code to be generated in
10838 the Cygwin environment as well as with the MinGW compiler.
10842 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003] ###
10844 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10845 via timing by performing a MAC computation even if incorrect
10846 block cipher padding has been found. This is a countermeasure
10847 against active attacks where the attacker has to distinguish
10848 between bad padding and a MAC verification error. [CVE-2003-0078][]
10850 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10851 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10852 Martin Vuagnoux (EPFL, Ilion)*
10854 * Make the no-err option work as intended. The intention with no-err
10855 is not to have the whole error stack handling routines removed from
10856 libcrypto, it's only intended to remove all the function name and
10857 reason texts, thereby removing some of the footprint that may not
10858 be interesting if those errors aren't displayed anyway.
10860 NOTE: it's still possible for any application or module to have its
10861 own set of error texts inserted. The routines are there, just not
10862 used by default when no-err is given.
10866 * Add support for FreeBSD on IA64.
10868 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10870 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10871 Kerberos function mit_des_cbc_cksum(). Before this change,
10872 the value returned by DES_cbc_cksum() was like the one from
10873 mit_des_cbc_cksum(), except the bytes were swapped.
10875 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10877 * Allow an application to disable the automatic SSL chain building.
10878 Before this a rather primitive chain build was always performed in
10879 ssl3_output_cert_chain(): an application had no way to send the
10880 correct chain if the automatic operation produced an incorrect result.
10882 Now the chain builder is disabled if either:
10884 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10886 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10888 The reasoning behind this is that an application would not want the
10889 auto chain building to take place if extra chain certificates are
10890 present and it might also want a means of sending no additional
10891 certificates (for example the chain has two certificates and the
10896 * Add the possibility to build without the ENGINE framework.
10898 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10900 * Under Win32 gmtime() can return NULL: check return value in
10901 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
10905 * DSA routines: under certain error conditions uninitialized BN objects
10906 could be freed. Solution: make sure initialization is performed early
10907 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
10908 Nils Larsch <nla@trustcenter.de> via PR#459)
10912 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
10913 checked on reconnect on the client side, therefore session resumption
10914 could still fail with a "ssl session id is different" error. This
10915 behaviour is masked when SSL_OP_ALL is used due to
10916 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
10917 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10918 followup to PR #377.
10922 * IA-32 assembler support enhancements: unified ELF targets, support
10923 for SCO/Caldera platforms, fix for Cygwin shared build.
10927 * Add support for FreeBSD on sparc64. As a consequence, support for
10928 FreeBSD on non-x86 processors is separate from x86 processors on
10929 the config script, much like the NetBSD support.
10931 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
10933 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002] ###
10935 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
10938 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
10939 code (06) was taken as the first octet of the session ID and the last
10940 octet was ignored consequently. As a result SSLv2 client side session
10941 caching could not have worked due to the session ID mismatch between
10943 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10948 * Change the declaration of needed Kerberos libraries to use EX_LIBS
10949 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
10954 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
10955 seems that in spite of existing for more than a year, many application
10956 author have done nothing to provide the necessary callbacks, which
10957 means that this particular engine will not work properly anywhere.
10958 This is a very unfortunate situation which forces us, in the name
10959 of usability, to give the hw_ncipher.c a static lock, which is part
10961 NOTE: This is for the 0.9.7 series ONLY. This hack will never
10962 appear in 0.9.8 or later. We EXPECT application authors to have
10963 dealt properly with this when 0.9.8 is released (unless we actually
10964 make such changes in the libcrypto locking code that changes will
10965 have to be made anyway).
10969 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
10970 octets have been read, EOF or an error occurs. Without this change
10971 some truncated ASN1 structures will not produce an error.
10975 * Disable Heimdal support, since it hasn't been fully implemented.
10976 Still give the possibility to force the use of Heimdal, but with
10977 warnings and a request that patches get sent to openssl-dev.
10981 * Add the VC-CE target, introduce the WINCE sysname, and add
10982 INSTALL.WCE and appropriate conditionals to make it build.
10984 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10986 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
10987 cygssl-x.y.z.dll, where x, y and z are the major, minor and
10988 edit numbers of the version.
10990 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
10992 * Introduce safe string copy and catenation functions
10993 (BUF_strlcpy() and BUF_strlcat()).
10995 *Ben Laurie (CHATS) and Richard Levitte*
10997 * Avoid using fixed-size buffers for one-line DNs.
10999 *Ben Laurie (CHATS)*
11001 * Add BUF_MEM_grow_clean() to avoid information leakage when
11002 resizing buffers containing secrets, and use where appropriate.
11004 *Ben Laurie (CHATS)*
11006 * Avoid using fixed size buffers for configuration file location.
11008 *Ben Laurie (CHATS)*
11010 * Avoid filename truncation for various CA files.
11012 *Ben Laurie (CHATS)*
11014 * Use sizeof in preference to magic numbers.
11016 *Ben Laurie (CHATS)*
11018 * Avoid filename truncation in cert requests.
11020 *Ben Laurie (CHATS)*
11022 * Add assertions to check for (supposedly impossible) buffer
11025 *Ben Laurie (CHATS)*
11027 * Don't cache truncated DNS entries in the local cache (this could
11028 potentially lead to a spoofing attack).
11030 *Ben Laurie (CHATS)*
11032 * Fix various buffers to be large enough for hex/decimal
11033 representations in a platform independent manner.
11035 *Ben Laurie (CHATS)*
11037 * Add CRYPTO_realloc_clean() to avoid information leakage when
11038 resizing buffers containing secrets, and use where appropriate.
11040 *Ben Laurie (CHATS)*
11042 * Add BIO_indent() to avoid much slightly worrying code to do
11045 *Ben Laurie (CHATS)*
11047 * Convert sprintf()/BIO_puts() to BIO_printf().
11049 *Ben Laurie (CHATS)*
11051 * buffer_gets() could terminate with the buffer only half
11054 *Ben Laurie (CHATS)*
11056 * Add assertions to prevent user-supplied crypto functions from
11057 overflowing internal buffers by having large block sizes, etc.
11059 *Ben Laurie (CHATS)*
11061 * New OPENSSL_assert() macro (similar to assert(), but enabled
11064 *Ben Laurie (CHATS)*
11066 * Eliminate unused copy of key in RC4.
11068 *Ben Laurie (CHATS)*
11070 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11072 *Ben Laurie (CHATS)*
11074 * Fix off-by-one error in EGD path.
11076 *Ben Laurie (CHATS)*
11078 * If RANDFILE path is too long, ignore instead of truncating.
11080 *Ben Laurie (CHATS)*
11082 * Eliminate unused and incorrectly sized X.509 structure
11085 *Ben Laurie (CHATS)*
11087 * Eliminate unused and dangerous function knumber().
11089 *Ben Laurie (CHATS)*
11091 * Eliminate unused and dangerous structure, KSSL_ERR.
11093 *Ben Laurie (CHATS)*
11095 * Protect against overlong session ID context length in an encoded
11096 session object. Since these are local, this does not appear to be
11099 *Ben Laurie (CHATS)*
11101 * Change from security patch (see 0.9.6e below) that did not affect
11102 the 0.9.6 release series:
11104 Remote buffer overflow in SSL3 protocol - an attacker could
11105 supply an oversized master key in Kerberos-enabled versions.
11108 *Ben Laurie (CHATS)*
11110 * Change the SSL kerb5 codes to match RFC 2712.
11114 * Make -nameopt work fully for req and add -reqopt switch.
11116 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11118 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11120 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11122 * Make sure tests can be performed even if the corresponding algorithms
11123 have been removed entirely. This was also the last step to make
11124 OpenSSL compilable with DJGPP under all reasonable conditions.
11126 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11128 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11129 to allow version independent disabling of normally unselected ciphers,
11130 which may be activated as a side-effect of selecting a single cipher.
11132 (E.g., cipher list string "RSA" enables ciphersuites that are left
11133 out of "ALL" because they do not provide symmetric encryption.
11134 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11136 *Lutz Jaenicke, Bodo Moeller*
11138 * Add appropriate support for separate platform-dependent build
11139 directories. The recommended way to make a platform-dependent
11140 build directory is the following (tested on Linux), maybe with
11143 # Place yourself outside of the OpenSSL source tree. In
11144 # this example, the environment variable OPENSSL_SOURCE
11145 # is assumed to contain the absolute OpenSSL source directory.
11146 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11147 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11148 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11149 mkdir -p `dirname $F`
11150 ln -s $OPENSSL_SOURCE/$F $F
11153 To be absolutely sure not to disturb the source tree, a "make clean"
11154 is a good thing. If it isn't successful, don't worry about it,
11155 it probably means the source directory is very clean.
11159 * Make sure any ENGINE control commands make local copies of string
11160 pointers passed to them whenever necessary. Otherwise it is possible
11161 the caller may have overwritten (or deallocated) the original string
11162 data when a later ENGINE operation tries to use the stored values.
11164 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11166 * Improve diagnostics in file reading and command-line digests.
11168 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11170 * Add AES modes CFB and OFB to the object database. Correct an
11171 error in AES-CFB decryption.
11175 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11176 allows existing EVP_CIPHER_CTX structures to be reused after
11177 calling EVP_*Final(). This behaviour is used by encryption
11178 BIOs and some applications. This has the side effect that
11179 applications must explicitly clean up cipher contexts with
11180 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11184 * Check the values of dna and dnb in bn_mul_recursive before calling
11185 bn_mul_comba (a non zero value means the a or b arrays do not contain
11186 n2 elements) and fallback to bn_mul_normal if either is not zero.
11190 * Fix escaping of non-ASCII characters when using the -subj option
11191 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11195 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11196 form for "surname", serialNumber has no short form.
11197 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11198 therefore remove "mail" short name for "internet 7".
11199 The OID for unique identifiers in X509 certificates is
11200 x500UniqueIdentifier, not uniqueIdentifier.
11201 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11205 * Add an "init" command to the ENGINE config module and auto initialize
11206 ENGINEs. Without any "init" command the ENGINE will be initialized
11207 after all ctrl commands have been executed on it. If init=1 the
11208 ENGINE is initialized at that point (ctrls before that point are run
11209 on the uninitialized ENGINE and after on the initialized one). If
11210 init=0 then the ENGINE will not be initialized at all.
11214 * Fix the 'app_verify_callback' interface so that the user-defined
11215 argument is actually passed to the callback: In the
11216 SSL_CTX_set_cert_verify_callback() prototype, the callback
11217 declaration has been changed from
11220 int (*cb)(X509_STORE_CTX *,void *);
11221 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11222 i=s->ctx->app_verify_callback(&ctx)
11223 has been changed into
11224 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11226 To update applications using SSL_CTX_set_cert_verify_callback(),
11227 a dummy argument can be added to their callback functions.
11229 *D. K. Smetters <smetters@parc.xerox.com>*
11231 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11233 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11235 * Add and OPENSSL_LOAD_CONF define which will cause
11236 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11237 This allows older applications to transparently support certain
11238 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11239 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11240 load the config file and OPENSSL_add_all_algorithms_conf() which will
11241 always load it have also been added.
11245 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11246 Adjust NIDs and EVP layer.
11248 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11250 * Config modules support in openssl utility.
11252 Most commands now load modules from the config file,
11253 though in a few (such as version) this isn't done
11254 because it couldn't be used for anything.
11256 In the case of ca and req the config file used is
11257 the same as the utility itself: that is the -config
11258 command line option can be used to specify an
11263 * Move default behaviour from OPENSSL_config(). If appname is NULL
11264 use "openssl_conf" if filename is NULL use default openssl config file.
11268 * Add an argument to OPENSSL_config() to allow the use of an alternative
11269 config section name. Add a new flag to tolerate a missing config file
11270 and move code to CONF_modules_load_file().
11274 * Support for crypto accelerator cards from Accelerated Encryption
11275 Processing, www.aep.ie. (Use engine 'aep')
11276 The support was copied from 0.9.6c [engine] and adapted/corrected
11277 to work with the new engine framework.
11279 *AEP Inc. and Richard Levitte*
11281 * Support for SureWare crypto accelerator cards from Baltimore
11282 Technologies. (Use engine 'sureware')
11283 The support was copied from 0.9.6c [engine] and adapted
11284 to work with the new engine framework.
11288 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11289 make the newer ENGINE framework commands for the CHIL engine work.
11291 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11293 * Make it possible to produce shared libraries on ReliantUNIX.
11295 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11297 * Add the configuration target debug-linux-ppro.
11298 Make 'openssl rsa' use the general key loading routines
11299 implemented in apps.c, and make those routines able to
11300 handle the key format FORMAT_NETSCAPE and the variant
11303 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11305 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11307 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11309 * Add -keyform to rsautl, and document -engine.
11311 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11313 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11314 BIO_R_NO_SUCH_FILE error code rather than the generic
11315 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11319 * Add new functions
11320 ERR_peek_last_error
11321 ERR_peek_last_error_line
11322 ERR_peek_last_error_line_data.
11323 These are similar to
11325 ERR_peek_error_line
11326 ERR_peek_error_line_data,
11327 but report on the latest error recorded rather than the first one
11328 still in the error queue.
11330 *Ben Laurie, Bodo Moeller*
11332 * default_algorithms option in ENGINE config module. This allows things
11334 default_algorithms = ALL
11335 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11339 * Preliminary ENGINE config module.
11343 * New experimental application configuration code.
11347 * Change the AES code to follow the same name structure as all other
11348 symmetric ciphers, and behave the same way. Move everything to
11349 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11351 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11353 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11355 *Ben Laurie and Theo de Raadt*
11357 * Add option to output public keys in req command.
11359 *Massimiliano Pala madwolf@openca.org*
11361 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11362 (up to about 10% better than before for P-192 and P-224).
11366 * New functions/macros
11368 SSL_CTX_set_msg_callback(ctx, cb)
11369 SSL_CTX_set_msg_callback_arg(ctx, arg)
11370 SSL_set_msg_callback(ssl, cb)
11371 SSL_set_msg_callback_arg(ssl, arg)
11373 to request calling a callback function
11375 void cb(int write_p, int version, int content_type,
11376 const void *buf, size_t len, SSL *ssl, void *arg)
11378 whenever a protocol message has been completely received
11379 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11380 protocol version according to which the SSL library interprets
11381 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11382 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11383 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11384 specification (change_cipher_spec(20), alert(21), handshake(22)).
11385 'buf' and 'len' point to the actual message, 'ssl' to the
11386 SSL object, and 'arg' is the application-defined value set by
11387 SSL[_CTX]_set_msg_callback_arg().
11389 'openssl s_client' and 'openssl s_server' have new '-msg' options
11390 to enable a callback that displays all protocol messages.
11394 * Change the shared library support so shared libraries are built as
11395 soon as the corresponding static library is finished, and thereby get
11396 openssl and the test programs linked against the shared library.
11397 This still only happens when the keyword "shard" has been given to
11398 the configuration scripts.
11400 NOTE: shared library support is still an experimental thing, and
11401 backward binary compatibility is still not guaranteed.
11403 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11405 * Add support for Subject Information Access extension.
11407 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11409 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11410 additional bytes when new memory had to be allocated, not just
11411 when reusing an existing buffer.
11415 * New command line and configuration option 'utf8' for the req command.
11416 This allows field values to be specified as UTF8 strings.
11420 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11421 runs for the former and machine-readable output for the latter.
11425 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11426 of the e-mail address in the DN (i.e., it will go into a certificate
11427 extension only). The new configuration file option 'email_in_dn = no'
11428 has the same effect.
11430 *Massimiliano Pala madwolf@openca.org*
11432 * Change all functions with names starting with des_ to be starting
11433 with DES_ instead. Add wrappers that are compatible with libdes,
11434 but are named _ossl_old_des_*. Finally, add macros that map the
11435 des_* symbols to the corresponding _ossl_old_des_* if libdes
11436 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11437 desired, the des_* symbols will be mapped to DES_*, with one
11440 Since we provide two compatibility mappings, the user needs to
11441 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11442 compatibility is desired. The default (i.e., when that macro
11443 isn't defined) is OpenSSL 0.9.6c compatibility.
11445 There are also macros that enable and disable the support of old
11446 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11447 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11448 are defined, the default will apply: to support the old des routines.
11450 In either case, one must include openssl/des.h to get the correct
11451 definitions. Do not try to just include openssl/des_old.h, that
11454 NOTE: This is a major break of an old API into a new one. Software
11455 authors are encouraged to switch to the DES_ style functions. Some
11456 time in the future, des_old.h and the libdes compatibility functions
11457 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11458 default), and then completely removed.
11462 * Test for certificates which contain unsupported critical extensions.
11463 If such a certificate is found during a verify operation it is
11464 rejected by default: this behaviour can be overridden by either
11465 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11466 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11467 X509_supported_extension() has also been added which returns 1 if a
11468 particular extension is supported.
11472 * Modify the behaviour of EVP cipher functions in similar way to digests
11473 to retain compatibility with existing code.
11477 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11478 compatibility with existing code. In particular the 'ctx' parameter does
11479 not have to be to be initialized before the call to EVP_DigestInit() and
11480 it is tidied up after a call to EVP_DigestFinal(). New function
11481 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11482 EVP_MD_CTX_copy() changed to not require the destination to be
11483 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11484 requires the destination to be valid.
11486 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11487 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11491 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11492 so that complete 'Handshake' protocol structures are kept in memory
11493 instead of overwriting 'msg_type' and 'length' with 'body' data.
11497 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11499 *Massimo Santin via Richard Levitte*
11501 * Major restructuring to the underlying ENGINE code. This includes
11502 reduction of linker bloat, separation of pure "ENGINE" manipulation
11503 (initialisation, etc) from functionality dealing with implementations
11504 of specific crypto interfaces. This change also introduces integrated
11505 support for symmetric ciphers and digest implementations - so ENGINEs
11506 can now accelerate these by providing EVP_CIPHER and EVP_MD
11507 implementations of their own. This is detailed in crypto/engine/README
11508 as it couldn't be adequately described here. However, there are a few
11509 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11510 were changed in the original introduction of ENGINE code have now
11511 reverted back - the hooking from this code to ENGINE is now a good
11512 deal more passive and at run-time, operations deal directly with
11513 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11514 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11515 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
11516 they were not being used by the framework as there is no concept of a
11517 BIGNUM_METHOD and they could not be generalised to the new
11518 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11519 ENGINE_cpy() has been removed as it cannot be consistently defined in
11524 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11528 * Change mkdef.pl to sort symbols that get the same entry number,
11529 and make sure the automatically generated functions ERR_load_*
11530 become part of libeay.num as well.
11534 * New function SSL_renegotiate_pending(). This returns true once
11535 renegotiation has been requested (either SSL_renegotiate() call
11536 or HelloRequest/ClientHello received from the peer) and becomes
11537 false once a handshake has been completed.
11538 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11539 sends a HelloRequest, but does not ensure that a handshake takes
11540 place. SSL_renegotiate_pending() is useful for checking if the
11541 client has followed the request.)
11545 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11546 By default, clients may request session resumption even during
11547 renegotiation (if session ID contexts permit); with this option,
11548 session resumption is possible only in the first handshake.
11550 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11551 more bits available for options that should not be part of
11552 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11556 * Add some demos for certificate and certificate request creation.
11560 * Make maximum certificate chain size accepted from the peer application
11561 settable (SSL*_get/set_max_cert_list()), as proposed by
11562 "Douglas E. Engert" <deengert@anl.gov>.
11566 * Add support for shared libraries for Unixware-7
11567 (Boyd Lynn Gerber <gerberb@zenez.com>).
11571 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11572 be done prior to destruction. Use this to unload error strings from
11573 ENGINEs that load their own error strings. NB: This adds two new API
11574 functions to "get" and "set" this destroy handler in an ENGINE.
11578 * Alter all existing ENGINE implementations (except "openssl" and
11579 "openbsd") to dynamically instantiate their own error strings. This
11580 makes them more flexible to be built both as statically-linked ENGINEs
11581 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11582 Also, add stub code to each that makes building them as self-contained
11583 shared-libraries easier (see README.ENGINE).
11587 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11588 implementations into applications that are completely implemented in
11589 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11590 commands that can be used to configure what shared-library to load and
11591 to control aspects of the way it is handled. Also, made an update to
11592 the README.ENGINE file that brings its information up-to-date and
11593 provides some information and instructions on the "dynamic" ENGINE
11594 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11598 * Make it possible to unload ranges of ERR strings with a new
11599 "ERR_unload_strings" function.
11603 * Add a copy() function to EVP_MD.
11607 * Make EVP_MD routines take a context pointer instead of just the
11608 md_data void pointer.
11612 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11613 that the digest can only process a single chunk of data
11614 (typically because it is provided by a piece of
11615 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11616 is only going to provide a single chunk of data, and hence the
11617 framework needn't accumulate the data for oneshot drivers.
11621 * As with "ERR", make it possible to replace the underlying "ex_data"
11622 functions. This change also alters the storage and management of global
11623 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11624 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11625 index counters. The API functions that use this state have been changed
11626 to take a "class_index" rather than pointers to the class's local STACK
11627 and counter, and there is now an API function to dynamically create new
11628 classes. This centralisation allows us to (a) plug a lot of the
11629 thread-safety problems that existed, and (b) makes it possible to clean
11630 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11631 such data would previously have always leaked in application code and
11632 workarounds were in place to make the memory debugging turn a blind eye
11633 to it. Application code that doesn't use this new function will still
11634 leak as before, but their memory debugging output will announce it now
11635 rather than letting it slide.
11637 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11638 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11639 has a return value to indicate success or failure.
11643 * Make it possible to replace the underlying "ERR" functions such that the
11644 global state (2 LHASH tables and 2 locks) is only used by the "default"
11645 implementation. This change also adds two functions to "get" and "set"
11646 the implementation prior to it being automatically set the first time
11647 any other ERR function takes place. Ie. an application can call "get",
11648 pass the return value to a module it has just loaded, and that module
11649 can call its own "set" function using that value. This means the
11650 module's "ERR" operations will use (and modify) the error state in the
11651 application and not in its own statically linked copy of OpenSSL code.
11655 * Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
11656 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11657 the operation, and provides a more encapsulated way for external code
11658 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11659 to use these functions rather than manually incrementing the counts.
11661 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11665 * Add EVP test program.
11669 * Add symmetric cipher support to ENGINE. Expect the API to change!
11673 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11674 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11675 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11676 These allow a CRL to be built without having to access X509_CRL fields
11677 directly. Modify 'ca' application to use new functions.
11681 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11682 bug workarounds. Rollback attack detection is a security feature.
11683 The problem will only arise on OpenSSL servers when TLSv1 is not
11684 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11685 Software authors not wanting to support TLSv1 will have special reasons
11686 for their choice and can explicitly enable this option.
11688 *Bodo Moeller, Lutz Jaenicke*
11690 * Rationalise EVP so it can be extended: don't include a union of
11691 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11692 (similar to those existing for EVP_CIPHER_CTX).
11697 EVP_MD_CTX_init(&md); /* new function call */
11698 EVP_DigestInit(&md, EVP_sha1());
11699 EVP_DigestUpdate(&md, in, len);
11700 EVP_DigestFinal(&md, out, NULL);
11701 EVP_MD_CTX_cleanup(&md); /* new function call */
11706 * Make DES key schedule conform to the usual scheme, as well as
11707 correcting its structure. This means that calls to DES functions
11708 now have to pass a pointer to a des_key_schedule instead of a
11709 plain des_key_schedule (which was actually always a pointer
11712 des_key_schedule ks;
11714 des_set_key_checked(..., &ks);
11715 des_ncbc_encrypt(..., &ks, ...);
11717 (Note that a later change renames 'des_...' into 'DES_...'.)
11721 * Initial reduction of linker bloat: the use of some functions, such as
11722 PEM causes large amounts of unused functions to be linked in due to
11723 poor organisation. For example pem_all.c contains every PEM function
11724 which has a knock on effect of linking in large amounts of (unused)
11725 ASN1 code. Grouping together similar functions and splitting unrelated
11726 functions prevents this.
11730 * Cleanup of EVP macros.
11734 * Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
11735 correct _ecb suffix.
11739 * Add initial OCSP responder support to ocsp application. The
11740 revocation information is handled using the text based index
11741 use by the ca application. The responder can either handle
11742 requests generated internally, supplied in files (for example
11743 via a CGI script) or using an internal minimal server.
11747 * Add configuration choices to get zlib compression for TLS.
11751 * Changes to Kerberos SSL for RFC 2712 compliance:
11752 1. Implemented real KerberosWrapper, instead of just using
11753 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11754 2. Implemented optional authenticator field of KerberosWrapper.
11756 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11757 and authenticator structs; see crypto/krb5/.
11759 Generalized Kerberos calls to support multiple Kerberos libraries.
11760 *Vern Staats <staatsvr@asc.hpc.mil>,
11761 Jeffrey Altman <jaltman@columbia.edu>
11762 via Richard Levitte*
11764 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11765 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11766 values for each of the key sizes rather than having just
11767 parameters (and 'speed' generating keys each time).
11771 * Speed up EVP routines.
11774 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11775 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11776 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11777 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11779 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11780 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11781 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11784 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11786 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11790 * Added the OS2-EMX target.
11792 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11794 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11795 to support NCONF routines in extension code. New function CONF_set_nconf()
11796 to allow functions which take an NCONF to also handle the old LHASH
11797 structure: this means that the old CONF compatible routines can be
11798 retained (in particular wrt extensions) without having to duplicate the
11799 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11803 * Enhance the general user interface with mechanisms for inner control
11804 and with possibilities to have yes/no kind of prompts.
11808 * Change all calls to low level digest routines in the library and
11809 applications to use EVP. Add missing calls to HMAC_cleanup() and
11810 don't assume HMAC_CTX can be copied using memcpy().
11812 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11814 * Add the possibility to control engines through control names but with
11815 arbitrary arguments instead of just a string.
11816 Change the key loaders to take a UI_METHOD instead of a callback
11817 function pointer. NOTE: this breaks binary compatibility with earlier
11818 versions of OpenSSL [engine].
11819 Adapt the nCipher code for these new conditions and add a card insertion
11824 * Enhance the general user interface with mechanisms to better support
11825 dialog box interfaces, application-defined prompts, the possibility
11826 to use defaults (for example default passwords from somewhere else)
11827 and interrupts/cancellations.
11831 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11832 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11836 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11837 tidy up some unnecessarily weird code in 'sk_new()').
11839 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11841 * Change the key loading routines for ENGINEs to use the same kind
11842 callback (pem_password_cb) as all other routines that need this
11847 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11848 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11849 than this minimum value is recommended.
11853 * New random seeder for OpenVMS, using the system process statistics
11854 that are easily reachable.
11858 * Windows apparently can't transparently handle global
11859 variables defined in DLLs. Initialisations such as:
11861 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11863 won't compile. This is used by the any applications that need to
11864 declare their own ASN1 modules. This was fixed by adding the option
11865 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11866 needed for static libraries under Win32.
11870 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11871 setting of purpose and trust fields. New X509_STORE trust and
11872 purpose functions and tidy up setting in other SSL functions.
11876 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11877 structure. These are inherited by X509_STORE_CTX when it is
11878 initialised. This allows various defaults to be set in the
11879 X509_STORE structure (such as flags for CRL checking and custom
11880 purpose or trust settings) for functions which only use X509_STORE_CTX
11881 internally such as S/MIME.
11883 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11884 trust settings if they are not set in X509_STORE. This allows X509_STORE
11885 purposes and trust (in S/MIME for example) to override any set by default.
11887 Add command line options for CRL checking to smime, s_client and s_server
11892 * Initial CRL based revocation checking. If the CRL checking flag(s)
11893 are set then the CRL is looked up in the X509_STORE structure and
11894 its validity and signature checked, then if the certificate is found
11895 in the CRL the verify fails with a revoked error.
11897 Various new CRL related callbacks added to X509_STORE_CTX structure.
11899 Command line options added to 'verify' application to support this.
11901 This needs some additional work, such as being able to handle multiple
11902 CRLs with different times, extension based lookup (rather than just
11903 by subject name) and ultimately more complete V2 CRL extension
11908 * Add a general user interface API (crypto/ui/). This is designed
11909 to replace things like des_read_password and friends (backward
11910 compatibility functions using this new API are provided).
11911 The purpose is to remove prompting functions from the DES code
11912 section as well as provide for prompting through dialog boxes in
11913 a window system and the like.
11917 * Add "ex_data" support to ENGINE so implementations can add state at a
11918 per-structure level rather than having to store it globally.
11922 * Make it possible for ENGINE structures to be copied when retrieved by
11923 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
11924 This causes the "original" ENGINE structure to act like a template,
11925 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
11926 operational state can be localised to each ENGINE structure, despite the
11927 fact they all share the same "methods". New ENGINE structures returned in
11928 this case have no functional references and the return value is the single
11929 structural reference. This matches the single structural reference returned
11930 by ENGINE_by_id() normally, when it is incremented on the pre-existing
11935 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
11936 needs to match any other type at all we need to manually clear the
11941 * Changes to the "openssl engine" utility to include;
11942 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
11943 about an ENGINE's available control commands.
11944 - executing control commands from command line arguments using the
11945 '-pre' and '-post' switches. '-post' is only used if '-t' is
11946 specified and the ENGINE is successfully initialised. The syntax for
11947 the individual commands are colon-separated, for example;
11948 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
11952 * New dynamic control command support for ENGINEs. ENGINEs can now
11953 declare their own commands (numbers), names (strings), descriptions,
11954 and input types for run-time discovery by calling applications. A
11955 subset of these commands are implicitly classed as "executable"
11956 depending on their input type, and only these can be invoked through
11957 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
11958 can be based on user input, config files, etc). The distinction is
11959 that "executable" commands cannot return anything other than a boolean
11960 result and can only support numeric or string input, whereas some
11961 discoverable commands may only be for direct use through
11962 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
11963 pointers, or other custom uses. The "executable" commands are to
11964 support parameterisations of ENGINE behaviour that can be
11965 unambiguously defined by ENGINEs and used consistently across any
11966 OpenSSL-based application. Commands have been added to all the
11967 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
11968 control over shared-library paths without source code alterations.
11972 * Changed all ENGINE implementations to dynamically allocate their
11973 ENGINEs rather than declaring them statically. Apart from this being
11974 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
11975 this also allows the implementations to compile without using the
11976 internal engine_int.h header.
11980 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
11981 'const' value. Any code that should be able to modify a RAND_METHOD
11982 should already have non-const pointers to it (ie. they should only
11983 modify their own ones).
11987 * Made a variety of little tweaks to the ENGINE code.
11988 - "atalla" and "ubsec" string definitions were moved from header files
11989 to C code. "nuron" string definitions were placed in variables
11990 rather than hard-coded - allowing parameterisation of these values
11991 later on via ctrl() commands.
11992 - Removed unused "#if 0"'d code.
11993 - Fixed engine list iteration code so it uses ENGINE_free() to release
11994 structural references.
11995 - Constified the RAND_METHOD element of ENGINE structures.
11996 - Constified various get/set functions as appropriate and added
11997 missing functions (including a catch-all ENGINE_cpy that duplicates
11998 all ENGINE values onto a new ENGINE except reference counts/state).
11999 - Removed NULL parameter checks in get/set functions. Setting a method
12000 or function to NULL is a way of cancelling out a previously set
12001 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12002 and doesn't justify the extra error symbols and code.
12003 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12004 flags from engine_int.h to engine.h.
12005 - Changed prototypes for ENGINE handler functions (init(), finish(),
12006 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12010 * Implement binary inversion algorithm for BN_mod_inverse in addition
12011 to the algorithm using long division. The binary algorithm can be
12012 used only if the modulus is odd. On 32-bit systems, it is faster
12013 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12014 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12015 up to 450 bits. In 64-bit environments, the binary algorithm
12016 appears to be advantageous for much longer moduli; here we use it
12017 for moduli up to 2048 bits.
12021 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12022 could not support the combine flag in choice fields.
12026 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12027 extensions from a certificate request to the certificate.
12031 * Allow multiple 'certopt' and 'nameopt' options to be separated
12032 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12033 file: this allows the display of the certificate about to be
12034 signed to be customised, to allow certain fields to be included
12035 or excluded and extension details. The old system didn't display
12036 multicharacter strings properly, omitted fields not in the policy
12037 and couldn't display additional details such as extensions.
12041 * Function EC_POINTs_mul for multiple scalar multiplication
12042 of an arbitrary number of elliptic curve points
12043 \sum scalars[i]*points[i],
12044 optionally including the generator defined for the EC_GROUP:
12045 scalar*generator + \sum scalars[i]*points[i].
12047 EC_POINT_mul is a simple wrapper function for the typical case
12048 that the point list has just one item (besides the optional
12053 * First EC_METHODs for curves over GF(p):
12055 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12056 operations and provides various method functions that can also
12057 operate with faster implementations of modular arithmetic.
12059 EC_GFp_mont_method() reuses most functions that are part of
12060 EC_GFp_simple_method, but uses Montgomery arithmetic.
12062 *Bodo Moeller; point addition and point doubling
12063 implementation directly derived from source code provided by
12064 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12066 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12067 crypto/ec/ec_lib.c):
12069 Curves are EC_GROUP objects (with an optional group generator)
12070 based on EC_METHODs that are built into the library.
12072 Points are EC_POINT objects based on EC_GROUP objects.
12074 Most of the framework would be able to handle curves over arbitrary
12075 finite fields, but as there are no obvious types for fields other
12076 than GF(p), some functions are limited to that for now.
12080 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12081 that the file contains a complete HTTP response.
12085 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12086 change the def and num file printf format specifier from "%-40sXXX"
12087 to "%-39s XXX". The latter will always guarantee a space after the
12088 field while the former will cause them to run together if the field
12089 is 40 of more characters long.
12093 * Constify the cipher and digest 'method' functions and structures
12094 and modify related functions to take constant EVP_MD and EVP_CIPHER
12099 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12100 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12104 * Modify EVP_Digest*() routines so they now return values. Although the
12105 internal software routines can never fail additional hardware versions
12110 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12112 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12113 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12116 ERR_R_NESTED_ASN1_ERROR
12118 ERR_R_MISSING_ASN1_EOS
12119 were 4 .. 9, conflicting with
12120 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12122 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12123 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12125 Add new error code 'ERR_R_INTERNAL_ERROR'.
12129 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12134 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12135 sets the subject name for a new request or supersedes the
12136 subject name in a given request. Formats that can be parsed are
12137 'CN=Some Name, OU=myOU, C=IT'
12139 'CN=Some Name/OU=myOU/C=IT'.
12141 Add options '-batch' and '-verbose' to 'openssl req'.
12143 *Massimiliano Pala <madwolf@hackmasters.net>*
12145 * Introduce the possibility to access global variables through
12146 functions on platform were that's the best way to handle exporting
12147 global variables in shared libraries. To enable this functionality,
12148 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12149 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12150 is normally done by Configure or something similar).
12152 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12153 in the source file (foo.c) like this:
12155 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12156 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12158 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12159 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12161 OPENSSL_DECLARE_GLOBAL(int,foo);
12162 #define foo OPENSSL_GLOBAL_REF(foo)
12163 OPENSSL_DECLARE_GLOBAL(double,bar);
12164 #define bar OPENSSL_GLOBAL_REF(bar)
12166 The #defines are very important, and therefore so is including the
12167 header file everywhere where the defined globals are used.
12169 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12170 of ASN.1 items, but that structure is a bit different.
12172 The largest change is in util/mkdef.pl which has been enhanced with
12173 better and easier to understand logic to choose which symbols should
12174 go into the Windows .def files as well as a number of fixes and code
12175 cleanup (among others, algorithm keywords are now sorted
12176 lexicographically to avoid constant rewrites).
12180 * In BN_div() keep a copy of the sign of 'num' before writing the
12181 result to 'rm' because if rm==num the value will be overwritten
12182 and produce the wrong result if 'num' is negative: this caused
12183 problems with BN_mod() and BN_nnmod().
12187 * Function OCSP_request_verify(). This checks the signature on an
12188 OCSP request and verifies the signer certificate. The signer
12189 certificate is just checked for a generic purpose and OCSP request
12194 * Add OCSP_check_validity() function to check the validity of OCSP
12195 responses. OCSP responses are prepared in real time and may only
12196 be a few seconds old. Simply checking that the current time lies
12197 between thisUpdate and nextUpdate max reject otherwise valid responses
12198 caused by either OCSP responder or client clock inaccuracy. Instead
12199 we allow thisUpdate and nextUpdate to fall within a certain period of
12200 the current time. The age of the response can also optionally be
12201 checked. Two new options -validity_period and -status_age added to
12206 * If signature or public key algorithm is unrecognized print out its
12207 OID rather that just UNKNOWN.
12211 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12212 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12213 ID to be generated from the issuer certificate alone which can then be
12214 passed to OCSP_id_issuer_cmp().
12218 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12219 ASN1 modules to export functions returning ASN1_ITEM pointers
12220 instead of the ASN1_ITEM structures themselves. This adds several
12221 new macros which allow the underlying ASN1 function/structure to
12222 be accessed transparently. As a result code should not use ASN1_ITEM
12223 references directly (such as &X509_it) but instead use the relevant
12224 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12225 use of the new ASN1 code on platforms where exporting structures
12226 is problematical (for example in shared libraries) but exporting
12227 functions returning pointers to structures is not.
12231 * Add support for overriding the generation of SSL/TLS session IDs.
12232 These callbacks can be registered either in an SSL_CTX or per SSL.
12233 The purpose of this is to allow applications to control, if they wish,
12234 the arbitrary values chosen for use as session IDs, particularly as it
12235 can be useful for session caching in multiple-server environments. A
12236 command-line switch for testing this (and any client code that wishes
12237 to use such a feature) has been added to "s_server".
12239 *Geoff Thorpe, Lutz Jaenicke*
12241 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12242 of the form '#if defined(...) || defined(...) || ...' and
12243 '#if !defined(...) && !defined(...) && ...'. This also avoids
12244 the growing number of special cases it was previously handling.
12248 * Make all configuration macros available for application by making
12249 sure they are available in opensslconf.h, by giving them names starting
12250 with "OPENSSL_" to avoid conflicts with other packages and by making
12251 sure e_os2.h will cover all platform-specific cases together with
12253 Additionally, it is now possible to define configuration/platform-
12254 specific names (called "system identities"). In the C code, these
12255 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
12256 macro with the name beginning with "OPENSSL_SYS_", which is determined
12257 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
12262 * New option -set_serial to 'req' and 'x509' this allows the serial
12263 number to use to be specified on the command line. Previously self
12264 signed certificates were hard coded with serial number 0 and the
12265 CA options of 'x509' had to use a serial number in a file which was
12270 * New options to 'ca' utility to support V2 CRL entry extensions.
12271 Currently CRL reason, invalidity date and hold instruction are
12272 supported. Add new CRL extensions to V3 code and some new objects.
12276 * New function EVP_CIPHER_CTX_set_padding() this is used to
12277 disable standard block padding (aka PKCS#5 padding) in the EVP
12278 API, which was previously mandatory. This means that the data is
12279 not padded in any way and so the total length much be a multiple
12280 of the block size, otherwise an error occurs.
12284 * Initial (incomplete) OCSP SSL support.
12288 * New function OCSP_parse_url(). This splits up a URL into its host,
12289 port and path components: primarily to parse OCSP URLs. New -url
12290 option to ocsp utility.
12294 * New nonce behavior. The return value of OCSP_check_nonce() now
12295 reflects the various checks performed. Applications can decide
12296 whether to tolerate certain situations such as an absent nonce
12297 in a response when one was present in a request: the ocsp application
12298 just prints out a warning. New function OCSP_add1_basic_nonce()
12299 this is to allow responders to include a nonce in a response even if
12300 the request is nonce-less.
12304 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12305 skipped when using openssl x509 multiple times on a single input file,
12306 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
12310 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12311 set string type: to handle setting ASN1_TIME structures. Fix ca
12312 utility to correctly initialize revocation date of CRLs.
12316 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12317 the clients preferred ciphersuites and rather use its own preferences.
12318 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12319 Internet Explorer by ensuring unchanged hash method during stepup.
12320 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12324 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12325 to aes and add a new 'exist' option to print out symbols that don't
12330 * Additional options to ocsp utility to allow flags to be set and
12331 additional certificates supplied.
12335 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12336 OCSP client a number of certificate to only verify the response
12341 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12342 handle the new API. Currently only ECB, CBC modes supported. Add new
12345 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12346 Encryption Standard (AES) Ciphersuites for Transport Layer
12347 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12348 not enabled by default and were not part of the "ALL" ciphersuite
12349 alias because they were not yet official; they could be
12350 explicitly requested by specifying the "AESdraft" ciphersuite
12351 group alias. In the final release of OpenSSL 0.9.7, the group
12352 alias is called "AES" and is part of "ALL".)
12354 *Ben Laurie, Steve Henson, Bodo Moeller*
12356 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12357 request to response.
12361 * Functions for OCSP responders. OCSP_request_onereq_count(),
12362 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12363 extract information from a certificate request. OCSP_response_create()
12364 creates a response and optionally adds a basic response structure.
12365 OCSP_basic_add1_status() adds a complete single response to a basic
12366 response and returns the OCSP_SINGLERESP structure just added (to allow
12367 extensions to be included for example). OCSP_basic_add1_cert() adds a
12368 certificate to a basic response and OCSP_basic_sign() signs a basic
12369 response with various flags. New helper functions ASN1_TIME_check()
12370 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12371 (converts ASN1_TIME to GeneralizedTime).
12375 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12376 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12377 structure from a certificate. X509_pubkey_digest() digests the public_key
12378 contents: this is used in various key identifiers.
12382 * Make sk_sort() tolerate a NULL argument.
12384 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12386 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12387 passed by the function are trusted implicitly. If any of them signed the
12388 response then it is assumed to be valid and is not verified.
12392 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12393 to data. This was previously part of the PKCS7 ASN1 code. This
12394 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12395 *Steve Henson, reported by Kenneth R. Robinette
12396 <support@securenetterm.com>*
12398 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12399 routines: without these tracing memory leaks is very painful.
12400 Fix leaks in PKCS12 and PKCS7 routines.
12404 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12405 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12406 effectively meant GeneralizedTime would never be used. Now it
12407 is initialised to -1 but X509_time_adj() now has to check the value
12408 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12409 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12410 *Steve Henson, reported by Kenneth R. Robinette
12411 <support@securenetterm.com>*
12413 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12414 result in a zero length in the ASN1_INTEGER structure which was
12415 not consistent with the structure when d2i_ASN1_INTEGER() was used
12416 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12417 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12418 where it did not print out a minus for negative ASN1_INTEGER.
12422 * Add summary printout to ocsp utility. The various functions which
12423 convert status values to strings have been renamed to:
12424 OCSP_response_status_str(), OCSP_cert_status_str() and
12425 OCSP_crl_reason_str() and are no longer static. New options
12426 to verify nonce values and to disable verification. OCSP response
12427 printout format cleaned up.
12431 * Add additional OCSP certificate checks. These are those specified
12432 in RFC2560. This consists of two separate checks: the CA of the
12433 certificate being checked must either be the OCSP signer certificate
12434 or the issuer of the OCSP signer certificate. In the latter case the
12435 OCSP signer certificate must contain the OCSP signing extended key
12436 usage. This check is performed by attempting to match the OCSP
12437 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12438 in the OCSP_CERTID structures of the response.
12442 * Initial OCSP certificate verification added to OCSP_basic_verify()
12443 and related routines. This uses the standard OpenSSL certificate
12444 verify routines to perform initial checks (just CA validity) and
12445 to obtain the certificate chain. Then additional checks will be
12446 performed on the chain. Currently the root CA is checked to see
12447 if it is explicitly trusted for OCSP signing. This is used to set
12448 a root CA as a global signing root: that is any certificate that
12449 chains to that CA is an acceptable OCSP signing certificate.
12453 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12454 extensions from a separate configuration file.
12455 As when reading extensions from the main configuration file,
12456 the '-extensions ...' option may be used for specifying the
12459 *Massimiliano Pala <madwolf@comune.modena.it>*
12461 * New OCSP utility. Allows OCSP requests to be generated or
12462 read. The request can be sent to a responder and the output
12463 parsed, outputted or printed in text form. Not complete yet:
12464 still needs to check the OCSP response validity.
12468 * New subcommands for 'openssl ca':
12469 'openssl ca -status <serial>' prints the status of the cert with
12470 the given serial number (according to the index file).
12471 'openssl ca -updatedb' updates the expiry status of certificates
12474 *Massimiliano Pala <madwolf@comune.modena.it>*
12476 * New '-newreq-nodes' command option to CA.pl. This is like
12477 '-newreq', but calls 'openssl req' with the '-nodes' option
12478 so that the resulting key is not encrypted.
12480 *Damien Miller <djm@mindrot.org>*
12482 * New configuration for the GNU Hurd.
12484 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12486 * Initial code to implement OCSP basic response verify. This
12487 is currently incomplete. Currently just finds the signer's
12488 certificate and verifies the signature on the response.
12492 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12493 value of OPENSSLDIR. This is available via the new '-d' option
12494 to 'openssl version', and is also included in 'openssl version -a'.
12498 * Allowing defining memory allocation callbacks that will be given
12499 file name and line number information in additional arguments
12500 (a const char* and an int). The basic functionality remains, as
12501 well as the original possibility to just replace malloc(),
12502 realloc() and free() by functions that do not know about these
12503 additional arguments. To register and find out the current
12504 settings for extended allocation functions, the following
12505 functions are provided:
12507 CRYPTO_set_mem_ex_functions
12508 CRYPTO_set_locked_mem_ex_functions
12509 CRYPTO_get_mem_ex_functions
12510 CRYPTO_get_locked_mem_ex_functions
12512 These work the same way as CRYPTO_set_mem_functions and friends.
12513 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
12514 extended allocation function is enabled.
12515 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
12516 a conventional allocation function is enabled.
12518 *Richard Levitte, Bodo Moeller*
12520 * Finish off removing the remaining LHASH function pointer casts.
12521 There should no longer be any prototype-casting required when using
12522 the LHASH abstraction, and any casts that remain are "bugs". See
12523 the callback types and macros at the head of lhash.h for details
12524 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12528 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12529 If /dev/[u]random devices are not available or do not return enough
12530 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12532 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12533 /etc/entropy will be queried once each in this sequence, querying stops
12534 when enough entropy was collected without querying more sockets.
12538 * Change the Unix RAND_poll() variant to be able to poll several
12539 random devices, as specified by DEVRANDOM, until a sufficient amount
12540 of data has been collected. We spend at most 10 ms on each file
12541 (select timeout) and read in non-blocking mode. DEVRANDOM now
12542 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12543 (previously it was just the string "/dev/urandom"), so on typical
12544 platforms the 10 ms delay will never occur.
12545 Also separate out the Unix variant to its own file, rand_unix.c.
12546 For VMS, there's a currently-empty rand_vms.c.
12550 * Move OCSP client related routines to ocsp_cl.c. These
12551 provide utility functions which an application needing
12552 to issue a request to an OCSP responder and analyse the
12553 response will typically need: as opposed to those which an
12554 OCSP responder itself would need which will be added later.
12556 OCSP_request_sign() signs an OCSP request with an API similar
12557 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12558 response. OCSP_response_get1_basic() extracts basic response
12559 from response. OCSP_resp_find_status(): finds and extracts status
12560 information from an OCSP_CERTID structure (which will be created
12561 when the request structure is built). These are built from lower
12562 level functions which work on OCSP_SINGLERESP structures but
12563 won't normally be used unless the application wishes to examine
12564 extensions in the OCSP response for example.
12566 Replace nonce routines with a pair of functions.
12567 OCSP_request_add1_nonce() adds a nonce value and optionally
12568 generates a random value. OCSP_check_nonce() checks the
12569 validity of the nonce in an OCSP response.
12573 * Change function OCSP_request_add() to OCSP_request_add0_id().
12574 This doesn't copy the supplied OCSP_CERTID and avoids the
12575 need to free up the newly created id. Change return type
12576 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12577 This can then be used to add extensions to the request.
12578 Deleted OCSP_request_new(), since most of its functionality
12579 is now in OCSP_REQUEST_new() (and the case insensitive name
12580 clash) apart from the ability to set the request name which
12581 will be added elsewhere.
12585 * Update OCSP API. Remove obsolete extensions argument from
12586 various functions. Extensions are now handled using the new
12587 OCSP extension code. New simple OCSP HTTP function which
12588 can be used to send requests and parse the response.
12592 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12593 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12594 uses the special reorder version of SET OF to sort the attributes
12595 and reorder them to match the encoded order. This resolves a long
12596 standing problem: a verify on a PKCS7 structure just after signing
12597 it used to fail because the attribute order did not match the
12598 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12599 it uses the received order. This is necessary to tolerate some broken
12600 software that does not order SET OF. This is handled by encoding
12601 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12602 to produce the required SET OF.
12606 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12607 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12608 files to get correct declarations of the ASN.1 item variables.
12612 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12613 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12614 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12615 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12616 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12617 ASN1_ITEM and no wrapper functions.
12621 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12622 replace the old function pointer based I/O routines. Change most of
12623 the *_d2i_bio() and *_d2i_fp() functions to use these.
12627 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12628 lines, recognize more "algorithms" that can be deselected, and make
12629 it complain about algorithm deselection that isn't recognised.
12633 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12634 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12635 to use new functions. Add NO_ASN1_OLD which can be set to remove
12636 some old style ASN1 functions: this can be used to determine if old
12637 code will still work when these eventually go away.
12641 * New extension functions for OCSP structures, these follow the
12642 same conventions as certificates and CRLs.
12646 * New function X509V3_add1_i2d(). This automatically encodes and
12647 adds an extension. Its behaviour can be customised with various
12648 flags to append, replace or delete. Various wrappers added for
12649 certificates and CRLs.
12653 * Fix to avoid calling the underlying ASN1 print routine when
12654 an extension cannot be parsed. Correct a typo in the
12655 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12659 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12660 entries for variables.
12664 * Add functionality to apps/openssl.c for detecting locking
12665 problems: As the program is single-threaded, all we have
12666 to do is register a locking callback using an array for
12667 storing which locks are currently held by the program.
12671 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12672 SSL_get_ex_data_X509_STORE_idx(), which is used in
12673 ssl_verify_cert_chain() and thus can be called at any time
12674 during TLS/SSL handshakes so that thread-safety is essential.
12675 Unfortunately, the ex_data design is not at all suited
12676 for multi-threaded use, so it probably should be abolished.
12680 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12682 *Broadcom, tweaked and integrated by Geoff Thorpe*
12684 * Move common extension printing code to new function
12685 X509V3_print_extensions(). Reorganise OCSP print routines and
12686 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12690 * New function X509_signature_print() to remove duplication in some
12695 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12696 set (this was treated exactly the same as SET OF previously). This
12697 is used to reorder the STACK representing the structure to match the
12698 encoding. This will be used to get round a problem where a PKCS7
12699 structure which was signed could not be verified because the STACK
12700 order did not reflect the encoded order.
12704 * Reimplement the OCSP ASN1 module using the new code.
12708 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12709 for its ASN1 operations. The old style function pointers still exist
12710 for now but they will eventually go away.
12714 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12715 completely replaces the old ASN1 functionality with a table driven
12716 encoder and decoder which interprets an ASN1_ITEM structure describing
12717 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12718 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12719 has also been converted to the new form.
12723 * Change BN_mod_exp_recp so that negative moduli are tolerated
12724 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12725 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12726 for negative moduli.
12730 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12731 of not touching the result's sign bit.
12735 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12740 * Changed the LHASH code to use prototypes for callbacks, and created
12741 macros to declare and implement thin (optionally static) functions
12742 that provide type-safety and avoid function pointer casting for the
12743 type-specific callbacks.
12747 * Added Kerberos Cipher Suites to be used with TLS, as written in
12749 *Veers Staats <staatsvr@asc.hpc.mil>,
12750 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12752 * Reformat the FAQ so the different questions and answers can be divided
12753 in sections depending on the subject.
12757 * Have the zlib compression code load ZLIB.DLL dynamically under
12762 * New function BN_mod_sqrt for computing square roots modulo a prime
12763 (using the probabilistic Tonelli-Shanks algorithm unless
12764 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12765 be handled deterministically).
12767 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12769 * Make BN_mod_inverse faster by explicitly handling small quotients
12770 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12771 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12775 * New function BN_kronecker.
12779 * Fix BN_gcd so that it works on negative inputs; the result is
12780 positive unless both parameters are zero.
12781 Previously something reasonably close to an infinite loop was
12782 possible because numbers could be growing instead of shrinking
12783 in the implementation of Euclid's algorithm.
12787 * Fix BN_is_word() and BN_is_one() macros to take into account the
12788 sign of the number in question.
12790 Fix BN_is_word(a,w) to work correctly for w == 0.
12792 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12793 because its test if the absolute value of 'a' equals 'w'.
12794 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12795 it exists mostly for use in the implementations of BN_is_zero(),
12796 BN_is_one(), and BN_is_word().
12800 * New function BN_swap.
12804 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12805 the exponentiation functions are more likely to produce reasonable
12806 results on negative inputs.
12810 * Change BN_mod_mul so that the result is always non-negative.
12811 Previously, it could be negative if one of the factors was negative;
12812 I don't think anyone really wanted that behaviour.
12816 * Move BN_mod_... functions into new file crypto/bn/bn_mod.c
12817 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
12818 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
12819 and add new functions:
12828 BN_mod_lshift1_quick
12830 BN_mod_lshift_quick
12832 These functions always generate non-negative results.
12834 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
12835 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
12837 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
12838 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
12839 be reduced modulo m.
12841 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12844 The following entry accidentally appeared in the CHANGES file
12845 distributed with OpenSSL 0.9.7. The modifications described in
12846 it do *not* apply to OpenSSL 0.9.7.
12848 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12849 was actually never needed) and in BN_mul(). The removal in BN_mul()
12850 required a small change in bn_mul_part_recursive() and the addition
12851 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12852 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12853 bn_sub_words() and bn_add_words() except they take arrays with
12859 * In 'openssl passwd', verify passwords read from the terminal
12860 unless the '-salt' option is used (which usually means that
12861 verification would just waste user's time since the resulting
12862 hash is going to be compared with some given password hash)
12863 or the new '-noverify' option is used.
12865 This is an incompatible change, but it does not affect
12866 non-interactive use of 'openssl passwd' (passwords on the command
12867 line, '-stdin' option, '-in ...' option) and thus should not
12868 cause any problems.
12872 * Remove all references to RSAref, since there's no more need for it.
12876 * Make DSO load along a path given through an environment variable
12877 (SHLIB_PATH) with shl_load().
12881 * Constify the ENGINE code as a result of BIGNUM constification.
12882 Also constify the RSA code and most things related to it. In a
12883 few places, most notable in the depth of the ASN.1 code, ugly
12884 casts back to non-const were required (to be solved at a later
12889 * Make it so the openssl application has all engines loaded by default.
12893 * Constify the BIGNUM routines a little more.
12897 * Add the following functions:
12899 ENGINE_load_cswift()
12901 ENGINE_load_atalla()
12902 ENGINE_load_nuron()
12903 ENGINE_load_builtin_engines()
12905 That way, an application can itself choose if external engines that
12906 are built-in in OpenSSL shall ever be used or not. The benefit is
12907 that applications won't have to be linked with libdl or other dso
12908 libraries unless it's really needed.
12910 Changed 'openssl engine' to load all engines on demand.
12911 Changed the engine header files to avoid the duplication of some
12912 declarations (they differed!).
12916 * 'openssl engine' can now list capabilities.
12920 * Better error reporting in 'openssl engine'.
12924 * Never call load_dh_param(NULL) in s_server.
12928 * Add engine application. It can currently list engines by name and
12929 identity, and test if they are actually available.
12933 * Improve RPM specification file by forcing symbolic linking and making
12934 sure the installed documentation is also owned by root.root.
12936 *Damien Miller <djm@mindrot.org>*
12938 * Give the OpenSSL applications more possibilities to make use of
12939 keys (public as well as private) handled by engines.
12943 * Add OCSP code that comes from CertCo.
12947 * Add VMS support for the Rijndael code.
12951 * Added untested support for Nuron crypto accelerator.
12955 * Add support for external cryptographic devices. This code was
12956 previously distributed separately as the "engine" branch.
12958 *Geoff Thorpe, Richard Levitte*
12960 * Rework the filename-translation in the DSO code. It is now possible to
12961 have far greater control over how a "name" is turned into a filename
12962 depending on the operating environment and any oddities about the
12963 different shared library filenames on each system.
12967 * Support threads on FreeBSD-elf in Configure.
12971 * Fix for SHA1 assembly problem with MASM: it produces
12972 warnings about corrupt line number information when assembling
12973 with debugging information. This is caused by the overlapping
12976 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
12979 NCONF_get_number() has no error checking at all. As a replacement,
12980 NCONF_get_number_e() is defined (_e for "error checking") and is
12981 promoted strongly. The old NCONF_get_number is kept around for
12982 binary backward compatibility.
12983 Make it possible for methods to load from something other than a BIO,
12984 by providing a function pointer that is given a name instead of a BIO.
12985 For example, this could be used to load configuration data from an
12990 * Fix for non blocking accept BIOs. Added new I/O special reason
12991 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
12992 with non blocking I/O was not possible because no retry code was
12993 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
12998 * Added the beginnings of Rijndael support.
13002 * Fix for bug in DirectoryString mask setting. Add support for
13003 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13004 to allow certificate printing to more controllable, additional
13005 'certopt' option to 'x509' to allow new printing options to be
13010 * Clean old EAY MD5 hack from e_os.h.
13014 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004] ###
13016 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13017 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13019 *Joe Orton, Steve Henson*
13021 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003] ###
13023 * Fix additional bug revealed by the NISCC test suite:
13025 Stop bug triggering large recursion when presented with
13026 certain ASN.1 tags [CVE-2003-0851][]
13030 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003] ###
13032 * Fix various bugs revealed by running the NISCC test suite:
13034 Stop out of bounds reads in the ASN1 code when presented with
13035 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13037 If verify callback ignores invalid public key errors don't try to check
13038 certificate signature with the NULL public key.
13043 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13044 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13049 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13050 extra data after the compression methods not only for TLS 1.0
13051 but also for SSL 3.0 (as required by the specification).
13053 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13055 * Change X509_certificate_type() to mark the key as exported/exportable
13056 when it's 512 *bits* long, not 512 bytes.
13060 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003] ###
13062 * Countermeasure against the Klima-Pokorny-Rosa extension of
13063 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13064 a protocol version number mismatch like a decryption error
13065 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13069 * Turn on RSA blinding by default in the default implementation
13070 to avoid a timing attack. Applications that don't want it can call
13071 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13072 They would be ill-advised to do so in most cases.
13074 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13076 * Change RSA blinding code so that it works when the PRNG is not
13077 seeded (in this case, the secret RSA exponent is abused as
13078 an unpredictable seed -- if it is not unpredictable, there
13079 is no point in blinding anyway). Make RSA blinding thread-safe
13080 by remembering the creator's thread ID in rsa->blinding and
13081 having all other threads use local one-time blinding factors
13082 (this requires more computation than sharing rsa->blinding, but
13083 avoids excessive locking; and if an RSA object is not shared
13084 between threads, blinding will still be very fast).
13088 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003] ###
13090 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13091 via timing by performing a MAC computation even if incorrect
13092 block cipher padding has been found. This is a countermeasure
13093 against active attacks where the attacker has to distinguish
13094 between bad padding and a MAC verification error. [CVE-2003-0078][]
13096 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13097 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13098 Martin Vuagnoux (EPFL, Ilion)*
13100 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002] ###
13102 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13103 memory from its contents. This is done with a counter that will
13104 place alternating values in each byte. This can be used to solve
13105 two issues: 1) the removal of calls to memset() by highly optimizing
13106 compilers, and 2) cleansing with other values than 0, since those can
13107 be read through on certain media, for example a swap space on disk.
13111 * Bugfix: client side session caching did not work with external caching,
13112 because the session->cipher setting was not restored when reloading
13113 from the external cache. This problem was masked, when
13114 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13115 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13119 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13120 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13122 *Zeev Lieber <zeev-l@yahoo.com>*
13124 * Undo an undocumented change introduced in 0.9.6e which caused
13125 repeated calls to OpenSSL_add_all_ciphers() and
13126 OpenSSL_add_all_digests() to be ignored, even after calling
13131 * Change the default configuration reader to deal with last line not
13132 being properly terminated.
13136 * Change X509_NAME_cmp() so it applies the special rules on handling
13137 DN values that are of type PrintableString, as well as RDNs of type
13138 emailAddress where the value has the type ia5String.
13140 *stefank@valicert.com via Richard Levitte*
13142 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13143 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13144 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13145 the bitwise-OR of the two for use by the majority of applications
13146 wanting this behaviour, and update the docs. The documented
13147 behaviour and actual behaviour were inconsistent and had been
13148 changing anyway, so this is more a bug-fix than a behavioural
13151 *Geoff Thorpe, diagnosed by Nadav Har'El*
13153 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13154 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13158 * Fix initialization code race conditions in
13159 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13160 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13161 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13162 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13163 ssl2_get_cipher_by_char(),
13164 ssl3_get_cipher_by_char().
13166 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13168 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13169 the cached sessions are flushed, as the remove_cb() might use ex_data
13170 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13171 (see [openssl.org #212]).
13173 *Geoff Thorpe, Lutz Jaenicke*
13175 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13176 length, instead of the encoding length to d2i_ASN1_OBJECT.
13180 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002] ###
13182 * [In 0.9.6g-engine release:]
13183 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
13185 *Lynn Gazis <lgazis@rainbow.com>*
13187 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002] ###
13189 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13190 and get fix the header length calculation.
13191 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13192 Alon Kantor <alonk@checkpoint.com> (and others),
13195 * Use proper error handling instead of 'assertions' in buffer
13196 overflow checks added in 0.9.6e. This prevents DoS (the
13197 assertions could call abort()).
13199 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13201 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002] ###
13203 * Add various sanity checks to asn1_get_length() to reject
13204 the ASN1 length bytes if they exceed sizeof(long), will appear
13205 negative or the content length exceeds the length of the
13208 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13210 * Fix cipher selection routines: ciphers without encryption had no flags
13211 for the cipher strength set and where therefore not handled correctly
13212 by the selection routines (PR #130).
13216 * Fix EVP_dsa_sha macro.
13221 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13222 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13223 that was added in OpenSSL 0.9.6d.
13225 As the countermeasure turned out to be incompatible with some
13226 broken SSL implementations, the new option is part of SSL_OP_ALL.
13227 SSL_OP_ALL is usually employed when compatibility with weird SSL
13228 implementations is desired (e.g. '-bugs' option to 's_client' and
13229 's_server'), so the new option is automatically set in many
13234 * Changes in security patch:
13236 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13237 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13238 Air Force Materiel Command, USAF, under agreement number
13241 * Add various sanity checks to asn1_get_length() to reject
13242 the ASN1 length bytes if they exceed sizeof(long), will appear
13243 negative or the content length exceeds the length of the
13244 supplied buffer. [CVE-2002-0659][]
13246 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13248 * Assertions for various potential buffer overflows, not known to
13249 happen in practice.
13251 *Ben Laurie (CHATS)*
13253 * Various temporary buffers to hold ASCII versions of integers were
13254 too small for 64 bit platforms. [CVE-2002-0655][]
13255 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
13257 * Remote buffer overflow in SSL3 protocol - an attacker could
13258 supply an oversized session ID to a client. [CVE-2002-0656][]
13260 *Ben Laurie (CHATS)*
13262 * Remote buffer overflow in SSL2 protocol - an attacker could
13263 supply an oversized client master key. [CVE-2002-0656][]
13265 *Ben Laurie (CHATS)*
13267 ### Changes between 0.9.6c and 0.9.6d [9 May 2002] ###
13269 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13270 encoded as NULL) with id-dsa-with-sha1.
13272 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13274 * Check various X509_...() return values in apps/req.c.
13276 *Nils Larsch <nla@trustcenter.de>*
13278 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13279 an end-of-file condition would erroneously be flagged, when the CRLF
13280 was just at the end of a processed block. The bug was discovered when
13281 processing data through a buffering memory BIO handing the data to a
13282 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13283 <ptsekov@syntrex.com> and Nedelcho Stanev.
13287 * Implement a countermeasure against a vulnerability recently found
13288 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13289 before application data chunks to avoid the use of known IVs
13290 with data potentially chosen by the attacker.
13294 * Fix length checks in ssl3_get_client_hello().
13298 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13299 to prevent ssl3_read_internal() from incorrectly assuming that
13300 ssl3_read_bytes() found application data while handshake
13301 processing was enabled when in fact s->s3->in_read_app_data was
13302 merely automatically cleared during the initial handshake.
13304 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13306 * Fix object definitions for Private and Enterprise: they were not
13307 recognized in their shortname (=lowercase) representation. Extend
13308 obj_dat.pl to issue an error when using undefined keywords instead
13309 of silently ignoring the problem (Svenning Sorensen
13310 <sss@sss.dnsalias.net>).
13314 * Fix DH_generate_parameters() so that it works for 'non-standard'
13315 generators, i.e. generators other than 2 and 5. (Previously, the
13316 code did not properly initialise the 'add' and 'rem' values to
13317 BN_generate_prime().)
13319 In the new general case, we do not insist that 'generator' is
13320 actually a primitive root: This requirement is rather pointless;
13321 a generator of the order-q subgroup is just as good, if not
13326 * Map new X509 verification errors to alerts. Discovered and submitted by
13327 Tom Wu <tom@arcot.com>.
13331 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13332 returning non-zero before the data has been completely received
13333 when using non-blocking I/O.
13335 *Bodo Moeller; problem pointed out by John Hughes*
13337 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13339 *Ben Laurie, Lutz Jaenicke*
13341 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13342 Yoram Zahavi <YoramZ@gilian.com>).
13346 * Add information about CygWin 1.3 and on, and preserve proper
13347 configuration for the versions before that.
13349 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13351 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13352 check whether we deal with a copy of a session and do not delete from
13353 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13354 <izhar@checkpoint.com>.
13358 * Do not store session data into the internal session cache, if it
13359 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13360 flag is set). Proposed by Aslam <aslam@funk.com>.
13364 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13369 * [In 0.9.6d-engine release:]
13370 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13372 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13374 * Add the configuration target linux-s390x.
13376 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13378 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13379 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13380 variable as an indication that a ClientHello message has been
13381 received. As the flag value will be lost between multiple
13382 invocations of ssl3_accept when using non-blocking I/O, the
13383 function may not be aware that a handshake has actually taken
13384 place, thus preventing a new session from being added to the
13387 To avoid this problem, we now set s->new_session to 2 instead of
13388 using a local variable.
13390 *Lutz Jaenicke, Bodo Moeller*
13392 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13393 if the SSL_R_LENGTH_MISMATCH error is detected.
13395 *Geoff Thorpe, Bodo Moeller*
13397 * New 'shared_ldflag' column in Configure platform table.
13401 * Fix EVP_CIPHER_mode macro.
13403 *"Dan S. Camper" <dan@bti.net>*
13405 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13406 type, we must throw them away by setting rr->length to 0.
13408 *D P Chang <dpc@qualys.com>*
13410 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001] ###
13412 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13413 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13414 worked incorrectly for those cases where range = 10..._2 and
13415 3*range is two bits longer than range.)
13419 * Only add signing time to PKCS7 structures if it is not already
13424 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13425 OBJ_ld_ce should be OBJ_id_ce.
13426 Also some ip-pda OIDs in crypto/objects/objects.txt were
13427 incorrect (cf. RFC 3039).
13429 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13431 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13432 returns early because it has nothing to do.
13434 *Andy Schneider <andy.schneider@bjss.co.uk>*
13436 * [In 0.9.6c-engine release:]
13437 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13439 *Andy Schneider <andy.schneider@bjss.co.uk>*
13441 * [In 0.9.6c-engine release:]
13442 Add support for Cryptographic Appliance's keyserver technology.
13443 (Use engine 'keyclient')
13445 *Cryptographic Appliances and Geoff Thorpe*
13447 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13448 is called via tools/c89.sh because arguments have to be
13449 rearranged (all '-L' options must appear before the first object
13452 *Richard Shapiro <rshapiro@abinitio.com>*
13454 * [In 0.9.6c-engine release:]
13455 Add support for Broadcom crypto accelerator cards, backported
13458 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13460 * [In 0.9.6c-engine release:]
13461 Add support for SureWare crypto accelerator cards from
13462 Baltimore Technologies. (Use engine 'sureware')
13464 *Baltimore Technologies and Mark Cox*
13466 * [In 0.9.6c-engine release:]
13467 Add support for crypto accelerator cards from Accelerated
13468 Encryption Processing, www.aep.ie. (Use engine 'aep')
13470 *AEP Inc. and Mark Cox*
13472 * Add a configuration entry for gcc on UnixWare.
13474 *Gary Benson <gbenson@redhat.com>*
13476 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13477 messages are stored in a single piece (fixed-length part and
13478 variable-length part combined) and fix various bugs found on the way.
13482 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13483 instead. BIO_gethostbyname() does not know what timeouts are
13484 appropriate, so entries would stay in cache even when they have
13486 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
13488 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13489 faced with a pathologically small ClientHello fragment that does
13490 not contain client_version: Instead of aborting with an error,
13491 simply choose the highest available protocol version (i.e.,
13492 TLS 1.0 unless it is disabled). In practice, ClientHello
13493 messages are never sent like this, but this change gives us
13494 strictly correct behaviour at least for TLS.
13498 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13499 never resets s->method to s->ctx->method when called from within
13500 one of the SSL handshake functions.
13502 *Bodo Moeller; problem pointed out by Niko Baric*
13504 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13505 (sent using the client's version number) if client_version is
13506 smaller than the protocol version in use. Also change
13507 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13508 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13509 the client will at least see that alert.
13513 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13518 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13519 client receives HelloRequest while in a handshake.
13521 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13523 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13524 should end in 'break', not 'goto end' which circumvents various
13525 cleanups done in state SSL_ST_OK. But session related stuff
13526 must be disabled for SSL_ST_OK in the case that we just sent a
13529 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13530 before just sending a HelloRequest.
13532 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13534 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13535 reveal whether illegal block cipher padding was found or a MAC
13536 verification error occurred. (Neither SSLerr() codes nor alerts
13537 are directly visible to potential attackers, but the information
13538 may leak via logfiles.)
13540 Similar changes are not required for the SSL 2.0 implementation
13541 because the number of padding bytes is sent in clear for SSL 2.0,
13542 and the extra bytes are just ignored. However ssl/s2_pkt.c
13543 failed to verify that the purported number of padding bytes is in
13548 * Add OpenUNIX-8 support including shared libraries
13549 (Boyd Lynn Gerber <gerberb@zenez.com>).
13553 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13554 'wristwatch attack' using huge encoding parameters (cf.
13555 James H. Manger's CRYPTO 2001 paper). Note that the
13556 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13557 encoding parameters and hence was not vulnerable.
13561 * BN_sqr() bug fix.
13563 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13565 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13566 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13567 followed by modular reduction.
13569 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13571 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13572 equivalent based on BN_pseudo_rand() instead of BN_rand().
13576 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13577 This function was broken, as the check for a new client hello message
13578 to handle SGC did not allow these large messages.
13579 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13583 * Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
13587 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13588 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13592 * Rework the configuration and shared library support for Tru64 Unix.
13593 The configuration part makes use of modern compiler features and
13594 still retains old compiler behavior for those that run older versions
13595 of the OS. The shared library support part includes a variant that
13596 uses the RPATH feature, and is available through the special
13597 configuration target "alpha-cc-rpath", which will never be selected
13600 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13602 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13603 with the same message size as in ssl3_get_certificate_request().
13604 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13605 messages might inadvertently be reject as too long.
13607 *Petr Lampa <lampa@fee.vutbr.cz>*
13609 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13613 * Modified SSL library such that the verify_callback that has been set
13614 specifically for an SSL object with SSL_set_verify() is actually being
13615 used. Before the change, a verify_callback set with this function was
13616 ignored and the verify_callback() set in the SSL_CTX at the time of
13617 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13618 to allow the necessary settings.
13622 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13623 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13624 done automatically (in contradiction to the requirements of the C
13625 standard). This made problems when used from OpenSSH.
13629 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13630 dh->length and always used
13632 BN_rand_range(priv_key, dh->p).
13634 BN_rand_range() is not necessary for Diffie-Hellman, and this
13635 specific range makes Diffie-Hellman unnecessarily inefficient if
13636 dh->length (recommended exponent length) is much smaller than the
13637 length of dh->p. We could use BN_rand_range() if the order of
13638 the subgroup was stored in the DH structure, but we only have
13643 BN_rand(priv_key, l, ...)
13645 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13652 RSA_eay_public_encrypt
13653 RSA_eay_private_decrypt
13654 RSA_eay_private_encrypt (signing)
13655 RSA_eay_public_decrypt (signature verification)
13657 (default implementations for RSA_public_encrypt,
13658 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13659 always reject numbers >= n.
13663 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13664 to synchronize access to 'locking_thread'. This is necessary on
13665 systems where access to 'locking_thread' (an 'unsigned long'
13666 variable) is not atomic.
13670 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13671 *before* setting the 'crypto_lock_rand' flag. The previous code had
13672 a race condition if 0 is a valid thread ID.
13674 *Travis Vitek <vitek@roguewave.com>*
13676 * Add support for shared libraries under Irix.
13678 *Albert Chin-A-Young <china@thewrittenword.com>*
13680 * Add configuration option to build on Linux on both big-endian and
13681 little-endian MIPS.
13683 *Ralf Baechle <ralf@uni-koblenz.de>*
13685 * Add the possibility to create shared libraries on HP-UX.
13689 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001] ###
13691 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13692 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13693 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13694 PRNG state recovery was possible based on the output of
13695 one PRNG request appropriately sized to gain knowledge on
13696 'md' followed by enough consecutive 1-byte PRNG requests
13697 to traverse all of 'state'.
13699 1. When updating 'md_local' (the current thread's copy of 'md')
13700 during PRNG output generation, hash all of the previous
13701 'md_local' value, not just the half used for PRNG output.
13703 2. Make the number of bytes from 'state' included into the hash
13704 independent from the number of PRNG bytes requested.
13706 The first measure alone would be sufficient to avoid
13707 Markku-Juhani's attack. (Actually it had never occurred
13708 to me that the half of 'md_local' used for chaining was the
13709 half from which PRNG output bytes were taken -- I had always
13710 assumed that the secret half would be used.) The second
13711 measure makes sure that additional data from 'state' is never
13712 mixed into 'md_local' in small portions; this heuristically
13713 further strengthens the PRNG.
13717 * Fix crypto/bn/asm/mips3.s.
13721 * When only the key is given to "enc", the IV is undefined. Print out
13722 an error message in this case.
13726 * Handle special case when X509_NAME is empty in X509 printing routines.
13730 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13731 positive and less than q.
13735 * Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
13736 used: it isn't thread safe and the add_lock_callback should handle
13739 *Paul Rose <Paul.Rose@bridge.com>*
13741 * Verify that incoming data obeys the block size in
13742 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13748 *Ulf Möller, Bodo Möller*
13750 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13751 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13752 when fixing the server behaviour for backwards-compatible 'client
13753 hello' messages. (Note that the attack is impractical against
13754 SSL 3.0 and TLS 1.0 anyway because length and version checking
13755 means that the probability of guessing a valid ciphertext is
13756 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13759 Before 0.9.5, the countermeasure (hide the error by generating a
13760 random 'decryption result') did not work properly because
13761 ERR_clear_error() was missing, meaning that SSL_get_error() would
13762 detect the supposedly ignored error.
13764 Both problems are now fixed.
13768 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13769 (previously it was 1024).
13773 * Fix for compatibility mode trust settings: ignore trust settings
13774 unless some valid trust or reject settings are present.
13778 * Fix for blowfish EVP: its a variable length cipher.
13782 * Fix various bugs related to DSA S/MIME verification. Handle missing
13783 parameters in DSA public key structures and return an error in the
13784 DSA routines if parameters are absent.
13788 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13789 in the current directory if neither $RANDFILE nor $HOME was set.
13790 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13791 caused some confusion to Windows users who haven't defined $HOME.
13792 Thus RAND_file_name() is changed again: e_os.h can define a
13793 DEFAULT_HOME, which will be used if $HOME is not set.
13794 For Windows, we use "C:"; on other platforms, we still require
13795 environment variables.
13797 * Move 'if (!initialized) RAND_poll()' into regions protected by
13798 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13799 having multiple threads call RAND_poll() concurrently.
13803 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13804 combination of a flag and a thread ID variable.
13805 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13806 flag), *other* threads can enter ssleay_add_bytes without obeying
13807 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13808 that they do not hold after the first thread unsets add_do_not_lock).
13812 * Change bctest again: '-x' expressions are not available in all
13813 versions of 'test'.
13817 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001] ###
13819 * Fix a couple of memory leaks in PKCS7_dataDecode()
13821 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13823 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13824 the default extension for executables, if any. Also, make the perl
13825 scripts that use symlink() to test if it really exists and use "cp"
13826 if it doesn't. All this made OpenSSL compilable and installable in
13831 * Fix for asn1_GetSequence() for indefinite length constructed data.
13832 If SEQUENCE is length is indefinite just set c->slen to the total
13833 amount of data available.
13835 *Steve Henson, reported by shige@FreeBSD.org*
13837 *This change does not apply to 0.9.7.*
13839 * Change bctest to avoid here-documents inside command substitution
13840 (workaround for FreeBSD /bin/sh bug).
13841 For compatibility with Ultrix, avoid shell functions (introduced
13842 in the bctest version that searches along $PATH).
13846 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13847 with des_encrypt() defined on some operating systems, like Solaris
13852 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13853 On the Importance of Eliminating Errors in Cryptographic
13854 Computations, J. Cryptology 14 (2001) 2, 101-119,
13855 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
13859 * MIPS assembler BIGNUM division bug fix.
13863 * Disabled incorrect Alpha assembler code.
13867 * Fix PKCS#7 decode routines so they correctly update the length
13868 after reading an EOC for the EXPLICIT tag.
13872 *This change does not apply to 0.9.7.*
13874 * Fix bug in PKCS#12 key generation routines. This was triggered
13875 if a 3DES key was generated with a 0 initial byte. Include
13876 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13877 (but broken) behaviour.
13881 * Enhance bctest to search for a working bc along $PATH and print
13884 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13886 * Fix memory leaks in err.c: free err_data string if necessary;
13887 don't write to the wrong index in ERR_set_error_data.
13891 * Implement ssl23_peek (analogous to ssl23_read), which previously
13896 * Replace rdtsc with _emit statements for VC++ version 5.
13898 *Jeremy Cooper <jeremy@baymoo.org>*
13900 * Make it possible to reuse SSLv2 sessions.
13904 * In copy_email() check for >= 0 as a return value for
13905 X509_NAME_get_index_by_NID() since 0 is a valid index.
13907 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
13909 * Avoid coredump with unsupported or invalid public keys by checking if
13910 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
13911 PKCS7_verify() fails with non detached data.
13915 * Don't use getenv in library functions when run as setuid/setgid.
13916 New function OPENSSL_issetugid().
13920 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
13921 due to incorrect handling of multi-threading:
13923 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
13925 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
13927 3. Count how many times MemCheck_off() has been called so that
13928 nested use can be treated correctly. This also avoids
13929 inband-signalling in the previous code (which relied on the
13930 assumption that thread ID 0 is impossible).
13934 * Add "-rand" option also to s_client and s_server.
13938 * Fix CPU detection on Irix 6.x.
13939 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
13940 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13942 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
13947 *This change does not apply to 0.9.7.*
13949 * Use the cached encoding of an X509_NAME structure rather than
13950 copying it. This is apparently the reason for the libsafe "errors"
13951 but the code is actually correct.
13955 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
13956 Bleichenbacher's DSA attack.
13957 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
13958 to be set and top=0 forces the highest bit to be set; top=-1 is new
13959 and leaves the highest bit random.
13961 *Ulf Moeller, Bodo Moeller*
13963 * In the NCONF_...-based implementations for CONF_... queries
13964 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
13965 a temporary CONF structure with the data component set to NULL
13966 (which gives segmentation faults in lh_retrieve).
13967 Instead, use NULL for the CONF pointer in CONF_get_string and
13968 CONF_get_number (which may use environment variables) and directly
13969 return NULL from CONF_get_section.
13973 * Fix potential buffer overrun for EBCDIC.
13977 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
13978 keyUsage if basicConstraints absent for a CA.
13982 * Make SMIME_write_PKCS7() write mail header values with a format that
13983 is more generally accepted (no spaces before the semicolon), since
13984 some programs can't parse those values properly otherwise. Also make
13985 sure BIO's that break lines after each write do not create invalid
13990 * Make the CRL encoding routines work with empty SEQUENCE OF. The
13991 macros previously used would not encode an empty SEQUENCE OF
13992 and break the signature.
13996 *This change does not apply to 0.9.7.*
13998 * Zero the premaster secret after deriving the master secret in
14003 * Add some EVP_add_digest_alias registrations (as found in
14004 OpenSSL_add_all_digests()) to SSL_library_init()
14005 aka OpenSSL_add_ssl_algorithms(). This provides improved
14006 compatibility with peers using X.509 certificates
14007 with unconventional AlgorithmIdentifier OIDs.
14011 * Fix for Irix with NO_ASM.
14013 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14015 * ./config script fixes.
14017 *Ulf Moeller, Richard Levitte*
14019 * Fix 'openssl passwd -1'.
14023 * Change PKCS12_key_gen_asc() so it can cope with non null
14024 terminated strings whose length is passed in the passlen
14025 parameter, for example from PEM callbacks. This was done
14026 by adding an extra length parameter to asc2uni().
14028 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14030 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14031 call failed, free the DSA structure.
14035 * Fix to uni2asc() to cope with zero length Unicode strings.
14036 These are present in some PKCS#12 files.
14040 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14041 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14042 when writing a 32767 byte record.
14044 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14046 * In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
14047 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
14049 (RSA objects have a reference count access to which is protected
14050 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14051 so they are meant to be shared between threads.)
14052 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14053 "Reddie, Steven" <Steven.Reddie@ca.com>*
14055 * Fix a deadlock in CRYPTO_mem_leaks().
14059 * Use better test patterns in bntest.
14063 * rand_win.c fix for Borland C.
14067 * BN_rshift bugfix for n == 0.
14071 * Add a 'bctest' script that checks for some known 'bc' bugs
14072 so that 'make test' does not abort just because 'bc' is broken.
14076 * Store verify_result within SSL_SESSION also for client side to
14077 avoid potential security hole. (Re-used sessions on the client side
14078 always resulted in verify_result==X509_V_OK, not using the original
14079 result of the server certificate verification.)
14083 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14084 SSL3_RT_APPLICATION_DATA, return 0.
14085 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14090 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14091 releases, have been re-implemented by renaming the previous
14092 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14093 and ssl3_read_internal, respectively, and adding 'peek' parameters
14094 to them. The new ssl[23]_{read,peek} functions are calls to
14095 ssl[23]_read_internal with the 'peek' flag set appropriately.
14096 A 'peek' parameter has also been added to ssl3_read_bytes, which
14097 does the actual work for ssl3_read_internal.
14101 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14102 the method-specific "init()" handler. Also clean up ex_data after
14103 calling the method-specific "finish()" handler. Previously, this was
14104 happening the other way round.
14108 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14109 The previous value, 12, was not always sufficient for BN_mod_exp().
14113 * Make sure that shared libraries get the internal name engine with
14114 the full version number and not just 0. This should mark the
14115 shared libraries as not backward compatible. Of course, this should
14116 be changed again when we can guarantee backward binary compatibility.
14120 * Fix typo in get_cert_by_subject() in by_dir.c
14122 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14124 * Rework the system to generate shared libraries:
14126 - Make note of the expected extension for the shared libraries and
14127 if there is a need for symbolic links from for example libcrypto.so.0
14128 to libcrypto.so.0.9.7. There is extended info in Configure for
14131 - Make as few rebuilds of the shared libraries as possible.
14133 - Still avoid linking the OpenSSL programs with the shared libraries.
14135 - When installing, install the shared libraries separately from the
14140 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14142 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14143 and not in SSL_clear because the latter is also used by the
14144 accept/connect functions; previously, the settings made by
14145 SSL_set_read_ahead would be lost during the handshake.
14147 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14149 * Correct util/mkdef.pl to be selective about disabled algorithms.
14150 Previously, it would create entries for disabled algorithms no
14155 * Added several new manual pages for SSL_* function.
14159 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000] ###
14161 * In ssl23_get_client_hello, generate an error message when faced
14162 with an initial SSL 3.0/TLS record that is too small to contain the
14163 first two bytes of the ClientHello message, i.e. client_version.
14164 (Note that this is a pathologic case that probably has never happened
14165 in real life.) The previous approach was to use the version number
14166 from the record header as a substitute; but our protocol choice
14167 should not depend on that one because it is not authenticated
14168 by the Finished messages.
14172 * More robust randomness gathering functions for Windows.
14174 *Jeffrey Altman <jaltman@columbia.edu>*
14176 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14177 not set then we don't setup the error code for issuer check errors
14178 to avoid possibly overwriting other errors which the callback does
14179 handle. If an application does set the flag then we assume it knows
14180 what it is doing and can handle the new informational codes
14185 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14186 a general "ANY" type, as such it should be able to decode anything
14187 including tagged types. However it didn't check the class so it would
14188 wrongly interpret tagged types in the same way as their universal
14189 counterpart and unknown types were just rejected. Changed so that the
14190 tagged and unknown types are handled in the same way as a SEQUENCE:
14191 that is the encoding is stored intact. There is also a new type
14192 "V_ASN1_OTHER" which is used when the class is not universal, in this
14193 case we have no idea what the actual type is so we just lump them all
14198 * On VMS, stdout may very well lead to a file that is written to
14199 in a record-oriented fashion. That means that every write() will
14200 write a separate record, which will be read separately by the
14201 programs trying to read from it. This can be very confusing.
14203 The solution is to put a BIO filter in the way that will buffer
14204 text until a linefeed is reached, and then write everything a
14205 line at a time, so every record written will be an actual line,
14206 not chunks of lines and not (usually doesn't happen, but I've
14207 seen it once) several lines in one record. BIO_f_linebuffer() is
14210 Currently, it's a VMS-only method, because that's where it has
14211 been tested well enough.
14215 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14216 it can return incorrect results.
14217 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14218 but it was in 0.9.6-beta[12].)
14222 * Disable the check for content being present when verifying detached
14223 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14224 include zero length content when signing messages.
14228 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14229 BIO_ctrl (for BIO pairs).
14233 * Add DSO method for VMS.
14237 * Bug fix: Montgomery multiplication could produce results with the
14242 * Add RPM specification openssl.spec and modify it to build three
14243 packages. The default package contains applications, application
14244 documentation and run-time libraries. The devel package contains
14245 include files, static libraries and function documentation. The
14246 doc package contains the contents of the doc directory. The original
14247 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14251 * Add a large number of documentation files for many SSL routines.
14253 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14255 * Add a configuration entry for Sony News 4.
14257 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14259 * Don't set the two most significant bits to one when generating a
14260 random number < q in the DSA library.
14264 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14265 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14266 the underlying transport is blocking) if a handshake took place.
14267 (The default behaviour is needed by applications such as s_client
14268 and s_server that use select() to determine when to use SSL_read;
14269 but for applications that know in advance when to expect data, it
14270 just makes things more complicated.)
14274 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14279 * Add a few more EBCDIC conditionals that make `req' and `x509'
14280 work better on such systems.
14282 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14284 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14285 Update PKCS12_parse() so it copies the friendlyName and the
14286 keyid to the certificates aux info.
14290 * Fix bug in PKCS7_verify() which caused an infinite loop
14291 if there was more than one signature.
14293 *Sven Uszpelkat <su@celocom.de>*
14295 * Major change in util/mkdef.pl to include extra information
14296 about each symbol, as well as presenting variables as well
14297 as functions. This change means that there's n more need
14298 to rebuild the .num files when some algorithms are excluded.
14302 * Allow the verify time to be set by an application,
14303 rather than always using the current time.
14307 * Phase 2 verify code reorganisation. The certificate
14308 verify code now looks up an issuer certificate by a
14309 number of criteria: subject name, authority key id
14310 and key usage. It also verifies self signed certificates
14311 by the same criteria. The main comparison function is
14312 X509_check_issued() which performs these checks.
14314 Lot of changes were necessary in order to support this
14315 without completely rewriting the lookup code.
14317 Authority and subject key identifier are now cached.
14319 The LHASH 'certs' is X509_STORE has now been replaced
14320 by a STACK_OF(X509_OBJECT). This is mainly because an
14321 LHASH can't store or retrieve multiple objects with
14322 the same hash value.
14324 As a result various functions (which were all internal
14325 use only) have changed to handle the new X509_STORE
14326 structure. This will break anything that messed round
14327 with X509_STORE internally.
14329 The functions X509_STORE_add_cert() now checks for an
14330 exact match, rather than just subject name.
14332 The X509_STORE API doesn't directly support the retrieval
14333 of multiple certificates matching a given criteria, however
14334 this can be worked round by performing a lookup first
14335 (which will fill the cache with candidate certificates)
14336 and then examining the cache for matches. This is probably
14337 the best we can do without throwing out X509_LOOKUP
14338 entirely (maybe later...).
14340 The X509_VERIFY_CTX structure has been enhanced considerably.
14342 All certificate lookup operations now go via a get_issuer()
14343 callback. Although this currently uses an X509_STORE it
14344 can be replaced by custom lookups. This is a simple way
14345 to bypass the X509_STORE hackery necessary to make this
14346 work and makes it possible to use more efficient techniques
14347 in future. A very simple version which uses a simple
14348 STACK for its trusted certificate store is also provided
14349 using X509_STORE_CTX_trusted_stack().
14351 The verify_cb() and verify() callbacks now have equivalents
14352 in the X509_STORE_CTX structure.
14354 X509_STORE_CTX also has a 'flags' field which can be used
14355 to customise the verify behaviour.
14359 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14360 excludes S/MIME capabilities.
14364 * When a certificate request is read in keep a copy of the
14365 original encoding of the signed data and use it when outputting
14366 again. Signatures then use the original encoding rather than
14367 a decoded, encoded version which may cause problems if the
14368 request is improperly encoded.
14372 * For consistency with other BIO_puts implementations, call
14373 buffer_write(b, ...) directly in buffer_puts instead of calling
14376 In BIO_puts, increment b->num_write as in BIO_write.
14378 *Peter.Sylvester@EdelWeb.fr*
14380 * Fix BN_mul_word for the case where the word is 0. (We have to use
14381 BN_zero, we may not return a BIGNUM with an array consisting of
14382 words set to zero.)
14386 * Avoid calling abort() from within the library when problems are
14387 detected, except if preprocessor symbols have been defined
14388 (such as REF_CHECK, BN_DEBUG etc.).
14392 * New openssl application 'rsautl'. This utility can be
14393 used for low level RSA operations. DER public key
14394 BIO/fp routines also added.
14398 * New Configure entry and patches for compiling on QNX 4.
14400 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14402 * A demo state-machine implementation was sponsored by
14403 Nuron (http://www.nuron.com/) and is now available in
14404 demos/state_machine.
14408 * New options added to the 'dgst' utility for signature
14409 generation and verification.
14413 * Unrecognized PKCS#7 content types are now handled via a
14414 catch all ASN1_TYPE structure. This allows unsupported
14415 types to be stored as a "blob" and an application can
14416 encode and decode it manually.
14420 * Fix various signed/unsigned issues to make a_strex.c
14421 compile under VC++.
14423 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14425 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14426 length if passed a buffer. ASN1_INTEGER_to_BN failed
14427 if passed a NULL BN and its argument was negative.
14429 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14431 * Modification to PKCS#7 encoding routines to output definite
14432 length encoding. Since currently the whole structures are in
14433 memory there's not real point in using indefinite length
14434 constructed encoding. However if OpenSSL is compiled with
14435 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14439 * Added BIO_vprintf() and BIO_vsnprintf().
14443 * Added more prefixes to parse for in the strings written
14444 through a logging bio, to cover all the levels that are available
14445 through syslog. The prefixes are now:
14447 PANIC, EMERG, EMR => LOG_EMERG
14448 ALERT, ALR => LOG_ALERT
14449 CRIT, CRI => LOG_CRIT
14450 ERROR, ERR => LOG_ERR
14451 WARNING, WARN, WAR => LOG_WARNING
14452 NOTICE, NOTE, NOT => LOG_NOTICE
14453 INFO, INF => LOG_INFO
14454 DEBUG, DBG => LOG_DEBUG
14456 and as before, if none of those prefixes are present at the
14457 beginning of the string, LOG_ERR is chosen.
14459 On Win32, the LOG_* levels are mapped according to this:
14461 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14462 LOG_WARNING => EVENTLOG_WARNING_TYPE
14463 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14468 * Made it possible to reconfigure with just the configuration
14469 argument "reconf" or "reconfigure". The command line arguments
14470 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14471 and are retrieved from there when reconfiguring.
14477 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14479 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14483 * The obj_dat.pl script was messing up the sorting of object
14484 names. The reason was that it compared the quoted version
14485 of strings as a result "OCSP" > "OCSP Signing" because
14486 " > SPACE. Changed script to store unquoted versions of
14487 names and add quotes on output. It was also omitting some
14488 names from the lookup table if they were given a default
14489 value (that is if SN is missing it is given the same
14490 value as LN and vice versa), these are now added on the
14491 grounds that if an object has a name we should be able to
14492 look it up. Finally added warning output when duplicate
14493 short or long names are found.
14497 * Changes needed for Tandem NSK.
14499 *Scott Uroff <scott@xypro.com>*
14501 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14502 RSA_padding_check_SSLv23(), special padding was never detected
14503 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14504 version rollback attacks was not effective.
14506 In s23_clnt.c, don't use special rollback-attack detection padding
14507 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14508 client; similarly, in s23_srvr.c, don't do the rollback check if
14509 SSL 2.0 is the only protocol enabled in the server.
14513 * Make it possible to get hexdumps of unprintable data with 'openssl
14514 asn1parse'. By implication, the functions ASN1_parse_dump() and
14515 BIO_dump_indent() are added.
14519 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14520 these print out strings and name structures based on various
14521 flags including RFC2253 support and proper handling of
14522 multibyte characters. Added options to the 'x509' utility
14523 to allow the various flags to be set.
14527 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14528 Also change the functions X509_cmp_current_time() and
14529 X509_gmtime_adj() work with an ASN1_TIME structure,
14530 this will enable certificates using GeneralizedTime in validity
14531 dates to be checked.
14535 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14536 negative public key encodings) on by default,
14537 NO_NEG_PUBKEY_BUG can be set to disable it.
14541 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14542 content octets. An i2c_ASN1_OBJECT is unnecessary because
14543 the encoding can be trivially obtained from the structure.
14547 * crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
14548 not read locks (CRYPTO_r_[un]lock).
14552 * A first attempt at creating official support for shared
14553 libraries through configuration. I've kept it so the
14554 default is static libraries only, and the OpenSSL programs
14555 are always statically linked for now, but there are
14556 preparations for dynamic linking in place.
14557 This has been tested on Linux and Tru64.
14561 * Randomness polling function for Win9x, as described in:
14562 Peter Gutmann, Software Generation of Practically Strong
14567 * Fix so PRNG is seeded in req if using an already existing
14572 * New options to smime application. -inform and -outform
14573 allow alternative formats for the S/MIME message including
14574 PEM and DER. The -content option allows the content to be
14575 specified separately. This should allow things like Netscape
14576 form signing output easier to verify.
14580 * Fix the ASN1 encoding of tags using the 'long form'.
14584 * New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
14585 STRING types. These convert content octets to and from the
14586 underlying type. The actual tag and length octets are
14587 already assumed to have been read in and checked. These
14588 are needed because all other string types have virtually
14589 identical handling apart from the tag. By having versions
14590 of the ASN1 functions that just operate on content octets
14591 IMPLICIT tagging can be handled properly. It also allows
14592 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14593 and ASN1_INTEGER are identical apart from the tag.
14597 * Change the handling of OID objects as follows:
14599 - New object identifiers are inserted in objects.txt, following
14600 the syntax given in objects.README.
14601 - objects.pl is used to process obj_mac.num and create a new
14603 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14606 This is currently kind of a hack, and the perl code in objects.pl
14607 isn't very elegant, but it works as I intended. The simplest way
14608 to check that it worked correctly is to look in obj_dat.h and
14609 check the array nid_objs and make sure the objects haven't moved
14610 around (this is important!). Additions are OK, as well as
14611 consistent name changes.
14615 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14619 * Addition of the command line parameter '-rand file' to 'openssl req'.
14620 The given file adds to whatever has already been seeded into the
14621 random pool through the RANDFILE configuration file option or
14622 environment variable, or the default random state file.
14626 * mkstack.pl now sorts each macro group into lexical order.
14627 Previously the output order depended on the order the files
14628 appeared in the directory, resulting in needless rewriting
14633 * Patches to make OpenSSL compile under Win32 again. Mostly
14634 work arounds for the VC++ problem that it treats func() as
14635 func(void). Also stripped out the parts of mkdef.pl that
14636 added extra typesafe functions: these no longer exist.
14640 * Reorganisation of the stack code. The macros are now all
14641 collected in safestack.h . Each macro is defined in terms of
14642 a "stack macro" of the form SKM_<name>(type, a, b). The
14643 DEBUG_SAFESTACK is now handled in terms of function casts,
14644 this has the advantage of retaining type safety without the
14645 use of additional functions. If DEBUG_SAFESTACK is not defined
14646 then the non typesafe macros are used instead. Also modified the
14647 mkstack.pl script to handle the new form. Needs testing to see
14648 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14649 the default if no major problems. Similar behaviour for ASN1_SET_OF
14650 and PKCS12_STACK_OF.
14654 * When some versions of IIS use the 'NET' form of private key the
14655 key derivation algorithm is different. Normally MD5(password) is
14656 used as a 128 bit RC4 key. In the modified case
14657 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14658 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14659 as the old Netscape_RSA functions except they have an additional
14660 'sgckey' parameter which uses the modified algorithm. Also added
14661 an -sgckey command line option to the rsa utility. Thanks to
14662 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14663 algorithm to openssl-dev.
14667 * The evp_local.h macros were using 'c.##kname' which resulted in
14668 invalid expansion on some systems (SCO 5.0.5 for example).
14669 Corrected to 'c.kname'.
14671 *Phillip Porch <root@theporch.com>*
14673 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14674 a STACK of email addresses from a certificate or request, these look
14675 in the subject name and the subject alternative name extensions and
14676 omit any duplicate addresses.
14680 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14681 This makes DSA verification about 2 % faster.
14685 * Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
14686 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14687 plus overhead for 1024 bit moduli).
14688 This makes exponentiations about 0.5 % faster for 1024 bit
14689 exponents (as measured by "openssl speed rsa2048").
14693 * Rename memory handling macros to avoid conflicts with other
14695 Malloc => OPENSSL_malloc
14696 Malloc_locked => OPENSSL_malloc_locked
14697 Realloc => OPENSSL_realloc
14698 Free => OPENSSL_free
14702 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14703 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14707 * CygWin32 support.
14709 *John Jarvie <jjarvie@newsguy.com>*
14711 * The type-safe stack code has been rejigged. It is now only compiled
14712 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14713 by default all type-specific stack functions are "#define"d back to
14714 standard stack functions. This results in more streamlined output
14715 but retains the type-safety checking possibilities of the original
14720 * The STACK code has been cleaned up, and certain type declarations
14721 that didn't make a lot of sense have been brought in line. This has
14722 also involved a cleanup of sorts in safestack.h to more correctly
14723 map type-safe stack functions onto their plain stack counterparts.
14724 This work has also resulted in a variety of "const"ifications of
14725 lots of the code, especially "_cmp" operations which should normally
14726 be prototyped with "const" parameters anyway.
14730 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14731 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14732 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14733 where all of 'md' is used each time the PRNG is used, but 'state'
14734 is used only indexed by a cyclic counter. As entropy may not be
14735 well distributed from the beginning, 'md' is important as a
14736 chaining variable. However, the output function chains only half
14737 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14738 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14739 in all of 'state' being rewritten, with the new values depending
14740 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14744 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14745 the handshake is continued after ssl_verify_cert_chain();
14746 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14747 can lead to 'unexplainable' connection aborts later.
14749 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14751 * Major EVP API cipher revision.
14752 Add hooks for extra EVP features. This allows various cipher
14753 parameters to be set in the EVP interface. Support added for variable
14754 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14755 setting of RC2 and RC5 parameters.
14757 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14760 Remove lots of duplicated code from the EVP library. For example *every*
14761 cipher init() function handles the 'iv' in the same way according to the
14762 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14763 for CFB and OFB modes they zero ctx->num.
14765 New functionality allows removal of S/MIME code RC2 hack.
14767 Most of the routines have the same form and so can be declared in terms
14770 By shifting this to the top level EVP_CipherInit() it can be removed from
14771 all individual ciphers. If the cipher wants to handle IVs or keys
14772 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14775 Change lots of functions like EVP_EncryptUpdate() to now return a
14776 value: although software versions of the algorithms cannot fail
14777 any installed hardware versions can.
14781 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14782 this option is set, tolerate broken clients that send the negotiated
14783 protocol version number instead of the requested protocol version
14788 * Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
14789 i.e. non-zero for export ciphersuites, zero otherwise.
14790 Previous versions had this flag inverted, inconsistent with
14791 rsa_tmp_cb (..._TMP_RSA_CB).
14793 *Bodo Moeller; problem reported by Amit Chopra*
14795 * Add missing DSA library text string. Work around for some IIS
14796 key files with invalid SEQUENCE encoding.
14800 * Add a document (doc/standards.txt) that list all kinds of standards
14801 and so on that are implemented in OpenSSL.
14805 * Enhance c_rehash script. Old version would mishandle certificates
14806 with the same subject name hash and wouldn't handle CRLs at all.
14807 Added -fingerprint option to crl utility, to support new c_rehash
14812 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14816 * Fix for SSL server purpose checking. Server checking was
14817 rejecting certificates which had extended key usage present
14818 but no ssl client purpose.
14820 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14822 * Make PKCS#12 code work with no password. The PKCS#12 spec
14823 is a little unclear about how a blank password is handled.
14824 Since the password in encoded as a BMPString with terminating
14825 double NULL a zero length password would end up as just the
14826 double NULL. However no password at all is different and is
14827 handled differently in the PKCS#12 key generation code. NS
14828 treats a blank password as zero length. MSIE treats it as no
14829 password on export: but it will try both on import. We now do
14830 the same: PKCS12_parse() tries zero length and no password if
14831 the password is set to "" or NULL (NULL is now a valid password:
14832 it wasn't before) as does the pkcs12 application.
14836 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14837 perror when PEM_read_bio_X509_REQ fails, the error message must
14838 be obtained from the error queue.
14842 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14843 it in ERR_remove_state if appropriate, and change ERR_get_state
14844 accordingly to avoid race conditions (this is necessary because
14845 thread_hash is no longer constant once set).
14849 * Bugfix for linux-elf makefile.one.
14853 * RSA_get_default_method() will now cause a default
14854 RSA_METHOD to be chosen if one doesn't exist already.
14855 Previously this was only set during a call to RSA_new()
14856 or RSA_new_method(NULL) meaning it was possible for
14857 RSA_get_default_method() to return NULL.
14861 * Added native name translation to the existing DSO code
14862 that will convert (if the flag to do so is set) filenames
14863 that are sufficiently small and have no path information
14864 into a canonical native form. Eg. "blah" converted to
14865 "libblah.so" or "blah.dll" etc.
14869 * New function ERR_error_string_n(e, buf, len) which is like
14870 ERR_error_string(e, buf), but writes at most 'len' bytes
14871 including the 0 terminator. For ERR_error_string_n, 'buf'
14874 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14876 * CONF library reworked to become more general. A new CONF
14877 configuration file reader "class" is implemented as well as a
14878 new functions (NCONF_*, for "New CONF") to handle it. The now
14879 old CONF_* functions are still there, but are reimplemented to
14880 work in terms of the new functions. Also, a set of functions
14881 to handle the internal storage of the configuration data is
14882 provided to make it easier to write new configuration file
14883 reader "classes" (I can definitely see something reading a
14884 configuration file in XML format, for example), called _CONF_*,
14885 or "the configuration storage API"...
14887 The new configuration file reading functions are:
14889 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14890 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14892 NCONF_default, NCONF_WIN32
14894 NCONF_dump_fp, NCONF_dump_bio
14896 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14897 NCONF_new creates a new CONF object. This works in the same way
14898 as other interfaces in OpenSSL, like the BIO interface.
14899 NCONF_dump_* dump the internal storage of the configuration file,
14900 which is useful for debugging. All other functions take the same
14901 arguments as the old CONF_* functions with the exception of the
14902 first that must be a `CONF *' instead of a `LHASH *'.
14904 To make it easier to use the new classes with the old CONF_* functions,
14905 the function CONF_set_default_method is provided.
14909 * Add '-tls1' option to 'openssl ciphers', which was already
14910 mentioned in the documentation but had not been implemented.
14911 (This option is not yet really useful because even the additional
14912 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
14916 * Initial DSO code added into libcrypto for letting OpenSSL (and
14917 OpenSSL-based applications) load shared libraries and bind to
14918 them in a portable way.
14920 *Geoff Thorpe, with contributions from Richard Levitte*
14922 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000] ###
14924 * Make sure _lrotl and _lrotr are only used with MSVC.
14926 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
14927 (the default implementation of RAND_status).
14929 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
14930 to '-clrext' (= clear extensions), as intended and documented.
14931 *Bodo Moeller; inconsistency pointed out by Michael Attili
14932 <attili@amaxo.com>*
14934 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
14935 was larger than the MD block size.
14937 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
14939 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
14940 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
14941 using the passed key: if the passed key was a private key the result
14942 of X509_print(), for example, would be to print out all the private key
14947 * des_quad_cksum() byte order bug fix.
14948 *Ulf Möller, using the problem description in krb4-0.9.7, where
14949 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
14951 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
14954 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
14956 * For easily testing in shell scripts whether some command
14957 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
14958 returns with exit code 0 iff no command of the given name is available.
14959 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
14960 the output goes to stdout and nothing is printed to stderr.
14961 Additional arguments are always ignored.
14963 Since for each cipher there is a command of the same name,
14964 the 'no-cipher' compilation switches can be tested this way.
14966 ('openssl no-XXX' is not able to detect pseudo-commands such
14967 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
14971 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
14975 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
14976 is set; it will be thrown away anyway because each handshake creates
14978 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
14979 to parameters -- in previous versions (since OpenSSL 0.9.3) the
14980 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
14981 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
14985 * New s_client option -ign_eof: EOF at stdin is ignored, and
14986 'Q' and 'R' lose their special meanings (quit/renegotiate).
14987 This is part of what -quiet does; unlike -quiet, -ign_eof
14988 does not suppress any output.
14992 * Add compatibility options to the purpose and trust code. The
14993 purpose X509_PURPOSE_ANY is "any purpose" which automatically
14994 accepts a certificate or CA, this was the previous behaviour,
14995 with all the associated security issues.
14997 X509_TRUST_COMPAT is the old trust behaviour: only and
14998 automatically trust self signed roots in certificate store. A
14999 new trust setting X509_TRUST_DEFAULT is used to specify that
15000 a purpose has no associated trust setting and it should instead
15001 use the value in the default purpose.
15005 * Fix the PKCS#8 DSA private key code so it decodes keys again
15006 and fix a memory leak.
15010 * In util/mkerr.pl (which implements 'make errors'), preserve
15011 reason strings from the previous version of the .c file, as
15012 the default to have only downcase letters (and digits) in
15013 automatically generated reasons codes is not always appropriate.
15017 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15018 using strerror. Previously, ERR_reason_error_string() returned
15019 library names as reason strings for SYSerr; but SYSerr is a special
15020 case where small numbers are errno values, not library numbers.
15024 * Add '-dsaparam' option to 'openssl dhparam' application. This
15025 converts DSA parameters into DH parameters. (When creating parameters,
15026 DSA_generate_parameters is used.)
15030 * Include 'length' (recommended exponent length) in C code generated
15031 by 'openssl dhparam -C'.
15035 * The second argument to set_label in perlasm was already being used
15036 so couldn't be used as a "file scope" flag. Moved to third argument
15041 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15042 instead of RAND_bytes for encryption IVs and salts.
15046 * Include RAND_status() into RAND_METHOD instead of implementing
15047 it only for md_rand.c Otherwise replacing the PRNG by calling
15048 RAND_set_rand_method would be impossible.
15052 * Don't let DSA_generate_key() enter an infinite loop if the random
15053 number generation fails.
15057 * New 'rand' application for creating pseudo-random output.
15061 * Added configuration support for Linux/IA64
15063 *Rolf Haberrecker <rolf@suse.de>*
15065 * Assembler module support for Mingw32.
15069 * Shared library support for HPUX (in shlib/).
15071 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15073 * Shared library support for Solaris gcc.
15075 *Lutz Behnke <behnke@trustcenter.de>*
15077 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000] ###
15079 * PKCS7_encrypt() was adding text MIME headers twice because they
15080 were added manually and by SMIME_crlf_copy().
15084 * In bntest.c don't call BN_rand with zero bits argument.
15086 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15088 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15089 case was implemented. This caused BN_div_recp() to fail occasionally.
15093 * Add an optional second argument to the set_label() in the perl
15094 assembly language builder. If this argument exists and is set
15095 to 1 it signals that the assembler should use a symbol whose
15096 scope is the entire file, not just the current function. This
15097 is needed with MASM which uses the format label:: for this scope.
15099 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15101 * Change the ASN1 types so they are typedefs by default. Before
15102 almost all types were #define'd to ASN1_STRING which was causing
15103 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15108 * Change names of new functions to the new get1/get0 naming
15109 convention: After 'get1', the caller owns a reference count
15110 and has to call ..._free; 'get0' returns a pointer to some
15111 data structure without incrementing reference counters.
15112 (Some of the existing 'get' functions increment a reference
15113 counter, some don't.)
15114 Similarly, 'set1' and 'add1' functions increase reference
15115 counters or duplicate objects.
15119 * Allow for the possibility of temp RSA key generation failure:
15120 the code used to assume it always worked and crashed on failure.
15124 * Fix potential buffer overrun problem in BIO_printf().
15125 *Ulf Möller, using public domain code by Patrick Powell; problem
15126 pointed out by David Sacerdote <das33@cornell.edu>*
15128 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15129 RAND_egd() and RAND_status(). In the command line application,
15130 the EGD socket can be specified like a seed file using RANDFILE
15135 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15136 Some CAs (e.g. Verisign) distribute certificates in this form.
15140 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15141 list to exclude them. This means that no special compilation option
15142 is needed to use anonymous DH: it just needs to be included in the
15147 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15148 EVP_MD_type. The old functionality is available in a new macro called
15149 EVP_MD_md(). Change code that uses it and update docs.
15153 * ..._ctrl functions now have corresponding ..._callback_ctrl functions
15154 where the 'void *' argument is replaced by a function pointer argument.
15155 Previously 'void *' was abused to point to functions, which works on
15156 many platforms, but is not correct. As these functions are usually
15157 called by macros defined in OpenSSL header files, most source code
15158 should work without changes.
15162 * <openssl/opensslconf.h> (which is created by Configure) now contains
15163 sections with information on -D... compiler switches used for
15164 compiling the library so that applications can see them. To enable
15165 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
15166 must be defined. E.g.,
15167 #define OPENSSL_ALGORITHM_DEFINES
15168 #include <openssl/opensslconf.h>
15169 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
15171 *Richard Levitte, Ulf and Bodo Möller*
15173 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15178 * Change the 'other' type in certificate aux info to a STACK_OF
15179 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15180 the required ASN1 format: arbitrary types determined by an OID.
15184 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15185 argument to 'req'. This is not because the function is newer or
15186 better than others it just uses the work 'NEW' in the certificate
15187 request header lines. Some software needs this.
15191 * Reorganise password command line arguments: now passwords can be
15192 obtained from various sources. Delete the PEM_cb function and make
15193 it the default behaviour: i.e. if the callback is NULL and the
15194 usrdata argument is not NULL interpret it as a null terminated pass
15195 phrase. If usrdata and the callback are NULL then the pass phrase
15196 is prompted for as usual.
15200 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15201 the support is automatically enabled. The resulting binaries will
15202 autodetect the card and use it if present.
15204 *Ben Laurie and Compaq Inc.*
15206 * Work around for Netscape hang bug. This sends certificate request
15207 and server done in one record. Since this is perfectly legal in the
15208 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15209 the bugs/SSLv3 entry for more info.
15213 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15217 * Add -rand argument to smime and pkcs12 applications and read/write
15222 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15226 * Add command line password options to the remaining applications.
15230 * Bug fix for BN_div_recp() for numerators with an even number of
15235 * More tests in bntest.c, and changed test_bn output.
15239 * ./config recognizes MacOS X now.
15243 * Bug fix for BN_div() when the first words of num and divisor are
15244 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
15248 * Add support for various broken PKCS#8 formats, and command line
15249 options to produce them.
15253 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15254 get temporary BIGNUMs from a BN_CTX.
15258 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15263 * Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
15264 include a #define from the old name to the new. The original intent
15265 was that statically linked binaries could for example just call
15266 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15267 link with digests. This never worked because SSLeay_add_all_digests()
15268 and SSLeay_add_all_ciphers() were in the same source file so calling
15269 one would link with the other. They are now in separate source files.
15273 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15277 * Use a less unusual form of the Miller-Rabin primality test (it used
15278 a binary algorithm for exponentiation integrated into the Miller-Rabin
15279 loop, our standard modexp algorithms are faster).
15283 * Support for the EBCDIC character set completed.
15285 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15287 * Source code cleanups: use const where appropriate, eliminate casts,
15288 use void * instead of char * in lhash.
15292 * Bugfix: ssl3_send_server_key_exchange was not restartable
15293 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15294 this the server could overwrite ephemeral keys that the client
15299 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15300 using 50 iterations of the Rabin-Miller test.
15302 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15303 iterations of the Rabin-Miller test as required by the appendix
15304 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15305 As BN_is_prime_fasttest includes trial division, DSA parameter
15306 generation becomes much faster.
15308 This implies a change for the callback functions in DSA_is_prime
15309 and DSA_generate_parameters: The callback function is called once
15310 for each positive witness in the Rabin-Miller test, not just
15311 occasionally in the inner loop; and the parameters to the
15312 callback function now provide an iteration count for the outer
15313 loop rather than for the current invocation of the inner loop.
15314 DSA_generate_parameters additionally can call the callback
15315 function with an 'iteration count' of -1, meaning that a
15316 candidate has passed the trial division test (when q is generated
15317 from an application-provided seed, trial division is skipped).
15321 * New function BN_is_prime_fasttest that optionally does trial
15322 division before starting the Rabin-Miller test and has
15323 an additional BN_CTX * argument (whereas BN_is_prime always
15324 has to allocate at least one BN_CTX).
15325 'callback(1, -1, cb_arg)' is called when a number has passed the
15326 trial division stage.
15330 * Fix for bug in CRL encoding. The validity dates weren't being handled
15335 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15339 * New function BN_pseudo_rand().
15343 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15344 bignum version of BN_from_montgomery() with the working code from
15345 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15350 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15351 made it impossible to use the same SSL_SESSION data structure in
15352 SSL2 clients in multiple threads.
15356 * The return value of RAND_load_file() no longer counts bytes obtained
15357 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15358 to seed the PRNG (previously an explicit byte count was required).
15360 *Ulf Möller, Bodo Möller*
15362 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15363 used (char *) instead of (void *) and had casts all over the place.
15367 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15371 * Retain source code compatibility for BN_prime_checks macro:
15372 BN_is_prime(..., BN_prime_checks, ...) now uses
15373 BN_prime_checks_for_size to determine the appropriate number of
15374 Rabin-Miller iterations.
15378 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15379 DH_CHECK_P_NOT_SAFE_PRIME.
15380 (Check if this is true? OpenPGP calls them "strong".)
15384 * Merge the functionality of "dh" and "gendh" programs into a new program
15385 "dhparam". The old programs are retained for now but will handle DH keys
15386 (instead of parameters) in future.
15390 * Make the ciphers, s_server and s_client programs check the return values
15391 when a new cipher list is set.
15395 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15396 ciphers. Before when the 56bit ciphers were enabled the sorting was
15399 The syntax for the cipher sorting has been extended to support sorting by
15400 cipher-strength (using the strength_bits hard coded in the tables).
15401 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15403 Fix a bug in the cipher-command parser: when supplying a cipher command
15404 string with an "undefined" symbol (neither command nor alphanumeric
15405 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15406 an error is flagged.
15408 Due to the strength-sorting extension, the code of the
15409 ssl_create_cipher_list() function was completely rearranged. I hope that
15410 the readability was also increased :-)
15412 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15414 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15415 for the first serial number and places 2 in the serial number file. This
15416 avoids problems when the root CA is created with serial number zero and
15417 the first user certificate has the same issuer name and serial number
15422 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15423 the new code. Add documentation for this stuff.
15427 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15428 X509_*() to X509at_*() on the grounds that they don't handle X509
15429 structures and behave in an analogous way to the X509v3 functions:
15430 they shouldn't be called directly but wrapper functions should be used
15433 So we also now have some wrapper functions that call the X509at functions
15434 when passed certificate requests. (TO DO: similar things can be done with
15435 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15436 things. Some of these need some d2i or i2d and print functionality
15437 because they handle more complex structures.)
15441 * Add missing #ifndefs that caused missing symbols when building libssl
15442 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15443 NO_RSA in ssl/s2*.c.
15445 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15447 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15448 has a return value which indicates the quality of the random data
15449 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15450 error queue. New function RAND_pseudo_bytes() generates output that is
15451 guaranteed to be unique but not unpredictable. RAND_add is like
15452 RAND_seed, but takes an extra argument for an entropy estimate
15453 (RAND_seed always assumes full entropy).
15457 * Do more iterations of Rabin-Miller probable prime test (specifically,
15458 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15459 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15460 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15461 false-positive rate of at most 2^-80 for random input.
15465 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15469 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15470 in the 0.9.5 release), this returns the chain
15471 from an X509_CTX structure with a dup of the stack and all
15472 the X509 reference counts upped: so the stack will exist
15473 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15476 Also make SSL_SESSION_print() print out the verify return
15481 * Add manpage for the pkcs12 command. Also change the default
15482 behaviour so MAC iteration counts are used unless the new
15483 -nomaciter option is used. This improves file security and
15484 only older versions of MSIE (4.0 for example) need it.
15488 * Honor the no-xxx Configure options when creating .DEF files.
15492 * Add PKCS#10 attributes to field table: challengePassword,
15493 unstructuredName and unstructuredAddress. These are taken from
15494 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15495 international characters are used.
15497 More changes to X509_ATTRIBUTE code: allow the setting of types
15498 based on strings. Remove the 'loc' parameter when adding
15499 attributes because these will be a SET OF encoding which is sorted
15504 * Initial changes to the 'req' utility to allow request generation
15505 automation. This will allow an application to just generate a template
15506 file containing all the field values and have req construct the
15509 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15510 used all over the place including certificate requests and PKCS#7
15511 structures. They are currently handled manually where necessary with
15512 some primitive wrappers for PKCS#7. The new functions behave in a
15513 manner analogous to the X509 extension functions: they allow
15514 attributes to be looked up by NID and added.
15516 Later something similar to the X509V3 code would be desirable to
15517 automatically handle the encoding, decoding and printing of the
15518 more complex types. The string types like challengePassword can
15519 be handled by the string table functions.
15521 Also modified the multi byte string table handling. Now there is
15522 a 'global mask' which masks out certain types. The table itself
15523 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15524 is useful when for example there is only one permissible type
15525 (as in countryName) and using the mask might result in no valid
15530 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15531 SSL_get_peer_finished to allow applications to obtain the latest
15532 Finished messages sent to the peer or expected from the peer,
15533 respectively. (SSL_get_peer_finished is usually the Finished message
15534 actually received from the peer, otherwise the protocol will be aborted.)
15536 As the Finished message are message digests of the complete handshake
15537 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15538 be used for external authentication procedures when the authentication
15539 provided by SSL/TLS is not desired or is not enough.
15543 * Enhanced support for Alpha Linux is added. Now ./config checks if
15544 the host supports BWX extension and if Compaq C is present on the
15545 $PATH. Just exploiting of the BWX extension results in 20-30%
15546 performance kick for some algorithms, e.g. DES and RC4 to mention
15547 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15552 * Add support for MS "fast SGC". This is arguably a violation of the
15553 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15554 weak crypto and after checking the certificate is SGC a second one
15555 with strong crypto. MS SGC stops the first handshake after receiving
15556 the server certificate message and sends a second client hello. Since
15557 a server will typically do all the time consuming operations before
15558 expecting any further messages from the client (server key exchange
15559 is the most expensive) there is little difference between the two.
15561 To get OpenSSL to support MS SGC we have to permit a second client
15562 hello message after we have sent server done. In addition we have to
15563 reset the MAC if we do get this second client hello.
15567 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15568 if a DER encoded private key is RSA or DSA traditional format. Changed
15569 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15570 format DER encoded private key. Newer code should use PKCS#8 format which
15571 has the key type encoded in the ASN1 structure. Added DER private key
15572 support to pkcs8 application.
15576 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15577 ciphersuites has been selected (as required by the SSL 3/TLS 1
15578 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15579 is set, we interpret this as a request to violate the specification
15580 (the worst that can happen is a handshake failure, and 'correct'
15581 behaviour would result in a handshake failure anyway).
15585 * In SSL_CTX_add_session, take into account that there might be multiple
15586 SSL_SESSION structures with the same session ID (e.g. when two threads
15587 concurrently obtain them from an external cache).
15588 The internal cache can handle only one SSL_SESSION with a given ID,
15589 so if there's a conflict, we now throw out the old one to achieve
15594 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15595 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15596 some routines that use cipher OIDs: some ciphers do not have OIDs
15597 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15602 * Simplify the trust setting structure and code. Now we just have
15603 two sequences of OIDs for trusted and rejected settings. These will
15604 typically have values the same as the extended key usage extension
15605 and any application specific purposes.
15607 The trust checking code now has a default behaviour: it will just
15608 check for an object with the same NID as the passed id. Functions can
15609 be provided to override either the default behaviour or the behaviour
15610 for a given id. SSL client, server and email already have functions
15611 in place for compatibility: they check the NID and also return "trusted"
15612 if the certificate is self signed.
15616 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15617 traditional format into an EVP_PKEY structure.
15621 * Add a password callback function PEM_cb() which either prompts for
15622 a password if usr_data is NULL or otherwise assumes it is a null
15623 terminated password. Allow passwords to be passed on command line
15624 environment or config files in a few more utilities.
15628 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15629 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15630 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15631 Update documentation.
15635 * Support for ASN1 "NULL" type. This could be handled before by using
15636 ASN1_TYPE but there wasn't any function that would try to read a NULL
15637 and produce an error if it couldn't. For compatibility we also have
15638 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15639 don't allocate anything because they don't need to.
15643 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15646 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15648 * Rebuild of the memory allocation routines used by OpenSSL code and
15649 possibly others as well. The purpose is to make an interface that
15650 provide hooks so anyone can build a separate set of allocation and
15651 deallocation routines to be used by OpenSSL, for example memory
15652 pool implementations, or something else, which was previously hard
15653 since Malloc(), Realloc() and Free() were defined as macros having
15654 the values malloc, realloc and free, respectively (except for Win32
15655 compilations). The same is provided for memory debugging code.
15656 OpenSSL already comes with functionality to find memory leaks, but
15657 this gives people a chance to debug other memory problems.
15659 With these changes, a new set of functions and macros have appeared:
15661 CRYPTO_set_mem_debug_functions() [F]
15662 CRYPTO_get_mem_debug_functions() [F]
15663 CRYPTO_dbg_set_options() [F]
15664 CRYPTO_dbg_get_options() [F]
15665 CRYPTO_malloc_debug_init() [M]
15667 The memory debug functions are NULL by default, unless the library
15668 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15669 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15670 gives the standard debugging functions that come with OpenSSL) or
15671 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15672 provided by the library user) must be used. When the standard
15673 debugging functions are used, CRYPTO_dbg_set_options can be used to
15674 request additional information:
15675 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15676 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15678 Also, things like CRYPTO_set_mem_functions will always give the
15679 expected result (the new set of functions is used for allocation
15680 and deallocation) at all times, regardless of platform and compiler
15683 To finish it up, some functions that were never use in any other
15684 way than through macros have a new API and new semantic:
15686 CRYPTO_dbg_malloc()
15687 CRYPTO_dbg_realloc()
15690 All macros of value have retained their old syntax.
15692 *Richard Levitte and Bodo Moeller*
15694 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15695 ordering of SMIMECapabilities wasn't in "strength order" and there
15696 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15701 * Some ASN1 types with illegal zero length encoding (INTEGER,
15702 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15704 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15706 * Merge in my S/MIME library for OpenSSL. This provides a simple
15707 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15708 functionality to handle multipart/signed properly) and a utility
15709 called 'smime' to call all this stuff. This is based on code I
15710 originally wrote for Celo who have kindly allowed it to be
15711 included in OpenSSL.
15715 * Add variants des_set_key_checked and des_set_key_unchecked of
15716 des_set_key (aka des_key_sched). Global variable des_check_key
15717 decides which of these is called by des_set_key; this way
15718 des_check_key behaves as it always did, but applications and
15719 the library itself, which was buggy for des_check_key == 1,
15720 have a cleaner way to pick the version they need.
15724 * New function PKCS12_newpass() which changes the password of a
15729 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15730 dynamic mix. In both cases the ids can be used as an index into the
15731 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15732 functions so they accept a list of the field values and the
15733 application doesn't need to directly manipulate the X509_TRUST
15738 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15743 * Modify the way the V3 extension code looks up extensions. This now
15744 works in a similar way to the object code: we have some "standard"
15745 extensions in a static table which is searched with OBJ_bsearch()
15746 and the application can add dynamic ones if needed. The file
15747 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15748 updated whenever a new extension is added to the core code and kept
15749 in ext_nid order. There is a simple program 'tabtest.c' which checks
15750 this. New extensions are not added too often so this file can readily
15751 be maintained manually.
15753 There are two big advantages in doing things this way. The extensions
15754 can be looked up immediately and no longer need to be "added" using
15755 X509V3_add_standard_extensions(): this function now does nothing.
15756 *Side note: I get *lots* of email saying the extension code doesn't
15757 work because people forget to call this function*
15758 Also no dynamic allocation is done unless new extensions are added:
15759 so if we don't add custom extensions there is no need to call
15760 X509V3_EXT_cleanup().
15764 * Modify enc utility's salting as follows: make salting the default. Add a
15765 magic header, so unsalted files fail gracefully instead of just decrypting
15766 to garbage. This is because not salting is a big security hole, so people
15767 should be discouraged from doing it.
15771 * Fixes and enhancements to the 'x509' utility. It allowed a message
15772 digest to be passed on the command line but it only used this
15773 parameter when signing a certificate. Modified so all relevant
15774 operations are affected by the digest parameter including the
15775 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15776 DSA key was used because it didn't fix the digest.
15780 * Initial certificate chain verify code. Currently tests the untrusted
15781 certificates for consistency with the verify purpose (which is set
15782 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15784 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15785 this is because it will reject chains with invalid extensions whereas
15786 every previous version of OpenSSL and SSLeay made no checks at all.
15788 Trust code: checks the root CA for the relevant trust settings. Trust
15789 settings have an initial value consistent with the verify purpose: e.g.
15790 if the verify purpose is for SSL client use it expects the CA to be
15791 trusted for SSL client use. However the default value can be changed to
15792 permit custom trust settings: one example of this would be to only trust
15793 certificates from a specific "secure" set of CAs.
15795 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15796 which should be used for version portability: especially since the
15797 verify structure is likely to change more often now.
15799 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15800 to set them. If not set then assume SSL clients will verify SSL servers
15803 Two new options to the verify program: -untrusted allows a set of
15804 untrusted certificates to be passed in and -purpose which sets the
15805 intended purpose of the certificate. If a purpose is set then the
15806 new chain verify code is used to check extension consistency.
15810 * Support for the authority information access extension.
15814 * Modify RSA and DSA PEM read routines to transparently handle
15815 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15816 public keys in a format compatible with certificate
15817 SubjectPublicKeyInfo structures. Unfortunately there were already
15818 functions called *_PublicKey_* which used various odd formats so
15819 these are retained for compatibility: however the DSA variants were
15820 never in a public release so they have been deleted. Changed dsa/rsa
15821 utilities to handle the new format: note no releases ever handled public
15822 keys so we should be OK.
15824 The primary motivation for this change is to avoid the same fiasco
15825 that dogs private keys: there are several incompatible private key
15826 formats some of which are standard and some OpenSSL specific and
15827 require various evil hacks to allow partial transparent handling and
15828 even then it doesn't work with DER formats. Given the option anything
15829 other than PKCS#8 should be dumped: but the other formats have to
15830 stay in the name of compatibility.
15832 With public keys and the benefit of hindsight one standard format
15833 is used which works with EVP_PKEY, RSA or DSA structures: though
15834 it clearly returns an error if you try to read the wrong kind of key.
15836 Added a -pubkey option to the 'x509' utility to output the public key.
15837 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
15838 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
15839 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
15840 that do the same as the EVP_PKEY_assign_*() except they up the
15841 reference count of the added key (they don't "swallow" the
15846 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15847 CRLs would fail if the file contained no certificates or no CRLs:
15848 added a new function to read in both types and return the number
15849 read: this means that if none are read it will be an error. The
15850 DER versions of the certificate and CRL reader would always fail
15851 because it isn't possible to mix certificates and CRLs in DER format
15852 without choking one or the other routine. Changed this to just read
15853 a certificate: this is the best we can do. Also modified the code
15854 in apps/verify.c to take notice of return codes: it was previously
15855 attempting to read in certificates from NULL pointers and ignoring
15856 any errors: this is one reason why the cert and CRL reader seemed
15857 to work. It doesn't check return codes from the default certificate
15858 routines: these may well fail if the certificates aren't installed.
15862 * Code to support otherName option in GeneralName.
15866 * First update to verify code. Change the verify utility
15867 so it warns if it is passed a self signed certificate:
15868 for consistency with the normal behaviour. X509_verify
15869 has been modified to it will now verify a self signed
15870 certificate if *exactly* the same certificate appears
15871 in the store: it was previously impossible to trust a
15872 single self signed certificate. This means that:
15873 openssl verify ss.pem
15874 now gives a warning about a self signed certificate but
15875 openssl verify -CAfile ss.pem ss.pem
15880 * For servers, store verify_result in SSL_SESSION data structure
15881 (and add it to external session representation).
15882 This is needed when client certificate verifications fails,
15883 but an application-provided verification callback (set by
15884 SSL_CTX_set_cert_verify_callback) allows accepting the session
15885 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15886 but returns 1): When the session is reused, we have to set
15887 ssl->verify_result to the appropriate error code to avoid
15890 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15892 * Fix a bug in the new PKCS#7 code: it didn't consider the
15893 case in PKCS7_dataInit() where the signed PKCS7 structure
15894 didn't contain any existing data because it was being created.
15896 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15898 * Add a salt to the key derivation routines in enc.c. This
15899 forms the first 8 bytes of the encrypted file. Also add a
15900 -S option to allow a salt to be input on the command line.
15904 * New function X509_cmp(). Oddly enough there wasn't a function
15905 to compare two certificates. We do this by working out the SHA1
15906 hash and comparing that. X509_cmp() will be needed by the trust
15911 * SSL_get1_session() is like SSL_get_session(), but increments
15912 the reference count in the SSL_SESSION returned.
15914 *Geoff Thorpe <geoff@eu.c2.net>*
15916 * Fix for 'req': it was adding a null to request attributes.
15917 Also change the X509_LOOKUP and X509_INFO code to handle
15918 certificate auxiliary information.
15922 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
15927 * Add the possibility to add extra information to the memory leak
15928 detecting output, to form tracebacks, showing from where each
15929 allocation was originated: CRYPTO_push_info("constant string") adds
15930 the string plus current file name and line number to a per-thread
15931 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
15932 is like calling CYRPTO_pop_info() until the stack is empty.
15933 Also updated memory leak detection code to be multi-thread-safe.
15937 * Add options -text and -noout to pkcs7 utility and delete the
15938 encryption options which never did anything. Update docs.
15942 * Add options to some of the utilities to allow the pass phrase
15943 to be included on either the command line (not recommended on
15944 OSes like Unix) or read from the environment. Update the
15945 manpages and fix a few bugs.
15949 * Add a few manpages for some of the openssl commands.
15953 * Fix the -revoke option in ca. It was freeing up memory twice,
15954 leaking and not finding already revoked certificates.
15958 * Extensive changes to support certificate auxiliary information.
15959 This involves the use of X509_CERT_AUX structure and X509_AUX
15960 functions. An X509_AUX function such as PEM_read_X509_AUX()
15961 can still read in a certificate file in the usual way but it
15962 will also read in any additional "auxiliary information". By
15963 doing things this way a fair degree of compatibility can be
15964 retained: existing certificates can have this information added
15965 using the new 'x509' options.
15967 Current auxiliary information includes an "alias" and some trust
15968 settings. The trust settings will ultimately be used in enhanced
15969 certificate chain verification routines: currently a certificate
15970 can only be trusted if it is self signed and then it is trusted
15975 * Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
15976 The problem was that one of the replacement routines had not been working
15977 since SSLeay releases. For now the offending routine has been replaced
15978 with non-optimised assembler. Even so, this now gives around 95%
15979 performance improvement for 1024 bit RSA signs.
15983 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
15984 handling. Most clients have the effective key size in bits equal to
15985 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
15986 A few however don't do this and instead use the size of the decrypted key
15987 to determine the RC2 key length and the AlgorithmIdentifier to determine
15988 the effective key length. In this case the effective key length can still
15989 be 40 bits but the key length can be 168 bits for example. This is fixed
15990 by manually forcing an RC2 key into the EVP_PKEY structure because the
15991 EVP code can't currently handle unusual RC2 key sizes: it always assumes
15992 the key length and effective key length are equal.
15996 * Add a bunch of functions that should simplify the creation of
15997 X509_NAME structures. Now you should be able to do:
15998 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
15999 and have it automatically work out the correct field type and fill in
16000 the structures. The more adventurous can try:
16001 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16002 and it will (hopefully) work out the correct multibyte encoding.
16006 * Change the 'req' utility to use the new field handling and multibyte
16007 copy routines. Before the DN field creation was handled in an ad hoc
16008 way in req, ca, and x509 which was rather broken and didn't support
16009 BMPStrings or UTF8Strings. Since some software doesn't implement
16010 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16011 using the dirstring_type option. See the new comment in the default
16012 openssl.cnf for more info.
16016 * Make crypto/rand/md_rand.c more robust:
16017 - Assure unique random numbers after fork().
16018 - Make sure that concurrent threads access the global counter and
16019 md serializably so that we never lose entropy in them
16020 or use exactly the same state in multiple threads.
16021 Access to the large state is not always serializable because
16022 the additional locking could be a performance killer, and
16023 md should be large enough anyway.
16027 * New file apps/app_rand.c with commonly needed functionality
16028 for handling the random seed file.
16030 Use the random seed file in some applications that previously did not:
16032 dsaparam -genkey (which also ignored its '-rand' option),
16035 x509 (when signing).
16036 Except on systems with /dev/urandom, it is crucial to have a random
16037 seed file at least for key creation, DSA signing, and for DH exchanges;
16038 for RSA signatures we could do without one.
16040 gendh and gendsa (unlike genrsa) used to read only the first byte
16041 of each file listed in the '-rand' option. The function as previously
16042 found in genrsa is now in app_rand.c and is used by all programs
16043 that support '-rand'.
16047 * In RAND_write_file, use mode 0600 for creating files;
16048 don't just chmod when it may be too late.
16052 * Report an error from X509_STORE_load_locations
16053 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16057 * New function ASN1_mbstring_copy() this copies a string in either
16058 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16059 into an ASN1_STRING type. A mask of permissible types is passed
16060 and it chooses the "minimal" type to use or an error if not type
16065 * Add function equivalents to the various macros in asn1.h. The old
16066 macros are retained with an M_ prefix. Code inside the library can
16067 use the M_ macros. External code (including the openssl utility)
16068 should *NOT* in order to be "shared library friendly".
16072 * Add various functions that can check a certificate's extensions
16073 to see if it usable for various purposes such as SSL client,
16074 server or S/MIME and CAs of these types. This is currently
16075 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16076 verification. Also added a -purpose flag to x509 utility to
16077 print out all the purposes.
16081 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16086 * New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
16087 for, obtain and decode and extension and obtain its critical flag.
16088 This allows all the necessary extension code to be handled in a
16089 single function call.
16093 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16094 platforms. See crypto/rc4/rc4_enc.c for further details.
16098 * New -noout option to asn1parse. This causes no output to be produced
16099 its main use is when combined with -strparse and -out to extract data
16100 from a file (which may not be in ASN.1 format).
16104 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16105 when producing the local key id.
16107 *Richard Levitte <levitte@stacken.kth.se>*
16109 * New option -dhparam in s_server. This allows a DH parameter file to be
16110 stated explicitly. If it is not stated then it tries the first server
16111 certificate file. The previous behaviour hard coded the filename
16116 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16117 a public key to be input or output. For example:
16118 openssl rsa -in key.pem -pubout -out pubkey.pem
16119 Also added necessary DSA public key functions to handle this.
16123 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16124 in the message. This was handled by allowing
16125 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16127 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16129 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16130 to the end of the strings whereas this didn't. This would cause problems
16131 if strings read with d2i_ASN1_bytes() were later modified.
16133 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16135 * Fix for base64 decode bug. When a base64 bio reads only one line of
16136 data and it contains EOF it will end up returning an error. This is
16137 caused by input 46 bytes long. The cause is due to the way base64
16138 BIOs find the start of base64 encoded data. They do this by trying a
16139 trial decode on each line until they find one that works. When they
16140 do a flag is set and it starts again knowing it can pass all the
16141 data directly through the decoder. Unfortunately it doesn't reset
16142 the context it uses. This means that if EOF is reached an attempt
16143 is made to pass two EOFs through the context and this causes the
16144 resulting error. This can also cause other problems as well. As is
16145 usual with these problems it takes *ages* to find and the fix is
16146 trivial: move one line.
16148 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) *
16150 * Ugly workaround to get s_client and s_server working under Windows. The
16151 old code wouldn't work because it needed to select() on sockets and the
16152 tty (for keypresses and to see if data could be written). Win32 only
16153 supports select() on sockets so we select() with a 1s timeout on the
16154 sockets and then see if any characters are waiting to be read, if none
16155 are present then we retry, we also assume we can always write data to
16156 the tty. This isn't nice because the code then blocks until we've
16157 received a complete line of data and it is effectively polling the
16158 keyboard at 1s intervals: however it's quite a bit better than not
16159 working at all :-) A dedicated Windows application might handle this
16160 with an event loop for example.
16164 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16165 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16166 will be called when RSA_sign() and RSA_verify() are used. This is useful
16167 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16168 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16169 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16170 This necessitated the support of an extra signature type NID_md5_sha1
16171 for SSL signatures and modifications to the SSL library to use it instead
16172 of calling RSA_public_decrypt() and RSA_private_encrypt().
16176 * Add new -verify -CAfile and -CApath options to the crl program, these
16177 will lookup a CRL issuers certificate and verify the signature in a
16178 similar way to the verify program. Tidy up the crl program so it
16179 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16180 less strict. It will now permit CRL extensions even if it is not
16181 a V2 CRL: this will allow it to tolerate some broken CRLs.
16185 * Initialize all non-automatic variables each time one of the openssl
16186 sub-programs is started (this is necessary as they may be started
16187 multiple times from the "OpenSSL>" prompt).
16189 *Lennart Bang, Bodo Moeller*
16191 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16192 removing all other RSA functionality (this is what NO_RSA does). This
16193 is so (for example) those in the US can disable those operations covered
16194 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16199 * Non-copying interface to BIO pairs.
16200 (still largely untested)
16204 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16205 ASCII string. This was handled independently in various places before.
16209 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16210 UTF8 strings a character at a time.
16214 * Use client_version from client hello to select the protocol
16215 (s23_srvr.c) and for RSA client key exchange verification
16216 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16220 * Add various utility functions to handle SPKACs, these were previously
16221 handled by poking round in the structure internals. Added new function
16222 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16223 print, verify and generate SPKACs. Based on an original idea from
16224 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16228 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16232 * Allow the config file extension section to be overwritten on the
16233 command line. Based on an original idea from Massimiliano Pala
16234 <madwolf@comune.modena.it>. The new option is called -extensions
16235 and can be applied to ca, req and x509. Also -reqexts to override
16236 the request extensions in req and -crlexts to override the crl extensions
16241 * Add new feature to the SPKAC handling in ca. Now you can include
16242 the same field multiple times by preceding it by "XXXX." for example:
16245 this is the same syntax as used in the req config file.
16249 * Allow certificate extensions to be added to certificate requests. These
16250 are specified in a 'req_extensions' option of the req section of the
16251 config file. They can be printed out with the -text option to req but
16252 are otherwise ignored at present.
16256 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16257 data read consists of only the final block it would not decrypted because
16258 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16259 A misplaced 'break' also meant the decrypted final block might not be
16260 copied until the next read.
16264 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16265 a few extra parameters to the DH structure: these will be useful if
16266 for example we want the value of 'q' or implement X9.42 DH.
16270 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16271 provides hooks that allow the default DSA functions or functions on a
16272 "per key" basis to be replaced. This allows hardware acceleration and
16273 hardware key storage to be handled without major modification to the
16274 library. Also added low level modexp hooks and CRYPTO_EX structure and
16275 associated functions.
16279 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16280 as "read only": it can't be written to and the buffer it points to will
16281 not be freed. Reading from a read only BIO is much more efficient than
16282 a normal memory BIO. This was added because there are several times when
16283 an area of memory needs to be read from a BIO. The previous method was
16284 to create a memory BIO and write the data to it, this results in two
16285 copies of the data and an O(n^2) reading algorithm. There is a new
16286 function BIO_new_mem_buf() which creates a read only memory BIO from
16287 an area of memory. Also modified the PKCS#7 routines to use read only
16292 * Bugfix: ssl23_get_client_hello did not work properly when called in
16293 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16294 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16295 but a retry condition occurred while trying to read the rest.
16299 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16300 NID_pkcs7_encrypted by default: this was wrong since this should almost
16301 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16302 the encrypted data type: this is a more sensible place to put it and it
16303 allows the PKCS#12 code to be tidied up that duplicated this
16308 * Changed obj_dat.pl script so it takes its input and output files on
16309 the command line. This should avoid shell escape redirection problems
16314 * Initial support for certificate extension requests, these are included
16315 in things like Xenroll certificate requests. Included functions to allow
16316 extensions to be obtained and added.
16320 * -crlf option to s_client and s_server for sending newlines as
16321 CRLF (as required by many protocols).
16325 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999] ###
16327 * Install libRSAglue.a when OpenSSL is built with RSAref.
16329 *Ralf S. Engelschall*
16331 * A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
16333 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16335 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16340 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16341 DH parameters/keys (q is lost during that conversion, but the resulting
16342 DH parameters contain its length).
16344 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16345 much faster than DH_generate_parameters (which creates parameters
16346 where p = 2*q + 1), and also the smaller q makes DH computations
16347 much more efficient (160-bit exponentiation instead of 1024-bit
16348 exponentiation); so this provides a convenient way to support DHE
16349 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16350 utter importance to use
16351 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16353 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16354 when such DH parameters are used, because otherwise small subgroup
16355 attacks may become possible!
16359 * Avoid memory leak in i2d_DHparams.
16363 * Allow the -k option to be used more than once in the enc program:
16364 this allows the same encrypted message to be read by multiple recipients.
16368 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16369 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16370 it will always use the numerical form of the OID, even if it has a short
16375 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16376 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16377 otherwise bn_mod_exp was called. In the case of hardware keys for example
16378 no private key components need be present and it might store extra data
16379 in the RSA structure, which cannot be accessed from bn_mod_exp.
16380 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16381 private key operations.
16385 * Added support for SPARC Linux.
16389 * pem_password_cb function type incompatibly changed from
16390 typedef int pem_password_cb(char *buf, int size, int rwflag);
16392 ....(char *buf, int size, int rwflag, void *userdata);
16393 so that applications can pass data to their callbacks:
16394 The PEM[_ASN1]_{read,write}... functions and macros now take an
16395 additional void * argument, which is just handed through whenever
16396 the password callback is called.
16398 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16400 New function SSL_CTX_set_default_passwd_cb_userdata.
16402 Compatibility note: As many C implementations push function arguments
16403 onto the stack in reverse order, the new library version is likely to
16404 interoperate with programs that have been compiled with the old
16405 pem_password_cb definition (PEM_whatever takes some data that
16406 happens to be on the stack as its last argument, and the callback
16407 just ignores this garbage); but there is no guarantee whatsoever that
16410 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16411 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16412 problems not only on Windows, but also on some Unix platforms.
16413 To avoid problematic command lines, these definitions are now in an
16414 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16415 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16419 * MIPS III/IV assembler module is reimplemented.
16423 * More DES library cleanups: remove references to srand/rand and
16424 delete an unused file.
16428 * Add support for the free Netwide assembler (NASM) under Win32,
16429 since not many people have MASM (ml) and it can be hard to obtain.
16430 This is currently experimental but it seems to work OK and pass all
16431 the tests. Check out INSTALL.W32 for info.
16435 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16436 without temporary keys kept an extra copy of the server key,
16437 and connections with temporary keys did not free everything in case
16442 * New function RSA_check_key and new openssl rsa option -check
16443 for verifying the consistency of RSA keys.
16445 *Ulf Moeller, Bodo Moeller*
16447 * Various changes to make Win32 compile work:
16448 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16449 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16450 comparison" warnings.
16451 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
16455 * Add a debugging option to PKCS#5 v2 key generation function: when
16456 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16457 derived keys are printed to stderr.
16461 * Copy the flags in ASN1_STRING_dup().
16463 *Roman E. Pavlov <pre@mo.msk.ru>*
16465 * The x509 application mishandled signing requests containing DSA
16466 keys when the signing key was also DSA and the parameters didn't match.
16468 It was supposed to omit the parameters when they matched the signing key:
16469 the verifying software was then supposed to automatically use the CA's
16470 parameters if they were absent from the end user certificate.
16472 Omitting parameters is no longer recommended. The test was also
16473 the wrong way round! This was probably due to unusual behaviour in
16474 EVP_cmp_parameters() which returns 1 if the parameters match.
16475 This meant that parameters were omitted when they *didn't* match and
16476 the certificate was useless. Certificates signed with 'ca' didn't have
16479 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16481 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16482 The interface is as follows:
16483 Applications can use
16484 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16485 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16486 "off" is now the default.
16487 The library internally uses
16488 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16489 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16490 to disable memory-checking temporarily.
16492 Some inconsistent states that previously were possible (and were
16493 even the default) are now avoided.
16495 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16496 with each memory chunk allocated; this is occasionally more helpful
16497 than just having a counter.
16499 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16501 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16506 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16507 which largely parallels "options", but is for changing API behaviour,
16508 whereas "options" are about protocol behaviour.
16509 Initial "mode" flags are:
16511 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16512 a single record has been written.
16513 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16514 retries use the same buffer location.
16515 (But all of the contents must be
16520 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16523 * Fix problems with no-hmac etc.
16525 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16527 * New functions RSA_get_default_method(), RSA_set_method() and
16528 RSA_get_method(). These allows replacement of RSA_METHODs without having
16529 to mess around with the internals of an RSA structure.
16533 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16534 Also really enable memory leak checks in openssl.c and in some
16537 *Chad C. Mulligan, Bodo Moeller*
16539 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16540 up the length of negative integers. This has now been simplified to just
16541 store the length when it is first determined and use it later, rather
16542 than trying to keep track of where data is copied and updating it to
16544 *Steve Henson, reported by Brien Wheeler
16545 <bwheeler@authentica-security.com>*
16547 * Add a new function PKCS7_signatureVerify. This allows the verification
16548 of a PKCS#7 signature but with the signing certificate passed to the
16549 function itself. This contrasts with PKCS7_dataVerify which assumes the
16550 certificate is present in the PKCS#7 structure. This isn't always the
16551 case: certificates can be omitted from a PKCS#7 structure and be
16552 distributed by "out of band" means (such as a certificate database).
16556 * Complete the PEM_* macros with DECLARE_PEM versions to replace the
16557 function prototypes in pem.h, also change util/mkdef.pl to add the
16558 necessary function names.
16562 * mk1mf.pl (used by Windows builds) did not properly read the
16563 options set by Configure in the top level Makefile, and Configure
16564 was not even able to write more than one option correctly.
16565 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16569 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16570 file to be loaded from a BIO or FILE pointer. The BIO version will
16571 for example allow memory BIOs to contain config info.
16575 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16576 Whoever hopes to achieve shared-library compatibility across versions
16577 must use this, not the compile-time macro.
16578 (Exercise 0.9.4: Which is the minimum library version required by
16580 Note: All this applies only to multi-threaded programs, others don't
16585 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16586 through a BIO pair triggered the default case, i.e.
16587 SSLerr(...,SSL_R_UNKNOWN_STATE).
16591 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16592 can use the SSL library even if none of the specific BIOs is
16597 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16598 for the encoded length.
16600 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16602 * Add initial documentation of the X509V3 functions.
16606 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16607 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16608 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16609 secure PKCS#8 private key format with a high iteration count.
16613 * Fix determination of Perl interpreter: A perl or perl5
16614 _directory_ in $PATH was also accepted as the interpreter.
16616 *Ralf S. Engelschall*
16618 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16619 wrong with it but it was very old and did things like calling
16620 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16621 unusual formatting.
16625 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16626 to use the new extension code.
16630 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16631 with macros. This should make it easier to change their form, add extra
16632 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16637 * Add to configuration table a new entry that can specify an alternative
16638 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16639 according to Mark Crispin <MRC@Panda.COM>.
16644 * DES CBC did not update the IV. Weird.
16648 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16649 Changing the behaviour of the former might break existing programs --
16650 where IV updating is needed, des_ncbc_encrypt can be used.
16653 * When bntest is run from "make test" it drives bc to check its
16654 calculations, as well as internally checking them. If an internal check
16655 fails, it needs to cause bc to give a non-zero result or make test carries
16656 on without noticing the failure. Fixed.
16660 * DES library cleanups.
16664 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16665 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16666 ciphers. NOTE: although the key derivation function has been verified
16667 against some published test vectors it has not been extensively tested
16668 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16673 * Instead of "mkdir -p", which is not fully portable, use new
16674 Perl script "util/mkdir-p.pl".
16678 * Rewrite the way password based encryption (PBE) is handled. It used to
16679 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16680 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16681 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16682 the 'parameter' field of the AlgorithmIdentifier is passed to the
16683 underlying key generation function so it must do its own ASN1 parsing.
16684 This has also changed the EVP_PBE_CipherInit() function which now has a
16685 'parameter' argument instead of literal salt and iteration count values
16686 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16690 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16691 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16692 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16693 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16694 value was just used as a "magic string" and not used directly its
16695 value doesn't matter.
16699 * Introduce some semblance of const correctness to BN. Shame C doesn't
16704 * "linux-sparc64" configuration (ultrapenguin).
16706 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16707 "linux-sparc" configuration.
16709 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16711 * config now generates no-xxx options for missing ciphers.
16715 * Support the EBCDIC character set (work in progress).
16716 File ebcdic.c not yet included because it has a different license.
16718 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16720 * Support BS2000/OSD-POSIX.
16722 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16724 * Make callbacks for key generation use void * instead of char *.
16728 * Make S/MIME samples compile (not yet tested).
16732 * Additional typesafe stacks.
16736 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16741 ### Changes between 0.9.3 and 0.9.3a [29 May 1999] ###
16743 * New configuration variant "sco5-gcc".
16745 * Updated some demos.
16747 *Sean O Riordain, Wade Scholine*
16749 * Add missing BIO_free at exit of pkcs12 application.
16753 * Fix memory leak in conf.c.
16757 * Updates for Win32 to assembler version of MD5.
16761 * Set #! path to perl in apps/der_chop to where we found it
16762 instead of using a fixed path.
16766 * SHA library changes for irix64-mips4-cc.
16770 * Improvements for VMS support.
16775 ### Changes between 0.9.2b and 0.9.3 [24 May 1999] ###
16777 * Bignum library bug fix. IRIX 6 passes "make test" now!
16778 This also avoids the problems with SC4.2 and unpatched SC5.
16780 *Andy Polyakov <appro@fy.chalmers.se>*
16782 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16783 These are required because of the typesafe stack would otherwise break
16784 existing code. If old code used a structure member which used to be STACK
16785 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16786 sk_num or sk_value it would produce an error because the num, data members
16787 are not present in STACK_OF. Now it just produces a warning. sk_set
16788 replaces the old method of assigning a value to sk_value
16789 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16790 that does this will no longer work (and should use sk_set instead) but
16791 this could be regarded as a "questionable" behaviour anyway.
16795 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16796 correctly handle encrypted S/MIME data.
16800 * Change type of various DES function arguments from des_cblock
16801 (which means, in function argument declarations, pointer to char)
16802 to des_cblock * (meaning pointer to array with 8 char elements),
16803 which allows the compiler to do more typechecking; it was like
16804 that back in SSLeay, but with lots of ugly casts.
16806 Introduce new type const_des_cblock.
16810 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16811 problems: find RecipientInfo structure that matches recipient certificate
16812 and initialise the ASN1 structures properly based on passed cipher.
16816 * Belatedly make the BN tests actually check the results.
16820 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16821 to and from BNs: it was completely broken. New compilation option
16822 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16823 key elements as negative integers.
16827 * Reorganize and speed up MD5.
16829 *Andy Polyakov <appro@fy.chalmers.se>*
16833 *Richard Levitte <richard@levitte.org>*
16835 * New option -out to asn1parse to allow the parsed structure to be
16836 output to a file. This is most useful when combined with the -strparse
16837 option to examine the output of things like OCTET STRINGS.
16841 * Make SSL library a little more fool-proof by not requiring any longer
16842 that SSL_set_{accept,connect}_state be called before
16843 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
16844 in many applications because usually everything *appeared* to work as
16845 intended anyway -- now it really works as intended).
16849 * Move openssl.cnf out of lib/.
16853 * Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
16854 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16855 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
16857 *Ralf S. Engelschall*
16859 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16860 handle PKCS#7 enveloped data properly.
16862 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16864 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16865 copying pointers. The cert_st handling is changed by this in
16866 various ways (and thus what used to be known as ctx->default_cert
16867 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
16868 any longer when s->cert does not give us what we need).
16869 ssl_cert_instantiate becomes obsolete by this change.
16870 As soon as we've got the new code right (possibly it already is?),
16871 we have solved a couple of bugs of the earlier code where s->cert
16872 was used as if it could not have been shared with other SSL structures.
16874 Note that using the SSL API in certain dirty ways now will result
16875 in different behaviour than observed with earlier library versions:
16876 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
16877 does not influence s as it used to.
16879 In order to clean up things more thoroughly, inside SSL_SESSION
16880 we don't use CERT any longer, but a new structure SESS_CERT
16881 that holds per-session data (if available); currently, this is
16882 the peer's certificate chain and, for clients, the server's certificate
16883 and temporary key. CERT holds only those values that can have
16884 meaningful defaults in an SSL_CTX.
16888 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16889 from the internal representation. Various PKCS#7 fixes: remove some
16890 evil casts and set the enc_dig_alg field properly based on the signing
16895 * Allow PKCS#12 password to be set from the command line or the
16896 environment. Let 'ca' get its config file name from the environment
16897 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16902 * Allow certificate policies extension to use an IA5STRING for the
16903 organization field. This is contrary to the PKIX definition but
16904 VeriSign uses it and IE5 only recognises this form. Document 'x509'
16909 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
16910 without disallowing inline assembler and the like for non-pedantic builds.
16914 * Support Borland C++ builder.
16916 *Janez Jere <jj@void.si>, modified by Ulf Möller*
16922 * SHA-1 cleanups and performance enhancements.
16924 *Andy Polyakov <appro@fy.chalmers.se>*
16926 * Sparc v8plus assembler for the bignum library.
16928 *Andy Polyakov <appro@fy.chalmers.se>*
16930 * Accept any -xxx and +xxx compiler options in Configure.
16934 * Update HPUX configuration.
16938 * Add missing sk_<type>_unshift() function to safestack.h
16940 *Ralf S. Engelschall*
16942 * New function SSL_CTX_use_certificate_chain_file that sets the
16943 "extra_cert"s in addition to the certificate. (This makes sense
16944 only for "PEM" format files, as chains as a whole are not
16949 * Support verify_depth from the SSL API.
16950 x509_vfy.c had what can be considered an off-by-one-error:
16951 Its depth (which was not part of the external interface)
16952 was actually counting the number of certificates in a chain;
16953 now it really counts the depth.
16957 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
16958 instead of X509err, which often resulted in confusing error
16959 messages since the error codes are not globally unique
16960 (e.g. an alleged error in ssl3_accept when a certificate
16961 didn't match the private key).
16963 * New function SSL_CTX_set_session_id_context that allows to set a default
16964 value (so that you don't need SSL_set_session_id_context for each
16965 connection using the SSL_CTX).
16969 * OAEP decoding bug fix.
16973 * Support INSTALL_PREFIX for package builders, as proposed by
16978 * New Configure options "threads" and "no-threads". For systems
16979 where the proper compiler options are known (currently Solaris
16980 and Linux), "threads" is the default.
16984 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
16988 * Install various scripts to $(OPENSSLDIR)/misc, not to
16989 $(INSTALLTOP)/bin -- they shouldn't clutter directories
16990 such as /usr/local/bin.
16994 * "make linux-shared" to build shared libraries.
16996 *Niels Poppe <niels@netbox.org>*
16998 * New Configure option no-<cipher> (rsa, idea, rc5, ...).
17002 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17003 extension adding in x509 utility.
17007 * Remove NOPROTO sections and error code comments.
17011 * Partial rewrite of the DEF file generator to now parse the ANSI
17016 * New Configure options --prefix=DIR and --openssldir=DIR.
17020 * Complete rewrite of the error code script(s). It is all now handled
17021 by one script at the top level which handles error code gathering,
17022 header rewriting and C source file generation. It should be much better
17023 than the old method: it now uses a modified version of Ulf's parser to
17024 read the ANSI prototypes in all header files (thus the old K&R definitions
17025 aren't needed for error creation any more) and do a better job of
17026 translating function codes into names. The old 'ASN1 error code embedded
17027 in a comment' is no longer necessary and it doesn't use .err files which
17028 have now been deleted. Also the error code call doesn't have to appear all
17029 on one line (which resulted in some large lines...).
17033 * Change #include filenames from <foo.h> to <openssl/foo.h>.
17037 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17038 0 (which usually indicates a closed connection), but continue reading.
17042 * Fix some race conditions.
17046 * Add support for CRL distribution points extension. Add Certificate
17047 Policies and CRL distribution points documentation.
17051 * Move the autogenerated header file parts to crypto/opensslconf.h.
17055 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17056 8 of keying material. Merlin has also confirmed interop with this fix
17057 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17059 *Merlin Hughes <merlin@baltimore.ie>*
17061 * Fix lots of warnings.
17063 *Richard Levitte <levitte@stacken.kth.se>*
17065 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17066 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17068 *Richard Levitte <levitte@stacken.kth.se>*
17070 * Fix problems with sizeof(long) == 8.
17072 *Andy Polyakov <appro@fy.chalmers.se>*
17074 * Change functions to ANSI C.
17078 * Fix typos in error codes.
17080 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17082 * Remove defunct assembler files from Configure.
17086 * SPARC v8 assembler BIGNUM implementation.
17088 *Andy Polyakov <appro@fy.chalmers.se>*
17090 * Support for Certificate Policies extension: both print and set.
17091 Various additions to support the r2i method this uses.
17095 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17096 return a const string when you are expecting an allocated buffer.
17100 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17101 types DirectoryString and DisplayText.
17105 * Add code to allow r2i extensions to access the configuration database,
17106 add an LHASH database driver and add several ctx helper functions.
17110 * Fix an evil bug in bn_expand2() which caused various BN functions to
17111 fail when they extended the size of a BIGNUM.
17115 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17116 support typesafe stack.
17120 * Fix typo in SSL_[gs]et_options().
17122 *Nils Frostberg <nils@medcom.se>*
17124 * Delete various functions and files that belonged to the (now obsolete)
17125 old X509V3 handling code.
17129 * New Configure option "rsaref".
17133 * Don't auto-generate pem.h.
17137 * Introduce type-safe ASN.1 SETs.
17141 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17143 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17145 * Introduce type-safe STACKs. This will almost certainly break lots of code
17146 that links with OpenSSL (well at least cause lots of warnings), but fear
17147 not: the conversion is trivial, and it eliminates loads of evil casts. A
17148 few STACKed things have been converted already. Feel free to convert more.
17149 In the fullness of time, I'll do away with the STACK type altogether.
17153 * Add `openssl ca -revoke <certfile>' facility which revokes a certificate
17154 specified in <certfile> by updating the entry in the index.txt file.
17155 This way one no longer has to edit the index.txt file manually for
17156 revoking a certificate. The -revoke option does the gory details now.
17158 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17160 * Fix `openssl crl -noout -text' combination where `-noout' killed the
17161 `-text' option at all and this way the `-noout -text' combination was
17162 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
17164 *Ralf S. Engelschall*
17166 * Make sure a corresponding plain text error message exists for the
17167 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17168 verify callback function determined that a certificate was revoked.
17170 *Ralf S. Engelschall*
17172 * Bugfix: In test/testenc, don't test "openssl <cipher>" for
17173 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17174 all available ciphers including rc5, which was forgotten until now.
17175 In order to let the testing shell script know which algorithms
17176 are available, a new (up to now undocumented) command
17177 "openssl list-cipher-commands" is used.
17181 * Bugfix: s_client occasionally would sleep in select() when
17182 it should have checked SSL_pending() first.
17186 * New functions DSA_do_sign and DSA_do_verify to provide access to
17187 the raw DSA values prior to ASN.1 encoding.
17191 * Tweaks to Configure
17193 *Niels Poppe <niels@netbox.org>*
17195 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17200 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17204 * New config option to avoid instructions that are illegal on the 80386.
17205 The default code is faster, but requires at least a 486.
17209 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17210 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17211 same as SSL2_VERSION anyway.
17215 * New "-showcerts" option for s_client.
17219 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17220 application. Various cleanups and fixes.
17224 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17225 modify error routines to work internally. Add error codes and PBE init
17226 to library startup routines.
17230 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17231 packing functions to asn1 and evp. Changed function names and error
17232 codes along the way.
17236 * PKCS12 integration: and so it begins... First of several patches to
17237 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17238 objects to objects.h
17242 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17243 and display support for Thawte strong extranet extension.
17247 * Add LinuxPPC support.
17249 *Jeff Dubrule <igor@pobox.org>*
17251 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17252 bn_div_words in alpha.s.
17254 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17256 * Make sure the RSA OAEP test is skipped under -DRSAref because
17257 OAEP isn't supported when OpenSSL is built with RSAref.
17259 *Ulf Moeller <ulf@fitug.de>*
17261 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17262 so they no longer are missing under -DNOPROTO.
17264 *Soren S. Jorvang <soren@t.dk>*
17267 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999] ###
17269 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17270 doesn't work when the session is reused. Coming soon!
17274 * Fix a security hole, that allows sessions to be reused in the wrong
17275 context thus bypassing client cert protection! All software that uses
17276 client certs and session caches in multiple contexts NEEDS PATCHING to
17277 allow session reuse! A fuller solution is in the works.
17279 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17281 * Some more source tree cleanups (removed obsolete files
17282 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17283 permission on "config" script to be executable) and a fix for the INSTALL
17286 *Ulf Moeller <ulf@fitug.de>*
17288 * Remove some legacy and erroneous uses of malloc, free instead of
17291 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17293 * Make rsa_oaep_test return non-zero on error.
17295 *Ulf Moeller <ulf@fitug.de>*
17297 * Add support for native Solaris shared libraries. Configure
17298 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17299 if someone would make that last step automatic.
17301 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17303 * ctx_size was not built with the right compiler during "make links". Fixed.
17307 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17308 except NULL ciphers". This means the default cipher list will no longer
17309 enable NULL ciphers. They need to be specifically enabled e.g. with
17310 the string "DEFAULT:eNULL".
17314 * Fix to RSA private encryption routines: if p < q then it would
17315 occasionally produce an invalid result. This will only happen with
17316 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17320 * Be less restrictive and allow also `perl util/perlpath.pl
17321 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
17322 because this way one can also use an interpreter named `perl5' (which is
17323 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17324 installed as `perl').
17326 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17328 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17330 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17332 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17333 advapi32.lib to Win32 build and change the pem test comparison
17334 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17335 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17336 and crypto/des/ede_cbcm_enc.c.
17340 * DES quad checksum was broken on big-endian architectures. Fixed.
17344 * Comment out two functions in bio.h that aren't implemented. Fix up the
17345 Win32 test batch file so it (might) work again. The Win32 test batch file
17346 is horrible: I feel ill....
17350 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17351 in e_os.h. Audit of header files to check ANSI and non ANSI
17352 sections: 10 functions were absent from non ANSI section and not exported
17353 from Windows DLLs. Fixed up libeay.num for new functions.
17357 * Make `openssl version' output lines consistent.
17359 *Ralf S. Engelschall*
17361 * Fix Win32 symbol export lists for BIO functions: Added
17362 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17363 to ms/libeay{16,32}.def.
17365 *Ralf S. Engelschall*
17367 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17368 fine under Unix and passes some trivial tests I've now added. But the
17369 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17370 added to make sure no one expects that this stuff really works in the
17371 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17372 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17375 *Ralf S. Engelschall*
17377 * Fix the generation of two part addresses in perl.
17379 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17381 * Add config entry for Linux on MIPS.
17383 *John Tobey <jtobey@channel1.com>*
17385 * Make links whenever Configure is run, unless we are on Windoze.
17389 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17390 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17395 * Add a useful kludge to allow package maintainers to specify compiler and
17396 other platforms details on the command line without having to patch the
17397 Configure script every time: One now can use ``perl Configure
17398 <id>:<details>'', i.e. platform ids are allowed to have details appended
17399 to them (separated by colons). This is treated as there would be a static
17400 pre-configured entry in Configure's %table under key <id> with value
17401 <details> and ``perl Configure <id>'' is called. So, when you want to
17402 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17403 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
17404 now, which overrides the FreeBSD-elf entry on-the-fly.
17406 *Ralf S. Engelschall*
17408 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17412 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17413 on the `perl Configure ...' command line. This way one can compile
17414 OpenSSL libraries with Position Independent Code (PIC) which is needed
17415 for linking it into DSOs.
17417 *Ralf S. Engelschall*
17419 * Remarkably, export ciphers were totally broken and no-one had noticed!
17424 * Cleaned up the LICENSE document: The official contact for any license
17425 questions now is the OpenSSL core team under openssl-core@openssl.org.
17426 And add a paragraph about the dual-license situation to make sure people
17427 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17428 to the OpenSSL toolkit.
17430 *Ralf S. Engelschall*
17432 * General source tree makefile cleanups: Made `making xxx in yyy...'
17433 display consistent in the source tree and replaced `/bin/rm' by `rm'.
17434 Additionally cleaned up the `make links' target: Remove unnecessary
17435 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17436 to speed processing and no longer clutter the display with confusing
17437 stuff. Instead only the actually done links are displayed.
17439 *Ralf S. Engelschall*
17441 * Permit null encryption ciphersuites, used for authentication only. It used
17442 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17443 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17448 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17449 signed attributes when verifying signatures (this would break them),
17450 the detached data encoding was wrong and public keys obtained using
17451 X509_get_pubkey() weren't freed.
17455 * Add text documentation for the BUFFER functions. Also added a work around
17456 to a Win95 console bug. This was triggered by the password read stuff: the
17457 last character typed gets carried over to the next fread(). If you were
17458 generating a new cert request using 'req' for example then the last
17459 character of the passphrase would be CR which would then enter the first
17464 * Added the new `Includes OpenSSL Cryptography Software' button as
17465 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17466 button and can be used by applications based on OpenSSL to show the
17467 relationship to the OpenSSL project.
17469 *Ralf S. Engelschall*
17471 * Remove confusing variables in function signatures in files
17472 ssl/ssl_lib.c and ssl/ssl.h.
17474 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17476 * Don't install bss_file.c under PREFIX/include/
17478 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17480 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17481 functions that return function pointers and has support for NT specific
17482 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17483 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17484 unsigned to signed types: this was killing the Win32 compile.
17488 * Add new certificate file to stack functions,
17489 SSL_add_dir_cert_subjects_to_stack() and
17490 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17491 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17492 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17493 This means that Apache-SSL and similar packages don't have to mess around
17494 to add as many CAs as they want to the preferred list.
17498 * Experiment with doxygen documentation. Currently only partially applied to
17500 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
17501 openssl.doxy as the configuration file.
17505 * Get rid of remaining C++-style comments which strict C compilers hate.
17507 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17509 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17510 compiled in by default: it has problems with large keys.
17514 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17515 DH private keys and/or callback functions which directly correspond to
17516 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17517 is needed for applications which have to configure certificates on a
17518 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17520 For the RSA certificate situation is makes no difference, but
17521 for the DSA certificate situation this fixes the "no shared cipher"
17522 problem where the OpenSSL cipher selection procedure failed because the
17523 temporary keys were not overtaken from the context and the API provided
17524 no way to reconfigure them.
17525 The new functions now let applications reconfigure the stuff and they
17526 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17527 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17528 non-public-API function ssl_cert_instantiate() is used as a helper
17529 function and also to reduce code redundancy inside ssl_rsa.c.
17531 *Ralf S. Engelschall*
17533 * Move s_server -dcert and -dkey options out of the undocumented feature
17534 area because they are useful for the DSA situation and should be
17535 recognized by the users.
17537 *Ralf S. Engelschall*
17539 * Fix the cipher decision scheme for export ciphers: the export bits are
17540 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17541 SSL_EXP_MASK. So, the original variable has to be used instead of the
17542 already masked variable.
17544 *Richard Levitte <levitte@stacken.kth.se>*
17546 * Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
17548 *Richard Levitte <levitte@stacken.kth.se>*
17550 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17551 from `int' to `unsigned int' because it's a length and initialized by
17552 EVP_DigestFinal() which expects an `unsigned int *'.
17554 *Richard Levitte <levitte@stacken.kth.se>*
17556 * Don't hard-code path to Perl interpreter on shebang line of Configure
17557 script. Instead use the usual Shell->Perl transition trick.
17559 *Ralf S. Engelschall*
17561 * Make `openssl x509 -noout -modulus' functional also for DSA certificates
17562 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17563 -noout -modulus' as it's already the case for `openssl rsa -noout
17564 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
17565 currently the public key is printed (a decision which was already done by
17566 `openssl dsa -modulus' in the past) which serves a similar purpose.
17567 Additionally the NO_RSA no longer completely removes the whole -modulus
17568 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17571 *Ralf S. Engelschall*
17573 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17574 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17576 *Arne Ansper <arne@ats.cyber.ee>*
17578 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17579 to be added. Now both 'req' and 'ca' can use new objects defined in the
17584 * Add cool BIO that does syslog (or event log on NT).
17586 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17588 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17589 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17590 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17591 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17595 * Add preliminary config info for new extension code.
17599 * Make RSA_NO_PADDING really use no padding.
17601 *Ulf Moeller <ulf@fitug.de>*
17603 * Generate errors when private/public key check is done.
17607 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17608 for some CRL extensions and new objects added.
17612 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17613 key usage extension and fuller support for authority key id.
17617 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17618 padding method for RSA, which is recommended for new applications in PKCS
17619 #1 v2.0 (RFC 2437, October 1998).
17620 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17621 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17622 against Bleichbacher's attack on RSA.
17623 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17626 * Updates to the new SSL compression code
17628 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17630 * Fix so that the version number in the master secret, when passed
17631 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17632 (because the server will not accept higher), that the version number
17633 is 0x03,0x01, not 0x03,0x00
17635 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17637 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17638 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17639 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17643 * Support for RAW extensions where an arbitrary extension can be
17644 created by including its DER encoding. See apps/openssl.cnf for
17649 * Make sure latest Perl versions don't interpret some generated C array
17650 code as Perl array code in the crypto/err/err_genc.pl script.
17652 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17654 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17655 not many people have the assembler. Various Win32 compilation fixes and
17656 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17657 build instructions.
17661 * Modify configure script 'Configure' to automatically create crypto/date.h
17662 file under Win32 and also build pem.h from pem.org. New script
17663 util/mkfiles.pl to create the MINFO file on environments that can't do a
17664 'make files': perl util/mkfiles.pl >MINFO should work.
17668 * Major rework of DES function declarations, in the pursuit of correctness
17669 and purity. As a result, many evil casts evaporated, and some weirdness,
17670 too. You may find this causes warnings in your code. Zapping your evil
17671 casts will probably fix them. Mostly.
17675 * Fix for a typo in asn1.h. Bug fix to object creation script
17676 obj_dat.pl. It considered a zero in an object definition to mean
17677 "end of object": none of the objects in objects.h have any zeros
17678 so it wasn't spotted.
17680 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17682 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17683 Masking (CBCM). In the absence of test vectors, the best I have been able
17684 to do is check that the decrypt undoes the encrypt, so far. Send me test
17685 vectors if you have them.
17689 * Correct calculation of key length for export ciphers (too much space was
17690 allocated for null ciphers). This has not been tested!
17694 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17695 message is now correct (it understands "crypto" and "ssl" on its
17696 command line). There is also now an "update" option. This will update
17697 the util/ssleay.num and util/libeay.num files with any new functions.
17699 perl util/mkdef.pl crypto ssl update
17700 it will update them.
17704 * Overhauled the Perl interface (perl/*):
17705 - ported BN stuff to OpenSSL's different BN library
17706 - made the perl/ source tree CVS-aware
17707 - renamed the package from SSLeay to OpenSSL (the files still contain
17708 their history because I've copied them in the repository)
17709 - removed obsolete files (the test scripts will be replaced
17710 by better Test::Harness variants in the future)
17712 *Ralf S. Engelschall*
17714 * First cut for a very conservative source tree cleanup:
17715 1. merge various obsolete readme texts into doc/ssleay.txt
17716 where we collect the old documents and readme texts.
17717 2. remove the first part of files where I'm already sure that we no
17718 longer need them because of three reasons: either they are just temporary
17719 files which were left by Eric or they are preserved original files where
17720 I've verified that the diff is also available in the CVS via "cvs diff
17721 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17722 the crypto/md/ stuff).
17724 *Ralf S. Engelschall*
17726 * More extension code. Incomplete support for subject and issuer alt
17727 name, issuer and authority key id. Change the i2v function parameters
17728 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17729 what that's for :-) Fix to ASN1 macro which messed up
17730 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17734 * Preliminary support for ENUMERATED type. This is largely copied from the
17739 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17741 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17743 * Make sure `make rehash' target really finds the `openssl' program.
17745 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17747 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17748 like to hear about it if this slows down other processors.
17752 * Add CygWin32 platform information to Configure script.
17754 *Alan Batie <batie@aahz.jf.intel.com>*
17756 * Fixed ms/32all.bat script: `no_asm' -> `no-asm'
17758 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17760 * New program nseq to manipulate netscape certificate sequences
17764 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17769 * Fixes to BN code. Previously the default was to define BN_RECURSION
17770 but the BN code had some problems that would cause failures when
17771 doing certificate verification and some other functions.
17773 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17775 * Add ASN1 and PEM code to support netscape certificate sequences.
17779 * Add ASN1 and PEM code to support netscape certificate sequences.
17783 * Add several PKIX and private extended key usage OIDs.
17787 * Modify the 'ca' program to handle the new extension code. Modify
17788 openssl.cnf for new extension format, add comments.
17792 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17793 and add a sample to openssl.cnf so req -x509 now adds appropriate
17798 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17799 error code, add initial support to X509_print() and x509 application.
17803 * Takes a deep breath and start adding X509 V3 extension support code. Add
17804 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17805 stuff is currently isolated and isn't even compiled yet.
17809 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17810 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17811 Removed the versions check from X509 routines when loading extensions:
17812 this allows certain broken certificates that don't set the version
17813 properly to be processed.
17817 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17818 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17819 can still be regenerated with "make depend".
17823 * Spelling mistake in C version of CAST-128.
17825 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17827 * Changes to the error generation code. The perl script err-code.pl
17828 now reads in the old error codes and retains the old numbers, only
17829 adding new ones if necessary. It also only changes the .err files if new
17830 codes are added. The makefiles have been modified to only insert errors
17831 when needed (to avoid needlessly modifying header files). This is done
17832 by only inserting errors if the .err file is newer than the auto generated
17833 C file. To rebuild all the error codes from scratch (the old behaviour)
17834 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17835 or delete all the .err files.
17839 * CAST-128 was incorrectly implemented for short keys. The C version has
17840 been fixed, but is untested. The assembler versions are also fixed, but
17841 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17842 to regenerate it if needed.
17843 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17844 Hagino <itojun@kame.net>*
17846 * File was opened incorrectly in randfile.c.
17848 *Ulf Möller <ulf@fitug.de>*
17850 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17851 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17852 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17853 al: it's just almost always a UTCTime. Note this patch adds new error
17854 codes so do a "make errors" if there are problems.
17858 * Correct Linux 1 recognition in config.
17860 *Ulf Möller <ulf@fitug.de>*
17862 * Remove pointless MD5 hash when using DSA keys in ca.
17864 *Anonymous <nobody@replay.com>*
17866 * Generate an error if given an empty string as a cert directory. Also
17867 generate an error if handed NULL (previously returned 0 to indicate an
17868 error, but didn't set one).
17870 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17872 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17876 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17877 parameters. This was causing a warning which killed off the Win32 compile.
17881 * Remove C++ style comments from crypto/bn/bn_local.h.
17883 *Neil Costigan <neil.costigan@celocom.com>*
17885 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17886 based on a text string, looking up short and long names and finally
17887 "dot" format. The "dot" format stuff didn't work. Added new function
17888 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17889 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17890 OID is not part of the table.
17894 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17895 X509_LOOKUP_by_alias().
17899 * Sort openssl functions by name.
17903 * Get the gendsa program working (hopefully) and add it to app list. Remove
17904 encryption from sample DSA keys (in case anyone is interested the password
17909 * Make _all_ *_free functions accept a NULL pointer.
17911 *Frans Heymans <fheymans@isaserver.be>*
17913 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
17916 *Anonymous <nobody@replay.com>*
17918 * s_server should send the CAfile as acceptable CAs, not its own cert.
17920 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17922 * Don't blow it for numeric -newkey arguments to apps/req.
17924 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17926 * Temp key "for export" tests were wrong in s3_srvr.c.
17928 *Anonymous <nobody@replay.com>*
17930 * Add prototype for temp key callback functions
17931 SSL_CTX_set_tmp_{rsa,dh}_callback().
17935 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
17936 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
17940 * X509_name_add_entry() freed the wrong thing after an error.
17942 *Arne Ansper <arne@ats.cyber.ee>*
17944 * rsa_eay.c would attempt to free a NULL context.
17946 *Arne Ansper <arne@ats.cyber.ee>*
17948 * BIO_s_socket() had a broken should_retry() on Windoze.
17950 *Arne Ansper <arne@ats.cyber.ee>*
17952 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
17954 *Arne Ansper <arne@ats.cyber.ee>*
17956 * Make sure the already existing X509_STORE->depth variable is initialized
17957 in X509_STORE_new(), but document the fact that this variable is still
17958 unused in the certificate verification process.
17960 *Ralf S. Engelschall*
17962 * Fix the various library and apps files to free up pkeys obtained from
17963 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
17967 * Fix reference counting in X509_PUBKEY_get(). This makes
17968 demos/maurice/example2.c work, amongst others, probably.
17970 *Steve Henson and Ben Laurie*
17972 * First cut of a cleanup for apps/. First the `ssleay' program is now named
17973 `openssl' and second, the shortcut symlinks for the `openssl <command>'
17974 are no longer created. This way we have a single and consistent command
17975 line interface `openssl <command>', similar to `cvs <command>'.
17977 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
17979 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
17980 BIT STRING wrapper always have zero unused bits.
17984 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
17988 * Make the top-level INSTALL documentation easier to understand.
17992 * Makefiles updated to exit if an error occurs in a sub-directory
17993 make (including if user presses ^C) [Paul Sutton]
17995 * Make Montgomery context stuff explicit in RSA data structure.
17999 * Fix build order of pem and err to allow for generated pem.h.
18003 * Fix renumbering bug in X509_NAME_delete_entry().
18007 * Enhanced the err-ins.pl script so it makes the error library number
18008 global and can add a library name. This is needed for external ASN1 and
18009 other error libraries.
18013 * Fixed sk_insert which never worked properly.
18017 * Fix ASN1 macros so they can handle indefinite length constructed
18018 EXPLICIT tags. Some non standard certificates use these: they can now
18023 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18024 into a single doc/ssleay.txt bundle. This way the information is still
18025 preserved but no longer messes up this directory. Now it's new room for
18026 the new set of documentation files.
18028 *Ralf S. Engelschall*
18030 * SETs were incorrectly DER encoded. This was a major pain, because they
18031 shared code with SEQUENCEs, which aren't coded the same. This means that
18032 almost everything to do with SETs or SEQUENCEs has either changed name or
18033 number of arguments.
18035 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18037 * Fix test data to work with the above.
18041 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18042 was already fixed by Eric for 0.9.1 it seems.
18044 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18046 * Autodetect FreeBSD3.
18050 * Fix various bugs in Configure. This affects the following platforms:
18054 unixware-2.0-pentium
18059 * Eliminate generated files from CVS. Reorder tests to regenerate files
18060 before they are needed.
18064 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18069 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998] ###
18071 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18072 changed SSLeay to OpenSSL in version strings.
18074 *Ralf S. Engelschall*
18076 * Some fixups to the top-level documents.
18080 * Fixed the nasty bug where rsaref.h was not found under compile-time
18081 because the symlink to include/ was missing.
18083 *Ralf S. Engelschall*
18085 * Incorporated the popular no-RSA/DSA-only patches
18086 which allow to compile a RSA-free SSLeay.
18088 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18090 * Fixed nasty rehash problem under `make -f Makefile.ssl links'
18091 when "ssleay" is still not found.
18093 *Ralf S. Engelschall*
18095 * Added more platforms to Configure: Cray T3E, HPUX 11,
18097 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18099 * Updated the README file.
18101 *Ralf S. Engelschall*
18103 * Added various .cvsignore files in the CVS repository subdirs
18104 to make a "cvs update" really silent.
18106 *Ralf S. Engelschall*
18108 * Recompiled the error-definition header files and added
18109 missing symbols to the Win32 linker tables.
18111 *Ralf S. Engelschall*
18113 * Cleaned up the top-level documents;
18114 o new files: CHANGES and LICENSE
18115 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18116 o merged COPYRIGHT into LICENSE
18117 o removed obsolete TODO file
18118 o renamed MICROSOFT to INSTALL.W32
18120 *Ralf S. Engelschall*
18122 * Removed dummy files from the 0.9.1b source tree:
18123 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18124 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18125 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18126 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18127 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18129 *Ralf S. Engelschall*
18131 * Added various platform portability fixes.
18135 * The Genesis of the OpenSSL rpject:
18136 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18137 Young and Tim J. Hudson created while they were working for C2Net until
18140 *The OpenSSL Project*
18143 ### Changes between 0.9.0b and 0.9.1b [not released] ###
18145 * Updated a few CA certificates under certs/
18149 * Changed some BIGNUM api stuff.
18153 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18154 DGUX x86, Linux Alpha, etc.
18158 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18159 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18164 * Add -strparse option to asn1pars program which parses nested
18167 *Dr Stephen Henson <shenson@bigfoot.com>*
18169 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18173 * DSA fix for "ca" program.
18177 * Added "-genkey" option to "dsaparam" program.
18181 * Added RIPE MD160 (rmd160) message digest.
18185 * Added -a (all) option to "ssleay version" command.
18189 * Added PLATFORM define which is the id given to Configure.
18193 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18197 * Extended the ASN.1 parser routines.
18201 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18205 * Added a BN_CTX to the BN library.
18209 * Fixed the weak key values in DES library
18213 * Changed API in EVP library for cipher aliases.
18217 * Added support for RC2/64bit cipher.
18221 * Converted the lhash library to the crypto/mem.c functions.
18225 * Added more recognized ASN.1 object ids.
18229 * Added more RSA padding checks for SSL/TLS.
18233 * Added BIO proxy/filter functionality.
18237 * Added extra_certs to SSL_CTX which can be used
18238 send extra CA certificates to the client in the CA cert chain sending
18239 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18243 * Now Fortezza is denied in the authentication phase because
18244 this is key exchange mechanism is not supported by SSLeay at all.
18248 * Additional PKCS1 checks.
18252 * Support the string "TLSv1" for all TLS v1 ciphers.
18256 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18257 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18261 * Fixed a few memory leaks.
18265 * Fixed various code and comment typos.
18269 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18270 bytes sent in the client random.
18272 *Edward Bishop <ebishop@spyglass.com>*
18277 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18278 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18279 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18280 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18281 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18282 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18283 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18284 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18285 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18286 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18287 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18288 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18289 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18290 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18291 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18292 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18293 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18294 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18295 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18296 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18297 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18298 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18299 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18300 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18301 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18302 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18303 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18304 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18305 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18306 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18307 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18308 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18309 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18310 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18311 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18312 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18313 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18314 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18315 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18316 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18317 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18318 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18319 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18320 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18321 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18322 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18323 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18324 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18325 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18326 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18327 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18328 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18329 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18330 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18331 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18332 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18333 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18334 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18335 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18336 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18337 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18338 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18339 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18340 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18341 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18342 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18343 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18344 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18345 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18346 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18347 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18348 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18349 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18350 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18351 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18352 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18353 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18354 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18355 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18356 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18357 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18358 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18359 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18360 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18361 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18362 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18363 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18364 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18365 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18366 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18367 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18368 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18369 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18370 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18371 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18372 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18373 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18374 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18375 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18376 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18377 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18378 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18379 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18380 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18381 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18382 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18383 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18384 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18385 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18386 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18387 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18388 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18389 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18390 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18391 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18392 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18393 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18394 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18395 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18396 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18397 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18398 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18399 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18400 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18401 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18402 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18403 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18404 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18405 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18406 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18407 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18408 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18409 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18410 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18411 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18412 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18413 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18414 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18415 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18416 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18417 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18418 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18419 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18420 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18421 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18422 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18423 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18424 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18425 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18426 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18427 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18428 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18429 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18430 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18431 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18432 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18433 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18434 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18435 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18436 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655