5 This is a high-level summary of the most important changes.
6 For a full list of changes, see the git commit log; for example,
7 https://github.com/openssl/openssl/commits/ and pick the appropriate
10 Changes between 1.1.1d and 1.1.1e [xx XXX xxxx]
11 *) Properly detect EOF while reading in libssl. Previously if we hit an EOF
12 while reading in libssl then we would report an error back to the
13 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
14 an error to the stack (which means we instead return SSL_ERROR_SSL) and
15 therefore give a hint as to what went wrong.
18 *) Check that ed25519 and ed448 are allowed by the security level. Previously
19 signature algorithms not using an MD were not being checked that they were
20 allowed by the security level.
23 *) Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
24 was not quite right. The behaviour was not consistent between resumption
25 and normal handshakes, and also not quite consistent with historical
26 behaviour. The behaviour in various scenarios has been clarified and
27 it has been updated to make it match historical behaviour as closely as
31 *) [VMS only] The header files that the VMS compilers include automatically,
32 __DECC_INCLUDE_PROLOGUE.H and __DECC_INCLUDE_EPILOGUE.H, use pragmas that
33 the C++ compiler doesn't understand. This is a shortcoming in the
34 compiler, but can be worked around with __cplusplus guards.
36 C++ applications that use OpenSSL libraries must be compiled using the
37 qualifier '/NAMES=(AS_IS,SHORTENED)' to be able to use all the OpenSSL
38 functions. Otherwise, only functions with symbols of less than 31
39 characters can be used, as the linker will not be able to successfully
40 resolve symbols with longer names.
43 *) Corrected the documentation of the return values from the EVP_DigestSign*
44 set of functions. The documentation mentioned negative values for some
45 errors, but this was never the case, so the mention of negative values
48 Code that followed the documentation and thereby check with something
49 like 'EVP_DigestSignInit(...) <= 0' will continue to work undisturbed.
52 *) Fixed an an overflow bug in the x64_64 Montgomery squaring procedure
53 used in exponentiation with 512-bit moduli. No EC algorithms are
54 affected. Analysis suggests that attacks against 2-prime RSA1024,
55 3-prime RSA1536, and DSA1024 as a result of this defect would be very
56 difficult to perform and are not believed likely. Attacks against DH512
57 are considered just feasible. However, for an attack the target would
58 have to re-use the DH512 private key, which is not recommended anyway.
59 Also applications directly using the low level API BN_mod_exp may be
60 affected if they use BN_FLG_CONSTTIME.
64 *) Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
65 The presence of this system service is determined at run-time.
68 *) Added newline escaping functionality to a filename when using openssl dgst.
69 This output format is to replicate the output format found in the '*sum'
70 checksum programs. This aims to preserve backward compatibility.
71 [Matt Eaton, Richard Levitte, and Paul Dale]
73 *) Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
77 Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
79 *) Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
80 number generator (RNG). This was intended to include protection in the
81 event of a fork() system call in order to ensure that the parent and child
82 processes did not share the same RNG state. However this protection was not
83 being used in the default case.
85 A partial mitigation for this issue is that the output from a high
86 precision timer is mixed into the RNG state so the likelihood of a parent
87 and child process sharing state is significantly reduced.
89 If an application already calls OPENSSL_init_crypto() explicitly using
90 OPENSSL_INIT_ATFORK then this problem does not occur at all.
94 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is
95 used even when parsing explicit parameters, when loading a serialized key
96 or calling `EC_GROUP_new_from_ecpkparameters()`/
97 `EC_GROUP_new_from_ecparameters()`.
98 This prevents bypass of security hardening and performance gains,
99 especially for curves with specialized EC_METHODs.
100 By default, if a key encoded with explicit parameters is loaded and later
101 serialized, the output is still encoded with explicit parameters, even if
102 internally a "named" EC_GROUP is used for computation.
105 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before
106 this change, EC_GROUP_set_generator would accept order and/or cofactor as
107 NULL. After this change, only the cofactor parameter can be NULL. It also
108 does some minimal sanity checks on the passed order.
112 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
113 An attack is simple, if the first CMS_recipientInfo is valid but the
114 second CMS_recipientInfo is chosen ciphertext. If the second
115 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
116 encryption key will be replaced by garbage, and the message cannot be
117 decoded, but if the RSA decryption fails, the correct encryption key is
118 used and the recipient will not notice the attack.
119 As a work around for this potential attack the length of the decrypted
120 key must be equal to the cipher default key length, in case the
121 certifiate is not given and all recipientInfo are tried out.
122 The old behaviour can be re-enabled in the CMS code by setting the
123 CMS_DEBUG_DECRYPT flag.
127 *) Early start up entropy quality from the DEVRANDOM seed source has been
128 improved for older Linux systems. The RAND subsystem will wait for
129 /dev/random to be producing output before seeding from /dev/urandom.
130 The seeded state is stored for future library initialisations using
131 a system global shared memory segment. The shared memory identifier
132 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
133 the desired value. The default identifier is 114.
136 *) Correct the extended master secret constant on EBCDIC systems. Without this
137 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
138 negotiate EMS will fail. Unfortunately this also means that TLS connections
139 between EBCDIC systems with this fix, and EBCDIC systems without this
140 fix will fail if they negotiate EMS.
143 *) Use Windows installation paths in the mingw builds
145 Mingw isn't a POSIX environment per se, which means that Windows
146 paths should be used for installation.
150 *) Changed DH_check to accept parameters with order q and 2q subgroups.
151 With order 2q subgroups the bit 0 of the private key is not secret
152 but DH_generate_key works around that by clearing bit 0 of the
153 private key for those. This avoids leaking bit 0 of the private key.
156 *) Significantly reduce secure memory usage by the randomness pools.
159 *) Revert the DEVRANDOM_WAIT feature for Linux systems
161 The DEVRANDOM_WAIT feature added a select() call to wait for the
162 /dev/random device to become readable before reading from the
165 It turned out that this change had negative side effects on
166 performance which were not acceptable. After some discussion it
167 was decided to revert this feature and leave it up to the OS
168 resp. the platform maintainer to ensure a proper initialization
169 during early boot time.
170 [Matthias St. Pierre]
172 Changes between 1.1.1b and 1.1.1c [28 May 2019]
174 *) Add build tests for C++. These are generated files that only do one
175 thing, to include one public OpenSSL head file each. This tests that
176 the public header files can be usefully included in a C++ application.
178 This test isn't enabled by default. It can be enabled with the option
179 'enable-buildtest-c++'.
182 *) Enable SHA3 pre-hashing for ECDSA and DSA.
185 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
186 This changes the size when using the genpkey app when no size is given. It
187 fixes an omission in earlier changes that changed all RSA, DSA and DH
188 generation apps to use 2048 bits by default.
191 *) Reorganize the manual pages to consistently have RETURN VALUES,
192 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
193 util/fix-doc-nits accordingly.
194 [Paul Yang, Joshua Lock]
196 *) Add the missing accessor EVP_PKEY_get0_engine()
199 *) Have apps like 's_client' and 's_server' output the signature scheme
200 along with other cipher suite parameters when debugging.
203 *) Make OPENSSL_config() error agnostic again.
206 *) Do the error handling in RSA decryption constant time.
209 *) Prevent over long nonces in ChaCha20-Poly1305.
211 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
212 for every encryption operation. RFC 7539 specifies that the nonce value
213 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
214 and front pads the nonce with 0 bytes if it is less than 12
215 bytes. However it also incorrectly allows a nonce to be set of up to 16
216 bytes. In this case only the last 12 bytes are significant and any
217 additional leading bytes are ignored.
219 It is a requirement of using this cipher that nonce values are
220 unique. Messages encrypted using a reused nonce value are susceptible to
221 serious confidentiality and integrity attacks. If an application changes
222 the default nonce length to be longer than 12 bytes and then makes a
223 change to the leading bytes of the nonce expecting the new value to be a
224 new unique nonce then such an application could inadvertently encrypt
225 messages with a reused nonce.
227 Additionally the ignored bytes in a long nonce are not covered by the
228 integrity guarantee of this cipher. Any application that relies on the
229 integrity of these ignored leading bytes of a long nonce may be further
230 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
231 is safe because no such use sets such a long nonce value. However user
232 applications that use this cipher directly and set a non-default nonce
233 length to be longer than 12 bytes may be vulnerable.
235 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
240 *) Add DEVRANDOM_WAIT feature for Linux systems
242 On older Linux systems where the getrandom() system call is not available,
243 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
244 Contrary to getrandom(), the /dev/urandom device will not block during
245 early boot when the kernel CSPRNG has not been seeded yet.
247 To mitigate this known weakness, use select() to wait for /dev/random to
248 become readable before reading from /dev/urandom.
250 *) Ensure that SM2 only uses SM3 as digest algorithm
253 Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
255 *) Added SCA hardening for modular field inversion in EC_GROUP through
256 a new dedicated field_inv() pointer in EC_METHOD.
257 This also addresses a leakage affecting conversions from projective
258 to affine coordinates.
259 [Billy Bob Brumley, Nicola Tuveri]
261 *) Change the info callback signals for the start and end of a post-handshake
262 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
263 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
264 confused by this and assume that a TLSv1.2 renegotiation has started. This
265 can break KeyUpdate handling. Instead we no longer signal the start and end
266 of a post handshake message exchange (although the messages themselves are
267 still signalled). This could break some applications that were expecting
268 the old signals. However without this KeyUpdate is not usable for many
272 *) Fix a bug in the computation of the endpoint-pair shared secret used
273 by DTLS over SCTP. This breaks interoperability with older versions
274 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
275 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
276 interoperability with such broken implementations. However, enabling
277 this switch breaks interoperability with correct implementations.
279 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
280 re-used X509_PUBKEY object if the second PUBKEY is malformed.
283 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
286 *) Remove the 'dist' target and add a tarball building script. The
287 'dist' target has fallen out of use, and it shouldn't be
288 necessary to configure just to create a source distribution.
291 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
293 *) Timing vulnerability in DSA signature generation
295 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
296 timing side channel attack. An attacker could use variations in the signing
297 algorithm to recover the private key.
299 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
303 *) Timing vulnerability in ECDSA signature generation
305 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
306 timing side channel attack. An attacker could use variations in the signing
307 algorithm to recover the private key.
309 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
313 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
314 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
315 are retained for backwards compatibility.
318 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
319 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
320 of two gigabytes and the error handling improved.
322 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
323 categorized as a normal bug, not a security issue, because the DRBG reseeds
324 automatically and is fully functional even without additional randomness
325 provided by the application.
327 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
329 *) Add a new ClientHello callback. Provides a callback interface that gives
330 the application the ability to adjust the nascent SSL object at the
331 earliest stage of ClientHello processing, immediately after extensions have
332 been collected but before they have been processed. In particular, this
333 callback can adjust the supported TLS versions in response to the contents
337 *) Add SM2 base algorithm support.
340 *) s390x assembly pack: add (improved) hardware-support for the following
341 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
342 aes-cfb/cfb8, aes-ecb.
345 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
346 parameter is no longer accepted, as it leads to a corrupt table. NULL
347 pem_str is reserved for alias entries only.
350 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
351 step for prime curves. The new implementation is based on formulae from
352 differential addition-and-doubling in homogeneous projective coordinates
353 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
354 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
355 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
356 to work in projective coordinates.
357 [Billy Bob Brumley, Nicola Tuveri]
359 *) Change generating and checking of primes so that the error rate of not
360 being prime depends on the intended use based on the size of the input.
361 For larger primes this will result in more rounds of Miller-Rabin.
362 The maximal error rate for primes with more than 1080 bits is lowered
364 [Kurt Roeckx, Annie Yousar]
366 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
369 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
370 moving between systems, and to avoid confusion when a Windows build is
371 done with mingw vs with MSVC. For POSIX installs, there's still a
372 symlink or copy named 'tsget' to avoid that confusion as well.
375 *) Revert blinding in ECDSA sign and instead make problematic addition
376 length-invariant. Switch even to fixed-length Montgomery multiplication.
379 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
380 step for binary curves. The new implementation is based on formulae from
381 differential addition-and-doubling in mixed Lopez-Dahab projective
382 coordinates, modified to independently blind the operands.
383 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
385 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
386 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
387 EC_METHODs to implement their own specialized "ladder step", to take
388 advantage of more favorable coordinate systems or more efficient
389 differential addition-and-doubling algorithms.
390 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
392 *) Modified the random device based seed sources to keep the relevant
393 file descriptors open rather than reopening them on each access.
394 This allows such sources to operate in a chroot() jail without
395 the associated device nodes being available. This behaviour can be
396 controlled using RAND_keep_random_devices_open().
399 *) Numerous side-channel attack mitigations have been applied. This may have
400 performance impacts for some algorithms for the benefit of improved
401 security. Specific changes are noted in this change log by their respective
405 *) AIX shared library support overhaul. Switch to AIX "natural" way of
406 handling shared libraries, which means collecting shared objects of
407 different versions and bitnesses in one common archive. This allows to
408 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
409 doesn't affect the way 3rd party applications are linked, only how
410 multi-version installation is managed.
413 *) Make ec_group_do_inverse_ord() more robust and available to other
414 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
415 mitigations are applied to the fallback BN_mod_inverse().
416 When using this function rather than BN_mod_inverse() directly, new
417 EC cryptosystem implementations are then safer-by-default.
420 *) Add coordinate blinding for EC_POINT and implement projective
421 coordinate blinding for generic prime curves as a countermeasure to
422 chosen point SCA attacks.
423 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
425 *) Add blinding to ECDSA and DSA signatures to protect against side channel
426 attacks discovered by Keegan Ryan (NCC Group).
429 *) Enforce checking in the pkeyutl command line app to ensure that the input
430 length does not exceed the maximum supported digest length when performing
431 a sign, verify or verifyrecover operation.
434 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
435 I/O in combination with something like select() or poll() will hang. This
436 can be turned off again using SSL_CTX_clear_mode().
437 Many applications do not properly handle non-application data records, and
438 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
439 around the problems in those applications, but can also break some.
440 It's recommended to read the manpages about SSL_read(), SSL_write(),
441 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
442 SSL_CTX_set_read_ahead() again.
445 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
446 now allow empty (zero character) pass phrases.
449 *) Apply blinding to binary field modular inversion and remove patent
450 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
453 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
454 binary and prime elliptic curves.
457 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
458 constant time fixed point multiplication.
461 *) Revise elliptic curve scalar multiplication with timing attack
462 defenses: ec_wNAF_mul redirects to a constant time implementation
463 when computing fixed point and variable point multiplication (which
464 in OpenSSL are mostly used with secret scalars in keygen, sign,
465 ECDH derive operations).
466 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
469 *) Updated CONTRIBUTING
472 *) Updated DRBG / RAND to request nonce and additional low entropy
473 randomness from the system.
474 [Matthias St. Pierre]
476 *) Updated 'openssl rehash' to use OpenSSL consistent default.
479 *) Moved the load of the ssl_conf module to libcrypto, which helps
480 loading engines that libssl uses before libssl is initialised.
483 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
486 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
487 [Ingo Schwarze, Rich Salz]
489 *) Added output of accepting IP address and port for 'openssl s_server'
492 *) Added a new API for TLSv1.3 ciphersuites:
493 SSL_CTX_set_ciphersuites()
494 SSL_set_ciphersuites()
497 *) Memory allocation failures consistently add an error to the error
501 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
502 in libcrypto when run as setuid/setgid.
505 *) Load any config file by default when libssl is used.
508 *) Added new public header file <openssl/rand_drbg.h> and documentation
509 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
510 [Matthias St. Pierre]
512 *) QNX support removed (cannot find contributors to get their approval
513 for the license change).
516 *) TLSv1.3 replay protection for early data has been implemented. See the
517 SSL_read_early_data() man page for further details.
520 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
521 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
522 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
523 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
524 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
525 configuration has been separated out. See the ciphers man page or the
526 SSL_CTX_set_ciphersuites() man page for more information.
529 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
530 in responder mode now supports the new "-multi" option, which
531 spawns the specified number of child processes to handle OCSP
532 requests. The "-timeout" option now also limits the OCSP
533 responder's patience to wait to receive the full client request
534 on a newly accepted connection. Child processes are respawned
535 as needed, and the CA index file is automatically reloaded
536 when changed. This makes it possible to run the "ocsp" responder
537 as a long-running service, making the OpenSSL CA somewhat more
538 feature-complete. In this mode, most diagnostic messages logged
539 after entering the event loop are logged via syslog(3) rather than
543 *) Added support for X448 and Ed448. Heavily based on original work by
547 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
548 objects loaded. This adds the functions OSSL_STORE_expect() and
549 OSSL_STORE_find() as well as needed tools to construct searches and
550 get the search data out of them.
553 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
554 version of OpenSSL should review their configuration settings to ensure
555 that they are still appropriate for TLSv1.3. For further information see:
556 https://wiki.openssl.org/index.php/TLS1.3
559 *) Grand redesign of the OpenSSL random generator
561 The default RAND method now utilizes an AES-CTR DRBG according to
562 NIST standard SP 800-90Ar1. The new random generator is essentially
563 a port of the default random generator from the OpenSSL FIPS 2.0
564 object module. It is a hybrid deterministic random bit generator
565 using an AES-CTR bit stream and which seeds and reseeds itself
566 automatically using trusted system entropy sources.
568 Some of its new features are:
569 o Support for multiple DRBG instances with seed chaining.
570 o The default RAND method makes use of a DRBG.
571 o There is a public and private DRBG instance.
572 o The DRBG instances are fork-safe.
573 o Keep all global DRBG instances on the secure heap if it is enabled.
574 o The public and private DRBG instance are per thread for lock free
576 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
578 *) Changed Configure so it only says what it does and doesn't dump
579 so much data. Instead, ./configdata.pm should be used as a script
580 to display all sorts of configuration data.
583 *) Added processing of "make variables" to Configure.
586 *) Added SHA512/224 and SHA512/256 algorithm support.
589 *) The last traces of Netware support, first removed in 1.1.0, have
593 *) Get rid of Makefile.shared, and in the process, make the processing
594 of certain files (rc.obj, or the .def/.map/.opt files produced from
595 the ordinal files) more visible and hopefully easier to trace and
596 debug (or make silent).
599 *) Make it possible to have environment variable assignments as
600 arguments to config / Configure.
603 *) Add multi-prime RSA (RFC 8017) support.
606 *) Add SM3 implemented according to GB/T 32905-2016
607 [ Jack Lloyd <jack.lloyd@ribose.com>,
608 Ronald Tse <ronald.tse@ribose.com>,
609 Erick Borsboom <erick.borsboom@ribose.com> ]
611 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
612 as documented in RFC6066.
613 Based on a patch from Tomasz Moń
614 [Filipe Raimundo da Silva]
616 *) Add SM4 implemented according to GB/T 32907-2016.
617 [ Jack Lloyd <jack.lloyd@ribose.com>,
618 Ronald Tse <ronald.tse@ribose.com>,
619 Erick Borsboom <erick.borsboom@ribose.com> ]
621 *) Reimplement -newreq-nodes and ERR_error_string_n; the
622 original author does not agree with the license change.
625 *) Add ARIA AEAD TLS support.
628 *) Some macro definitions to support VS6 have been removed. Visual
629 Studio 6 has not worked since 1.1.0
632 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
633 without clearing the errors.
636 *) Add "atfork" functions. If building on a system that without
637 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
638 requirements. The RAND facility now uses/requires this.
644 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
645 not possible to disable entirely. However, it's still possible to
646 disable the console reading UI method, UI_OpenSSL() (use UI_null()
649 To disable, configure with 'no-ui-console'. 'no-ui' is still
650 possible to use as an alias. Check at compile time with the
651 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
652 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
655 *) Add a STORE module, which implements a uniform and URI based reader of
656 stores that can contain keys, certificates, CRLs and numerous other
657 objects. The main API is loosely based on a few stdio functions,
658 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
659 OSSL_STORE_error and OSSL_STORE_close.
660 The implementation uses backends called "loaders" to implement arbitrary
661 URI schemes. There is one built in "loader" for the 'file' scheme.
664 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
665 then adjusted to work on FreeBSD 8.4 as well.
666 Enable by configuring with 'enable-devcryptoeng'. This is done by default
667 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
670 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
671 util/mkerr.pl, which is adapted to allow those prefixes, leading to
672 error code calls like this:
674 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
676 With this change, we claim the namespaces OSSL and OPENSSL in a manner
677 that can be encoded in C. For the foreseeable future, this will only
679 [Richard Levitte and Tim Hudson]
681 *) Removed BSD cryptodev engine.
684 *) Add a build target 'build_all_generated', to build all generated files
685 and only that. This can be used to prepare everything that requires
686 things like perl for a system that lacks perl and then move everything
687 to that system and do the rest of the build there.
690 *) In the UI interface, make it possible to duplicate the user data. This
691 can be used by engines that need to retain the data for a longer time
692 than just the call where this user data is passed.
695 *) Ignore the '-named_curve auto' value for compatibility of applications
697 [Tomas Mraz <tmraz@fedoraproject.org>]
699 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
700 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
701 alerts across multiple records (some of which could be empty). In practice
702 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
703 prohibits this altogether and other libraries (BoringSSL, NSS) do not
704 support this at all. Supporting it adds significant complexity to the
705 record layer, and its removal is unlikely to cause interoperability
709 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
710 with Z. These are meant to replace LONG and ZLONG and to be size safe.
711 The use of LONG and ZLONG is discouraged and scheduled for deprecation
715 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
716 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
717 [Richard Levitte, Andy Polyakov]
719 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
723 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
724 platform rather than 'mingw'.
727 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
728 success if they are asked to add an object which already exists
729 in the store. This change cascades to other functions which load
730 certificates and CRLs.
733 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
734 facilitate stack unwinding even from assembly subroutines.
737 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
738 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
741 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
742 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
743 which is the minimum version we support.
746 *) Certificate time validation (X509_cmp_time) enforces stricter
747 compliance with RFC 5280. Fractional seconds and timezone offsets
748 are no longer allowed.
751 *) Add support for ARIA
754 *) s_client will now send the Server Name Indication (SNI) extension by
755 default unless the new "-noservername" option is used. The server name is
756 based on the host provided to the "-connect" option unless overridden by
760 *) Add support for SipHash
763 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
764 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
765 prevent issues where no progress is being made and the peer continually
766 sends unrecognised record types, using up resources processing them.
769 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
770 using the algorithm defined in
771 https://www.akkadia.org/drepper/SHA-crypt.txt
774 *) Heartbeat support has been removed; the ABI is changed for now.
775 [Richard Levitte, Rich Salz]
777 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
780 *) The RSA "null" method, which was partially supported to avoid patent
781 issues, has been replaced to always returns NULL.
785 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
787 *) Client DoS due to large DH parameter
789 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
790 malicious server can send a very large prime value to the client. This will
791 cause the client to spend an unreasonably long period of time generating a
792 key for this prime resulting in a hang until the client has finished. This
793 could be exploited in a Denial Of Service attack.
795 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
799 *) Cache timing vulnerability in RSA Key Generation
801 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
802 a cache timing side channel attack. An attacker with sufficient access to
803 mount cache timing attacks during the RSA key generation process could
804 recover the private key.
806 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
807 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
811 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
812 parameter is no longer accepted, as it leads to a corrupt table. NULL
813 pem_str is reserved for alias entries only.
816 *) Revert blinding in ECDSA sign and instead make problematic addition
817 length-invariant. Switch even to fixed-length Montgomery multiplication.
820 *) Change generating and checking of primes so that the error rate of not
821 being prime depends on the intended use based on the size of the input.
822 For larger primes this will result in more rounds of Miller-Rabin.
823 The maximal error rate for primes with more than 1080 bits is lowered
825 [Kurt Roeckx, Annie Yousar]
827 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
830 *) Add blinding to ECDSA and DSA signatures to protect against side channel
831 attacks discovered by Keegan Ryan (NCC Group).
834 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
835 now allow empty (zero character) pass phrases.
838 *) Certificate time validation (X509_cmp_time) enforces stricter
839 compliance with RFC 5280. Fractional seconds and timezone offsets
840 are no longer allowed.
843 *) Fixed a text canonicalisation bug in CMS
845 Where a CMS detached signature is used with text content the text goes
846 through a canonicalisation process first prior to signing or verifying a
847 signature. This process strips trailing space at the end of lines, converts
848 line terminators to CRLF and removes additional trailing line terminators
849 at the end of a file. A bug in the canonicalisation process meant that
850 some characters, such as form-feed, were incorrectly treated as whitespace
851 and removed. This is contrary to the specification (RFC5485). This fix
852 could mean that detached text data signed with an earlier version of
853 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
854 signed with a fixed OpenSSL may fail to verify with an earlier version of
855 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
856 and use the "-binary" flag (for the "cms" command line application) or set
857 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
860 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
862 *) Constructed ASN.1 types with a recursive definition could exceed the stack
864 Constructed ASN.1 types with a recursive definition (such as can be found
865 in PKCS7) could eventually exceed the stack given malicious input with
866 excessive recursion. This could result in a Denial Of Service attack. There
867 are no such structures used within SSL/TLS that come from untrusted sources
868 so this is considered safe.
870 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
875 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
877 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
878 effectively reduced to only comparing the least significant bit of each
879 byte. This allows an attacker to forge messages that would be considered as
880 authenticated in an amount of tries lower than that guaranteed by the
881 security claims of the scheme. The module can only be compiled by the
882 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
884 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
889 *) Add a build target 'build_all_generated', to build all generated files
890 and only that. This can be used to prepare everything that requires
891 things like perl for a system that lacks perl and then move everything
892 to that system and do the rest of the build there.
895 *) Backport SSL_OP_NO_RENGOTIATION
897 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
898 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
899 changes this is no longer possible in 1.1.0. Therefore the new
900 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
901 1.1.0 to provide equivalent functionality.
903 Note that if an application built against 1.1.0h headers (or above) is run
904 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
905 accepted but nothing will happen, i.e. renegotiation will not be prevented.
908 *) Removed the OS390-Unix config target. It relied on a script that doesn't
912 *) rsaz_1024_mul_avx2 overflow bug on x86_64
914 There is an overflow bug in the AVX2 Montgomery multiplication procedure
915 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
916 Analysis suggests that attacks against RSA and DSA as a result of this
917 defect would be very difficult to perform and are not believed likely.
918 Attacks against DH1024 are considered just feasible, because most of the
919 work necessary to deduce information about a private key may be performed
920 offline. The amount of resources required for such an attack would be
921 significant. However, for an attack on TLS to be meaningful, the server
922 would have to share the DH1024 private key among multiple clients, which is
923 no longer an option since CVE-2016-0701.
925 This only affects processors that support the AVX2 but not ADX extensions
926 like Intel Haswell (4th generation).
928 This issue was reported to OpenSSL by David Benjamin (Google). The issue
929 was originally found via the OSS-Fuzz project.
933 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
935 *) bn_sqrx8x_internal carry bug on x86_64
937 There is a carry propagating bug in the x86_64 Montgomery squaring
938 procedure. No EC algorithms are affected. Analysis suggests that attacks
939 against RSA and DSA as a result of this defect would be very difficult to
940 perform and are not believed likely. Attacks against DH are considered just
941 feasible (although very difficult) because most of the work necessary to
942 deduce information about a private key may be performed offline. The amount
943 of resources required for such an attack would be very significant and
944 likely only accessible to a limited number of attackers. An attacker would
945 additionally need online access to an unpatched system using the target
946 private key in a scenario with persistent DH parameters and a private
947 key that is shared between multiple clients.
949 This only affects processors that support the BMI1, BMI2 and ADX extensions
950 like Intel Broadwell (5th generation) and later or AMD Ryzen.
952 This issue was reported to OpenSSL by the OSS-Fuzz project.
956 *) Malformed X.509 IPAddressFamily could cause OOB read
958 If an X.509 certificate has a malformed IPAddressFamily extension,
959 OpenSSL could do a one-byte buffer overread. The most likely result
960 would be an erroneous display of the certificate in text format.
962 This issue was reported to OpenSSL by the OSS-Fuzz project.
966 Changes between 1.1.0e and 1.1.0f [25 May 2017]
968 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
969 platform rather than 'mingw'.
972 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
973 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
974 which is the minimum version we support.
977 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
979 *) Encrypt-Then-Mac renegotiation crash
981 During a renegotiation handshake if the Encrypt-Then-Mac extension is
982 negotiated where it was not in the original handshake (or vice-versa) then
983 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
984 and servers are affected.
986 This issue was reported to OpenSSL by Joe Orton (Red Hat).
990 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
992 *) Truncated packet could crash via OOB read
994 If one side of an SSL/TLS path is running on a 32-bit host and a specific
995 cipher is being used, then a truncated packet can cause that host to
996 perform an out-of-bounds read, usually resulting in a crash.
998 This issue was reported to OpenSSL by Robert Święcki of Google.
1002 *) Bad (EC)DHE parameters cause a client crash
1004 If a malicious server supplies bad parameters for a DHE or ECDHE key
1005 exchange then this can result in the client attempting to dereference a
1006 NULL pointer leading to a client crash. This could be exploited in a Denial
1009 This issue was reported to OpenSSL by Guido Vranken.
1013 *) BN_mod_exp may produce incorrect results on x86_64
1015 There is a carry propagating bug in the x86_64 Montgomery squaring
1016 procedure. No EC algorithms are affected. Analysis suggests that attacks
1017 against RSA and DSA as a result of this defect would be very difficult to
1018 perform and are not believed likely. Attacks against DH are considered just
1019 feasible (although very difficult) because most of the work necessary to
1020 deduce information about a private key may be performed offline. The amount
1021 of resources required for such an attack would be very significant and
1022 likely only accessible to a limited number of attackers. An attacker would
1023 additionally need online access to an unpatched system using the target
1024 private key in a scenario with persistent DH parameters and a private
1025 key that is shared between multiple clients. For example this can occur by
1026 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1027 similar to CVE-2015-3193 but must be treated as a separate problem.
1029 This issue was reported to OpenSSL by the OSS-Fuzz project.
1033 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1035 *) ChaCha20/Poly1305 heap-buffer-overflow
1037 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1038 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1039 crash. This issue is not considered to be exploitable beyond a DoS.
1041 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1045 *) CMS Null dereference
1047 Applications parsing invalid CMS structures can crash with a NULL pointer
1048 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1049 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1050 structure callback if an attempt is made to free certain invalid encodings.
1051 Only CHOICE structures using a callback which do not handle NULL value are
1054 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1058 *) Montgomery multiplication may produce incorrect results
1060 There is a carry propagating bug in the Broadwell-specific Montgomery
1061 multiplication procedure that handles input lengths divisible by, but
1062 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1063 and DH private keys are impossible. This is because the subroutine in
1064 question is not used in operations with the private key itself and an input
1065 of the attacker's direct choice. Otherwise the bug can manifest itself as
1066 transient authentication and key negotiation failures or reproducible
1067 erroneous outcome of public-key operations with specially crafted input.
1068 Among EC algorithms only Brainpool P-512 curves are affected and one
1069 presumably can attack ECDH key negotiation. Impact was not analyzed in
1070 detail, because pre-requisites for attack are considered unlikely. Namely
1071 multiple clients have to choose the curve in question and the server has to
1072 share the private key among them, neither of which is default behaviour.
1073 Even then only clients that chose the curve will be affected.
1075 This issue was publicly reported as transient failures and was not
1076 initially recognized as a security issue. Thanks to Richard Morgan for
1077 providing reproducible case.
1081 *) Removed automatic addition of RPATH in shared libraries and executables,
1082 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1085 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1087 *) Fix Use After Free for large message sizes
1089 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1090 message larger than approx 16k is received then the underlying buffer to
1091 store the incoming message is reallocated and moved. Unfortunately a
1092 dangling pointer to the old location is left which results in an attempt to
1093 write to the previously freed location. This is likely to result in a
1094 crash, however it could potentially lead to execution of arbitrary code.
1096 This issue only affects OpenSSL 1.1.0a.
1098 This issue was reported to OpenSSL by Robert Święcki.
1102 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1104 *) OCSP Status Request extension unbounded memory growth
1106 A malicious client can send an excessively large OCSP Status Request
1107 extension. If that client continually requests renegotiation, sending a
1108 large OCSP Status Request extension each time, then there will be unbounded
1109 memory growth on the server. This will eventually lead to a Denial Of
1110 Service attack through memory exhaustion. Servers with a default
1111 configuration are vulnerable even if they do not support OCSP. Builds using
1112 the "no-ocsp" build time option are not affected.
1114 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1118 *) SSL_peek() hang on empty record
1120 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1121 sends an empty record. This could be exploited by a malicious peer in a
1122 Denial Of Service attack.
1124 This issue was reported to OpenSSL by Alex Gaynor.
1128 *) Excessive allocation of memory in tls_get_message_header() and
1129 dtls1_preprocess_fragment()
1131 A (D)TLS message includes 3 bytes for its length in the header for the
1132 message. This would allow for messages up to 16Mb in length. Messages of
1133 this length are excessive and OpenSSL includes a check to ensure that a
1134 peer is sending reasonably sized messages in order to avoid too much memory
1135 being consumed to service a connection. A flaw in the logic of version
1136 1.1.0 means that memory for the message is allocated too early, prior to
1137 the excessive message length check. Due to way memory is allocated in
1138 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1139 to service a connection. This could lead to a Denial of Service through
1140 memory exhaustion. However, the excessive message length check still takes
1141 place, and this would cause the connection to immediately fail. Assuming
1142 that the application calls SSL_free() on the failed connection in a timely
1143 manner then the 21Mb of allocated memory will then be immediately freed
1144 again. Therefore the excessive memory allocation will be transitory in
1145 nature. This then means that there is only a security impact if:
1147 1) The application does not call SSL_free() in a timely manner in the event
1148 that the connection fails
1150 2) The application is working in a constrained environment where there is
1151 very little free memory
1153 3) The attacker initiates multiple connection attempts such that there are
1154 multiple connections in a state where memory has been allocated for the
1155 connection; SSL_free() has not yet been called; and there is insufficient
1156 memory to service the multiple requests.
1158 Except in the instance of (1) above any Denial Of Service is likely to be
1159 transitory because as soon as the connection fails the memory is
1160 subsequently freed again in the SSL_free() call. However there is an
1161 increased risk during this period of application crashes due to the lack of
1162 memory - which would then mean a more serious Denial of Service.
1164 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1165 (CVE-2016-6307 and CVE-2016-6308)
1168 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1169 had to be removed. Primary reason is that vendor assembler can't
1170 assemble our modules with -KPIC flag. As result it, assembly
1171 support, was not even available as option. But its lack means
1172 lack of side-channel resistant code, which is incompatible with
1173 security by todays standards. Fortunately gcc is readily available
1174 prepackaged option, which we firmly point at...
1177 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1179 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1180 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1181 (to any value) allows Windows user to access PKCS#12 file generated
1182 with Windows CryptoAPI and protected with non-ASCII password, as well
1183 as files generated under UTF-8 locale on Linux also protected with
1187 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1188 have been disabled by default and removed from DEFAULT, just like RC4.
1189 See the RC4 item below to re-enable both.
1192 *) The method for finding the storage location for the Windows RAND seed file
1193 has changed. First we check %RANDFILE%. If that is not set then we check
1194 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1195 all else fails we fall back to C:\.
1198 *) The EVP_EncryptUpdate() function has had its return type changed from void
1199 to int. A return of 0 indicates and error while a return of 1 indicates
1203 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1204 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1205 off the constant time implementation for RSA, DSA and DH have been made
1206 no-ops and deprecated.
1209 *) Windows RAND implementation was simplified to only get entropy by
1210 calling CryptGenRandom(). Various other RAND-related tickets
1212 [Joseph Wylie Yandle, Rich Salz]
1214 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1215 and OPENSSL_LH_, respectively. The old names are available
1216 with API compatibility. They new names are now completely documented.
1219 *) Unify TYPE_up_ref(obj) methods signature.
1220 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1221 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1222 int (instead of void) like all others TYPE_up_ref() methods.
1223 So now these methods also check the return value of CRYPTO_atomic_add(),
1224 and the validity of object reference counter.
1225 [fdasilvayy@gmail.com]
1227 *) With Windows Visual Studio builds, the .pdb files are installed
1228 alongside the installed libraries and executables. For a static
1229 library installation, ossl_static.pdb is the associate compiler
1230 generated .pdb file to be used when linking programs.
1233 *) Remove openssl.spec. Packaging files belong with the packagers.
1236 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1237 recognise x86_64 architectures automatically. You can still decide
1238 to build for a different bitness with the environment variable
1239 KERNEL_BITS (can be 32 or 64), for example:
1241 KERNEL_BITS=32 ./config
1245 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1246 256 bit AES and HMAC with SHA256.
1249 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1252 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1255 *) To enable users to have their own config files and build file templates,
1256 Configure looks in the directory indicated by the environment variable
1257 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1258 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1259 name and is used as is.
1262 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1263 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1264 X509_CERT_FILE_CTX was removed.
1267 *) "shared" builds are now the default. To create only static libraries use
1268 the "no-shared" Configure option.
1271 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1272 All of these option have not worked for some while and are fundamental
1276 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1277 global cleanup functions are no longer required because they are handled
1278 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1279 Explicitly de-initing can cause problems (e.g. where a library that uses
1280 OpenSSL de-inits, but an application is still using it). The affected
1281 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1282 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1283 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1284 COMP_zlib_cleanup().
1287 *) --strict-warnings no longer enables runtime debugging options
1288 such as REF_DEBUG. Instead, debug options are automatically
1289 enabled with '--debug' builds.
1290 [Andy Polyakov, Emilia Käsper]
1292 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1293 have been moved out of the public header files. New functions for managing
1294 these have been added.
1297 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1298 objects have been moved out of the public header files. New
1299 functions for managing these have been added.
1302 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1303 have been moved out of the public header files. New functions for managing
1304 these have been added.
1307 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1308 moved out of the public header files. New functions for managing these
1312 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1315 *) Removed the mk1mf build scripts.
1318 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1319 it is always safe to #include a header now.
1322 *) Removed the aged BC-32 config and all its supporting scripts
1325 *) Removed support for Ultrix, Netware, and OS/2.
1328 *) Add support for HKDF.
1329 [Alessandro Ghedini]
1331 *) Add support for blake2b and blake2s
1334 *) Added support for "pipelining". Ciphers that have the
1335 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1336 encryptions/decryptions simultaneously. There are currently no built-in
1337 ciphers with this property but the expectation is that engines will be able
1338 to offer it to significantly improve throughput. Support has been extended
1339 into libssl so that multiple records for a single connection can be
1340 processed in one go (for >=TLS 1.1).
1343 *) Added the AFALG engine. This is an async capable engine which is able to
1344 offload work to the Linux kernel. In this initial version it only supports
1345 AES128-CBC. The kernel must be version 4.1.0 or greater.
1348 *) OpenSSL now uses a new threading API. It is no longer necessary to
1349 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1350 are two supported threading models: pthreads and windows threads. It is
1351 also possible to configure OpenSSL at compile time for "no-threads". The
1352 old threading API should no longer be used. The functions have been
1353 replaced with "no-op" compatibility macros.
1354 [Alessandro Ghedini, Matt Caswell]
1356 *) Modify behavior of ALPN to invoke callback after SNI/servername
1357 callback, such that updates to the SSL_CTX affect ALPN.
1360 *) Add SSL_CIPHER queries for authentication and key-exchange.
1363 *) Changes to the DEFAULT cipherlist:
1364 - Prefer (EC)DHE handshakes over plain RSA.
1365 - Prefer AEAD ciphers over legacy ciphers.
1366 - Prefer ECDSA over RSA when both certificates are available.
1367 - Prefer TLSv1.2 ciphers/PRF.
1368 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1372 *) Change the ECC default curve list to be this, in order: x25519,
1373 secp256r1, secp521r1, secp384r1.
1376 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1377 disabled by default. They can be re-enabled using the
1378 enable-weak-ssl-ciphers option to Configure.
1381 *) If the server has ALPN configured, but supports no protocols that the
1382 client advertises, send a fatal "no_application_protocol" alert.
1383 This behaviour is SHALL in RFC 7301, though it isn't universally
1384 implemented by other servers.
1387 *) Add X25519 support.
1388 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1389 for public and private key encoding using the format documented in
1390 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1391 key generation and key derivation.
1393 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1397 *) Deprecate SRP_VBASE_get_by_user.
1398 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1399 In order to fix an unavoidable memory leak (CVE-2016-0798),
1400 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1401 seed, even if the seed is configured.
1403 Users should use SRP_VBASE_get1_by_user instead. Note that in
1404 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1405 also that even though configuring the SRP seed attempts to hide
1406 invalid usernames by continuing the handshake with fake
1407 credentials, this behaviour is not constant time and no strong
1408 guarantees are made that the handshake is indistinguishable from
1409 that of a valid user.
1412 *) Configuration change; it's now possible to build dynamic engines
1413 without having to build shared libraries and vice versa. This
1414 only applies to the engines in engines/, those in crypto/engine/
1415 will always be built into libcrypto (i.e. "static").
1417 Building dynamic engines is enabled by default; to disable, use
1418 the configuration option "disable-dynamic-engine".
1420 The only requirements for building dynamic engines are the
1421 presence of the DSO module and building with position independent
1422 code, so they will also automatically be disabled if configuring
1423 with "disable-dso" or "disable-pic".
1425 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1426 are also taken away from openssl/opensslconf.h, as they are
1430 *) Configuration change; if there is a known flag to compile
1431 position independent code, it will always be applied on the
1432 libcrypto and libssl object files, and never on the application
1433 object files. This means other libraries that use routines from
1434 libcrypto / libssl can be made into shared libraries regardless
1435 of how OpenSSL was configured.
1437 If this isn't desirable, the configuration options "disable-pic"
1438 or "no-pic" can be used to disable the use of PIC. This will
1439 also disable building shared libraries and dynamic engines.
1442 *) Removed JPAKE code. It was experimental and has no wide use.
1445 *) The INSTALL_PREFIX Makefile variable has been renamed to
1446 DESTDIR. That makes for less confusion on what this variable
1447 is for. Also, the configuration option --install_prefix is
1451 *) Heartbeat for TLS has been removed and is disabled by default
1452 for DTLS; configure with enable-heartbeats. Code that uses the
1453 old #define's might need to be updated.
1454 [Emilia Käsper, Rich Salz]
1456 *) Rename REF_CHECK to REF_DEBUG.
1459 *) New "unified" build system
1461 The "unified" build system is aimed to be a common system for all
1462 platforms we support. With it comes new support for VMS.
1464 This system builds supports building in a different directory tree
1465 than the source tree. It produces one Makefile (for unix family
1466 or lookalikes), or one descrip.mms (for VMS).
1468 The source of information to make the Makefile / descrip.mms is
1469 small files called 'build.info', holding the necessary
1470 information for each directory with source to compile, and a
1471 template in Configurations, like unix-Makefile.tmpl or
1474 With this change, the library names were also renamed on Windows
1475 and on VMS. They now have names that are closer to the standard
1476 on Unix, and include the major version number, and in certain
1477 cases, the architecture they are built for. See "Notes on shared
1478 libraries" in INSTALL.
1480 We rely heavily on the perl module Text::Template.
1483 *) Added support for auto-initialisation and de-initialisation of the library.
1484 OpenSSL no longer requires explicit init or deinit routines to be called,
1485 except in certain circumstances. See the OPENSSL_init_crypto() and
1486 OPENSSL_init_ssl() man pages for further information.
1489 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1490 "peer" argument is now expected to be a BIO_ADDR object.
1492 *) Rewrite of BIO networking library. The BIO library lacked consistent
1493 support of IPv6, and adding it required some more extensive
1494 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1495 which hold all types of addresses and chains of address information.
1496 It also introduces a new API, with functions like BIO_socket,
1497 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1498 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1499 have been adapted accordingly.
1502 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1506 *) CRIME protection: disable compression by default, even if OpenSSL is
1507 compiled with zlib enabled. Applications can still enable compression
1508 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1509 using the SSL_CONF library to configure compression.
1512 *) The signature of the session callback configured with
1513 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1514 was explicitly marked as 'const unsigned char*' instead of
1518 *) Always DPURIFY. Remove the use of uninitialized memory in the
1519 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1522 *) Removed many obsolete configuration items, including
1523 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1524 MD2_CHAR, MD2_INT, MD2_LONG
1526 IDEA_SHORT, IDEA_LONG
1527 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1528 [Rich Salz, with advice from Andy Polyakov]
1530 *) Many BN internals have been moved to an internal header file.
1531 [Rich Salz with help from Andy Polyakov]
1533 *) Configuration and writing out the results from it has changed.
1534 Files such as Makefile include/openssl/opensslconf.h and are now
1535 produced through general templates, such as Makefile.in and
1536 crypto/opensslconf.h.in and some help from the perl module
1539 Also, the center of configuration information is no longer
1540 Makefile. Instead, Configure produces a perl module in
1541 configdata.pm which holds most of the config data (in the hash
1542 table %config), the target data that comes from the target
1543 configuration in one of the Configurations/*.conf files (in
1547 *) To clarify their intended purposes, the Configure options
1548 --prefix and --openssldir change their semantics, and become more
1549 straightforward and less interdependent.
1551 --prefix shall be used exclusively to give the location INSTALLTOP
1552 where programs, scripts, libraries, include files and manuals are
1553 going to be installed. The default is now /usr/local.
1555 --openssldir shall be used exclusively to give the default
1556 location OPENSSLDIR where certificates, private keys, CRLs are
1557 managed. This is also where the default openssl.cnf gets
1559 If the directory given with this option is a relative path, the
1560 values of both the --prefix value and the --openssldir value will
1561 be combined to become OPENSSLDIR.
1562 The default for --openssldir is INSTALLTOP/ssl.
1564 Anyone who uses --openssldir to specify where OpenSSL is to be
1565 installed MUST change to use --prefix instead.
1568 *) The GOST engine was out of date and therefore it has been removed. An up
1569 to date GOST engine is now being maintained in an external repository.
1570 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1571 support for GOST ciphersuites (these are only activated if a GOST engine
1575 *) EGD is no longer supported by default; use enable-egd when
1577 [Ben Kaduk and Rich Salz]
1579 *) The distribution now has Makefile.in files, which are used to
1580 create Makefile's when Configure is run. *Configure must be run
1581 before trying to build now.*
1584 *) The return value for SSL_CIPHER_description() for error conditions
1588 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1590 Obtaining and performing DNSSEC validation of TLSA records is
1591 the application's responsibility. The application provides
1592 the TLSA records of its choice to OpenSSL, and these are then
1593 used to authenticate the peer.
1595 The TLSA records need not even come from DNS. They can, for
1596 example, be used to implement local end-entity certificate or
1597 trust-anchor "pinning", where the "pin" data takes the form
1598 of TLSA records, which can augment or replace verification
1599 based on the usual WebPKI public certification authorities.
1602 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1603 continues to support deprecated interfaces in default builds.
1604 However, applications are strongly advised to compile their
1605 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1606 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1607 or the 1.1.0 releases.
1609 In environments in which all applications have been ported to
1610 not use any deprecated interfaces OpenSSL's Configure script
1611 should be used with the --api=1.1.0 option to entirely remove
1612 support for the deprecated features from the library and
1613 unconditionally disable them in the installed headers.
1614 Essentially the same effect can be achieved with the "no-deprecated"
1615 argument to Configure, except that this will always restrict
1616 the build to just the latest API, rather than a fixed API
1619 As applications are ported to future revisions of the API,
1620 they should update their compile-time OPENSSL_API_COMPAT define
1621 accordingly, but in most cases should be able to continue to
1622 compile with later releases.
1624 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1625 0x10000000L and 0x00908000L, respectively. However those
1626 versions did not support the OPENSSL_API_COMPAT feature, and
1627 so applications are not typically tested for explicit support
1628 of just the undeprecated features of either release.
1631 *) Add support for setting the minimum and maximum supported protocol.
1632 It can bet set via the SSL_set_min_proto_version() and
1633 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1634 MaxProtocol. It's recommended to use the new APIs to disable
1635 protocols instead of disabling individual protocols using
1636 SSL_set_options() or SSL_CONF's Protocol. This change also
1637 removes support for disabling TLS 1.2 in the OpenSSL TLS
1638 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1641 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1644 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1645 and integrates ECDSA and ECDH functionality into EC. Implementations can
1646 now redirect key generation and no longer need to convert to or from
1649 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1650 include the ec.h header file instead.
1653 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1654 ciphers who are no longer supported and drops support the ephemeral RSA key
1655 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1658 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1659 opaque. For HMAC_CTX, the following constructors and destructors
1662 HMAC_CTX *HMAC_CTX_new(void);
1663 void HMAC_CTX_free(HMAC_CTX *ctx);
1665 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1666 destroy such methods has been added. See EVP_MD_meth_new(3) and
1667 EVP_CIPHER_meth_new(3) for documentation.
1670 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1671 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1672 EVP_MD_CTX_reset() should be called instead to reinitialise
1673 an already created structure.
1674 2) For consistency with the majority of our object creators and
1675 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1676 EVP_MD_CTX_(new|free). The old names are retained as macros
1677 for deprecated builds.
1680 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1681 cryptographic operations to be performed asynchronously as long as an
1682 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1683 further details. Libssl has also had this capability integrated with the
1684 introduction of the new mode SSL_MODE_ASYNC and associated error
1685 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1686 pages. This work was developed in partnership with Intel Corp.
1689 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1690 always enabled now. If you want to disable the support you should
1691 exclude it using the list of supported ciphers. This also means that the
1692 "-no_ecdhe" option has been removed from s_server.
1695 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1696 SSL_{CTX_}set1_curves() which can set a list.
1699 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1700 curve you want to support using SSL_{CTX_}set1_curves().
1703 *) State machine rewrite. The state machine code has been significantly
1704 refactored in order to remove much duplication of code and solve issues
1705 with the old code (see ssl/statem/README for further details). This change
1706 does have some associated API changes. Notably the SSL_state() function
1707 has been removed and replaced by SSL_get_state which now returns an
1708 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1709 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1713 *) All instances of the string "ssleay" in the public API were replaced
1714 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1715 Some error codes related to internal RSA_eay API's were renamed.
1718 *) The demo files in crypto/threads were moved to demo/threads.
1721 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1723 [Matt Caswell, Rich Salz]
1725 *) New ASN.1 embed macro.
1727 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1728 structure is not allocated: it is part of the parent. That is instead of
1736 This reduces memory fragmentation and make it impossible to accidentally
1737 set a mandatory field to NULL.
1739 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1740 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1741 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1745 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1748 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1749 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1750 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1751 DES and RC4 ciphersuites.
1754 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1755 This changes the decoding behaviour for some invalid messages,
1756 though the change is mostly in the more lenient direction, and
1757 legacy behaviour is preserved as much as possible.
1760 *) Fix no-stdio build.
1761 [ David Woodhouse <David.Woodhouse@intel.com> and also
1762 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1764 *) New testing framework
1765 The testing framework has been largely rewritten and is now using
1766 perl and the perl modules Test::Harness and an extended variant of
1767 Test::More called OpenSSL::Test to do its work. All test scripts in
1768 test/ have been rewritten into test recipes, and all direct calls to
1769 executables in test/Makefile have become individual recipes using the
1770 simplified testing OpenSSL::Test::Simple.
1772 For documentation on our testing modules, do:
1774 perldoc test/testlib/OpenSSL/Test/Simple.pm
1775 perldoc test/testlib/OpenSSL/Test.pm
1779 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1780 are used; the latter aborts on memory leaks (usually checked on exit).
1781 Some undocumented "set malloc, etc., hooks" functions were removed
1782 and others were changed. All are now documented.
1785 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1787 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1789 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1790 from RFC4279, RFC4785, RFC5487, RFC5489.
1792 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1793 original RSA_PSK patch.
1796 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1797 era flag was never set throughout the codebase (only read). Also removed
1798 SSL3_FLAGS_POP_BUFFER which was only used if
1799 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1802 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1803 to be "oneline" instead of "compat".
1806 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1807 not aware of clients that still exhibit this bug, and the workaround
1808 hasn't been working properly for a while.
1811 *) The return type of BIO_number_read() and BIO_number_written() as well as
1812 the corresponding num_read and num_write members in the BIO structure has
1813 changed from unsigned long to uint64_t. On platforms where an unsigned
1814 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1818 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1819 OpenSSL without support for them. It also means that maintaining
1820 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1821 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1824 *) Removed support for the two export grade static DH ciphersuites
1825 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1826 were newly added (along with a number of other static DH ciphersuites) to
1827 1.0.2. However the two export ones have *never* worked since they were
1828 introduced. It seems strange in any case to be adding new export
1829 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1832 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1833 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1834 and turned into macros which simply call the new preferred function names
1835 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1836 should use the new names instead. Also as part of this change the ssl23.h
1837 header file has been removed.
1840 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1841 code and the associated standard is no longer considered fit-for-purpose.
1844 *) RT2547 was closed. When generating a private key, try to make the
1845 output file readable only by the owner. This behavior change might
1846 be noticeable when interacting with other software.
1848 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1852 *) Added HTTP GET support to the ocsp command.
1855 *) Changed default digest for the dgst and enc commands from MD5 to
1859 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1862 *) Added support for TLS extended master secret from
1863 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1864 initial patch which was a great help during development.
1867 *) All libssl internal structures have been removed from the public header
1868 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1869 now redundant). Users should not attempt to access internal structures
1870 directly. Instead they should use the provided API functions.
1873 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1874 Access to deprecated functions can be re-enabled by running config with
1875 "enable-deprecated". In addition applications wishing to use deprecated
1876 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1877 will, by default, disable some transitive includes that previously existed
1878 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1881 *) Added support for OCB mode. OpenSSL has been granted a patent license
1882 compatible with the OpenSSL license for use of OCB. Details are available
1883 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1884 for OCB can be removed by calling config with no-ocb.
1887 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1888 compatible client hello.
1891 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1892 done while fixing the error code for the key-too-small case.
1893 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1895 *) CA.sh has been removed; use CA.pl instead.
1898 *) Removed old DES API.
1901 *) Remove various unsupported platforms:
1907 Sinix/ReliantUNIX RM400
1912 16-bit platforms such as WIN16
1915 *) Clean up OPENSSL_NO_xxx #define's
1916 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1917 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1918 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1919 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1920 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1921 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1922 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1923 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1924 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1925 Remove MS_STATIC; it's a relic from platforms <32 bits.
1928 *) Cleaned up dead code
1929 Remove all but one '#ifdef undef' which is to be looked at.
1932 *) Clean up calling of xxx_free routines.
1933 Just like free(), fix most of the xxx_free routines to accept
1934 NULL. Remove the non-null checks from callers. Save much code.
1937 *) Add secure heap for storage of private keys (when possible).
1938 Add BIO_s_secmem(), CBIGNUM, etc.
1939 Contributed by Akamai Technologies under our Corporate CLA.
1942 *) Experimental support for a new, fast, unbiased prime candidate generator,
1943 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1944 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1946 *) New output format NSS in the sess_id command line tool. This allows
1947 exporting the session id and the master key in NSS keylog format.
1948 [Martin Kaiser <martin@kaiser.cx>]
1950 *) Harmonize version and its documentation. -f flag is used to display
1952 [mancha <mancha1@zoho.com>]
1954 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1955 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1956 [mancha <mancha1@zoho.com>]
1958 *) Fix some double frees. These are not thought to be exploitable.
1959 [mancha <mancha1@zoho.com>]
1961 *) A missing bounds check in the handling of the TLS heartbeat extension
1962 can be used to reveal up to 64k of memory to a connected client or
1965 Thanks for Neel Mehta of Google Security for discovering this bug and to
1966 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1967 preparing the fix (CVE-2014-0160)
1968 [Adam Langley, Bodo Moeller]
1970 *) Fix for the attack described in the paper "Recovering OpenSSL
1971 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1972 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1973 http://eprint.iacr.org/2014/140
1975 Thanks to Yuval Yarom and Naomi Benger for discovering this
1976 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1977 [Yuval Yarom and Naomi Benger]
1979 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1980 this fixes a limitation in previous versions of OpenSSL.
1983 *) Experimental encrypt-then-mac support.
1985 Experimental support for encrypt then mac from
1986 draft-gutmann-tls-encrypt-then-mac-02.txt
1988 To enable it set the appropriate extension number (0x42 for the test
1989 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1991 For non-compliant peers (i.e. just about everything) this should have no
1994 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1998 *) Add EVP support for key wrapping algorithms, to avoid problems with
1999 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2000 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2001 algorithms and include tests cases.
2004 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2008 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2009 MGF1 digest and OAEP label.
2012 *) Make openssl verify return errors.
2013 [Chris Palmer <palmer@google.com> and Ben Laurie]
2015 *) New function ASN1_TIME_diff to calculate the difference between two
2016 ASN1_TIME structures or one structure and the current time.
2019 *) Update fips_test_suite to support multiple command line options. New
2020 test to induce all self test errors in sequence and check expected
2024 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2025 sign or verify all in one operation.
2028 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2029 test programs and fips_test_suite. Includes functionality to parse
2030 the minimal script output of fipsalgest.pl directly.
2033 *) Add authorisation parameter to FIPS_module_mode_set().
2036 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2039 *) Use separate DRBG fields for internal and external flags. New function
2040 FIPS_drbg_health_check() to perform on demand health checking. Add
2041 generation tests to fips_test_suite with reduced health check interval to
2042 demonstrate periodic health checking. Add "nodh" option to
2043 fips_test_suite to skip very slow DH test.
2046 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2050 *) More extensive health check for DRBG checking many more failure modes.
2051 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2052 combination: call this in fips_test_suite.
2055 *) Add support for canonical generation of DSA parameter 'g'. See
2058 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2059 POST to handle HMAC cases.
2062 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2063 to return numerical and string versions of the FIPS module number.
2066 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2067 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2068 outside the validated module in the FIPS capable OpenSSL.
2071 *) Minor change to DRBG entropy callback semantics. In some cases
2072 there is no multiple of the block length between min_len and
2073 max_len. Allow the callback to return more than max_len bytes
2074 of entropy but discard any extra: it is the callback's responsibility
2075 to ensure that the extra data discarded does not impact the
2076 requested amount of entropy.
2079 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2080 information in FIPS186-3, SP800-57 and SP800-131A.
2083 *) CCM support via EVP. Interface is very similar to GCM case except we
2084 must supply all data in one chunk (i.e. no update, final) and the
2085 message length must be supplied if AAD is used. Add algorithm test
2089 *) Initial version of POST overhaul. Add POST callback to allow the status
2090 of POST to be monitored and/or failures induced. Modify fips_test_suite
2091 to use callback. Always run all selftests even if one fails.
2094 *) XTS support including algorithm test driver in the fips_gcmtest program.
2095 Note: this does increase the maximum key length from 32 to 64 bytes but
2096 there should be no binary compatibility issues as existing applications
2097 will never use XTS mode.
2100 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2101 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2102 performs algorithm blocking for unapproved PRNG types. Also do not
2103 set PRNG type in FIPS_mode_set(): leave this to the application.
2104 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2105 the standard OpenSSL PRNG: set additional data to a date time vector.
2108 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2109 This shouldn't present any incompatibility problems because applications
2110 shouldn't be using these directly and any that are will need to rethink
2111 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2114 *) Extensive self tests and health checking required by SP800-90 DRBG.
2115 Remove strength parameter from FIPS_drbg_instantiate and always
2116 instantiate at maximum supported strength.
2119 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2122 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2125 *) New function DH_compute_key_padded() to compute a DH key and pad with
2126 leading zeroes if needed: this complies with SP800-56A et al.
2129 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2130 anything, incomplete, subject to change and largely untested at present.
2133 *) Modify fipscanisteronly build option to only build the necessary object
2134 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2137 *) Add experimental option FIPSSYMS to give all symbols in
2138 fipscanister.o and FIPS or fips prefix. This will avoid
2139 conflicts with future versions of OpenSSL. Add perl script
2140 util/fipsas.pl to preprocess assembly language source files
2141 and rename any affected symbols.
2144 *) Add selftest checks and algorithm block of non-fips algorithms in
2145 FIPS mode. Remove DES2 from selftests.
2148 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2149 return internal method without any ENGINE dependencies. Add new
2150 tiny fips sign and verify functions.
2153 *) New build option no-ec2m to disable characteristic 2 code.
2156 *) New build option "fipscanisteronly". This only builds fipscanister.o
2157 and (currently) associated fips utilities. Uses the file Makefile.fips
2158 instead of Makefile.org as the prototype.
2161 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2162 Update fips_gcmtest to use IV generator.
2165 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2166 setting output buffer to NULL. The *Final function must be
2167 called although it will not retrieve any additional data. The tag
2168 can be set or retrieved with a ctrl. The IV length is by default 12
2169 bytes (96 bits) but can be set to an alternative value. If the IV
2170 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2174 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2175 underlying do_cipher function handles all cipher semantics itself
2176 including padding and finalisation. This is useful if (for example)
2177 an ENGINE cipher handles block padding itself. The behaviour of
2178 do_cipher is subtly changed if this flag is set: the return value
2179 is the number of characters written to the output buffer (zero is
2180 no longer an error code) or a negative error code. Also if the
2181 input buffer is NULL and length 0 finalisation should be performed.
2184 *) If a candidate issuer certificate is already part of the constructed
2185 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2188 *) Improve forward-security support: add functions
2190 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2191 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2193 for use by SSL/TLS servers; the callback function will be called whenever a
2194 new session is created, and gets to decide whether the session may be
2195 cached to make it resumable (return 0) or not (return 1). (As by the
2196 SSL/TLS protocol specifications, the session_id sent by the server will be
2197 empty to indicate that the session is not resumable; also, the server will
2198 not generate RFC 4507 (RFC 5077) session tickets.)
2200 A simple reasonable callback implementation is to return is_forward_secure.
2201 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2202 by the SSL/TLS server library, indicating whether it can provide forward
2204 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2206 *) New -verify_name option in command line utilities to set verification
2210 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2211 Add CMAC pkey methods.
2214 *) Experimental renegotiation in s_server -www mode. If the client
2215 browses /reneg connection is renegotiated. If /renegcert it is
2216 renegotiated requesting a certificate.
2219 *) Add an "external" session cache for debugging purposes to s_server. This
2220 should help trace issues which normally are only apparent in deployed
2221 multi-process servers.
2224 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2225 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2226 BIO_set_cipher() and some obscure PEM functions were changed so they
2227 can now return an error. The RAND changes required a change to the
2228 RAND_METHOD structure.
2231 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2232 a gcc attribute to warn if the result of a function is ignored. This
2233 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2234 whose return value is often ignored.
2237 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2238 These allow SCTs (signed certificate timestamps) to be requested and
2239 validated when establishing a connection.
2240 [Rob Percival <robpercival@google.com>]
2242 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2244 *) Prevent padding oracle in AES-NI CBC MAC check
2246 A MITM attacker can use a padding oracle attack to decrypt traffic
2247 when the connection uses an AES CBC cipher and the server support
2250 This issue was introduced as part of the fix for Lucky 13 padding
2251 attack (CVE-2013-0169). The padding check was rewritten to be in
2252 constant time by making sure that always the same bytes are read and
2253 compared against either the MAC or padding bytes. But it no longer
2254 checked that there was enough data to have both the MAC and padding
2257 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2261 *) Fix EVP_EncodeUpdate overflow
2263 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2264 Base64 encoding of binary data. If an attacker is able to supply very large
2265 amounts of input data then a length check can overflow resulting in a heap
2268 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2269 the PEM_write_bio* family of functions. These are mainly used within the
2270 OpenSSL command line applications, so any application which processes data
2271 from an untrusted source and outputs it as a PEM file should be considered
2272 vulnerable to this issue. User applications that call these APIs directly
2273 with large amounts of untrusted data may also be vulnerable.
2275 This issue was reported by Guido Vranken.
2279 *) Fix EVP_EncryptUpdate overflow
2281 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2282 is able to supply very large amounts of input data after a previous call to
2283 EVP_EncryptUpdate() with a partial block then a length check can overflow
2284 resulting in a heap corruption. Following an analysis of all OpenSSL
2285 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2286 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2287 the first called function after an EVP_EncryptInit(), and therefore that
2288 specific call must be safe. The second form is where the length passed to
2289 EVP_EncryptUpdate() can be seen from the code to be some small value and
2290 therefore there is no possibility of an overflow. Since all instances are
2291 one of these two forms, it is believed that there can be no overflows in
2292 internal code due to this problem. It should be noted that
2293 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2294 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2295 of these calls have also been analysed too and it is believed there are no
2296 instances in internal usage where an overflow could occur.
2298 This issue was reported by Guido Vranken.
2302 *) Prevent ASN.1 BIO excessive memory allocation
2304 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2305 a short invalid encoding can cause allocation of large amounts of memory
2306 potentially consuming excessive resources or exhausting memory.
2308 Any application parsing untrusted data through d2i BIO functions is
2309 affected. The memory based functions such as d2i_X509() are *not* affected.
2310 Since the memory based functions are used by the TLS library, TLS
2311 applications are not affected.
2313 This issue was reported by Brian Carpenter.
2319 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2320 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2321 in arbitrary stack data being returned in the buffer.
2323 This issue was reported by Guido Vranken.
2327 *) Modify behavior of ALPN to invoke callback after SNI/servername
2328 callback, such that updates to the SSL_CTX affect ALPN.
2331 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2335 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2336 methods are enabled and ssl2 is disabled the methods return NULL.
2339 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2341 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2342 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2343 provide any "EXPORT" or "LOW" strength ciphers.
2346 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2347 is by default disabled at build-time. Builds that are not configured with
2348 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2349 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2350 will need to explicitly call either of:
2352 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2354 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2356 as appropriate. Even if either of those is used, or the application
2357 explicitly uses the version-specific SSLv2_method() or its client and
2358 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2359 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2360 ciphers, and SSLv2 56-bit DES are no longer available.
2364 *) Fix a double-free in DSA code
2366 A double free bug was discovered when OpenSSL parses malformed DSA private
2367 keys and could lead to a DoS attack or memory corruption for applications
2368 that receive DSA private keys from untrusted sources. This scenario is
2371 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2376 *) Disable SRP fake user seed to address a server memory leak.
2378 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2380 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2381 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2382 was changed to ignore the "fake user" SRP seed, even if the seed
2385 Users should use SRP_VBASE_get1_by_user instead. Note that in
2386 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2387 also that even though configuring the SRP seed attempts to hide
2388 invalid usernames by continuing the handshake with fake
2389 credentials, this behaviour is not constant time and no strong
2390 guarantees are made that the handshake is indistinguishable from
2391 that of a valid user.
2395 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2397 In the BN_hex2bn function the number of hex digits is calculated using an
2398 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2399 large values of |i| this can result in |bn_expand| not allocating any
2400 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2401 field as NULL leading to a subsequent NULL ptr deref. For very large values
2402 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2403 In this case memory is allocated to the internal BIGNUM data field, but it
2404 is insufficiently sized leading to heap corruption. A similar issue exists
2405 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2406 is ever called by user applications with very large untrusted hex/dec data.
2407 This is anticipated to be a rare occurrence.
2409 All OpenSSL internal usage of these functions use data that is not expected
2410 to be untrusted, e.g. config file data or application command line
2411 arguments. If user developed applications generate config file data based
2412 on untrusted data then it is possible that this could also lead to security
2413 consequences. This is also anticipated to be rare.
2415 This issue was reported to OpenSSL by Guido Vranken.
2419 *) Fix memory issues in BIO_*printf functions
2421 The internal |fmtstr| function used in processing a "%s" format string in
2422 the BIO_*printf functions could overflow while calculating the length of a
2423 string and cause an OOB read when printing very long strings.
2425 Additionally the internal |doapr_outch| function can attempt to write to an
2426 OOB memory location (at an offset from the NULL pointer) in the event of a
2427 memory allocation failure. In 1.0.2 and below this could be caused where
2428 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2429 could be in processing a very long "%s" format string. Memory leaks can
2432 The first issue may mask the second issue dependent on compiler behaviour.
2433 These problems could enable attacks where large amounts of untrusted data
2434 is passed to the BIO_*printf functions. If applications use these functions
2435 in this way then they could be vulnerable. OpenSSL itself uses these
2436 functions when printing out human-readable dumps of ASN.1 data. Therefore
2437 applications that print this data could be vulnerable if the data is from
2438 untrusted sources. OpenSSL command line applications could also be
2439 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2440 as command line arguments.
2442 Libssl is not considered directly vulnerable. Additionally certificates etc
2443 received via remote connections via libssl are also unlikely to be able to
2444 trigger these issues because of message size limits enforced within libssl.
2446 This issue was reported to OpenSSL Guido Vranken.
2450 *) Side channel attack on modular exponentiation
2452 A side-channel attack was found which makes use of cache-bank conflicts on
2453 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2454 of RSA keys. The ability to exploit this issue is limited as it relies on
2455 an attacker who has control of code in a thread running on the same
2456 hyper-threaded core as the victim thread which is performing decryptions.
2458 This issue was reported to OpenSSL by Yuval Yarom, The University of
2459 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2460 Nadia Heninger, University of Pennsylvania with more information at
2461 http://cachebleed.info.
2465 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2466 if no keysize is specified with default_bits. This fixes an
2467 omission in an earlier change that changed all RSA/DSA key generation
2468 apps to use 2048 bits by default.
2471 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2472 *) DH small subgroups
2474 Historically OpenSSL only ever generated DH parameters based on "safe"
2475 primes. More recently (in version 1.0.2) support was provided for
2476 generating X9.42 style parameter files such as those required for RFC 5114
2477 support. The primes used in such files may not be "safe". Where an
2478 application is using DH configured with parameters based on primes that are
2479 not "safe" then an attacker could use this fact to find a peer's private
2480 DH exponent. This attack requires that the attacker complete multiple
2481 handshakes in which the peer uses the same private DH exponent. For example
2482 this could be used to discover a TLS server's private DH exponent if it's
2483 reusing the private DH exponent or it's using a static DH ciphersuite.
2485 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2486 TLS. It is not on by default. If the option is not set then the server
2487 reuses the same private DH exponent for the life of the server process and
2488 would be vulnerable to this attack. It is believed that many popular
2489 applications do set this option and would therefore not be at risk.
2491 The fix for this issue adds an additional check where a "q" parameter is
2492 available (as is the case in X9.42 based parameters). This detects the
2493 only known attack, and is the only possible defense for static DH
2494 ciphersuites. This could have some performance impact.
2496 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2497 default and cannot be disabled. This could have some performance impact.
2499 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2503 *) SSLv2 doesn't block disabled ciphers
2505 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2506 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2507 been disabled, provided that the SSLv2 protocol was not also disabled via
2510 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2511 and Sebastian Schinzel.
2515 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2517 *) BN_mod_exp may produce incorrect results on x86_64
2519 There is a carry propagating bug in the x86_64 Montgomery squaring
2520 procedure. No EC algorithms are affected. Analysis suggests that attacks
2521 against RSA and DSA as a result of this defect would be very difficult to
2522 perform and are not believed likely. Attacks against DH are considered just
2523 feasible (although very difficult) because most of the work necessary to
2524 deduce information about a private key may be performed offline. The amount
2525 of resources required for such an attack would be very significant and
2526 likely only accessible to a limited number of attackers. An attacker would
2527 additionally need online access to an unpatched system using the target
2528 private key in a scenario with persistent DH parameters and a private
2529 key that is shared between multiple clients. For example this can occur by
2530 default in OpenSSL DHE based SSL/TLS ciphersuites.
2532 This issue was reported to OpenSSL by Hanno Böck.
2536 *) Certificate verify crash with missing PSS parameter
2538 The signature verification routines will crash with a NULL pointer
2539 dereference if presented with an ASN.1 signature using the RSA PSS
2540 algorithm and absent mask generation function parameter. Since these
2541 routines are used to verify certificate signature algorithms this can be
2542 used to crash any certificate verification operation and exploited in a
2543 DoS attack. Any application which performs certificate verification is
2544 vulnerable including OpenSSL clients and servers which enable client
2547 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2551 *) X509_ATTRIBUTE memory leak
2553 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2554 memory. This structure is used by the PKCS#7 and CMS routines so any
2555 application which reads PKCS#7 or CMS data from untrusted sources is
2556 affected. SSL/TLS is not affected.
2558 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2563 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2564 This changes the decoding behaviour for some invalid messages,
2565 though the change is mostly in the more lenient direction, and
2566 legacy behaviour is preserved as much as possible.
2569 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2571 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2573 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2575 *) Alternate chains certificate forgery
2577 During certificate verification, OpenSSL will attempt to find an
2578 alternative certificate chain if the first attempt to build such a chain
2579 fails. An error in the implementation of this logic can mean that an
2580 attacker could cause certain checks on untrusted certificates to be
2581 bypassed, such as the CA flag, enabling them to use a valid leaf
2582 certificate to act as a CA and "issue" an invalid certificate.
2584 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2588 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2590 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2591 incompatibility in the handling of HMAC. The previous ABI has now been
2595 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2597 *) Malformed ECParameters causes infinite loop
2599 When processing an ECParameters structure OpenSSL enters an infinite loop
2600 if the curve specified is over a specially malformed binary polynomial
2603 This can be used to perform denial of service against any
2604 system which processes public keys, certificate requests or
2605 certificates. This includes TLS clients and TLS servers with
2606 client authentication enabled.
2608 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2612 *) Exploitable out-of-bounds read in X509_cmp_time
2614 X509_cmp_time does not properly check the length of the ASN1_TIME
2615 string and can read a few bytes out of bounds. In addition,
2616 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2619 An attacker can use this to craft malformed certificates and CRLs of
2620 various sizes and potentially cause a segmentation fault, resulting in
2621 a DoS on applications that verify certificates or CRLs. TLS clients
2622 that verify CRLs are affected. TLS clients and servers with client
2623 authentication enabled may be affected if they use custom verification
2626 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2627 independently by Hanno Böck.
2631 *) PKCS7 crash with missing EnvelopedContent
2633 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2634 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2635 with missing content and trigger a NULL pointer dereference on parsing.
2637 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2638 structures from untrusted sources are affected. OpenSSL clients and
2639 servers are not affected.
2641 This issue was reported to OpenSSL by Michal Zalewski (Google).
2645 *) CMS verify infinite loop with unknown hash function
2647 When verifying a signedData message the CMS code can enter an infinite loop
2648 if presented with an unknown hash function OID. This can be used to perform
2649 denial of service against any system which verifies signedData messages using
2651 This issue was reported to OpenSSL by Johannes Bauer.
2655 *) Race condition handling NewSessionTicket
2657 If a NewSessionTicket is received by a multi-threaded client when attempting to
2658 reuse a previous ticket then a race condition can occur potentially leading to
2659 a double free of the ticket data.
2663 *) Only support 256-bit or stronger elliptic curves with the
2664 'ecdh_auto' setting (server) or by default (client). Of supported
2665 curves, prefer P-256 (both).
2668 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2670 *) ClientHello sigalgs DoS fix
2672 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2673 invalid signature algorithms extension a NULL pointer dereference will
2674 occur. This can be exploited in a DoS attack against the server.
2676 This issue was was reported to OpenSSL by David Ramos of Stanford
2679 [Stephen Henson and Matt Caswell]
2681 *) Multiblock corrupted pointer fix
2683 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2684 feature only applies on 64 bit x86 architecture platforms that support AES
2685 NI instructions. A defect in the implementation of "multiblock" can cause
2686 OpenSSL's internal write buffer to become incorrectly set to NULL when
2687 using non-blocking IO. Typically, when the user application is using a
2688 socket BIO for writing, this will only result in a failed connection.
2689 However if some other BIO is used then it is likely that a segmentation
2690 fault will be triggered, thus enabling a potential DoS attack.
2692 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2696 *) Segmentation fault in DTLSv1_listen fix
2698 The DTLSv1_listen function is intended to be stateless and processes the
2699 initial ClientHello from many peers. It is common for user code to loop
2700 over the call to DTLSv1_listen until a valid ClientHello is received with
2701 an associated cookie. A defect in the implementation of DTLSv1_listen means
2702 that state is preserved in the SSL object from one invocation to the next
2703 that can lead to a segmentation fault. Errors processing the initial
2704 ClientHello can trigger this scenario. An example of such an error could be
2705 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2708 This issue was reported to OpenSSL by Per Allansson.
2712 *) Segmentation fault in ASN1_TYPE_cmp fix
2714 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2715 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2716 certificate signature algorithm consistency this can be used to crash any
2717 certificate verification operation and exploited in a DoS attack. Any
2718 application which performs certificate verification is vulnerable including
2719 OpenSSL clients and servers which enable client authentication.
2723 *) Segmentation fault for invalid PSS parameters fix
2725 The signature verification routines will crash with a NULL pointer
2726 dereference if presented with an ASN.1 signature using the RSA PSS
2727 algorithm and invalid parameters. Since these routines are used to verify
2728 certificate signature algorithms this can be used to crash any
2729 certificate verification operation and exploited in a DoS attack. Any
2730 application which performs certificate verification is vulnerable including
2731 OpenSSL clients and servers which enable client authentication.
2733 This issue was was reported to OpenSSL by Brian Carpenter.
2737 *) ASN.1 structure reuse memory corruption fix
2739 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2740 memory corruption via an invalid write. Such reuse is and has been
2741 strongly discouraged and is believed to be rare.
2743 Applications that parse structures containing CHOICE or ANY DEFINED BY
2744 components may be affected. Certificate parsing (d2i_X509 and related
2745 functions) are however not affected. OpenSSL clients and servers are
2750 *) PKCS7 NULL pointer dereferences fix
2752 The PKCS#7 parsing code does not handle missing outer ContentInfo
2753 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2754 missing content and trigger a NULL pointer dereference on parsing.
2756 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2757 otherwise parse PKCS#7 structures from untrusted sources are
2758 affected. OpenSSL clients and servers are not affected.
2760 This issue was reported to OpenSSL by Michal Zalewski (Google).
2764 *) DoS via reachable assert in SSLv2 servers fix
2766 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2767 servers that both support SSLv2 and enable export cipher suites by sending
2768 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2770 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2771 (OpenSSL development team).
2775 *) Empty CKE with client auth and DHE fix
2777 If client auth is used then a server can seg fault in the event of a DHE
2778 ciphersuite being selected and a zero length ClientKeyExchange message
2779 being sent by the client. This could be exploited in a DoS attack.
2783 *) Handshake with unseeded PRNG fix
2785 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2786 with an unseeded PRNG. The conditions are:
2787 - The client is on a platform where the PRNG has not been seeded
2788 automatically, and the user has not seeded manually
2789 - A protocol specific client method version has been used (i.e. not
2790 SSL_client_methodv23)
2791 - A ciphersuite is used that does not require additional random data from
2792 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2794 If the handshake succeeds then the client random that has been used will
2795 have been generated from a PRNG with insufficient entropy and therefore the
2796 output may be predictable.
2798 For example using the following command with an unseeded openssl will
2799 succeed on an unpatched platform:
2801 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2805 *) Use After Free following d2i_ECPrivatekey error fix
2807 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2808 could cause a use after free condition. This, in turn, could cause a double
2809 free in several private key parsing functions (such as d2i_PrivateKey
2810 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2811 for applications that receive EC private keys from untrusted
2812 sources. This scenario is considered rare.
2814 This issue was discovered by the BoringSSL project and fixed in their
2819 *) X509_to_X509_REQ NULL pointer deref fix
2821 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2822 the certificate key is invalid. This function is rarely used in practice.
2824 This issue was discovered by Brian Carpenter.
2828 *) Removed the export ciphers from the DEFAULT ciphers
2831 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2833 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2834 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2835 So far those who have to target multiple platforms would compromise
2836 and argue that binary targeting say ARMv5 would still execute on
2837 ARMv8. "Universal" build resolves this compromise by providing
2838 near-optimal performance even on newer platforms.
2841 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2842 (other platforms pending).
2843 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2845 *) Add support for the SignedCertificateTimestampList certificate and
2846 OCSP response extensions from RFC6962.
2849 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2850 for corner cases. (Certain input points at infinity could lead to
2851 bogus results, with non-infinity inputs mapped to infinity too.)
2854 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2855 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2856 common cases are optimized and there still is room for further
2857 improvements. Vector Permutation AES for Altivec is also added.
2860 *) Add support for little-endian ppc64 Linux target.
2861 [Marcelo Cerri (IBM)]
2863 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2864 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2865 are optimized and there still is room for further improvements.
2866 Both 32- and 64-bit modes are supported.
2867 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2869 *) Improved ARMv7 NEON support.
2872 *) Support for SPARC Architecture 2011 crypto extensions, first
2873 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2874 SHA256/512, MD5, GHASH and modular exponentiation.
2875 [Andy Polyakov, David Miller]
2877 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2879 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2881 *) Support for new and upcoming Intel processors, including AVX2,
2882 BMI and SHA ISA extensions. This includes additional "stitched"
2883 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2886 This work was sponsored by Intel Corp.
2889 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2890 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2891 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2894 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2895 this fixes a limitation in previous versions of OpenSSL.
2898 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2899 MGF1 digest and OAEP label.
2902 *) Add EVP support for key wrapping algorithms, to avoid problems with
2903 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2904 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2905 algorithms and include tests cases.
2908 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2910 [Douglas E. Engert, Steve Henson]
2912 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2913 difference in days and seconds between two tm or ASN1_TIME structures.
2916 *) Add -rev test option to s_server to just reverse order of characters
2917 received by client and send back to server. Also prints an abbreviated
2918 summary of the connection parameters.
2921 *) New option -brief for s_client and s_server to print out a brief summary
2922 of connection parameters.
2925 *) Add callbacks for arbitrary TLS extensions.
2926 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2928 *) New option -crl_download in several openssl utilities to download CRLs
2929 from CRLDP extension in certificates.
2932 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2935 *) New function X509_CRL_diff to generate a delta CRL from the difference
2936 of two full CRLs. Add support to "crl" utility.
2939 *) New functions to set lookup_crls function and to retrieve
2940 X509_STORE from X509_STORE_CTX.
2943 *) Print out deprecated issuer and subject unique ID fields in
2947 *) Extend OCSP I/O functions so they can be used for simple general purpose
2948 HTTP as well as OCSP. New wrapper function which can be used to download
2949 CRLs using the OCSP API.
2952 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2955 *) SSL_CONF* functions. These provide a common framework for application
2956 configuration using configuration files or command lines.
2959 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2960 message callback and prints the results. Needs compile time option
2961 "enable-ssl-trace". New options to s_client and s_server to enable
2965 *) New ctrl and macro to retrieve supported points extensions.
2966 Print out extension in s_server and s_client.
2969 *) New functions to retrieve certificate signature and signature
2973 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2977 *) New Suite B modes for TLS code. These use and enforce the requirements
2978 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2979 only use Suite B curves. The Suite B modes can be set by using the
2980 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2983 *) New chain verification flags for Suite B levels of security. Check
2984 algorithms are acceptable when flags are set in X509_verify_cert.
2987 *) Make tls1_check_chain return a set of flags indicating checks passed
2988 by a certificate chain. Add additional tests to handle client
2989 certificates: checks for matching certificate type and issuer name
2993 *) If an attempt is made to use a signature algorithm not in the peer
2994 preference list abort the handshake. If client has no suitable
2995 signature algorithms in response to a certificate request do not
2996 use the certificate.
2999 *) If server EC tmp key is not in client preference list abort handshake.
3002 *) Add support for certificate stores in CERT structure. This makes it
3003 possible to have different stores per SSL structure or one store in
3004 the parent SSL_CTX. Include distinct stores for certificate chain
3005 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3006 to build and store a certificate chain in CERT structure: returning
3007 an error if the chain cannot be built: this will allow applications
3008 to test if a chain is correctly configured.
3010 Note: if the CERT based stores are not set then the parent SSL_CTX
3011 store is used to retain compatibility with existing behaviour.
3015 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3016 mask based on the current session, check mask when sending client
3017 hello and checking the requested ciphersuite.
3020 *) New ctrls to retrieve and set certificate types in a certificate
3021 request message. Print out received values in s_client. If certificate
3022 types is not set with custom values set sensible values based on
3023 supported signature algorithms.
3026 *) Support for distinct client and server supported signature algorithms.
3029 *) Add certificate callback. If set this is called whenever a certificate
3030 is required by client or server. An application can decide which
3031 certificate chain to present based on arbitrary criteria: for example
3032 supported signature algorithms. Add very simple example to s_server.
3033 This fixes many of the problems and restrictions of the existing client
3034 certificate callback: for example you can now clear an existing
3035 certificate and specify the whole chain.
3038 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3039 the certificate can be used for (if anything). Set valid_flags field
3040 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3041 to have similar checks in it.
3043 Add new "cert_flags" field to CERT structure and include a "strict mode".
3044 This enforces some TLS certificate requirements (such as only permitting
3045 certificate signature algorithms contained in the supported algorithms
3046 extension) which some implementations ignore: this option should be used
3047 with caution as it could cause interoperability issues.
3050 *) Update and tidy signature algorithm extension processing. Work out
3051 shared signature algorithms based on preferences and peer algorithms
3052 and print them out in s_client and s_server. Abort handshake if no
3053 shared signature algorithms.
3056 *) Add new functions to allow customised supported signature algorithms
3057 for SSL and SSL_CTX structures. Add options to s_client and s_server
3061 *) New function SSL_certs_clear() to delete all references to certificates
3062 from an SSL structure. Before this once a certificate had been added
3063 it couldn't be removed.
3066 *) Integrate hostname, email address and IP address checking with certificate
3067 verification. New verify options supporting checking in openssl utility.
3070 *) Fixes and wildcard matching support to hostname and email checking
3071 functions. Add manual page.
3072 [Florian Weimer (Red Hat Product Security Team)]
3074 *) New functions to check a hostname email or IP address against a
3075 certificate. Add options x509 utility to print results of checks against
3079 *) Fix OCSP checking.
3080 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3082 *) Initial experimental support for explicitly trusted non-root CAs.
3083 OpenSSL still tries to build a complete chain to a root but if an
3084 intermediate CA has a trust setting included that is used. The first
3085 setting is used: whether to trust (e.g., -addtrust option to the x509
3089 *) Add -trusted_first option which attempts to find certificates in the
3090 trusted store even if an untrusted chain is also supplied.
3093 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3094 platform support for Linux and Android.
3097 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3100 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3101 When in FIPS mode the approved implementations are used as normal,
3102 when not in FIPS mode the internal unapproved versions are used instead.
3103 This means that the FIPS capable OpenSSL isn't forced to use the
3104 (often lower performance) FIPS implementations outside FIPS mode.
3107 *) Transparently support X9.42 DH parameters when calling
3108 PEM_read_bio_DHparameters. This means existing applications can handle
3109 the new parameter format automatically.
3112 *) Initial experimental support for X9.42 DH parameter format: mainly
3113 to support use of 'q' parameter for RFC5114 parameters.
3116 *) Add DH parameters from RFC5114 including test data to dhtest.
3119 *) Support for automatic EC temporary key parameter selection. If enabled
3120 the most preferred EC parameters are automatically used instead of
3121 hardcoded fixed parameters. Now a server just has to call:
3122 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3123 support ECDH and use the most appropriate parameters.
3126 *) Enhance and tidy EC curve and point format TLS extension code. Use
3127 static structures instead of allocation if default values are used.
3128 New ctrls to set curves we wish to support and to retrieve shared curves.
3129 Print out shared curves in s_server. New options to s_server and s_client
3130 to set list of supported curves.
3133 *) New ctrls to retrieve supported signature algorithms and
3134 supported curve values as an array of NIDs. Extend openssl utility
3135 to print out received values.
3138 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3139 between NIDs and the more common NIST names such as "P-256". Enhance
3140 ecparam utility and ECC method to recognise the NIST names for curves.
3143 *) Enhance SSL/TLS certificate chain handling to support different
3144 chains for each certificate instead of one chain in the parent SSL_CTX.
3147 *) Support for fixed DH ciphersuite client authentication: where both
3148 server and client use DH certificates with common parameters.
3151 *) Support for fixed DH ciphersuites: those requiring DH server
3155 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3157 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3158 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3159 X509_CINF_get_signature were reverted post internal team review.
3161 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3163 *) Build fixes for the Windows and OpenVMS platforms
3164 [Matt Caswell and Richard Levitte]
3166 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3168 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3169 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3170 dereference. This could lead to a Denial Of Service attack. Thanks to
3171 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3175 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3176 dtls1_buffer_record function under certain conditions. In particular this
3177 could occur if an attacker sent repeated DTLS records with the same
3178 sequence number but for the next epoch. The memory leak could be exploited
3179 by an attacker in a Denial of Service attack through memory exhaustion.
3180 Thanks to Chris Mueller for reporting this issue.
3184 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3185 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3186 method would be set to NULL which could later result in a NULL pointer
3187 dereference. Thanks to Frank Schmirler for reporting this issue.
3191 *) Abort handshake if server key exchange message is omitted for ephemeral
3194 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3195 reporting this issue.
3199 *) Remove non-export ephemeral RSA code on client and server. This code
3200 violated the TLS standard by allowing the use of temporary RSA keys in
3201 non-export ciphersuites and could be used by a server to effectively
3202 downgrade the RSA key length used to a value smaller than the server
3203 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3204 INRIA or reporting this issue.
3208 *) Fixed issue where DH client certificates are accepted without verification.
3209 An OpenSSL server will accept a DH certificate for client authentication
3210 without the certificate verify message. This effectively allows a client to
3211 authenticate without the use of a private key. This only affects servers
3212 which trust a client certificate authority which issues certificates
3213 containing DH keys: these are extremely rare and hardly ever encountered.
3214 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3219 *) Ensure that the session ID context of an SSL is updated when its
3220 SSL_CTX is updated via SSL_set_SSL_CTX.
3222 The session ID context is typically set from the parent SSL_CTX,
3223 and can vary with the CTX.
3226 *) Fix various certificate fingerprint issues.
3228 By using non-DER or invalid encodings outside the signed portion of a
3229 certificate the fingerprint can be changed without breaking the signature.
3230 Although no details of the signed portion of the certificate can be changed
3231 this can cause problems with some applications: e.g. those using the
3232 certificate fingerprint for blacklists.
3234 1. Reject signatures with non zero unused bits.
3236 If the BIT STRING containing the signature has non zero unused bits reject
3237 the signature. All current signature algorithms require zero unused bits.
3239 2. Check certificate algorithm consistency.
3241 Check the AlgorithmIdentifier inside TBS matches the one in the
3242 certificate signature. NB: this will result in signature failure
3243 errors for some broken certificates.
3245 Thanks to Konrad Kraszewski from Google for reporting this issue.
3247 3. Check DSA/ECDSA signatures use DER.
3249 Re-encode DSA/ECDSA signatures and compare with the original received
3250 signature. Return an error if there is a mismatch.
3252 This will reject various cases including garbage after signature
3253 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3254 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3255 (negative or with leading zeroes).
3257 Further analysis was conducted and fixes were developed by Stephen Henson
3258 of the OpenSSL core team.
3263 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3264 results on some platforms, including x86_64. This bug occurs at random
3265 with a very low probability, and is not known to be exploitable in any
3266 way, though its exact impact is difficult to determine. Thanks to Pieter
3267 Wuille (Blockstream) who reported this issue and also suggested an initial
3268 fix. Further analysis was conducted by the OpenSSL development team and
3269 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3270 the OpenSSL core team.
3274 *) Do not resume sessions on the server if the negotiated protocol
3275 version does not match the session's version. Resuming with a different
3276 version, while not strictly forbidden by the RFC, is of questionable
3277 sanity and breaks all known clients.
3278 [David Benjamin, Emilia Käsper]
3280 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3281 early CCS messages during renegotiation. (Note that because
3282 renegotiation is encrypted, this early CCS was not exploitable.)
3285 *) Tighten client-side session ticket handling during renegotiation:
3286 ensure that the client only accepts a session ticket if the server sends
3287 the extension anew in the ServerHello. Previously, a TLS client would
3288 reuse the old extension state and thus accept a session ticket if one was
3289 announced in the initial ServerHello.
3291 Similarly, ensure that the client requires a session ticket if one
3292 was advertised in the ServerHello. Previously, a TLS client would
3293 ignore a missing NewSessionTicket message.
3296 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3298 *) SRTP Memory Leak.
3300 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3301 sends a carefully crafted handshake message, to cause OpenSSL to fail
3302 to free up to 64k of memory causing a memory leak. This could be
3303 exploited in a Denial Of Service attack. This issue affects OpenSSL
3304 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3305 whether SRTP is used or configured. Implementations of OpenSSL that
3306 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3308 The fix was developed by the OpenSSL team.
3312 *) Session Ticket Memory Leak.
3314 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3315 integrity of that ticket is first verified. In the event of a session
3316 ticket integrity check failing, OpenSSL will fail to free memory
3317 causing a memory leak. By sending a large number of invalid session
3318 tickets an attacker could exploit this issue in a Denial Of Service
3323 *) Build option no-ssl3 is incomplete.
3325 When OpenSSL is configured with "no-ssl3" as a build option, servers
3326 could accept and complete a SSL 3.0 handshake, and clients could be
3327 configured to send them.
3329 [Akamai and the OpenSSL team]
3331 *) Add support for TLS_FALLBACK_SCSV.
3332 Client applications doing fallback retries should call
3333 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3335 [Adam Langley, Bodo Moeller]
3337 *) Add additional DigestInfo checks.
3339 Re-encode DigestInto in DER and check against the original when
3340 verifying RSA signature: this will reject any improperly encoded
3341 DigestInfo structures.
3343 Note: this is a precautionary measure and no attacks are currently known.
3347 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3349 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3350 SRP code can be overrun an internal buffer. Add sanity check that
3351 g, A, B < N to SRP code.
3353 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3354 Group for discovering this issue.
3358 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3359 TLS 1.0 instead of higher protocol versions when the ClientHello message
3360 is badly fragmented. This allows a man-in-the-middle attacker to force a
3361 downgrade to TLS 1.0 even if both the server and the client support a
3362 higher protocol version, by modifying the client's TLS records.
3364 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3365 researching this issue.
3369 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3370 to a denial of service attack. A malicious server can crash the client
3371 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3372 ciphersuite and sending carefully crafted handshake messages.
3374 Thanks to Felix Gröbert (Google) for discovering and researching this
3379 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3380 to leak memory. This can be exploited through a Denial of Service attack.
3381 Thanks to Adam Langley for discovering and researching this issue.
3385 *) An attacker can force openssl to consume large amounts of memory whilst
3386 processing DTLS handshake messages. This can be exploited through a
3387 Denial of Service attack.
3388 Thanks to Adam Langley for discovering and researching this issue.
3392 *) An attacker can force an error condition which causes openssl to crash
3393 whilst processing DTLS packets due to memory being freed twice. This
3394 can be exploited through a Denial of Service attack.
3395 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3400 *) If a multithreaded client connects to a malicious server using a resumed
3401 session and the server sends an ec point format extension it could write
3402 up to 255 bytes to freed memory.
3404 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3409 *) A malicious server can crash an OpenSSL client with a null pointer
3410 dereference (read) by specifying an SRP ciphersuite even though it was not
3411 properly negotiated with the client. This can be exploited through a
3412 Denial of Service attack.
3414 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3415 discovering and researching this issue.
3419 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3420 X509_name_oneline, X509_name_print_ex et al. to leak some information
3421 from the stack. Applications may be affected if they echo pretty printing
3422 output to the attacker.
3424 Thanks to Ivan Fratric (Google) for discovering this issue.
3426 [Emilia Käsper, and Steve Henson]
3428 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3429 for corner cases. (Certain input points at infinity could lead to
3430 bogus results, with non-infinity inputs mapped to infinity too.)
3433 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3435 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3436 handshake can force the use of weak keying material in OpenSSL
3437 SSL/TLS clients and servers.
3439 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3440 researching this issue. (CVE-2014-0224)
3441 [KIKUCHI Masashi, Steve Henson]
3443 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3444 OpenSSL DTLS client the code can be made to recurse eventually crashing
3447 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3449 [Imre Rad, Steve Henson]
3451 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3452 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3453 client or server. This is potentially exploitable to run arbitrary
3454 code on a vulnerable client or server.
3456 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3457 [Jüri Aedla, Steve Henson]
3459 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3460 are subject to a denial of service attack.
3462 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3463 this issue. (CVE-2014-3470)
3464 [Felix Gröbert, Ivan Fratric, Steve Henson]
3466 *) Harmonize version and its documentation. -f flag is used to display
3468 [mancha <mancha1@zoho.com>]
3470 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3471 in i2d_ECPrivateKey.
3472 [mancha <mancha1@zoho.com>]
3474 *) Fix some double frees. These are not thought to be exploitable.
3475 [mancha <mancha1@zoho.com>]
3477 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3479 *) A missing bounds check in the handling of the TLS heartbeat extension
3480 can be used to reveal up to 64k of memory to a connected client or
3483 Thanks for Neel Mehta of Google Security for discovering this bug and to
3484 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3485 preparing the fix (CVE-2014-0160)
3486 [Adam Langley, Bodo Moeller]
3488 *) Fix for the attack described in the paper "Recovering OpenSSL
3489 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3490 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3491 http://eprint.iacr.org/2014/140
3493 Thanks to Yuval Yarom and Naomi Benger for discovering this
3494 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3495 [Yuval Yarom and Naomi Benger]
3497 *) TLS pad extension: draft-agl-tls-padding-03
3499 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3500 TLS client Hello record length value would otherwise be > 255 and
3501 less that 512 pad with a dummy extension containing zeroes so it
3502 is at least 512 bytes long.
3504 [Adam Langley, Steve Henson]
3506 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3508 *) Fix for TLS record tampering bug. A carefully crafted invalid
3509 handshake could crash OpenSSL with a NULL pointer exception.
3510 Thanks to Anton Johansson for reporting this issues.
3513 *) Keep original DTLS digest and encryption contexts in retransmission
3514 structures so we can use the previous session parameters if they need
3515 to be resent. (CVE-2013-6450)
3518 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3519 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3520 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3521 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3522 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3523 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3524 [Rob Stradling, Adam Langley]
3526 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3528 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3529 supporting platforms or when small records were transferred.
3530 [Andy Polyakov, Steve Henson]
3532 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3534 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3536 This addresses the flaw in CBC record processing discovered by
3537 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3538 at: http://www.isg.rhul.ac.uk/tls/
3540 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3541 Security Group at Royal Holloway, University of London
3542 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3543 Emilia Käsper for the initial patch.
3545 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3547 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3548 ciphersuites which can be exploited in a denial of service attack.
3549 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3550 and detecting this bug and to Wolfgang Ettlinger
3551 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3555 *) Return an error when checking OCSP signatures when key is NULL.
3556 This fixes a DoS attack. (CVE-2013-0166)
3559 *) Make openssl verify return errors.
3560 [Chris Palmer <palmer@google.com> and Ben Laurie]
3562 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3563 the right response is stapled. Also change SSL_get_certificate()
3564 so it returns the certificate actually sent.
3565 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3566 [Rob Stradling <rob.stradling@comodo.com>]
3568 *) Fix possible deadlock when decoding public keys.
3571 *) Don't use TLS 1.0 record version number in initial client hello
3575 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3577 *) Sanity check record length before skipping explicit IV in TLS
3578 1.2, 1.1 and DTLS to fix DoS attack.
3580 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3581 fuzzing as a service testing platform.
3585 *) Initialise tkeylen properly when encrypting CMS messages.
3586 Thanks to Solar Designer of Openwall for reporting this issue.
3589 *) In FIPS mode don't try to use composite ciphers as they are not
3593 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3595 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3596 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3597 mean any application compiled against OpenSSL 1.0.0 headers setting
3598 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3599 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3600 0x10000000L Any application which was previously compiled against
3601 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3602 will need to be recompiled as a result. Letting be results in
3603 inability to disable specifically TLS 1.1 and in client context,
3604 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3607 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3608 disable just protocol X, but all protocols above X *if* there are
3609 protocols *below* X still enabled. In more practical terms it means
3610 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3611 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3612 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3616 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3618 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3619 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3620 in CRYPTO_realloc_clean.
3622 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3623 issue and to Adam Langley <agl@chromium.org> for fixing it.
3625 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3627 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3630 *) Workarounds for some broken servers that "hang" if a client hello
3631 record length exceeds 255 bytes.
3633 1. Do not use record version number > TLS 1.0 in initial client
3634 hello: some (but not all) hanging servers will now work.
3635 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3636 the number of ciphers sent in the client hello. This should be
3637 set to an even number, such as 50, for example by passing:
3638 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3639 Most broken servers should now work.
3640 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3641 TLS 1.2 client support entirely.
3644 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3647 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3649 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3650 STRING form instead of a DigestInfo.
3653 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3654 and the RSA_sign/RSA_verify functions. This was made more apparent when
3655 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3656 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3657 the correct format in RSA_verify so both forms transparently work.
3660 *) Some servers which support TLS 1.0 can choke if we initially indicate
3661 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3662 encrypted premaster secret. As a workaround use the maximum permitted
3663 client version in client hello, this should keep such servers happy
3664 and still work with previous versions of OpenSSL.
3667 *) Add support for TLS/DTLS heartbeats.
3668 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3670 *) Add support for SCTP.
3671 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3673 *) Improved PRNG seeding for VOS.
3674 [Paul Green <Paul.Green@stratus.com>]
3676 *) Extensive assembler packs updates, most notably:
3678 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3679 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3680 - x86_64: bit-sliced AES implementation;
3681 - ARM: NEON support, contemporary platforms optimizations;
3682 - s390x: z196 support;
3683 - *: GHASH and GF(2^m) multiplication implementations;
3687 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3688 (removal of unnecessary code)
3689 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3691 *) Add TLS key material exporter from RFC 5705.
3694 *) Add DTLS-SRTP negotiation from RFC 5764.
3697 *) Add Next Protocol Negotiation,
3698 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3699 disabled with a no-npn flag to config or Configure. Code donated
3701 [Adam Langley <agl@google.com> and Ben Laurie]
3703 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3704 NIST-P256, NIST-P521, with constant-time single point multiplication on
3705 typical inputs. Compiler support for the nonstandard type __uint128_t is
3706 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3707 Code made available under Apache License version 2.0.
3709 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3710 line to include this in your build of OpenSSL, and run "make depend" (or
3711 "make update"). This enables the following EC_METHODs:
3713 EC_GFp_nistp224_method()
3714 EC_GFp_nistp256_method()
3715 EC_GFp_nistp521_method()
3717 EC_GROUP_new_by_curve_name() will automatically use these (while
3718 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3720 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3722 *) Use type ossl_ssize_t instead of ssize_t which isn't available on
3723 all platforms. Move ssize_t definition from e_os.h to the public
3724 header file e_os2.h as it now appears in public header file cms.h
3727 *) New -sigopt option to the ca, req and x509 utilities. Additional
3728 signature parameters can be passed using this option and in
3732 *) Add RSA PSS signing function. This will generate and set the
3733 appropriate AlgorithmIdentifiers for PSS based on those in the
3734 corresponding EVP_MD_CTX structure. No application support yet.
3737 *) Support for companion algorithm specific ASN1 signing routines.
3738 New function ASN1_item_sign_ctx() signs a pre-initialised
3739 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3740 the appropriate parameters.
3743 *) Add new algorithm specific ASN1 verification initialisation function
3744 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3745 handling will be the same no matter what EVP_PKEY_METHOD is used.
3746 Add a PSS handler to support verification of PSS signatures: checked
3747 against a number of sample certificates.
3750 *) Add signature printing for PSS. Add PSS OIDs.
3751 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3753 *) Add algorithm specific signature printing. An individual ASN1 method
3754 can now print out signatures instead of the standard hex dump.
3756 More complex signatures (e.g. PSS) can print out more meaningful
3757 information. Include DSA version that prints out the signature
3761 *) Password based recipient info support for CMS library: implementing
3765 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3766 neatly separates the code into cipher and PBE sections and is required
3767 for some algorithms that split PBES2 into separate pieces (such as
3768 password based CMS).
3771 *) Session-handling fixes:
3772 - Fix handling of connections that are resuming with a session ID,
3773 but also support Session Tickets.
3774 - Fix a bug that suppressed issuing of a new ticket if the client
3775 presented a ticket with an expired session.
3776 - Try to set the ticket lifetime hint to something reasonable.
3777 - Make tickets shorter by excluding irrelevant information.
3778 - On the client side, don't ignore renewed tickets.
3779 [Adam Langley, Bodo Moeller (Google)]
3781 *) Fix PSK session representation.
3784 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3786 This work was sponsored by Intel.
3789 *) Add GCM support to TLS library. Some custom code is needed to split
3790 the IV between the fixed (from PRF) and explicit (from TLS record)
3791 portions. This adds all GCM ciphersuites supported by RFC5288 and
3792 RFC5289. Generalise some AES* cipherstrings to include GCM and
3793 add a special AESGCM string for GCM only.
3796 *) Expand range of ctrls for AES GCM. Permit setting invocation
3797 field on decrypt and retrieval of invocation field only on encrypt.
3800 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3801 As required by RFC5289 these ciphersuites cannot be used if for
3802 versions of TLS earlier than 1.2.
3805 *) For FIPS capable OpenSSL interpret a NULL default public key method
3806 as unset and return the appropriate default but do *not* set the default.
3807 This means we can return the appropriate method in applications that
3808 switch between FIPS and non-FIPS modes.
3811 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3812 ENGINE is used then we cannot handle that in the FIPS module so we
3813 keep original code iff non-FIPS operations are allowed.
3816 *) Add -attime option to openssl utilities.
3817 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3819 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3822 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3823 FIPS EC methods unconditionally for now.
3826 *) New build option no-ec2m to disable characteristic 2 code.
3829 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3830 all cases can be covered as some introduce binary incompatibilities.
3833 *) Redirect RSA operations to FIPS module including keygen,
3834 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3837 *) Add similar low level API blocking to ciphers.
3840 *) Low level digest APIs are not approved in FIPS mode: any attempt
3841 to use these will cause a fatal error. Applications that *really* want
3842 to use them can use the private_* version instead.
3845 *) Redirect cipher operations to FIPS module for FIPS builds.
3848 *) Redirect digest operations to FIPS module for FIPS builds.
3851 *) Update build system to add "fips" flag which will link in fipscanister.o
3852 for static and shared library builds embedding a signature if needed.
3855 *) Output TLS supported curves in preference order instead of numerical
3856 order. This is currently hardcoded for the highest order curves first.
3857 This should be configurable so applications can judge speed vs strength.
3860 *) Add TLS v1.2 server support for client authentication.
3863 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3867 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3868 FIPS modules versions.
3871 *) Add TLS v1.2 client side support for client authentication. Keep cache
3872 of handshake records longer as we don't know the hash algorithm to use
3873 until after the certificate request message is received.
3876 *) Initial TLS v1.2 client support. Add a default signature algorithms
3877 extension including all the algorithms we support. Parse new signature
3878 format in client key exchange. Relax some ECC signing restrictions for
3879 TLS v1.2 as indicated in RFC5246.
3882 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3883 to new signature format when needed using client digest preference.
3884 All server ciphersuites should now work correctly in TLS v1.2. No client
3885 support yet and no support for client certificates.
3888 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3889 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3890 ciphersuites. At present only RSA key exchange ciphersuites work with
3891 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3892 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3893 and version checking.
3896 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3897 with this defined it will not be affected by any changes to ssl internal
3898 structures. Add several utility functions to allow openssl application
3899 to work with OPENSSL_NO_SSL_INTERN defined.
3902 *) A long standing patch to add support for SRP from EdelWeb (Peter
3903 Sylvester and Christophe Renou) was integrated.
3904 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3905 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3908 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3911 *) Permit abbreviated handshakes when renegotiating using the function
3912 SSL_renegotiate_abbreviated().
3913 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3915 *) Add call to ENGINE_register_all_complete() to
3916 ENGINE_load_builtin_engines(), so some implementations get used
3917 automatically instead of needing explicit application support.
3920 *) Add support for TLS key exporter as described in RFC5705.
3921 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3923 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3924 a few changes are required:
3926 Add SSL_OP_NO_TLSv1_1 flag.
3927 Add TLSv1_1 methods.
3928 Update version checking logic to handle version 1.1.
3929 Add explicit IV handling (ported from DTLS code).
3930 Add command line options to s_client/s_server.
3933 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3935 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3936 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3937 content decryption and always return the same error. Note: this attack
3938 needs on average 2^20 messages so it only affects automated senders. The
3939 old behaviour can be re-enabled in the CMS code by setting the
3940 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3941 an MMA defence is not necessary.
3942 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3943 this issue. (CVE-2012-0884)
3946 *) Fix CVE-2011-4619: make sure we really are receiving a
3947 client hello before rejecting multiple SGC restarts. Thanks to
3948 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3951 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3953 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3954 Thanks to Antonio Martin, Enterprise Secure Access Research and
3955 Development, Cisco Systems, Inc. for discovering this bug and
3956 preparing a fix. (CVE-2012-0050)
3959 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3961 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3962 of the Vaudenay padding oracle attack on CBC mode encryption
3963 which enables an efficient plaintext recovery attack against
3964 the OpenSSL implementation of DTLS. Their attack exploits timing
3965 differences arising during decryption processing. A research
3966 paper describing this attack can be found at:
3967 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3968 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3969 Security Group at Royal Holloway, University of London
3970 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3971 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3972 for preparing the fix. (CVE-2011-4108)
3973 [Robin Seggelmann, Michael Tuexen]
3975 *) Clear bytes used for block padding of SSL 3.0 records.
3977 [Adam Langley (Google)]
3979 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3980 Kadianakis <desnacked@gmail.com> for discovering this issue and
3981 Adam Langley for preparing the fix. (CVE-2011-4619)
3982 [Adam Langley (Google)]
3984 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3985 [Andrey Kulikov <amdeich@gmail.com>]
3987 *) Prevent malformed RFC3779 data triggering an assertion failure.
3988 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3989 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3990 [Rob Austein <sra@hactrn.net>]
3992 *) Improved PRNG seeding for VOS.
3993 [Paul Green <Paul.Green@stratus.com>]
3995 *) Fix ssl_ciph.c set-up race.
3996 [Adam Langley (Google)]
3998 *) Fix spurious failures in ecdsatest.c.
3999 [Emilia Käsper (Google)]
4001 *) Fix the BIO_f_buffer() implementation (which was mixing different
4002 interpretations of the '..._len' fields).
4003 [Adam Langley (Google)]
4005 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4006 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4007 threads won't reuse the same blinding coefficients.
4009 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4010 lock to call BN_BLINDING_invert_ex, and avoids one use of
4011 BN_BLINDING_update for each BN_BLINDING structure (previously,
4012 the last update always remained unused).
4013 [Emilia Käsper (Google)]
4015 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4016 [Bob Buckholz (Google)]
4018 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4020 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4021 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4022 [Kaspar Brand <ossl@velox.ch>]
4024 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4025 for multi-threaded use of ECDH. (CVE-2011-3210)
4026 [Adam Langley (Google)]
4028 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4031 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4032 signature public key algorithm by using OID xref utilities instead.
4033 Before this you could only use some ECC ciphersuites with SHA1 only.
4036 *) Add protection against ECDSA timing attacks as mentioned in the paper
4037 by Billy Bob Brumley and Nicola Tuveri, see:
4039 http://eprint.iacr.org/2011/232.pdf
4041 [Billy Bob Brumley and Nicola Tuveri]
4043 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4045 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4046 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4048 *) Fix bug in string printing code: if *any* escaping is enabled we must
4049 escape the escape character (backslash) or the resulting string is
4053 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4055 *) Disable code workaround for ancient and obsolete Netscape browsers
4056 and servers: an attacker can use it in a ciphersuite downgrade attack.
4057 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4060 *) Fixed J-PAKE implementation error, originally discovered by
4061 Sebastien Martini, further info and confirmation from Stefan
4062 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4065 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4067 *) Fix extension code to avoid race conditions which can result in a buffer
4068 overrun vulnerability: resumed sessions must not be modified as they can
4069 be shared by multiple threads. CVE-2010-3864
4072 *) Fix WIN32 build system to correctly link an ENGINE directory into
4076 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4078 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4080 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4082 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4084 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4085 context. The operation can be customised via the ctrl mechanism in
4086 case ENGINEs want to include additional functionality.
4089 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4092 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4093 output hashes compatible with older versions of OpenSSL.
4094 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4096 *) Fix compression algorithm handling: if resuming a session use the
4097 compression algorithm of the resumed session instead of determining
4098 it from client hello again. Don't allow server to change algorithm.
4101 *) Add load_crls() function to apps tidying load_certs() too. Add option
4102 to verify utility to allow additional CRLs to be included.
4105 *) Update OCSP request code to permit adding custom headers to the request:
4106 some responders need this.
4109 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4111 [Julia Lawall <julia@diku.dk>]
4113 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4114 needlessly dereferenced structures, used obsolete functions and
4115 didn't handle all updated verify codes correctly.
4118 *) Disable MD2 in the default configuration.
4121 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4122 indicate the initial BIO being pushed or popped. This makes it possible
4123 to determine whether the BIO is the one explicitly called or as a result
4124 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4125 it handles reference counts correctly and doesn't zero out the I/O bio
4126 when it is not being explicitly popped. WARNING: applications which
4127 included workarounds for the old buggy behaviour will need to be modified
4128 or they could free up already freed BIOs.
4131 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4132 renaming to all platforms (within the 0.9.8 branch, this was
4133 done conditionally on Netware platforms to avoid a name clash).
4134 [Guenter <lists@gknw.net>]
4136 *) Add ECDHE and PSK support to DTLS.
4137 [Michael Tuexen <tuexen@fh-muenster.de>]
4139 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4143 *) Add "missing" function EVP_MD_flags() (without this the only way to
4144 retrieve a digest flags is by accessing the structure directly. Update
4145 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4146 or cipher is registered as in the "from" argument. Print out all
4147 registered digests in the dgst usage message instead of manually
4148 attempting to work them out.
4151 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4152 this allows the use of compression and extensions. Change default cipher
4153 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4154 by default unless an application cipher string requests it.
4157 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4158 key ids to find matching certificates and keys but some PKCS#12 files
4159 don't follow the (somewhat unwritten) rules and this strategy fails.
4160 Now just gather all certificates together and the first private key
4161 then look for the first certificate that matches the key.
4164 *) Support use of registered digest and cipher names for dgst and cipher
4165 commands instead of having to add each one as a special case. So now
4172 openssl dgst -sha256 foo
4174 and this works for ENGINE based algorithms too.
4178 *) Update Gost ENGINE to support parameter files.
4179 [Victor B. Wagner <vitus@cryptocom.ru>]
4181 *) Support GeneralizedTime in ca utility.
4182 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4184 *) Enhance the hash format used for certificate directory links. The new
4185 form uses the canonical encoding (meaning equivalent names will work
4186 even if they aren't identical) and uses SHA1 instead of MD5. This form
4187 is incompatible with the older format and as a result c_rehash should
4188 be used to rebuild symbolic links.
4191 *) Make PKCS#8 the default write format for private keys, replacing the
4192 traditional format. This form is standardised, more secure and doesn't
4193 include an implicit MD5 dependency.
4196 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4197 committed to OpenSSL should pass this lot as a minimum.
4200 *) Add session ticket override functionality for use by EAP-FAST.
4201 [Jouni Malinen <j@w1.fi>]
4203 *) Modify HMAC functions to return a value. Since these can be implemented
4204 in an ENGINE errors can occur.
4207 *) Type-checked OBJ_bsearch_ex.
4210 *) Type-checked OBJ_bsearch. Also some constification necessitated
4211 by type-checking. Still to come: TXT_DB, bsearch(?),
4212 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4216 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4217 seconds to a tm structure directly, instead of going through OS
4218 specific date routines. This avoids any issues with OS routines such
4219 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4220 and X509_time_adj_ex() to cover the extended range. The existing
4221 X509_time_adj() is still usable and will no longer have any date issues.
4224 *) Delta CRL support. New use deltas option which will attempt to locate
4225 and search any appropriate delta CRLs available.
4227 This work was sponsored by Google.
4230 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4231 code and add additional score elements. Validate alternate CRL paths
4232 as part of the CRL checking and indicate a new error "CRL path validation
4233 error" in this case. Applications wanting additional details can use
4234 the verify callback and check the new "parent" field. If this is not
4235 NULL CRL path validation is taking place. Existing applications won't
4236 see this because it requires extended CRL support which is off by
4239 This work was sponsored by Google.
4242 *) Support for freshest CRL extension.
4244 This work was sponsored by Google.
4247 *) Initial indirect CRL support. Currently only supported in the CRLs
4248 passed directly and not via lookup. Process certificate issuer
4249 CRL entry extension and lookup CRL entries by bother issuer name
4250 and serial number. Check and process CRL issuer entry in IDP extension.
4252 This work was sponsored by Google.
4255 *) Add support for distinct certificate and CRL paths. The CRL issuer
4256 certificate is validated separately in this case. Only enabled if
4257 an extended CRL support flag is set: this flag will enable additional
4258 CRL functionality in future.
4260 This work was sponsored by Google.
4263 *) Add support for policy mappings extension.
4265 This work was sponsored by Google.
4268 *) Fixes to pathlength constraint, self issued certificate handling,
4269 policy processing to align with RFC3280 and PKITS tests.
4271 This work was sponsored by Google.
4274 *) Support for name constraints certificate extension. DN, email, DNS
4275 and URI types are currently supported.
4277 This work was sponsored by Google.
4280 *) To cater for systems that provide a pointer-based thread ID rather
4281 than numeric, deprecate the current numeric thread ID mechanism and
4282 replace it with a structure and associated callback type. This
4283 mechanism allows a numeric "hash" to be extracted from a thread ID in
4284 either case, and on platforms where pointers are larger than 'long',
4285 mixing is done to help ensure the numeric 'hash' is usable even if it
4286 can't be guaranteed unique. The default mechanism is to use "&errno"
4287 as a pointer-based thread ID to distinguish between threads.
4289 Applications that want to provide their own thread IDs should now use
4290 CRYPTO_THREADID_set_callback() to register a callback that will call
4291 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4293 Note that ERR_remove_state() is now deprecated, because it is tied
4294 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4295 to free the current thread's error state should be replaced by
4296 ERR_remove_thread_state(NULL).
4298 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4299 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4300 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4301 application was previously providing a numeric thread callback that
4302 was inappropriate for distinguishing threads, then uniqueness might
4303 have been obtained with &errno that happened immediately in the
4304 intermediate development versions of OpenSSL; this is no longer the
4305 case, the numeric thread callback will now override the automatic use
4307 [Geoff Thorpe, with help from Bodo Moeller]
4309 *) Initial support for different CRL issuing certificates. This covers a
4310 simple case where the self issued certificates in the chain exist and
4311 the real CRL issuer is higher in the existing chain.
4313 This work was sponsored by Google.
4316 *) Removed effectively defunct crypto/store from the build.
4319 *) Revamp of STACK to provide stronger type-checking. Still to come:
4320 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4321 ASN1_STRING, CONF_VALUE.
4324 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4325 RAM on SSL connections. This option can save about 34k per idle SSL.
4328 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4329 STACK, TXT_DB, bsearch, qsort.
4332 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4333 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4334 support for data, signedData, compressedData, digestedData and
4335 encryptedData, envelopedData types included. Scripts to check against
4336 RFC4134 examples draft and interop and consistency checks of many
4337 content types and variants.
4340 *) Add options to enc utility to support use of zlib compression BIO.
4343 *) Extend mk1mf to support importing of options and assembly language
4344 files from Configure script, currently only included in VC-WIN32.
4345 The assembly language rules can now optionally generate the source
4346 files from the associated perl scripts.
4349 *) Implement remaining functionality needed to support GOST ciphersuites.
4350 Interop testing has been performed using CryptoPro implementations.
4351 [Victor B. Wagner <vitus@cryptocom.ru>]
4353 *) s390x assembler pack.
4356 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4360 *) Implement Opaque PRF Input TLS extension as specified in
4361 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4362 official specification yet and no extension type assignment by
4363 IANA exists, this extension (for now) will have to be explicitly
4364 enabled when building OpenSSL by providing the extension number
4365 to use. For example, specify an option
4367 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4369 to the "config" or "Configure" script to enable the extension,
4370 assuming extension number 0x9527 (which is a completely arbitrary
4371 and unofficial assignment based on the MD5 hash of the Internet
4372 Draft). Note that by doing so, you potentially lose
4373 interoperability with other TLS implementations since these might
4374 be using the same extension number for other purposes.
4376 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4377 opaque PRF input value to use in the handshake. This will create
4378 an internal copy of the length-'len' string at 'src', and will
4379 return non-zero for success.
4381 To get more control and flexibility, provide a callback function
4384 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4385 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4389 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4392 Callback function 'cb' will be called in handshakes, and is
4393 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4394 Argument 'arg' is for application purposes (the value as given to
4395 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4396 be provided to the callback function). The callback function
4397 has to return non-zero to report success: usually 1 to use opaque
4398 PRF input just if possible, or 2 to enforce use of the opaque PRF
4399 input. In the latter case, the library will abort the handshake
4400 if opaque PRF input is not successfully negotiated.
4402 Arguments 'peerinput' and 'len' given to the callback function
4403 will always be NULL and 0 in the case of a client. A server will
4404 see the client's opaque PRF input through these variables if
4405 available (NULL and 0 otherwise). Note that if the server
4406 provides an opaque PRF input, the length must be the same as the
4407 length of the client's opaque PRF input.
4409 Note that the callback function will only be called when creating
4410 a new session (session resumption can resume whatever was
4411 previously negotiated), and will not be called in SSL 2.0
4412 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4413 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4414 for applications that need to enforce opaque PRF input.
4418 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4421 [Victor B. Wagner <vitus@cryptocom.ru>]
4423 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4424 RFC4507bis. The encrypted ticket format is an encrypted encoded
4425 SSL_SESSION structure, that way new session features are automatically
4428 If a client application caches session in an SSL_SESSION structure
4429 support is transparent because tickets are now stored in the encoded
4432 The SSL_CTX structure automatically generates keys for ticket
4433 protection in servers so again support should be possible
4434 with no application modification.
4436 If a client or server wishes to disable RFC4507 support then the option
4437 SSL_OP_NO_TICKET can be set.
4439 Add a TLS extension debugging callback to allow the contents of any client
4440 or server extensions to be examined.
4442 This work was sponsored by Google.
4445 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4446 OpenSSL should now compile cleanly on gcc 4.2
4447 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4449 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4450 support including streaming MAC support: this is required for GOST
4451 ciphersuite support.
4452 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4454 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4455 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4456 to output in BER and PEM format.
4459 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4460 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4461 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4462 ENGINE support for HMAC keys which are unextractable. New -mac and
4463 -macopt options to dgst utility.
4466 *) New option -sigopt to dgst utility. Update dgst to use
4467 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4468 alternative signing parameters such as X9.31 or PSS in the dgst
4472 *) Change ssl_cipher_apply_rule(), the internal function that does
4473 the work each time a ciphersuite string requests enabling
4474 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4475 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4476 the order of disabled ciphersuites such that those ciphersuites
4477 that most recently went from enabled to disabled not only stay
4478 in order with respect to each other, but also have higher priority
4479 than other disabled ciphersuites the next time ciphersuites are
4482 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4483 the same ciphersuites as with "HIGH" alone, but in a specific
4484 order where the PSK ciphersuites come first (since they are the
4485 most recently disabled ciphersuites when "HIGH" is parsed).
4487 Also, change ssl_create_cipher_list() (using this new
4488 functionality) such that between otherwise identical
4489 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4493 *) Change ssl_create_cipher_list() so that it automatically
4494 arranges the ciphersuites in reasonable order before starting
4495 to process the rule string. Thus, the definition for "DEFAULT"
4496 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4497 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4498 This makes it much easier to arrive at a reasonable default order
4499 in applications for which anonymous ciphers are OK (meaning
4500 that you can't actually use DEFAULT).
4501 [Bodo Moeller; suggested by Victor Duchovni]
4503 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4504 processing) into multiple integers instead of setting
4505 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4506 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4507 (These masks as well as the individual bit definitions are hidden
4508 away into the non-exported interface ssl/ssl_locl.h, so this
4509 change to the definition of the SSL_CIPHER structure shouldn't
4510 affect applications.) This give us more bits for each of these
4511 categories, so there is no longer a need to coagulate AES128 and
4512 AES256 into a single algorithm bit, and to coagulate Camellia128
4513 and Camellia256 into a single algorithm bit, which has led to all
4516 Thus, among other things, the kludge introduced in 0.9.7m and
4517 0.9.8e for masking out AES256 independently of AES128 or masking
4518 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4520 With the change, we also introduce new ciphersuite aliases that
4521 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4525 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4526 Use the leftmost N bytes of the signature input if the input is
4527 larger than the prime q (with N being the size in bytes of q).
4530 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4531 it yet and it is largely untested.
4534 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4537 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4538 some compilers (gcc 4.2 and later) reject their use. Safestack is
4539 reimplemented. Update ASN1 to avoid use of legacy functions.
4542 *) Win32/64 targets are linked with Winsock2.
4545 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4546 to external functions. This can be used to increase CRL handling
4547 efficiency especially when CRLs are very large by (for example) storing
4548 the CRL revoked certificates in a database.
4551 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4552 new CRLs added to a directory can be used. New command line option
4553 -verify_return_error to s_client and s_server. This causes real errors
4554 to be returned by the verify callback instead of carrying on no matter
4555 what. This reflects the way a "real world" verify callback would behave.
4558 *) GOST engine, supporting several GOST algorithms and public key formats.
4559 Kindly donated by Cryptocom.
4562 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4563 partitioned by DP are handled but no indirect CRL or reason partitioning
4564 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4565 selected via a scoring technique which handles IDP and AKID in CRLs.
4568 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4569 will ultimately be used for all verify operations: this will remove the
4570 X509_STORE dependency on certificate verification and allow alternative
4571 lookup methods. X509_STORE based implementations of these two callbacks.
4574 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4575 Modify get_crl() to find a valid (unexpired) CRL if possible.
4578 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4579 this would be called X509_CRL_cmp() but that name is already used by
4580 a function that just compares CRL issuer names. Cache several CRL
4581 extensions in X509_CRL structure and cache CRLDP in X509.
4584 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4585 this maps equivalent X509_NAME structures into a consistent structure.
4586 Name comparison can then be performed rapidly using memcmp().
4589 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4593 *) Allow digests to supply their own micalg string for S/MIME type using
4594 the ctrl EVP_MD_CTRL_MICALG.
4597 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4598 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4599 ctrl. It can then customise the structure before and/or after signing
4603 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4604 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4605 to free up any added signature OIDs.
4608 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4609 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4610 digest and cipher tables. New options added to openssl utility:
4611 list-message-digest-algorithms and list-cipher-algorithms.
4614 *) Change the array representation of binary polynomials: the list
4615 of degrees of non-zero coefficients is now terminated with -1.
4616 Previously it was terminated with 0, which was also part of the
4617 value; thus, the array representation was not applicable to
4618 polynomials where t^0 has coefficient zero. This change makes
4619 the array representation useful in a more general context.
4622 *) Various modifications and fixes to SSL/TLS cipher string
4623 handling. For ECC, the code now distinguishes between fixed ECDH
4624 with RSA certificates on the one hand and with ECDSA certificates
4625 on the other hand, since these are separate ciphersuites. The
4626 unused code for Fortezza ciphersuites has been removed.
4628 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4629 (not "ECDHE"). For consistency with the code for DH
4630 certificates, use of ECDH certificates is now considered ECDH
4631 authentication, not RSA or ECDSA authentication (the latter is
4632 merely the CA's signing algorithm and not actively used in the
4635 The temporary ciphersuite alias "ECCdraft" is no longer
4636 available, and ECC ciphersuites are no longer excluded from "ALL"
4637 and "DEFAULT". The following aliases now exist for RFC 4492
4638 ciphersuites, most of these by analogy with the DH case:
4640 kECDHr - ECDH cert, signed with RSA
4641 kECDHe - ECDH cert, signed with ECDSA
4642 kECDH - ECDH cert (signed with either RSA or ECDSA)
4643 kEECDH - ephemeral ECDH
4644 ECDH - ECDH cert or ephemeral ECDH
4650 AECDH - anonymous ECDH
4651 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4655 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4656 Use correct micalg parameters depending on digest(s) in signed message.
4659 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4660 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4663 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4664 an engine to register a method. Add ENGINE lookups for methods and
4665 functional reference processing.
4668 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4669 EVP_{Sign,Verify}* which allow an application to customise the signature
4673 *) New -resign option to smime utility. This adds one or more signers
4674 to an existing PKCS#7 signedData structure. Also -md option to use an
4675 alternative message digest algorithm for signing.
4678 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4679 create PKCS7 structures containing multiple signers. Update smime
4680 application to support multiple signers.
4683 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4687 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4688 Reorganize PBE internals to lookup from a static table using NIDs,
4689 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4690 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4691 PRF which will be automatically used with PBES2.
4694 *) Replace the algorithm specific calls to generate keys in "req" with the
4698 *) Update PKCS#7 enveloped data routines to use new API. This is now
4699 supported by any public key method supporting the encrypt operation. A
4700 ctrl is added to allow the public key algorithm to examine or modify
4701 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4705 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4706 a default digest type to use. In most cases this will be SHA1 but some
4707 algorithms (such as GOST) need to specify an alternative digest. The
4708 return value indicates how strong the preference is 1 means optional and
4709 2 is mandatory (that is it is the only supported type). Modify
4710 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4711 use the default md. Update openssl utilities to use the default digest
4712 type for signing if it is not explicitly indicated.
4715 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4716 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4717 signing method from the key type. This effectively removes the link
4718 between digests and public key types.
4721 *) Add an OID cross reference table and utility functions. Its purpose is to
4722 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4723 rsaEncryption. This will allow some of the algorithm specific hackery
4724 needed to use the correct OID to be removed.
4727 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4728 structures for PKCS7_sign(). They are now set up by the relevant public
4732 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4735 *) Add support for key derivation (agreement) in the API, DH method and
4739 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4740 public and private key formats. As a side effect these add additional
4741 command line functionality not previously available: DSA signatures can be
4742 generated and verified using pkeyutl and DH key support and generation in
4747 [Oliver Tappe <zooey@hirschkaefer.de>]
4749 *) New make target "install_html_docs" installs HTML renditions of the
4751 [Oliver Tappe <zooey@hirschkaefer.de>]
4753 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4754 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4755 support key and parameter generation and add initial key generation
4756 functionality for RSA.
4759 *) Add functions for main EVP_PKEY_method operations. The undocumented
4760 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4761 EVP_PKEY_{encrypt,decrypt}_old.
4764 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4765 key API, doesn't do much yet.
4768 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4769 public key algorithms. New option to openssl utility:
4770 "list-public-key-algorithms" to print out info.
4773 *) Implement the Supported Elliptic Curves Extension for
4774 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4777 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4778 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4781 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4782 utilities such as rsa, dsa, dsaparam etc except they process any key
4786 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4787 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4788 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4792 *) Initial support for pluggable public key ASN1.
4793 De-spaghettify the public key ASN1 handling. Move public and private
4794 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4795 algorithm specific handling to a single module within the relevant
4796 algorithm directory. Add functions to allow (near) opaque processing
4797 of public and private key structures.
4800 *) Implement the Supported Point Formats Extension for
4801 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4804 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4805 for the psk identity [hint] and the psk callback functions to the
4806 SSL_SESSION, SSL and SSL_CTX structure.
4809 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4813 SSL_CTX_use_psk_identity_hint
4814 SSL_get_psk_identity_hint
4815 SSL_get_psk_identity
4816 SSL_use_psk_identity_hint
4818 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4820 *) Add RFC 3161 compliant time stamp request creation, response generation
4821 and response verification functionality.
4822 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4824 *) Add initial support for TLS extensions, specifically for the server_name
4825 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4826 have new members for a host name. The SSL data structure has an
4827 additional member SSL_CTX *initial_ctx so that new sessions can be
4828 stored in that context to allow for session resumption, even after the
4829 SSL has been switched to a new SSL_CTX in reaction to a client's
4830 server_name extension.
4832 New functions (subject to change):
4834 SSL_get_servername()
4835 SSL_get_servername_type()
4838 New CTRL codes and macros (subject to change):
4840 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4841 - SSL_CTX_set_tlsext_servername_callback()
4842 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4843 - SSL_CTX_set_tlsext_servername_arg()
4844 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4846 openssl s_client has a new '-servername ...' option.
4848 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4849 '-key2 ...', '-servername_fatal' (subject to change). This allows
4850 testing the HostName extension for a specific single host name ('-cert'
4851 and '-key' remain fallbacks for handshakes without HostName
4852 negotiation). If the unrecognized_name alert has to be sent, this by
4853 default is a warning; it becomes fatal with the '-servername_fatal'
4856 [Peter Sylvester, Remy Allais, Christophe Renou]
4858 *) Whirlpool hash implementation is added.
4861 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4862 bn(64,32). Because of instruction set limitations it doesn't have
4863 any negative impact on performance. This was done mostly in order
4864 to make it possible to share assembler modules, such as bn_mul_mont
4865 implementations, between 32- and 64-bit builds without hassle.
4868 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4869 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4873 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4874 dedicated Montgomery multiplication procedure, is introduced.
4875 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4876 "64-bit" performance on certain 32-bit targets.
4879 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4880 in SSL structures. New SSL ctrl to set maximum send fragment size.
4881 Save memory by setting the I/O buffer sizes dynamically instead of
4882 using the maximum available value.
4885 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4886 in addition to the text details.
4889 *) Very, very preliminary EXPERIMENTAL support for printing of general
4890 ASN1 structures. This currently produces rather ugly output and doesn't
4891 handle several customised structures at all.
4894 *) Integrated support for PVK file format and some related formats such
4895 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4896 these in the 'rsa' and 'dsa' utilities.
4899 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4902 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4903 place for the (very old) "NETSCAPE" format certificates which are now
4904 handled using new ASN1 code equivalents.
4907 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4908 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4909 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4912 *) Modify CRL distribution points extension code to print out previously
4913 unsupported fields. Enhance extension setting code to allow setting of
4917 *) Add print and set support for Issuing Distribution Point CRL extension.
4920 *) Change 'Configure' script to enable Camellia by default.
4923 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4925 *) When rejecting SSL/TLS records due to an incorrect version number, never
4926 update s->server with a new major version number. As of
4927 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4928 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4929 the previous behavior could result in a read attempt at NULL when
4930 receiving specific incorrect SSL/TLS records once record payload
4931 protection is active. (CVE-2010-0740)
4932 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4934 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4935 could be crashed if the relevant tables were not present (e.g. chrooted).
4936 [Tomas Hoger <thoger@redhat.com>]
4938 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4940 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4941 [Martin Olsson, Neel Mehta]
4943 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4944 accommodate for stack sorting, always a write lock!).
4947 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4948 excessive delays in the RAND_poll(): over a minute. As a workaround
4949 include a time check in the inner Heap32Next loop too.
4952 *) The code that handled flushing of data in SSL/TLS originally used the
4953 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4954 the problem outlined in PR#1949. The fix suggested there however can
4955 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4956 of Apache). So instead simplify the code to flush unconditionally.
4957 This should be fine since flushing with no data to flush is a no op.
4960 *) Handle TLS versions 2.0 and later properly and correctly use the
4961 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4962 off ancient servers have a habit of sticking around for a while...
4965 *) Modify compression code so it frees up structures without using the
4966 ex_data callbacks. This works around a problem where some applications
4967 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4968 restarting) then use compression (e.g. SSL with compression) later.
4969 This results in significant per-connection memory leaks and
4970 has caused some security issues including CVE-2008-1678 and
4974 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4975 change when encrypting or decrypting.
4978 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4979 connect and renegotiate with servers which do not support RI.
4980 Until RI is more widely deployed this option is enabled by default.
4983 *) Add "missing" ssl ctrls to clear options and mode.
4986 *) If client attempts to renegotiate and doesn't support RI respond with
4987 a no_renegotiation alert as required by RFC5746. Some renegotiating
4988 TLS clients will continue a connection gracefully when they receive
4989 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4990 waiting for a server hello which it will never receive. Now we treat a
4991 received no_renegotiation alert as a fatal error. This is because
4992 applications requesting a renegotiation might well expect it to succeed
4993 and would have no code in place to handle the server denying it so the
4994 only safe thing to do is to terminate the connection.
4997 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4998 peer supports secure renegotiation and 0 otherwise. Print out peer
4999 renegotiation support in s_client/s_server.
5002 *) Replace the highly broken and deprecated SPKAC certification method with
5003 the updated NID creation version. This should correctly handle UTF8.
5006 *) Implement RFC5746. Re-enable renegotiation but require the extension
5007 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
5008 turns out to be a bad idea. It has been replaced by
5009 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5010 SSL_CTX_set_options(). This is really not recommended unless you
5011 know what you are doing.
5012 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5014 *) Fixes to stateless session resumption handling. Use initial_ctx when
5015 issuing and attempting to decrypt tickets in case it has changed during
5016 servername handling. Use a non-zero length session ID when attempting
5017 stateless session resumption: this makes it possible to determine if
5018 a resumption has occurred immediately after receiving server hello
5019 (several places in OpenSSL subtly assume this) instead of later in
5023 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5024 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5025 fixes for a few places where the return code is not checked
5027 [Julia Lawall <julia@diku.dk>]
5029 *) Add --strict-warnings option to Configure script to include devteam
5030 warnings in other configurations.
5033 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5034 makes it possible to install openssl libraries in locations which
5035 have names other than "lib", for example "/usr/lib64" which some
5037 [Steve Henson, based on patch from Jeremy Utley]
5039 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5040 X690 8.9.12 and can produce some misleading textual output of OIDs.
5041 [Steve Henson, reported by Dan Kaminsky]
5043 *) Delete MD2 from algorithm tables. This follows the recommendation in
5044 several standards that it is not used in new applications due to
5045 several cryptographic weaknesses. For binary compatibility reasons
5046 the MD2 API is still compiled in by default.
5049 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5053 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5054 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5056 [Guenter <lists@gknw.net>]
5058 *) Fix the server certificate chain building code to use X509_verify_cert(),
5059 it used to have an ad-hoc builder which was unable to cope with anything
5060 other than a simple chain.
5061 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5063 *) Don't check self signed certificate signatures in X509_verify_cert()
5064 by default (a flag can override this): it just wastes time without
5065 adding any security. As a useful side effect self signed root CAs
5066 with non-FIPS digests are now usable in FIPS mode.
5069 *) In dtls1_process_out_of_seq_message() the check if the current message
5070 is already buffered was missing. For every new message was memory
5071 allocated, allowing an attacker to perform an denial of service attack
5072 with sending out of seq handshake messages until there is no memory
5073 left. Additionally every future message was buffered, even if the
5074 sequence number made no sense and would be part of another handshake.
5075 So only messages with sequence numbers less than 10 in advance will be
5076 buffered. (CVE-2009-1378)
5077 [Robin Seggelmann, discovered by Daniel Mentz]
5079 *) Records are buffered if they arrive with a future epoch to be
5080 processed after finishing the corresponding handshake. There is
5081 currently no limitation to this buffer allowing an attacker to perform
5082 a DOS attack with sending records with future epochs until there is no
5083 memory left. This patch adds the pqueue_size() function to determine
5084 the size of a buffer and limits the record buffer to 100 entries.
5086 [Robin Seggelmann, discovered by Daniel Mentz]
5088 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5089 parent structure is freed. (CVE-2009-1379)
5092 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5093 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5096 [Ilya O. <vrghost@gmail.com>]
5098 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5100 *) Disable renegotiation completely - this fixes a severe security
5101 problem (CVE-2009-3555) at the cost of breaking all
5102 renegotiation. Renegotiation can be re-enabled by setting
5103 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5104 run-time. This is really not recommended unless you know what
5108 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5110 *) Don't set val to NULL when freeing up structures, it is freed up by
5111 underlying code. If sizeof(void *) > sizeof(long) this can result in
5112 zeroing past the valid field. (CVE-2009-0789)
5113 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5115 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5116 checked correctly. This would allow some invalid signed attributes to
5117 appear to verify correctly. (CVE-2009-0591)
5118 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5120 *) Reject UniversalString and BMPString types with invalid lengths. This
5121 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5122 a legal length. (CVE-2009-0590)
5125 *) Set S/MIME signing as the default purpose rather than setting it
5126 unconditionally. This allows applications to override it at the store
5130 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5131 to handle some structures.
5134 *) Improve efficiency of mem_gets: don't search whole buffer each time
5136 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5138 *) New -hex option for openssl rand.
5141 *) Print out UTF8String and NumericString when parsing ASN1.
5144 *) Support NumericString type for name components.
5147 *) Allow CC in the environment to override the automatically chosen
5148 compiler. Note that nothing is done to ensure flags work with the
5152 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5154 *) Properly check EVP_VerifyFinal() and similar return values
5156 [Ben Laurie, Bodo Moeller, Google Security Team]
5158 *) Enable TLS extensions by default.
5161 *) Allow the CHIL engine to be loaded, whether the application is
5162 multithreaded or not. (This does not release the developer from the
5163 obligation to set up the dynamic locking callbacks.)
5164 [Sander Temme <sander@temme.net>]
5166 *) Use correct exit code if there is an error in dgst command.
5167 [Steve Henson; problem pointed out by Roland Dirlewanger]
5169 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5170 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5173 *) Add experimental JPAKE support, including demo authentication in
5174 s_client and s_server.
5177 *) Set the comparison function in v3_addr_canonize().
5178 [Rob Austein <sra@hactrn.net>]
5180 *) Add support for XMPP STARTTLS in s_client.
5181 [Philip Paeps <philip@freebsd.org>]
5183 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5184 to ensure that even with this option, only ciphersuites in the
5185 server's preference list will be accepted. (Note that the option
5186 applies only when resuming a session, so the earlier behavior was
5187 just about the algorithm choice for symmetric cryptography.)
5190 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5192 *) Fix NULL pointer dereference if a DTLS server received
5193 ChangeCipherSpec as first record (CVE-2009-1386).
5196 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5197 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5200 *) The fix in 0.9.8c that supposedly got rid of unsafe
5201 double-checked locking was incomplete for RSA blinding,
5202 addressing just one layer of what turns out to have been
5203 doubly unsafe triple-checked locking.
5205 So now fix this for real by retiring the MONT_HELPER macro
5206 in crypto/rsa/rsa_eay.c.
5208 [Bodo Moeller; problem pointed out by Marius Schilder]
5210 *) Various precautionary measures:
5212 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5214 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5215 (NB: This would require knowledge of the secret session ticket key
5216 to exploit, in which case you'd be SOL either way.)
5218 - Change bn_nist.c so that it will properly handle input BIGNUMs
5219 outside the expected range.
5221 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5224 [Neel Mehta, Bodo Moeller]
5226 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5227 the load fails. Useful for distros.
5228 [Ben Laurie and the FreeBSD team]
5230 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5233 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5236 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5238 This work was sponsored by Logica.
5241 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5242 keystores. Support for SSL/TLS client authentication too.
5243 Not compiled unless enable-capieng specified to Configure.
5245 This work was sponsored by Logica.
5248 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5249 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5250 attribute creation routines such as certificate requests and PKCS#12
5254 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5256 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5257 handshake which could lead to a client crash as found using the
5258 Codenomicon TLS test suite (CVE-2008-1672)
5259 [Steve Henson, Mark Cox]
5261 *) Fix double free in TLS server name extensions which could lead to
5262 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5265 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5267 Clear the error queue to ensure that error entries left from
5268 older function calls do not interfere with the correct operation.
5269 [Lutz Jaenicke, Erik de Castro Lopo]
5271 *) Remove root CA certificates of commercial CAs:
5273 The OpenSSL project does not recommend any specific CA and does not
5274 have any policy with respect to including or excluding any CA.
5275 Therefore it does not make any sense to ship an arbitrary selection
5276 of root CA certificates with the OpenSSL software.
5279 *) RSA OAEP patches to fix two separate invalid memory reads.
5280 The first one involves inputs when 'lzero' is greater than
5281 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5282 before the beginning of from). The second one involves inputs where
5283 the 'db' section contains nothing but zeroes (there is a one-byte
5284 invalid read after the end of 'db').
5285 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5287 *) Partial backport from 0.9.9-dev:
5289 Introduce bn_mul_mont (dedicated Montgomery multiplication
5290 procedure) as a candidate for BIGNUM assembler implementation.
5291 While 0.9.9-dev uses assembler for various architectures, only
5292 x86_64 is available by default here in the 0.9.8 branch, and
5293 32-bit x86 is available through a compile-time setting.
5295 To try the 32-bit x86 assembler implementation, use Configure
5296 option "enable-montasm" (which exists only for this backport).
5298 As "enable-montasm" for 32-bit x86 disclaims code stability
5299 anyway, in this constellation we activate additional code
5300 backported from 0.9.9-dev for further performance improvements,
5301 namely BN_from_montgomery_word. (To enable this otherwise,
5302 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5304 [Andy Polyakov (backport partially by Bodo Moeller)]
5306 *) Add TLS session ticket callback. This allows an application to set
5307 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5308 values. This is useful for key rollover for example where several key
5309 sets may exist with different names.
5312 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5313 This was broken until now in 0.9.8 releases, such that the only way
5314 a registered ENGINE could be used (assuming it initialises
5315 successfully on the host) was to explicitly set it as the default
5316 for the relevant algorithms. This is in contradiction with 0.9.7
5317 behaviour and the documentation. With this fix, when an ENGINE is
5318 registered into a given algorithm's table of implementations, the
5319 'uptodate' flag is reset so that auto-discovery will be used next
5320 time a new context for that algorithm attempts to select an
5322 [Ian Lister (tweaked by Geoff Thorpe)]
5324 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5325 implementation in the following ways:
5327 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5330 Lack of BER streaming support means one pass streaming processing is
5331 only supported if data is detached: setting the streaming flag is
5332 ignored for embedded content.
5334 CMS support is disabled by default and must be explicitly enabled
5335 with the enable-cms configuration option.
5338 *) Update the GMP engine glue to do direct copies between BIGNUM and
5339 mpz_t when openssl and GMP use the same limb size. Otherwise the
5340 existing "conversion via a text string export" trick is still used.
5341 [Paul Sheer <paulsheer@gmail.com>]
5343 *) Zlib compression BIO. This is a filter BIO which compressed and
5344 uncompresses any data passed through it.
5347 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5348 RFC3394 compatible AES key wrapping.
5351 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5352 sets string data without copying. X509_ALGOR_set0() and
5353 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5354 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5355 from an X509_ATTRIBUTE structure optionally checking it occurs only
5356 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5360 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5361 to get the expected BN_FLG_CONSTTIME behavior.
5362 [Bodo Moeller (Google)]
5366 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5367 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5368 - added some more tests to do_tests.pl
5369 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5370 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5371 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5372 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5373 - various changes to netware.pl to enable gcc-cross builds on Win32
5375 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5376 - various changes to fix missing prototype warnings
5377 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5378 - added AES, WHIRLPOOL and CPUID assembler code to build files
5379 - added missing AES assembler make rules to mk1mf.pl
5380 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5381 [Guenter Knauf <eflash@gmx.net>]
5383 *) Implement certificate status request TLS extension defined in RFC3546.
5384 A client can set the appropriate parameters and receive the encoded
5385 OCSP response via a callback. A server can query the supplied parameters
5386 and set the encoded OCSP response in the callback. Add simplified examples
5387 to s_client and s_server.
5390 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5392 *) Fix various bugs:
5393 + Binary incompatibility of ssl_ctx_st structure
5394 + DTLS interoperation with non-compliant servers
5395 + Don't call get_session_cb() without proposed session
5396 + Fix ia64 assembler code
5397 [Andy Polyakov, Steve Henson]
5399 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5401 *) DTLS Handshake overhaul. There were longstanding issues with
5402 OpenSSL DTLS implementation, which were making it impossible for
5403 RFC 4347 compliant client to communicate with OpenSSL server.
5404 Unfortunately just fixing these incompatibilities would "cut off"
5405 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5406 server keeps tolerating non RFC compliant syntax. The opposite is
5407 not true, 0.9.8f client can not communicate with earlier server.
5408 This update even addresses CVE-2007-4995.
5411 *) Changes to avoid need for function casts in OpenSSL: some compilers
5412 (gcc 4.2 and later) reject their use.
5413 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5416 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5417 RFC4507bis. The encrypted ticket format is an encrypted encoded
5418 SSL_SESSION structure, that way new session features are automatically
5421 If a client application caches session in an SSL_SESSION structure
5422 support is transparent because tickets are now stored in the encoded
5425 The SSL_CTX structure automatically generates keys for ticket
5426 protection in servers so again support should be possible
5427 with no application modification.
5429 If a client or server wishes to disable RFC4507 support then the option
5430 SSL_OP_NO_TICKET can be set.
5432 Add a TLS extension debugging callback to allow the contents of any client
5433 or server extensions to be examined.
5435 This work was sponsored by Google.
5438 *) Add initial support for TLS extensions, specifically for the server_name
5439 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5440 have new members for a host name. The SSL data structure has an
5441 additional member SSL_CTX *initial_ctx so that new sessions can be
5442 stored in that context to allow for session resumption, even after the
5443 SSL has been switched to a new SSL_CTX in reaction to a client's
5444 server_name extension.
5446 New functions (subject to change):
5448 SSL_get_servername()
5449 SSL_get_servername_type()
5452 New CTRL codes and macros (subject to change):
5454 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5455 - SSL_CTX_set_tlsext_servername_callback()
5456 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5457 - SSL_CTX_set_tlsext_servername_arg()
5458 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5460 openssl s_client has a new '-servername ...' option.
5462 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5463 '-key2 ...', '-servername_fatal' (subject to change). This allows
5464 testing the HostName extension for a specific single host name ('-cert'
5465 and '-key' remain fallbacks for handshakes without HostName
5466 negotiation). If the unrecognized_name alert has to be sent, this by
5467 default is a warning; it becomes fatal with the '-servername_fatal'
5470 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5472 *) Add AES and SSE2 assembly language support to VC++ build.
5475 *) Mitigate attack on final subtraction in Montgomery reduction.
5478 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5479 (which previously caused an internal error).
5482 *) Squeeze another 10% out of IGE mode when in != out.
5485 *) AES IGE mode speedup.
5486 [Dean Gaudet (Google)]
5488 *) Add the Korean symmetric 128-bit cipher SEED (see
5489 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5490 add SEED ciphersuites from RFC 4162:
5492 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5493 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5494 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5495 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5497 To minimize changes between patchlevels in the OpenSSL 0.9.8
5498 series, SEED remains excluded from compilation unless OpenSSL
5499 is configured with 'enable-seed'.
5500 [KISA, Bodo Moeller]
5502 *) Mitigate branch prediction attacks, which can be practical if a
5503 single processor is shared, allowing a spy process to extract
5504 information. For detailed background information, see
5505 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5506 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5507 and Necessary Software Countermeasures"). The core of the change
5508 are new versions BN_div_no_branch() and
5509 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5510 respectively, which are slower, but avoid the security-relevant
5511 conditional branches. These are automatically called by BN_div()
5512 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5513 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5514 remove a conditional branch.
5516 BN_FLG_CONSTTIME is the new name for the previous
5517 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5518 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5519 in the exponent causes BN_mod_exp_mont() to use the alternative
5520 implementation in BN_mod_exp_mont_consttime().) The old name
5521 remains as a deprecated alias.
5523 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5524 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5525 constant-time implementations for more than just exponentiation.
5526 Here too the old name is kept as a deprecated alias.
5528 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5529 the BN_BLINDING structure gets an independent copy of the
5530 modulus. This means that the previous "BIGNUM *m" argument to
5531 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5532 essentially becomes "const BIGNUM *m", although we can't actually
5533 change this in the header file before 0.9.9. It allows
5534 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5535 enable BN_FLG_CONSTTIME.
5537 [Matthew D Wood (Intel Corp)]
5539 *) In the SSL/TLS server implementation, be strict about session ID
5540 context matching (which matters if an application uses a single
5541 external cache for different purposes). Previously,
5542 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5543 set. This did ensure strict client verification, but meant that,
5544 with applications using a single external cache for quite
5545 different requirements, clients could circumvent ciphersuite
5546 restrictions for a given session ID context by starting a session
5547 in a different context.
5550 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5551 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5552 authentication-only ciphersuites.
5555 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5556 not complete and could lead to a possible single byte overflow
5557 (CVE-2007-5135) [Ben Laurie]
5559 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5561 *) Since AES128 and AES256 (and similarly Camellia128 and
5562 Camellia256) share a single mask bit in the logic of
5563 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5564 kludge to work properly if AES128 is available and AES256 isn't
5565 (or if Camellia128 is available and Camellia256 isn't).
5568 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5569 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5570 When a point or a seed is encoded in a BIT STRING, we need to
5571 prevent the removal of trailing zero bits to get the proper DER
5572 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5573 of a NamedBitList, for which trailing 0 bits need to be removed.)
5576 *) Have SSL/TLS server implementation tolerate "mismatched" record
5577 protocol version while receiving ClientHello even if the
5578 ClientHello is fragmented. (The server can't insist on the
5579 particular protocol version it has chosen before the ServerHello
5580 message has informed the client about his choice.)
5583 *) Add RFC 3779 support.
5584 [Rob Austein for ARIN, Ben Laurie]
5586 *) Load error codes if they are not already present instead of using a
5587 static variable. This allows them to be cleanly unloaded and reloaded.
5588 Improve header file function name parsing.
5591 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5592 or CAPABILITY handshake as required by RFCs.
5595 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5597 *) Introduce limits to prevent malicious keys being able to
5598 cause a denial of service. (CVE-2006-2940)
5599 [Steve Henson, Bodo Moeller]
5601 *) Fix ASN.1 parsing of certain invalid structures that can result
5602 in a denial of service. (CVE-2006-2937) [Steve Henson]
5604 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5605 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5607 *) Fix SSL client code which could crash if connecting to a
5608 malicious SSLv2 server. (CVE-2006-4343)
5609 [Tavis Ormandy and Will Drewry, Google Security Team]
5611 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5612 match only those. Before that, "AES256-SHA" would be interpreted
5613 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5614 the same strength classification in 0.9.7h) as we currently only
5615 have a single AES bit in the ciphersuite description bitmap.
5616 That change, however, also applied to ciphersuite strings such as
5617 "RC4-MD5" that intentionally matched multiple ciphersuites --
5618 namely, SSL 2.0 ciphersuites in addition to the more common ones
5619 from SSL 3.0/TLS 1.0.
5621 So we change the selection algorithm again: Naming an explicit
5622 ciphersuite selects this one ciphersuite, and any other similar
5623 ciphersuite (same bitmap) from *other* protocol versions.
5624 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5625 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5627 Since SSL 2.0 does not have any ciphersuites for which the
5628 128/256 bit distinction would be relevant, this works for now.
5629 The proper fix will be to use different bits for AES128 and
5630 AES256, which would have avoided the problems from the beginning;
5631 however, bits are scarce, so we can only do this in a new release
5632 (not just a patchlevel) when we can change the SSL_CIPHER
5633 definition to split the single 'unsigned long mask' bitmap into
5634 multiple values to extend the available space.
5638 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5640 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5641 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5643 *) Add AES IGE and biIGE modes.
5646 *) Change the Unix randomness entropy gathering to use poll() when
5647 possible instead of select(), since the latter has some
5648 undesirable limitations.
5649 [Darryl Miles via Richard Levitte and Bodo Moeller]
5651 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5652 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5653 cannot be implicitly activated as part of, e.g., the "AES" alias.
5654 However, please upgrade to OpenSSL 0.9.9[-dev] for
5655 non-experimental use of the ECC ciphersuites to get TLS extension
5656 support, which is required for curve and point format negotiation
5657 to avoid potential handshake problems.
5660 *) Disable rogue ciphersuites:
5662 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5663 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5664 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5666 The latter two were purportedly from
5667 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5670 Also deactivate the remaining ciphersuites from
5671 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5672 unofficial, and the ID has long expired.
5675 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5676 dual-core machines) and other potential thread-safety issues.
5679 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5680 versions), which is now available for royalty-free use
5681 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5682 Also, add Camellia TLS ciphersuites from RFC 4132.
5684 To minimize changes between patchlevels in the OpenSSL 0.9.8
5685 series, Camellia remains excluded from compilation unless OpenSSL
5686 is configured with 'enable-camellia'.
5689 *) Disable the padding bug check when compression is in use. The padding
5690 bug check assumes the first packet is of even length, this is not
5691 necessarily true if compression is enabled and can result in false
5692 positives causing handshake failure. The actual bug test is ancient
5693 code so it is hoped that implementations will either have fixed it by
5694 now or any which still have the bug do not support compression.
5697 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5699 *) When applying a cipher rule check to see if string match is an explicit
5700 cipher suite and only match that one cipher suite if it is.
5703 *) Link in manifests for VC++ if needed.
5704 [Austin Ziegler <halostatue@gmail.com>]
5706 *) Update support for ECC-based TLS ciphersuites according to
5707 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5708 TLS extensions, which are supported starting with the 0.9.9
5709 branch, not in the OpenSSL 0.9.8 branch).
5712 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5713 opaque EVP_CIPHER_CTX handling.
5716 *) Fixes and enhancements to zlib compression code. We now only use
5717 "zlib1.dll" and use the default __cdecl calling convention on Win32
5718 to conform with the standards mentioned here:
5719 http://www.zlib.net/DLL_FAQ.txt
5720 Static zlib linking now works on Windows and the new --with-zlib-include
5721 --with-zlib-lib options to Configure can be used to supply the location
5722 of the headers and library. Gracefully handle case where zlib library
5726 *) Several fixes and enhancements to the OID generation code. The old code
5727 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5728 handle numbers larger than ULONG_MAX, truncated printing and had a
5729 non standard OBJ_obj2txt() behaviour.
5732 *) Add support for building of engines under engine/ as shared libraries
5733 under VC++ build system.
5736 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5737 Hopefully, we will not see any false combination of paths any more.
5740 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5742 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5743 (part of SSL_OP_ALL). This option used to disable the
5744 countermeasure against man-in-the-middle protocol-version
5745 rollback in the SSL 2.0 server implementation, which is a bad
5746 idea. (CVE-2005-2969)
5748 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5749 for Information Security, National Institute of Advanced Industrial
5750 Science and Technology [AIST], Japan)]
5752 *) Add two function to clear and return the verify parameter flags.
5755 *) Keep cipherlists sorted in the source instead of sorting them at
5756 runtime, thus removing the need for a lock.
5759 *) Avoid some small subgroup attacks in Diffie-Hellman.
5760 [Nick Mathewson and Ben Laurie]
5762 *) Add functions for well-known primes.
5765 *) Extended Windows CE support.
5766 [Satoshi Nakamura and Andy Polyakov]
5768 *) Initialize SSL_METHOD structures at compile time instead of during
5769 runtime, thus removing the need for a lock.
5772 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5773 attempting to decrypt each encrypted key in turn. Add support to
5777 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5779 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5782 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5785 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5786 key into the same file any more.
5789 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5792 *) Add -utf8 command line and config file option to 'ca'.
5793 [Stefan <stf@udoma.org]
5795 *) Removed the macro des_crypt(), as it seems to conflict with some
5796 libraries. Use DES_crypt().
5799 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5800 involves renaming the source and generated shared-libs for
5801 both. The engines will accept the corrected or legacy ids
5802 ('ncipher' and '4758_cca' respectively) when binding. NB,
5803 this only applies when building 'shared'.
5804 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5806 *) Add attribute functions to EVP_PKEY structure. Modify
5807 PKCS12_create() to recognize a CSP name attribute and
5808 use it. Make -CSP option work again in pkcs12 utility.
5811 *) Add new functionality to the bn blinding code:
5812 - automatic re-creation of the BN_BLINDING parameters after
5813 a fixed number of uses (currently 32)
5814 - add new function for parameter creation
5815 - introduce flags to control the update behaviour of the
5816 BN_BLINDING parameters
5817 - hide BN_BLINDING structure
5818 Add a second BN_BLINDING slot to the RSA structure to improve
5819 performance when a single RSA object is shared among several
5823 *) Add support for DTLS.
5824 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5826 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5827 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5830 *) Remove buggy and incomplete DH cert support from
5831 ssl/ssl_rsa.c and ssl/s3_both.c
5834 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5835 the apps/openssl applications.
5838 *) Compile clean with "-Wall -Wmissing-prototypes
5839 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5840 DEBUG_SAFESTACK must also be set.
5843 *) Change ./Configure so that certain algorithms can be disabled by default.
5844 The new counterpiece to "no-xxx" is "enable-xxx".
5846 The patented RC5 and MDC2 algorithms will now be disabled unless
5847 "enable-rc5" and "enable-mdc2", respectively, are specified.
5849 (IDEA remains enabled despite being patented. This is because IDEA
5850 is frequently required for interoperability, and there is no license
5851 fee for non-commercial use. As before, "no-idea" can be used to
5852 avoid this algorithm.)
5856 *) Add processing of proxy certificates (see RFC 3820). This work was
5857 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5858 EGEE (Enabling Grids for E-science in Europe).
5861 *) RC4 performance overhaul on modern architectures/implementations, such
5862 as Intel P4, IA-64 and AMD64.
5865 *) New utility extract-section.pl. This can be used specify an alternative
5866 section number in a pod file instead of having to treat each file as
5867 a separate case in Makefile. This can be done by adding two lines to the
5870 =for comment openssl_section:XXX
5872 The blank line is mandatory.
5876 *) New arguments -certform, -keyform and -pass for s_client and s_server
5877 to allow alternative format key and certificate files and passphrase
5881 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5882 update associated structures and add various utility functions.
5884 Add new policy related verify parameters, include policy checking in
5885 standard verify code. Enhance 'smime' application with extra parameters
5886 to support policy checking and print out.
5889 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5890 Nehemiah processors. These extensions support AES encryption in hardware
5891 as well as RNG (though RNG support is currently disabled).
5892 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5894 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5897 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5898 [Andy Polyakov and a number of other people]
5900 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5901 implementation contributed by IBM.
5902 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5904 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5905 exponent rather than 'unsigned long'. There is a corresponding change to
5906 the new 'rsa_keygen' element of the RSA_METHOD structure.
5907 [Jelte Jansen, Geoff Thorpe]
5909 *) Functionality for creating the initial serial number file is now
5910 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5912 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5913 number file to 1, which is bound to cause problems. To avoid
5914 the problems while respecting compatibility between different 0.9.7
5915 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5916 CA.pl for serial number initialization. With the new release 0.9.8,
5917 we can fix the problem directly in the 'ca' utility.)
5920 *) Reduced header interdependencies by declaring more opaque objects in
5921 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5922 give fewer recursive includes, which could break lazy source code - so
5923 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5924 developers should define this symbol when building and using openssl to
5925 ensure they track the recommended behaviour, interfaces, [etc], but
5926 backwards-compatible behaviour prevails when this isn't defined.
5929 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5932 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5933 This will generate a random key of the appropriate length based on the
5934 cipher context. The EVP_CIPHER can provide its own random key generation
5935 routine to support keys of a specific form. This is used in the des and
5936 3des routines to generate a key of the correct parity. Update S/MIME
5937 code to use new functions and hence generate correct parity DES keys.
5938 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5939 valid (weak or incorrect parity).
5942 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5943 as looking them up. This is useful when the verified structure may contain
5944 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5945 present unless the new PKCS7_NO_CRL flag is asserted.
5948 *) Extend ASN1 oid configuration module. It now additionally accepts the
5951 shortName = some long name, 1.2.3.4
5954 *) Reimplemented the BN_CTX implementation. There is now no more static
5955 limitation on the number of variables it can handle nor the depth of the
5956 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5957 information can now expand as required, and rather than having a single
5958 static array of bignums, BN_CTX now uses a linked-list of such arrays
5959 allowing it to expand on demand whilst maintaining the usefulness of
5960 BN_CTX's "bundling".
5963 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5964 to allow all RSA operations to function using a single BN_CTX.
5967 *) Preliminary support for certificate policy evaluation and checking. This
5968 is initially intended to pass the tests outlined in "Conformance Testing
5969 of Relying Party Client Certificate Path Processing Logic" v1.07.
5972 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5973 remained unused and not that useful. A variety of other little bignum
5974 tweaks and fixes have also been made continuing on from the audit (see
5978 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5979 associated ASN1, EVP and SSL functions and old ASN1 macros.
5982 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5983 and this should never fail. So the return value from the use of
5984 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5985 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5988 *) BN_CTX_get() should return zero-valued bignums, providing the same
5989 initialised value as BN_new().
5990 [Geoff Thorpe, suggested by Ulf Möller]
5992 *) Support for inhibitAnyPolicy certificate extension.
5995 *) An audit of the BIGNUM code is underway, for which debugging code is
5996 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5997 is considered valid when processing BIGNUMs, and causes execution to
5998 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5999 further steps are taken to deliberately pollute unused data in BIGNUM
6000 structures to try and expose faulty code further on. For now, openssl will
6001 (in its default mode of operation) continue to tolerate the inconsistent
6002 forms that it has tolerated in the past, but authors and packagers should
6003 consider trying openssl and their own applications when compiled with
6004 these debugging symbols defined. It will help highlight potential bugs in
6005 their own code, and will improve the test coverage for OpenSSL itself. At
6006 some point, these tighter rules will become openssl's default to improve
6007 maintainability, though the assert()s and other overheads will remain only
6008 in debugging configurations. See bn.h for more details.
6009 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6011 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6012 that can only be obtained through BN_CTX_new() (which implicitly
6013 initialises it). The presence of this function only made it possible
6014 to overwrite an existing structure (and cause memory leaks).
6017 *) Because of the callback-based approach for implementing LHASH as a
6018 template type, lh_insert() adds opaque objects to hash-tables and
6019 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6020 to clean up those corresponding objects before destroying the hash table
6021 (and losing the object pointers). So some over-zealous constifications in
6022 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6023 objects as "const" and the lh_doall[_arg] callback wrappers are not
6024 prototyped to have "const" restrictions on the object pointers they are
6025 given (and so aren't required to cast them away any more).
6028 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6029 (speed) prefers to use its own implementation. The two implementations
6030 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6031 its object type properly exposed (MS_TM) instead of casting to/from "char
6032 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6033 aren't necessarily the greatest nomenclatures - but this is what was used
6034 internally to the implementation so I've used that for now.
6037 *) Ensure that deprecated functions do not get compiled when
6038 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6039 the self-tests were still using deprecated key-generation functions so
6040 these have been updated also.
6043 *) Reorganise PKCS#7 code to separate the digest location functionality
6044 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6045 New function PKCS7_set_digest() to set the digest type for PKCS#7
6046 digestedData type. Add additional code to correctly generate the
6047 digestedData type and add support for this type in PKCS7 initialization
6051 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6052 structure of type "other".
6055 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6056 sure the loop does correctly stop and breaking ("division by zero")
6057 modulus operations are not performed. The (pre-generated) prime
6058 table crypto/bn/bn_prime.h was already correct, but it could not be
6059 re-generated on some platforms because of the "division by zero"
6060 situation in the script.
6061 [Ralf S. Engelschall]
6063 *) Update support for ECC-based TLS ciphersuites according to
6064 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6065 SHA-1 now is only used for "small" curves (where the
6066 representation of a field element takes up to 24 bytes); for
6067 larger curves, the field element resulting from ECDH is directly
6068 used as premaster secret.
6069 [Douglas Stebila (Sun Microsystems Laboratories)]
6071 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6072 curve secp160r1 to the tests.
6073 [Douglas Stebila (Sun Microsystems Laboratories)]
6075 *) Add the possibility to load symbols globally with DSO.
6076 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6078 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6079 control of the error stack.
6082 *) Add support for STORE in ENGINE.
6085 *) Add the STORE type. The intention is to provide a common interface
6086 to certificate and key stores, be they simple file-based stores, or
6087 HSM-type store, or LDAP stores, or...
6088 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6091 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6092 pass a list of arguments to any function as well as provide a way
6093 for a function to pass data back to the caller.
6096 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6097 works like BUF_strdup() but can be used to duplicate a portion of
6098 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6102 *) Add the function sk_find_ex() which works like sk_find(), but will
6103 return an index to an element even if an exact match couldn't be
6104 found. The index is guaranteed to point at the element where the
6105 searched-for key would be inserted to preserve sorting order.
6108 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6109 takes an extra flags argument for optional functionality. Currently,
6110 the following flags are defined:
6112 OBJ_BSEARCH_VALUE_ON_NOMATCH
6113 This one gets OBJ_bsearch_ex() to return a pointer to the first
6114 element where the comparing function returns a negative or zero
6117 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6118 This one gets OBJ_bsearch_ex() to return a pointer to the first
6119 element where the comparing function returns zero. This is useful
6120 if there are more than one element where the comparing function
6124 *) Make it possible to create self-signed certificates with 'openssl ca'
6125 in such a way that the self-signed certificate becomes part of the
6126 CA database and uses the same mechanisms for serial number generation
6127 as all other certificate signing. The new flag '-selfsign' enables
6128 this functionality. Adapt CA.sh and CA.pl.in.
6131 *) Add functionality to check the public key of a certificate request
6132 against a given private. This is useful to check that a certificate
6133 request can be signed by that key (self-signing).
6136 *) Make it possible to have multiple active certificates with the same
6137 subject in the CA index file. This is done only if the keyword
6138 'unique_subject' is set to 'no' in the main CA section (default
6139 if 'CA_default') of the configuration file. The value is saved
6140 with the database itself in a separate index attribute file,
6141 named like the index file with '.attr' appended to the name.
6144 *) Generate multi-valued AVAs using '+' notation in config files for
6148 *) Support for nameConstraints certificate extension.
6151 *) Support for policyConstraints certificate extension.
6154 *) Support for policyMappings certificate extension.
6157 *) Make sure the default DSA_METHOD implementation only uses its
6158 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6159 and change its own handlers to be NULL so as to remove unnecessary
6160 indirection. This lets alternative implementations fallback to the
6161 default implementation more easily.
6164 *) Support for directoryName in GeneralName related extensions
6168 *) Make it possible to link applications using Makefile.shared.
6169 Make that possible even when linking against static libraries!
6172 *) Support for single pass processing for S/MIME signing. This now
6173 means that S/MIME signing can be done from a pipe, in addition
6174 cleartext signing (multipart/signed type) is effectively streaming
6175 and the signed data does not need to be all held in memory.
6177 This is done with a new flag PKCS7_STREAM. When this flag is set
6178 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6179 is done after the data is output (and digests calculated) in
6180 SMIME_write_PKCS7().
6183 *) Add full support for -rpath/-R, both in shared libraries and
6184 applications, at least on the platforms where it's known how
6188 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6189 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6190 will now compute a table of multiples of the generator that
6191 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6192 faster (notably in the case of a single point multiplication,
6193 scalar * generator).
6194 [Nils Larsch, Bodo Moeller]
6196 *) IPv6 support for certificate extensions. The various extensions
6197 which use the IP:a.b.c.d can now take IPv6 addresses using the
6198 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6202 *) Added an ENGINE that implements RSA by performing private key
6203 exponentiations with the GMP library. The conversions to and from
6204 GMP's mpz_t format aren't optimised nor are any montgomery forms
6205 cached, and on x86 it appears OpenSSL's own performance has caught up.
6206 However there are likely to be other architectures where GMP could
6207 provide a boost. This ENGINE is not built in by default, but it can be
6208 specified at Configure time and should be accompanied by the necessary
6209 linker additions, eg;
6210 ./config -DOPENSSL_USE_GMP -lgmp
6213 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6214 testing availability of engines with "-t" - the old behaviour is
6215 produced by increasing the feature's verbosity with "-tt".
6218 *) ECDSA routines: under certain error conditions uninitialized BN objects
6219 could be freed. Solution: make sure initialization is performed early
6220 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6224 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6225 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6226 software implementations. For DSA and DH, parameter generation can
6227 also be overridden by providing the appropriate method callbacks.
6230 *) Change the "progress" mechanism used in key-generation and
6231 primality testing to functions that take a new BN_GENCB pointer in
6232 place of callback/argument pairs. The new API functions have "_ex"
6233 postfixes and the older functions are reimplemented as wrappers for
6234 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6235 declarations of the old functions to help (graceful) attempts to
6236 migrate to the new functions. Also, the new key-generation API
6237 functions operate on a caller-supplied key-structure and return
6238 success/failure rather than returning a key or NULL - this is to
6239 help make "keygen" another member function of RSA_METHOD etc.
6241 Example for using the new callback interface:
6243 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6247 BN_GENCB_set(&my_cb, my_callback, my_arg);
6249 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6250 /* For the meaning of a, b in calls to my_callback(), see the
6251 * documentation of the function that calls the callback.
6252 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6253 * my_callback should return 1 if it wants BN_is_prime_ex()
6254 * to continue, or 0 to stop.
6259 *) Change the ZLIB compression method to be stateful, and make it
6260 available to TLS with the number defined in
6261 draft-ietf-tls-compression-04.txt.
6264 *) Add the ASN.1 structures and functions for CertificatePair, which
6265 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6267 CertificatePair ::= SEQUENCE {
6268 forward [0] Certificate OPTIONAL,
6269 reverse [1] Certificate OPTIONAL,
6270 -- at least one of the pair shall be present -- }
6272 Also implement the PEM functions to read and write certificate
6273 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6275 This needed to be defined, mostly for the sake of the LDAP
6276 attribute crossCertificatePair, but may prove useful elsewhere as
6280 *) Make it possible to inhibit symlinking of shared libraries in
6281 Makefile.shared, for Cygwin's sake.
6284 *) Extend the BIGNUM API by creating a function
6285 void BN_set_negative(BIGNUM *a, int neg);
6286 and a macro that behave like
6287 int BN_is_negative(const BIGNUM *a);
6289 to avoid the need to access 'a->neg' directly in applications.
6292 *) Implement fast modular reduction for pseudo-Mersenne primes
6293 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6294 EC_GROUP_new_curve_GFp() will now automatically use this
6296 [Nils Larsch <nla@trustcenter.de>]
6298 *) Add new lock type (CRYPTO_LOCK_BN).
6301 *) Change the ENGINE framework to automatically load engines
6302 dynamically from specific directories unless they could be
6303 found to already be built in or loaded. Move all the
6304 current engines except for the cryptodev one to a new
6306 The engines in engines/ are built as shared libraries if
6307 the "shared" options was given to ./Configure or ./config.
6308 Otherwise, they are inserted in libcrypto.a.
6309 /usr/local/ssl/engines is the default directory for dynamic
6310 engines, but that can be overridden at configure time through
6311 the usual use of --prefix and/or --openssldir, and at run
6312 time with the environment variable OPENSSL_ENGINES.
6313 [Geoff Thorpe and Richard Levitte]
6315 *) Add Makefile.shared, a helper makefile to build shared
6316 libraries. Adapt Makefile.org.
6319 *) Add version info to Win32 DLLs.
6320 [Peter 'Luna' Runestig" <peter@runestig.com>]
6322 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6323 can be added using this API to created arbitrary PKCS#12
6324 files while avoiding the low level API.
6326 New options to PKCS12_create(), key or cert can be NULL and
6327 will then be omitted from the output file. The encryption
6328 algorithm NIDs can be set to -1 for no encryption, the mac
6329 iteration count can be set to 0 to omit the mac.
6331 Enhance pkcs12 utility by making the -nokeys and -nocerts
6332 options work when creating a PKCS#12 file. New option -nomac
6333 to omit the mac, NONE can be set for an encryption algorithm.
6334 New code is modified to use the enhanced PKCS12_create()
6335 instead of the low level API.
6338 *) Extend ASN1 encoder to support indefinite length constructed
6339 encoding. This can output sequences tags and octet strings in
6340 this form. Modify pk7_asn1.c to support indefinite length
6341 encoding. This is experimental and needs additional code to
6342 be useful, such as an ASN1 bio and some enhanced streaming
6345 Extend template encode functionality so that tagging is passed
6346 down to the template encoder.
6349 *) Let 'openssl req' fail if an argument to '-newkey' is not
6350 recognized instead of using RSA as a default.
6353 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6354 As these are not official, they are not included in "ALL";
6355 the "ECCdraft" ciphersuite group alias can be used to select them.
6356 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6358 *) Add ECDH engine support.
6359 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6361 *) Add ECDH in new directory crypto/ecdh/.
6362 [Douglas Stebila (Sun Microsystems Laboratories)]
6364 *) Let BN_rand_range() abort with an error after 100 iterations
6365 without success (which indicates a broken PRNG).
6368 *) Change BN_mod_sqrt() so that it verifies that the input value
6369 is really the square of the return value. (Previously,
6370 BN_mod_sqrt would show GIGO behaviour.)
6373 *) Add named elliptic curves over binary fields from X9.62, SECG,
6374 and WAP/WTLS; add OIDs that were still missing.
6376 [Sheueling Chang Shantz and Douglas Stebila
6377 (Sun Microsystems Laboratories)]
6379 *) Extend the EC library for elliptic curves over binary fields
6380 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6383 EC_GF2m_simple_method
6387 EC_GROUP_new_curve_GF2m
6388 EC_GROUP_set_curve_GF2m
6389 EC_GROUP_get_curve_GF2m
6390 EC_POINT_set_affine_coordinates_GF2m
6391 EC_POINT_get_affine_coordinates_GF2m
6392 EC_POINT_set_compressed_coordinates_GF2m
6394 Point compression for binary fields is disabled by default for
6395 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6398 As binary polynomials are represented as BIGNUMs, various members
6399 of the EC_GROUP and EC_POINT data structures can be shared
6400 between the implementations for prime fields and binary fields;
6401 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6402 are essentially identical to their ..._GFp counterparts.
6403 (For simplicity, the '..._GFp' prefix has been dropped from
6404 various internal method names.)
6406 An internal 'field_div' method (similar to 'field_mul' and
6407 'field_sqr') has been added; this is used only for binary fields.
6409 [Sheueling Chang Shantz and Douglas Stebila
6410 (Sun Microsystems Laboratories)]
6412 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6413 through methods ('mul', 'precompute_mult').
6415 The generic implementations (now internally called 'ec_wNAF_mul'
6416 and 'ec_wNAF_precomputed_mult') remain the default if these
6417 methods are undefined.
6419 [Sheueling Chang Shantz and Douglas Stebila
6420 (Sun Microsystems Laboratories)]
6422 *) New function EC_GROUP_get_degree, which is defined through
6423 EC_METHOD. For curves over prime fields, this returns the bit
6424 length of the modulus.
6426 [Sheueling Chang Shantz and Douglas Stebila
6427 (Sun Microsystems Laboratories)]
6429 *) New functions EC_GROUP_dup, EC_POINT_dup.
6430 (These simply call ..._new and ..._copy).
6432 [Sheueling Chang Shantz and Douglas Stebila
6433 (Sun Microsystems Laboratories)]
6435 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6436 Polynomials are represented as BIGNUMs (where the sign bit is not
6437 used) in the following functions [macros]:
6440 BN_GF2m_sub [= BN_GF2m_add]
6441 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6442 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6443 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6445 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6446 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6447 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6448 BN_GF2m_cmp [= BN_ucmp]
6450 (Note that only the 'mod' functions are actually for fields GF(2^m).
6451 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6453 For some functions, an the irreducible polynomial defining a
6454 field can be given as an 'unsigned int[]' with strictly
6455 decreasing elements giving the indices of those bits that are set;
6456 i.e., p[] represents the polynomial
6457 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6459 p[0] > p[1] > ... > p[k] = 0.
6460 This applies to the following functions:
6465 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6466 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6468 BN_GF2m_mod_sqrt_arr
6469 BN_GF2m_mod_solve_quad_arr
6473 Conversion can be performed by the following functions:
6478 bntest.c has additional tests for binary polynomial arithmetic.
6480 Two implementations for BN_GF2m_mod_div() are available.
6481 The default algorithm simply uses BN_GF2m_mod_inv() and
6482 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6483 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6484 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6486 [Sheueling Chang Shantz and Douglas Stebila
6487 (Sun Microsystems Laboratories)]
6489 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6490 functionality is disabled at compile-time.
6491 [Douglas Stebila <douglas.stebila@sun.com>]
6493 *) Change default behaviour of 'openssl asn1parse' so that more
6494 information is visible when viewing, e.g., a certificate:
6496 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6497 mode the content of non-printable OCTET STRINGs is output in a
6498 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6499 avoid the appearance of a printable string.
6500 [Nils Larsch <nla@trustcenter.de>]
6502 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6504 EC_GROUP_set_asn1_flag()
6505 EC_GROUP_get_asn1_flag()
6506 EC_GROUP_set_point_conversion_form()
6507 EC_GROUP_get_point_conversion_form()
6508 These control ASN1 encoding details:
6509 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6510 has been set to OPENSSL_EC_NAMED_CURVE.
6511 - Points are encoded in uncompressed form by default; options for
6512 asn1_for are as for point2oct, namely
6513 POINT_CONVERSION_COMPRESSED
6514 POINT_CONVERSION_UNCOMPRESSED
6515 POINT_CONVERSION_HYBRID
6517 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6520 EC_GROUP_get0_seed()
6521 EC_GROUP_get_seed_len()
6522 This is used only for ASN1 purposes (so far).
6523 [Nils Larsch <nla@trustcenter.de>]
6525 *) Add 'field_type' member to EC_METHOD, which holds the NID
6526 of the appropriate field type OID. The new function
6527 EC_METHOD_get_field_type() returns this value.
6528 [Nils Larsch <nla@trustcenter.de>]
6533 EC_POINT_point2hex()
6534 EC_POINT_hex2point()
6535 providing useful interfaces to EC_POINT_point2oct() and
6536 EC_POINT_oct2point().
6537 [Nils Larsch <nla@trustcenter.de>]
6539 *) Change internals of the EC library so that the functions
6540 EC_GROUP_set_generator()
6541 EC_GROUP_get_generator()
6542 EC_GROUP_get_order()
6543 EC_GROUP_get_cofactor()
6544 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6545 to methods, which would lead to unnecessary code duplication when
6546 adding different types of curves.
6547 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6549 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6550 arithmetic, and such that modified wNAFs are generated
6551 (which avoid length expansion in many cases).
6554 *) Add a function EC_GROUP_check_discriminant() (defined via
6555 EC_METHOD) that verifies that the curve discriminant is non-zero.
6557 Add a function EC_GROUP_check() that makes some sanity tests
6558 on a EC_GROUP, its generator and order. This includes
6559 EC_GROUP_check_discriminant().
6560 [Nils Larsch <nla@trustcenter.de>]
6562 *) Add ECDSA in new directory crypto/ecdsa/.
6564 Add applications 'openssl ecparam' and 'openssl ecdsa'
6565 (these are based on 'openssl dsaparam' and 'openssl dsa').
6567 ECDSA support is also included in various other files across the
6568 library. Most notably,
6569 - 'openssl req' now has a '-newkey ecdsa:file' option;
6570 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6571 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6572 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6573 them suitable for ECDSA where domain parameters must be
6574 extracted before the specific public key;
6575 - ECDSA engine support has been added.
6576 [Nils Larsch <nla@trustcenter.de>]
6578 *) Include some named elliptic curves, and add OIDs from X9.62,
6579 SECG, and WAP/WTLS. Each curve can be obtained from the new
6581 EC_GROUP_new_by_curve_name(),
6582 and the list of available named curves can be obtained with
6583 EC_get_builtin_curves().
6584 Also add a 'curve_name' member to EC_GROUP objects, which can be
6586 EC_GROUP_set_curve_name()
6587 EC_GROUP_get_curve_name()
6588 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6590 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6591 was actually never needed) and in BN_mul(). The removal in BN_mul()
6592 required a small change in bn_mul_part_recursive() and the addition
6593 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6594 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6595 bn_sub_words() and bn_add_words() except they take arrays with
6599 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6601 *) Cleanse PEM buffers before freeing them since they may contain
6603 [Benjamin Bennett <ben@psc.edu>]
6605 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6606 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6607 authentication-only ciphersuites.
6610 *) Since AES128 and AES256 share a single mask bit in the logic of
6611 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6612 kludge to work properly if AES128 is available and AES256 isn't.
6615 *) Expand security boundary to match 1.1.1 module.
6618 *) Remove redundant features: hash file source, editing of test vectors
6619 modify fipsld to use external fips_premain.c signature.
6622 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6623 run algorithm test programs.
6626 *) Make algorithm test programs more tolerant of whitespace.
6629 *) Have SSL/TLS server implementation tolerate "mismatched" record
6630 protocol version while receiving ClientHello even if the
6631 ClientHello is fragmented. (The server can't insist on the
6632 particular protocol version it has chosen before the ServerHello
6633 message has informed the client about his choice.)
6636 *) Load error codes if they are not already present instead of using a
6637 static variable. This allows them to be cleanly unloaded and reloaded.
6640 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6642 *) Introduce limits to prevent malicious keys being able to
6643 cause a denial of service. (CVE-2006-2940)
6644 [Steve Henson, Bodo Moeller]
6646 *) Fix ASN.1 parsing of certain invalid structures that can result
6647 in a denial of service. (CVE-2006-2937) [Steve Henson]
6649 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6650 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6652 *) Fix SSL client code which could crash if connecting to a
6653 malicious SSLv2 server. (CVE-2006-4343)
6654 [Tavis Ormandy and Will Drewry, Google Security Team]
6656 *) Change ciphersuite string processing so that an explicit
6657 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6658 will no longer include "AES128-SHA"), and any other similar
6659 ciphersuite (same bitmap) from *other* protocol versions (so that
6660 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6661 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6662 changes from 0.9.8b and 0.9.8d.
6665 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6667 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6668 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6670 *) Change the Unix randomness entropy gathering to use poll() when
6671 possible instead of select(), since the latter has some
6672 undesirable limitations.
6673 [Darryl Miles via Richard Levitte and Bodo Moeller]
6675 *) Disable rogue ciphersuites:
6677 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6678 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6679 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6681 The latter two were purportedly from
6682 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6685 Also deactivate the remaining ciphersuites from
6686 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6687 unofficial, and the ID has long expired.
6690 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6691 dual-core machines) and other potential thread-safety issues.
6694 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6696 *) Adapt fipsld and the build system to link against the validated FIPS
6697 module in FIPS mode.
6700 *) Fixes for VC++ 2005 build under Windows.
6703 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6704 from a Windows bash shell such as MSYS. It is autodetected from the
6705 "config" script when run from a VC++ environment. Modify standard VC++
6706 build to use fipscanister.o from the GNU make build.
6709 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6711 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6712 The value now differs depending on if you build for FIPS or not.
6713 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6714 safely run with a non-FIPSed libcrypto, as it may crash because of
6715 the difference induced by this change.
6718 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6720 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6721 (part of SSL_OP_ALL). This option used to disable the
6722 countermeasure against man-in-the-middle protocol-version
6723 rollback in the SSL 2.0 server implementation, which is a bad
6724 idea. (CVE-2005-2969)
6726 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6727 for Information Security, National Institute of Advanced Industrial
6728 Science and Technology [AIST], Japan)]
6730 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6731 mainly for FIPS compliance and not fully integrated at this stage.
6734 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6735 the exponentiation using a fixed-length exponent. (Otherwise,
6736 the information leaked through timing could expose the secret key
6737 after many signatures; cf. Bleichenbacher's attack on DSA with
6741 *) Make a new fixed-window mod_exp implementation the default for
6742 RSA, DSA, and DH private-key operations so that the sequence of
6743 squares and multiplies and the memory access pattern are
6744 independent of the particular secret key. This will mitigate
6745 cache-timing and potential related attacks.
6747 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6748 and this is automatically used by BN_mod_exp_mont() if the new flag
6749 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6750 will use this BN flag for private exponents unless the flag
6751 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6752 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6754 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6756 *) Change the client implementation for SSLv23_method() and
6757 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6758 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6759 (Previously, the SSL 2.0 backwards compatible Client Hello
6760 message format would be used even with SSL_OP_NO_SSLv2.)
6763 *) Add support for smime-type MIME parameter in S/MIME messages which some
6767 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6768 a threadsafe manner. Modify rsa code to use new function and add calls
6769 to dsa and dh code (which had race conditions before).
6772 *) Include the fixed error library code in the C error file definitions
6773 instead of fixing them up at runtime. This keeps the error code
6774 structures constant.
6777 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6779 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6782 *) Fixes for newer kerberos headers. NB: the casts are needed because
6783 the 'length' field is signed on one version and unsigned on another
6784 with no (?) obvious way to tell the difference, without these VC++
6785 complains. Also the "definition" of FAR (blank) is no longer included
6786 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6787 some needed definitions.
6790 *) Undo Cygwin change.
6793 *) Added support for proxy certificates according to RFC 3820.
6794 Because they may be a security thread to unaware applications,
6795 they must be explicitly allowed in run-time. See
6796 docs/HOWTO/proxy_certificates.txt for further information.
6799 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6801 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6802 server and client random values. Previously
6803 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6804 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6806 This change has negligible security impact because:
6808 1. Server and client random values still have 24 bytes of pseudo random
6811 2. Server and client random values are sent in the clear in the initial
6814 3. The master secret is derived using the premaster secret (48 bytes in
6815 size for static RSA ciphersuites) as well as client server and random
6818 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6821 [Stephen Henson, reported by UK NISCC]
6823 *) Use Windows randomness collection on Cygwin.
6826 *) Fix hang in EGD/PRNGD query when communication socket is closed
6827 prematurely by EGD/PRNGD.
6828 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6830 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6833 *) Back-port of selected performance improvements from development
6834 branch, as well as improved support for PowerPC platforms.
6837 *) Add lots of checks for memory allocation failure, error codes to indicate
6838 failure and freeing up memory if a failure occurs.
6839 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6841 *) Add new -passin argument to dgst.
6844 *) Perform some character comparisons of different types in X509_NAME_cmp:
6845 this is needed for some certificates that re-encode DNs into UTF8Strings
6846 (in violation of RFC3280) and can't or won't issue name rollover
6850 *) Make an explicit check during certificate validation to see that
6851 the CA setting in each certificate on the chain is correct. As a
6852 side effect always do the following basic checks on extensions,
6853 not just when there's an associated purpose to the check:
6855 - if there is an unhandled critical extension (unless the user
6856 has chosen to ignore this fault)
6857 - if the path length has been exceeded (if one is set at all)
6858 - that certain extensions fit the associated purpose (if one has
6862 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6864 *) Avoid a race condition when CRLs are checked in a multi threaded
6865 environment. This would happen due to the reordering of the revoked
6866 entries during signature checking and serial number lookup. Now the
6867 encoding is cached and the serial number sort performed under a lock.
6868 Add new STACK function sk_is_sorted().
6871 *) Add Delta CRL to the extension code.
6874 *) Various fixes to s3_pkt.c so alerts are sent properly.
6875 [David Holmes <d.holmes@f5.com>]
6877 *) Reduce the chances of duplicate issuer name and serial numbers (in
6878 violation of RFC3280) using the OpenSSL certificate creation utilities.
6879 This is done by creating a random 64 bit value for the initial serial
6880 number when a serial number file is created or when a self signed
6881 certificate is created using 'openssl req -x509'. The initial serial
6882 number file is created using 'openssl x509 -next_serial' in CA.pl
6883 rather than being initialized to 1.
6886 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6888 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6889 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6890 [Joe Orton, Steve Henson]
6892 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6894 [Joe Orton, Steve Henson]
6896 *) Make it possible to have multiple active certificates with the same
6897 subject in the CA index file. This is done only if the keyword
6898 'unique_subject' is set to 'no' in the main CA section (default
6899 if 'CA_default') of the configuration file. The value is saved
6900 with the database itself in a separate index attribute file,
6901 named like the index file with '.attr' appended to the name.
6904 *) X509 verify fixes. Disable broken certificate workarounds when
6905 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6906 keyUsage extension present. Don't accept CRLs with unhandled critical
6907 extensions: since verify currently doesn't process CRL extensions this
6908 rejects a CRL with *any* critical extensions. Add new verify error codes
6912 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6913 A clarification of RFC2560 will require the use of OCTET STRINGs and
6914 some implementations cannot handle the current raw format. Since OpenSSL
6915 copies and compares OCSP nonces as opaque blobs without any attempt at
6916 parsing them this should not create any compatibility issues.
6919 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6920 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6921 this HMAC (and other) operations are several times slower than OpenSSL
6925 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6926 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6928 *) Use the correct content when signing type "other".
6931 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6933 *) Fix various bugs revealed by running the NISCC test suite:
6935 Stop out of bounds reads in the ASN1 code when presented with
6936 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6938 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6940 If verify callback ignores invalid public key errors don't try to check
6941 certificate signature with the NULL public key.
6945 *) New -ignore_err option in ocsp application to stop the server
6946 exiting on the first error in a request.
6949 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6950 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6954 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6955 extra data after the compression methods not only for TLS 1.0
6956 but also for SSL 3.0 (as required by the specification).
6957 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6959 *) Change X509_certificate_type() to mark the key as exported/exportable
6960 when it's 512 *bits* long, not 512 bytes.
6963 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6964 blocks during encryption.
6967 *) Various fixes to base64 BIO and non blocking I/O. On write
6968 flushes were not handled properly if the BIO retried. On read
6969 data was not being buffered properly and had various logic bugs.
6970 This also affects blocking I/O when the data being decoded is a
6974 *) Various S/MIME bugfixes and compatibility changes:
6975 output correct application/pkcs7 MIME type if
6976 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6977 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6978 of files as .eml work). Correctly handle very long lines in MIME
6982 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6984 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6985 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6986 a protocol version number mismatch like a decryption error
6987 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6990 *) Turn on RSA blinding by default in the default implementation
6991 to avoid a timing attack. Applications that don't want it can call
6992 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6993 They would be ill-advised to do so in most cases.
6994 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6996 *) Change RSA blinding code so that it works when the PRNG is not
6997 seeded (in this case, the secret RSA exponent is abused as
6998 an unpredictable seed -- if it is not unpredictable, there
6999 is no point in blinding anyway). Make RSA blinding thread-safe
7000 by remembering the creator's thread ID in rsa->blinding and
7001 having all other threads use local one-time blinding factors
7002 (this requires more computation than sharing rsa->blinding, but
7003 avoids excessive locking; and if an RSA object is not shared
7004 between threads, blinding will still be very fast).
7007 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
7008 ENGINE as defaults for all supported algorithms irrespective of
7009 the 'flags' parameter. 'flags' is now honoured, so applications
7010 should make sure they are passing it correctly.
7013 *) Target "mingw" now allows native Windows code to be generated in
7014 the Cygwin environment as well as with the MinGW compiler.
7017 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7019 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7020 via timing by performing a MAC computation even if incorrect
7021 block cipher padding has been found. This is a countermeasure
7022 against active attacks where the attacker has to distinguish
7023 between bad padding and a MAC verification error. (CVE-2003-0078)
7025 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7026 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7027 Martin Vuagnoux (EPFL, Ilion)]
7029 *) Make the no-err option work as intended. The intention with no-err
7030 is not to have the whole error stack handling routines removed from
7031 libcrypto, it's only intended to remove all the function name and
7032 reason texts, thereby removing some of the footprint that may not
7033 be interesting if those errors aren't displayed anyway.
7035 NOTE: it's still possible for any application or module to have its
7036 own set of error texts inserted. The routines are there, just not
7037 used by default when no-err is given.
7040 *) Add support for FreeBSD on IA64.
7041 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7043 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7044 Kerberos function mit_des_cbc_cksum(). Before this change,
7045 the value returned by DES_cbc_cksum() was like the one from
7046 mit_des_cbc_cksum(), except the bytes were swapped.
7047 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7049 *) Allow an application to disable the automatic SSL chain building.
7050 Before this a rather primitive chain build was always performed in
7051 ssl3_output_cert_chain(): an application had no way to send the
7052 correct chain if the automatic operation produced an incorrect result.
7054 Now the chain builder is disabled if either:
7056 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7058 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7060 The reasoning behind this is that an application would not want the
7061 auto chain building to take place if extra chain certificates are
7062 present and it might also want a means of sending no additional
7063 certificates (for example the chain has two certificates and the
7067 *) Add the possibility to build without the ENGINE framework.
7068 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7070 *) Under Win32 gmtime() can return NULL: check return value in
7071 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7074 *) DSA routines: under certain error conditions uninitialized BN objects
7075 could be freed. Solution: make sure initialization is performed early
7076 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7077 Nils Larsch <nla@trustcenter.de> via PR#459)
7080 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7081 checked on reconnect on the client side, therefore session resumption
7082 could still fail with a "ssl session id is different" error. This
7083 behaviour is masked when SSL_OP_ALL is used due to
7084 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7085 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7086 followup to PR #377.
7089 *) IA-32 assembler support enhancements: unified ELF targets, support
7090 for SCO/Caldera platforms, fix for Cygwin shared build.
7093 *) Add support for FreeBSD on sparc64. As a consequence, support for
7094 FreeBSD on non-x86 processors is separate from x86 processors on
7095 the config script, much like the NetBSD support.
7096 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7098 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7100 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7103 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7104 code (06) was taken as the first octet of the session ID and the last
7105 octet was ignored consequently. As a result SSLv2 client side session
7106 caching could not have worked due to the session ID mismatch between
7108 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7112 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7113 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7117 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7118 seems that in spite of existing for more than a year, many application
7119 author have done nothing to provide the necessary callbacks, which
7120 means that this particular engine will not work properly anywhere.
7121 This is a very unfortunate situation which forces us, in the name
7122 of usability, to give the hw_ncipher.c a static lock, which is part
7124 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7125 appear in 0.9.8 or later. We EXPECT application authors to have
7126 dealt properly with this when 0.9.8 is released (unless we actually
7127 make such changes in the libcrypto locking code that changes will
7128 have to be made anyway).
7131 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7132 octets have been read, EOF or an error occurs. Without this change
7133 some truncated ASN1 structures will not produce an error.
7136 *) Disable Heimdal support, since it hasn't been fully implemented.
7137 Still give the possibility to force the use of Heimdal, but with
7138 warnings and a request that patches get sent to openssl-dev.
7141 *) Add the VC-CE target, introduce the WINCE sysname, and add
7142 INSTALL.WCE and appropriate conditionals to make it build.
7143 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7145 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7146 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7147 edit numbers of the version.
7148 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7150 *) Introduce safe string copy and catenation functions
7151 (BUF_strlcpy() and BUF_strlcat()).
7152 [Ben Laurie (CHATS) and Richard Levitte]
7154 *) Avoid using fixed-size buffers for one-line DNs.
7155 [Ben Laurie (CHATS)]
7157 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7158 resizing buffers containing secrets, and use where appropriate.
7159 [Ben Laurie (CHATS)]
7161 *) Avoid using fixed size buffers for configuration file location.
7162 [Ben Laurie (CHATS)]
7164 *) Avoid filename truncation for various CA files.
7165 [Ben Laurie (CHATS)]
7167 *) Use sizeof in preference to magic numbers.
7168 [Ben Laurie (CHATS)]
7170 *) Avoid filename truncation in cert requests.
7171 [Ben Laurie (CHATS)]
7173 *) Add assertions to check for (supposedly impossible) buffer
7175 [Ben Laurie (CHATS)]
7177 *) Don't cache truncated DNS entries in the local cache (this could
7178 potentially lead to a spoofing attack).
7179 [Ben Laurie (CHATS)]
7181 *) Fix various buffers to be large enough for hex/decimal
7182 representations in a platform independent manner.
7183 [Ben Laurie (CHATS)]
7185 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7186 resizing buffers containing secrets, and use where appropriate.
7187 [Ben Laurie (CHATS)]
7189 *) Add BIO_indent() to avoid much slightly worrying code to do
7191 [Ben Laurie (CHATS)]
7193 *) Convert sprintf()/BIO_puts() to BIO_printf().
7194 [Ben Laurie (CHATS)]
7196 *) buffer_gets() could terminate with the buffer only half
7198 [Ben Laurie (CHATS)]
7200 *) Add assertions to prevent user-supplied crypto functions from
7201 overflowing internal buffers by having large block sizes, etc.
7202 [Ben Laurie (CHATS)]
7204 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7206 [Ben Laurie (CHATS)]
7208 *) Eliminate unused copy of key in RC4.
7209 [Ben Laurie (CHATS)]
7211 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7212 [Ben Laurie (CHATS)]
7214 *) Fix off-by-one error in EGD path.
7215 [Ben Laurie (CHATS)]
7217 *) If RANDFILE path is too long, ignore instead of truncating.
7218 [Ben Laurie (CHATS)]
7220 *) Eliminate unused and incorrectly sized X.509 structure
7222 [Ben Laurie (CHATS)]
7224 *) Eliminate unused and dangerous function knumber().
7225 [Ben Laurie (CHATS)]
7227 *) Eliminate unused and dangerous structure, KSSL_ERR.
7228 [Ben Laurie (CHATS)]
7230 *) Protect against overlong session ID context length in an encoded
7231 session object. Since these are local, this does not appear to be
7233 [Ben Laurie (CHATS)]
7235 *) Change from security patch (see 0.9.6e below) that did not affect
7236 the 0.9.6 release series:
7238 Remote buffer overflow in SSL3 protocol - an attacker could
7239 supply an oversized master key in Kerberos-enabled versions.
7241 [Ben Laurie (CHATS)]
7243 *) Change the SSL kerb5 codes to match RFC 2712.
7246 *) Make -nameopt work fully for req and add -reqopt switch.
7247 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7249 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7250 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7252 *) Make sure tests can be performed even if the corresponding algorithms
7253 have been removed entirely. This was also the last step to make
7254 OpenSSL compilable with DJGPP under all reasonable conditions.
7255 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7257 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7258 to allow version independent disabling of normally unselected ciphers,
7259 which may be activated as a side-effect of selecting a single cipher.
7261 (E.g., cipher list string "RSA" enables ciphersuites that are left
7262 out of "ALL" because they do not provide symmetric encryption.
7263 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7264 [Lutz Jaenicke, Bodo Moeller]
7266 *) Add appropriate support for separate platform-dependent build
7267 directories. The recommended way to make a platform-dependent
7268 build directory is the following (tested on Linux), maybe with
7271 # Place yourself outside of the OpenSSL source tree. In
7272 # this example, the environment variable OPENSSL_SOURCE
7273 # is assumed to contain the absolute OpenSSL source directory.
7274 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7275 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7276 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7277 mkdir -p `dirname $F`
7278 ln -s $OPENSSL_SOURCE/$F $F
7281 To be absolutely sure not to disturb the source tree, a "make clean"
7282 is a good thing. If it isn't successful, don't worry about it,
7283 it probably means the source directory is very clean.
7286 *) Make sure any ENGINE control commands make local copies of string
7287 pointers passed to them whenever necessary. Otherwise it is possible
7288 the caller may have overwritten (or deallocated) the original string
7289 data when a later ENGINE operation tries to use the stored values.
7290 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7292 *) Improve diagnostics in file reading and command-line digests.
7293 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7295 *) Add AES modes CFB and OFB to the object database. Correct an
7296 error in AES-CFB decryption.
7299 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7300 allows existing EVP_CIPHER_CTX structures to be reused after
7301 calling EVP_*Final(). This behaviour is used by encryption
7302 BIOs and some applications. This has the side effect that
7303 applications must explicitly clean up cipher contexts with
7304 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7307 *) Check the values of dna and dnb in bn_mul_recursive before calling
7308 bn_mul_comba (a non zero value means the a or b arrays do not contain
7309 n2 elements) and fallback to bn_mul_normal if either is not zero.
7312 *) Fix escaping of non-ASCII characters when using the -subj option
7313 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7316 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7317 form for "surname", serialNumber has no short form.
7318 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7319 therefore remove "mail" short name for "internet 7".
7320 The OID for unique identifiers in X509 certificates is
7321 x500UniqueIdentifier, not uniqueIdentifier.
7322 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7325 *) Add an "init" command to the ENGINE config module and auto initialize
7326 ENGINEs. Without any "init" command the ENGINE will be initialized
7327 after all ctrl commands have been executed on it. If init=1 the
7328 ENGINE is initialized at that point (ctrls before that point are run
7329 on the uninitialized ENGINE and after on the initialized one). If
7330 init=0 then the ENGINE will not be initialized at all.
7333 *) Fix the 'app_verify_callback' interface so that the user-defined
7334 argument is actually passed to the callback: In the
7335 SSL_CTX_set_cert_verify_callback() prototype, the callback
7336 declaration has been changed from
7339 int (*cb)(X509_STORE_CTX *,void *);
7340 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7341 i=s->ctx->app_verify_callback(&ctx)
7342 has been changed into
7343 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7345 To update applications using SSL_CTX_set_cert_verify_callback(),
7346 a dummy argument can be added to their callback functions.
7347 [D. K. Smetters <smetters@parc.xerox.com>]
7349 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7350 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7352 *) Add and OPENSSL_LOAD_CONF define which will cause
7353 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7354 This allows older applications to transparently support certain
7355 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7356 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7357 load the config file and OPENSSL_add_all_algorithms_conf() which will
7358 always load it have also been added.
7361 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7362 Adjust NIDs and EVP layer.
7363 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7365 *) Config modules support in openssl utility.
7367 Most commands now load modules from the config file,
7368 though in a few (such as version) this isn't done
7369 because it couldn't be used for anything.
7371 In the case of ca and req the config file used is
7372 the same as the utility itself: that is the -config
7373 command line option can be used to specify an
7377 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7378 use "openssl_conf" if filename is NULL use default openssl config file.
7381 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7382 config section name. Add a new flag to tolerate a missing config file
7383 and move code to CONF_modules_load_file().
7386 *) Support for crypto accelerator cards from Accelerated Encryption
7387 Processing, www.aep.ie. (Use engine 'aep')
7388 The support was copied from 0.9.6c [engine] and adapted/corrected
7389 to work with the new engine framework.
7390 [AEP Inc. and Richard Levitte]
7392 *) Support for SureWare crypto accelerator cards from Baltimore
7393 Technologies. (Use engine 'sureware')
7394 The support was copied from 0.9.6c [engine] and adapted
7395 to work with the new engine framework.
7398 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7399 make the newer ENGINE framework commands for the CHIL engine work.
7400 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7402 *) Make it possible to produce shared libraries on ReliantUNIX.
7403 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7405 *) Add the configuration target debug-linux-ppro.
7406 Make 'openssl rsa' use the general key loading routines
7407 implemented in apps.c, and make those routines able to
7408 handle the key format FORMAT_NETSCAPE and the variant
7410 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7412 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7413 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7415 *) Add -keyform to rsautl, and document -engine.
7416 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7418 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7419 BIO_R_NO_SUCH_FILE error code rather than the generic
7420 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7423 *) Add new functions
7425 ERR_peek_last_error_line
7426 ERR_peek_last_error_line_data.
7427 These are similar to
7430 ERR_peek_error_line_data,
7431 but report on the latest error recorded rather than the first one
7432 still in the error queue.
7433 [Ben Laurie, Bodo Moeller]
7435 *) default_algorithms option in ENGINE config module. This allows things
7437 default_algorithms = ALL
7438 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7441 *) Preliminary ENGINE config module.
7444 *) New experimental application configuration code.
7447 *) Change the AES code to follow the same name structure as all other
7448 symmetric ciphers, and behave the same way. Move everything to
7449 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7450 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7452 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7453 [Ben Laurie and Theo de Raadt]
7455 *) Add option to output public keys in req command.
7456 [Massimiliano Pala madwolf@openca.org]
7458 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7459 (up to about 10% better than before for P-192 and P-224).
7462 *) New functions/macros
7464 SSL_CTX_set_msg_callback(ctx, cb)
7465 SSL_CTX_set_msg_callback_arg(ctx, arg)
7466 SSL_set_msg_callback(ssl, cb)
7467 SSL_set_msg_callback_arg(ssl, arg)
7469 to request calling a callback function
7471 void cb(int write_p, int version, int content_type,
7472 const void *buf, size_t len, SSL *ssl, void *arg)
7474 whenever a protocol message has been completely received
7475 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7476 protocol version according to which the SSL library interprets
7477 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7478 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7479 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7480 specification (change_cipher_spec(20), alert(21), handshake(22)).
7481 'buf' and 'len' point to the actual message, 'ssl' to the
7482 SSL object, and 'arg' is the application-defined value set by
7483 SSL[_CTX]_set_msg_callback_arg().
7485 'openssl s_client' and 'openssl s_server' have new '-msg' options
7486 to enable a callback that displays all protocol messages.
7489 *) Change the shared library support so shared libraries are built as
7490 soon as the corresponding static library is finished, and thereby get
7491 openssl and the test programs linked against the shared library.
7492 This still only happens when the keyword "shard" has been given to
7493 the configuration scripts.
7495 NOTE: shared library support is still an experimental thing, and
7496 backward binary compatibility is still not guaranteed.
7497 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7499 *) Add support for Subject Information Access extension.
7500 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7502 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7503 additional bytes when new memory had to be allocated, not just
7504 when reusing an existing buffer.
7507 *) New command line and configuration option 'utf8' for the req command.
7508 This allows field values to be specified as UTF8 strings.
7511 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7512 runs for the former and machine-readable output for the latter.
7515 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7516 of the e-mail address in the DN (i.e., it will go into a certificate
7517 extension only). The new configuration file option 'email_in_dn = no'
7518 has the same effect.
7519 [Massimiliano Pala madwolf@openca.org]
7521 *) Change all functions with names starting with des_ to be starting
7522 with DES_ instead. Add wrappers that are compatible with libdes,
7523 but are named _ossl_old_des_*. Finally, add macros that map the
7524 des_* symbols to the corresponding _ossl_old_des_* if libdes
7525 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7526 desired, the des_* symbols will be mapped to DES_*, with one
7529 Since we provide two compatibility mappings, the user needs to
7530 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7531 compatibility is desired. The default (i.e., when that macro
7532 isn't defined) is OpenSSL 0.9.6c compatibility.
7534 There are also macros that enable and disable the support of old
7535 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7536 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7537 are defined, the default will apply: to support the old des routines.
7539 In either case, one must include openssl/des.h to get the correct
7540 definitions. Do not try to just include openssl/des_old.h, that
7543 NOTE: This is a major break of an old API into a new one. Software
7544 authors are encouraged to switch to the DES_ style functions. Some
7545 time in the future, des_old.h and the libdes compatibility functions
7546 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7547 default), and then completely removed.
7550 *) Test for certificates which contain unsupported critical extensions.
7551 If such a certificate is found during a verify operation it is
7552 rejected by default: this behaviour can be overridden by either
7553 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7554 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7555 X509_supported_extension() has also been added which returns 1 if a
7556 particular extension is supported.
7559 *) Modify the behaviour of EVP cipher functions in similar way to digests
7560 to retain compatibility with existing code.
7563 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7564 compatibility with existing code. In particular the 'ctx' parameter does
7565 not have to be to be initialized before the call to EVP_DigestInit() and
7566 it is tidied up after a call to EVP_DigestFinal(). New function
7567 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7568 EVP_MD_CTX_copy() changed to not require the destination to be
7569 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7570 requires the destination to be valid.
7572 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7573 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7576 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7577 so that complete 'Handshake' protocol structures are kept in memory
7578 instead of overwriting 'msg_type' and 'length' with 'body' data.
7581 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7582 [Massimo Santin via Richard Levitte]
7584 *) Major restructuring to the underlying ENGINE code. This includes
7585 reduction of linker bloat, separation of pure "ENGINE" manipulation
7586 (initialisation, etc) from functionality dealing with implementations
7587 of specific crypto interfaces. This change also introduces integrated
7588 support for symmetric ciphers and digest implementations - so ENGINEs
7589 can now accelerate these by providing EVP_CIPHER and EVP_MD
7590 implementations of their own. This is detailed in crypto/engine/README
7591 as it couldn't be adequately described here. However, there are a few
7592 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7593 were changed in the original introduction of ENGINE code have now
7594 reverted back - the hooking from this code to ENGINE is now a good
7595 deal more passive and at run-time, operations deal directly with
7596 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7597 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7598 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7599 they were not being used by the framework as there is no concept of a
7600 BIGNUM_METHOD and they could not be generalised to the new
7601 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7602 ENGINE_cpy() has been removed as it cannot be consistently defined in
7606 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7609 *) Change mkdef.pl to sort symbols that get the same entry number,
7610 and make sure the automatically generated functions ERR_load_*
7611 become part of libeay.num as well.
7614 *) New function SSL_renegotiate_pending(). This returns true once
7615 renegotiation has been requested (either SSL_renegotiate() call
7616 or HelloRequest/ClientHello received from the peer) and becomes
7617 false once a handshake has been completed.
7618 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7619 sends a HelloRequest, but does not ensure that a handshake takes
7620 place. SSL_renegotiate_pending() is useful for checking if the
7621 client has followed the request.)
7624 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7625 By default, clients may request session resumption even during
7626 renegotiation (if session ID contexts permit); with this option,
7627 session resumption is possible only in the first handshake.
7629 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7630 more bits available for options that should not be part of
7631 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7634 *) Add some demos for certificate and certificate request creation.
7637 *) Make maximum certificate chain size accepted from the peer application
7638 settable (SSL*_get/set_max_cert_list()), as proposed by
7639 "Douglas E. Engert" <deengert@anl.gov>.
7642 *) Add support for shared libraries for Unixware-7
7643 (Boyd Lynn Gerber <gerberb@zenez.com>).
7646 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7647 be done prior to destruction. Use this to unload error strings from
7648 ENGINEs that load their own error strings. NB: This adds two new API
7649 functions to "get" and "set" this destroy handler in an ENGINE.
7652 *) Alter all existing ENGINE implementations (except "openssl" and
7653 "openbsd") to dynamically instantiate their own error strings. This
7654 makes them more flexible to be built both as statically-linked ENGINEs
7655 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7656 Also, add stub code to each that makes building them as self-contained
7657 shared-libraries easier (see README.ENGINE).
7660 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7661 implementations into applications that are completely implemented in
7662 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7663 commands that can be used to configure what shared-library to load and
7664 to control aspects of the way it is handled. Also, made an update to
7665 the README.ENGINE file that brings its information up-to-date and
7666 provides some information and instructions on the "dynamic" ENGINE
7667 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7670 *) Make it possible to unload ranges of ERR strings with a new
7671 "ERR_unload_strings" function.
7674 *) Add a copy() function to EVP_MD.
7677 *) Make EVP_MD routines take a context pointer instead of just the
7678 md_data void pointer.
7681 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7682 that the digest can only process a single chunk of data
7683 (typically because it is provided by a piece of
7684 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7685 is only going to provide a single chunk of data, and hence the
7686 framework needn't accumulate the data for oneshot drivers.
7689 *) As with "ERR", make it possible to replace the underlying "ex_data"
7690 functions. This change also alters the storage and management of global
7691 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7692 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7693 index counters. The API functions that use this state have been changed
7694 to take a "class_index" rather than pointers to the class's local STACK
7695 and counter, and there is now an API function to dynamically create new
7696 classes. This centralisation allows us to (a) plug a lot of the
7697 thread-safety problems that existed, and (b) makes it possible to clean
7698 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7699 such data would previously have always leaked in application code and
7700 workarounds were in place to make the memory debugging turn a blind eye
7701 to it. Application code that doesn't use this new function will still
7702 leak as before, but their memory debugging output will announce it now
7703 rather than letting it slide.
7705 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7706 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7707 has a return value to indicate success or failure.
7710 *) Make it possible to replace the underlying "ERR" functions such that the
7711 global state (2 LHASH tables and 2 locks) is only used by the "default"
7712 implementation. This change also adds two functions to "get" and "set"
7713 the implementation prior to it being automatically set the first time
7714 any other ERR function takes place. Ie. an application can call "get",
7715 pass the return value to a module it has just loaded, and that module
7716 can call its own "set" function using that value. This means the
7717 module's "ERR" operations will use (and modify) the error state in the
7718 application and not in its own statically linked copy of OpenSSL code.
7721 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7722 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7723 the operation, and provides a more encapsulated way for external code
7724 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7725 to use these functions rather than manually incrementing the counts.
7727 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7730 *) Add EVP test program.
7733 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7736 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7737 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7738 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7739 These allow a CRL to be built without having to access X509_CRL fields
7740 directly. Modify 'ca' application to use new functions.
7743 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7744 bug workarounds. Rollback attack detection is a security feature.
7745 The problem will only arise on OpenSSL servers when TLSv1 is not
7746 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7747 Software authors not wanting to support TLSv1 will have special reasons
7748 for their choice and can explicitly enable this option.
7749 [Bodo Moeller, Lutz Jaenicke]
7751 *) Rationalise EVP so it can be extended: don't include a union of
7752 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7753 (similar to those existing for EVP_CIPHER_CTX).
7758 EVP_MD_CTX_init(&md); /* new function call */
7759 EVP_DigestInit(&md, EVP_sha1());
7760 EVP_DigestUpdate(&md, in, len);
7761 EVP_DigestFinal(&md, out, NULL);
7762 EVP_MD_CTX_cleanup(&md); /* new function call */
7766 *) Make DES key schedule conform to the usual scheme, as well as
7767 correcting its structure. This means that calls to DES functions
7768 now have to pass a pointer to a des_key_schedule instead of a
7769 plain des_key_schedule (which was actually always a pointer
7772 des_key_schedule ks;
7774 des_set_key_checked(..., &ks);
7775 des_ncbc_encrypt(..., &ks, ...);
7777 (Note that a later change renames 'des_...' into 'DES_...'.)
7780 *) Initial reduction of linker bloat: the use of some functions, such as
7781 PEM causes large amounts of unused functions to be linked in due to
7782 poor organisation. For example pem_all.c contains every PEM function
7783 which has a knock on effect of linking in large amounts of (unused)
7784 ASN1 code. Grouping together similar functions and splitting unrelated
7785 functions prevents this.
7788 *) Cleanup of EVP macros.
7791 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7792 correct _ecb suffix.
7795 *) Add initial OCSP responder support to ocsp application. The
7796 revocation information is handled using the text based index
7797 use by the ca application. The responder can either handle
7798 requests generated internally, supplied in files (for example
7799 via a CGI script) or using an internal minimal server.
7802 *) Add configuration choices to get zlib compression for TLS.
7805 *) Changes to Kerberos SSL for RFC 2712 compliance:
7806 1. Implemented real KerberosWrapper, instead of just using
7807 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7808 2. Implemented optional authenticator field of KerberosWrapper.
7810 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7811 and authenticator structs; see crypto/krb5/.
7813 Generalized Kerberos calls to support multiple Kerberos libraries.
7814 [Vern Staats <staatsvr@asc.hpc.mil>,
7815 Jeffrey Altman <jaltman@columbia.edu>
7816 via Richard Levitte]
7818 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7819 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7820 values for each of the key sizes rather than having just
7821 parameters (and 'speed' generating keys each time).
7824 *) Speed up EVP routines.
7827 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7828 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7829 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7830 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7832 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7833 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7834 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7837 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7839 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7842 *) Added the OS2-EMX target.
7843 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7845 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7846 to support NCONF routines in extension code. New function CONF_set_nconf()
7847 to allow functions which take an NCONF to also handle the old LHASH
7848 structure: this means that the old CONF compatible routines can be
7849 retained (in particular wrt extensions) without having to duplicate the
7850 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7853 *) Enhance the general user interface with mechanisms for inner control
7854 and with possibilities to have yes/no kind of prompts.
7857 *) Change all calls to low level digest routines in the library and
7858 applications to use EVP. Add missing calls to HMAC_cleanup() and
7859 don't assume HMAC_CTX can be copied using memcpy().
7860 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7862 *) Add the possibility to control engines through control names but with
7863 arbitrary arguments instead of just a string.
7864 Change the key loaders to take a UI_METHOD instead of a callback
7865 function pointer. NOTE: this breaks binary compatibility with earlier
7866 versions of OpenSSL [engine].
7867 Adapt the nCipher code for these new conditions and add a card insertion
7871 *) Enhance the general user interface with mechanisms to better support
7872 dialog box interfaces, application-defined prompts, the possibility
7873 to use defaults (for example default passwords from somewhere else)
7874 and interrupts/cancellations.
7877 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7878 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7881 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7882 tidy up some unnecessarily weird code in 'sk_new()').
7883 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7885 *) Change the key loading routines for ENGINEs to use the same kind
7886 callback (pem_password_cb) as all other routines that need this
7890 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7891 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7892 than this minimum value is recommended.
7895 *) New random seeder for OpenVMS, using the system process statistics
7896 that are easily reachable.
7899 *) Windows apparently can't transparently handle global
7900 variables defined in DLLs. Initialisations such as:
7902 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7904 won't compile. This is used by the any applications that need to
7905 declare their own ASN1 modules. This was fixed by adding the option
7906 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7907 needed for static libraries under Win32.
7910 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7911 setting of purpose and trust fields. New X509_STORE trust and
7912 purpose functions and tidy up setting in other SSL functions.
7915 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7916 structure. These are inherited by X509_STORE_CTX when it is
7917 initialised. This allows various defaults to be set in the
7918 X509_STORE structure (such as flags for CRL checking and custom
7919 purpose or trust settings) for functions which only use X509_STORE_CTX
7920 internally such as S/MIME.
7922 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7923 trust settings if they are not set in X509_STORE. This allows X509_STORE
7924 purposes and trust (in S/MIME for example) to override any set by default.
7926 Add command line options for CRL checking to smime, s_client and s_server
7930 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7931 are set then the CRL is looked up in the X509_STORE structure and
7932 its validity and signature checked, then if the certificate is found
7933 in the CRL the verify fails with a revoked error.
7935 Various new CRL related callbacks added to X509_STORE_CTX structure.
7937 Command line options added to 'verify' application to support this.
7939 This needs some additional work, such as being able to handle multiple
7940 CRLs with different times, extension based lookup (rather than just
7941 by subject name) and ultimately more complete V2 CRL extension
7945 *) Add a general user interface API (crypto/ui/). This is designed
7946 to replace things like des_read_password and friends (backward
7947 compatibility functions using this new API are provided).
7948 The purpose is to remove prompting functions from the DES code
7949 section as well as provide for prompting through dialog boxes in
7950 a window system and the like.
7953 *) Add "ex_data" support to ENGINE so implementations can add state at a
7954 per-structure level rather than having to store it globally.
7957 *) Make it possible for ENGINE structures to be copied when retrieved by
7958 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7959 This causes the "original" ENGINE structure to act like a template,
7960 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7961 operational state can be localised to each ENGINE structure, despite the
7962 fact they all share the same "methods". New ENGINE structures returned in
7963 this case have no functional references and the return value is the single
7964 structural reference. This matches the single structural reference returned
7965 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7969 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7970 needs to match any other type at all we need to manually clear the
7974 *) Changes to the "openssl engine" utility to include;
7975 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7976 about an ENGINE's available control commands.
7977 - executing control commands from command line arguments using the
7978 '-pre' and '-post' switches. '-post' is only used if '-t' is
7979 specified and the ENGINE is successfully initialised. The syntax for
7980 the individual commands are colon-separated, for example;
7981 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7984 *) New dynamic control command support for ENGINEs. ENGINEs can now
7985 declare their own commands (numbers), names (strings), descriptions,
7986 and input types for run-time discovery by calling applications. A
7987 subset of these commands are implicitly classed as "executable"
7988 depending on their input type, and only these can be invoked through
7989 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7990 can be based on user input, config files, etc). The distinction is
7991 that "executable" commands cannot return anything other than a boolean
7992 result and can only support numeric or string input, whereas some
7993 discoverable commands may only be for direct use through
7994 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7995 pointers, or other custom uses. The "executable" commands are to
7996 support parameterisations of ENGINE behaviour that can be
7997 unambiguously defined by ENGINEs and used consistently across any
7998 OpenSSL-based application. Commands have been added to all the
7999 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
8000 control over shared-library paths without source code alterations.
8003 *) Changed all ENGINE implementations to dynamically allocate their
8004 ENGINEs rather than declaring them statically. Apart from this being
8005 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
8006 this also allows the implementations to compile without using the
8007 internal engine_int.h header.
8010 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8011 'const' value. Any code that should be able to modify a RAND_METHOD
8012 should already have non-const pointers to it (ie. they should only
8013 modify their own ones).
8016 *) Made a variety of little tweaks to the ENGINE code.
8017 - "atalla" and "ubsec" string definitions were moved from header files
8018 to C code. "nuron" string definitions were placed in variables
8019 rather than hard-coded - allowing parameterisation of these values
8020 later on via ctrl() commands.
8021 - Removed unused "#if 0"'d code.
8022 - Fixed engine list iteration code so it uses ENGINE_free() to release
8023 structural references.
8024 - Constified the RAND_METHOD element of ENGINE structures.
8025 - Constified various get/set functions as appropriate and added
8026 missing functions (including a catch-all ENGINE_cpy that duplicates
8027 all ENGINE values onto a new ENGINE except reference counts/state).
8028 - Removed NULL parameter checks in get/set functions. Setting a method
8029 or function to NULL is a way of cancelling out a previously set
8030 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8031 and doesn't justify the extra error symbols and code.
8032 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8033 flags from engine_int.h to engine.h.
8034 - Changed prototypes for ENGINE handler functions (init(), finish(),
8035 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8038 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8039 to the algorithm using long division. The binary algorithm can be
8040 used only if the modulus is odd. On 32-bit systems, it is faster
8041 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8042 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8043 up to 450 bits. In 64-bit environments, the binary algorithm
8044 appears to be advantageous for much longer moduli; here we use it
8045 for moduli up to 2048 bits.
8048 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8049 could not support the combine flag in choice fields.
8052 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8053 extensions from a certificate request to the certificate.
8056 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8057 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8058 file: this allows the display of the certificate about to be
8059 signed to be customised, to allow certain fields to be included
8060 or excluded and extension details. The old system didn't display
8061 multicharacter strings properly, omitted fields not in the policy
8062 and couldn't display additional details such as extensions.
8065 *) Function EC_POINTs_mul for multiple scalar multiplication
8066 of an arbitrary number of elliptic curve points
8067 \sum scalars[i]*points[i],
8068 optionally including the generator defined for the EC_GROUP:
8069 scalar*generator + \sum scalars[i]*points[i].
8071 EC_POINT_mul is a simple wrapper function for the typical case
8072 that the point list has just one item (besides the optional
8076 *) First EC_METHODs for curves over GF(p):
8078 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8079 operations and provides various method functions that can also
8080 operate with faster implementations of modular arithmetic.
8082 EC_GFp_mont_method() reuses most functions that are part of
8083 EC_GFp_simple_method, but uses Montgomery arithmetic.
8085 [Bodo Moeller; point addition and point doubling
8086 implementation directly derived from source code provided by
8087 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8089 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8090 crypto/ec/ec_lib.c):
8092 Curves are EC_GROUP objects (with an optional group generator)
8093 based on EC_METHODs that are built into the library.
8095 Points are EC_POINT objects based on EC_GROUP objects.
8097 Most of the framework would be able to handle curves over arbitrary
8098 finite fields, but as there are no obvious types for fields other
8099 than GF(p), some functions are limited to that for now.
8102 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8103 that the file contains a complete HTTP response.
8106 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8107 change the def and num file printf format specifier from "%-40sXXX"
8108 to "%-39s XXX". The latter will always guarantee a space after the
8109 field while the former will cause them to run together if the field
8110 is 40 of more characters long.
8113 *) Constify the cipher and digest 'method' functions and structures
8114 and modify related functions to take constant EVP_MD and EVP_CIPHER
8118 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8119 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8122 *) Modify EVP_Digest*() routines so they now return values. Although the
8123 internal software routines can never fail additional hardware versions
8127 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8129 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8130 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8133 ERR_R_NESTED_ASN1_ERROR
8135 ERR_R_MISSING_ASN1_EOS
8136 were 4 .. 9, conflicting with
8137 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8139 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8140 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8142 Add new error code 'ERR_R_INTERNAL_ERROR'.
8145 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8149 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8150 sets the subject name for a new request or supersedes the
8151 subject name in a given request. Formats that can be parsed are
8152 'CN=Some Name, OU=myOU, C=IT'
8154 'CN=Some Name/OU=myOU/C=IT'.
8156 Add options '-batch' and '-verbose' to 'openssl req'.
8157 [Massimiliano Pala <madwolf@hackmasters.net>]
8159 *) Introduce the possibility to access global variables through
8160 functions on platform were that's the best way to handle exporting
8161 global variables in shared libraries. To enable this functionality,
8162 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8163 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8164 is normally done by Configure or something similar).
8166 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8167 in the source file (foo.c) like this:
8169 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8170 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8172 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8173 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8175 OPENSSL_DECLARE_GLOBAL(int,foo);
8176 #define foo OPENSSL_GLOBAL_REF(foo)
8177 OPENSSL_DECLARE_GLOBAL(double,bar);
8178 #define bar OPENSSL_GLOBAL_REF(bar)
8180 The #defines are very important, and therefore so is including the
8181 header file everywhere where the defined globals are used.
8183 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8184 of ASN.1 items, but that structure is a bit different.
8186 The largest change is in util/mkdef.pl which has been enhanced with
8187 better and easier to understand logic to choose which symbols should
8188 go into the Windows .def files as well as a number of fixes and code
8189 cleanup (among others, algorithm keywords are now sorted
8190 lexicographically to avoid constant rewrites).
8193 *) In BN_div() keep a copy of the sign of 'num' before writing the
8194 result to 'rm' because if rm==num the value will be overwritten
8195 and produce the wrong result if 'num' is negative: this caused
8196 problems with BN_mod() and BN_nnmod().
8199 *) Function OCSP_request_verify(). This checks the signature on an
8200 OCSP request and verifies the signer certificate. The signer
8201 certificate is just checked for a generic purpose and OCSP request
8205 *) Add OCSP_check_validity() function to check the validity of OCSP
8206 responses. OCSP responses are prepared in real time and may only
8207 be a few seconds old. Simply checking that the current time lies
8208 between thisUpdate and nextUpdate max reject otherwise valid responses
8209 caused by either OCSP responder or client clock inaccuracy. Instead
8210 we allow thisUpdate and nextUpdate to fall within a certain period of
8211 the current time. The age of the response can also optionally be
8212 checked. Two new options -validity_period and -status_age added to
8216 *) If signature or public key algorithm is unrecognized print out its
8217 OID rather that just UNKNOWN.
8220 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8221 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8222 ID to be generated from the issuer certificate alone which can then be
8223 passed to OCSP_id_issuer_cmp().
8226 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8227 ASN1 modules to export functions returning ASN1_ITEM pointers
8228 instead of the ASN1_ITEM structures themselves. This adds several
8229 new macros which allow the underlying ASN1 function/structure to
8230 be accessed transparently. As a result code should not use ASN1_ITEM
8231 references directly (such as &X509_it) but instead use the relevant
8232 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8233 use of the new ASN1 code on platforms where exporting structures
8234 is problematical (for example in shared libraries) but exporting
8235 functions returning pointers to structures is not.
8238 *) Add support for overriding the generation of SSL/TLS session IDs.
8239 These callbacks can be registered either in an SSL_CTX or per SSL.
8240 The purpose of this is to allow applications to control, if they wish,
8241 the arbitrary values chosen for use as session IDs, particularly as it
8242 can be useful for session caching in multiple-server environments. A
8243 command-line switch for testing this (and any client code that wishes
8244 to use such a feature) has been added to "s_server".
8245 [Geoff Thorpe, Lutz Jaenicke]
8247 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8248 of the form '#if defined(...) || defined(...) || ...' and
8249 '#if !defined(...) && !defined(...) && ...'. This also avoids
8250 the growing number of special cases it was previously handling.
8253 *) Make all configuration macros available for application by making
8254 sure they are available in opensslconf.h, by giving them names starting
8255 with "OPENSSL_" to avoid conflicts with other packages and by making
8256 sure e_os2.h will cover all platform-specific cases together with
8258 Additionally, it is now possible to define configuration/platform-
8259 specific names (called "system identities"). In the C code, these
8260 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8261 macro with the name beginning with "OPENSSL_SYS_", which is determined
8262 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8266 *) New option -set_serial to 'req' and 'x509' this allows the serial
8267 number to use to be specified on the command line. Previously self
8268 signed certificates were hard coded with serial number 0 and the
8269 CA options of 'x509' had to use a serial number in a file which was
8273 *) New options to 'ca' utility to support V2 CRL entry extensions.
8274 Currently CRL reason, invalidity date and hold instruction are
8275 supported. Add new CRL extensions to V3 code and some new objects.
8278 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8279 disable standard block padding (aka PKCS#5 padding) in the EVP
8280 API, which was previously mandatory. This means that the data is
8281 not padded in any way and so the total length much be a multiple
8282 of the block size, otherwise an error occurs.
8285 *) Initial (incomplete) OCSP SSL support.
8288 *) New function OCSP_parse_url(). This splits up a URL into its host,
8289 port and path components: primarily to parse OCSP URLs. New -url
8290 option to ocsp utility.
8293 *) New nonce behavior. The return value of OCSP_check_nonce() now
8294 reflects the various checks performed. Applications can decide
8295 whether to tolerate certain situations such as an absent nonce
8296 in a response when one was present in a request: the ocsp application
8297 just prints out a warning. New function OCSP_add1_basic_nonce()
8298 this is to allow responders to include a nonce in a response even if
8299 the request is nonce-less.
8302 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8303 skipped when using openssl x509 multiple times on a single input file,
8304 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8307 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8308 set string type: to handle setting ASN1_TIME structures. Fix ca
8309 utility to correctly initialize revocation date of CRLs.
8312 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8313 the clients preferred ciphersuites and rather use its own preferences.
8314 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8315 Internet Explorer by ensuring unchanged hash method during stepup.
8316 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8319 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8320 to aes and add a new 'exist' option to print out symbols that don't
8324 *) Additional options to ocsp utility to allow flags to be set and
8325 additional certificates supplied.
8328 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8329 OCSP client a number of certificate to only verify the response
8333 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8334 handle the new API. Currently only ECB, CBC modes supported. Add new
8337 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8338 Encryption Standard (AES) Ciphersuites for Transport Layer
8339 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8340 not enabled by default and were not part of the "ALL" ciphersuite
8341 alias because they were not yet official; they could be
8342 explicitly requested by specifying the "AESdraft" ciphersuite
8343 group alias. In the final release of OpenSSL 0.9.7, the group
8344 alias is called "AES" and is part of "ALL".)
8345 [Ben Laurie, Steve Henson, Bodo Moeller]
8347 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8348 request to response.
8351 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8352 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8353 extract information from a certificate request. OCSP_response_create()
8354 creates a response and optionally adds a basic response structure.
8355 OCSP_basic_add1_status() adds a complete single response to a basic
8356 response and returns the OCSP_SINGLERESP structure just added (to allow
8357 extensions to be included for example). OCSP_basic_add1_cert() adds a
8358 certificate to a basic response and OCSP_basic_sign() signs a basic
8359 response with various flags. New helper functions ASN1_TIME_check()
8360 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8361 (converts ASN1_TIME to GeneralizedTime).
8364 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8365 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8366 structure from a certificate. X509_pubkey_digest() digests the public_key
8367 contents: this is used in various key identifiers.
8370 *) Make sk_sort() tolerate a NULL argument.
8371 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8373 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8374 passed by the function are trusted implicitly. If any of them signed the
8375 response then it is assumed to be valid and is not verified.
8378 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8379 to data. This was previously part of the PKCS7 ASN1 code. This
8380 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8381 [Steve Henson, reported by Kenneth R. Robinette
8382 <support@securenetterm.com>]
8384 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8385 routines: without these tracing memory leaks is very painful.
8386 Fix leaks in PKCS12 and PKCS7 routines.
8389 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8390 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8391 effectively meant GeneralizedTime would never be used. Now it
8392 is initialised to -1 but X509_time_adj() now has to check the value
8393 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8394 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8395 [Steve Henson, reported by Kenneth R. Robinette
8396 <support@securenetterm.com>]
8398 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8399 result in a zero length in the ASN1_INTEGER structure which was
8400 not consistent with the structure when d2i_ASN1_INTEGER() was used
8401 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8402 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8403 where it did not print out a minus for negative ASN1_INTEGER.
8406 *) Add summary printout to ocsp utility. The various functions which
8407 convert status values to strings have been renamed to:
8408 OCSP_response_status_str(), OCSP_cert_status_str() and
8409 OCSP_crl_reason_str() and are no longer static. New options
8410 to verify nonce values and to disable verification. OCSP response
8411 printout format cleaned up.
8414 *) Add additional OCSP certificate checks. These are those specified
8415 in RFC2560. This consists of two separate checks: the CA of the
8416 certificate being checked must either be the OCSP signer certificate
8417 or the issuer of the OCSP signer certificate. In the latter case the
8418 OCSP signer certificate must contain the OCSP signing extended key
8419 usage. This check is performed by attempting to match the OCSP
8420 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8421 in the OCSP_CERTID structures of the response.
8424 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8425 and related routines. This uses the standard OpenSSL certificate
8426 verify routines to perform initial checks (just CA validity) and
8427 to obtain the certificate chain. Then additional checks will be
8428 performed on the chain. Currently the root CA is checked to see
8429 if it is explicitly trusted for OCSP signing. This is used to set
8430 a root CA as a global signing root: that is any certificate that
8431 chains to that CA is an acceptable OCSP signing certificate.
8434 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8435 extensions from a separate configuration file.
8436 As when reading extensions from the main configuration file,
8437 the '-extensions ...' option may be used for specifying the
8439 [Massimiliano Pala <madwolf@comune.modena.it>]
8441 *) New OCSP utility. Allows OCSP requests to be generated or
8442 read. The request can be sent to a responder and the output
8443 parsed, outputted or printed in text form. Not complete yet:
8444 still needs to check the OCSP response validity.
8447 *) New subcommands for 'openssl ca':
8448 'openssl ca -status <serial>' prints the status of the cert with
8449 the given serial number (according to the index file).
8450 'openssl ca -updatedb' updates the expiry status of certificates
8452 [Massimiliano Pala <madwolf@comune.modena.it>]
8454 *) New '-newreq-nodes' command option to CA.pl. This is like
8455 '-newreq', but calls 'openssl req' with the '-nodes' option
8456 so that the resulting key is not encrypted.
8457 [Damien Miller <djm@mindrot.org>]
8459 *) New configuration for the GNU Hurd.
8460 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8462 *) Initial code to implement OCSP basic response verify. This
8463 is currently incomplete. Currently just finds the signer's
8464 certificate and verifies the signature on the response.
8467 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8468 value of OPENSSLDIR. This is available via the new '-d' option
8469 to 'openssl version', and is also included in 'openssl version -a'.
8472 *) Allowing defining memory allocation callbacks that will be given
8473 file name and line number information in additional arguments
8474 (a const char* and an int). The basic functionality remains, as
8475 well as the original possibility to just replace malloc(),
8476 realloc() and free() by functions that do not know about these
8477 additional arguments. To register and find out the current
8478 settings for extended allocation functions, the following
8479 functions are provided:
8481 CRYPTO_set_mem_ex_functions
8482 CRYPTO_set_locked_mem_ex_functions
8483 CRYPTO_get_mem_ex_functions
8484 CRYPTO_get_locked_mem_ex_functions
8486 These work the same way as CRYPTO_set_mem_functions and friends.
8487 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8488 extended allocation function is enabled.
8489 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8490 a conventional allocation function is enabled.
8491 [Richard Levitte, Bodo Moeller]
8493 *) Finish off removing the remaining LHASH function pointer casts.
8494 There should no longer be any prototype-casting required when using
8495 the LHASH abstraction, and any casts that remain are "bugs". See
8496 the callback types and macros at the head of lhash.h for details
8497 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8500 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8501 If /dev/[u]random devices are not available or do not return enough
8502 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8504 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8505 /etc/entropy will be queried once each in this sequence, querying stops
8506 when enough entropy was collected without querying more sockets.
8509 *) Change the Unix RAND_poll() variant to be able to poll several
8510 random devices, as specified by DEVRANDOM, until a sufficient amount
8511 of data has been collected. We spend at most 10 ms on each file
8512 (select timeout) and read in non-blocking mode. DEVRANDOM now
8513 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8514 (previously it was just the string "/dev/urandom"), so on typical
8515 platforms the 10 ms delay will never occur.
8516 Also separate out the Unix variant to its own file, rand_unix.c.
8517 For VMS, there's a currently-empty rand_vms.c.
8520 *) Move OCSP client related routines to ocsp_cl.c. These
8521 provide utility functions which an application needing
8522 to issue a request to an OCSP responder and analyse the
8523 response will typically need: as opposed to those which an
8524 OCSP responder itself would need which will be added later.
8526 OCSP_request_sign() signs an OCSP request with an API similar
8527 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8528 response. OCSP_response_get1_basic() extracts basic response
8529 from response. OCSP_resp_find_status(): finds and extracts status
8530 information from an OCSP_CERTID structure (which will be created
8531 when the request structure is built). These are built from lower
8532 level functions which work on OCSP_SINGLERESP structures but
8533 won't normally be used unless the application wishes to examine
8534 extensions in the OCSP response for example.
8536 Replace nonce routines with a pair of functions.
8537 OCSP_request_add1_nonce() adds a nonce value and optionally
8538 generates a random value. OCSP_check_nonce() checks the
8539 validity of the nonce in an OCSP response.
8542 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8543 This doesn't copy the supplied OCSP_CERTID and avoids the
8544 need to free up the newly created id. Change return type
8545 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8546 This can then be used to add extensions to the request.
8547 Deleted OCSP_request_new(), since most of its functionality
8548 is now in OCSP_REQUEST_new() (and the case insensitive name
8549 clash) apart from the ability to set the request name which
8550 will be added elsewhere.
8553 *) Update OCSP API. Remove obsolete extensions argument from
8554 various functions. Extensions are now handled using the new
8555 OCSP extension code. New simple OCSP HTTP function which
8556 can be used to send requests and parse the response.
8559 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8560 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8561 uses the special reorder version of SET OF to sort the attributes
8562 and reorder them to match the encoded order. This resolves a long
8563 standing problem: a verify on a PKCS7 structure just after signing
8564 it used to fail because the attribute order did not match the
8565 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8566 it uses the received order. This is necessary to tolerate some broken
8567 software that does not order SET OF. This is handled by encoding
8568 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8569 to produce the required SET OF.
8572 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8573 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8574 files to get correct declarations of the ASN.1 item variables.
8577 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8578 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8579 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8580 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8581 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8582 ASN1_ITEM and no wrapper functions.
8585 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8586 replace the old function pointer based I/O routines. Change most of
8587 the *_d2i_bio() and *_d2i_fp() functions to use these.
8590 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8591 lines, recognize more "algorithms" that can be deselected, and make
8592 it complain about algorithm deselection that isn't recognised.
8595 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8596 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8597 to use new functions. Add NO_ASN1_OLD which can be set to remove
8598 some old style ASN1 functions: this can be used to determine if old
8599 code will still work when these eventually go away.
8602 *) New extension functions for OCSP structures, these follow the
8603 same conventions as certificates and CRLs.
8606 *) New function X509V3_add1_i2d(). This automatically encodes and
8607 adds an extension. Its behaviour can be customised with various
8608 flags to append, replace or delete. Various wrappers added for
8609 certificates and CRLs.
8612 *) Fix to avoid calling the underlying ASN1 print routine when
8613 an extension cannot be parsed. Correct a typo in the
8614 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8617 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8618 entries for variables.
8621 *) Add functionality to apps/openssl.c for detecting locking
8622 problems: As the program is single-threaded, all we have
8623 to do is register a locking callback using an array for
8624 storing which locks are currently held by the program.
8627 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8628 SSL_get_ex_data_X509_STORE_idx(), which is used in
8629 ssl_verify_cert_chain() and thus can be called at any time
8630 during TLS/SSL handshakes so that thread-safety is essential.
8631 Unfortunately, the ex_data design is not at all suited
8632 for multi-threaded use, so it probably should be abolished.
8635 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8636 [Broadcom, tweaked and integrated by Geoff Thorpe]
8638 *) Move common extension printing code to new function
8639 X509V3_print_extensions(). Reorganise OCSP print routines and
8640 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8643 *) New function X509_signature_print() to remove duplication in some
8647 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8648 set (this was treated exactly the same as SET OF previously). This
8649 is used to reorder the STACK representing the structure to match the
8650 encoding. This will be used to get round a problem where a PKCS7
8651 structure which was signed could not be verified because the STACK
8652 order did not reflect the encoded order.
8655 *) Reimplement the OCSP ASN1 module using the new code.
8658 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8659 for its ASN1 operations. The old style function pointers still exist
8660 for now but they will eventually go away.
8663 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8664 completely replaces the old ASN1 functionality with a table driven
8665 encoder and decoder which interprets an ASN1_ITEM structure describing
8666 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8667 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8668 has also been converted to the new form.
8671 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8672 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8673 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8674 for negative moduli.
8677 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8678 of not touching the result's sign bit.
8681 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8685 *) Changed the LHASH code to use prototypes for callbacks, and created
8686 macros to declare and implement thin (optionally static) functions
8687 that provide type-safety and avoid function pointer casting for the
8688 type-specific callbacks.
8691 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8693 [Veers Staats <staatsvr@asc.hpc.mil>,
8694 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8696 *) Reformat the FAQ so the different questions and answers can be divided
8697 in sections depending on the subject.
8700 *) Have the zlib compression code load ZLIB.DLL dynamically under
8704 *) New function BN_mod_sqrt for computing square roots modulo a prime
8705 (using the probabilistic Tonelli-Shanks algorithm unless
8706 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8707 be handled deterministically).
8708 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8710 *) Make BN_mod_inverse faster by explicitly handling small quotients
8711 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8712 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8715 *) New function BN_kronecker.
8718 *) Fix BN_gcd so that it works on negative inputs; the result is
8719 positive unless both parameters are zero.
8720 Previously something reasonably close to an infinite loop was
8721 possible because numbers could be growing instead of shrinking
8722 in the implementation of Euclid's algorithm.
8725 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8726 sign of the number in question.
8728 Fix BN_is_word(a,w) to work correctly for w == 0.
8730 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8731 because its test if the absolute value of 'a' equals 'w'.
8732 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8733 it exists mostly for use in the implementations of BN_is_zero(),
8734 BN_is_one(), and BN_is_word().
8737 *) New function BN_swap.
8740 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8741 the exponentiation functions are more likely to produce reasonable
8742 results on negative inputs.
8745 *) Change BN_mod_mul so that the result is always non-negative.
8746 Previously, it could be negative if one of the factors was negative;
8747 I don't think anyone really wanted that behaviour.
8750 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8751 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8752 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8753 and add new functions:
8762 BN_mod_lshift1_quick
8766 These functions always generate non-negative results.
8768 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8769 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8771 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8772 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8773 be reduced modulo m.
8774 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8777 The following entry accidentally appeared in the CHANGES file
8778 distributed with OpenSSL 0.9.7. The modifications described in
8779 it do *not* apply to OpenSSL 0.9.7.
8781 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8782 was actually never needed) and in BN_mul(). The removal in BN_mul()
8783 required a small change in bn_mul_part_recursive() and the addition
8784 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8785 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8786 bn_sub_words() and bn_add_words() except they take arrays with
8791 *) In 'openssl passwd', verify passwords read from the terminal
8792 unless the '-salt' option is used (which usually means that
8793 verification would just waste user's time since the resulting
8794 hash is going to be compared with some given password hash)
8795 or the new '-noverify' option is used.
8797 This is an incompatible change, but it does not affect
8798 non-interactive use of 'openssl passwd' (passwords on the command
8799 line, '-stdin' option, '-in ...' option) and thus should not
8803 *) Remove all references to RSAref, since there's no more need for it.
8806 *) Make DSO load along a path given through an environment variable
8807 (SHLIB_PATH) with shl_load().
8810 *) Constify the ENGINE code as a result of BIGNUM constification.
8811 Also constify the RSA code and most things related to it. In a
8812 few places, most notable in the depth of the ASN.1 code, ugly
8813 casts back to non-const were required (to be solved at a later
8817 *) Make it so the openssl application has all engines loaded by default.
8820 *) Constify the BIGNUM routines a little more.
8823 *) Add the following functions:
8825 ENGINE_load_cswift()
8827 ENGINE_load_atalla()
8829 ENGINE_load_builtin_engines()
8831 That way, an application can itself choose if external engines that
8832 are built-in in OpenSSL shall ever be used or not. The benefit is
8833 that applications won't have to be linked with libdl or other dso
8834 libraries unless it's really needed.
8836 Changed 'openssl engine' to load all engines on demand.
8837 Changed the engine header files to avoid the duplication of some
8838 declarations (they differed!).
8841 *) 'openssl engine' can now list capabilities.
8844 *) Better error reporting in 'openssl engine'.
8847 *) Never call load_dh_param(NULL) in s_server.
8850 *) Add engine application. It can currently list engines by name and
8851 identity, and test if they are actually available.
8854 *) Improve RPM specification file by forcing symbolic linking and making
8855 sure the installed documentation is also owned by root.root.
8856 [Damien Miller <djm@mindrot.org>]
8858 *) Give the OpenSSL applications more possibilities to make use of
8859 keys (public as well as private) handled by engines.
8862 *) Add OCSP code that comes from CertCo.
8865 *) Add VMS support for the Rijndael code.
8868 *) Added untested support for Nuron crypto accelerator.
8871 *) Add support for external cryptographic devices. This code was
8872 previously distributed separately as the "engine" branch.
8873 [Geoff Thorpe, Richard Levitte]
8875 *) Rework the filename-translation in the DSO code. It is now possible to
8876 have far greater control over how a "name" is turned into a filename
8877 depending on the operating environment and any oddities about the
8878 different shared library filenames on each system.
8881 *) Support threads on FreeBSD-elf in Configure.
8884 *) Fix for SHA1 assembly problem with MASM: it produces
8885 warnings about corrupt line number information when assembling
8886 with debugging information. This is caused by the overlapping
8888 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8891 NCONF_get_number() has no error checking at all. As a replacement,
8892 NCONF_get_number_e() is defined (_e for "error checking") and is
8893 promoted strongly. The old NCONF_get_number is kept around for
8894 binary backward compatibility.
8895 Make it possible for methods to load from something other than a BIO,
8896 by providing a function pointer that is given a name instead of a BIO.
8897 For example, this could be used to load configuration data from an
8901 *) Fix for non blocking accept BIOs. Added new I/O special reason
8902 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8903 with non blocking I/O was not possible because no retry code was
8904 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8908 *) Added the beginnings of Rijndael support.
8911 *) Fix for bug in DirectoryString mask setting. Add support for
8912 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8913 to allow certificate printing to more controllable, additional
8914 'certopt' option to 'x509' to allow new printing options to be
8918 *) Clean old EAY MD5 hack from e_os.h.
8921 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8923 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8924 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8925 [Joe Orton, Steve Henson]
8927 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8929 *) Fix additional bug revealed by the NISCC test suite:
8931 Stop bug triggering large recursion when presented with
8932 certain ASN.1 tags (CVE-2003-0851)
8935 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8937 *) Fix various bugs revealed by running the NISCC test suite:
8939 Stop out of bounds reads in the ASN1 code when presented with
8940 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8942 If verify callback ignores invalid public key errors don't try to check
8943 certificate signature with the NULL public key.
8947 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8948 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8952 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8953 extra data after the compression methods not only for TLS 1.0
8954 but also for SSL 3.0 (as required by the specification).
8955 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8957 *) Change X509_certificate_type() to mark the key as exported/exportable
8958 when it's 512 *bits* long, not 512 bytes.
8961 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8963 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8964 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8965 a protocol version number mismatch like a decryption error
8966 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8969 *) Turn on RSA blinding by default in the default implementation
8970 to avoid a timing attack. Applications that don't want it can call
8971 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8972 They would be ill-advised to do so in most cases.
8973 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8975 *) Change RSA blinding code so that it works when the PRNG is not
8976 seeded (in this case, the secret RSA exponent is abused as
8977 an unpredictable seed -- if it is not unpredictable, there
8978 is no point in blinding anyway). Make RSA blinding thread-safe
8979 by remembering the creator's thread ID in rsa->blinding and
8980 having all other threads use local one-time blinding factors
8981 (this requires more computation than sharing rsa->blinding, but
8982 avoids excessive locking; and if an RSA object is not shared
8983 between threads, blinding will still be very fast).
8986 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8988 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8989 via timing by performing a MAC computation even if incorrect
8990 block cipher padding has been found. This is a countermeasure
8991 against active attacks where the attacker has to distinguish
8992 between bad padding and a MAC verification error. (CVE-2003-0078)
8994 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8995 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8996 Martin Vuagnoux (EPFL, Ilion)]
8998 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
9000 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
9001 memory from its contents. This is done with a counter that will
9002 place alternating values in each byte. This can be used to solve
9003 two issues: 1) the removal of calls to memset() by highly optimizing
9004 compilers, and 2) cleansing with other values than 0, since those can
9005 be read through on certain media, for example a swap space on disk.
9008 *) Bugfix: client side session caching did not work with external caching,
9009 because the session->cipher setting was not restored when reloading
9010 from the external cache. This problem was masked, when
9011 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9012 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9015 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9016 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9017 [Zeev Lieber <zeev-l@yahoo.com>]
9019 *) Undo an undocumented change introduced in 0.9.6e which caused
9020 repeated calls to OpenSSL_add_all_ciphers() and
9021 OpenSSL_add_all_digests() to be ignored, even after calling
9025 *) Change the default configuration reader to deal with last line not
9026 being properly terminated.
9029 *) Change X509_NAME_cmp() so it applies the special rules on handling
9030 DN values that are of type PrintableString, as well as RDNs of type
9031 emailAddress where the value has the type ia5String.
9032 [stefank@valicert.com via Richard Levitte]
9034 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9035 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9036 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9037 the bitwise-OR of the two for use by the majority of applications
9038 wanting this behaviour, and update the docs. The documented
9039 behaviour and actual behaviour were inconsistent and had been
9040 changing anyway, so this is more a bug-fix than a behavioural
9042 [Geoff Thorpe, diagnosed by Nadav Har'El]
9044 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9045 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9048 *) Fix initialization code race conditions in
9049 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9050 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9051 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9052 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9053 ssl2_get_cipher_by_char(),
9054 ssl3_get_cipher_by_char().
9055 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9057 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9058 the cached sessions are flushed, as the remove_cb() might use ex_data
9059 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9060 (see [openssl.org #212]).
9061 [Geoff Thorpe, Lutz Jaenicke]
9063 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9064 length, instead of the encoding length to d2i_ASN1_OBJECT.
9067 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9069 *) [In 0.9.6g-engine release:]
9070 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9071 [Lynn Gazis <lgazis@rainbow.com>]
9073 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9075 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9076 and get fix the header length calculation.
9077 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9078 Alon Kantor <alonk@checkpoint.com> (and others),
9081 *) Use proper error handling instead of 'assertions' in buffer
9082 overflow checks added in 0.9.6e. This prevents DoS (the
9083 assertions could call abort()).
9084 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9086 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9088 *) Add various sanity checks to asn1_get_length() to reject
9089 the ASN1 length bytes if they exceed sizeof(long), will appear
9090 negative or the content length exceeds the length of the
9092 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9094 *) Fix cipher selection routines: ciphers without encryption had no flags
9095 for the cipher strength set and where therefore not handled correctly
9096 by the selection routines (PR #130).
9099 *) Fix EVP_dsa_sha macro.
9103 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9104 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9105 that was added in OpenSSL 0.9.6d.
9107 As the countermeasure turned out to be incompatible with some
9108 broken SSL implementations, the new option is part of SSL_OP_ALL.
9109 SSL_OP_ALL is usually employed when compatibility with weird SSL
9110 implementations is desired (e.g. '-bugs' option to 's_client' and
9111 's_server'), so the new option is automatically set in many
9115 *) Changes in security patch:
9117 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9118 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9119 Air Force Materiel Command, USAF, under agreement number
9122 *) Add various sanity checks to asn1_get_length() to reject
9123 the ASN1 length bytes if they exceed sizeof(long), will appear
9124 negative or the content length exceeds the length of the
9125 supplied buffer. (CVE-2002-0659)
9126 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9128 *) Assertions for various potential buffer overflows, not known to
9130 [Ben Laurie (CHATS)]
9132 *) Various temporary buffers to hold ASCII versions of integers were
9133 too small for 64 bit platforms. (CVE-2002-0655)
9134 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9136 *) Remote buffer overflow in SSL3 protocol - an attacker could
9137 supply an oversized session ID to a client. (CVE-2002-0656)
9138 [Ben Laurie (CHATS)]
9140 *) Remote buffer overflow in SSL2 protocol - an attacker could
9141 supply an oversized client master key. (CVE-2002-0656)
9142 [Ben Laurie (CHATS)]
9144 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9146 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9147 encoded as NULL) with id-dsa-with-sha1.
9148 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9150 *) Check various X509_...() return values in apps/req.c.
9151 [Nils Larsch <nla@trustcenter.de>]
9153 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9154 an end-of-file condition would erroneously be flagged, when the CRLF
9155 was just at the end of a processed block. The bug was discovered when
9156 processing data through a buffering memory BIO handing the data to a
9157 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9158 <ptsekov@syntrex.com> and Nedelcho Stanev.
9161 *) Implement a countermeasure against a vulnerability recently found
9162 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9163 before application data chunks to avoid the use of known IVs
9164 with data potentially chosen by the attacker.
9167 *) Fix length checks in ssl3_get_client_hello().
9170 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9171 to prevent ssl3_read_internal() from incorrectly assuming that
9172 ssl3_read_bytes() found application data while handshake
9173 processing was enabled when in fact s->s3->in_read_app_data was
9174 merely automatically cleared during the initial handshake.
9175 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9177 *) Fix object definitions for Private and Enterprise: they were not
9178 recognized in their shortname (=lowercase) representation. Extend
9179 obj_dat.pl to issue an error when using undefined keywords instead
9180 of silently ignoring the problem (Svenning Sorensen
9181 <sss@sss.dnsalias.net>).
9184 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9185 generators, i.e. generators other than 2 and 5. (Previously, the
9186 code did not properly initialise the 'add' and 'rem' values to
9187 BN_generate_prime().)
9189 In the new general case, we do not insist that 'generator' is
9190 actually a primitive root: This requirement is rather pointless;
9191 a generator of the order-q subgroup is just as good, if not
9195 *) Map new X509 verification errors to alerts. Discovered and submitted by
9196 Tom Wu <tom@arcot.com>.
9199 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9200 returning non-zero before the data has been completely received
9201 when using non-blocking I/O.
9202 [Bodo Moeller; problem pointed out by John Hughes]
9204 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9205 [Ben Laurie, Lutz Jaenicke]
9207 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9208 Yoram Zahavi <YoramZ@gilian.com>).
9211 *) Add information about CygWin 1.3 and on, and preserve proper
9212 configuration for the versions before that.
9213 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9215 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9216 check whether we deal with a copy of a session and do not delete from
9217 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9218 <izhar@checkpoint.com>.
9221 *) Do not store session data into the internal session cache, if it
9222 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9223 flag is set). Proposed by Aslam <aslam@funk.com>.
9226 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9230 *) [In 0.9.6d-engine release:]
9231 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9232 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9234 *) Add the configuration target linux-s390x.
9235 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9237 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9238 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9239 variable as an indication that a ClientHello message has been
9240 received. As the flag value will be lost between multiple
9241 invocations of ssl3_accept when using non-blocking I/O, the
9242 function may not be aware that a handshake has actually taken
9243 place, thus preventing a new session from being added to the
9246 To avoid this problem, we now set s->new_session to 2 instead of
9247 using a local variable.
9248 [Lutz Jaenicke, Bodo Moeller]
9250 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9251 if the SSL_R_LENGTH_MISMATCH error is detected.
9252 [Geoff Thorpe, Bodo Moeller]
9254 *) New 'shared_ldflag' column in Configure platform table.
9257 *) Fix EVP_CIPHER_mode macro.
9258 ["Dan S. Camper" <dan@bti.net>]
9260 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9261 type, we must throw them away by setting rr->length to 0.
9262 [D P Chang <dpc@qualys.com>]
9264 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9266 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9267 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9268 worked incorrectly for those cases where range = 10..._2 and
9269 3*range is two bits longer than range.)
9272 *) Only add signing time to PKCS7 structures if it is not already
9276 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9277 OBJ_ld_ce should be OBJ_id_ce.
9278 Also some ip-pda OIDs in crypto/objects/objects.txt were
9279 incorrect (cf. RFC 3039).
9280 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9282 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9283 returns early because it has nothing to do.
9284 [Andy Schneider <andy.schneider@bjss.co.uk>]
9286 *) [In 0.9.6c-engine release:]
9287 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9288 [Andy Schneider <andy.schneider@bjss.co.uk>]
9290 *) [In 0.9.6c-engine release:]
9291 Add support for Cryptographic Appliance's keyserver technology.
9292 (Use engine 'keyclient')
9293 [Cryptographic Appliances and Geoff Thorpe]
9295 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9296 is called via tools/c89.sh because arguments have to be
9297 rearranged (all '-L' options must appear before the first object
9299 [Richard Shapiro <rshapiro@abinitio.com>]
9301 *) [In 0.9.6c-engine release:]
9302 Add support for Broadcom crypto accelerator cards, backported
9304 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9306 *) [In 0.9.6c-engine release:]
9307 Add support for SureWare crypto accelerator cards from
9308 Baltimore Technologies. (Use engine 'sureware')
9309 [Baltimore Technologies and Mark Cox]
9311 *) [In 0.9.6c-engine release:]
9312 Add support for crypto accelerator cards from Accelerated
9313 Encryption Processing, www.aep.ie. (Use engine 'aep')
9314 [AEP Inc. and Mark Cox]
9316 *) Add a configuration entry for gcc on UnixWare.
9317 [Gary Benson <gbenson@redhat.com>]
9319 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9320 messages are stored in a single piece (fixed-length part and
9321 variable-length part combined) and fix various bugs found on the way.
9324 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9325 instead. BIO_gethostbyname() does not know what timeouts are
9326 appropriate, so entries would stay in cache even when they have
9328 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9330 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9331 faced with a pathologically small ClientHello fragment that does
9332 not contain client_version: Instead of aborting with an error,
9333 simply choose the highest available protocol version (i.e.,
9334 TLS 1.0 unless it is disabled). In practice, ClientHello
9335 messages are never sent like this, but this change gives us
9336 strictly correct behaviour at least for TLS.
9339 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9340 never resets s->method to s->ctx->method when called from within
9341 one of the SSL handshake functions.
9342 [Bodo Moeller; problem pointed out by Niko Baric]
9344 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9345 (sent using the client's version number) if client_version is
9346 smaller than the protocol version in use. Also change
9347 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9348 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9349 the client will at least see that alert.
9352 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9356 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9357 client receives HelloRequest while in a handshake.
9358 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9360 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9361 should end in 'break', not 'goto end' which circumvents various
9362 cleanups done in state SSL_ST_OK. But session related stuff
9363 must be disabled for SSL_ST_OK in the case that we just sent a
9366 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9367 before just sending a HelloRequest.
9368 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9370 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9371 reveal whether illegal block cipher padding was found or a MAC
9372 verification error occurred. (Neither SSLerr() codes nor alerts
9373 are directly visible to potential attackers, but the information
9374 may leak via logfiles.)
9376 Similar changes are not required for the SSL 2.0 implementation
9377 because the number of padding bytes is sent in clear for SSL 2.0,
9378 and the extra bytes are just ignored. However ssl/s2_pkt.c
9379 failed to verify that the purported number of padding bytes is in
9383 *) Add OpenUNIX-8 support including shared libraries
9384 (Boyd Lynn Gerber <gerberb@zenez.com>).
9387 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9388 'wristwatch attack' using huge encoding parameters (cf.
9389 James H. Manger's CRYPTO 2001 paper). Note that the
9390 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9391 encoding parameters and hence was not vulnerable.
9394 *) BN_sqr() bug fix.
9395 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9397 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9398 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9399 followed by modular reduction.
9400 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9402 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9403 equivalent based on BN_pseudo_rand() instead of BN_rand().
9406 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9407 This function was broken, as the check for a new client hello message
9408 to handle SGC did not allow these large messages.
9409 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9412 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9415 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9416 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9419 *) Rework the configuration and shared library support for Tru64 Unix.
9420 The configuration part makes use of modern compiler features and
9421 still retains old compiler behavior for those that run older versions
9422 of the OS. The shared library support part includes a variant that
9423 uses the RPATH feature, and is available through the special
9424 configuration target "alpha-cc-rpath", which will never be selected
9426 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9428 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9429 with the same message size as in ssl3_get_certificate_request().
9430 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9431 messages might inadvertently be reject as too long.
9432 [Petr Lampa <lampa@fee.vutbr.cz>]
9434 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9437 *) Modified SSL library such that the verify_callback that has been set
9438 specifically for an SSL object with SSL_set_verify() is actually being
9439 used. Before the change, a verify_callback set with this function was
9440 ignored and the verify_callback() set in the SSL_CTX at the time of
9441 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9442 to allow the necessary settings.
9445 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9446 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9447 done automatically (in contradiction to the requirements of the C
9448 standard). This made problems when used from OpenSSH.
9451 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9452 dh->length and always used
9454 BN_rand_range(priv_key, dh->p).
9456 BN_rand_range() is not necessary for Diffie-Hellman, and this
9457 specific range makes Diffie-Hellman unnecessarily inefficient if
9458 dh->length (recommended exponent length) is much smaller than the
9459 length of dh->p. We could use BN_rand_range() if the order of
9460 the subgroup was stored in the DH structure, but we only have
9465 BN_rand(priv_key, l, ...)
9467 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9473 RSA_eay_public_encrypt
9474 RSA_eay_private_decrypt
9475 RSA_eay_private_encrypt (signing)
9476 RSA_eay_public_decrypt (signature verification)
9478 (default implementations for RSA_public_encrypt,
9479 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9480 always reject numbers >= n.
9483 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9484 to synchronize access to 'locking_thread'. This is necessary on
9485 systems where access to 'locking_thread' (an 'unsigned long'
9486 variable) is not atomic.
9489 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9490 *before* setting the 'crypto_lock_rand' flag. The previous code had
9491 a race condition if 0 is a valid thread ID.
9492 [Travis Vitek <vitek@roguewave.com>]
9494 *) Add support for shared libraries under Irix.
9495 [Albert Chin-A-Young <china@thewrittenword.com>]
9497 *) Add configuration option to build on Linux on both big-endian and
9499 [Ralf Baechle <ralf@uni-koblenz.de>]
9501 *) Add the possibility to create shared libraries on HP-UX.
9504 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9506 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9507 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9508 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9509 PRNG state recovery was possible based on the output of
9510 one PRNG request appropriately sized to gain knowledge on
9511 'md' followed by enough consecutive 1-byte PRNG requests
9512 to traverse all of 'state'.
9514 1. When updating 'md_local' (the current thread's copy of 'md')
9515 during PRNG output generation, hash all of the previous
9516 'md_local' value, not just the half used for PRNG output.
9518 2. Make the number of bytes from 'state' included into the hash
9519 independent from the number of PRNG bytes requested.
9521 The first measure alone would be sufficient to avoid
9522 Markku-Juhani's attack. (Actually it had never occurred
9523 to me that the half of 'md_local' used for chaining was the
9524 half from which PRNG output bytes were taken -- I had always
9525 assumed that the secret half would be used.) The second
9526 measure makes sure that additional data from 'state' is never
9527 mixed into 'md_local' in small portions; this heuristically
9528 further strengthens the PRNG.
9531 *) Fix crypto/bn/asm/mips3.s.
9534 *) When only the key is given to "enc", the IV is undefined. Print out
9535 an error message in this case.
9538 *) Handle special case when X509_NAME is empty in X509 printing routines.
9541 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9542 positive and less than q.
9545 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9546 used: it isn't thread safe and the add_lock_callback should handle
9548 [Paul Rose <Paul.Rose@bridge.com>]
9550 *) Verify that incoming data obeys the block size in
9551 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9555 [Ulf Möller, Bodo Möller]
9557 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9558 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9559 when fixing the server behaviour for backwards-compatible 'client
9560 hello' messages. (Note that the attack is impractical against
9561 SSL 3.0 and TLS 1.0 anyway because length and version checking
9562 means that the probability of guessing a valid ciphertext is
9563 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9566 Before 0.9.5, the countermeasure (hide the error by generating a
9567 random 'decryption result') did not work properly because
9568 ERR_clear_error() was missing, meaning that SSL_get_error() would
9569 detect the supposedly ignored error.
9571 Both problems are now fixed.
9574 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9575 (previously it was 1024).
9578 *) Fix for compatibility mode trust settings: ignore trust settings
9579 unless some valid trust or reject settings are present.
9582 *) Fix for blowfish EVP: its a variable length cipher.
9585 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9586 parameters in DSA public key structures and return an error in the
9587 DSA routines if parameters are absent.
9590 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9591 in the current directory if neither $RANDFILE nor $HOME was set.
9592 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9593 caused some confusion to Windows users who haven't defined $HOME.
9594 Thus RAND_file_name() is changed again: e_os.h can define a
9595 DEFAULT_HOME, which will be used if $HOME is not set.
9596 For Windows, we use "C:"; on other platforms, we still require
9597 environment variables.
9599 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9600 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9601 having multiple threads call RAND_poll() concurrently.
9604 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9605 combination of a flag and a thread ID variable.
9606 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9607 flag), *other* threads can enter ssleay_add_bytes without obeying
9608 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9609 that they do not hold after the first thread unsets add_do_not_lock).
9612 *) Change bctest again: '-x' expressions are not available in all
9616 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9618 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9619 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9621 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9622 the default extension for executables, if any. Also, make the perl
9623 scripts that use symlink() to test if it really exists and use "cp"
9624 if it doesn't. All this made OpenSSL compilable and installable in
9628 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9629 If SEQUENCE is length is indefinite just set c->slen to the total
9630 amount of data available.
9631 [Steve Henson, reported by shige@FreeBSD.org]
9632 [This change does not apply to 0.9.7.]
9634 *) Change bctest to avoid here-documents inside command substitution
9635 (workaround for FreeBSD /bin/sh bug).
9636 For compatibility with Ultrix, avoid shell functions (introduced
9637 in the bctest version that searches along $PATH).
9640 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9641 with des_encrypt() defined on some operating systems, like Solaris
9645 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9646 On the Importance of Eliminating Errors in Cryptographic
9647 Computations, J. Cryptology 14 (2001) 2, 101-119,
9648 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9651 *) MIPS assembler BIGNUM division bug fix.
9654 *) Disabled incorrect Alpha assembler code.
9657 *) Fix PKCS#7 decode routines so they correctly update the length
9658 after reading an EOC for the EXPLICIT tag.
9660 [This change does not apply to 0.9.7.]
9662 *) Fix bug in PKCS#12 key generation routines. This was triggered
9663 if a 3DES key was generated with a 0 initial byte. Include
9664 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9665 (but broken) behaviour.
9668 *) Enhance bctest to search for a working bc along $PATH and print
9670 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9672 *) Fix memory leaks in err.c: free err_data string if necessary;
9673 don't write to the wrong index in ERR_set_error_data.
9676 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9680 *) Replace rdtsc with _emit statements for VC++ version 5.
9681 [Jeremy Cooper <jeremy@baymoo.org>]
9683 *) Make it possible to reuse SSLv2 sessions.
9686 *) In copy_email() check for >= 0 as a return value for
9687 X509_NAME_get_index_by_NID() since 0 is a valid index.
9688 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9690 *) Avoid coredump with unsupported or invalid public keys by checking if
9691 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9692 PKCS7_verify() fails with non detached data.
9695 *) Don't use getenv in library functions when run as setuid/setgid.
9696 New function OPENSSL_issetugid().
9699 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9700 due to incorrect handling of multi-threading:
9702 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9704 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9706 3. Count how many times MemCheck_off() has been called so that
9707 nested use can be treated correctly. This also avoids
9708 inband-signalling in the previous code (which relied on the
9709 assumption that thread ID 0 is impossible).
9712 *) Add "-rand" option also to s_client and s_server.
9715 *) Fix CPU detection on Irix 6.x.
9716 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9717 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9719 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9722 [This change does not apply to 0.9.7.]
9724 *) Use the cached encoding of an X509_NAME structure rather than
9725 copying it. This is apparently the reason for the libsafe "errors"
9726 but the code is actually correct.
9729 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9730 Bleichenbacher's DSA attack.
9731 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9732 to be set and top=0 forces the highest bit to be set; top=-1 is new
9733 and leaves the highest bit random.
9734 [Ulf Moeller, Bodo Moeller]
9736 *) In the NCONF_...-based implementations for CONF_... queries
9737 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9738 a temporary CONF structure with the data component set to NULL
9739 (which gives segmentation faults in lh_retrieve).
9740 Instead, use NULL for the CONF pointer in CONF_get_string and
9741 CONF_get_number (which may use environment variables) and directly
9742 return NULL from CONF_get_section.
9745 *) Fix potential buffer overrun for EBCDIC.
9748 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9749 keyUsage if basicConstraints absent for a CA.
9752 *) Make SMIME_write_PKCS7() write mail header values with a format that
9753 is more generally accepted (no spaces before the semicolon), since
9754 some programs can't parse those values properly otherwise. Also make
9755 sure BIO's that break lines after each write do not create invalid
9759 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9760 macros previously used would not encode an empty SEQUENCE OF
9761 and break the signature.
9763 [This change does not apply to 0.9.7.]
9765 *) Zero the premaster secret after deriving the master secret in
9769 *) Add some EVP_add_digest_alias registrations (as found in
9770 OpenSSL_add_all_digests()) to SSL_library_init()
9771 aka OpenSSL_add_ssl_algorithms(). This provides improved
9772 compatibility with peers using X.509 certificates
9773 with unconventional AlgorithmIdentifier OIDs.
9776 *) Fix for Irix with NO_ASM.
9777 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9779 *) ./config script fixes.
9780 [Ulf Moeller, Richard Levitte]
9782 *) Fix 'openssl passwd -1'.
9785 *) Change PKCS12_key_gen_asc() so it can cope with non null
9786 terminated strings whose length is passed in the passlen
9787 parameter, for example from PEM callbacks. This was done
9788 by adding an extra length parameter to asc2uni().
9789 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9791 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9792 call failed, free the DSA structure.
9795 *) Fix to uni2asc() to cope with zero length Unicode strings.
9796 These are present in some PKCS#12 files.
9799 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9800 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9801 when writing a 32767 byte record.
9802 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9804 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9805 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9807 (RSA objects have a reference count access to which is protected
9808 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9809 so they are meant to be shared between threads.)
9810 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9811 "Reddie, Steven" <Steven.Reddie@ca.com>]
9813 *) Fix a deadlock in CRYPTO_mem_leaks().
9816 *) Use better test patterns in bntest.
9819 *) rand_win.c fix for Borland C.
9822 *) BN_rshift bugfix for n == 0.
9825 *) Add a 'bctest' script that checks for some known 'bc' bugs
9826 so that 'make test' does not abort just because 'bc' is broken.
9829 *) Store verify_result within SSL_SESSION also for client side to
9830 avoid potential security hole. (Re-used sessions on the client side
9831 always resulted in verify_result==X509_V_OK, not using the original
9832 result of the server certificate verification.)
9835 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9836 SSL3_RT_APPLICATION_DATA, return 0.
9837 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9841 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9842 releases, have been re-implemented by renaming the previous
9843 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9844 and ssl3_read_internal, respectively, and adding 'peek' parameters
9845 to them. The new ssl[23]_{read,peek} functions are calls to
9846 ssl[23]_read_internal with the 'peek' flag set appropriately.
9847 A 'peek' parameter has also been added to ssl3_read_bytes, which
9848 does the actual work for ssl3_read_internal.
9851 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9852 the method-specific "init()" handler. Also clean up ex_data after
9853 calling the method-specific "finish()" handler. Previously, this was
9854 happening the other way round.
9857 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9858 The previous value, 12, was not always sufficient for BN_mod_exp().
9861 *) Make sure that shared libraries get the internal name engine with
9862 the full version number and not just 0. This should mark the
9863 shared libraries as not backward compatible. Of course, this should
9864 be changed again when we can guarantee backward binary compatibility.
9867 *) Fix typo in get_cert_by_subject() in by_dir.c
9868 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9870 *) Rework the system to generate shared libraries:
9872 - Make note of the expected extension for the shared libraries and
9873 if there is a need for symbolic links from for example libcrypto.so.0
9874 to libcrypto.so.0.9.7. There is extended info in Configure for
9877 - Make as few rebuilds of the shared libraries as possible.
9879 - Still avoid linking the OpenSSL programs with the shared libraries.
9881 - When installing, install the shared libraries separately from the
9885 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9887 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9888 and not in SSL_clear because the latter is also used by the
9889 accept/connect functions; previously, the settings made by
9890 SSL_set_read_ahead would be lost during the handshake.
9891 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9893 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9894 Previously, it would create entries for disabled algorithms no
9898 *) Added several new manual pages for SSL_* function.
9901 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9903 *) In ssl23_get_client_hello, generate an error message when faced
9904 with an initial SSL 3.0/TLS record that is too small to contain the
9905 first two bytes of the ClientHello message, i.e. client_version.
9906 (Note that this is a pathologic case that probably has never happened
9907 in real life.) The previous approach was to use the version number
9908 from the record header as a substitute; but our protocol choice
9909 should not depend on that one because it is not authenticated
9910 by the Finished messages.
9913 *) More robust randomness gathering functions for Windows.
9914 [Jeffrey Altman <jaltman@columbia.edu>]
9916 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9917 not set then we don't setup the error code for issuer check errors
9918 to avoid possibly overwriting other errors which the callback does
9919 handle. If an application does set the flag then we assume it knows
9920 what it is doing and can handle the new informational codes
9924 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9925 a general "ANY" type, as such it should be able to decode anything
9926 including tagged types. However it didn't check the class so it would
9927 wrongly interpret tagged types in the same way as their universal
9928 counterpart and unknown types were just rejected. Changed so that the
9929 tagged and unknown types are handled in the same way as a SEQUENCE:
9930 that is the encoding is stored intact. There is also a new type
9931 "V_ASN1_OTHER" which is used when the class is not universal, in this
9932 case we have no idea what the actual type is so we just lump them all
9936 *) On VMS, stdout may very well lead to a file that is written to
9937 in a record-oriented fashion. That means that every write() will
9938 write a separate record, which will be read separately by the
9939 programs trying to read from it. This can be very confusing.
9941 The solution is to put a BIO filter in the way that will buffer
9942 text until a linefeed is reached, and then write everything a
9943 line at a time, so every record written will be an actual line,
9944 not chunks of lines and not (usually doesn't happen, but I've
9945 seen it once) several lines in one record. BIO_f_linebuffer() is
9948 Currently, it's a VMS-only method, because that's where it has
9949 been tested well enough.
9952 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9953 it can return incorrect results.
9954 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9955 but it was in 0.9.6-beta[12].)
9958 *) Disable the check for content being present when verifying detached
9959 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9960 include zero length content when signing messages.
9963 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9964 BIO_ctrl (for BIO pairs).
9967 *) Add DSO method for VMS.
9970 *) Bug fix: Montgomery multiplication could produce results with the
9974 *) Add RPM specification openssl.spec and modify it to build three
9975 packages. The default package contains applications, application
9976 documentation and run-time libraries. The devel package contains
9977 include files, static libraries and function documentation. The
9978 doc package contains the contents of the doc directory. The original
9979 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9982 *) Add a large number of documentation files for many SSL routines.
9983 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9985 *) Add a configuration entry for Sony News 4.
9986 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9988 *) Don't set the two most significant bits to one when generating a
9989 random number < q in the DSA library.
9992 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9993 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9994 the underlying transport is blocking) if a handshake took place.
9995 (The default behaviour is needed by applications such as s_client
9996 and s_server that use select() to determine when to use SSL_read;
9997 but for applications that know in advance when to expect data, it
9998 just makes things more complicated.)
10001 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
10005 *) Add a few more EBCDIC conditionals that make `req' and `x509'
10006 work better on such systems.
10007 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10009 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10010 Update PKCS12_parse() so it copies the friendlyName and the
10011 keyid to the certificates aux info.
10014 *) Fix bug in PKCS7_verify() which caused an infinite loop
10015 if there was more than one signature.
10016 [Sven Uszpelkat <su@celocom.de>]
10018 *) Major change in util/mkdef.pl to include extra information
10019 about each symbol, as well as presenting variables as well
10020 as functions. This change means that there's n more need
10021 to rebuild the .num files when some algorithms are excluded.
10024 *) Allow the verify time to be set by an application,
10025 rather than always using the current time.
10028 *) Phase 2 verify code reorganisation. The certificate
10029 verify code now looks up an issuer certificate by a
10030 number of criteria: subject name, authority key id
10031 and key usage. It also verifies self signed certificates
10032 by the same criteria. The main comparison function is
10033 X509_check_issued() which performs these checks.
10035 Lot of changes were necessary in order to support this
10036 without completely rewriting the lookup code.
10038 Authority and subject key identifier are now cached.
10040 The LHASH 'certs' is X509_STORE has now been replaced
10041 by a STACK_OF(X509_OBJECT). This is mainly because an
10042 LHASH can't store or retrieve multiple objects with
10043 the same hash value.
10045 As a result various functions (which were all internal
10046 use only) have changed to handle the new X509_STORE
10047 structure. This will break anything that messed round
10048 with X509_STORE internally.
10050 The functions X509_STORE_add_cert() now checks for an
10051 exact match, rather than just subject name.
10053 The X509_STORE API doesn't directly support the retrieval
10054 of multiple certificates matching a given criteria, however
10055 this can be worked round by performing a lookup first
10056 (which will fill the cache with candidate certificates)
10057 and then examining the cache for matches. This is probably
10058 the best we can do without throwing out X509_LOOKUP
10059 entirely (maybe later...).
10061 The X509_VERIFY_CTX structure has been enhanced considerably.
10063 All certificate lookup operations now go via a get_issuer()
10064 callback. Although this currently uses an X509_STORE it
10065 can be replaced by custom lookups. This is a simple way
10066 to bypass the X509_STORE hackery necessary to make this
10067 work and makes it possible to use more efficient techniques
10068 in future. A very simple version which uses a simple
10069 STACK for its trusted certificate store is also provided
10070 using X509_STORE_CTX_trusted_stack().
10072 The verify_cb() and verify() callbacks now have equivalents
10073 in the X509_STORE_CTX structure.
10075 X509_STORE_CTX also has a 'flags' field which can be used
10076 to customise the verify behaviour.
10079 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10080 excludes S/MIME capabilities.
10083 *) When a certificate request is read in keep a copy of the
10084 original encoding of the signed data and use it when outputting
10085 again. Signatures then use the original encoding rather than
10086 a decoded, encoded version which may cause problems if the
10087 request is improperly encoded.
10090 *) For consistency with other BIO_puts implementations, call
10091 buffer_write(b, ...) directly in buffer_puts instead of calling
10094 In BIO_puts, increment b->num_write as in BIO_write.
10095 [Peter.Sylvester@EdelWeb.fr]
10097 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10098 BN_zero, we may not return a BIGNUM with an array consisting of
10099 words set to zero.)
10102 *) Avoid calling abort() from within the library when problems are
10103 detected, except if preprocessor symbols have been defined
10104 (such as REF_CHECK, BN_DEBUG etc.).
10107 *) New openssl application 'rsautl'. This utility can be
10108 used for low level RSA operations. DER public key
10109 BIO/fp routines also added.
10112 *) New Configure entry and patches for compiling on QNX 4.
10113 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10115 *) A demo state-machine implementation was sponsored by
10116 Nuron (http://www.nuron.com/) and is now available in
10117 demos/state_machine.
10120 *) New options added to the 'dgst' utility for signature
10121 generation and verification.
10124 *) Unrecognized PKCS#7 content types are now handled via a
10125 catch all ASN1_TYPE structure. This allows unsupported
10126 types to be stored as a "blob" and an application can
10127 encode and decode it manually.
10130 *) Fix various signed/unsigned issues to make a_strex.c
10131 compile under VC++.
10132 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10134 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10135 length if passed a buffer. ASN1_INTEGER_to_BN failed
10136 if passed a NULL BN and its argument was negative.
10137 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10139 *) Modification to PKCS#7 encoding routines to output definite
10140 length encoding. Since currently the whole structures are in
10141 memory there's not real point in using indefinite length
10142 constructed encoding. However if OpenSSL is compiled with
10143 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10146 *) Added BIO_vprintf() and BIO_vsnprintf().
10149 *) Added more prefixes to parse for in the strings written
10150 through a logging bio, to cover all the levels that are available
10151 through syslog. The prefixes are now:
10153 PANIC, EMERG, EMR => LOG_EMERG
10154 ALERT, ALR => LOG_ALERT
10155 CRIT, CRI => LOG_CRIT
10156 ERROR, ERR => LOG_ERR
10157 WARNING, WARN, WAR => LOG_WARNING
10158 NOTICE, NOTE, NOT => LOG_NOTICE
10159 INFO, INF => LOG_INFO
10160 DEBUG, DBG => LOG_DEBUG
10162 and as before, if none of those prefixes are present at the
10163 beginning of the string, LOG_ERR is chosen.
10165 On Win32, the LOG_* levels are mapped according to this:
10167 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10168 LOG_WARNING => EVENTLOG_WARNING_TYPE
10169 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10173 *) Made it possible to reconfigure with just the configuration
10174 argument "reconf" or "reconfigure". The command line arguments
10175 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10176 and are retrieved from there when reconfiguring.
10179 *) MD4 implemented.
10180 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10182 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10185 *) The obj_dat.pl script was messing up the sorting of object
10186 names. The reason was that it compared the quoted version
10187 of strings as a result "OCSP" > "OCSP Signing" because
10188 " > SPACE. Changed script to store unquoted versions of
10189 names and add quotes on output. It was also omitting some
10190 names from the lookup table if they were given a default
10191 value (that is if SN is missing it is given the same
10192 value as LN and vice versa), these are now added on the
10193 grounds that if an object has a name we should be able to
10194 look it up. Finally added warning output when duplicate
10195 short or long names are found.
10198 *) Changes needed for Tandem NSK.
10199 [Scott Uroff <scott@xypro.com>]
10201 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10202 RSA_padding_check_SSLv23(), special padding was never detected
10203 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10204 version rollback attacks was not effective.
10206 In s23_clnt.c, don't use special rollback-attack detection padding
10207 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10208 client; similarly, in s23_srvr.c, don't do the rollback check if
10209 SSL 2.0 is the only protocol enabled in the server.
10212 *) Make it possible to get hexdumps of unprintable data with 'openssl
10213 asn1parse'. By implication, the functions ASN1_parse_dump() and
10214 BIO_dump_indent() are added.
10217 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10218 these print out strings and name structures based on various
10219 flags including RFC2253 support and proper handling of
10220 multibyte characters. Added options to the 'x509' utility
10221 to allow the various flags to be set.
10224 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10225 Also change the functions X509_cmp_current_time() and
10226 X509_gmtime_adj() work with an ASN1_TIME structure,
10227 this will enable certificates using GeneralizedTime in validity
10228 dates to be checked.
10231 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10232 negative public key encodings) on by default,
10233 NO_NEG_PUBKEY_BUG can be set to disable it.
10236 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10237 content octets. An i2c_ASN1_OBJECT is unnecessary because
10238 the encoding can be trivially obtained from the structure.
10241 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10242 not read locks (CRYPTO_r_[un]lock).
10245 *) A first attempt at creating official support for shared
10246 libraries through configuration. I've kept it so the
10247 default is static libraries only, and the OpenSSL programs
10248 are always statically linked for now, but there are
10249 preparations for dynamic linking in place.
10250 This has been tested on Linux and Tru64.
10253 *) Randomness polling function for Win9x, as described in:
10254 Peter Gutmann, Software Generation of Practically Strong
10258 *) Fix so PRNG is seeded in req if using an already existing
10262 *) New options to smime application. -inform and -outform
10263 allow alternative formats for the S/MIME message including
10264 PEM and DER. The -content option allows the content to be
10265 specified separately. This should allow things like Netscape
10266 form signing output easier to verify.
10269 *) Fix the ASN1 encoding of tags using the 'long form'.
10272 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10273 STRING types. These convert content octets to and from the
10274 underlying type. The actual tag and length octets are
10275 already assumed to have been read in and checked. These
10276 are needed because all other string types have virtually
10277 identical handling apart from the tag. By having versions
10278 of the ASN1 functions that just operate on content octets
10279 IMPLICIT tagging can be handled properly. It also allows
10280 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10281 and ASN1_INTEGER are identical apart from the tag.
10284 *) Change the handling of OID objects as follows:
10286 - New object identifiers are inserted in objects.txt, following
10287 the syntax given in objects.README.
10288 - objects.pl is used to process obj_mac.num and create a new
10290 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10293 This is currently kind of a hack, and the perl code in objects.pl
10294 isn't very elegant, but it works as I intended. The simplest way
10295 to check that it worked correctly is to look in obj_dat.h and
10296 check the array nid_objs and make sure the objects haven't moved
10297 around (this is important!). Additions are OK, as well as
10298 consistent name changes.
10301 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10304 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10305 The given file adds to whatever has already been seeded into the
10306 random pool through the RANDFILE configuration file option or
10307 environment variable, or the default random state file.
10310 *) mkstack.pl now sorts each macro group into lexical order.
10311 Previously the output order depended on the order the files
10312 appeared in the directory, resulting in needless rewriting
10316 *) Patches to make OpenSSL compile under Win32 again. Mostly
10317 work arounds for the VC++ problem that it treats func() as
10318 func(void). Also stripped out the parts of mkdef.pl that
10319 added extra typesafe functions: these no longer exist.
10322 *) Reorganisation of the stack code. The macros are now all
10323 collected in safestack.h . Each macro is defined in terms of
10324 a "stack macro" of the form SKM_<name>(type, a, b). The
10325 DEBUG_SAFESTACK is now handled in terms of function casts,
10326 this has the advantage of retaining type safety without the
10327 use of additional functions. If DEBUG_SAFESTACK is not defined
10328 then the non typesafe macros are used instead. Also modified the
10329 mkstack.pl script to handle the new form. Needs testing to see
10330 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10331 the default if no major problems. Similar behaviour for ASN1_SET_OF
10332 and PKCS12_STACK_OF.
10335 *) When some versions of IIS use the 'NET' form of private key the
10336 key derivation algorithm is different. Normally MD5(password) is
10337 used as a 128 bit RC4 key. In the modified case
10338 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10339 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10340 as the old Netscape_RSA functions except they have an additional
10341 'sgckey' parameter which uses the modified algorithm. Also added
10342 an -sgckey command line option to the rsa utility. Thanks to
10343 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10344 algorithm to openssl-dev.
10347 *) The evp_local.h macros were using 'c.##kname' which resulted in
10348 invalid expansion on some systems (SCO 5.0.5 for example).
10349 Corrected to 'c.kname'.
10350 [Phillip Porch <root@theporch.com>]
10352 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10353 a STACK of email addresses from a certificate or request, these look
10354 in the subject name and the subject alternative name extensions and
10355 omit any duplicate addresses.
10358 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10359 This makes DSA verification about 2 % faster.
10362 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10363 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10364 plus overhead for 1024 bit moduli).
10365 This makes exponentiations about 0.5 % faster for 1024 bit
10366 exponents (as measured by "openssl speed rsa2048").
10369 *) Rename memory handling macros to avoid conflicts with other
10371 Malloc => OPENSSL_malloc
10372 Malloc_locked => OPENSSL_malloc_locked
10373 Realloc => OPENSSL_realloc
10374 Free => OPENSSL_free
10377 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10378 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10381 *) CygWin32 support.
10382 [John Jarvie <jjarvie@newsguy.com>]
10384 *) The type-safe stack code has been rejigged. It is now only compiled
10385 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10386 by default all type-specific stack functions are "#define"d back to
10387 standard stack functions. This results in more streamlined output
10388 but retains the type-safety checking possibilities of the original
10392 *) The STACK code has been cleaned up, and certain type declarations
10393 that didn't make a lot of sense have been brought in line. This has
10394 also involved a cleanup of sorts in safestack.h to more correctly
10395 map type-safe stack functions onto their plain stack counterparts.
10396 This work has also resulted in a variety of "const"ifications of
10397 lots of the code, especially "_cmp" operations which should normally
10398 be prototyped with "const" parameters anyway.
10401 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10402 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10403 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10404 where all of 'md' is used each time the PRNG is used, but 'state'
10405 is used only indexed by a cyclic counter. As entropy may not be
10406 well distributed from the beginning, 'md' is important as a
10407 chaining variable. However, the output function chains only half
10408 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10409 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10410 in all of 'state' being rewritten, with the new values depending
10411 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10414 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10415 the handshake is continued after ssl_verify_cert_chain();
10416 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10417 can lead to 'unexplainable' connection aborts later.
10418 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10420 *) Major EVP API cipher revision.
10421 Add hooks for extra EVP features. This allows various cipher
10422 parameters to be set in the EVP interface. Support added for variable
10423 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10424 setting of RC2 and RC5 parameters.
10426 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10429 Remove lots of duplicated code from the EVP library. For example *every*
10430 cipher init() function handles the 'iv' in the same way according to the
10431 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10432 for CFB and OFB modes they zero ctx->num.
10434 New functionality allows removal of S/MIME code RC2 hack.
10436 Most of the routines have the same form and so can be declared in terms
10439 By shifting this to the top level EVP_CipherInit() it can be removed from
10440 all individual ciphers. If the cipher wants to handle IVs or keys
10441 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10444 Change lots of functions like EVP_EncryptUpdate() to now return a
10445 value: although software versions of the algorithms cannot fail
10446 any installed hardware versions can.
10449 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10450 this option is set, tolerate broken clients that send the negotiated
10451 protocol version number instead of the requested protocol version
10455 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10456 i.e. non-zero for export ciphersuites, zero otherwise.
10457 Previous versions had this flag inverted, inconsistent with
10458 rsa_tmp_cb (..._TMP_RSA_CB).
10459 [Bodo Moeller; problem reported by Amit Chopra]
10461 *) Add missing DSA library text string. Work around for some IIS
10462 key files with invalid SEQUENCE encoding.
10465 *) Add a document (doc/standards.txt) that list all kinds of standards
10466 and so on that are implemented in OpenSSL.
10469 *) Enhance c_rehash script. Old version would mishandle certificates
10470 with the same subject name hash and wouldn't handle CRLs at all.
10471 Added -fingerprint option to crl utility, to support new c_rehash
10475 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10478 *) Fix for SSL server purpose checking. Server checking was
10479 rejecting certificates which had extended key usage present
10480 but no ssl client purpose.
10481 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10483 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10484 is a little unclear about how a blank password is handled.
10485 Since the password in encoded as a BMPString with terminating
10486 double NULL a zero length password would end up as just the
10487 double NULL. However no password at all is different and is
10488 handled differently in the PKCS#12 key generation code. NS
10489 treats a blank password as zero length. MSIE treats it as no
10490 password on export: but it will try both on import. We now do
10491 the same: PKCS12_parse() tries zero length and no password if
10492 the password is set to "" or NULL (NULL is now a valid password:
10493 it wasn't before) as does the pkcs12 application.
10496 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10497 perror when PEM_read_bio_X509_REQ fails, the error message must
10498 be obtained from the error queue.
10501 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10502 it in ERR_remove_state if appropriate, and change ERR_get_state
10503 accordingly to avoid race conditions (this is necessary because
10504 thread_hash is no longer constant once set).
10507 *) Bugfix for linux-elf makefile.one.
10510 *) RSA_get_default_method() will now cause a default
10511 RSA_METHOD to be chosen if one doesn't exist already.
10512 Previously this was only set during a call to RSA_new()
10513 or RSA_new_method(NULL) meaning it was possible for
10514 RSA_get_default_method() to return NULL.
10517 *) Added native name translation to the existing DSO code
10518 that will convert (if the flag to do so is set) filenames
10519 that are sufficiently small and have no path information
10520 into a canonical native form. Eg. "blah" converted to
10521 "libblah.so" or "blah.dll" etc.
10524 *) New function ERR_error_string_n(e, buf, len) which is like
10525 ERR_error_string(e, buf), but writes at most 'len' bytes
10526 including the 0 terminator. For ERR_error_string_n, 'buf'
10528 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10530 *) CONF library reworked to become more general. A new CONF
10531 configuration file reader "class" is implemented as well as a
10532 new functions (NCONF_*, for "New CONF") to handle it. The now
10533 old CONF_* functions are still there, but are reimplemented to
10534 work in terms of the new functions. Also, a set of functions
10535 to handle the internal storage of the configuration data is
10536 provided to make it easier to write new configuration file
10537 reader "classes" (I can definitely see something reading a
10538 configuration file in XML format, for example), called _CONF_*,
10539 or "the configuration storage API"...
10541 The new configuration file reading functions are:
10543 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10544 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10546 NCONF_default, NCONF_WIN32
10548 NCONF_dump_fp, NCONF_dump_bio
10550 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10551 NCONF_new creates a new CONF object. This works in the same way
10552 as other interfaces in OpenSSL, like the BIO interface.
10553 NCONF_dump_* dump the internal storage of the configuration file,
10554 which is useful for debugging. All other functions take the same
10555 arguments as the old CONF_* functions with the exception of the
10556 first that must be a `CONF *' instead of a `LHASH *'.
10558 To make it easier to use the new classes with the old CONF_* functions,
10559 the function CONF_set_default_method is provided.
10562 *) Add '-tls1' option to 'openssl ciphers', which was already
10563 mentioned in the documentation but had not been implemented.
10564 (This option is not yet really useful because even the additional
10565 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10568 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10569 OpenSSL-based applications) load shared libraries and bind to
10570 them in a portable way.
10571 [Geoff Thorpe, with contributions from Richard Levitte]
10573 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10575 *) Make sure _lrotl and _lrotr are only used with MSVC.
10577 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10578 (the default implementation of RAND_status).
10580 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10581 to '-clrext' (= clear extensions), as intended and documented.
10582 [Bodo Moeller; inconsistency pointed out by Michael Attili
10583 <attili@amaxo.com>]
10585 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10586 was larger than the MD block size.
10587 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10589 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10590 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10591 using the passed key: if the passed key was a private key the result
10592 of X509_print(), for example, would be to print out all the private key
10596 *) des_quad_cksum() byte order bug fix.
10597 [Ulf Möller, using the problem description in krb4-0.9.7, where
10598 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10600 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10602 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10604 *) For easily testing in shell scripts whether some command
10605 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10606 returns with exit code 0 iff no command of the given name is available.
10607 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10608 the output goes to stdout and nothing is printed to stderr.
10609 Additional arguments are always ignored.
10611 Since for each cipher there is a command of the same name,
10612 the 'no-cipher' compilation switches can be tested this way.
10614 ('openssl no-XXX' is not able to detect pseudo-commands such
10615 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10618 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10621 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10622 is set; it will be thrown away anyway because each handshake creates
10624 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10625 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10626 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10627 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10630 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10631 'Q' and 'R' lose their special meanings (quit/renegotiate).
10632 This is part of what -quiet does; unlike -quiet, -ign_eof
10633 does not suppress any output.
10636 *) Add compatibility options to the purpose and trust code. The
10637 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10638 accepts a certificate or CA, this was the previous behaviour,
10639 with all the associated security issues.
10641 X509_TRUST_COMPAT is the old trust behaviour: only and
10642 automatically trust self signed roots in certificate store. A
10643 new trust setting X509_TRUST_DEFAULT is used to specify that
10644 a purpose has no associated trust setting and it should instead
10645 use the value in the default purpose.
10648 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10649 and fix a memory leak.
10652 *) In util/mkerr.pl (which implements 'make errors'), preserve
10653 reason strings from the previous version of the .c file, as
10654 the default to have only downcase letters (and digits) in
10655 automatically generated reasons codes is not always appropriate.
10658 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10659 using strerror. Previously, ERR_reason_error_string() returned
10660 library names as reason strings for SYSerr; but SYSerr is a special
10661 case where small numbers are errno values, not library numbers.
10664 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10665 converts DSA parameters into DH parameters. (When creating parameters,
10666 DSA_generate_parameters is used.)
10669 *) Include 'length' (recommended exponent length) in C code generated
10670 by 'openssl dhparam -C'.
10673 *) The second argument to set_label in perlasm was already being used
10674 so couldn't be used as a "file scope" flag. Moved to third argument
10678 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10679 instead of RAND_bytes for encryption IVs and salts.
10682 *) Include RAND_status() into RAND_METHOD instead of implementing
10683 it only for md_rand.c Otherwise replacing the PRNG by calling
10684 RAND_set_rand_method would be impossible.
10687 *) Don't let DSA_generate_key() enter an infinite loop if the random
10688 number generation fails.
10691 *) New 'rand' application for creating pseudo-random output.
10694 *) Added configuration support for Linux/IA64
10695 [Rolf Haberrecker <rolf@suse.de>]
10697 *) Assembler module support for Mingw32.
10700 *) Shared library support for HPUX (in shlib/).
10701 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10703 *) Shared library support for Solaris gcc.
10704 [Lutz Behnke <behnke@trustcenter.de>]
10706 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10708 *) PKCS7_encrypt() was adding text MIME headers twice because they
10709 were added manually and by SMIME_crlf_copy().
10712 *) In bntest.c don't call BN_rand with zero bits argument.
10713 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10715 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10716 case was implemented. This caused BN_div_recp() to fail occasionally.
10719 *) Add an optional second argument to the set_label() in the perl
10720 assembly language builder. If this argument exists and is set
10721 to 1 it signals that the assembler should use a symbol whose
10722 scope is the entire file, not just the current function. This
10723 is needed with MASM which uses the format label:: for this scope.
10724 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10726 *) Change the ASN1 types so they are typedefs by default. Before
10727 almost all types were #define'd to ASN1_STRING which was causing
10728 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10732 *) Change names of new functions to the new get1/get0 naming
10733 convention: After 'get1', the caller owns a reference count
10734 and has to call ..._free; 'get0' returns a pointer to some
10735 data structure without incrementing reference counters.
10736 (Some of the existing 'get' functions increment a reference
10737 counter, some don't.)
10738 Similarly, 'set1' and 'add1' functions increase reference
10739 counters or duplicate objects.
10742 *) Allow for the possibility of temp RSA key generation failure:
10743 the code used to assume it always worked and crashed on failure.
10746 *) Fix potential buffer overrun problem in BIO_printf().
10747 [Ulf Möller, using public domain code by Patrick Powell; problem
10748 pointed out by David Sacerdote <das33@cornell.edu>]
10750 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10751 RAND_egd() and RAND_status(). In the command line application,
10752 the EGD socket can be specified like a seed file using RANDFILE
10756 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10757 Some CAs (e.g. Verisign) distribute certificates in this form.
10760 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10761 list to exclude them. This means that no special compilation option
10762 is needed to use anonymous DH: it just needs to be included in the
10766 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10767 EVP_MD_type. The old functionality is available in a new macro called
10768 EVP_MD_md(). Change code that uses it and update docs.
10771 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10772 where the 'void *' argument is replaced by a function pointer argument.
10773 Previously 'void *' was abused to point to functions, which works on
10774 many platforms, but is not correct. As these functions are usually
10775 called by macros defined in OpenSSL header files, most source code
10776 should work without changes.
10779 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10780 sections with information on -D... compiler switches used for
10781 compiling the library so that applications can see them. To enable
10782 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10783 must be defined. E.g.,
10784 #define OPENSSL_ALGORITHM_DEFINES
10785 #include <openssl/opensslconf.h>
10786 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10787 [Richard Levitte, Ulf and Bodo Möller]
10789 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10793 *) Change the 'other' type in certificate aux info to a STACK_OF
10794 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10795 the required ASN1 format: arbitrary types determined by an OID.
10798 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10799 argument to 'req'. This is not because the function is newer or
10800 better than others it just uses the work 'NEW' in the certificate
10801 request header lines. Some software needs this.
10804 *) Reorganise password command line arguments: now passwords can be
10805 obtained from various sources. Delete the PEM_cb function and make
10806 it the default behaviour: i.e. if the callback is NULL and the
10807 usrdata argument is not NULL interpret it as a null terminated pass
10808 phrase. If usrdata and the callback are NULL then the pass phrase
10809 is prompted for as usual.
10812 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10813 the support is automatically enabled. The resulting binaries will
10814 autodetect the card and use it if present.
10815 [Ben Laurie and Compaq Inc.]
10817 *) Work around for Netscape hang bug. This sends certificate request
10818 and server done in one record. Since this is perfectly legal in the
10819 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10820 the bugs/SSLv3 entry for more info.
10823 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10826 *) Add -rand argument to smime and pkcs12 applications and read/write
10830 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10833 *) Add command line password options to the remaining applications.
10836 *) Bug fix for BN_div_recp() for numerators with an even number of
10840 *) More tests in bntest.c, and changed test_bn output.
10843 *) ./config recognizes MacOS X now.
10846 *) Bug fix for BN_div() when the first words of num and divisor are
10847 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10850 *) Add support for various broken PKCS#8 formats, and command line
10851 options to produce them.
10854 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10855 get temporary BIGNUMs from a BN_CTX.
10858 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10862 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10863 include a #define from the old name to the new. The original intent
10864 was that statically linked binaries could for example just call
10865 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10866 link with digests. This never worked because SSLeay_add_all_digests()
10867 and SSLeay_add_all_ciphers() were in the same source file so calling
10868 one would link with the other. They are now in separate source files.
10871 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10874 *) Use a less unusual form of the Miller-Rabin primality test (it used
10875 a binary algorithm for exponentiation integrated into the Miller-Rabin
10876 loop, our standard modexp algorithms are faster).
10879 *) Support for the EBCDIC character set completed.
10880 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10882 *) Source code cleanups: use const where appropriate, eliminate casts,
10883 use void * instead of char * in lhash.
10886 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10887 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10888 this the server could overwrite ephemeral keys that the client
10892 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10893 using 50 iterations of the Rabin-Miller test.
10895 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10896 iterations of the Rabin-Miller test as required by the appendix
10897 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10898 As BN_is_prime_fasttest includes trial division, DSA parameter
10899 generation becomes much faster.
10901 This implies a change for the callback functions in DSA_is_prime
10902 and DSA_generate_parameters: The callback function is called once
10903 for each positive witness in the Rabin-Miller test, not just
10904 occasionally in the inner loop; and the parameters to the
10905 callback function now provide an iteration count for the outer
10906 loop rather than for the current invocation of the inner loop.
10907 DSA_generate_parameters additionally can call the callback
10908 function with an 'iteration count' of -1, meaning that a
10909 candidate has passed the trial division test (when q is generated
10910 from an application-provided seed, trial division is skipped).
10913 *) New function BN_is_prime_fasttest that optionally does trial
10914 division before starting the Rabin-Miller test and has
10915 an additional BN_CTX * argument (whereas BN_is_prime always
10916 has to allocate at least one BN_CTX).
10917 'callback(1, -1, cb_arg)' is called when a number has passed the
10918 trial division stage.
10921 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10925 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10928 *) New function BN_pseudo_rand().
10931 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10932 bignum version of BN_from_montgomery() with the working code from
10933 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10937 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10938 made it impossible to use the same SSL_SESSION data structure in
10939 SSL2 clients in multiple threads.
10942 *) The return value of RAND_load_file() no longer counts bytes obtained
10943 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10944 to seed the PRNG (previously an explicit byte count was required).
10945 [Ulf Möller, Bodo Möller]
10947 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10948 used (char *) instead of (void *) and had casts all over the place.
10951 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10954 *) Retain source code compatibility for BN_prime_checks macro:
10955 BN_is_prime(..., BN_prime_checks, ...) now uses
10956 BN_prime_checks_for_size to determine the appropriate number of
10957 Rabin-Miller iterations.
10960 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10961 DH_CHECK_P_NOT_SAFE_PRIME.
10962 (Check if this is true? OpenPGP calls them "strong".)
10965 *) Merge the functionality of "dh" and "gendh" programs into a new program
10966 "dhparam". The old programs are retained for now but will handle DH keys
10967 (instead of parameters) in future.
10970 *) Make the ciphers, s_server and s_client programs check the return values
10971 when a new cipher list is set.
10974 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10975 ciphers. Before when the 56bit ciphers were enabled the sorting was
10978 The syntax for the cipher sorting has been extended to support sorting by
10979 cipher-strength (using the strength_bits hard coded in the tables).
10980 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10982 Fix a bug in the cipher-command parser: when supplying a cipher command
10983 string with an "undefined" symbol (neither command nor alphanumeric
10984 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10985 an error is flagged.
10987 Due to the strength-sorting extension, the code of the
10988 ssl_create_cipher_list() function was completely rearranged. I hope that
10989 the readability was also increased :-)
10990 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10992 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10993 for the first serial number and places 2 in the serial number file. This
10994 avoids problems when the root CA is created with serial number zero and
10995 the first user certificate has the same issuer name and serial number
10999 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
11000 the new code. Add documentation for this stuff.
11003 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
11004 X509_*() to X509at_*() on the grounds that they don't handle X509
11005 structures and behave in an analogous way to the X509v3 functions:
11006 they shouldn't be called directly but wrapper functions should be used
11009 So we also now have some wrapper functions that call the X509at functions
11010 when passed certificate requests. (TO DO: similar things can be done with
11011 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11012 things. Some of these need some d2i or i2d and print functionality
11013 because they handle more complex structures.)
11016 *) Add missing #ifndefs that caused missing symbols when building libssl
11017 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11018 NO_RSA in ssl/s2*.c.
11019 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11021 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11022 has a return value which indicates the quality of the random data
11023 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11024 error queue. New function RAND_pseudo_bytes() generates output that is
11025 guaranteed to be unique but not unpredictable. RAND_add is like
11026 RAND_seed, but takes an extra argument for an entropy estimate
11027 (RAND_seed always assumes full entropy).
11030 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11031 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11032 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11033 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11034 false-positive rate of at most 2^-80 for random input.
11037 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11040 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11041 in the 0.9.5 release), this returns the chain
11042 from an X509_CTX structure with a dup of the stack and all
11043 the X509 reference counts upped: so the stack will exist
11044 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11047 Also make SSL_SESSION_print() print out the verify return
11051 *) Add manpage for the pkcs12 command. Also change the default
11052 behaviour so MAC iteration counts are used unless the new
11053 -nomaciter option is used. This improves file security and
11054 only older versions of MSIE (4.0 for example) need it.
11057 *) Honor the no-xxx Configure options when creating .DEF files.
11060 *) Add PKCS#10 attributes to field table: challengePassword,
11061 unstructuredName and unstructuredAddress. These are taken from
11062 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11063 international characters are used.
11065 More changes to X509_ATTRIBUTE code: allow the setting of types
11066 based on strings. Remove the 'loc' parameter when adding
11067 attributes because these will be a SET OF encoding which is sorted
11071 *) Initial changes to the 'req' utility to allow request generation
11072 automation. This will allow an application to just generate a template
11073 file containing all the field values and have req construct the
11076 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11077 used all over the place including certificate requests and PKCS#7
11078 structures. They are currently handled manually where necessary with
11079 some primitive wrappers for PKCS#7. The new functions behave in a
11080 manner analogous to the X509 extension functions: they allow
11081 attributes to be looked up by NID and added.
11083 Later something similar to the X509V3 code would be desirable to
11084 automatically handle the encoding, decoding and printing of the
11085 more complex types. The string types like challengePassword can
11086 be handled by the string table functions.
11088 Also modified the multi byte string table handling. Now there is
11089 a 'global mask' which masks out certain types. The table itself
11090 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11091 is useful when for example there is only one permissible type
11092 (as in countryName) and using the mask might result in no valid
11096 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11097 SSL_get_peer_finished to allow applications to obtain the latest
11098 Finished messages sent to the peer or expected from the peer,
11099 respectively. (SSL_get_peer_finished is usually the Finished message
11100 actually received from the peer, otherwise the protocol will be aborted.)
11102 As the Finished message are message digests of the complete handshake
11103 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11104 be used for external authentication procedures when the authentication
11105 provided by SSL/TLS is not desired or is not enough.
11108 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11109 the host supports BWX extension and if Compaq C is present on the
11110 $PATH. Just exploiting of the BWX extension results in 20-30%
11111 performance kick for some algorithms, e.g. DES and RC4 to mention
11112 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11116 *) Add support for MS "fast SGC". This is arguably a violation of the
11117 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11118 weak crypto and after checking the certificate is SGC a second one
11119 with strong crypto. MS SGC stops the first handshake after receiving
11120 the server certificate message and sends a second client hello. Since
11121 a server will typically do all the time consuming operations before
11122 expecting any further messages from the client (server key exchange
11123 is the most expensive) there is little difference between the two.
11125 To get OpenSSL to support MS SGC we have to permit a second client
11126 hello message after we have sent server done. In addition we have to
11127 reset the MAC if we do get this second client hello.
11130 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11131 if a DER encoded private key is RSA or DSA traditional format. Changed
11132 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11133 format DER encoded private key. Newer code should use PKCS#8 format which
11134 has the key type encoded in the ASN1 structure. Added DER private key
11135 support to pkcs8 application.
11138 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11139 ciphersuites has been selected (as required by the SSL 3/TLS 1
11140 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11141 is set, we interpret this as a request to violate the specification
11142 (the worst that can happen is a handshake failure, and 'correct'
11143 behaviour would result in a handshake failure anyway).
11146 *) In SSL_CTX_add_session, take into account that there might be multiple
11147 SSL_SESSION structures with the same session ID (e.g. when two threads
11148 concurrently obtain them from an external cache).
11149 The internal cache can handle only one SSL_SESSION with a given ID,
11150 so if there's a conflict, we now throw out the old one to achieve
11154 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11155 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11156 some routines that use cipher OIDs: some ciphers do not have OIDs
11157 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11161 *) Simplify the trust setting structure and code. Now we just have
11162 two sequences of OIDs for trusted and rejected settings. These will
11163 typically have values the same as the extended key usage extension
11164 and any application specific purposes.
11166 The trust checking code now has a default behaviour: it will just
11167 check for an object with the same NID as the passed id. Functions can
11168 be provided to override either the default behaviour or the behaviour
11169 for a given id. SSL client, server and email already have functions
11170 in place for compatibility: they check the NID and also return "trusted"
11171 if the certificate is self signed.
11174 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11175 traditional format into an EVP_PKEY structure.
11178 *) Add a password callback function PEM_cb() which either prompts for
11179 a password if usr_data is NULL or otherwise assumes it is a null
11180 terminated password. Allow passwords to be passed on command line
11181 environment or config files in a few more utilities.
11184 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11185 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11186 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11187 Update documentation.
11190 *) Support for ASN1 "NULL" type. This could be handled before by using
11191 ASN1_TYPE but there wasn't any function that would try to read a NULL
11192 and produce an error if it couldn't. For compatibility we also have
11193 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11194 don't allocate anything because they don't need to.
11197 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11199 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11201 *) Rebuild of the memory allocation routines used by OpenSSL code and
11202 possibly others as well. The purpose is to make an interface that
11203 provide hooks so anyone can build a separate set of allocation and
11204 deallocation routines to be used by OpenSSL, for example memory
11205 pool implementations, or something else, which was previously hard
11206 since Malloc(), Realloc() and Free() were defined as macros having
11207 the values malloc, realloc and free, respectively (except for Win32
11208 compilations). The same is provided for memory debugging code.
11209 OpenSSL already comes with functionality to find memory leaks, but
11210 this gives people a chance to debug other memory problems.
11212 With these changes, a new set of functions and macros have appeared:
11214 CRYPTO_set_mem_debug_functions() [F]
11215 CRYPTO_get_mem_debug_functions() [F]
11216 CRYPTO_dbg_set_options() [F]
11217 CRYPTO_dbg_get_options() [F]
11218 CRYPTO_malloc_debug_init() [M]
11220 The memory debug functions are NULL by default, unless the library
11221 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11222 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11223 gives the standard debugging functions that come with OpenSSL) or
11224 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11225 provided by the library user) must be used. When the standard
11226 debugging functions are used, CRYPTO_dbg_set_options can be used to
11227 request additional information:
11228 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11229 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11231 Also, things like CRYPTO_set_mem_functions will always give the
11232 expected result (the new set of functions is used for allocation
11233 and deallocation) at all times, regardless of platform and compiler
11236 To finish it up, some functions that were never use in any other
11237 way than through macros have a new API and new semantic:
11239 CRYPTO_dbg_malloc()
11240 CRYPTO_dbg_realloc()
11243 All macros of value have retained their old syntax.
11244 [Richard Levitte and Bodo Moeller]
11246 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11247 ordering of SMIMECapabilities wasn't in "strength order" and there
11248 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11252 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11253 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11254 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11256 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11257 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11258 functionality to handle multipart/signed properly) and a utility
11259 called 'smime' to call all this stuff. This is based on code I
11260 originally wrote for Celo who have kindly allowed it to be
11261 included in OpenSSL.
11264 *) Add variants des_set_key_checked and des_set_key_unchecked of
11265 des_set_key (aka des_key_sched). Global variable des_check_key
11266 decides which of these is called by des_set_key; this way
11267 des_check_key behaves as it always did, but applications and
11268 the library itself, which was buggy for des_check_key == 1,
11269 have a cleaner way to pick the version they need.
11272 *) New function PKCS12_newpass() which changes the password of a
11276 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11277 dynamic mix. In both cases the ids can be used as an index into the
11278 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11279 functions so they accept a list of the field values and the
11280 application doesn't need to directly manipulate the X509_TRUST
11284 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11288 *) Modify the way the V3 extension code looks up extensions. This now
11289 works in a similar way to the object code: we have some "standard"
11290 extensions in a static table which is searched with OBJ_bsearch()
11291 and the application can add dynamic ones if needed. The file
11292 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11293 updated whenever a new extension is added to the core code and kept
11294 in ext_nid order. There is a simple program 'tabtest.c' which checks
11295 this. New extensions are not added too often so this file can readily
11296 be maintained manually.
11298 There are two big advantages in doing things this way. The extensions
11299 can be looked up immediately and no longer need to be "added" using
11300 X509V3_add_standard_extensions(): this function now does nothing.
11301 [Side note: I get *lots* of email saying the extension code doesn't
11302 work because people forget to call this function]
11303 Also no dynamic allocation is done unless new extensions are added:
11304 so if we don't add custom extensions there is no need to call
11305 X509V3_EXT_cleanup().
11308 *) Modify enc utility's salting as follows: make salting the default. Add a
11309 magic header, so unsalted files fail gracefully instead of just decrypting
11310 to garbage. This is because not salting is a big security hole, so people
11311 should be discouraged from doing it.
11314 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11315 digest to be passed on the command line but it only used this
11316 parameter when signing a certificate. Modified so all relevant
11317 operations are affected by the digest parameter including the
11318 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11319 DSA key was used because it didn't fix the digest.
11322 *) Initial certificate chain verify code. Currently tests the untrusted
11323 certificates for consistency with the verify purpose (which is set
11324 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11326 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11327 this is because it will reject chains with invalid extensions whereas
11328 every previous version of OpenSSL and SSLeay made no checks at all.
11330 Trust code: checks the root CA for the relevant trust settings. Trust
11331 settings have an initial value consistent with the verify purpose: e.g.
11332 if the verify purpose is for SSL client use it expects the CA to be
11333 trusted for SSL client use. However the default value can be changed to
11334 permit custom trust settings: one example of this would be to only trust
11335 certificates from a specific "secure" set of CAs.
11337 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11338 which should be used for version portability: especially since the
11339 verify structure is likely to change more often now.
11341 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11342 to set them. If not set then assume SSL clients will verify SSL servers
11345 Two new options to the verify program: -untrusted allows a set of
11346 untrusted certificates to be passed in and -purpose which sets the
11347 intended purpose of the certificate. If a purpose is set then the
11348 new chain verify code is used to check extension consistency.
11351 *) Support for the authority information access extension.
11354 *) Modify RSA and DSA PEM read routines to transparently handle
11355 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11356 public keys in a format compatible with certificate
11357 SubjectPublicKeyInfo structures. Unfortunately there were already
11358 functions called *_PublicKey_* which used various odd formats so
11359 these are retained for compatibility: however the DSA variants were
11360 never in a public release so they have been deleted. Changed dsa/rsa
11361 utilities to handle the new format: note no releases ever handled public
11362 keys so we should be OK.
11364 The primary motivation for this change is to avoid the same fiasco
11365 that dogs private keys: there are several incompatible private key
11366 formats some of which are standard and some OpenSSL specific and
11367 require various evil hacks to allow partial transparent handling and
11368 even then it doesn't work with DER formats. Given the option anything
11369 other than PKCS#8 should be dumped: but the other formats have to
11370 stay in the name of compatibility.
11372 With public keys and the benefit of hindsight one standard format
11373 is used which works with EVP_PKEY, RSA or DSA structures: though
11374 it clearly returns an error if you try to read the wrong kind of key.
11376 Added a -pubkey option to the 'x509' utility to output the public key.
11377 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11378 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11379 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11380 that do the same as the EVP_PKEY_assign_*() except they up the
11381 reference count of the added key (they don't "swallow" the
11385 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11386 CRLs would fail if the file contained no certificates or no CRLs:
11387 added a new function to read in both types and return the number
11388 read: this means that if none are read it will be an error. The
11389 DER versions of the certificate and CRL reader would always fail
11390 because it isn't possible to mix certificates and CRLs in DER format
11391 without choking one or the other routine. Changed this to just read
11392 a certificate: this is the best we can do. Also modified the code
11393 in apps/verify.c to take notice of return codes: it was previously
11394 attempting to read in certificates from NULL pointers and ignoring
11395 any errors: this is one reason why the cert and CRL reader seemed
11396 to work. It doesn't check return codes from the default certificate
11397 routines: these may well fail if the certificates aren't installed.
11400 *) Code to support otherName option in GeneralName.
11403 *) First update to verify code. Change the verify utility
11404 so it warns if it is passed a self signed certificate:
11405 for consistency with the normal behaviour. X509_verify
11406 has been modified to it will now verify a self signed
11407 certificate if *exactly* the same certificate appears
11408 in the store: it was previously impossible to trust a
11409 single self signed certificate. This means that:
11410 openssl verify ss.pem
11411 now gives a warning about a self signed certificate but
11412 openssl verify -CAfile ss.pem ss.pem
11416 *) For servers, store verify_result in SSL_SESSION data structure
11417 (and add it to external session representation).
11418 This is needed when client certificate verifications fails,
11419 but an application-provided verification callback (set by
11420 SSL_CTX_set_cert_verify_callback) allows accepting the session
11421 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11422 but returns 1): When the session is reused, we have to set
11423 ssl->verify_result to the appropriate error code to avoid
11425 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11427 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11428 case in PKCS7_dataInit() where the signed PKCS7 structure
11429 didn't contain any existing data because it was being created.
11430 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11432 *) Add a salt to the key derivation routines in enc.c. This
11433 forms the first 8 bytes of the encrypted file. Also add a
11434 -S option to allow a salt to be input on the command line.
11437 *) New function X509_cmp(). Oddly enough there wasn't a function
11438 to compare two certificates. We do this by working out the SHA1
11439 hash and comparing that. X509_cmp() will be needed by the trust
11443 *) SSL_get1_session() is like SSL_get_session(), but increments
11444 the reference count in the SSL_SESSION returned.
11445 [Geoff Thorpe <geoff@eu.c2.net>]
11447 *) Fix for 'req': it was adding a null to request attributes.
11448 Also change the X509_LOOKUP and X509_INFO code to handle
11449 certificate auxiliary information.
11452 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11456 *) Add the possibility to add extra information to the memory leak
11457 detecting output, to form tracebacks, showing from where each
11458 allocation was originated: CRYPTO_push_info("constant string") adds
11459 the string plus current file name and line number to a per-thread
11460 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11461 is like calling CYRPTO_pop_info() until the stack is empty.
11462 Also updated memory leak detection code to be multi-thread-safe.
11465 *) Add options -text and -noout to pkcs7 utility and delete the
11466 encryption options which never did anything. Update docs.
11469 *) Add options to some of the utilities to allow the pass phrase
11470 to be included on either the command line (not recommended on
11471 OSes like Unix) or read from the environment. Update the
11472 manpages and fix a few bugs.
11475 *) Add a few manpages for some of the openssl commands.
11478 *) Fix the -revoke option in ca. It was freeing up memory twice,
11479 leaking and not finding already revoked certificates.
11482 *) Extensive changes to support certificate auxiliary information.
11483 This involves the use of X509_CERT_AUX structure and X509_AUX
11484 functions. An X509_AUX function such as PEM_read_X509_AUX()
11485 can still read in a certificate file in the usual way but it
11486 will also read in any additional "auxiliary information". By
11487 doing things this way a fair degree of compatibility can be
11488 retained: existing certificates can have this information added
11489 using the new 'x509' options.
11491 Current auxiliary information includes an "alias" and some trust
11492 settings. The trust settings will ultimately be used in enhanced
11493 certificate chain verification routines: currently a certificate
11494 can only be trusted if it is self signed and then it is trusted
11498 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11499 The problem was that one of the replacement routines had not been working
11500 since SSLeay releases. For now the offending routine has been replaced
11501 with non-optimised assembler. Even so, this now gives around 95%
11502 performance improvement for 1024 bit RSA signs.
11505 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11506 handling. Most clients have the effective key size in bits equal to
11507 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11508 A few however don't do this and instead use the size of the decrypted key
11509 to determine the RC2 key length and the AlgorithmIdentifier to determine
11510 the effective key length. In this case the effective key length can still
11511 be 40 bits but the key length can be 168 bits for example. This is fixed
11512 by manually forcing an RC2 key into the EVP_PKEY structure because the
11513 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11514 the key length and effective key length are equal.
11517 *) Add a bunch of functions that should simplify the creation of
11518 X509_NAME structures. Now you should be able to do:
11519 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11520 and have it automatically work out the correct field type and fill in
11521 the structures. The more adventurous can try:
11522 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11523 and it will (hopefully) work out the correct multibyte encoding.
11526 *) Change the 'req' utility to use the new field handling and multibyte
11527 copy routines. Before the DN field creation was handled in an ad hoc
11528 way in req, ca, and x509 which was rather broken and didn't support
11529 BMPStrings or UTF8Strings. Since some software doesn't implement
11530 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11531 using the dirstring_type option. See the new comment in the default
11532 openssl.cnf for more info.
11535 *) Make crypto/rand/md_rand.c more robust:
11536 - Assure unique random numbers after fork().
11537 - Make sure that concurrent threads access the global counter and
11538 md serializably so that we never lose entropy in them
11539 or use exactly the same state in multiple threads.
11540 Access to the large state is not always serializable because
11541 the additional locking could be a performance killer, and
11542 md should be large enough anyway.
11545 *) New file apps/app_rand.c with commonly needed functionality
11546 for handling the random seed file.
11548 Use the random seed file in some applications that previously did not:
11550 dsaparam -genkey (which also ignored its '-rand' option),
11553 x509 (when signing).
11554 Except on systems with /dev/urandom, it is crucial to have a random
11555 seed file at least for key creation, DSA signing, and for DH exchanges;
11556 for RSA signatures we could do without one.
11558 gendh and gendsa (unlike genrsa) used to read only the first byte
11559 of each file listed in the '-rand' option. The function as previously
11560 found in genrsa is now in app_rand.c and is used by all programs
11561 that support '-rand'.
11564 *) In RAND_write_file, use mode 0600 for creating files;
11565 don't just chmod when it may be too late.
11568 *) Report an error from X509_STORE_load_locations
11569 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11572 *) New function ASN1_mbstring_copy() this copies a string in either
11573 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11574 into an ASN1_STRING type. A mask of permissible types is passed
11575 and it chooses the "minimal" type to use or an error if not type
11579 *) Add function equivalents to the various macros in asn1.h. The old
11580 macros are retained with an M_ prefix. Code inside the library can
11581 use the M_ macros. External code (including the openssl utility)
11582 should *NOT* in order to be "shared library friendly".
11585 *) Add various functions that can check a certificate's extensions
11586 to see if it usable for various purposes such as SSL client,
11587 server or S/MIME and CAs of these types. This is currently
11588 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11589 verification. Also added a -purpose flag to x509 utility to
11590 print out all the purposes.
11593 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11597 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11598 for, obtain and decode and extension and obtain its critical flag.
11599 This allows all the necessary extension code to be handled in a
11600 single function call.
11603 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11604 platforms. See crypto/rc4/rc4_enc.c for further details.
11607 *) New -noout option to asn1parse. This causes no output to be produced
11608 its main use is when combined with -strparse and -out to extract data
11609 from a file (which may not be in ASN.1 format).
11612 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11613 when producing the local key id.
11614 [Richard Levitte <levitte@stacken.kth.se>]
11616 *) New option -dhparam in s_server. This allows a DH parameter file to be
11617 stated explicitly. If it is not stated then it tries the first server
11618 certificate file. The previous behaviour hard coded the filename
11622 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11623 a public key to be input or output. For example:
11624 openssl rsa -in key.pem -pubout -out pubkey.pem
11625 Also added necessary DSA public key functions to handle this.
11628 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11629 in the message. This was handled by allowing
11630 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11631 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11633 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11634 to the end of the strings whereas this didn't. This would cause problems
11635 if strings read with d2i_ASN1_bytes() were later modified.
11636 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11638 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11639 data and it contains EOF it will end up returning an error. This is
11640 caused by input 46 bytes long. The cause is due to the way base64
11641 BIOs find the start of base64 encoded data. They do this by trying a
11642 trial decode on each line until they find one that works. When they
11643 do a flag is set and it starts again knowing it can pass all the
11644 data directly through the decoder. Unfortunately it doesn't reset
11645 the context it uses. This means that if EOF is reached an attempt
11646 is made to pass two EOFs through the context and this causes the
11647 resulting error. This can also cause other problems as well. As is
11648 usual with these problems it takes *ages* to find and the fix is
11649 trivial: move one line.
11650 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11652 *) Ugly workaround to get s_client and s_server working under Windows. The
11653 old code wouldn't work because it needed to select() on sockets and the
11654 tty (for keypresses and to see if data could be written). Win32 only
11655 supports select() on sockets so we select() with a 1s timeout on the
11656 sockets and then see if any characters are waiting to be read, if none
11657 are present then we retry, we also assume we can always write data to
11658 the tty. This isn't nice because the code then blocks until we've
11659 received a complete line of data and it is effectively polling the
11660 keyboard at 1s intervals: however it's quite a bit better than not
11661 working at all :-) A dedicated Windows application might handle this
11662 with an event loop for example.
11665 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11666 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11667 will be called when RSA_sign() and RSA_verify() are used. This is useful
11668 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11669 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11670 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11671 This necessitated the support of an extra signature type NID_md5_sha1
11672 for SSL signatures and modifications to the SSL library to use it instead
11673 of calling RSA_public_decrypt() and RSA_private_encrypt().
11676 *) Add new -verify -CAfile and -CApath options to the crl program, these
11677 will lookup a CRL issuers certificate and verify the signature in a
11678 similar way to the verify program. Tidy up the crl program so it
11679 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11680 less strict. It will now permit CRL extensions even if it is not
11681 a V2 CRL: this will allow it to tolerate some broken CRLs.
11684 *) Initialize all non-automatic variables each time one of the openssl
11685 sub-programs is started (this is necessary as they may be started
11686 multiple times from the "OpenSSL>" prompt).
11687 [Lennart Bang, Bodo Moeller]
11689 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11690 removing all other RSA functionality (this is what NO_RSA does). This
11691 is so (for example) those in the US can disable those operations covered
11692 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11696 *) Non-copying interface to BIO pairs.
11697 (still largely untested)
11700 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
11701 ASCII string. This was handled independently in various places before.
11704 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11705 UTF8 strings a character at a time.
11708 *) Use client_version from client hello to select the protocol
11709 (s23_srvr.c) and for RSA client key exchange verification
11710 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11713 *) Add various utility functions to handle SPKACs, these were previously
11714 handled by poking round in the structure internals. Added new function
11715 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11716 print, verify and generate SPKACs. Based on an original idea from
11717 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11720 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11723 *) Allow the config file extension section to be overwritten on the
11724 command line. Based on an original idea from Massimiliano Pala
11725 <madwolf@comune.modena.it>. The new option is called -extensions
11726 and can be applied to ca, req and x509. Also -reqexts to override
11727 the request extensions in req and -crlexts to override the crl extensions
11731 *) Add new feature to the SPKAC handling in ca. Now you can include
11732 the same field multiple times by preceding it by "XXXX." for example:
11735 this is the same syntax as used in the req config file.
11738 *) Allow certificate extensions to be added to certificate requests. These
11739 are specified in a 'req_extensions' option of the req section of the
11740 config file. They can be printed out with the -text option to req but
11741 are otherwise ignored at present.
11744 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11745 data read consists of only the final block it would not decrypted because
11746 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11747 A misplaced 'break' also meant the decrypted final block might not be
11748 copied until the next read.
11751 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11752 a few extra parameters to the DH structure: these will be useful if
11753 for example we want the value of 'q' or implement X9.42 DH.
11756 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11757 provides hooks that allow the default DSA functions or functions on a
11758 "per key" basis to be replaced. This allows hardware acceleration and
11759 hardware key storage to be handled without major modification to the
11760 library. Also added low level modexp hooks and CRYPTO_EX structure and
11761 associated functions.
11764 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11765 as "read only": it can't be written to and the buffer it points to will
11766 not be freed. Reading from a read only BIO is much more efficient than
11767 a normal memory BIO. This was added because there are several times when
11768 an area of memory needs to be read from a BIO. The previous method was
11769 to create a memory BIO and write the data to it, this results in two
11770 copies of the data and an O(n^2) reading algorithm. There is a new
11771 function BIO_new_mem_buf() which creates a read only memory BIO from
11772 an area of memory. Also modified the PKCS#7 routines to use read only
11776 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11777 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11778 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11779 but a retry condition occurred while trying to read the rest.
11782 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11783 NID_pkcs7_encrypted by default: this was wrong since this should almost
11784 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11785 the encrypted data type: this is a more sensible place to put it and it
11786 allows the PKCS#12 code to be tidied up that duplicated this
11790 *) Changed obj_dat.pl script so it takes its input and output files on
11791 the command line. This should avoid shell escape redirection problems
11795 *) Initial support for certificate extension requests, these are included
11796 in things like Xenroll certificate requests. Included functions to allow
11797 extensions to be obtained and added.
11800 *) -crlf option to s_client and s_server for sending newlines as
11801 CRLF (as required by many protocols).
11804 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11806 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11807 [Ralf S. Engelschall]
11809 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11810 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11812 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11816 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11817 DH parameters/keys (q is lost during that conversion, but the resulting
11818 DH parameters contain its length).
11820 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11821 much faster than DH_generate_parameters (which creates parameters
11822 where p = 2*q + 1), and also the smaller q makes DH computations
11823 much more efficient (160-bit exponentiation instead of 1024-bit
11824 exponentiation); so this provides a convenient way to support DHE
11825 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11826 utter importance to use
11827 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11829 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11830 when such DH parameters are used, because otherwise small subgroup
11831 attacks may become possible!
11834 *) Avoid memory leak in i2d_DHparams.
11837 *) Allow the -k option to be used more than once in the enc program:
11838 this allows the same encrypted message to be read by multiple recipients.
11841 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11842 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11843 it will always use the numerical form of the OID, even if it has a short
11847 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11848 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11849 otherwise bn_mod_exp was called. In the case of hardware keys for example
11850 no private key components need be present and it might store extra data
11851 in the RSA structure, which cannot be accessed from bn_mod_exp.
11852 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11853 private key operations.
11856 *) Added support for SPARC Linux.
11859 *) pem_password_cb function type incompatibly changed from
11860 typedef int pem_password_cb(char *buf, int size, int rwflag);
11862 ....(char *buf, int size, int rwflag, void *userdata);
11863 so that applications can pass data to their callbacks:
11864 The PEM[_ASN1]_{read,write}... functions and macros now take an
11865 additional void * argument, which is just handed through whenever
11866 the password callback is called.
11867 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11869 New function SSL_CTX_set_default_passwd_cb_userdata.
11871 Compatibility note: As many C implementations push function arguments
11872 onto the stack in reverse order, the new library version is likely to
11873 interoperate with programs that have been compiled with the old
11874 pem_password_cb definition (PEM_whatever takes some data that
11875 happens to be on the stack as its last argument, and the callback
11876 just ignores this garbage); but there is no guarantee whatsoever that
11879 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11880 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11881 problems not only on Windows, but also on some Unix platforms.
11882 To avoid problematic command lines, these definitions are now in an
11883 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11884 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11887 *) MIPS III/IV assembler module is reimplemented.
11890 *) More DES library cleanups: remove references to srand/rand and
11891 delete an unused file.
11894 *) Add support for the free Netwide assembler (NASM) under Win32,
11895 since not many people have MASM (ml) and it can be hard to obtain.
11896 This is currently experimental but it seems to work OK and pass all
11897 the tests. Check out INSTALL.W32 for info.
11900 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11901 without temporary keys kept an extra copy of the server key,
11902 and connections with temporary keys did not free everything in case
11906 *) New function RSA_check_key and new openssl rsa option -check
11907 for verifying the consistency of RSA keys.
11908 [Ulf Moeller, Bodo Moeller]
11910 *) Various changes to make Win32 compile work:
11911 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11912 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11913 comparison" warnings.
11914 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11917 *) Add a debugging option to PKCS#5 v2 key generation function: when
11918 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11919 derived keys are printed to stderr.
11922 *) Copy the flags in ASN1_STRING_dup().
11923 [Roman E. Pavlov <pre@mo.msk.ru>]
11925 *) The x509 application mishandled signing requests containing DSA
11926 keys when the signing key was also DSA and the parameters didn't match.
11928 It was supposed to omit the parameters when they matched the signing key:
11929 the verifying software was then supposed to automatically use the CA's
11930 parameters if they were absent from the end user certificate.
11932 Omitting parameters is no longer recommended. The test was also
11933 the wrong way round! This was probably due to unusual behaviour in
11934 EVP_cmp_parameters() which returns 1 if the parameters match.
11935 This meant that parameters were omitted when they *didn't* match and
11936 the certificate was useless. Certificates signed with 'ca' didn't have
11938 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11940 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11941 The interface is as follows:
11942 Applications can use
11943 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11944 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11945 "off" is now the default.
11946 The library internally uses
11947 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11948 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11949 to disable memory-checking temporarily.
11951 Some inconsistent states that previously were possible (and were
11952 even the default) are now avoided.
11954 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11955 with each memory chunk allocated; this is occasionally more helpful
11956 than just having a counter.
11958 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11960 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11964 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11965 which largely parallels "options", but is for changing API behaviour,
11966 whereas "options" are about protocol behaviour.
11967 Initial "mode" flags are:
11969 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11970 a single record has been written.
11971 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11972 retries use the same buffer location.
11973 (But all of the contents must be
11977 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11980 *) Fix problems with no-hmac etc.
11981 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11983 *) New functions RSA_get_default_method(), RSA_set_method() and
11984 RSA_get_method(). These allows replacement of RSA_METHODs without having
11985 to mess around with the internals of an RSA structure.
11988 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11989 Also really enable memory leak checks in openssl.c and in some
11991 [Chad C. Mulligan, Bodo Moeller]
11993 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11994 up the length of negative integers. This has now been simplified to just
11995 store the length when it is first determined and use it later, rather
11996 than trying to keep track of where data is copied and updating it to
11998 [Steve Henson, reported by Brien Wheeler
11999 <bwheeler@authentica-security.com>]
12001 *) Add a new function PKCS7_signatureVerify. This allows the verification
12002 of a PKCS#7 signature but with the signing certificate passed to the
12003 function itself. This contrasts with PKCS7_dataVerify which assumes the
12004 certificate is present in the PKCS#7 structure. This isn't always the
12005 case: certificates can be omitted from a PKCS#7 structure and be
12006 distributed by "out of band" means (such as a certificate database).
12009 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12010 function prototypes in pem.h, also change util/mkdef.pl to add the
12011 necessary function names.
12014 *) mk1mf.pl (used by Windows builds) did not properly read the
12015 options set by Configure in the top level Makefile, and Configure
12016 was not even able to write more than one option correctly.
12017 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12020 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12021 file to be loaded from a BIO or FILE pointer. The BIO version will
12022 for example allow memory BIOs to contain config info.
12025 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12026 Whoever hopes to achieve shared-library compatibility across versions
12027 must use this, not the compile-time macro.
12028 (Exercise 0.9.4: Which is the minimum library version required by
12030 Note: All this applies only to multi-threaded programs, others don't
12034 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12035 through a BIO pair triggered the default case, i.e.
12036 SSLerr(...,SSL_R_UNKNOWN_STATE).
12039 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12040 can use the SSL library even if none of the specific BIOs is
12044 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12045 for the encoded length.
12046 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12048 *) Add initial documentation of the X509V3 functions.
12051 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12052 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12053 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12054 secure PKCS#8 private key format with a high iteration count.
12057 *) Fix determination of Perl interpreter: A perl or perl5
12058 _directory_ in $PATH was also accepted as the interpreter.
12059 [Ralf S. Engelschall]
12061 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12062 wrong with it but it was very old and did things like calling
12063 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12064 unusual formatting.
12067 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12068 to use the new extension code.
12071 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12072 with macros. This should make it easier to change their form, add extra
12073 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12077 *) Add to configuration table a new entry that can specify an alternative
12078 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12079 according to Mark Crispin <MRC@Panda.COM>.
12083 *) DES CBC did not update the IV. Weird.
12086 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12087 Changing the behaviour of the former might break existing programs --
12088 where IV updating is needed, des_ncbc_encrypt can be used.
12091 *) When bntest is run from "make test" it drives bc to check its
12092 calculations, as well as internally checking them. If an internal check
12093 fails, it needs to cause bc to give a non-zero result or make test carries
12094 on without noticing the failure. Fixed.
12097 *) DES library cleanups.
12100 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12101 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12102 ciphers. NOTE: although the key derivation function has been verified
12103 against some published test vectors it has not been extensively tested
12104 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12108 *) Instead of "mkdir -p", which is not fully portable, use new
12109 Perl script "util/mkdir-p.pl".
12112 *) Rewrite the way password based encryption (PBE) is handled. It used to
12113 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12114 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12115 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12116 the 'parameter' field of the AlgorithmIdentifier is passed to the
12117 underlying key generation function so it must do its own ASN1 parsing.
12118 This has also changed the EVP_PBE_CipherInit() function which now has a
12119 'parameter' argument instead of literal salt and iteration count values
12120 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12123 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12124 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12125 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12126 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12127 value was just used as a "magic string" and not used directly its
12128 value doesn't matter.
12131 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12135 *) "linux-sparc64" configuration (ultrapenguin).
12136 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12137 "linux-sparc" configuration.
12138 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12140 *) config now generates no-xxx options for missing ciphers.
12143 *) Support the EBCDIC character set (work in progress).
12144 File ebcdic.c not yet included because it has a different license.
12145 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12147 *) Support BS2000/OSD-POSIX.
12148 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12150 *) Make callbacks for key generation use void * instead of char *.
12153 *) Make S/MIME samples compile (not yet tested).
12156 *) Additional typesafe stacks.
12159 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12163 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12165 *) New configuration variant "sco5-gcc".
12167 *) Updated some demos.
12168 [Sean O Riordain, Wade Scholine]
12170 *) Add missing BIO_free at exit of pkcs12 application.
12173 *) Fix memory leak in conf.c.
12176 *) Updates for Win32 to assembler version of MD5.
12179 *) Set #! path to perl in apps/der_chop to where we found it
12180 instead of using a fixed path.
12183 *) SHA library changes for irix64-mips4-cc.
12186 *) Improvements for VMS support.
12190 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12192 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12193 This also avoids the problems with SC4.2 and unpatched SC5.
12194 [Andy Polyakov <appro@fy.chalmers.se>]
12196 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12197 These are required because of the typesafe stack would otherwise break
12198 existing code. If old code used a structure member which used to be STACK
12199 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12200 sk_num or sk_value it would produce an error because the num, data members
12201 are not present in STACK_OF. Now it just produces a warning. sk_set
12202 replaces the old method of assigning a value to sk_value
12203 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12204 that does this will no longer work (and should use sk_set instead) but
12205 this could be regarded as a "questionable" behaviour anyway.
12208 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12209 correctly handle encrypted S/MIME data.
12212 *) Change type of various DES function arguments from des_cblock
12213 (which means, in function argument declarations, pointer to char)
12214 to des_cblock * (meaning pointer to array with 8 char elements),
12215 which allows the compiler to do more typechecking; it was like
12216 that back in SSLeay, but with lots of ugly casts.
12218 Introduce new type const_des_cblock.
12221 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12222 problems: find RecipientInfo structure that matches recipient certificate
12223 and initialise the ASN1 structures properly based on passed cipher.
12226 *) Belatedly make the BN tests actually check the results.
12229 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12230 to and from BNs: it was completely broken. New compilation option
12231 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12232 key elements as negative integers.
12235 *) Reorganize and speed up MD5.
12236 [Andy Polyakov <appro@fy.chalmers.se>]
12239 [Richard Levitte <richard@levitte.org>]
12241 *) New option -out to asn1parse to allow the parsed structure to be
12242 output to a file. This is most useful when combined with the -strparse
12243 option to examine the output of things like OCTET STRINGS.
12246 *) Make SSL library a little more fool-proof by not requiring any longer
12247 that SSL_set_{accept,connect}_state be called before
12248 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12249 in many applications because usually everything *appeared* to work as
12250 intended anyway -- now it really works as intended).
12253 *) Move openssl.cnf out of lib/.
12256 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12257 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12258 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12259 [Ralf S. Engelschall]
12261 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12262 handle PKCS#7 enveloped data properly.
12263 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12265 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12266 copying pointers. The cert_st handling is changed by this in
12267 various ways (and thus what used to be known as ctx->default_cert
12268 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12269 any longer when s->cert does not give us what we need).
12270 ssl_cert_instantiate becomes obsolete by this change.
12271 As soon as we've got the new code right (possibly it already is?),
12272 we have solved a couple of bugs of the earlier code where s->cert
12273 was used as if it could not have been shared with other SSL structures.
12275 Note that using the SSL API in certain dirty ways now will result
12276 in different behaviour than observed with earlier library versions:
12277 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12278 does not influence s as it used to.
12280 In order to clean up things more thoroughly, inside SSL_SESSION
12281 we don't use CERT any longer, but a new structure SESS_CERT
12282 that holds per-session data (if available); currently, this is
12283 the peer's certificate chain and, for clients, the server's certificate
12284 and temporary key. CERT holds only those values that can have
12285 meaningful defaults in an SSL_CTX.
12288 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12289 from the internal representation. Various PKCS#7 fixes: remove some
12290 evil casts and set the enc_dig_alg field properly based on the signing
12294 *) Allow PKCS#12 password to be set from the command line or the
12295 environment. Let 'ca' get its config file name from the environment
12296 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12300 *) Allow certificate policies extension to use an IA5STRING for the
12301 organization field. This is contrary to the PKIX definition but
12302 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12306 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12307 without disallowing inline assembler and the like for non-pedantic builds.
12310 *) Support Borland C++ builder.
12311 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12313 *) Support Mingw32.
12316 *) SHA-1 cleanups and performance enhancements.
12317 [Andy Polyakov <appro@fy.chalmers.se>]
12319 *) Sparc v8plus assembler for the bignum library.
12320 [Andy Polyakov <appro@fy.chalmers.se>]
12322 *) Accept any -xxx and +xxx compiler options in Configure.
12325 *) Update HPUX configuration.
12328 *) Add missing sk_<type>_unshift() function to safestack.h
12329 [Ralf S. Engelschall]
12331 *) New function SSL_CTX_use_certificate_chain_file that sets the
12332 "extra_cert"s in addition to the certificate. (This makes sense
12333 only for "PEM" format files, as chains as a whole are not
12337 *) Support verify_depth from the SSL API.
12338 x509_vfy.c had what can be considered an off-by-one-error:
12339 Its depth (which was not part of the external interface)
12340 was actually counting the number of certificates in a chain;
12341 now it really counts the depth.
12344 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12345 instead of X509err, which often resulted in confusing error
12346 messages since the error codes are not globally unique
12347 (e.g. an alleged error in ssl3_accept when a certificate
12348 didn't match the private key).
12350 *) New function SSL_CTX_set_session_id_context that allows to set a default
12351 value (so that you don't need SSL_set_session_id_context for each
12352 connection using the SSL_CTX).
12355 *) OAEP decoding bug fix.
12358 *) Support INSTALL_PREFIX for package builders, as proposed by
12362 *) New Configure options "threads" and "no-threads". For systems
12363 where the proper compiler options are known (currently Solaris
12364 and Linux), "threads" is the default.
12367 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12370 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12371 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12372 such as /usr/local/bin.
12375 *) "make linux-shared" to build shared libraries.
12376 [Niels Poppe <niels@netbox.org>]
12378 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12381 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12382 extension adding in x509 utility.
12385 *) Remove NOPROTO sections and error code comments.
12388 *) Partial rewrite of the DEF file generator to now parse the ANSI
12392 *) New Configure options --prefix=DIR and --openssldir=DIR.
12395 *) Complete rewrite of the error code script(s). It is all now handled
12396 by one script at the top level which handles error code gathering,
12397 header rewriting and C source file generation. It should be much better
12398 than the old method: it now uses a modified version of Ulf's parser to
12399 read the ANSI prototypes in all header files (thus the old K&R definitions
12400 aren't needed for error creation any more) and do a better job of
12401 translating function codes into names. The old 'ASN1 error code embedded
12402 in a comment' is no longer necessary and it doesn't use .err files which
12403 have now been deleted. Also the error code call doesn't have to appear all
12404 on one line (which resulted in some large lines...).
12407 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12410 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12411 0 (which usually indicates a closed connection), but continue reading.
12414 *) Fix some race conditions.
12417 *) Add support for CRL distribution points extension. Add Certificate
12418 Policies and CRL distribution points documentation.
12421 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12424 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12425 8 of keying material. Merlin has also confirmed interop with this fix
12426 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12427 [Merlin Hughes <merlin@baltimore.ie>]
12429 *) Fix lots of warnings.
12430 [Richard Levitte <levitte@stacken.kth.se>]
12432 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12433 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12434 [Richard Levitte <levitte@stacken.kth.se>]
12436 *) Fix problems with sizeof(long) == 8.
12437 [Andy Polyakov <appro@fy.chalmers.se>]
12439 *) Change functions to ANSI C.
12442 *) Fix typos in error codes.
12443 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12445 *) Remove defunct assembler files from Configure.
12448 *) SPARC v8 assembler BIGNUM implementation.
12449 [Andy Polyakov <appro@fy.chalmers.se>]
12451 *) Support for Certificate Policies extension: both print and set.
12452 Various additions to support the r2i method this uses.
12455 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12456 return a const string when you are expecting an allocated buffer.
12459 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12460 types DirectoryString and DisplayText.
12463 *) Add code to allow r2i extensions to access the configuration database,
12464 add an LHASH database driver and add several ctx helper functions.
12467 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12468 fail when they extended the size of a BIGNUM.
12471 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12472 support typesafe stack.
12475 *) Fix typo in SSL_[gs]et_options().
12476 [Nils Frostberg <nils@medcom.se>]
12478 *) Delete various functions and files that belonged to the (now obsolete)
12479 old X509V3 handling code.
12482 *) New Configure option "rsaref".
12485 *) Don't auto-generate pem.h.
12488 *) Introduce type-safe ASN.1 SETs.
12491 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12492 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12494 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12495 that links with OpenSSL (well at least cause lots of warnings), but fear
12496 not: the conversion is trivial, and it eliminates loads of evil casts. A
12497 few STACKed things have been converted already. Feel free to convert more.
12498 In the fullness of time, I'll do away with the STACK type altogether.
12501 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12502 specified in <certfile> by updating the entry in the index.txt file.
12503 This way one no longer has to edit the index.txt file manually for
12504 revoking a certificate. The -revoke option does the gory details now.
12505 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12507 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12508 `-text' option at all and this way the `-noout -text' combination was
12509 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12510 [Ralf S. Engelschall]
12512 *) Make sure a corresponding plain text error message exists for the
12513 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12514 verify callback function determined that a certificate was revoked.
12515 [Ralf S. Engelschall]
12517 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12518 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12519 all available ciphers including rc5, which was forgotten until now.
12520 In order to let the testing shell script know which algorithms
12521 are available, a new (up to now undocumented) command
12522 "openssl list-cipher-commands" is used.
12525 *) Bugfix: s_client occasionally would sleep in select() when
12526 it should have checked SSL_pending() first.
12529 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12530 the raw DSA values prior to ASN.1 encoding.
12533 *) Tweaks to Configure
12534 [Niels Poppe <niels@netbox.org>]
12536 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12540 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12543 *) New config option to avoid instructions that are illegal on the 80386.
12544 The default code is faster, but requires at least a 486.
12547 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12548 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12549 same as SSL2_VERSION anyway.
12552 *) New "-showcerts" option for s_client.
12555 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12556 application. Various cleanups and fixes.
12559 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12560 modify error routines to work internally. Add error codes and PBE init
12561 to library startup routines.
12564 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12565 packing functions to asn1 and evp. Changed function names and error
12566 codes along the way.
12569 *) PKCS12 integration: and so it begins... First of several patches to
12570 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12571 objects to objects.h
12574 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12575 and display support for Thawte strong extranet extension.
12578 *) Add LinuxPPC support.
12579 [Jeff Dubrule <igor@pobox.org>]
12581 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12582 bn_div_words in alpha.s.
12583 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12585 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12586 OAEP isn't supported when OpenSSL is built with RSAref.
12587 [Ulf Moeller <ulf@fitug.de>]
12589 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12590 so they no longer are missing under -DNOPROTO.
12591 [Soren S. Jorvang <soren@t.dk>]
12594 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12596 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12597 doesn't work when the session is reused. Coming soon!
12600 *) Fix a security hole, that allows sessions to be reused in the wrong
12601 context thus bypassing client cert protection! All software that uses
12602 client certs and session caches in multiple contexts NEEDS PATCHING to
12603 allow session reuse! A fuller solution is in the works.
12604 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12606 *) Some more source tree cleanups (removed obsolete files
12607 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12608 permission on "config" script to be executable) and a fix for the INSTALL
12610 [Ulf Moeller <ulf@fitug.de>]
12612 *) Remove some legacy and erroneous uses of malloc, free instead of
12614 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12616 *) Make rsa_oaep_test return non-zero on error.
12617 [Ulf Moeller <ulf@fitug.de>]
12619 *) Add support for native Solaris shared libraries. Configure
12620 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12621 if someone would make that last step automatic.
12622 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12624 *) ctx_size was not built with the right compiler during "make links". Fixed.
12627 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12628 except NULL ciphers". This means the default cipher list will no longer
12629 enable NULL ciphers. They need to be specifically enabled e.g. with
12630 the string "DEFAULT:eNULL".
12633 *) Fix to RSA private encryption routines: if p < q then it would
12634 occasionally produce an invalid result. This will only happen with
12635 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12638 *) Be less restrictive and allow also `perl util/perlpath.pl
12639 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12640 because this way one can also use an interpreter named `perl5' (which is
12641 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12642 installed as `perl').
12643 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12645 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12646 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12648 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12649 advapi32.lib to Win32 build and change the pem test comparison
12650 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12651 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12652 and crypto/des/ede_cbcm_enc.c.
12655 *) DES quad checksum was broken on big-endian architectures. Fixed.
12658 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12659 Win32 test batch file so it (might) work again. The Win32 test batch file
12660 is horrible: I feel ill....
12663 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12664 in e_os.h. Audit of header files to check ANSI and non ANSI
12665 sections: 10 functions were absent from non ANSI section and not exported
12666 from Windows DLLs. Fixed up libeay.num for new functions.
12669 *) Make `openssl version' output lines consistent.
12670 [Ralf S. Engelschall]
12672 *) Fix Win32 symbol export lists for BIO functions: Added
12673 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12674 to ms/libeay{16,32}.def.
12675 [Ralf S. Engelschall]
12677 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12678 fine under Unix and passes some trivial tests I've now added. But the
12679 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12680 added to make sure no one expects that this stuff really works in the
12681 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12682 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12684 [Ralf S. Engelschall]
12686 *) Fix the generation of two part addresses in perl.
12687 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12689 *) Add config entry for Linux on MIPS.
12690 [John Tobey <jtobey@channel1.com>]
12692 *) Make links whenever Configure is run, unless we are on Windoze.
12695 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12696 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12700 *) Add a useful kludge to allow package maintainers to specify compiler and
12701 other platforms details on the command line without having to patch the
12702 Configure script every time: One now can use ``perl Configure
12703 <id>:<details>'', i.e. platform ids are allowed to have details appended
12704 to them (separated by colons). This is treated as there would be a static
12705 pre-configured entry in Configure's %table under key <id> with value
12706 <details> and ``perl Configure <id>'' is called. So, when you want to
12707 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12708 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12709 now, which overrides the FreeBSD-elf entry on-the-fly.
12710 [Ralf S. Engelschall]
12712 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12715 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12716 on the `perl Configure ...' command line. This way one can compile
12717 OpenSSL libraries with Position Independent Code (PIC) which is needed
12718 for linking it into DSOs.
12719 [Ralf S. Engelschall]
12721 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12725 *) Cleaned up the LICENSE document: The official contact for any license
12726 questions now is the OpenSSL core team under openssl-core@openssl.org.
12727 And add a paragraph about the dual-license situation to make sure people
12728 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12729 to the OpenSSL toolkit.
12730 [Ralf S. Engelschall]
12732 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12733 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12734 Additionally cleaned up the `make links' target: Remove unnecessary
12735 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12736 to speed processing and no longer clutter the display with confusing
12737 stuff. Instead only the actually done links are displayed.
12738 [Ralf S. Engelschall]
12740 *) Permit null encryption ciphersuites, used for authentication only. It used
12741 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12742 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12746 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12747 signed attributes when verifying signatures (this would break them),
12748 the detached data encoding was wrong and public keys obtained using
12749 X509_get_pubkey() weren't freed.
12752 *) Add text documentation for the BUFFER functions. Also added a work around
12753 to a Win95 console bug. This was triggered by the password read stuff: the
12754 last character typed gets carried over to the next fread(). If you were
12755 generating a new cert request using 'req' for example then the last
12756 character of the passphrase would be CR which would then enter the first
12760 *) Added the new `Includes OpenSSL Cryptography Software' button as
12761 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12762 button and can be used by applications based on OpenSSL to show the
12763 relationship to the OpenSSL project.
12764 [Ralf S. Engelschall]
12766 *) Remove confusing variables in function signatures in files
12767 ssl/ssl_lib.c and ssl/ssl.h.
12768 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12770 *) Don't install bss_file.c under PREFIX/include/
12771 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12773 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12774 functions that return function pointers and has support for NT specific
12775 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12776 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12777 unsigned to signed types: this was killing the Win32 compile.
12780 *) Add new certificate file to stack functions,
12781 SSL_add_dir_cert_subjects_to_stack() and
12782 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12783 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12784 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12785 This means that Apache-SSL and similar packages don't have to mess around
12786 to add as many CAs as they want to the preferred list.
12789 *) Experiment with doxygen documentation. Currently only partially applied to
12791 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12792 openssl.doxy as the configuration file.
12795 *) Get rid of remaining C++-style comments which strict C compilers hate.
12796 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12798 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12799 compiled in by default: it has problems with large keys.
12802 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12803 DH private keys and/or callback functions which directly correspond to
12804 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12805 is needed for applications which have to configure certificates on a
12806 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12808 For the RSA certificate situation is makes no difference, but
12809 for the DSA certificate situation this fixes the "no shared cipher"
12810 problem where the OpenSSL cipher selection procedure failed because the
12811 temporary keys were not overtaken from the context and the API provided
12812 no way to reconfigure them.
12813 The new functions now let applications reconfigure the stuff and they
12814 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12815 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12816 non-public-API function ssl_cert_instantiate() is used as a helper
12817 function and also to reduce code redundancy inside ssl_rsa.c.
12818 [Ralf S. Engelschall]
12820 *) Move s_server -dcert and -dkey options out of the undocumented feature
12821 area because they are useful for the DSA situation and should be
12822 recognized by the users.
12823 [Ralf S. Engelschall]
12825 *) Fix the cipher decision scheme for export ciphers: the export bits are
12826 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12827 SSL_EXP_MASK. So, the original variable has to be used instead of the
12828 already masked variable.
12829 [Richard Levitte <levitte@stacken.kth.se>]
12831 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12832 [Richard Levitte <levitte@stacken.kth.se>]
12834 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12835 from `int' to `unsigned int' because it's a length and initialized by
12836 EVP_DigestFinal() which expects an `unsigned int *'.
12837 [Richard Levitte <levitte@stacken.kth.se>]
12839 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12840 script. Instead use the usual Shell->Perl transition trick.
12841 [Ralf S. Engelschall]
12843 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12844 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12845 -noout -modulus' as it's already the case for `openssl rsa -noout
12846 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12847 currently the public key is printed (a decision which was already done by
12848 `openssl dsa -modulus' in the past) which serves a similar purpose.
12849 Additionally the NO_RSA no longer completely removes the whole -modulus
12850 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12852 [Ralf S. Engelschall]
12854 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12855 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12856 [Arne Ansper <arne@ats.cyber.ee>]
12858 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12859 to be added. Now both 'req' and 'ca' can use new objects defined in the
12863 *) Add cool BIO that does syslog (or event log on NT).
12864 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12866 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12867 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12868 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12869 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12872 *) Add preliminary config info for new extension code.
12875 *) Make RSA_NO_PADDING really use no padding.
12876 [Ulf Moeller <ulf@fitug.de>]
12878 *) Generate errors when private/public key check is done.
12881 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12882 for some CRL extensions and new objects added.
12885 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12886 key usage extension and fuller support for authority key id.
12889 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12890 padding method for RSA, which is recommended for new applications in PKCS
12891 #1 v2.0 (RFC 2437, October 1998).
12892 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12893 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12894 against Bleichbacher's attack on RSA.
12895 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12898 *) Updates to the new SSL compression code
12899 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12901 *) Fix so that the version number in the master secret, when passed
12902 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12903 (because the server will not accept higher), that the version number
12904 is 0x03,0x01, not 0x03,0x00
12905 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12907 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12908 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12909 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12912 *) Support for RAW extensions where an arbitrary extension can be
12913 created by including its DER encoding. See apps/openssl.cnf for
12917 *) Make sure latest Perl versions don't interpret some generated C array
12918 code as Perl array code in the crypto/err/err_genc.pl script.
12919 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12921 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12922 not many people have the assembler. Various Win32 compilation fixes and
12923 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12924 build instructions.
12927 *) Modify configure script 'Configure' to automatically create crypto/date.h
12928 file under Win32 and also build pem.h from pem.org. New script
12929 util/mkfiles.pl to create the MINFO file on environments that can't do a
12930 'make files': perl util/mkfiles.pl >MINFO should work.
12933 *) Major rework of DES function declarations, in the pursuit of correctness
12934 and purity. As a result, many evil casts evaporated, and some weirdness,
12935 too. You may find this causes warnings in your code. Zapping your evil
12936 casts will probably fix them. Mostly.
12939 *) Fix for a typo in asn1.h. Bug fix to object creation script
12940 obj_dat.pl. It considered a zero in an object definition to mean
12941 "end of object": none of the objects in objects.h have any zeros
12942 so it wasn't spotted.
12943 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12945 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12946 Masking (CBCM). In the absence of test vectors, the best I have been able
12947 to do is check that the decrypt undoes the encrypt, so far. Send me test
12948 vectors if you have them.
12951 *) Correct calculation of key length for export ciphers (too much space was
12952 allocated for null ciphers). This has not been tested!
12955 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12956 message is now correct (it understands "crypto" and "ssl" on its
12957 command line). There is also now an "update" option. This will update
12958 the util/ssleay.num and util/libeay.num files with any new functions.
12960 perl util/mkdef.pl crypto ssl update
12961 it will update them.
12964 *) Overhauled the Perl interface (perl/*):
12965 - ported BN stuff to OpenSSL's different BN library
12966 - made the perl/ source tree CVS-aware
12967 - renamed the package from SSLeay to OpenSSL (the files still contain
12968 their history because I've copied them in the repository)
12969 - removed obsolete files (the test scripts will be replaced
12970 by better Test::Harness variants in the future)
12971 [Ralf S. Engelschall]
12973 *) First cut for a very conservative source tree cleanup:
12974 1. merge various obsolete readme texts into doc/ssleay.txt
12975 where we collect the old documents and readme texts.
12976 2. remove the first part of files where I'm already sure that we no
12977 longer need them because of three reasons: either they are just temporary
12978 files which were left by Eric or they are preserved original files where
12979 I've verified that the diff is also available in the CVS via "cvs diff
12980 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12981 the crypto/md/ stuff).
12982 [Ralf S. Engelschall]
12984 *) More extension code. Incomplete support for subject and issuer alt
12985 name, issuer and authority key id. Change the i2v function parameters
12986 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12987 what that's for :-) Fix to ASN1 macro which messed up
12988 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12991 *) Preliminary support for ENUMERATED type. This is largely copied from the
12995 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12996 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12998 *) Make sure `make rehash' target really finds the `openssl' program.
12999 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13001 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
13002 like to hear about it if this slows down other processors.
13005 *) Add CygWin32 platform information to Configure script.
13006 [Alan Batie <batie@aahz.jf.intel.com>]
13008 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
13009 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13011 *) New program nseq to manipulate netscape certificate sequences
13014 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13018 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13019 but the BN code had some problems that would cause failures when
13020 doing certificate verification and some other functions.
13021 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13023 *) Add ASN1 and PEM code to support netscape certificate sequences.
13026 *) Add ASN1 and PEM code to support netscape certificate sequences.
13029 *) Add several PKIX and private extended key usage OIDs.
13032 *) Modify the 'ca' program to handle the new extension code. Modify
13033 openssl.cnf for new extension format, add comments.
13036 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13037 and add a sample to openssl.cnf so req -x509 now adds appropriate
13041 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13042 error code, add initial support to X509_print() and x509 application.
13045 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13046 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13047 stuff is currently isolated and isn't even compiled yet.
13050 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13051 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13052 Removed the versions check from X509 routines when loading extensions:
13053 this allows certain broken certificates that don't set the version
13054 properly to be processed.
13057 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13058 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13059 can still be regenerated with "make depend".
13062 *) Spelling mistake in C version of CAST-128.
13063 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13065 *) Changes to the error generation code. The perl script err-code.pl
13066 now reads in the old error codes and retains the old numbers, only
13067 adding new ones if necessary. It also only changes the .err files if new
13068 codes are added. The makefiles have been modified to only insert errors
13069 when needed (to avoid needlessly modifying header files). This is done
13070 by only inserting errors if the .err file is newer than the auto generated
13071 C file. To rebuild all the error codes from scratch (the old behaviour)
13072 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13073 or delete all the .err files.
13076 *) CAST-128 was incorrectly implemented for short keys. The C version has
13077 been fixed, but is untested. The assembler versions are also fixed, but
13078 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13079 to regenerate it if needed.
13080 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13081 Hagino <itojun@kame.net>]
13083 *) File was opened incorrectly in randfile.c.
13084 [Ulf Möller <ulf@fitug.de>]
13086 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13087 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13088 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13089 al: it's just almost always a UTCTime. Note this patch adds new error
13090 codes so do a "make errors" if there are problems.
13093 *) Correct Linux 1 recognition in config.
13094 [Ulf Möller <ulf@fitug.de>]
13096 *) Remove pointless MD5 hash when using DSA keys in ca.
13097 [Anonymous <nobody@replay.com>]
13099 *) Generate an error if given an empty string as a cert directory. Also
13100 generate an error if handed NULL (previously returned 0 to indicate an
13101 error, but didn't set one).
13102 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13104 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13107 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13108 parameters. This was causing a warning which killed off the Win32 compile.
13111 *) Remove C++ style comments from crypto/bn/bn_local.h.
13112 [Neil Costigan <neil.costigan@celocom.com>]
13114 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13115 based on a text string, looking up short and long names and finally
13116 "dot" format. The "dot" format stuff didn't work. Added new function
13117 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13118 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13119 OID is not part of the table.
13122 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13123 X509_LOOKUP_by_alias().
13126 *) Sort openssl functions by name.
13129 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13130 encryption from sample DSA keys (in case anyone is interested the password
13134 *) Make _all_ *_free functions accept a NULL pointer.
13135 [Frans Heymans <fheymans@isaserver.be>]
13137 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13139 [Anonymous <nobody@replay.com>]
13141 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13142 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13144 *) Don't blow it for numeric -newkey arguments to apps/req.
13145 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13147 *) Temp key "for export" tests were wrong in s3_srvr.c.
13148 [Anonymous <nobody@replay.com>]
13150 *) Add prototype for temp key callback functions
13151 SSL_CTX_set_tmp_{rsa,dh}_callback().
13154 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13155 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13158 *) X509_name_add_entry() freed the wrong thing after an error.
13159 [Arne Ansper <arne@ats.cyber.ee>]
13161 *) rsa_eay.c would attempt to free a NULL context.
13162 [Arne Ansper <arne@ats.cyber.ee>]
13164 *) BIO_s_socket() had a broken should_retry() on Windoze.
13165 [Arne Ansper <arne@ats.cyber.ee>]
13167 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13168 [Arne Ansper <arne@ats.cyber.ee>]
13170 *) Make sure the already existing X509_STORE->depth variable is initialized
13171 in X509_STORE_new(), but document the fact that this variable is still
13172 unused in the certificate verification process.
13173 [Ralf S. Engelschall]
13175 *) Fix the various library and apps files to free up pkeys obtained from
13176 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13179 *) Fix reference counting in X509_PUBKEY_get(). This makes
13180 demos/maurice/example2.c work, amongst others, probably.
13181 [Steve Henson and Ben Laurie]
13183 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13184 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13185 are no longer created. This way we have a single and consistent command
13186 line interface `openssl <command>', similar to `cvs <command>'.
13187 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13189 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13190 BIT STRING wrapper always have zero unused bits.
13193 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13196 *) Make the top-level INSTALL documentation easier to understand.
13199 *) Makefiles updated to exit if an error occurs in a sub-directory
13200 make (including if user presses ^C) [Paul Sutton]
13202 *) Make Montgomery context stuff explicit in RSA data structure.
13205 *) Fix build order of pem and err to allow for generated pem.h.
13208 *) Fix renumbering bug in X509_NAME_delete_entry().
13211 *) Enhanced the err-ins.pl script so it makes the error library number
13212 global and can add a library name. This is needed for external ASN1 and
13213 other error libraries.
13216 *) Fixed sk_insert which never worked properly.
13219 *) Fix ASN1 macros so they can handle indefinite length constructed
13220 EXPLICIT tags. Some non standard certificates use these: they can now
13224 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13225 into a single doc/ssleay.txt bundle. This way the information is still
13226 preserved but no longer messes up this directory. Now it's new room for
13227 the new set of documentation files.
13228 [Ralf S. Engelschall]
13230 *) SETs were incorrectly DER encoded. This was a major pain, because they
13231 shared code with SEQUENCEs, which aren't coded the same. This means that
13232 almost everything to do with SETs or SEQUENCEs has either changed name or
13233 number of arguments.
13234 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13236 *) Fix test data to work with the above.
13239 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13240 was already fixed by Eric for 0.9.1 it seems.
13241 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13243 *) Autodetect FreeBSD3.
13246 *) Fix various bugs in Configure. This affects the following platforms:
13250 unixware-2.0-pentium
13254 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13255 before they are needed.
13258 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13262 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13264 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13265 changed SSLeay to OpenSSL in version strings.
13266 [Ralf S. Engelschall]
13268 *) Some fixups to the top-level documents.
13271 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13272 because the symlink to include/ was missing.
13273 [Ralf S. Engelschall]
13275 *) Incorporated the popular no-RSA/DSA-only patches
13276 which allow to compile a RSA-free SSLeay.
13277 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13279 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13280 when "ssleay" is still not found.
13281 [Ralf S. Engelschall]
13283 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13284 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13286 *) Updated the README file.
13287 [Ralf S. Engelschall]
13289 *) Added various .cvsignore files in the CVS repository subdirs
13290 to make a "cvs update" really silent.
13291 [Ralf S. Engelschall]
13293 *) Recompiled the error-definition header files and added
13294 missing symbols to the Win32 linker tables.
13295 [Ralf S. Engelschall]
13297 *) Cleaned up the top-level documents;
13298 o new files: CHANGES and LICENSE
13299 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13300 o merged COPYRIGHT into LICENSE
13301 o removed obsolete TODO file
13302 o renamed MICROSOFT to INSTALL.W32
13303 [Ralf S. Engelschall]
13305 *) Removed dummy files from the 0.9.1b source tree:
13306 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13307 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13308 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13309 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13310 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13311 [Ralf S. Engelschall]
13313 *) Added various platform portability fixes.
13316 *) The Genesis of the OpenSSL rpject:
13317 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13318 Young and Tim J. Hudson created while they were working for C2Net until
13320 [The OpenSSL Project]
13323 Changes between 0.9.0b and 0.9.1b [not released]
13325 *) Updated a few CA certificates under certs/
13328 *) Changed some BIGNUM api stuff.
13331 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13332 DGUX x86, Linux Alpha, etc.
13335 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13336 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13340 *) Add -strparse option to asn1pars program which parses nested
13342 [Dr Stephen Henson <shenson@bigfoot.com>]
13344 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13347 *) DSA fix for "ca" program.
13350 *) Added "-genkey" option to "dsaparam" program.
13353 *) Added RIPE MD160 (rmd160) message digest.
13356 *) Added -a (all) option to "ssleay version" command.
13359 *) Added PLATFORM define which is the id given to Configure.
13362 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13365 *) Extended the ASN.1 parser routines.
13368 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13371 *) Added a BN_CTX to the BN library.
13374 *) Fixed the weak key values in DES library
13377 *) Changed API in EVP library for cipher aliases.
13380 *) Added support for RC2/64bit cipher.
13383 *) Converted the lhash library to the crypto/mem.c functions.
13386 *) Added more recognized ASN.1 object ids.
13389 *) Added more RSA padding checks for SSL/TLS.
13392 *) Added BIO proxy/filter functionality.
13395 *) Added extra_certs to SSL_CTX which can be used
13396 send extra CA certificates to the client in the CA cert chain sending
13397 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13400 *) Now Fortezza is denied in the authentication phase because
13401 this is key exchange mechanism is not supported by SSLeay at all.
13404 *) Additional PKCS1 checks.
13407 *) Support the string "TLSv1" for all TLS v1 ciphers.
13410 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13411 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13414 *) Fixed a few memory leaks.
13417 *) Fixed various code and comment typos.
13420 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13421 bytes sent in the client random.
13422 [Edward Bishop <ebishop@spyglass.com>]