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 1.1.1k and 1.1.1l [xx XXX xxxx]
14 Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
16 *) Fixed a problem with verifying a certificate chain when using the
17 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks
18 of the certificates present in a certificate chain. It is not set by
21 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
22 the chain that have explicitly encoded elliptic curve parameters was added
23 as an additional strict check.
25 An error in the implementation of this check meant that the result of a
26 previous check to confirm that certificates in the chain are valid CA
27 certificates was overwritten. This effectively bypasses the check
28 that non-CA certificates must not be able to issue other certificates.
30 If a "purpose" has been configured then there is a subsequent opportunity
31 for checks that the certificate is a valid CA. All of the named "purpose"
32 values implemented in libcrypto perform this check. Therefore, where
33 a purpose is set the certificate chain will still be rejected even when the
34 strict flag has been used. A purpose is set by default in libssl client and
35 server certificate verification routines, but it can be overridden or
36 removed by an application.
38 In order to be affected, an application must explicitly set the
39 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
40 for the certificate verification or, in the case of TLS client or server
41 applications, override the default purpose.
45 *) Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
46 crafted renegotiation ClientHello message from a client. If a TLSv1.2
47 renegotiation ClientHello omits the signature_algorithms extension (where
48 it was present in the initial ClientHello), but includes a
49 signature_algorithms_cert extension then a NULL pointer dereference will
50 result, leading to a crash and a denial of service attack.
52 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
53 (which is the default configuration). OpenSSL TLS clients are not impacted
56 [Peter Kästle and Samuel Sapalski]
58 Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
60 *) Fixed the X509_issuer_and_serial_hash() function. It attempts to
61 create a unique hash value based on the issuer and serial number data
62 contained within an X509 certificate. However it was failing to correctly
63 handle any errors that may occur while parsing the issuer field (which might
64 occur if the issuer field is maliciously constructed). This may subsequently
65 result in a NULL pointer deref and a crash leading to a potential denial of
70 *) Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
71 padding mode to correctly check for rollback attacks. This is considered a
72 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
76 *) Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
77 functions. Previously they could overflow the output length argument in some
78 cases where the input length is close to the maximum permissable length for
79 an integer on the platform. In such cases the return value from the function
80 call would be 1 (indicating success), but the output length value would be
81 negative. This could cause applications to behave incorrectly or crash.
85 *) Fixed SRP_Calc_client_key so that it runs in constant time. The previous
86 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
87 could be exploited in a side channel attack to recover the password. Since
88 the attack is local host only this is outside of the current OpenSSL
89 threat model and therefore no CVE is assigned.
91 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
95 Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
97 *) Fixed NULL pointer deref in the GENERAL_NAME_cmp function
98 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
99 If an attacker can control both items being compared then this could lead
100 to a possible denial of service attack. OpenSSL itself uses the
101 GENERAL_NAME_cmp function for two purposes:
102 1) Comparing CRL distribution point names between an available CRL and a
103 CRL distribution point embedded in an X509 certificate
104 2) When verifying that a timestamp response token signer matches the
105 timestamp authority name (exposed via the API functions
106 TS_RESP_verify_response and TS_RESP_verify_token)
110 *) Add support for Apple Silicon M1 Macs with the darwin64-arm64-cc target.
113 *) The security callback, which can be customised by application code, supports
114 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
115 in the "other" parameter. In most places this is what is passed. All these
116 places occur server side. However there was one client side call of this
117 security operation and it passed a DH object instead. This is incorrect
118 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
119 of the other locations. Therefore this client side call has been changed to
120 pass an EVP_PKEY instead.
123 *) In 1.1.1h, an expired trusted (root) certificate was not anymore rejected
124 when validating a certificate path. This check is restored in 1.1.1i.
127 Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
129 *) Certificates with explicit curve parameters are now disallowed in
130 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
133 *) The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
134 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
135 conversely, silently ignore DTLS protocol version bounds when configuring
136 TLS-based contexts. The commands can be repeated to set bounds of both
137 types. The same applies with the corresponding "min_protocol" and
138 "max_protocol" command-line switches, in case some application uses both TLS
141 SSL_CTX instances that are created for a fixed protocol version (e.g.
142 TLSv1_server_method()) also silently ignore version bounds. Previously
143 attempts to apply bounds to these protocol versions would result in an
144 error. Now only the "version-flexible" SSL_CTX instances are subject to
145 limits in configuration files in command-line options.
148 *) Handshake now fails if Extended Master Secret extension is dropped
152 *) Accidentally, an expired trusted (root) certificate is not anymore rejected
153 when validating a certificate path.
156 *) The Oracle Developer Studio compiler will start reporting deprecated APIs
158 Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
160 *) Fixed segmentation fault in SSL_check_chain()
161 Server or client applications that call the SSL_check_chain() function
162 during or after a TLS 1.3 handshake may crash due to a NULL pointer
163 dereference as a result of incorrect handling of the
164 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
165 or unrecognised signature algorithm is received from the peer. This could
166 be exploited by a malicious peer in a Denial of Service attack.
170 *) Added AES consttime code for no-asm configurations
171 an optional constant time support for AES was added
172 when building openssl for no-asm.
173 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
174 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
175 At this time this feature is by default disabled.
176 It will be enabled by default in 3.0.
179 Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
181 *) Revert the change of EOF detection while reading in libssl to avoid
182 regressions in applications depending on the current way of reporting
183 the EOF. As the existing method is not fully accurate the change to
184 reporting the EOF via SSL_ERROR_SSL is kept on the current development
185 branch and will be present in the 3.0 release.
188 *) Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
189 when primes for RSA keys are computed.
190 Since we previously always generated primes == 2 (mod 3) for RSA keys,
191 the 2-prime and 3-prime RSA modules were easy to distinguish, since
192 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
193 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
194 This avoids possible fingerprinting of newly generated RSA modules.
197 Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
198 *) Properly detect EOF while reading in libssl. Previously if we hit an EOF
199 while reading in libssl then we would report an error back to the
200 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
201 an error to the stack (which means we instead return SSL_ERROR_SSL) and
202 therefore give a hint as to what went wrong.
205 *) Check that ed25519 and ed448 are allowed by the security level. Previously
206 signature algorithms not using an MD were not being checked that they were
207 allowed by the security level.
210 *) Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
211 was not quite right. The behaviour was not consistent between resumption
212 and normal handshakes, and also not quite consistent with historical
213 behaviour. The behaviour in various scenarios has been clarified and
214 it has been updated to make it match historical behaviour as closely as
218 *) [VMS only] The header files that the VMS compilers include automatically,
219 __DECC_INCLUDE_PROLOGUE.H and __DECC_INCLUDE_EPILOGUE.H, use pragmas that
220 the C++ compiler doesn't understand. This is a shortcoming in the
221 compiler, but can be worked around with __cplusplus guards.
223 C++ applications that use OpenSSL libraries must be compiled using the
224 qualifier '/NAMES=(AS_IS,SHORTENED)' to be able to use all the OpenSSL
225 functions. Otherwise, only functions with symbols of less than 31
226 characters can be used, as the linker will not be able to successfully
227 resolve symbols with longer names.
230 *) Corrected the documentation of the return values from the EVP_DigestSign*
231 set of functions. The documentation mentioned negative values for some
232 errors, but this was never the case, so the mention of negative values
235 Code that followed the documentation and thereby check with something
236 like 'EVP_DigestSignInit(...) <= 0' will continue to work undisturbed.
239 *) Fixed an an overflow bug in the x64_64 Montgomery squaring procedure
240 used in exponentiation with 512-bit moduli. No EC algorithms are
241 affected. Analysis suggests that attacks against 2-prime RSA1024,
242 3-prime RSA1536, and DSA1024 as a result of this defect would be very
243 difficult to perform and are not believed likely. Attacks against DH512
244 are considered just feasible. However, for an attack the target would
245 have to re-use the DH512 private key, which is not recommended anyway.
246 Also applications directly using the low level API BN_mod_exp may be
247 affected if they use BN_FLG_CONSTTIME.
251 *) Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
252 The presence of this system service is determined at run-time.
255 *) Added newline escaping functionality to a filename when using openssl dgst.
256 This output format is to replicate the output format found in the '*sum'
257 checksum programs. This aims to preserve backward compatibility.
258 [Matt Eaton, Richard Levitte, and Paul Dale]
260 *) Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
264 Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
266 *) Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
267 number generator (RNG). This was intended to include protection in the
268 event of a fork() system call in order to ensure that the parent and child
269 processes did not share the same RNG state. However this protection was not
270 being used in the default case.
272 A partial mitigation for this issue is that the output from a high
273 precision timer is mixed into the RNG state so the likelihood of a parent
274 and child process sharing state is significantly reduced.
276 If an application already calls OPENSSL_init_crypto() explicitly using
277 OPENSSL_INIT_ATFORK then this problem does not occur at all.
279 [Matthias St. Pierre]
281 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is
282 used even when parsing explicit parameters, when loading a serialized key
283 or calling `EC_GROUP_new_from_ecpkparameters()`/
284 `EC_GROUP_new_from_ecparameters()`.
285 This prevents bypass of security hardening and performance gains,
286 especially for curves with specialized EC_METHODs.
287 By default, if a key encoded with explicit parameters is loaded and later
288 serialized, the output is still encoded with explicit parameters, even if
289 internally a "named" EC_GROUP is used for computation.
292 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before
293 this change, EC_GROUP_set_generator would accept order and/or cofactor as
294 NULL. After this change, only the cofactor parameter can be NULL. It also
295 does some minimal sanity checks on the passed order.
299 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
300 An attack is simple, if the first CMS_recipientInfo is valid but the
301 second CMS_recipientInfo is chosen ciphertext. If the second
302 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
303 encryption key will be replaced by garbage, and the message cannot be
304 decoded, but if the RSA decryption fails, the correct encryption key is
305 used and the recipient will not notice the attack.
306 As a work around for this potential attack the length of the decrypted
307 key must be equal to the cipher default key length, in case the
308 certifiate is not given and all recipientInfo are tried out.
309 The old behaviour can be re-enabled in the CMS code by setting the
310 CMS_DEBUG_DECRYPT flag.
314 *) Early start up entropy quality from the DEVRANDOM seed source has been
315 improved for older Linux systems. The RAND subsystem will wait for
316 /dev/random to be producing output before seeding from /dev/urandom.
317 The seeded state is stored for future library initialisations using
318 a system global shared memory segment. The shared memory identifier
319 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
320 the desired value. The default identifier is 114.
323 *) Correct the extended master secret constant on EBCDIC systems. Without this
324 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
325 negotiate EMS will fail. Unfortunately this also means that TLS connections
326 between EBCDIC systems with this fix, and EBCDIC systems without this
327 fix will fail if they negotiate EMS.
330 *) Use Windows installation paths in the mingw builds
332 Mingw isn't a POSIX environment per se, which means that Windows
333 paths should be used for installation.
337 *) Changed DH_check to accept parameters with order q and 2q subgroups.
338 With order 2q subgroups the bit 0 of the private key is not secret
339 but DH_generate_key works around that by clearing bit 0 of the
340 private key for those. This avoids leaking bit 0 of the private key.
343 *) Significantly reduce secure memory usage by the randomness pools.
346 *) Revert the DEVRANDOM_WAIT feature for Linux systems
348 The DEVRANDOM_WAIT feature added a select() call to wait for the
349 /dev/random device to become readable before reading from the
352 It turned out that this change had negative side effects on
353 performance which were not acceptable. After some discussion it
354 was decided to revert this feature and leave it up to the OS
355 resp. the platform maintainer to ensure a proper initialization
356 during early boot time.
357 [Matthias St. Pierre]
359 Changes between 1.1.1b and 1.1.1c [28 May 2019]
361 *) Add build tests for C++. These are generated files that only do one
362 thing, to include one public OpenSSL head file each. This tests that
363 the public header files can be usefully included in a C++ application.
365 This test isn't enabled by default. It can be enabled with the option
366 'enable-buildtest-c++'.
369 *) Enable SHA3 pre-hashing for ECDSA and DSA.
372 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
373 This changes the size when using the genpkey app when no size is given. It
374 fixes an omission in earlier changes that changed all RSA, DSA and DH
375 generation apps to use 2048 bits by default.
378 *) Reorganize the manual pages to consistently have RETURN VALUES,
379 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
380 util/fix-doc-nits accordingly.
381 [Paul Yang, Joshua Lock]
383 *) Add the missing accessor EVP_PKEY_get0_engine()
386 *) Have apps like 's_client' and 's_server' output the signature scheme
387 along with other cipher suite parameters when debugging.
390 *) Make OPENSSL_config() error agnostic again.
393 *) Do the error handling in RSA decryption constant time.
396 *) Prevent over long nonces in ChaCha20-Poly1305.
398 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
399 for every encryption operation. RFC 7539 specifies that the nonce value
400 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
401 and front pads the nonce with 0 bytes if it is less than 12
402 bytes. However it also incorrectly allows a nonce to be set of up to 16
403 bytes. In this case only the last 12 bytes are significant and any
404 additional leading bytes are ignored.
406 It is a requirement of using this cipher that nonce values are
407 unique. Messages encrypted using a reused nonce value are susceptible to
408 serious confidentiality and integrity attacks. If an application changes
409 the default nonce length to be longer than 12 bytes and then makes a
410 change to the leading bytes of the nonce expecting the new value to be a
411 new unique nonce then such an application could inadvertently encrypt
412 messages with a reused nonce.
414 Additionally the ignored bytes in a long nonce are not covered by the
415 integrity guarantee of this cipher. Any application that relies on the
416 integrity of these ignored leading bytes of a long nonce may be further
417 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
418 is safe because no such use sets such a long nonce value. However user
419 applications that use this cipher directly and set a non-default nonce
420 length to be longer than 12 bytes may be vulnerable.
422 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
427 *) Add DEVRANDOM_WAIT feature for Linux systems
429 On older Linux systems where the getrandom() system call is not available,
430 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
431 Contrary to getrandom(), the /dev/urandom device will not block during
432 early boot when the kernel CSPRNG has not been seeded yet.
434 To mitigate this known weakness, use select() to wait for /dev/random to
435 become readable before reading from /dev/urandom.
437 *) Ensure that SM2 only uses SM3 as digest algorithm
440 Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
442 *) Added SCA hardening for modular field inversion in EC_GROUP through
443 a new dedicated field_inv() pointer in EC_METHOD.
444 This also addresses a leakage affecting conversions from projective
445 to affine coordinates.
446 [Billy Bob Brumley, Nicola Tuveri]
448 *) Change the info callback signals for the start and end of a post-handshake
449 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
450 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
451 confused by this and assume that a TLSv1.2 renegotiation has started. This
452 can break KeyUpdate handling. Instead we no longer signal the start and end
453 of a post handshake message exchange (although the messages themselves are
454 still signalled). This could break some applications that were expecting
455 the old signals. However without this KeyUpdate is not usable for many
459 *) Fix a bug in the computation of the endpoint-pair shared secret used
460 by DTLS over SCTP. This breaks interoperability with older versions
461 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
462 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
463 interoperability with such broken implementations. However, enabling
464 this switch breaks interoperability with correct implementations.
466 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
467 re-used X509_PUBKEY object if the second PUBKEY is malformed.
470 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
473 *) Remove the 'dist' target and add a tarball building script. The
474 'dist' target has fallen out of use, and it shouldn't be
475 necessary to configure just to create a source distribution.
478 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
480 *) Timing vulnerability in DSA signature generation
482 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
483 timing side channel attack. An attacker could use variations in the signing
484 algorithm to recover the private key.
486 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
490 *) Timing vulnerability in ECDSA signature generation
492 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
493 timing side channel attack. An attacker could use variations in the signing
494 algorithm to recover the private key.
496 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
500 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
501 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
502 are retained for backwards compatibility.
505 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
506 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
507 of two gigabytes and the error handling improved.
509 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
510 categorized as a normal bug, not a security issue, because the DRBG reseeds
511 automatically and is fully functional even without additional randomness
512 provided by the application.
514 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
516 *) Add a new ClientHello callback. Provides a callback interface that gives
517 the application the ability to adjust the nascent SSL object at the
518 earliest stage of ClientHello processing, immediately after extensions have
519 been collected but before they have been processed. In particular, this
520 callback can adjust the supported TLS versions in response to the contents
524 *) Add SM2 base algorithm support.
527 *) s390x assembly pack: add (improved) hardware-support for the following
528 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
529 aes-cfb/cfb8, aes-ecb.
532 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
533 parameter is no longer accepted, as it leads to a corrupt table. NULL
534 pem_str is reserved for alias entries only.
537 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
538 step for prime curves. The new implementation is based on formulae from
539 differential addition-and-doubling in homogeneous projective coordinates
540 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
541 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
542 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
543 to work in projective coordinates.
544 [Billy Bob Brumley, Nicola Tuveri]
546 *) Change generating and checking of primes so that the error rate of not
547 being prime depends on the intended use based on the size of the input.
548 For larger primes this will result in more rounds of Miller-Rabin.
549 The maximal error rate for primes with more than 1080 bits is lowered
551 [Kurt Roeckx, Annie Yousar]
553 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
556 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
557 moving between systems, and to avoid confusion when a Windows build is
558 done with mingw vs with MSVC. For POSIX installs, there's still a
559 symlink or copy named 'tsget' to avoid that confusion as well.
562 *) Revert blinding in ECDSA sign and instead make problematic addition
563 length-invariant. Switch even to fixed-length Montgomery multiplication.
566 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
567 step for binary curves. The new implementation is based on formulae from
568 differential addition-and-doubling in mixed Lopez-Dahab projective
569 coordinates, modified to independently blind the operands.
570 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
572 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
573 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
574 EC_METHODs to implement their own specialized "ladder step", to take
575 advantage of more favorable coordinate systems or more efficient
576 differential addition-and-doubling algorithms.
577 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
579 *) Modified the random device based seed sources to keep the relevant
580 file descriptors open rather than reopening them on each access.
581 This allows such sources to operate in a chroot() jail without
582 the associated device nodes being available. This behaviour can be
583 controlled using RAND_keep_random_devices_open().
586 *) Numerous side-channel attack mitigations have been applied. This may have
587 performance impacts for some algorithms for the benefit of improved
588 security. Specific changes are noted in this change log by their respective
592 *) AIX shared library support overhaul. Switch to AIX "natural" way of
593 handling shared libraries, which means collecting shared objects of
594 different versions and bitnesses in one common archive. This allows to
595 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
596 doesn't affect the way 3rd party applications are linked, only how
597 multi-version installation is managed.
600 *) Make ec_group_do_inverse_ord() more robust and available to other
601 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
602 mitigations are applied to the fallback BN_mod_inverse().
603 When using this function rather than BN_mod_inverse() directly, new
604 EC cryptosystem implementations are then safer-by-default.
607 *) Add coordinate blinding for EC_POINT and implement projective
608 coordinate blinding for generic prime curves as a countermeasure to
609 chosen point SCA attacks.
610 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
612 *) Add blinding to ECDSA and DSA signatures to protect against side channel
613 attacks discovered by Keegan Ryan (NCC Group).
616 *) Enforce checking in the pkeyutl command line app to ensure that the input
617 length does not exceed the maximum supported digest length when performing
618 a sign, verify or verifyrecover operation.
621 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
622 I/O in combination with something like select() or poll() will hang. This
623 can be turned off again using SSL_CTX_clear_mode().
624 Many applications do not properly handle non-application data records, and
625 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
626 around the problems in those applications, but can also break some.
627 It's recommended to read the manpages about SSL_read(), SSL_write(),
628 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
629 SSL_CTX_set_read_ahead() again.
632 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
633 now allow empty (zero character) pass phrases.
636 *) Apply blinding to binary field modular inversion and remove patent
637 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
640 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
641 binary and prime elliptic curves.
644 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
645 constant time fixed point multiplication.
648 *) Revise elliptic curve scalar multiplication with timing attack
649 defenses: ec_wNAF_mul redirects to a constant time implementation
650 when computing fixed point and variable point multiplication (which
651 in OpenSSL are mostly used with secret scalars in keygen, sign,
652 ECDH derive operations).
653 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
656 *) Updated CONTRIBUTING
659 *) Updated DRBG / RAND to request nonce and additional low entropy
660 randomness from the system.
661 [Matthias St. Pierre]
663 *) Updated 'openssl rehash' to use OpenSSL consistent default.
666 *) Moved the load of the ssl_conf module to libcrypto, which helps
667 loading engines that libssl uses before libssl is initialised.
670 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
673 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
674 [Ingo Schwarze, Rich Salz]
676 *) Added output of accepting IP address and port for 'openssl s_server'
679 *) Added a new API for TLSv1.3 ciphersuites:
680 SSL_CTX_set_ciphersuites()
681 SSL_set_ciphersuites()
684 *) Memory allocation failures consistently add an error to the error
688 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
689 in libcrypto when run as setuid/setgid.
692 *) Load any config file by default when libssl is used.
695 *) Added new public header file <openssl/rand_drbg.h> and documentation
696 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
697 [Matthias St. Pierre]
699 *) QNX support removed (cannot find contributors to get their approval
700 for the license change).
703 *) TLSv1.3 replay protection for early data has been implemented. See the
704 SSL_read_early_data() man page for further details.
707 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
708 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
709 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
710 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
711 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
712 configuration has been separated out. See the ciphers man page or the
713 SSL_CTX_set_ciphersuites() man page for more information.
716 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
717 in responder mode now supports the new "-multi" option, which
718 spawns the specified number of child processes to handle OCSP
719 requests. The "-timeout" option now also limits the OCSP
720 responder's patience to wait to receive the full client request
721 on a newly accepted connection. Child processes are respawned
722 as needed, and the CA index file is automatically reloaded
723 when changed. This makes it possible to run the "ocsp" responder
724 as a long-running service, making the OpenSSL CA somewhat more
725 feature-complete. In this mode, most diagnostic messages logged
726 after entering the event loop are logged via syslog(3) rather than
730 *) Added support for X448 and Ed448. Heavily based on original work by
734 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
735 objects loaded. This adds the functions OSSL_STORE_expect() and
736 OSSL_STORE_find() as well as needed tools to construct searches and
737 get the search data out of them.
740 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
741 version of OpenSSL should review their configuration settings to ensure
742 that they are still appropriate for TLSv1.3. For further information see:
743 https://wiki.openssl.org/index.php/TLS1.3
746 *) Grand redesign of the OpenSSL random generator
748 The default RAND method now utilizes an AES-CTR DRBG according to
749 NIST standard SP 800-90Ar1. The new random generator is essentially
750 a port of the default random generator from the OpenSSL FIPS 2.0
751 object module. It is a hybrid deterministic random bit generator
752 using an AES-CTR bit stream and which seeds and reseeds itself
753 automatically using trusted system entropy sources.
755 Some of its new features are:
756 o Support for multiple DRBG instances with seed chaining.
757 o The default RAND method makes use of a DRBG.
758 o There is a public and private DRBG instance.
759 o The DRBG instances are fork-safe.
760 o Keep all global DRBG instances on the secure heap if it is enabled.
761 o The public and private DRBG instance are per thread for lock free
763 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
765 *) Changed Configure so it only says what it does and doesn't dump
766 so much data. Instead, ./configdata.pm should be used as a script
767 to display all sorts of configuration data.
770 *) Added processing of "make variables" to Configure.
773 *) Added SHA512/224 and SHA512/256 algorithm support.
776 *) The last traces of Netware support, first removed in 1.1.0, have
780 *) Get rid of Makefile.shared, and in the process, make the processing
781 of certain files (rc.obj, or the .def/.map/.opt files produced from
782 the ordinal files) more visible and hopefully easier to trace and
783 debug (or make silent).
786 *) Make it possible to have environment variable assignments as
787 arguments to config / Configure.
790 *) Add multi-prime RSA (RFC 8017) support.
793 *) Add SM3 implemented according to GB/T 32905-2016
794 [ Jack Lloyd <jack.lloyd@ribose.com>,
795 Ronald Tse <ronald.tse@ribose.com>,
796 Erick Borsboom <erick.borsboom@ribose.com> ]
798 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
799 as documented in RFC6066.
800 Based on a patch from Tomasz Moń
801 [Filipe Raimundo da Silva]
803 *) Add SM4 implemented according to GB/T 32907-2016.
804 [ Jack Lloyd <jack.lloyd@ribose.com>,
805 Ronald Tse <ronald.tse@ribose.com>,
806 Erick Borsboom <erick.borsboom@ribose.com> ]
808 *) Reimplement -newreq-nodes and ERR_error_string_n; the
809 original author does not agree with the license change.
812 *) Add ARIA AEAD TLS support.
815 *) Some macro definitions to support VS6 have been removed. Visual
816 Studio 6 has not worked since 1.1.0
819 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
820 without clearing the errors.
823 *) Add "atfork" functions. If building on a system that without
824 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
825 requirements. The RAND facility now uses/requires this.
831 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
832 not possible to disable entirely. However, it's still possible to
833 disable the console reading UI method, UI_OpenSSL() (use UI_null()
836 To disable, configure with 'no-ui-console'. 'no-ui' is still
837 possible to use as an alias. Check at compile time with the
838 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
839 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
842 *) Add a STORE module, which implements a uniform and URI based reader of
843 stores that can contain keys, certificates, CRLs and numerous other
844 objects. The main API is loosely based on a few stdio functions,
845 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
846 OSSL_STORE_error and OSSL_STORE_close.
847 The implementation uses backends called "loaders" to implement arbitrary
848 URI schemes. There is one built in "loader" for the 'file' scheme.
851 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
852 then adjusted to work on FreeBSD 8.4 as well.
853 Enable by configuring with 'enable-devcryptoeng'. This is done by default
854 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
857 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
858 util/mkerr.pl, which is adapted to allow those prefixes, leading to
859 error code calls like this:
861 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
863 With this change, we claim the namespaces OSSL and OPENSSL in a manner
864 that can be encoded in C. For the foreseeable future, this will only
866 [Richard Levitte and Tim Hudson]
868 *) Removed BSD cryptodev engine.
871 *) Add a build target 'build_all_generated', to build all generated files
872 and only that. This can be used to prepare everything that requires
873 things like perl for a system that lacks perl and then move everything
874 to that system and do the rest of the build there.
877 *) In the UI interface, make it possible to duplicate the user data. This
878 can be used by engines that need to retain the data for a longer time
879 than just the call where this user data is passed.
882 *) Ignore the '-named_curve auto' value for compatibility of applications
884 [Tomas Mraz <tmraz@fedoraproject.org>]
886 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
887 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
888 alerts across multiple records (some of which could be empty). In practice
889 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
890 prohibits this altogether and other libraries (BoringSSL, NSS) do not
891 support this at all. Supporting it adds significant complexity to the
892 record layer, and its removal is unlikely to cause interoperability
896 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
897 with Z. These are meant to replace LONG and ZLONG and to be size safe.
898 The use of LONG and ZLONG is discouraged and scheduled for deprecation
902 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
903 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
904 [Richard Levitte, Andy Polyakov]
906 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
910 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
911 platform rather than 'mingw'.
914 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
915 success if they are asked to add an object which already exists
916 in the store. This change cascades to other functions which load
917 certificates and CRLs.
920 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
921 facilitate stack unwinding even from assembly subroutines.
924 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
925 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
928 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
929 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
930 which is the minimum version we support.
933 *) Certificate time validation (X509_cmp_time) enforces stricter
934 compliance with RFC 5280. Fractional seconds and timezone offsets
935 are no longer allowed.
938 *) Add support for ARIA
941 *) s_client will now send the Server Name Indication (SNI) extension by
942 default unless the new "-noservername" option is used. The server name is
943 based on the host provided to the "-connect" option unless overridden by
947 *) Add support for SipHash
950 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
951 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
952 prevent issues where no progress is being made and the peer continually
953 sends unrecognised record types, using up resources processing them.
956 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
957 using the algorithm defined in
958 https://www.akkadia.org/drepper/SHA-crypt.txt
961 *) Heartbeat support has been removed; the ABI is changed for now.
962 [Richard Levitte, Rich Salz]
964 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
967 *) The RSA "null" method, which was partially supported to avoid patent
968 issues, has been replaced to always returns NULL.
972 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
974 *) Client DoS due to large DH parameter
976 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
977 malicious server can send a very large prime value to the client. This will
978 cause the client to spend an unreasonably long period of time generating a
979 key for this prime resulting in a hang until the client has finished. This
980 could be exploited in a Denial Of Service attack.
982 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
986 *) Cache timing vulnerability in RSA Key Generation
988 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
989 a cache timing side channel attack. An attacker with sufficient access to
990 mount cache timing attacks during the RSA key generation process could
991 recover the private key.
993 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
994 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
998 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
999 parameter is no longer accepted, as it leads to a corrupt table. NULL
1000 pem_str is reserved for alias entries only.
1003 *) Revert blinding in ECDSA sign and instead make problematic addition
1004 length-invariant. Switch even to fixed-length Montgomery multiplication.
1007 *) Change generating and checking of primes so that the error rate of not
1008 being prime depends on the intended use based on the size of the input.
1009 For larger primes this will result in more rounds of Miller-Rabin.
1010 The maximal error rate for primes with more than 1080 bits is lowered
1012 [Kurt Roeckx, Annie Yousar]
1014 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1017 *) Add blinding to ECDSA and DSA signatures to protect against side channel
1018 attacks discovered by Keegan Ryan (NCC Group).
1021 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1022 now allow empty (zero character) pass phrases.
1025 *) Certificate time validation (X509_cmp_time) enforces stricter
1026 compliance with RFC 5280. Fractional seconds and timezone offsets
1027 are no longer allowed.
1030 *) Fixed a text canonicalisation bug in CMS
1032 Where a CMS detached signature is used with text content the text goes
1033 through a canonicalisation process first prior to signing or verifying a
1034 signature. This process strips trailing space at the end of lines, converts
1035 line terminators to CRLF and removes additional trailing line terminators
1036 at the end of a file. A bug in the canonicalisation process meant that
1037 some characters, such as form-feed, were incorrectly treated as whitespace
1038 and removed. This is contrary to the specification (RFC5485). This fix
1039 could mean that detached text data signed with an earlier version of
1040 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
1041 signed with a fixed OpenSSL may fail to verify with an earlier version of
1042 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
1043 and use the "-binary" flag (for the "cms" command line application) or set
1044 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
1047 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
1049 *) Constructed ASN.1 types with a recursive definition could exceed the stack
1051 Constructed ASN.1 types with a recursive definition (such as can be found
1052 in PKCS7) could eventually exceed the stack given malicious input with
1053 excessive recursion. This could result in a Denial Of Service attack. There
1054 are no such structures used within SSL/TLS that come from untrusted sources
1055 so this is considered safe.
1057 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
1062 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
1064 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
1065 effectively reduced to only comparing the least significant bit of each
1066 byte. This allows an attacker to forge messages that would be considered as
1067 authenticated in an amount of tries lower than that guaranteed by the
1068 security claims of the scheme. The module can only be compiled by the
1069 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
1071 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
1076 *) Add a build target 'build_all_generated', to build all generated files
1077 and only that. This can be used to prepare everything that requires
1078 things like perl for a system that lacks perl and then move everything
1079 to that system and do the rest of the build there.
1082 *) Backport SSL_OP_NO_RENGOTIATION
1084 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
1085 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
1086 changes this is no longer possible in 1.1.0. Therefore the new
1087 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
1088 1.1.0 to provide equivalent functionality.
1090 Note that if an application built against 1.1.0h headers (or above) is run
1091 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
1092 accepted but nothing will happen, i.e. renegotiation will not be prevented.
1095 *) Removed the OS390-Unix config target. It relied on a script that doesn't
1099 *) rsaz_1024_mul_avx2 overflow bug on x86_64
1101 There is an overflow bug in the AVX2 Montgomery multiplication procedure
1102 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
1103 Analysis suggests that attacks against RSA and DSA as a result of this
1104 defect would be very difficult to perform and are not believed likely.
1105 Attacks against DH1024 are considered just feasible, because most of the
1106 work necessary to deduce information about a private key may be performed
1107 offline. The amount of resources required for such an attack would be
1108 significant. However, for an attack on TLS to be meaningful, the server
1109 would have to share the DH1024 private key among multiple clients, which is
1110 no longer an option since CVE-2016-0701.
1112 This only affects processors that support the AVX2 but not ADX extensions
1113 like Intel Haswell (4th generation).
1115 This issue was reported to OpenSSL by David Benjamin (Google). The issue
1116 was originally found via the OSS-Fuzz project.
1120 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
1122 *) bn_sqrx8x_internal carry bug on x86_64
1124 There is a carry propagating bug in the x86_64 Montgomery squaring
1125 procedure. No EC algorithms are affected. Analysis suggests that attacks
1126 against RSA and DSA as a result of this defect would be very difficult to
1127 perform and are not believed likely. Attacks against DH are considered just
1128 feasible (although very difficult) because most of the work necessary to
1129 deduce information about a private key may be performed offline. The amount
1130 of resources required for such an attack would be very significant and
1131 likely only accessible to a limited number of attackers. An attacker would
1132 additionally need online access to an unpatched system using the target
1133 private key in a scenario with persistent DH parameters and a private
1134 key that is shared between multiple clients.
1136 This only affects processors that support the BMI1, BMI2 and ADX extensions
1137 like Intel Broadwell (5th generation) and later or AMD Ryzen.
1139 This issue was reported to OpenSSL by the OSS-Fuzz project.
1143 *) Malformed X.509 IPAddressFamily could cause OOB read
1145 If an X.509 certificate has a malformed IPAddressFamily extension,
1146 OpenSSL could do a one-byte buffer overread. The most likely result
1147 would be an erroneous display of the certificate in text format.
1149 This issue was reported to OpenSSL by the OSS-Fuzz project.
1153 Changes between 1.1.0e and 1.1.0f [25 May 2017]
1155 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1156 platform rather than 'mingw'.
1159 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1160 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1161 which is the minimum version we support.
1164 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
1166 *) Encrypt-Then-Mac renegotiation crash
1168 During a renegotiation handshake if the Encrypt-Then-Mac extension is
1169 negotiated where it was not in the original handshake (or vice-versa) then
1170 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
1171 and servers are affected.
1173 This issue was reported to OpenSSL by Joe Orton (Red Hat).
1177 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
1179 *) Truncated packet could crash via OOB read
1181 If one side of an SSL/TLS path is running on a 32-bit host and a specific
1182 cipher is being used, then a truncated packet can cause that host to
1183 perform an out-of-bounds read, usually resulting in a crash.
1185 This issue was reported to OpenSSL by Robert Święcki of Google.
1189 *) Bad (EC)DHE parameters cause a client crash
1191 If a malicious server supplies bad parameters for a DHE or ECDHE key
1192 exchange then this can result in the client attempting to dereference a
1193 NULL pointer leading to a client crash. This could be exploited in a Denial
1196 This issue was reported to OpenSSL by Guido Vranken.
1200 *) BN_mod_exp may produce incorrect results on x86_64
1202 There is a carry propagating bug in the x86_64 Montgomery squaring
1203 procedure. No EC algorithms are affected. Analysis suggests that attacks
1204 against RSA and DSA as a result of this defect would be very difficult to
1205 perform and are not believed likely. Attacks against DH are considered just
1206 feasible (although very difficult) because most of the work necessary to
1207 deduce information about a private key may be performed offline. The amount
1208 of resources required for such an attack would be very significant and
1209 likely only accessible to a limited number of attackers. An attacker would
1210 additionally need online access to an unpatched system using the target
1211 private key in a scenario with persistent DH parameters and a private
1212 key that is shared between multiple clients. For example this can occur by
1213 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1214 similar to CVE-2015-3193 but must be treated as a separate problem.
1216 This issue was reported to OpenSSL by the OSS-Fuzz project.
1220 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1222 *) ChaCha20/Poly1305 heap-buffer-overflow
1224 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1225 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1226 crash. This issue is not considered to be exploitable beyond a DoS.
1228 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1232 *) CMS Null dereference
1234 Applications parsing invalid CMS structures can crash with a NULL pointer
1235 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1236 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1237 structure callback if an attempt is made to free certain invalid encodings.
1238 Only CHOICE structures using a callback which do not handle NULL value are
1241 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1245 *) Montgomery multiplication may produce incorrect results
1247 There is a carry propagating bug in the Broadwell-specific Montgomery
1248 multiplication procedure that handles input lengths divisible by, but
1249 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1250 and DH private keys are impossible. This is because the subroutine in
1251 question is not used in operations with the private key itself and an input
1252 of the attacker's direct choice. Otherwise the bug can manifest itself as
1253 transient authentication and key negotiation failures or reproducible
1254 erroneous outcome of public-key operations with specially crafted input.
1255 Among EC algorithms only Brainpool P-512 curves are affected and one
1256 presumably can attack ECDH key negotiation. Impact was not analyzed in
1257 detail, because pre-requisites for attack are considered unlikely. Namely
1258 multiple clients have to choose the curve in question and the server has to
1259 share the private key among them, neither of which is default behaviour.
1260 Even then only clients that chose the curve will be affected.
1262 This issue was publicly reported as transient failures and was not
1263 initially recognized as a security issue. Thanks to Richard Morgan for
1264 providing reproducible case.
1268 *) Removed automatic addition of RPATH in shared libraries and executables,
1269 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1272 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1274 *) Fix Use After Free for large message sizes
1276 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1277 message larger than approx 16k is received then the underlying buffer to
1278 store the incoming message is reallocated and moved. Unfortunately a
1279 dangling pointer to the old location is left which results in an attempt to
1280 write to the previously freed location. This is likely to result in a
1281 crash, however it could potentially lead to execution of arbitrary code.
1283 This issue only affects OpenSSL 1.1.0a.
1285 This issue was reported to OpenSSL by Robert Święcki.
1289 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1291 *) OCSP Status Request extension unbounded memory growth
1293 A malicious client can send an excessively large OCSP Status Request
1294 extension. If that client continually requests renegotiation, sending a
1295 large OCSP Status Request extension each time, then there will be unbounded
1296 memory growth on the server. This will eventually lead to a Denial Of
1297 Service attack through memory exhaustion. Servers with a default
1298 configuration are vulnerable even if they do not support OCSP. Builds using
1299 the "no-ocsp" build time option are not affected.
1301 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1305 *) SSL_peek() hang on empty record
1307 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1308 sends an empty record. This could be exploited by a malicious peer in a
1309 Denial Of Service attack.
1311 This issue was reported to OpenSSL by Alex Gaynor.
1315 *) Excessive allocation of memory in tls_get_message_header() and
1316 dtls1_preprocess_fragment()
1318 A (D)TLS message includes 3 bytes for its length in the header for the
1319 message. This would allow for messages up to 16Mb in length. Messages of
1320 this length are excessive and OpenSSL includes a check to ensure that a
1321 peer is sending reasonably sized messages in order to avoid too much memory
1322 being consumed to service a connection. A flaw in the logic of version
1323 1.1.0 means that memory for the message is allocated too early, prior to
1324 the excessive message length check. Due to way memory is allocated in
1325 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1326 to service a connection. This could lead to a Denial of Service through
1327 memory exhaustion. However, the excessive message length check still takes
1328 place, and this would cause the connection to immediately fail. Assuming
1329 that the application calls SSL_free() on the failed connection in a timely
1330 manner then the 21Mb of allocated memory will then be immediately freed
1331 again. Therefore the excessive memory allocation will be transitory in
1332 nature. This then means that there is only a security impact if:
1334 1) The application does not call SSL_free() in a timely manner in the event
1335 that the connection fails
1337 2) The application is working in a constrained environment where there is
1338 very little free memory
1340 3) The attacker initiates multiple connection attempts such that there are
1341 multiple connections in a state where memory has been allocated for the
1342 connection; SSL_free() has not yet been called; and there is insufficient
1343 memory to service the multiple requests.
1345 Except in the instance of (1) above any Denial Of Service is likely to be
1346 transitory because as soon as the connection fails the memory is
1347 subsequently freed again in the SSL_free() call. However there is an
1348 increased risk during this period of application crashes due to the lack of
1349 memory - which would then mean a more serious Denial of Service.
1351 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1352 (CVE-2016-6307 and CVE-2016-6308)
1355 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1356 had to be removed. Primary reason is that vendor assembler can't
1357 assemble our modules with -KPIC flag. As result it, assembly
1358 support, was not even available as option. But its lack means
1359 lack of side-channel resistant code, which is incompatible with
1360 security by todays standards. Fortunately gcc is readily available
1361 prepackaged option, which we firmly point at...
1364 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1366 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1367 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1368 (to any value) allows Windows user to access PKCS#12 file generated
1369 with Windows CryptoAPI and protected with non-ASCII password, as well
1370 as files generated under UTF-8 locale on Linux also protected with
1374 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1375 have been disabled by default and removed from DEFAULT, just like RC4.
1376 See the RC4 item below to re-enable both.
1379 *) The method for finding the storage location for the Windows RAND seed file
1380 has changed. First we check %RANDFILE%. If that is not set then we check
1381 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1382 all else fails we fall back to C:\.
1385 *) The EVP_EncryptUpdate() function has had its return type changed from void
1386 to int. A return of 0 indicates and error while a return of 1 indicates
1390 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1391 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1392 off the constant time implementation for RSA, DSA and DH have been made
1393 no-ops and deprecated.
1396 *) Windows RAND implementation was simplified to only get entropy by
1397 calling CryptGenRandom(). Various other RAND-related tickets
1399 [Joseph Wylie Yandle, Rich Salz]
1401 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1402 and OPENSSL_LH_, respectively. The old names are available
1403 with API compatibility. They new names are now completely documented.
1406 *) Unify TYPE_up_ref(obj) methods signature.
1407 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1408 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1409 int (instead of void) like all others TYPE_up_ref() methods.
1410 So now these methods also check the return value of CRYPTO_atomic_add(),
1411 and the validity of object reference counter.
1412 [fdasilvayy@gmail.com]
1414 *) With Windows Visual Studio builds, the .pdb files are installed
1415 alongside the installed libraries and executables. For a static
1416 library installation, ossl_static.pdb is the associate compiler
1417 generated .pdb file to be used when linking programs.
1420 *) Remove openssl.spec. Packaging files belong with the packagers.
1423 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1424 recognise x86_64 architectures automatically. You can still decide
1425 to build for a different bitness with the environment variable
1426 KERNEL_BITS (can be 32 or 64), for example:
1428 KERNEL_BITS=32 ./config
1432 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1433 256 bit AES and HMAC with SHA256.
1436 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1439 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1442 *) To enable users to have their own config files and build file templates,
1443 Configure looks in the directory indicated by the environment variable
1444 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1445 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1446 name and is used as is.
1449 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1450 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1451 X509_CERT_FILE_CTX was removed.
1454 *) "shared" builds are now the default. To create only static libraries use
1455 the "no-shared" Configure option.
1458 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1459 All of these option have not worked for some while and are fundamental
1463 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1464 global cleanup functions are no longer required because they are handled
1465 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1466 Explicitly de-initing can cause problems (e.g. where a library that uses
1467 OpenSSL de-inits, but an application is still using it). The affected
1468 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1469 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1470 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1471 COMP_zlib_cleanup().
1474 *) --strict-warnings no longer enables runtime debugging options
1475 such as REF_DEBUG. Instead, debug options are automatically
1476 enabled with '--debug' builds.
1477 [Andy Polyakov, Emilia Käsper]
1479 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1480 have been moved out of the public header files. New functions for managing
1481 these have been added.
1484 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1485 objects have been moved out of the public header files. New
1486 functions for managing these have been added.
1489 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1490 have been moved out of the public header files. New functions for managing
1491 these have been added.
1494 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1495 moved out of the public header files. New functions for managing these
1499 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1502 *) Removed the mk1mf build scripts.
1505 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1506 it is always safe to #include a header now.
1509 *) Removed the aged BC-32 config and all its supporting scripts
1512 *) Removed support for Ultrix, Netware, and OS/2.
1515 *) Add support for HKDF.
1516 [Alessandro Ghedini]
1518 *) Add support for blake2b and blake2s
1521 *) Added support for "pipelining". Ciphers that have the
1522 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1523 encryptions/decryptions simultaneously. There are currently no built-in
1524 ciphers with this property but the expectation is that engines will be able
1525 to offer it to significantly improve throughput. Support has been extended
1526 into libssl so that multiple records for a single connection can be
1527 processed in one go (for >=TLS 1.1).
1530 *) Added the AFALG engine. This is an async capable engine which is able to
1531 offload work to the Linux kernel. In this initial version it only supports
1532 AES128-CBC. The kernel must be version 4.1.0 or greater.
1535 *) OpenSSL now uses a new threading API. It is no longer necessary to
1536 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1537 are two supported threading models: pthreads and windows threads. It is
1538 also possible to configure OpenSSL at compile time for "no-threads". The
1539 old threading API should no longer be used. The functions have been
1540 replaced with "no-op" compatibility macros.
1541 [Alessandro Ghedini, Matt Caswell]
1543 *) Modify behavior of ALPN to invoke callback after SNI/servername
1544 callback, such that updates to the SSL_CTX affect ALPN.
1547 *) Add SSL_CIPHER queries for authentication and key-exchange.
1550 *) Changes to the DEFAULT cipherlist:
1551 - Prefer (EC)DHE handshakes over plain RSA.
1552 - Prefer AEAD ciphers over legacy ciphers.
1553 - Prefer ECDSA over RSA when both certificates are available.
1554 - Prefer TLSv1.2 ciphers/PRF.
1555 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1559 *) Change the ECC default curve list to be this, in order: x25519,
1560 secp256r1, secp521r1, secp384r1.
1563 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1564 disabled by default. They can be re-enabled using the
1565 enable-weak-ssl-ciphers option to Configure.
1568 *) If the server has ALPN configured, but supports no protocols that the
1569 client advertises, send a fatal "no_application_protocol" alert.
1570 This behaviour is SHALL in RFC 7301, though it isn't universally
1571 implemented by other servers.
1574 *) Add X25519 support.
1575 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1576 for public and private key encoding using the format documented in
1577 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1578 key generation and key derivation.
1580 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1584 *) Deprecate SRP_VBASE_get_by_user.
1585 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1586 In order to fix an unavoidable memory leak (CVE-2016-0798),
1587 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1588 seed, even if the seed is configured.
1590 Users should use SRP_VBASE_get1_by_user instead. Note that in
1591 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1592 also that even though configuring the SRP seed attempts to hide
1593 invalid usernames by continuing the handshake with fake
1594 credentials, this behaviour is not constant time and no strong
1595 guarantees are made that the handshake is indistinguishable from
1596 that of a valid user.
1599 *) Configuration change; it's now possible to build dynamic engines
1600 without having to build shared libraries and vice versa. This
1601 only applies to the engines in engines/, those in crypto/engine/
1602 will always be built into libcrypto (i.e. "static").
1604 Building dynamic engines is enabled by default; to disable, use
1605 the configuration option "disable-dynamic-engine".
1607 The only requirements for building dynamic engines are the
1608 presence of the DSO module and building with position independent
1609 code, so they will also automatically be disabled if configuring
1610 with "disable-dso" or "disable-pic".
1612 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1613 are also taken away from openssl/opensslconf.h, as they are
1617 *) Configuration change; if there is a known flag to compile
1618 position independent code, it will always be applied on the
1619 libcrypto and libssl object files, and never on the application
1620 object files. This means other libraries that use routines from
1621 libcrypto / libssl can be made into shared libraries regardless
1622 of how OpenSSL was configured.
1624 If this isn't desirable, the configuration options "disable-pic"
1625 or "no-pic" can be used to disable the use of PIC. This will
1626 also disable building shared libraries and dynamic engines.
1629 *) Removed JPAKE code. It was experimental and has no wide use.
1632 *) The INSTALL_PREFIX Makefile variable has been renamed to
1633 DESTDIR. That makes for less confusion on what this variable
1634 is for. Also, the configuration option --install_prefix is
1638 *) Heartbeat for TLS has been removed and is disabled by default
1639 for DTLS; configure with enable-heartbeats. Code that uses the
1640 old #define's might need to be updated.
1641 [Emilia Käsper, Rich Salz]
1643 *) Rename REF_CHECK to REF_DEBUG.
1646 *) New "unified" build system
1648 The "unified" build system is aimed to be a common system for all
1649 platforms we support. With it comes new support for VMS.
1651 This system builds supports building in a different directory tree
1652 than the source tree. It produces one Makefile (for unix family
1653 or lookalikes), or one descrip.mms (for VMS).
1655 The source of information to make the Makefile / descrip.mms is
1656 small files called 'build.info', holding the necessary
1657 information for each directory with source to compile, and a
1658 template in Configurations, like unix-Makefile.tmpl or
1661 With this change, the library names were also renamed on Windows
1662 and on VMS. They now have names that are closer to the standard
1663 on Unix, and include the major version number, and in certain
1664 cases, the architecture they are built for. See "Notes on shared
1665 libraries" in INSTALL.
1667 We rely heavily on the perl module Text::Template.
1670 *) Added support for auto-initialisation and de-initialisation of the library.
1671 OpenSSL no longer requires explicit init or deinit routines to be called,
1672 except in certain circumstances. See the OPENSSL_init_crypto() and
1673 OPENSSL_init_ssl() man pages for further information.
1676 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1677 "peer" argument is now expected to be a BIO_ADDR object.
1679 *) Rewrite of BIO networking library. The BIO library lacked consistent
1680 support of IPv6, and adding it required some more extensive
1681 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1682 which hold all types of addresses and chains of address information.
1683 It also introduces a new API, with functions like BIO_socket,
1684 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1685 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1686 have been adapted accordingly.
1689 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1693 *) CRIME protection: disable compression by default, even if OpenSSL is
1694 compiled with zlib enabled. Applications can still enable compression
1695 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1696 using the SSL_CONF library to configure compression.
1699 *) The signature of the session callback configured with
1700 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1701 was explicitly marked as 'const unsigned char*' instead of
1705 *) Always DPURIFY. Remove the use of uninitialized memory in the
1706 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1709 *) Removed many obsolete configuration items, including
1710 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1711 MD2_CHAR, MD2_INT, MD2_LONG
1713 IDEA_SHORT, IDEA_LONG
1714 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1715 [Rich Salz, with advice from Andy Polyakov]
1717 *) Many BN internals have been moved to an internal header file.
1718 [Rich Salz with help from Andy Polyakov]
1720 *) Configuration and writing out the results from it has changed.
1721 Files such as Makefile include/openssl/opensslconf.h and are now
1722 produced through general templates, such as Makefile.in and
1723 crypto/opensslconf.h.in and some help from the perl module
1726 Also, the center of configuration information is no longer
1727 Makefile. Instead, Configure produces a perl module in
1728 configdata.pm which holds most of the config data (in the hash
1729 table %config), the target data that comes from the target
1730 configuration in one of the Configurations/*.conf files (in
1734 *) To clarify their intended purposes, the Configure options
1735 --prefix and --openssldir change their semantics, and become more
1736 straightforward and less interdependent.
1738 --prefix shall be used exclusively to give the location INSTALLTOP
1739 where programs, scripts, libraries, include files and manuals are
1740 going to be installed. The default is now /usr/local.
1742 --openssldir shall be used exclusively to give the default
1743 location OPENSSLDIR where certificates, private keys, CRLs are
1744 managed. This is also where the default openssl.cnf gets
1746 If the directory given with this option is a relative path, the
1747 values of both the --prefix value and the --openssldir value will
1748 be combined to become OPENSSLDIR.
1749 The default for --openssldir is INSTALLTOP/ssl.
1751 Anyone who uses --openssldir to specify where OpenSSL is to be
1752 installed MUST change to use --prefix instead.
1755 *) The GOST engine was out of date and therefore it has been removed. An up
1756 to date GOST engine is now being maintained in an external repository.
1757 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1758 support for GOST ciphersuites (these are only activated if a GOST engine
1762 *) EGD is no longer supported by default; use enable-egd when
1764 [Ben Kaduk and Rich Salz]
1766 *) The distribution now has Makefile.in files, which are used to
1767 create Makefile's when Configure is run. *Configure must be run
1768 before trying to build now.*
1771 *) The return value for SSL_CIPHER_description() for error conditions
1775 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1777 Obtaining and performing DNSSEC validation of TLSA records is
1778 the application's responsibility. The application provides
1779 the TLSA records of its choice to OpenSSL, and these are then
1780 used to authenticate the peer.
1782 The TLSA records need not even come from DNS. They can, for
1783 example, be used to implement local end-entity certificate or
1784 trust-anchor "pinning", where the "pin" data takes the form
1785 of TLSA records, which can augment or replace verification
1786 based on the usual WebPKI public certification authorities.
1789 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1790 continues to support deprecated interfaces in default builds.
1791 However, applications are strongly advised to compile their
1792 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1793 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1794 or the 1.1.0 releases.
1796 In environments in which all applications have been ported to
1797 not use any deprecated interfaces OpenSSL's Configure script
1798 should be used with the --api=1.1.0 option to entirely remove
1799 support for the deprecated features from the library and
1800 unconditionally disable them in the installed headers.
1801 Essentially the same effect can be achieved with the "no-deprecated"
1802 argument to Configure, except that this will always restrict
1803 the build to just the latest API, rather than a fixed API
1806 As applications are ported to future revisions of the API,
1807 they should update their compile-time OPENSSL_API_COMPAT define
1808 accordingly, but in most cases should be able to continue to
1809 compile with later releases.
1811 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1812 0x10000000L and 0x00908000L, respectively. However those
1813 versions did not support the OPENSSL_API_COMPAT feature, and
1814 so applications are not typically tested for explicit support
1815 of just the undeprecated features of either release.
1818 *) Add support for setting the minimum and maximum supported protocol.
1819 It can bet set via the SSL_set_min_proto_version() and
1820 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1821 MaxProtocol. It's recommended to use the new APIs to disable
1822 protocols instead of disabling individual protocols using
1823 SSL_set_options() or SSL_CONF's Protocol. This change also
1824 removes support for disabling TLS 1.2 in the OpenSSL TLS
1825 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1828 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1831 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1832 and integrates ECDSA and ECDH functionality into EC. Implementations can
1833 now redirect key generation and no longer need to convert to or from
1836 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1837 include the ec.h header file instead.
1840 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1841 ciphers who are no longer supported and drops support the ephemeral RSA key
1842 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1845 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1846 opaque. For HMAC_CTX, the following constructors and destructors
1849 HMAC_CTX *HMAC_CTX_new(void);
1850 void HMAC_CTX_free(HMAC_CTX *ctx);
1852 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1853 destroy such methods has been added. See EVP_MD_meth_new(3) and
1854 EVP_CIPHER_meth_new(3) for documentation.
1857 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1858 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1859 EVP_MD_CTX_reset() should be called instead to reinitialise
1860 an already created structure.
1861 2) For consistency with the majority of our object creators and
1862 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1863 EVP_MD_CTX_(new|free). The old names are retained as macros
1864 for deprecated builds.
1867 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1868 cryptographic operations to be performed asynchronously as long as an
1869 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1870 further details. Libssl has also had this capability integrated with the
1871 introduction of the new mode SSL_MODE_ASYNC and associated error
1872 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1873 pages. This work was developed in partnership with Intel Corp.
1876 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1877 always enabled now. If you want to disable the support you should
1878 exclude it using the list of supported ciphers. This also means that the
1879 "-no_ecdhe" option has been removed from s_server.
1882 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1883 SSL_{CTX_}set1_curves() which can set a list.
1886 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1887 curve you want to support using SSL_{CTX_}set1_curves().
1890 *) State machine rewrite. The state machine code has been significantly
1891 refactored in order to remove much duplication of code and solve issues
1892 with the old code (see ssl/statem/README for further details). This change
1893 does have some associated API changes. Notably the SSL_state() function
1894 has been removed and replaced by SSL_get_state which now returns an
1895 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1896 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1900 *) All instances of the string "ssleay" in the public API were replaced
1901 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1902 Some error codes related to internal RSA_eay API's were renamed.
1905 *) The demo files in crypto/threads were moved to demo/threads.
1908 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1910 [Matt Caswell, Rich Salz]
1912 *) New ASN.1 embed macro.
1914 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1915 structure is not allocated: it is part of the parent. That is instead of
1923 This reduces memory fragmentation and make it impossible to accidentally
1924 set a mandatory field to NULL.
1926 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1927 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1928 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1932 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1935 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1936 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1937 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1938 DES and RC4 ciphersuites.
1941 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1942 This changes the decoding behaviour for some invalid messages,
1943 though the change is mostly in the more lenient direction, and
1944 legacy behaviour is preserved as much as possible.
1947 *) Fix no-stdio build.
1948 [ David Woodhouse <David.Woodhouse@intel.com> and also
1949 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1951 *) New testing framework
1952 The testing framework has been largely rewritten and is now using
1953 perl and the perl modules Test::Harness and an extended variant of
1954 Test::More called OpenSSL::Test to do its work. All test scripts in
1955 test/ have been rewritten into test recipes, and all direct calls to
1956 executables in test/Makefile have become individual recipes using the
1957 simplified testing OpenSSL::Test::Simple.
1959 For documentation on our testing modules, do:
1961 perldoc test/testlib/OpenSSL/Test/Simple.pm
1962 perldoc test/testlib/OpenSSL/Test.pm
1966 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1967 are used; the latter aborts on memory leaks (usually checked on exit).
1968 Some undocumented "set malloc, etc., hooks" functions were removed
1969 and others were changed. All are now documented.
1972 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1974 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1976 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1977 from RFC4279, RFC4785, RFC5487, RFC5489.
1979 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1980 original RSA_PSK patch.
1983 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1984 era flag was never set throughout the codebase (only read). Also removed
1985 SSL3_FLAGS_POP_BUFFER which was only used if
1986 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1989 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1990 to be "oneline" instead of "compat".
1993 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1994 not aware of clients that still exhibit this bug, and the workaround
1995 hasn't been working properly for a while.
1998 *) The return type of BIO_number_read() and BIO_number_written() as well as
1999 the corresponding num_read and num_write members in the BIO structure has
2000 changed from unsigned long to uint64_t. On platforms where an unsigned
2001 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
2005 *) Given the pervasive nature of TLS extensions it is inadvisable to run
2006 OpenSSL without support for them. It also means that maintaining
2007 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
2008 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
2011 *) Removed support for the two export grade static DH ciphersuites
2012 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
2013 were newly added (along with a number of other static DH ciphersuites) to
2014 1.0.2. However the two export ones have *never* worked since they were
2015 introduced. It seems strange in any case to be adding new export
2016 ciphersuites, and given "logjam" it also does not seem correct to fix them.
2019 *) Version negotiation has been rewritten. In particular SSLv23_method(),
2020 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
2021 and turned into macros which simply call the new preferred function names
2022 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
2023 should use the new names instead. Also as part of this change the ssl23.h
2024 header file has been removed.
2027 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
2028 code and the associated standard is no longer considered fit-for-purpose.
2031 *) RT2547 was closed. When generating a private key, try to make the
2032 output file readable only by the owner. This behavior change might
2033 be noticeable when interacting with other software.
2035 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
2039 *) Added HTTP GET support to the ocsp command.
2042 *) Changed default digest for the dgst and enc commands from MD5 to
2046 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
2049 *) Added support for TLS extended master secret from
2050 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
2051 initial patch which was a great help during development.
2054 *) All libssl internal structures have been removed from the public header
2055 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
2056 now redundant). Users should not attempt to access internal structures
2057 directly. Instead they should use the provided API functions.
2060 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
2061 Access to deprecated functions can be re-enabled by running config with
2062 "enable-deprecated". In addition applications wishing to use deprecated
2063 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
2064 will, by default, disable some transitive includes that previously existed
2065 in the header files (e.g. ec.h will no longer, by default, include bn.h)
2068 *) Added support for OCB mode. OpenSSL has been granted a patent license
2069 compatible with the OpenSSL license for use of OCB. Details are available
2070 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
2071 for OCB can be removed by calling config with no-ocb.
2074 *) SSLv2 support has been removed. It still supports receiving a SSLv2
2075 compatible client hello.
2078 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
2079 done while fixing the error code for the key-too-small case.
2080 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
2082 *) CA.sh has been removed; use CA.pl instead.
2085 *) Removed old DES API.
2088 *) Remove various unsupported platforms:
2094 Sinix/ReliantUNIX RM400
2099 16-bit platforms such as WIN16
2102 *) Clean up OPENSSL_NO_xxx #define's
2103 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
2104 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
2105 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
2106 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
2107 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
2108 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
2109 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
2110 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
2111 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
2112 Remove MS_STATIC; it's a relic from platforms <32 bits.
2115 *) Cleaned up dead code
2116 Remove all but one '#ifdef undef' which is to be looked at.
2119 *) Clean up calling of xxx_free routines.
2120 Just like free(), fix most of the xxx_free routines to accept
2121 NULL. Remove the non-null checks from callers. Save much code.
2124 *) Add secure heap for storage of private keys (when possible).
2125 Add BIO_s_secmem(), CBIGNUM, etc.
2126 Contributed by Akamai Technologies under our Corporate CLA.
2129 *) Experimental support for a new, fast, unbiased prime candidate generator,
2130 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
2131 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
2133 *) New output format NSS in the sess_id command line tool. This allows
2134 exporting the session id and the master key in NSS keylog format.
2135 [Martin Kaiser <martin@kaiser.cx>]
2137 *) Harmonize version and its documentation. -f flag is used to display
2139 [mancha <mancha1@zoho.com>]
2141 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2142 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
2143 [mancha <mancha1@zoho.com>]
2145 *) Fix some double frees. These are not thought to be exploitable.
2146 [mancha <mancha1@zoho.com>]
2148 *) A missing bounds check in the handling of the TLS heartbeat extension
2149 can be used to reveal up to 64k of memory to a connected client or
2152 Thanks for Neel Mehta of Google Security for discovering this bug and to
2153 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2154 preparing the fix (CVE-2014-0160)
2155 [Adam Langley, Bodo Moeller]
2157 *) Fix for the attack described in the paper "Recovering OpenSSL
2158 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2159 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2160 http://eprint.iacr.org/2014/140
2162 Thanks to Yuval Yarom and Naomi Benger for discovering this
2163 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2164 [Yuval Yarom and Naomi Benger]
2166 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2167 this fixes a limitation in previous versions of OpenSSL.
2170 *) Experimental encrypt-then-mac support.
2172 Experimental support for encrypt then mac from
2173 draft-gutmann-tls-encrypt-then-mac-02.txt
2175 To enable it set the appropriate extension number (0x42 for the test
2176 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
2178 For non-compliant peers (i.e. just about everything) this should have no
2181 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
2185 *) Add EVP support for key wrapping algorithms, to avoid problems with
2186 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2187 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2188 algorithms and include tests cases.
2191 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2195 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2196 MGF1 digest and OAEP label.
2199 *) Make openssl verify return errors.
2200 [Chris Palmer <palmer@google.com> and Ben Laurie]
2202 *) New function ASN1_TIME_diff to calculate the difference between two
2203 ASN1_TIME structures or one structure and the current time.
2206 *) Update fips_test_suite to support multiple command line options. New
2207 test to induce all self test errors in sequence and check expected
2211 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2212 sign or verify all in one operation.
2215 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2216 test programs and fips_test_suite. Includes functionality to parse
2217 the minimal script output of fipsalgest.pl directly.
2220 *) Add authorisation parameter to FIPS_module_mode_set().
2223 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2226 *) Use separate DRBG fields for internal and external flags. New function
2227 FIPS_drbg_health_check() to perform on demand health checking. Add
2228 generation tests to fips_test_suite with reduced health check interval to
2229 demonstrate periodic health checking. Add "nodh" option to
2230 fips_test_suite to skip very slow DH test.
2233 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2237 *) More extensive health check for DRBG checking many more failure modes.
2238 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2239 combination: call this in fips_test_suite.
2242 *) Add support for canonical generation of DSA parameter 'g'. See
2245 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2246 POST to handle HMAC cases.
2249 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2250 to return numerical and string versions of the FIPS module number.
2253 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2254 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2255 outside the validated module in the FIPS capable OpenSSL.
2258 *) Minor change to DRBG entropy callback semantics. In some cases
2259 there is no multiple of the block length between min_len and
2260 max_len. Allow the callback to return more than max_len bytes
2261 of entropy but discard any extra: it is the callback's responsibility
2262 to ensure that the extra data discarded does not impact the
2263 requested amount of entropy.
2266 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2267 information in FIPS186-3, SP800-57 and SP800-131A.
2270 *) CCM support via EVP. Interface is very similar to GCM case except we
2271 must supply all data in one chunk (i.e. no update, final) and the
2272 message length must be supplied if AAD is used. Add algorithm test
2276 *) Initial version of POST overhaul. Add POST callback to allow the status
2277 of POST to be monitored and/or failures induced. Modify fips_test_suite
2278 to use callback. Always run all selftests even if one fails.
2281 *) XTS support including algorithm test driver in the fips_gcmtest program.
2282 Note: this does increase the maximum key length from 32 to 64 bytes but
2283 there should be no binary compatibility issues as existing applications
2284 will never use XTS mode.
2287 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2288 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2289 performs algorithm blocking for unapproved PRNG types. Also do not
2290 set PRNG type in FIPS_mode_set(): leave this to the application.
2291 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2292 the standard OpenSSL PRNG: set additional data to a date time vector.
2295 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2296 This shouldn't present any incompatibility problems because applications
2297 shouldn't be using these directly and any that are will need to rethink
2298 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2301 *) Extensive self tests and health checking required by SP800-90 DRBG.
2302 Remove strength parameter from FIPS_drbg_instantiate and always
2303 instantiate at maximum supported strength.
2306 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2309 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2312 *) New function DH_compute_key_padded() to compute a DH key and pad with
2313 leading zeroes if needed: this complies with SP800-56A et al.
2316 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2317 anything, incomplete, subject to change and largely untested at present.
2320 *) Modify fipscanisteronly build option to only build the necessary object
2321 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2324 *) Add experimental option FIPSSYMS to give all symbols in
2325 fipscanister.o and FIPS or fips prefix. This will avoid
2326 conflicts with future versions of OpenSSL. Add perl script
2327 util/fipsas.pl to preprocess assembly language source files
2328 and rename any affected symbols.
2331 *) Add selftest checks and algorithm block of non-fips algorithms in
2332 FIPS mode. Remove DES2 from selftests.
2335 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2336 return internal method without any ENGINE dependencies. Add new
2337 tiny fips sign and verify functions.
2340 *) New build option no-ec2m to disable characteristic 2 code.
2343 *) New build option "fipscanisteronly". This only builds fipscanister.o
2344 and (currently) associated fips utilities. Uses the file Makefile.fips
2345 instead of Makefile.org as the prototype.
2348 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2349 Update fips_gcmtest to use IV generator.
2352 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2353 setting output buffer to NULL. The *Final function must be
2354 called although it will not retrieve any additional data. The tag
2355 can be set or retrieved with a ctrl. The IV length is by default 12
2356 bytes (96 bits) but can be set to an alternative value. If the IV
2357 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2361 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2362 underlying do_cipher function handles all cipher semantics itself
2363 including padding and finalisation. This is useful if (for example)
2364 an ENGINE cipher handles block padding itself. The behaviour of
2365 do_cipher is subtly changed if this flag is set: the return value
2366 is the number of characters written to the output buffer (zero is
2367 no longer an error code) or a negative error code. Also if the
2368 input buffer is NULL and length 0 finalisation should be performed.
2371 *) If a candidate issuer certificate is already part of the constructed
2372 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2375 *) Improve forward-security support: add functions
2377 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2378 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2380 for use by SSL/TLS servers; the callback function will be called whenever a
2381 new session is created, and gets to decide whether the session may be
2382 cached to make it resumable (return 0) or not (return 1). (As by the
2383 SSL/TLS protocol specifications, the session_id sent by the server will be
2384 empty to indicate that the session is not resumable; also, the server will
2385 not generate RFC 4507 (RFC 5077) session tickets.)
2387 A simple reasonable callback implementation is to return is_forward_secure.
2388 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2389 by the SSL/TLS server library, indicating whether it can provide forward
2391 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2393 *) New -verify_name option in command line utilities to set verification
2397 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2398 Add CMAC pkey methods.
2401 *) Experimental renegotiation in s_server -www mode. If the client
2402 browses /reneg connection is renegotiated. If /renegcert it is
2403 renegotiated requesting a certificate.
2406 *) Add an "external" session cache for debugging purposes to s_server. This
2407 should help trace issues which normally are only apparent in deployed
2408 multi-process servers.
2411 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2412 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2413 BIO_set_cipher() and some obscure PEM functions were changed so they
2414 can now return an error. The RAND changes required a change to the
2415 RAND_METHOD structure.
2418 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2419 a gcc attribute to warn if the result of a function is ignored. This
2420 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2421 whose return value is often ignored.
2424 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2425 These allow SCTs (signed certificate timestamps) to be requested and
2426 validated when establishing a connection.
2427 [Rob Percival <robpercival@google.com>]
2429 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2431 *) Prevent padding oracle in AES-NI CBC MAC check
2433 A MITM attacker can use a padding oracle attack to decrypt traffic
2434 when the connection uses an AES CBC cipher and the server support
2437 This issue was introduced as part of the fix for Lucky 13 padding
2438 attack (CVE-2013-0169). The padding check was rewritten to be in
2439 constant time by making sure that always the same bytes are read and
2440 compared against either the MAC or padding bytes. But it no longer
2441 checked that there was enough data to have both the MAC and padding
2444 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2448 *) Fix EVP_EncodeUpdate overflow
2450 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2451 Base64 encoding of binary data. If an attacker is able to supply very large
2452 amounts of input data then a length check can overflow resulting in a heap
2455 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2456 the PEM_write_bio* family of functions. These are mainly used within the
2457 OpenSSL command line applications, so any application which processes data
2458 from an untrusted source and outputs it as a PEM file should be considered
2459 vulnerable to this issue. User applications that call these APIs directly
2460 with large amounts of untrusted data may also be vulnerable.
2462 This issue was reported by Guido Vranken.
2466 *) Fix EVP_EncryptUpdate overflow
2468 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2469 is able to supply very large amounts of input data after a previous call to
2470 EVP_EncryptUpdate() with a partial block then a length check can overflow
2471 resulting in a heap corruption. Following an analysis of all OpenSSL
2472 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2473 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2474 the first called function after an EVP_EncryptInit(), and therefore that
2475 specific call must be safe. The second form is where the length passed to
2476 EVP_EncryptUpdate() can be seen from the code to be some small value and
2477 therefore there is no possibility of an overflow. Since all instances are
2478 one of these two forms, it is believed that there can be no overflows in
2479 internal code due to this problem. It should be noted that
2480 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2481 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2482 of these calls have also been analysed too and it is believed there are no
2483 instances in internal usage where an overflow could occur.
2485 This issue was reported by Guido Vranken.
2489 *) Prevent ASN.1 BIO excessive memory allocation
2491 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2492 a short invalid encoding can cause allocation of large amounts of memory
2493 potentially consuming excessive resources or exhausting memory.
2495 Any application parsing untrusted data through d2i BIO functions is
2496 affected. The memory based functions such as d2i_X509() are *not* affected.
2497 Since the memory based functions are used by the TLS library, TLS
2498 applications are not affected.
2500 This issue was reported by Brian Carpenter.
2506 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2507 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2508 in arbitrary stack data being returned in the buffer.
2510 This issue was reported by Guido Vranken.
2514 *) Modify behavior of ALPN to invoke callback after SNI/servername
2515 callback, such that updates to the SSL_CTX affect ALPN.
2518 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2522 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2523 methods are enabled and ssl2 is disabled the methods return NULL.
2526 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2528 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2529 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2530 provide any "EXPORT" or "LOW" strength ciphers.
2533 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2534 is by default disabled at build-time. Builds that are not configured with
2535 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2536 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2537 will need to explicitly call either of:
2539 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2541 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2543 as appropriate. Even if either of those is used, or the application
2544 explicitly uses the version-specific SSLv2_method() or its client and
2545 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2546 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2547 ciphers, and SSLv2 56-bit DES are no longer available.
2551 *) Fix a double-free in DSA code
2553 A double free bug was discovered when OpenSSL parses malformed DSA private
2554 keys and could lead to a DoS attack or memory corruption for applications
2555 that receive DSA private keys from untrusted sources. This scenario is
2558 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2563 *) Disable SRP fake user seed to address a server memory leak.
2565 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2567 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2568 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2569 was changed to ignore the "fake user" SRP seed, even if the seed
2572 Users should use SRP_VBASE_get1_by_user instead. Note that in
2573 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2574 also that even though configuring the SRP seed attempts to hide
2575 invalid usernames by continuing the handshake with fake
2576 credentials, this behaviour is not constant time and no strong
2577 guarantees are made that the handshake is indistinguishable from
2578 that of a valid user.
2582 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2584 In the BN_hex2bn function the number of hex digits is calculated using an
2585 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2586 large values of |i| this can result in |bn_expand| not allocating any
2587 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2588 field as NULL leading to a subsequent NULL ptr deref. For very large values
2589 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2590 In this case memory is allocated to the internal BIGNUM data field, but it
2591 is insufficiently sized leading to heap corruption. A similar issue exists
2592 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2593 is ever called by user applications with very large untrusted hex/dec data.
2594 This is anticipated to be a rare occurrence.
2596 All OpenSSL internal usage of these functions use data that is not expected
2597 to be untrusted, e.g. config file data or application command line
2598 arguments. If user developed applications generate config file data based
2599 on untrusted data then it is possible that this could also lead to security
2600 consequences. This is also anticipated to be rare.
2602 This issue was reported to OpenSSL by Guido Vranken.
2606 *) Fix memory issues in BIO_*printf functions
2608 The internal |fmtstr| function used in processing a "%s" format string in
2609 the BIO_*printf functions could overflow while calculating the length of a
2610 string and cause an OOB read when printing very long strings.
2612 Additionally the internal |doapr_outch| function can attempt to write to an
2613 OOB memory location (at an offset from the NULL pointer) in the event of a
2614 memory allocation failure. In 1.0.2 and below this could be caused where
2615 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2616 could be in processing a very long "%s" format string. Memory leaks can
2619 The first issue may mask the second issue dependent on compiler behaviour.
2620 These problems could enable attacks where large amounts of untrusted data
2621 is passed to the BIO_*printf functions. If applications use these functions
2622 in this way then they could be vulnerable. OpenSSL itself uses these
2623 functions when printing out human-readable dumps of ASN.1 data. Therefore
2624 applications that print this data could be vulnerable if the data is from
2625 untrusted sources. OpenSSL command line applications could also be
2626 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2627 as command line arguments.
2629 Libssl is not considered directly vulnerable. Additionally certificates etc
2630 received via remote connections via libssl are also unlikely to be able to
2631 trigger these issues because of message size limits enforced within libssl.
2633 This issue was reported to OpenSSL Guido Vranken.
2637 *) Side channel attack on modular exponentiation
2639 A side-channel attack was found which makes use of cache-bank conflicts on
2640 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2641 of RSA keys. The ability to exploit this issue is limited as it relies on
2642 an attacker who has control of code in a thread running on the same
2643 hyper-threaded core as the victim thread which is performing decryptions.
2645 This issue was reported to OpenSSL by Yuval Yarom, The University of
2646 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2647 Nadia Heninger, University of Pennsylvania with more information at
2648 http://cachebleed.info.
2652 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2653 if no keysize is specified with default_bits. This fixes an
2654 omission in an earlier change that changed all RSA/DSA key generation
2655 apps to use 2048 bits by default.
2658 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2659 *) DH small subgroups
2661 Historically OpenSSL only ever generated DH parameters based on "safe"
2662 primes. More recently (in version 1.0.2) support was provided for
2663 generating X9.42 style parameter files such as those required for RFC 5114
2664 support. The primes used in such files may not be "safe". Where an
2665 application is using DH configured with parameters based on primes that are
2666 not "safe" then an attacker could use this fact to find a peer's private
2667 DH exponent. This attack requires that the attacker complete multiple
2668 handshakes in which the peer uses the same private DH exponent. For example
2669 this could be used to discover a TLS server's private DH exponent if it's
2670 reusing the private DH exponent or it's using a static DH ciphersuite.
2672 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2673 TLS. It is not on by default. If the option is not set then the server
2674 reuses the same private DH exponent for the life of the server process and
2675 would be vulnerable to this attack. It is believed that many popular
2676 applications do set this option and would therefore not be at risk.
2678 The fix for this issue adds an additional check where a "q" parameter is
2679 available (as is the case in X9.42 based parameters). This detects the
2680 only known attack, and is the only possible defense for static DH
2681 ciphersuites. This could have some performance impact.
2683 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2684 default and cannot be disabled. This could have some performance impact.
2686 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2690 *) SSLv2 doesn't block disabled ciphers
2692 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2693 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2694 been disabled, provided that the SSLv2 protocol was not also disabled via
2697 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2698 and Sebastian Schinzel.
2702 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2704 *) BN_mod_exp may produce incorrect results on x86_64
2706 There is a carry propagating bug in the x86_64 Montgomery squaring
2707 procedure. No EC algorithms are affected. Analysis suggests that attacks
2708 against RSA and DSA as a result of this defect would be very difficult to
2709 perform and are not believed likely. Attacks against DH are considered just
2710 feasible (although very difficult) because most of the work necessary to
2711 deduce information about a private key may be performed offline. The amount
2712 of resources required for such an attack would be very significant and
2713 likely only accessible to a limited number of attackers. An attacker would
2714 additionally need online access to an unpatched system using the target
2715 private key in a scenario with persistent DH parameters and a private
2716 key that is shared between multiple clients. For example this can occur by
2717 default in OpenSSL DHE based SSL/TLS ciphersuites.
2719 This issue was reported to OpenSSL by Hanno Böck.
2723 *) Certificate verify crash with missing PSS parameter
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 absent mask generation function parameter. Since these
2728 routines are used to verify certificate signature algorithms this can be
2729 used to crash any certificate verification operation and exploited in a
2730 DoS attack. Any application which performs certificate verification is
2731 vulnerable including OpenSSL clients and servers which enable client
2734 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2738 *) X509_ATTRIBUTE memory leak
2740 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2741 memory. This structure is used by the PKCS#7 and CMS routines so any
2742 application which reads PKCS#7 or CMS data from untrusted sources is
2743 affected. SSL/TLS is not affected.
2745 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2750 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2751 This changes the decoding behaviour for some invalid messages,
2752 though the change is mostly in the more lenient direction, and
2753 legacy behaviour is preserved as much as possible.
2756 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2758 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2760 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2762 *) Alternate chains certificate forgery
2764 During certificate verification, OpenSSL will attempt to find an
2765 alternative certificate chain if the first attempt to build such a chain
2766 fails. An error in the implementation of this logic can mean that an
2767 attacker could cause certain checks on untrusted certificates to be
2768 bypassed, such as the CA flag, enabling them to use a valid leaf
2769 certificate to act as a CA and "issue" an invalid certificate.
2771 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2775 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2777 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2778 incompatibility in the handling of HMAC. The previous ABI has now been
2782 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2784 *) Malformed ECParameters causes infinite loop
2786 When processing an ECParameters structure OpenSSL enters an infinite loop
2787 if the curve specified is over a specially malformed binary polynomial
2790 This can be used to perform denial of service against any
2791 system which processes public keys, certificate requests or
2792 certificates. This includes TLS clients and TLS servers with
2793 client authentication enabled.
2795 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2799 *) Exploitable out-of-bounds read in X509_cmp_time
2801 X509_cmp_time does not properly check the length of the ASN1_TIME
2802 string and can read a few bytes out of bounds. In addition,
2803 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2806 An attacker can use this to craft malformed certificates and CRLs of
2807 various sizes and potentially cause a segmentation fault, resulting in
2808 a DoS on applications that verify certificates or CRLs. TLS clients
2809 that verify CRLs are affected. TLS clients and servers with client
2810 authentication enabled may be affected if they use custom verification
2813 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2814 independently by Hanno Böck.
2818 *) PKCS7 crash with missing EnvelopedContent
2820 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2821 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2822 with missing content and trigger a NULL pointer dereference on parsing.
2824 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2825 structures from untrusted sources are affected. OpenSSL clients and
2826 servers are not affected.
2828 This issue was reported to OpenSSL by Michal Zalewski (Google).
2832 *) CMS verify infinite loop with unknown hash function
2834 When verifying a signedData message the CMS code can enter an infinite loop
2835 if presented with an unknown hash function OID. This can be used to perform
2836 denial of service against any system which verifies signedData messages using
2838 This issue was reported to OpenSSL by Johannes Bauer.
2842 *) Race condition handling NewSessionTicket
2844 If a NewSessionTicket is received by a multi-threaded client when attempting to
2845 reuse a previous ticket then a race condition can occur potentially leading to
2846 a double free of the ticket data.
2850 *) Only support 256-bit or stronger elliptic curves with the
2851 'ecdh_auto' setting (server) or by default (client). Of supported
2852 curves, prefer P-256 (both).
2855 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2857 *) ClientHello sigalgs DoS fix
2859 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2860 invalid signature algorithms extension a NULL pointer dereference will
2861 occur. This can be exploited in a DoS attack against the server.
2863 This issue was was reported to OpenSSL by David Ramos of Stanford
2866 [Stephen Henson and Matt Caswell]
2868 *) Multiblock corrupted pointer fix
2870 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2871 feature only applies on 64 bit x86 architecture platforms that support AES
2872 NI instructions. A defect in the implementation of "multiblock" can cause
2873 OpenSSL's internal write buffer to become incorrectly set to NULL when
2874 using non-blocking IO. Typically, when the user application is using a
2875 socket BIO for writing, this will only result in a failed connection.
2876 However if some other BIO is used then it is likely that a segmentation
2877 fault will be triggered, thus enabling a potential DoS attack.
2879 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2883 *) Segmentation fault in DTLSv1_listen fix
2885 The DTLSv1_listen function is intended to be stateless and processes the
2886 initial ClientHello from many peers. It is common for user code to loop
2887 over the call to DTLSv1_listen until a valid ClientHello is received with
2888 an associated cookie. A defect in the implementation of DTLSv1_listen means
2889 that state is preserved in the SSL object from one invocation to the next
2890 that can lead to a segmentation fault. Errors processing the initial
2891 ClientHello can trigger this scenario. An example of such an error could be
2892 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2895 This issue was reported to OpenSSL by Per Allansson.
2899 *) Segmentation fault in ASN1_TYPE_cmp fix
2901 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2902 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2903 certificate signature algorithm consistency this can be used to crash any
2904 certificate verification operation and exploited in a DoS attack. Any
2905 application which performs certificate verification is vulnerable including
2906 OpenSSL clients and servers which enable client authentication.
2910 *) Segmentation fault for invalid PSS parameters fix
2912 The signature verification routines will crash with a NULL pointer
2913 dereference if presented with an ASN.1 signature using the RSA PSS
2914 algorithm and invalid parameters. Since these routines are used to verify
2915 certificate signature algorithms this can be used to crash any
2916 certificate verification operation and exploited in a DoS attack. Any
2917 application which performs certificate verification is vulnerable including
2918 OpenSSL clients and servers which enable client authentication.
2920 This issue was was reported to OpenSSL by Brian Carpenter.
2924 *) ASN.1 structure reuse memory corruption fix
2926 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2927 memory corruption via an invalid write. Such reuse is and has been
2928 strongly discouraged and is believed to be rare.
2930 Applications that parse structures containing CHOICE or ANY DEFINED BY
2931 components may be affected. Certificate parsing (d2i_X509 and related
2932 functions) are however not affected. OpenSSL clients and servers are
2937 *) PKCS7 NULL pointer dereferences fix
2939 The PKCS#7 parsing code does not handle missing outer ContentInfo
2940 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2941 missing content and trigger a NULL pointer dereference on parsing.
2943 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2944 otherwise parse PKCS#7 structures from untrusted sources are
2945 affected. OpenSSL clients and servers are not affected.
2947 This issue was reported to OpenSSL by Michal Zalewski (Google).
2951 *) DoS via reachable assert in SSLv2 servers fix
2953 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2954 servers that both support SSLv2 and enable export cipher suites by sending
2955 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2957 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2958 (OpenSSL development team).
2962 *) Empty CKE with client auth and DHE fix
2964 If client auth is used then a server can seg fault in the event of a DHE
2965 ciphersuite being selected and a zero length ClientKeyExchange message
2966 being sent by the client. This could be exploited in a DoS attack.
2970 *) Handshake with unseeded PRNG fix
2972 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2973 with an unseeded PRNG. The conditions are:
2974 - The client is on a platform where the PRNG has not been seeded
2975 automatically, and the user has not seeded manually
2976 - A protocol specific client method version has been used (i.e. not
2977 SSL_client_methodv23)
2978 - A ciphersuite is used that does not require additional random data from
2979 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2981 If the handshake succeeds then the client random that has been used will
2982 have been generated from a PRNG with insufficient entropy and therefore the
2983 output may be predictable.
2985 For example using the following command with an unseeded openssl will
2986 succeed on an unpatched platform:
2988 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2992 *) Use After Free following d2i_ECPrivatekey error fix
2994 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2995 could cause a use after free condition. This, in turn, could cause a double
2996 free in several private key parsing functions (such as d2i_PrivateKey
2997 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2998 for applications that receive EC private keys from untrusted
2999 sources. This scenario is considered rare.
3001 This issue was discovered by the BoringSSL project and fixed in their
3006 *) X509_to_X509_REQ NULL pointer deref fix
3008 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
3009 the certificate key is invalid. This function is rarely used in practice.
3011 This issue was discovered by Brian Carpenter.
3015 *) Removed the export ciphers from the DEFAULT ciphers
3018 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
3020 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
3021 ARMv5 through ARMv8, as opposite to "locking" it to single one.
3022 So far those who have to target multiple platforms would compromise
3023 and argue that binary targeting say ARMv5 would still execute on
3024 ARMv8. "Universal" build resolves this compromise by providing
3025 near-optimal performance even on newer platforms.
3028 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
3029 (other platforms pending).
3030 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
3032 *) Add support for the SignedCertificateTimestampList certificate and
3033 OCSP response extensions from RFC6962.
3036 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3037 for corner cases. (Certain input points at infinity could lead to
3038 bogus results, with non-infinity inputs mapped to infinity too.)
3041 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
3042 This covers AES, SHA256/512 and GHASH. "Initial" means that most
3043 common cases are optimized and there still is room for further
3044 improvements. Vector Permutation AES for Altivec is also added.
3047 *) Add support for little-endian ppc64 Linux target.
3048 [Marcelo Cerri (IBM)]
3050 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
3051 SHA1, SHA256 and GHASH. "Initial" means that most common cases
3052 are optimized and there still is room for further improvements.
3053 Both 32- and 64-bit modes are supported.
3054 [Andy Polyakov, Ard Biesheuvel (Linaro)]
3056 *) Improved ARMv7 NEON support.
3059 *) Support for SPARC Architecture 2011 crypto extensions, first
3060 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
3061 SHA256/512, MD5, GHASH and modular exponentiation.
3062 [Andy Polyakov, David Miller]
3064 *) Accelerated modular exponentiation for Intel processors, a.k.a.
3066 [Shay Gueron & Vlad Krasnov (Intel Corp)]
3068 *) Support for new and upcoming Intel processors, including AVX2,
3069 BMI and SHA ISA extensions. This includes additional "stitched"
3070 implementations, AESNI-SHA256 and GCM, and multi-buffer support
3073 This work was sponsored by Intel Corp.
3076 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
3077 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
3078 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
3081 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3082 this fixes a limitation in previous versions of OpenSSL.
3085 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3086 MGF1 digest and OAEP label.
3089 *) Add EVP support for key wrapping algorithms, to avoid problems with
3090 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3091 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3092 algorithms and include tests cases.
3095 *) Add functions to allocate and set the fields of an ECDSA_METHOD
3097 [Douglas E. Engert, Steve Henson]
3099 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
3100 difference in days and seconds between two tm or ASN1_TIME structures.
3103 *) Add -rev test option to s_server to just reverse order of characters
3104 received by client and send back to server. Also prints an abbreviated
3105 summary of the connection parameters.
3108 *) New option -brief for s_client and s_server to print out a brief summary
3109 of connection parameters.
3112 *) Add callbacks for arbitrary TLS extensions.
3113 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
3115 *) New option -crl_download in several openssl utilities to download CRLs
3116 from CRLDP extension in certificates.
3119 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
3122 *) New function X509_CRL_diff to generate a delta CRL from the difference
3123 of two full CRLs. Add support to "crl" utility.
3126 *) New functions to set lookup_crls function and to retrieve
3127 X509_STORE from X509_STORE_CTX.
3130 *) Print out deprecated issuer and subject unique ID fields in
3134 *) Extend OCSP I/O functions so they can be used for simple general purpose
3135 HTTP as well as OCSP. New wrapper function which can be used to download
3136 CRLs using the OCSP API.
3139 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
3142 *) SSL_CONF* functions. These provide a common framework for application
3143 configuration using configuration files or command lines.
3146 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
3147 message callback and prints the results. Needs compile time option
3148 "enable-ssl-trace". New options to s_client and s_server to enable
3152 *) New ctrl and macro to retrieve supported points extensions.
3153 Print out extension in s_server and s_client.
3156 *) New functions to retrieve certificate signature and signature
3160 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
3164 *) New Suite B modes for TLS code. These use and enforce the requirements
3165 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
3166 only use Suite B curves. The Suite B modes can be set by using the
3167 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
3170 *) New chain verification flags for Suite B levels of security. Check
3171 algorithms are acceptable when flags are set in X509_verify_cert.
3174 *) Make tls1_check_chain return a set of flags indicating checks passed
3175 by a certificate chain. Add additional tests to handle client
3176 certificates: checks for matching certificate type and issuer name
3180 *) If an attempt is made to use a signature algorithm not in the peer
3181 preference list abort the handshake. If client has no suitable
3182 signature algorithms in response to a certificate request do not
3183 use the certificate.
3186 *) If server EC tmp key is not in client preference list abort handshake.
3189 *) Add support for certificate stores in CERT structure. This makes it
3190 possible to have different stores per SSL structure or one store in
3191 the parent SSL_CTX. Include distinct stores for certificate chain
3192 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3193 to build and store a certificate chain in CERT structure: returning
3194 an error if the chain cannot be built: this will allow applications
3195 to test if a chain is correctly configured.
3197 Note: if the CERT based stores are not set then the parent SSL_CTX
3198 store is used to retain compatibility with existing behaviour.
3202 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3203 mask based on the current session, check mask when sending client
3204 hello and checking the requested ciphersuite.
3207 *) New ctrls to retrieve and set certificate types in a certificate
3208 request message. Print out received values in s_client. If certificate
3209 types is not set with custom values set sensible values based on
3210 supported signature algorithms.
3213 *) Support for distinct client and server supported signature algorithms.
3216 *) Add certificate callback. If set this is called whenever a certificate
3217 is required by client or server. An application can decide which
3218 certificate chain to present based on arbitrary criteria: for example
3219 supported signature algorithms. Add very simple example to s_server.
3220 This fixes many of the problems and restrictions of the existing client
3221 certificate callback: for example you can now clear an existing
3222 certificate and specify the whole chain.
3225 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3226 the certificate can be used for (if anything). Set valid_flags field
3227 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3228 to have similar checks in it.
3230 Add new "cert_flags" field to CERT structure and include a "strict mode".
3231 This enforces some TLS certificate requirements (such as only permitting
3232 certificate signature algorithms contained in the supported algorithms
3233 extension) which some implementations ignore: this option should be used
3234 with caution as it could cause interoperability issues.
3237 *) Update and tidy signature algorithm extension processing. Work out
3238 shared signature algorithms based on preferences and peer algorithms
3239 and print them out in s_client and s_server. Abort handshake if no
3240 shared signature algorithms.
3243 *) Add new functions to allow customised supported signature algorithms
3244 for SSL and SSL_CTX structures. Add options to s_client and s_server
3248 *) New function SSL_certs_clear() to delete all references to certificates
3249 from an SSL structure. Before this once a certificate had been added
3250 it couldn't be removed.
3253 *) Integrate hostname, email address and IP address checking with certificate
3254 verification. New verify options supporting checking in openssl utility.
3257 *) Fixes and wildcard matching support to hostname and email checking
3258 functions. Add manual page.
3259 [Florian Weimer (Red Hat Product Security Team)]
3261 *) New functions to check a hostname email or IP address against a
3262 certificate. Add options x509 utility to print results of checks against
3266 *) Fix OCSP checking.
3267 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3269 *) Initial experimental support for explicitly trusted non-root CAs.
3270 OpenSSL still tries to build a complete chain to a root but if an
3271 intermediate CA has a trust setting included that is used. The first
3272 setting is used: whether to trust (e.g., -addtrust option to the x509
3276 *) Add -trusted_first option which attempts to find certificates in the
3277 trusted store even if an untrusted chain is also supplied.
3280 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3281 platform support for Linux and Android.
3284 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3287 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3288 When in FIPS mode the approved implementations are used as normal,
3289 when not in FIPS mode the internal unapproved versions are used instead.
3290 This means that the FIPS capable OpenSSL isn't forced to use the
3291 (often lower performance) FIPS implementations outside FIPS mode.
3294 *) Transparently support X9.42 DH parameters when calling
3295 PEM_read_bio_DHparameters. This means existing applications can handle
3296 the new parameter format automatically.
3299 *) Initial experimental support for X9.42 DH parameter format: mainly
3300 to support use of 'q' parameter for RFC5114 parameters.
3303 *) Add DH parameters from RFC5114 including test data to dhtest.
3306 *) Support for automatic EC temporary key parameter selection. If enabled
3307 the most preferred EC parameters are automatically used instead of
3308 hardcoded fixed parameters. Now a server just has to call:
3309 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3310 support ECDH and use the most appropriate parameters.
3313 *) Enhance and tidy EC curve and point format TLS extension code. Use
3314 static structures instead of allocation if default values are used.
3315 New ctrls to set curves we wish to support and to retrieve shared curves.
3316 Print out shared curves in s_server. New options to s_server and s_client
3317 to set list of supported curves.
3320 *) New ctrls to retrieve supported signature algorithms and
3321 supported curve values as an array of NIDs. Extend openssl utility
3322 to print out received values.
3325 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3326 between NIDs and the more common NIST names such as "P-256". Enhance
3327 ecparam utility and ECC method to recognise the NIST names for curves.
3330 *) Enhance SSL/TLS certificate chain handling to support different
3331 chains for each certificate instead of one chain in the parent SSL_CTX.
3334 *) Support for fixed DH ciphersuite client authentication: where both
3335 server and client use DH certificates with common parameters.
3338 *) Support for fixed DH ciphersuites: those requiring DH server
3342 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3344 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3345 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3346 X509_CINF_get_signature were reverted post internal team review.
3348 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3350 *) Build fixes for the Windows and OpenVMS platforms
3351 [Matt Caswell and Richard Levitte]
3353 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3355 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3356 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3357 dereference. This could lead to a Denial Of Service attack. Thanks to
3358 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3362 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3363 dtls1_buffer_record function under certain conditions. In particular this
3364 could occur if an attacker sent repeated DTLS records with the same
3365 sequence number but for the next epoch. The memory leak could be exploited
3366 by an attacker in a Denial of Service attack through memory exhaustion.
3367 Thanks to Chris Mueller for reporting this issue.
3371 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3372 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3373 method would be set to NULL which could later result in a NULL pointer
3374 dereference. Thanks to Frank Schmirler for reporting this issue.
3378 *) Abort handshake if server key exchange message is omitted for ephemeral
3381 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3382 reporting this issue.
3386 *) Remove non-export ephemeral RSA code on client and server. This code
3387 violated the TLS standard by allowing the use of temporary RSA keys in
3388 non-export ciphersuites and could be used by a server to effectively
3389 downgrade the RSA key length used to a value smaller than the server
3390 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3391 INRIA or reporting this issue.
3395 *) Fixed issue where DH client certificates are accepted without verification.
3396 An OpenSSL server will accept a DH certificate for client authentication
3397 without the certificate verify message. This effectively allows a client to
3398 authenticate without the use of a private key. This only affects servers
3399 which trust a client certificate authority which issues certificates
3400 containing DH keys: these are extremely rare and hardly ever encountered.
3401 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3406 *) Ensure that the session ID context of an SSL is updated when its
3407 SSL_CTX is updated via SSL_set_SSL_CTX.
3409 The session ID context is typically set from the parent SSL_CTX,
3410 and can vary with the CTX.
3413 *) Fix various certificate fingerprint issues.
3415 By using non-DER or invalid encodings outside the signed portion of a
3416 certificate the fingerprint can be changed without breaking the signature.
3417 Although no details of the signed portion of the certificate can be changed
3418 this can cause problems with some applications: e.g. those using the
3419 certificate fingerprint for blacklists.
3421 1. Reject signatures with non zero unused bits.
3423 If the BIT STRING containing the signature has non zero unused bits reject
3424 the signature. All current signature algorithms require zero unused bits.
3426 2. Check certificate algorithm consistency.
3428 Check the AlgorithmIdentifier inside TBS matches the one in the
3429 certificate signature. NB: this will result in signature failure
3430 errors for some broken certificates.
3432 Thanks to Konrad Kraszewski from Google for reporting this issue.
3434 3. Check DSA/ECDSA signatures use DER.
3436 Re-encode DSA/ECDSA signatures and compare with the original received
3437 signature. Return an error if there is a mismatch.
3439 This will reject various cases including garbage after signature
3440 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3441 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3442 (negative or with leading zeroes).
3444 Further analysis was conducted and fixes were developed by Stephen Henson
3445 of the OpenSSL core team.
3450 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3451 results on some platforms, including x86_64. This bug occurs at random
3452 with a very low probability, and is not known to be exploitable in any
3453 way, though its exact impact is difficult to determine. Thanks to Pieter
3454 Wuille (Blockstream) who reported this issue and also suggested an initial
3455 fix. Further analysis was conducted by the OpenSSL development team and
3456 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3457 the OpenSSL core team.
3461 *) Do not resume sessions on the server if the negotiated protocol
3462 version does not match the session's version. Resuming with a different
3463 version, while not strictly forbidden by the RFC, is of questionable
3464 sanity and breaks all known clients.
3465 [David Benjamin, Emilia Käsper]
3467 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3468 early CCS messages during renegotiation. (Note that because
3469 renegotiation is encrypted, this early CCS was not exploitable.)
3472 *) Tighten client-side session ticket handling during renegotiation:
3473 ensure that the client only accepts a session ticket if the server sends
3474 the extension anew in the ServerHello. Previously, a TLS client would
3475 reuse the old extension state and thus accept a session ticket if one was
3476 announced in the initial ServerHello.
3478 Similarly, ensure that the client requires a session ticket if one
3479 was advertised in the ServerHello. Previously, a TLS client would
3480 ignore a missing NewSessionTicket message.
3483 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3485 *) SRTP Memory Leak.
3487 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3488 sends a carefully crafted handshake message, to cause OpenSSL to fail
3489 to free up to 64k of memory causing a memory leak. This could be
3490 exploited in a Denial Of Service attack. This issue affects OpenSSL
3491 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3492 whether SRTP is used or configured. Implementations of OpenSSL that
3493 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3495 The fix was developed by the OpenSSL team.
3499 *) Session Ticket Memory Leak.
3501 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3502 integrity of that ticket is first verified. In the event of a session
3503 ticket integrity check failing, OpenSSL will fail to free memory
3504 causing a memory leak. By sending a large number of invalid session
3505 tickets an attacker could exploit this issue in a Denial Of Service
3510 *) Build option no-ssl3 is incomplete.
3512 When OpenSSL is configured with "no-ssl3" as a build option, servers
3513 could accept and complete a SSL 3.0 handshake, and clients could be
3514 configured to send them.
3516 [Akamai and the OpenSSL team]
3518 *) Add support for TLS_FALLBACK_SCSV.
3519 Client applications doing fallback retries should call
3520 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3522 [Adam Langley, Bodo Moeller]
3524 *) Add additional DigestInfo checks.
3526 Re-encode DigestInto in DER and check against the original when
3527 verifying RSA signature: this will reject any improperly encoded
3528 DigestInfo structures.
3530 Note: this is a precautionary measure and no attacks are currently known.
3534 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3536 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3537 SRP code can be overrun an internal buffer. Add sanity check that
3538 g, A, B < N to SRP code.
3540 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3541 Group for discovering this issue.
3545 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3546 TLS 1.0 instead of higher protocol versions when the ClientHello message
3547 is badly fragmented. This allows a man-in-the-middle attacker to force a
3548 downgrade to TLS 1.0 even if both the server and the client support a
3549 higher protocol version, by modifying the client's TLS records.
3551 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3552 researching this issue.
3556 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3557 to a denial of service attack. A malicious server can crash the client
3558 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3559 ciphersuite and sending carefully crafted handshake messages.
3561 Thanks to Felix Gröbert (Google) for discovering and researching this
3566 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3567 to leak memory. This can be exploited through a Denial of Service attack.
3568 Thanks to Adam Langley for discovering and researching this issue.
3572 *) An attacker can force openssl to consume large amounts of memory whilst
3573 processing DTLS handshake messages. This can be exploited through a
3574 Denial of Service attack.
3575 Thanks to Adam Langley for discovering and researching this issue.
3579 *) An attacker can force an error condition which causes openssl to crash
3580 whilst processing DTLS packets due to memory being freed twice. This
3581 can be exploited through a Denial of Service attack.
3582 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3587 *) If a multithreaded client connects to a malicious server using a resumed
3588 session and the server sends an ec point format extension it could write
3589 up to 255 bytes to freed memory.
3591 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3596 *) A malicious server can crash an OpenSSL client with a null pointer
3597 dereference (read) by specifying an SRP ciphersuite even though it was not
3598 properly negotiated with the client. This can be exploited through a
3599 Denial of Service attack.
3601 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3602 discovering and researching this issue.
3606 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3607 X509_name_oneline, X509_name_print_ex et al. to leak some information
3608 from the stack. Applications may be affected if they echo pretty printing
3609 output to the attacker.
3611 Thanks to Ivan Fratric (Google) for discovering this issue.
3613 [Emilia Käsper, and Steve Henson]
3615 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3616 for corner cases. (Certain input points at infinity could lead to
3617 bogus results, with non-infinity inputs mapped to infinity too.)
3620 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3622 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3623 handshake can force the use of weak keying material in OpenSSL
3624 SSL/TLS clients and servers.
3626 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3627 researching this issue. (CVE-2014-0224)
3628 [KIKUCHI Masashi, Steve Henson]
3630 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3631 OpenSSL DTLS client the code can be made to recurse eventually crashing
3634 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3636 [Imre Rad, Steve Henson]
3638 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3639 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3640 client or server. This is potentially exploitable to run arbitrary
3641 code on a vulnerable client or server.
3643 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3644 [Jüri Aedla, Steve Henson]
3646 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3647 are subject to a denial of service attack.
3649 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3650 this issue. (CVE-2014-3470)
3651 [Felix Gröbert, Ivan Fratric, Steve Henson]
3653 *) Harmonize version and its documentation. -f flag is used to display
3655 [mancha <mancha1@zoho.com>]
3657 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3658 in i2d_ECPrivateKey.
3659 [mancha <mancha1@zoho.com>]
3661 *) Fix some double frees. These are not thought to be exploitable.
3662 [mancha <mancha1@zoho.com>]
3664 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3666 *) A missing bounds check in the handling of the TLS heartbeat extension
3667 can be used to reveal up to 64k of memory to a connected client or
3670 Thanks for Neel Mehta of Google Security for discovering this bug and to
3671 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3672 preparing the fix (CVE-2014-0160)
3673 [Adam Langley, Bodo Moeller]
3675 *) Fix for the attack described in the paper "Recovering OpenSSL
3676 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3677 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3678 http://eprint.iacr.org/2014/140
3680 Thanks to Yuval Yarom and Naomi Benger for discovering this
3681 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3682 [Yuval Yarom and Naomi Benger]
3684 *) TLS pad extension: draft-agl-tls-padding-03
3686 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3687 TLS client Hello record length value would otherwise be > 255 and
3688 less that 512 pad with a dummy extension containing zeroes so it
3689 is at least 512 bytes long.
3691 [Adam Langley, Steve Henson]
3693 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3695 *) Fix for TLS record tampering bug. A carefully crafted invalid
3696 handshake could crash OpenSSL with a NULL pointer exception.
3697 Thanks to Anton Johansson for reporting this issues.
3700 *) Keep original DTLS digest and encryption contexts in retransmission
3701 structures so we can use the previous session parameters if they need
3702 to be resent. (CVE-2013-6450)
3705 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3706 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3707 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3708 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3709 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3710 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3711 [Rob Stradling, Adam Langley]
3713 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3715 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3716 supporting platforms or when small records were transferred.
3717 [Andy Polyakov, Steve Henson]
3719 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3721 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3723 This addresses the flaw in CBC record processing discovered by
3724 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3725 at: http://www.isg.rhul.ac.uk/tls/
3727 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3728 Security Group at Royal Holloway, University of London
3729 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3730 Emilia Käsper for the initial patch.
3732 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3734 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3735 ciphersuites which can be exploited in a denial of service attack.
3736 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3737 and detecting this bug and to Wolfgang Ettlinger
3738 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3742 *) Return an error when checking OCSP signatures when key is NULL.
3743 This fixes a DoS attack. (CVE-2013-0166)
3746 *) Make openssl verify return errors.
3747 [Chris Palmer <palmer@google.com> and Ben Laurie]
3749 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3750 the right response is stapled. Also change SSL_get_certificate()
3751 so it returns the certificate actually sent.
3752 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3753 [Rob Stradling <rob.stradling@comodo.com>]
3755 *) Fix possible deadlock when decoding public keys.
3758 *) Don't use TLS 1.0 record version number in initial client hello
3762 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3764 *) Sanity check record length before skipping explicit IV in TLS
3765 1.2, 1.1 and DTLS to fix DoS attack.
3767 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3768 fuzzing as a service testing platform.
3772 *) Initialise tkeylen properly when encrypting CMS messages.
3773 Thanks to Solar Designer of Openwall for reporting this issue.
3776 *) In FIPS mode don't try to use composite ciphers as they are not
3780 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3782 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3783 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3784 mean any application compiled against OpenSSL 1.0.0 headers setting
3785 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3786 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3787 0x10000000L Any application which was previously compiled against
3788 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3789 will need to be recompiled as a result. Letting be results in
3790 inability to disable specifically TLS 1.1 and in client context,
3791 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3794 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3795 disable just protocol X, but all protocols above X *if* there are
3796 protocols *below* X still enabled. In more practical terms it means
3797 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3798 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3799 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3803 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3805 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3806 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3807 in CRYPTO_realloc_clean.
3809 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3810 issue and to Adam Langley <agl@chromium.org> for fixing it.
3812 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3814 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3817 *) Workarounds for some broken servers that "hang" if a client hello
3818 record length exceeds 255 bytes.
3820 1. Do not use record version number > TLS 1.0 in initial client
3821 hello: some (but not all) hanging servers will now work.
3822 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3823 the number of ciphers sent in the client hello. This should be
3824 set to an even number, such as 50, for example by passing:
3825 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3826 Most broken servers should now work.
3827 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3828 TLS 1.2 client support entirely.
3831 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3834 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3836 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3837 STRING form instead of a DigestInfo.
3840 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3841 and the RSA_sign/RSA_verify functions. This was made more apparent when
3842 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3843 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3844 the correct format in RSA_verify so both forms transparently work.
3847 *) Some servers which support TLS 1.0 can choke if we initially indicate
3848 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3849 encrypted premaster secret. As a workaround use the maximum permitted
3850 client version in client hello, this should keep such servers happy
3851 and still work with previous versions of OpenSSL.
3854 *) Add support for TLS/DTLS heartbeats.
3855 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3857 *) Add support for SCTP.
3858 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3860 *) Improved PRNG seeding for VOS.
3861 [Paul Green <Paul.Green@stratus.com>]
3863 *) Extensive assembler packs updates, most notably:
3865 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3866 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3867 - x86_64: bit-sliced AES implementation;
3868 - ARM: NEON support, contemporary platforms optimizations;
3869 - s390x: z196 support;
3870 - *: GHASH and GF(2^m) multiplication implementations;
3874 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3875 (removal of unnecessary code)
3876 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3878 *) Add TLS key material exporter from RFC 5705.
3881 *) Add DTLS-SRTP negotiation from RFC 5764.
3884 *) Add Next Protocol Negotiation,
3885 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3886 disabled with a no-npn flag to config or Configure. Code donated
3888 [Adam Langley <agl@google.com> and Ben Laurie]
3890 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3891 NIST-P256, NIST-P521, with constant-time single point multiplication on
3892 typical inputs. Compiler support for the nonstandard type __uint128_t is
3893 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3894 Code made available under Apache License version 2.0.
3896 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3897 line to include this in your build of OpenSSL, and run "make depend" (or
3898 "make update"). This enables the following EC_METHODs:
3900 EC_GFp_nistp224_method()
3901 EC_GFp_nistp256_method()
3902 EC_GFp_nistp521_method()
3904 EC_GROUP_new_by_curve_name() will automatically use these (while
3905 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3907 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3909 *) Use type ossl_ssize_t instead of ssize_t which isn't available on
3910 all platforms. Move ssize_t definition from e_os.h to the public
3911 header file e_os2.h as it now appears in public header file cms.h
3914 *) New -sigopt option to the ca, req and x509 utilities. Additional
3915 signature parameters can be passed using this option and in
3919 *) Add RSA PSS signing function. This will generate and set the
3920 appropriate AlgorithmIdentifiers for PSS based on those in the
3921 corresponding EVP_MD_CTX structure. No application support yet.
3924 *) Support for companion algorithm specific ASN1 signing routines.
3925 New function ASN1_item_sign_ctx() signs a pre-initialised
3926 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3927 the appropriate parameters.
3930 *) Add new algorithm specific ASN1 verification initialisation function
3931 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3932 handling will be the same no matter what EVP_PKEY_METHOD is used.
3933 Add a PSS handler to support verification of PSS signatures: checked
3934 against a number of sample certificates.
3937 *) Add signature printing for PSS. Add PSS OIDs.
3938 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3940 *) Add algorithm specific signature printing. An individual ASN1 method
3941 can now print out signatures instead of the standard hex dump.
3943 More complex signatures (e.g. PSS) can print out more meaningful
3944 information. Include DSA version that prints out the signature
3948 *) Password based recipient info support for CMS library: implementing
3952 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3953 neatly separates the code into cipher and PBE sections and is required
3954 for some algorithms that split PBES2 into separate pieces (such as
3955 password based CMS).
3958 *) Session-handling fixes:
3959 - Fix handling of connections that are resuming with a session ID,
3960 but also support Session Tickets.
3961 - Fix a bug that suppressed issuing of a new ticket if the client
3962 presented a ticket with an expired session.
3963 - Try to set the ticket lifetime hint to something reasonable.
3964 - Make tickets shorter by excluding irrelevant information.
3965 - On the client side, don't ignore renewed tickets.
3966 [Adam Langley, Bodo Moeller (Google)]
3968 *) Fix PSK session representation.
3971 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3973 This work was sponsored by Intel.
3976 *) Add GCM support to TLS library. Some custom code is needed to split
3977 the IV between the fixed (from PRF) and explicit (from TLS record)
3978 portions. This adds all GCM ciphersuites supported by RFC5288 and
3979 RFC5289. Generalise some AES* cipherstrings to include GCM and
3980 add a special AESGCM string for GCM only.
3983 *) Expand range of ctrls for AES GCM. Permit setting invocation
3984 field on decrypt and retrieval of invocation field only on encrypt.
3987 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3988 As required by RFC5289 these ciphersuites cannot be used if for
3989 versions of TLS earlier than 1.2.
3992 *) For FIPS capable OpenSSL interpret a NULL default public key method
3993 as unset and return the appropriate default but do *not* set the default.
3994 This means we can return the appropriate method in applications that
3995 switch between FIPS and non-FIPS modes.
3998 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3999 ENGINE is used then we cannot handle that in the FIPS module so we
4000 keep original code iff non-FIPS operations are allowed.
4003 *) Add -attime option to openssl utilities.
4004 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
4006 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
4009 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
4010 FIPS EC methods unconditionally for now.
4013 *) New build option no-ec2m to disable characteristic 2 code.
4016 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
4017 all cases can be covered as some introduce binary incompatibilities.
4020 *) Redirect RSA operations to FIPS module including keygen,
4021 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
4024 *) Add similar low level API blocking to ciphers.
4027 *) Low level digest APIs are not approved in FIPS mode: any attempt
4028 to use these will cause a fatal error. Applications that *really* want
4029 to use them can use the private_* version instead.
4032 *) Redirect cipher operations to FIPS module for FIPS builds.
4035 *) Redirect digest operations to FIPS module for FIPS builds.
4038 *) Update build system to add "fips" flag which will link in fipscanister.o
4039 for static and shared library builds embedding a signature if needed.
4042 *) Output TLS supported curves in preference order instead of numerical
4043 order. This is currently hardcoded for the highest order curves first.
4044 This should be configurable so applications can judge speed vs strength.
4047 *) Add TLS v1.2 server support for client authentication.
4050 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
4054 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
4055 FIPS modules versions.
4058 *) Add TLS v1.2 client side support for client authentication. Keep cache
4059 of handshake records longer as we don't know the hash algorithm to use
4060 until after the certificate request message is received.
4063 *) Initial TLS v1.2 client support. Add a default signature algorithms
4064 extension including all the algorithms we support. Parse new signature
4065 format in client key exchange. Relax some ECC signing restrictions for
4066 TLS v1.2 as indicated in RFC5246.
4069 *) Add server support for TLS v1.2 signature algorithms extension. Switch
4070 to new signature format when needed using client digest preference.
4071 All server ciphersuites should now work correctly in TLS v1.2. No client
4072 support yet and no support for client certificates.
4075 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
4076 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
4077 ciphersuites. At present only RSA key exchange ciphersuites work with
4078 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
4079 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
4080 and version checking.
4083 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
4084 with this defined it will not be affected by any changes to ssl internal
4085 structures. Add several utility functions to allow openssl application
4086 to work with OPENSSL_NO_SSL_INTERN defined.
4089 *) A long standing patch to add support for SRP from EdelWeb (Peter
4090 Sylvester and Christophe Renou) was integrated.
4091 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
4092 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
4095 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
4098 *) Permit abbreviated handshakes when renegotiating using the function
4099 SSL_renegotiate_abbreviated().
4100 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4102 *) Add call to ENGINE_register_all_complete() to
4103 ENGINE_load_builtin_engines(), so some implementations get used
4104 automatically instead of needing explicit application support.
4107 *) Add support for TLS key exporter as described in RFC5705.
4108 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
4110 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
4111 a few changes are required:
4113 Add SSL_OP_NO_TLSv1_1 flag.
4114 Add TLSv1_1 methods.
4115 Update version checking logic to handle version 1.1.
4116 Add explicit IV handling (ported from DTLS code).
4117 Add command line options to s_client/s_server.
4120 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
4122 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
4123 in CMS and PKCS7 code. When RSA decryption fails use a random key for
4124 content decryption and always return the same error. Note: this attack
4125 needs on average 2^20 messages so it only affects automated senders. The
4126 old behaviour can be re-enabled in the CMS code by setting the
4127 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
4128 an MMA defence is not necessary.
4129 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
4130 this issue. (CVE-2012-0884)
4133 *) Fix CVE-2011-4619: make sure we really are receiving a
4134 client hello before rejecting multiple SGC restarts. Thanks to
4135 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
4138 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
4140 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
4141 Thanks to Antonio Martin, Enterprise Secure Access Research and
4142 Development, Cisco Systems, Inc. for discovering this bug and
4143 preparing a fix. (CVE-2012-0050)
4146 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
4148 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
4149 of the Vaudenay padding oracle attack on CBC mode encryption
4150 which enables an efficient plaintext recovery attack against
4151 the OpenSSL implementation of DTLS. Their attack exploits timing
4152 differences arising during decryption processing. A research
4153 paper describing this attack can be found at:
4154 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
4155 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4156 Security Group at Royal Holloway, University of London
4157 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
4158 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
4159 for preparing the fix. (CVE-2011-4108)
4160 [Robin Seggelmann, Michael Tuexen]
4162 *) Clear bytes used for block padding of SSL 3.0 records.
4164 [Adam Langley (Google)]
4166 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
4167 Kadianakis <desnacked@gmail.com> for discovering this issue and
4168 Adam Langley for preparing the fix. (CVE-2011-4619)
4169 [Adam Langley (Google)]
4171 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
4172 [Andrey Kulikov <amdeich@gmail.com>]
4174 *) Prevent malformed RFC3779 data triggering an assertion failure.
4175 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
4176 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
4177 [Rob Austein <sra@hactrn.net>]
4179 *) Improved PRNG seeding for VOS.
4180 [Paul Green <Paul.Green@stratus.com>]
4182 *) Fix ssl_ciph.c set-up race.
4183 [Adam Langley (Google)]
4185 *) Fix spurious failures in ecdsatest.c.
4186 [Emilia Käsper (Google)]
4188 *) Fix the BIO_f_buffer() implementation (which was mixing different
4189 interpretations of the '..._len' fields).
4190 [Adam Langley (Google)]
4192 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4193 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4194 threads won't reuse the same blinding coefficients.
4196 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4197 lock to call BN_BLINDING_invert_ex, and avoids one use of
4198 BN_BLINDING_update for each BN_BLINDING structure (previously,
4199 the last update always remained unused).
4200 [Emilia Käsper (Google)]
4202 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4203 [Bob Buckholz (Google)]
4205 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4207 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4208 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4209 [Kaspar Brand <ossl@velox.ch>]
4211 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4212 for multi-threaded use of ECDH. (CVE-2011-3210)
4213 [Adam Langley (Google)]
4215 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4218 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4219 signature public key algorithm by using OID xref utilities instead.
4220 Before this you could only use some ECC ciphersuites with SHA1 only.
4223 *) Add protection against ECDSA timing attacks as mentioned in the paper
4224 by Billy Bob Brumley and Nicola Tuveri, see:
4226 http://eprint.iacr.org/2011/232.pdf
4228 [Billy Bob Brumley and Nicola Tuveri]
4230 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4232 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4233 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4235 *) Fix bug in string printing code: if *any* escaping is enabled we must
4236 escape the escape character (backslash) or the resulting string is
4240 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4242 *) Disable code workaround for ancient and obsolete Netscape browsers
4243 and servers: an attacker can use it in a ciphersuite downgrade attack.
4244 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4247 *) Fixed J-PAKE implementation error, originally discovered by
4248 Sebastien Martini, further info and confirmation from Stefan
4249 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4252 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4254 *) Fix extension code to avoid race conditions which can result in a buffer
4255 overrun vulnerability: resumed sessions must not be modified as they can
4256 be shared by multiple threads. CVE-2010-3864
4259 *) Fix WIN32 build system to correctly link an ENGINE directory into
4263 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4265 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4267 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4269 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4271 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4272 context. The operation can be customised via the ctrl mechanism in
4273 case ENGINEs want to include additional functionality.
4276 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4279 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4280 output hashes compatible with older versions of OpenSSL.
4281 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4283 *) Fix compression algorithm handling: if resuming a session use the
4284 compression algorithm of the resumed session instead of determining
4285 it from client hello again. Don't allow server to change algorithm.
4288 *) Add load_crls() function to apps tidying load_certs() too. Add option
4289 to verify utility to allow additional CRLs to be included.
4292 *) Update OCSP request code to permit adding custom headers to the request:
4293 some responders need this.
4296 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4298 [Julia Lawall <julia@diku.dk>]
4300 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4301 needlessly dereferenced structures, used obsolete functions and
4302 didn't handle all updated verify codes correctly.
4305 *) Disable MD2 in the default configuration.
4308 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4309 indicate the initial BIO being pushed or popped. This makes it possible
4310 to determine whether the BIO is the one explicitly called or as a result
4311 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4312 it handles reference counts correctly and doesn't zero out the I/O bio
4313 when it is not being explicitly popped. WARNING: applications which
4314 included workarounds for the old buggy behaviour will need to be modified
4315 or they could free up already freed BIOs.
4318 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4319 renaming to all platforms (within the 0.9.8 branch, this was
4320 done conditionally on Netware platforms to avoid a name clash).
4321 [Guenter <lists@gknw.net>]
4323 *) Add ECDHE and PSK support to DTLS.
4324 [Michael Tuexen <tuexen@fh-muenster.de>]
4326 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4330 *) Add "missing" function EVP_MD_flags() (without this the only way to
4331 retrieve a digest flags is by accessing the structure directly. Update
4332 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4333 or cipher is registered as in the "from" argument. Print out all
4334 registered digests in the dgst usage message instead of manually
4335 attempting to work them out.
4338 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4339 this allows the use of compression and extensions. Change default cipher
4340 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4341 by default unless an application cipher string requests it.
4344 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4345 key ids to find matching certificates and keys but some PKCS#12 files
4346 don't follow the (somewhat unwritten) rules and this strategy fails.
4347 Now just gather all certificates together and the first private key
4348 then look for the first certificate that matches the key.
4351 *) Support use of registered digest and cipher names for dgst and cipher
4352 commands instead of having to add each one as a special case. So now
4359 openssl dgst -sha256 foo
4361 and this works for ENGINE based algorithms too.
4365 *) Update Gost ENGINE to support parameter files.
4366 [Victor B. Wagner <vitus@cryptocom.ru>]
4368 *) Support GeneralizedTime in ca utility.
4369 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4371 *) Enhance the hash format used for certificate directory links. The new
4372 form uses the canonical encoding (meaning equivalent names will work
4373 even if they aren't identical) and uses SHA1 instead of MD5. This form
4374 is incompatible with the older format and as a result c_rehash should
4375 be used to rebuild symbolic links.
4378 *) Make PKCS#8 the default write format for private keys, replacing the
4379 traditional format. This form is standardised, more secure and doesn't
4380 include an implicit MD5 dependency.
4383 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4384 committed to OpenSSL should pass this lot as a minimum.
4387 *) Add session ticket override functionality for use by EAP-FAST.
4388 [Jouni Malinen <j@w1.fi>]
4390 *) Modify HMAC functions to return a value. Since these can be implemented
4391 in an ENGINE errors can occur.
4394 *) Type-checked OBJ_bsearch_ex.
4397 *) Type-checked OBJ_bsearch. Also some constification necessitated
4398 by type-checking. Still to come: TXT_DB, bsearch(?),
4399 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4403 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4404 seconds to a tm structure directly, instead of going through OS
4405 specific date routines. This avoids any issues with OS routines such
4406 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4407 and X509_time_adj_ex() to cover the extended range. The existing
4408 X509_time_adj() is still usable and will no longer have any date issues.
4411 *) Delta CRL support. New use deltas option which will attempt to locate
4412 and search any appropriate delta CRLs available.
4414 This work was sponsored by Google.
4417 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4418 code and add additional score elements. Validate alternate CRL paths
4419 as part of the CRL checking and indicate a new error "CRL path validation
4420 error" in this case. Applications wanting additional details can use
4421 the verify callback and check the new "parent" field. If this is not
4422 NULL CRL path validation is taking place. Existing applications won't
4423 see this because it requires extended CRL support which is off by
4426 This work was sponsored by Google.
4429 *) Support for freshest CRL extension.
4431 This work was sponsored by Google.
4434 *) Initial indirect CRL support. Currently only supported in the CRLs
4435 passed directly and not via lookup. Process certificate issuer
4436 CRL entry extension and lookup CRL entries by bother issuer name
4437 and serial number. Check and process CRL issuer entry in IDP extension.
4439 This work was sponsored by Google.
4442 *) Add support for distinct certificate and CRL paths. The CRL issuer
4443 certificate is validated separately in this case. Only enabled if
4444 an extended CRL support flag is set: this flag will enable additional
4445 CRL functionality in future.
4447 This work was sponsored by Google.
4450 *) Add support for policy mappings extension.
4452 This work was sponsored by Google.
4455 *) Fixes to pathlength constraint, self issued certificate handling,
4456 policy processing to align with RFC3280 and PKITS tests.
4458 This work was sponsored by Google.
4461 *) Support for name constraints certificate extension. DN, email, DNS
4462 and URI types are currently supported.
4464 This work was sponsored by Google.
4467 *) To cater for systems that provide a pointer-based thread ID rather
4468 than numeric, deprecate the current numeric thread ID mechanism and
4469 replace it with a structure and associated callback type. This
4470 mechanism allows a numeric "hash" to be extracted from a thread ID in
4471 either case, and on platforms where pointers are larger than 'long',
4472 mixing is done to help ensure the numeric 'hash' is usable even if it
4473 can't be guaranteed unique. The default mechanism is to use "&errno"
4474 as a pointer-based thread ID to distinguish between threads.
4476 Applications that want to provide their own thread IDs should now use
4477 CRYPTO_THREADID_set_callback() to register a callback that will call
4478 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4480 Note that ERR_remove_state() is now deprecated, because it is tied
4481 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4482 to free the current thread's error state should be replaced by
4483 ERR_remove_thread_state(NULL).
4485 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4486 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4487 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4488 application was previously providing a numeric thread callback that
4489 was inappropriate for distinguishing threads, then uniqueness might
4490 have been obtained with &errno that happened immediately in the
4491 intermediate development versions of OpenSSL; this is no longer the
4492 case, the numeric thread callback will now override the automatic use
4494 [Geoff Thorpe, with help from Bodo Moeller]
4496 *) Initial support for different CRL issuing certificates. This covers a
4497 simple case where the self issued certificates in the chain exist and
4498 the real CRL issuer is higher in the existing chain.
4500 This work was sponsored by Google.
4503 *) Removed effectively defunct crypto/store from the build.
4506 *) Revamp of STACK to provide stronger type-checking. Still to come:
4507 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4508 ASN1_STRING, CONF_VALUE.
4511 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4512 RAM on SSL connections. This option can save about 34k per idle SSL.
4515 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4516 STACK, TXT_DB, bsearch, qsort.
4519 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4520 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4521 support for data, signedData, compressedData, digestedData and
4522 encryptedData, envelopedData types included. Scripts to check against
4523 RFC4134 examples draft and interop and consistency checks of many
4524 content types and variants.
4527 *) Add options to enc utility to support use of zlib compression BIO.
4530 *) Extend mk1mf to support importing of options and assembly language
4531 files from Configure script, currently only included in VC-WIN32.
4532 The assembly language rules can now optionally generate the source
4533 files from the associated perl scripts.
4536 *) Implement remaining functionality needed to support GOST ciphersuites.
4537 Interop testing has been performed using CryptoPro implementations.
4538 [Victor B. Wagner <vitus@cryptocom.ru>]
4540 *) s390x assembler pack.
4543 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4547 *) Implement Opaque PRF Input TLS extension as specified in
4548 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4549 official specification yet and no extension type assignment by
4550 IANA exists, this extension (for now) will have to be explicitly
4551 enabled when building OpenSSL by providing the extension number
4552 to use. For example, specify an option
4554 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4556 to the "config" or "Configure" script to enable the extension,
4557 assuming extension number 0x9527 (which is a completely arbitrary
4558 and unofficial assignment based on the MD5 hash of the Internet
4559 Draft). Note that by doing so, you potentially lose
4560 interoperability with other TLS implementations since these might
4561 be using the same extension number for other purposes.
4563 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4564 opaque PRF input value to use in the handshake. This will create
4565 an internal copy of the length-'len' string at 'src', and will
4566 return non-zero for success.
4568 To get more control and flexibility, provide a callback function
4571 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4572 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4576 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4579 Callback function 'cb' will be called in handshakes, and is
4580 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4581 Argument 'arg' is for application purposes (the value as given to
4582 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4583 be provided to the callback function). The callback function
4584 has to return non-zero to report success: usually 1 to use opaque
4585 PRF input just if possible, or 2 to enforce use of the opaque PRF
4586 input. In the latter case, the library will abort the handshake
4587 if opaque PRF input is not successfully negotiated.
4589 Arguments 'peerinput' and 'len' given to the callback function
4590 will always be NULL and 0 in the case of a client. A server will
4591 see the client's opaque PRF input through these variables if
4592 available (NULL and 0 otherwise). Note that if the server
4593 provides an opaque PRF input, the length must be the same as the
4594 length of the client's opaque PRF input.
4596 Note that the callback function will only be called when creating
4597 a new session (session resumption can resume whatever was
4598 previously negotiated), and will not be called in SSL 2.0
4599 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4600 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4601 for applications that need to enforce opaque PRF input.
4605 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4608 [Victor B. Wagner <vitus@cryptocom.ru>]
4610 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4611 RFC4507bis. The encrypted ticket format is an encrypted encoded
4612 SSL_SESSION structure, that way new session features are automatically
4615 If a client application caches session in an SSL_SESSION structure
4616 support is transparent because tickets are now stored in the encoded
4619 The SSL_CTX structure automatically generates keys for ticket
4620 protection in servers so again support should be possible
4621 with no application modification.
4623 If a client or server wishes to disable RFC4507 support then the option
4624 SSL_OP_NO_TICKET can be set.
4626 Add a TLS extension debugging callback to allow the contents of any client
4627 or server extensions to be examined.
4629 This work was sponsored by Google.
4632 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4633 OpenSSL should now compile cleanly on gcc 4.2
4634 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4636 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4637 support including streaming MAC support: this is required for GOST
4638 ciphersuite support.
4639 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4641 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4642 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4643 to output in BER and PEM format.
4646 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4647 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4648 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4649 ENGINE support for HMAC keys which are unextractable. New -mac and
4650 -macopt options to dgst utility.
4653 *) New option -sigopt to dgst utility. Update dgst to use
4654 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4655 alternative signing parameters such as X9.31 or PSS in the dgst
4659 *) Change ssl_cipher_apply_rule(), the internal function that does
4660 the work each time a ciphersuite string requests enabling
4661 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4662 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4663 the order of disabled ciphersuites such that those ciphersuites
4664 that most recently went from enabled to disabled not only stay
4665 in order with respect to each other, but also have higher priority
4666 than other disabled ciphersuites the next time ciphersuites are
4669 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4670 the same ciphersuites as with "HIGH" alone, but in a specific
4671 order where the PSK ciphersuites come first (since they are the
4672 most recently disabled ciphersuites when "HIGH" is parsed).
4674 Also, change ssl_create_cipher_list() (using this new
4675 functionality) such that between otherwise identical
4676 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4680 *) Change ssl_create_cipher_list() so that it automatically
4681 arranges the ciphersuites in reasonable order before starting
4682 to process the rule string. Thus, the definition for "DEFAULT"
4683 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4684 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4685 This makes it much easier to arrive at a reasonable default order
4686 in applications for which anonymous ciphers are OK (meaning
4687 that you can't actually use DEFAULT).
4688 [Bodo Moeller; suggested by Victor Duchovni]
4690 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4691 processing) into multiple integers instead of setting
4692 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4693 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4694 (These masks as well as the individual bit definitions are hidden
4695 away into the non-exported interface ssl/ssl_locl.h, so this
4696 change to the definition of the SSL_CIPHER structure shouldn't
4697 affect applications.) This give us more bits for each of these
4698 categories, so there is no longer a need to coagulate AES128 and
4699 AES256 into a single algorithm bit, and to coagulate Camellia128
4700 and Camellia256 into a single algorithm bit, which has led to all
4703 Thus, among other things, the kludge introduced in 0.9.7m and
4704 0.9.8e for masking out AES256 independently of AES128 or masking
4705 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4707 With the change, we also introduce new ciphersuite aliases that
4708 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4712 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4713 Use the leftmost N bytes of the signature input if the input is
4714 larger than the prime q (with N being the size in bytes of q).
4717 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4718 it yet and it is largely untested.
4721 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4724 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4725 some compilers (gcc 4.2 and later) reject their use. Safestack is
4726 reimplemented. Update ASN1 to avoid use of legacy functions.
4729 *) Win32/64 targets are linked with Winsock2.
4732 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4733 to external functions. This can be used to increase CRL handling
4734 efficiency especially when CRLs are very large by (for example) storing
4735 the CRL revoked certificates in a database.
4738 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4739 new CRLs added to a directory can be used. New command line option
4740 -verify_return_error to s_client and s_server. This causes real errors
4741 to be returned by the verify callback instead of carrying on no matter
4742 what. This reflects the way a "real world" verify callback would behave.
4745 *) GOST engine, supporting several GOST algorithms and public key formats.
4746 Kindly donated by Cryptocom.
4749 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4750 partitioned by DP are handled but no indirect CRL or reason partitioning
4751 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4752 selected via a scoring technique which handles IDP and AKID in CRLs.
4755 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4756 will ultimately be used for all verify operations: this will remove the
4757 X509_STORE dependency on certificate verification and allow alternative
4758 lookup methods. X509_STORE based implementations of these two callbacks.
4761 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4762 Modify get_crl() to find a valid (unexpired) CRL if possible.
4765 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4766 this would be called X509_CRL_cmp() but that name is already used by
4767 a function that just compares CRL issuer names. Cache several CRL
4768 extensions in X509_CRL structure and cache CRLDP in X509.
4771 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4772 this maps equivalent X509_NAME structures into a consistent structure.
4773 Name comparison can then be performed rapidly using memcmp().
4776 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4780 *) Allow digests to supply their own micalg string for S/MIME type using
4781 the ctrl EVP_MD_CTRL_MICALG.
4784 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4785 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4786 ctrl. It can then customise the structure before and/or after signing
4790 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4791 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4792 to free up any added signature OIDs.
4795 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4796 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4797 digest and cipher tables. New options added to openssl utility:
4798 list-message-digest-algorithms and list-cipher-algorithms.
4801 *) Change the array representation of binary polynomials: the list
4802 of degrees of non-zero coefficients is now terminated with -1.
4803 Previously it was terminated with 0, which was also part of the
4804 value; thus, the array representation was not applicable to
4805 polynomials where t^0 has coefficient zero. This change makes
4806 the array representation useful in a more general context.
4809 *) Various modifications and fixes to SSL/TLS cipher string
4810 handling. For ECC, the code now distinguishes between fixed ECDH
4811 with RSA certificates on the one hand and with ECDSA certificates
4812 on the other hand, since these are separate ciphersuites. The
4813 unused code for Fortezza ciphersuites has been removed.
4815 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4816 (not "ECDHE"). For consistency with the code for DH
4817 certificates, use of ECDH certificates is now considered ECDH
4818 authentication, not RSA or ECDSA authentication (the latter is
4819 merely the CA's signing algorithm and not actively used in the
4822 The temporary ciphersuite alias "ECCdraft" is no longer
4823 available, and ECC ciphersuites are no longer excluded from "ALL"
4824 and "DEFAULT". The following aliases now exist for RFC 4492
4825 ciphersuites, most of these by analogy with the DH case:
4827 kECDHr - ECDH cert, signed with RSA
4828 kECDHe - ECDH cert, signed with ECDSA
4829 kECDH - ECDH cert (signed with either RSA or ECDSA)
4830 kEECDH - ephemeral ECDH
4831 ECDH - ECDH cert or ephemeral ECDH
4837 AECDH - anonymous ECDH
4838 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4842 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4843 Use correct micalg parameters depending on digest(s) in signed message.
4846 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4847 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4850 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4851 an engine to register a method. Add ENGINE lookups for methods and
4852 functional reference processing.
4855 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4856 EVP_{Sign,Verify}* which allow an application to customise the signature
4860 *) New -resign option to smime utility. This adds one or more signers
4861 to an existing PKCS#7 signedData structure. Also -md option to use an
4862 alternative message digest algorithm for signing.
4865 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4866 create PKCS7 structures containing multiple signers. Update smime
4867 application to support multiple signers.
4870 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4874 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4875 Reorganize PBE internals to lookup from a static table using NIDs,
4876 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4877 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4878 PRF which will be automatically used with PBES2.
4881 *) Replace the algorithm specific calls to generate keys in "req" with the
4885 *) Update PKCS#7 enveloped data routines to use new API. This is now
4886 supported by any public key method supporting the encrypt operation. A
4887 ctrl is added to allow the public key algorithm to examine or modify
4888 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4892 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4893 a default digest type to use. In most cases this will be SHA1 but some
4894 algorithms (such as GOST) need to specify an alternative digest. The
4895 return value indicates how strong the preference is 1 means optional and
4896 2 is mandatory (that is it is the only supported type). Modify
4897 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4898 use the default md. Update openssl utilities to use the default digest
4899 type for signing if it is not explicitly indicated.
4902 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4903 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4904 signing method from the key type. This effectively removes the link
4905 between digests and public key types.
4908 *) Add an OID cross reference table and utility functions. Its purpose is to
4909 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4910 rsaEncryption. This will allow some of the algorithm specific hackery
4911 needed to use the correct OID to be removed.
4914 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4915 structures for PKCS7_sign(). They are now set up by the relevant public
4919 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4922 *) Add support for key derivation (agreement) in the API, DH method and
4926 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4927 public and private key formats. As a side effect these add additional
4928 command line functionality not previously available: DSA signatures can be
4929 generated and verified using pkeyutl and DH key support and generation in
4934 [Oliver Tappe <zooey@hirschkaefer.de>]
4936 *) New make target "install_html_docs" installs HTML renditions of the
4938 [Oliver Tappe <zooey@hirschkaefer.de>]
4940 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4941 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4942 support key and parameter generation and add initial key generation
4943 functionality for RSA.
4946 *) Add functions for main EVP_PKEY_method operations. The undocumented
4947 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4948 EVP_PKEY_{encrypt,decrypt}_old.
4951 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4952 key API, doesn't do much yet.
4955 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4956 public key algorithms. New option to openssl utility:
4957 "list-public-key-algorithms" to print out info.
4960 *) Implement the Supported Elliptic Curves Extension for
4961 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4964 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4965 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4968 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4969 utilities such as rsa, dsa, dsaparam etc except they process any key
4973 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4974 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4975 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4979 *) Initial support for pluggable public key ASN1.
4980 De-spaghettify the public key ASN1 handling. Move public and private
4981 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4982 algorithm specific handling to a single module within the relevant
4983 algorithm directory. Add functions to allow (near) opaque processing
4984 of public and private key structures.
4987 *) Implement the Supported Point Formats Extension for
4988 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4991 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4992 for the psk identity [hint] and the psk callback functions to the
4993 SSL_SESSION, SSL and SSL_CTX structure.
4996 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
5000 SSL_CTX_use_psk_identity_hint
5001 SSL_get_psk_identity_hint
5002 SSL_get_psk_identity
5003 SSL_use_psk_identity_hint
5005 [Mika Kousa and Pasi Eronen of Nokia Corporation]
5007 *) Add RFC 3161 compliant time stamp request creation, response generation
5008 and response verification functionality.
5009 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
5011 *) Add initial support for TLS extensions, specifically for the server_name
5012 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5013 have new members for a host name. The SSL data structure has an
5014 additional member SSL_CTX *initial_ctx so that new sessions can be
5015 stored in that context to allow for session resumption, even after the
5016 SSL has been switched to a new SSL_CTX in reaction to a client's
5017 server_name extension.
5019 New functions (subject to change):
5021 SSL_get_servername()
5022 SSL_get_servername_type()
5025 New CTRL codes and macros (subject to change):
5027 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5028 - SSL_CTX_set_tlsext_servername_callback()
5029 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5030 - SSL_CTX_set_tlsext_servername_arg()
5031 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5033 openssl s_client has a new '-servername ...' option.
5035 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5036 '-key2 ...', '-servername_fatal' (subject to change). This allows
5037 testing the HostName extension for a specific single host name ('-cert'
5038 and '-key' remain fallbacks for handshakes without HostName
5039 negotiation). If the unrecognized_name alert has to be sent, this by
5040 default is a warning; it becomes fatal with the '-servername_fatal'
5043 [Peter Sylvester, Remy Allais, Christophe Renou]
5045 *) Whirlpool hash implementation is added.
5048 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
5049 bn(64,32). Because of instruction set limitations it doesn't have
5050 any negative impact on performance. This was done mostly in order
5051 to make it possible to share assembler modules, such as bn_mul_mont
5052 implementations, between 32- and 64-bit builds without hassle.
5055 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
5056 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
5060 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
5061 dedicated Montgomery multiplication procedure, is introduced.
5062 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
5063 "64-bit" performance on certain 32-bit targets.
5066 *) New option SSL_OP_NO_COMP to disable use of compression selectively
5067 in SSL structures. New SSL ctrl to set maximum send fragment size.
5068 Save memory by setting the I/O buffer sizes dynamically instead of
5069 using the maximum available value.
5072 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
5073 in addition to the text details.
5076 *) Very, very preliminary EXPERIMENTAL support for printing of general
5077 ASN1 structures. This currently produces rather ugly output and doesn't
5078 handle several customised structures at all.
5081 *) Integrated support for PVK file format and some related formats such
5082 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
5083 these in the 'rsa' and 'dsa' utilities.
5086 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
5089 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
5090 place for the (very old) "NETSCAPE" format certificates which are now
5091 handled using new ASN1 code equivalents.
5094 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
5095 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
5096 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
5099 *) Modify CRL distribution points extension code to print out previously
5100 unsupported fields. Enhance extension setting code to allow setting of
5104 *) Add print and set support for Issuing Distribution Point CRL extension.
5107 *) Change 'Configure' script to enable Camellia by default.
5110 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
5112 *) When rejecting SSL/TLS records due to an incorrect version number, never
5113 update s->server with a new major version number. As of
5114 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
5115 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
5116 the previous behavior could result in a read attempt at NULL when
5117 receiving specific incorrect SSL/TLS records once record payload
5118 protection is active. (CVE-2010-0740)
5119 [Bodo Moeller, Adam Langley <agl@chromium.org>]
5121 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
5122 could be crashed if the relevant tables were not present (e.g. chrooted).
5123 [Tomas Hoger <thoger@redhat.com>]
5125 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
5127 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
5128 [Martin Olsson, Neel Mehta]
5130 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
5131 accommodate for stack sorting, always a write lock!).
5134 *) On some versions of WIN32 Heap32Next is very slow. This can cause
5135 excessive delays in the RAND_poll(): over a minute. As a workaround
5136 include a time check in the inner Heap32Next loop too.
5139 *) The code that handled flushing of data in SSL/TLS originally used the
5140 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
5141 the problem outlined in PR#1949. The fix suggested there however can
5142 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
5143 of Apache). So instead simplify the code to flush unconditionally.
5144 This should be fine since flushing with no data to flush is a no op.
5147 *) Handle TLS versions 2.0 and later properly and correctly use the
5148 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
5149 off ancient servers have a habit of sticking around for a while...
5152 *) Modify compression code so it frees up structures without using the
5153 ex_data callbacks. This works around a problem where some applications
5154 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
5155 restarting) then use compression (e.g. SSL with compression) later.
5156 This results in significant per-connection memory leaks and
5157 has caused some security issues including CVE-2008-1678 and
5161 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
5162 change when encrypting or decrypting.
5165 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
5166 connect and renegotiate with servers which do not support RI.
5167 Until RI is more widely deployed this option is enabled by default.
5170 *) Add "missing" ssl ctrls to clear options and mode.
5173 *) If client attempts to renegotiate and doesn't support RI respond with
5174 a no_renegotiation alert as required by RFC5746. Some renegotiating
5175 TLS clients will continue a connection gracefully when they receive
5176 the alert. Unfortunately OpenSSL mishandled this alert and would hang
5177 waiting for a server hello which it will never receive. Now we treat a
5178 received no_renegotiation alert as a fatal error. This is because
5179 applications requesting a renegotiation might well expect it to succeed
5180 and would have no code in place to handle the server denying it so the
5181 only safe thing to do is to terminate the connection.
5184 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
5185 peer supports secure renegotiation and 0 otherwise. Print out peer
5186 renegotiation support in s_client/s_server.
5189 *) Replace the highly broken and deprecated SPKAC certification method with
5190 the updated NID creation version. This should correctly handle UTF8.
5193 *) Implement RFC5746. Re-enable renegotiation but require the extension
5194 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
5195 turns out to be a bad idea. It has been replaced by
5196 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5197 SSL_CTX_set_options(). This is really not recommended unless you
5198 know what you are doing.
5199 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5201 *) Fixes to stateless session resumption handling. Use initial_ctx when
5202 issuing and attempting to decrypt tickets in case it has changed during
5203 servername handling. Use a non-zero length session ID when attempting
5204 stateless session resumption: this makes it possible to determine if
5205 a resumption has occurred immediately after receiving server hello
5206 (several places in OpenSSL subtly assume this) instead of later in
5210 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5211 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5212 fixes for a few places where the return code is not checked
5214 [Julia Lawall <julia@diku.dk>]
5216 *) Add --strict-warnings option to Configure script to include devteam
5217 warnings in other configurations.
5220 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5221 makes it possible to install openssl libraries in locations which
5222 have names other than "lib", for example "/usr/lib64" which some
5224 [Steve Henson, based on patch from Jeremy Utley]
5226 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5227 X690 8.9.12 and can produce some misleading textual output of OIDs.
5228 [Steve Henson, reported by Dan Kaminsky]
5230 *) Delete MD2 from algorithm tables. This follows the recommendation in
5231 several standards that it is not used in new applications due to
5232 several cryptographic weaknesses. For binary compatibility reasons
5233 the MD2 API is still compiled in by default.
5236 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5240 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5241 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5243 [Guenter <lists@gknw.net>]
5245 *) Fix the server certificate chain building code to use X509_verify_cert(),
5246 it used to have an ad-hoc builder which was unable to cope with anything
5247 other than a simple chain.
5248 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5250 *) Don't check self signed certificate signatures in X509_verify_cert()
5251 by default (a flag can override this): it just wastes time without
5252 adding any security. As a useful side effect self signed root CAs
5253 with non-FIPS digests are now usable in FIPS mode.
5256 *) In dtls1_process_out_of_seq_message() the check if the current message
5257 is already buffered was missing. For every new message was memory
5258 allocated, allowing an attacker to perform an denial of service attack
5259 with sending out of seq handshake messages until there is no memory
5260 left. Additionally every future message was buffered, even if the
5261 sequence number made no sense and would be part of another handshake.
5262 So only messages with sequence numbers less than 10 in advance will be
5263 buffered. (CVE-2009-1378)
5264 [Robin Seggelmann, discovered by Daniel Mentz]
5266 *) Records are buffered if they arrive with a future epoch to be
5267 processed after finishing the corresponding handshake. There is
5268 currently no limitation to this buffer allowing an attacker to perform
5269 a DOS attack with sending records with future epochs until there is no
5270 memory left. This patch adds the pqueue_size() function to determine
5271 the size of a buffer and limits the record buffer to 100 entries.
5273 [Robin Seggelmann, discovered by Daniel Mentz]
5275 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5276 parent structure is freed. (CVE-2009-1379)
5279 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5280 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5283 [Ilya O. <vrghost@gmail.com>]
5285 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5287 *) Disable renegotiation completely - this fixes a severe security
5288 problem (CVE-2009-3555) at the cost of breaking all
5289 renegotiation. Renegotiation can be re-enabled by setting
5290 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5291 run-time. This is really not recommended unless you know what
5295 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5297 *) Don't set val to NULL when freeing up structures, it is freed up by
5298 underlying code. If sizeof(void *) > sizeof(long) this can result in
5299 zeroing past the valid field. (CVE-2009-0789)
5300 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5302 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5303 checked correctly. This would allow some invalid signed attributes to
5304 appear to verify correctly. (CVE-2009-0591)
5305 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5307 *) Reject UniversalString and BMPString types with invalid lengths. This
5308 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5309 a legal length. (CVE-2009-0590)
5312 *) Set S/MIME signing as the default purpose rather than setting it
5313 unconditionally. This allows applications to override it at the store
5317 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5318 to handle some structures.
5321 *) Improve efficiency of mem_gets: don't search whole buffer each time
5323 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5325 *) New -hex option for openssl rand.
5328 *) Print out UTF8String and NumericString when parsing ASN1.
5331 *) Support NumericString type for name components.
5334 *) Allow CC in the environment to override the automatically chosen
5335 compiler. Note that nothing is done to ensure flags work with the
5339 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5341 *) Properly check EVP_VerifyFinal() and similar return values
5343 [Ben Laurie, Bodo Moeller, Google Security Team]
5345 *) Enable TLS extensions by default.
5348 *) Allow the CHIL engine to be loaded, whether the application is
5349 multithreaded or not. (This does not release the developer from the
5350 obligation to set up the dynamic locking callbacks.)
5351 [Sander Temme <sander@temme.net>]
5353 *) Use correct exit code if there is an error in dgst command.
5354 [Steve Henson; problem pointed out by Roland Dirlewanger]
5356 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5357 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5360 *) Add experimental JPAKE support, including demo authentication in
5361 s_client and s_server.
5364 *) Set the comparison function in v3_addr_canonize().
5365 [Rob Austein <sra@hactrn.net>]
5367 *) Add support for XMPP STARTTLS in s_client.
5368 [Philip Paeps <philip@freebsd.org>]
5370 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5371 to ensure that even with this option, only ciphersuites in the
5372 server's preference list will be accepted. (Note that the option
5373 applies only when resuming a session, so the earlier behavior was
5374 just about the algorithm choice for symmetric cryptography.)
5377 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5379 *) Fix NULL pointer dereference if a DTLS server received
5380 ChangeCipherSpec as first record (CVE-2009-1386).
5383 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5384 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5387 *) The fix in 0.9.8c that supposedly got rid of unsafe
5388 double-checked locking was incomplete for RSA blinding,
5389 addressing just one layer of what turns out to have been
5390 doubly unsafe triple-checked locking.
5392 So now fix this for real by retiring the MONT_HELPER macro
5393 in crypto/rsa/rsa_eay.c.
5395 [Bodo Moeller; problem pointed out by Marius Schilder]
5397 *) Various precautionary measures:
5399 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5401 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5402 (NB: This would require knowledge of the secret session ticket key
5403 to exploit, in which case you'd be SOL either way.)
5405 - Change bn_nist.c so that it will properly handle input BIGNUMs
5406 outside the expected range.
5408 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5411 [Neel Mehta, Bodo Moeller]
5413 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5414 the load fails. Useful for distros.
5415 [Ben Laurie and the FreeBSD team]
5417 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5420 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5423 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5425 This work was sponsored by Logica.
5428 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5429 keystores. Support for SSL/TLS client authentication too.
5430 Not compiled unless enable-capieng specified to Configure.
5432 This work was sponsored by Logica.
5435 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5436 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5437 attribute creation routines such as certificate requests and PKCS#12
5441 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5443 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5444 handshake which could lead to a client crash as found using the
5445 Codenomicon TLS test suite (CVE-2008-1672)
5446 [Steve Henson, Mark Cox]
5448 *) Fix double free in TLS server name extensions which could lead to
5449 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5452 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5454 Clear the error queue to ensure that error entries left from
5455 older function calls do not interfere with the correct operation.
5456 [Lutz Jaenicke, Erik de Castro Lopo]
5458 *) Remove root CA certificates of commercial CAs:
5460 The OpenSSL project does not recommend any specific CA and does not
5461 have any policy with respect to including or excluding any CA.
5462 Therefore it does not make any sense to ship an arbitrary selection
5463 of root CA certificates with the OpenSSL software.
5466 *) RSA OAEP patches to fix two separate invalid memory reads.
5467 The first one involves inputs when 'lzero' is greater than
5468 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5469 before the beginning of from). The second one involves inputs where
5470 the 'db' section contains nothing but zeroes (there is a one-byte
5471 invalid read after the end of 'db').
5472 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5474 *) Partial backport from 0.9.9-dev:
5476 Introduce bn_mul_mont (dedicated Montgomery multiplication
5477 procedure) as a candidate for BIGNUM assembler implementation.
5478 While 0.9.9-dev uses assembler for various architectures, only
5479 x86_64 is available by default here in the 0.9.8 branch, and
5480 32-bit x86 is available through a compile-time setting.
5482 To try the 32-bit x86 assembler implementation, use Configure
5483 option "enable-montasm" (which exists only for this backport).
5485 As "enable-montasm" for 32-bit x86 disclaims code stability
5486 anyway, in this constellation we activate additional code
5487 backported from 0.9.9-dev for further performance improvements,
5488 namely BN_from_montgomery_word. (To enable this otherwise,
5489 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5491 [Andy Polyakov (backport partially by Bodo Moeller)]
5493 *) Add TLS session ticket callback. This allows an application to set
5494 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5495 values. This is useful for key rollover for example where several key
5496 sets may exist with different names.
5499 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5500 This was broken until now in 0.9.8 releases, such that the only way
5501 a registered ENGINE could be used (assuming it initialises
5502 successfully on the host) was to explicitly set it as the default
5503 for the relevant algorithms. This is in contradiction with 0.9.7
5504 behaviour and the documentation. With this fix, when an ENGINE is
5505 registered into a given algorithm's table of implementations, the
5506 'uptodate' flag is reset so that auto-discovery will be used next
5507 time a new context for that algorithm attempts to select an
5509 [Ian Lister (tweaked by Geoff Thorpe)]
5511 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5512 implementation in the following ways:
5514 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5517 Lack of BER streaming support means one pass streaming processing is
5518 only supported if data is detached: setting the streaming flag is
5519 ignored for embedded content.
5521 CMS support is disabled by default and must be explicitly enabled
5522 with the enable-cms configuration option.
5525 *) Update the GMP engine glue to do direct copies between BIGNUM and
5526 mpz_t when openssl and GMP use the same limb size. Otherwise the
5527 existing "conversion via a text string export" trick is still used.
5528 [Paul Sheer <paulsheer@gmail.com>]
5530 *) Zlib compression BIO. This is a filter BIO which compressed and
5531 uncompresses any data passed through it.
5534 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5535 RFC3394 compatible AES key wrapping.
5538 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5539 sets string data without copying. X509_ALGOR_set0() and
5540 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5541 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5542 from an X509_ATTRIBUTE structure optionally checking it occurs only
5543 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5547 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5548 to get the expected BN_FLG_CONSTTIME behavior.
5549 [Bodo Moeller (Google)]
5553 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5554 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5555 - added some more tests to do_tests.pl
5556 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5557 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5558 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5559 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5560 - various changes to netware.pl to enable gcc-cross builds on Win32
5562 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5563 - various changes to fix missing prototype warnings
5564 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5565 - added AES, WHIRLPOOL and CPUID assembler code to build files
5566 - added missing AES assembler make rules to mk1mf.pl
5567 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5568 [Guenter Knauf <eflash@gmx.net>]
5570 *) Implement certificate status request TLS extension defined in RFC3546.
5571 A client can set the appropriate parameters and receive the encoded
5572 OCSP response via a callback. A server can query the supplied parameters
5573 and set the encoded OCSP response in the callback. Add simplified examples
5574 to s_client and s_server.
5577 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5579 *) Fix various bugs:
5580 + Binary incompatibility of ssl_ctx_st structure
5581 + DTLS interoperation with non-compliant servers
5582 + Don't call get_session_cb() without proposed session
5583 + Fix ia64 assembler code
5584 [Andy Polyakov, Steve Henson]
5586 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5588 *) DTLS Handshake overhaul. There were longstanding issues with
5589 OpenSSL DTLS implementation, which were making it impossible for
5590 RFC 4347 compliant client to communicate with OpenSSL server.
5591 Unfortunately just fixing these incompatibilities would "cut off"
5592 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5593 server keeps tolerating non RFC compliant syntax. The opposite is
5594 not true, 0.9.8f client can not communicate with earlier server.
5595 This update even addresses CVE-2007-4995.
5598 *) Changes to avoid need for function casts in OpenSSL: some compilers
5599 (gcc 4.2 and later) reject their use.
5600 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5603 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5604 RFC4507bis. The encrypted ticket format is an encrypted encoded
5605 SSL_SESSION structure, that way new session features are automatically
5608 If a client application caches session in an SSL_SESSION structure
5609 support is transparent because tickets are now stored in the encoded
5612 The SSL_CTX structure automatically generates keys for ticket
5613 protection in servers so again support should be possible
5614 with no application modification.
5616 If a client or server wishes to disable RFC4507 support then the option
5617 SSL_OP_NO_TICKET can be set.
5619 Add a TLS extension debugging callback to allow the contents of any client
5620 or server extensions to be examined.
5622 This work was sponsored by Google.
5625 *) Add initial support for TLS extensions, specifically for the server_name
5626 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5627 have new members for a host name. The SSL data structure has an
5628 additional member SSL_CTX *initial_ctx so that new sessions can be
5629 stored in that context to allow for session resumption, even after the
5630 SSL has been switched to a new SSL_CTX in reaction to a client's
5631 server_name extension.
5633 New functions (subject to change):
5635 SSL_get_servername()
5636 SSL_get_servername_type()
5639 New CTRL codes and macros (subject to change):
5641 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5642 - SSL_CTX_set_tlsext_servername_callback()
5643 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5644 - SSL_CTX_set_tlsext_servername_arg()
5645 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5647 openssl s_client has a new '-servername ...' option.
5649 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5650 '-key2 ...', '-servername_fatal' (subject to change). This allows
5651 testing the HostName extension for a specific single host name ('-cert'
5652 and '-key' remain fallbacks for handshakes without HostName
5653 negotiation). If the unrecognized_name alert has to be sent, this by
5654 default is a warning; it becomes fatal with the '-servername_fatal'
5657 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5659 *) Add AES and SSE2 assembly language support to VC++ build.
5662 *) Mitigate attack on final subtraction in Montgomery reduction.
5665 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5666 (which previously caused an internal error).
5669 *) Squeeze another 10% out of IGE mode when in != out.
5672 *) AES IGE mode speedup.
5673 [Dean Gaudet (Google)]
5675 *) Add the Korean symmetric 128-bit cipher SEED (see
5676 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5677 add SEED ciphersuites from RFC 4162:
5679 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5680 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5681 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5682 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5684 To minimize changes between patchlevels in the OpenSSL 0.9.8
5685 series, SEED remains excluded from compilation unless OpenSSL
5686 is configured with 'enable-seed'.
5687 [KISA, Bodo Moeller]
5689 *) Mitigate branch prediction attacks, which can be practical if a
5690 single processor is shared, allowing a spy process to extract
5691 information. For detailed background information, see
5692 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5693 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5694 and Necessary Software Countermeasures"). The core of the change
5695 are new versions BN_div_no_branch() and
5696 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5697 respectively, which are slower, but avoid the security-relevant
5698 conditional branches. These are automatically called by BN_div()
5699 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5700 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5701 remove a conditional branch.
5703 BN_FLG_CONSTTIME is the new name for the previous
5704 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5705 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5706 in the exponent causes BN_mod_exp_mont() to use the alternative
5707 implementation in BN_mod_exp_mont_consttime().) The old name
5708 remains as a deprecated alias.
5710 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5711 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5712 constant-time implementations for more than just exponentiation.
5713 Here too the old name is kept as a deprecated alias.
5715 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5716 the BN_BLINDING structure gets an independent copy of the
5717 modulus. This means that the previous "BIGNUM *m" argument to
5718 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5719 essentially becomes "const BIGNUM *m", although we can't actually
5720 change this in the header file before 0.9.9. It allows
5721 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5722 enable BN_FLG_CONSTTIME.
5724 [Matthew D Wood (Intel Corp)]
5726 *) In the SSL/TLS server implementation, be strict about session ID
5727 context matching (which matters if an application uses a single
5728 external cache for different purposes). Previously,
5729 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5730 set. This did ensure strict client verification, but meant that,
5731 with applications using a single external cache for quite
5732 different requirements, clients could circumvent ciphersuite
5733 restrictions for a given session ID context by starting a session
5734 in a different context.
5737 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5738 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5739 authentication-only ciphersuites.
5742 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5743 not complete and could lead to a possible single byte overflow
5744 (CVE-2007-5135) [Ben Laurie]
5746 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5748 *) Since AES128 and AES256 (and similarly Camellia128 and
5749 Camellia256) share a single mask bit in the logic of
5750 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5751 kludge to work properly if AES128 is available and AES256 isn't
5752 (or if Camellia128 is available and Camellia256 isn't).
5755 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5756 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5757 When a point or a seed is encoded in a BIT STRING, we need to
5758 prevent the removal of trailing zero bits to get the proper DER
5759 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5760 of a NamedBitList, for which trailing 0 bits need to be removed.)
5763 *) Have SSL/TLS server implementation tolerate "mismatched" record
5764 protocol version while receiving ClientHello even if the
5765 ClientHello is fragmented. (The server can't insist on the
5766 particular protocol version it has chosen before the ServerHello
5767 message has informed the client about his choice.)
5770 *) Add RFC 3779 support.
5771 [Rob Austein for ARIN, Ben Laurie]
5773 *) Load error codes if they are not already present instead of using a
5774 static variable. This allows them to be cleanly unloaded and reloaded.
5775 Improve header file function name parsing.
5778 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5779 or CAPABILITY handshake as required by RFCs.
5782 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5784 *) Introduce limits to prevent malicious keys being able to
5785 cause a denial of service. (CVE-2006-2940)
5786 [Steve Henson, Bodo Moeller]
5788 *) Fix ASN.1 parsing of certain invalid structures that can result
5789 in a denial of service. (CVE-2006-2937) [Steve Henson]
5791 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5792 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5794 *) Fix SSL client code which could crash if connecting to a
5795 malicious SSLv2 server. (CVE-2006-4343)
5796 [Tavis Ormandy and Will Drewry, Google Security Team]
5798 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5799 match only those. Before that, "AES256-SHA" would be interpreted
5800 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5801 the same strength classification in 0.9.7h) as we currently only
5802 have a single AES bit in the ciphersuite description bitmap.
5803 That change, however, also applied to ciphersuite strings such as
5804 "RC4-MD5" that intentionally matched multiple ciphersuites --
5805 namely, SSL 2.0 ciphersuites in addition to the more common ones
5806 from SSL 3.0/TLS 1.0.
5808 So we change the selection algorithm again: Naming an explicit
5809 ciphersuite selects this one ciphersuite, and any other similar
5810 ciphersuite (same bitmap) from *other* protocol versions.
5811 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5812 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5814 Since SSL 2.0 does not have any ciphersuites for which the
5815 128/256 bit distinction would be relevant, this works for now.
5816 The proper fix will be to use different bits for AES128 and
5817 AES256, which would have avoided the problems from the beginning;
5818 however, bits are scarce, so we can only do this in a new release
5819 (not just a patchlevel) when we can change the SSL_CIPHER
5820 definition to split the single 'unsigned long mask' bitmap into
5821 multiple values to extend the available space.
5825 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5827 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5828 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5830 *) Add AES IGE and biIGE modes.
5833 *) Change the Unix randomness entropy gathering to use poll() when
5834 possible instead of select(), since the latter has some
5835 undesirable limitations.
5836 [Darryl Miles via Richard Levitte and Bodo Moeller]
5838 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5839 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5840 cannot be implicitly activated as part of, e.g., the "AES" alias.
5841 However, please upgrade to OpenSSL 0.9.9[-dev] for
5842 non-experimental use of the ECC ciphersuites to get TLS extension
5843 support, which is required for curve and point format negotiation
5844 to avoid potential handshake problems.
5847 *) Disable rogue ciphersuites:
5849 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5850 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5851 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5853 The latter two were purportedly from
5854 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5857 Also deactivate the remaining ciphersuites from
5858 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5859 unofficial, and the ID has long expired.
5862 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5863 dual-core machines) and other potential thread-safety issues.
5866 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5867 versions), which is now available for royalty-free use
5868 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5869 Also, add Camellia TLS ciphersuites from RFC 4132.
5871 To minimize changes between patchlevels in the OpenSSL 0.9.8
5872 series, Camellia remains excluded from compilation unless OpenSSL
5873 is configured with 'enable-camellia'.
5876 *) Disable the padding bug check when compression is in use. The padding
5877 bug check assumes the first packet is of even length, this is not
5878 necessarily true if compression is enabled and can result in false
5879 positives causing handshake failure. The actual bug test is ancient
5880 code so it is hoped that implementations will either have fixed it by
5881 now or any which still have the bug do not support compression.
5884 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5886 *) When applying a cipher rule check to see if string match is an explicit
5887 cipher suite and only match that one cipher suite if it is.
5890 *) Link in manifests for VC++ if needed.
5891 [Austin Ziegler <halostatue@gmail.com>]
5893 *) Update support for ECC-based TLS ciphersuites according to
5894 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5895 TLS extensions, which are supported starting with the 0.9.9
5896 branch, not in the OpenSSL 0.9.8 branch).
5899 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5900 opaque EVP_CIPHER_CTX handling.
5903 *) Fixes and enhancements to zlib compression code. We now only use
5904 "zlib1.dll" and use the default __cdecl calling convention on Win32
5905 to conform with the standards mentioned here:
5906 http://www.zlib.net/DLL_FAQ.txt
5907 Static zlib linking now works on Windows and the new --with-zlib-include
5908 --with-zlib-lib options to Configure can be used to supply the location
5909 of the headers and library. Gracefully handle case where zlib library
5913 *) Several fixes and enhancements to the OID generation code. The old code
5914 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5915 handle numbers larger than ULONG_MAX, truncated printing and had a
5916 non standard OBJ_obj2txt() behaviour.
5919 *) Add support for building of engines under engine/ as shared libraries
5920 under VC++ build system.
5923 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5924 Hopefully, we will not see any false combination of paths any more.
5927 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5929 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5930 (part of SSL_OP_ALL). This option used to disable the
5931 countermeasure against man-in-the-middle protocol-version
5932 rollback in the SSL 2.0 server implementation, which is a bad
5933 idea. (CVE-2005-2969)
5935 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5936 for Information Security, National Institute of Advanced Industrial
5937 Science and Technology [AIST], Japan)]
5939 *) Add two function to clear and return the verify parameter flags.
5942 *) Keep cipherlists sorted in the source instead of sorting them at
5943 runtime, thus removing the need for a lock.
5946 *) Avoid some small subgroup attacks in Diffie-Hellman.
5947 [Nick Mathewson and Ben Laurie]
5949 *) Add functions for well-known primes.
5952 *) Extended Windows CE support.
5953 [Satoshi Nakamura and Andy Polyakov]
5955 *) Initialize SSL_METHOD structures at compile time instead of during
5956 runtime, thus removing the need for a lock.
5959 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5960 attempting to decrypt each encrypted key in turn. Add support to
5964 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5966 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5969 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5972 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5973 key into the same file any more.
5976 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5979 *) Add -utf8 command line and config file option to 'ca'.
5980 [Stefan <stf@udoma.org]
5982 *) Removed the macro des_crypt(), as it seems to conflict with some
5983 libraries. Use DES_crypt().
5986 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5987 involves renaming the source and generated shared-libs for
5988 both. The engines will accept the corrected or legacy ids
5989 ('ncipher' and '4758_cca' respectively) when binding. NB,
5990 this only applies when building 'shared'.
5991 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5993 *) Add attribute functions to EVP_PKEY structure. Modify
5994 PKCS12_create() to recognize a CSP name attribute and
5995 use it. Make -CSP option work again in pkcs12 utility.
5998 *) Add new functionality to the bn blinding code:
5999 - automatic re-creation of the BN_BLINDING parameters after
6000 a fixed number of uses (currently 32)
6001 - add new function for parameter creation
6002 - introduce flags to control the update behaviour of the
6003 BN_BLINDING parameters
6004 - hide BN_BLINDING structure
6005 Add a second BN_BLINDING slot to the RSA structure to improve
6006 performance when a single RSA object is shared among several
6010 *) Add support for DTLS.
6011 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
6013 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
6014 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
6017 *) Remove buggy and incomplete DH cert support from
6018 ssl/ssl_rsa.c and ssl/s3_both.c
6021 *) Use SHA-1 instead of MD5 as the default digest algorithm for
6022 the apps/openssl applications.
6025 *) Compile clean with "-Wall -Wmissing-prototypes
6026 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
6027 DEBUG_SAFESTACK must also be set.
6030 *) Change ./Configure so that certain algorithms can be disabled by default.
6031 The new counterpiece to "no-xxx" is "enable-xxx".
6033 The patented RC5 and MDC2 algorithms will now be disabled unless
6034 "enable-rc5" and "enable-mdc2", respectively, are specified.
6036 (IDEA remains enabled despite being patented. This is because IDEA
6037 is frequently required for interoperability, and there is no license
6038 fee for non-commercial use. As before, "no-idea" can be used to
6039 avoid this algorithm.)
6043 *) Add processing of proxy certificates (see RFC 3820). This work was
6044 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
6045 EGEE (Enabling Grids for E-science in Europe).
6048 *) RC4 performance overhaul on modern architectures/implementations, such
6049 as Intel P4, IA-64 and AMD64.
6052 *) New utility extract-section.pl. This can be used specify an alternative
6053 section number in a pod file instead of having to treat each file as
6054 a separate case in Makefile. This can be done by adding two lines to the
6057 =for comment openssl_section:XXX
6059 The blank line is mandatory.
6063 *) New arguments -certform, -keyform and -pass for s_client and s_server
6064 to allow alternative format key and certificate files and passphrase
6068 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
6069 update associated structures and add various utility functions.
6071 Add new policy related verify parameters, include policy checking in
6072 standard verify code. Enhance 'smime' application with extra parameters
6073 to support policy checking and print out.
6076 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
6077 Nehemiah processors. These extensions support AES encryption in hardware
6078 as well as RNG (though RNG support is currently disabled).
6079 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
6081 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
6084 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
6085 [Andy Polyakov and a number of other people]
6087 *) Improved PowerPC platform support. Most notably BIGNUM assembler
6088 implementation contributed by IBM.
6089 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
6091 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
6092 exponent rather than 'unsigned long'. There is a corresponding change to
6093 the new 'rsa_keygen' element of the RSA_METHOD structure.
6094 [Jelte Jansen, Geoff Thorpe]
6096 *) Functionality for creating the initial serial number file is now
6097 moved from CA.pl to the 'ca' utility with a new option -create_serial.
6099 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
6100 number file to 1, which is bound to cause problems. To avoid
6101 the problems while respecting compatibility between different 0.9.7
6102 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
6103 CA.pl for serial number initialization. With the new release 0.9.8,
6104 we can fix the problem directly in the 'ca' utility.)
6107 *) Reduced header interdependencies by declaring more opaque objects in
6108 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
6109 give fewer recursive includes, which could break lazy source code - so
6110 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
6111 developers should define this symbol when building and using openssl to
6112 ensure they track the recommended behaviour, interfaces, [etc], but
6113 backwards-compatible behaviour prevails when this isn't defined.
6116 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
6119 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
6120 This will generate a random key of the appropriate length based on the
6121 cipher context. The EVP_CIPHER can provide its own random key generation
6122 routine to support keys of a specific form. This is used in the des and
6123 3des routines to generate a key of the correct parity. Update S/MIME
6124 code to use new functions and hence generate correct parity DES keys.
6125 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
6126 valid (weak or incorrect parity).
6129 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
6130 as looking them up. This is useful when the verified structure may contain
6131 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
6132 present unless the new PKCS7_NO_CRL flag is asserted.
6135 *) Extend ASN1 oid configuration module. It now additionally accepts the
6138 shortName = some long name, 1.2.3.4
6141 *) Reimplemented the BN_CTX implementation. There is now no more static
6142 limitation on the number of variables it can handle nor the depth of the
6143 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
6144 information can now expand as required, and rather than having a single
6145 static array of bignums, BN_CTX now uses a linked-list of such arrays
6146 allowing it to expand on demand whilst maintaining the usefulness of
6147 BN_CTX's "bundling".
6150 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
6151 to allow all RSA operations to function using a single BN_CTX.
6154 *) Preliminary support for certificate policy evaluation and checking. This
6155 is initially intended to pass the tests outlined in "Conformance Testing
6156 of Relying Party Client Certificate Path Processing Logic" v1.07.
6159 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
6160 remained unused and not that useful. A variety of other little bignum
6161 tweaks and fixes have also been made continuing on from the audit (see
6165 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
6166 associated ASN1, EVP and SSL functions and old ASN1 macros.
6169 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
6170 and this should never fail. So the return value from the use of
6171 BN_set_word() (which can fail due to needless expansion) is now deprecated;
6172 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
6175 *) BN_CTX_get() should return zero-valued bignums, providing the same
6176 initialised value as BN_new().
6177 [Geoff Thorpe, suggested by Ulf Möller]
6179 *) Support for inhibitAnyPolicy certificate extension.
6182 *) An audit of the BIGNUM code is underway, for which debugging code is
6183 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
6184 is considered valid when processing BIGNUMs, and causes execution to
6185 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
6186 further steps are taken to deliberately pollute unused data in BIGNUM
6187 structures to try and expose faulty code further on. For now, openssl will
6188 (in its default mode of operation) continue to tolerate the inconsistent
6189 forms that it has tolerated in the past, but authors and packagers should
6190 consider trying openssl and their own applications when compiled with
6191 these debugging symbols defined. It will help highlight potential bugs in
6192 their own code, and will improve the test coverage for OpenSSL itself. At
6193 some point, these tighter rules will become openssl's default to improve
6194 maintainability, though the assert()s and other overheads will remain only
6195 in debugging configurations. See bn.h for more details.
6196 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6198 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6199 that can only be obtained through BN_CTX_new() (which implicitly
6200 initialises it). The presence of this function only made it possible
6201 to overwrite an existing structure (and cause memory leaks).
6204 *) Because of the callback-based approach for implementing LHASH as a
6205 template type, lh_insert() adds opaque objects to hash-tables and
6206 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6207 to clean up those corresponding objects before destroying the hash table
6208 (and losing the object pointers). So some over-zealous constifications in
6209 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6210 objects as "const" and the lh_doall[_arg] callback wrappers are not
6211 prototyped to have "const" restrictions on the object pointers they are
6212 given (and so aren't required to cast them away any more).
6215 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6216 (speed) prefers to use its own implementation. The two implementations
6217 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6218 its object type properly exposed (MS_TM) instead of casting to/from "char
6219 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6220 aren't necessarily the greatest nomenclatures - but this is what was used
6221 internally to the implementation so I've used that for now.
6224 *) Ensure that deprecated functions do not get compiled when
6225 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6226 the self-tests were still using deprecated key-generation functions so
6227 these have been updated also.
6230 *) Reorganise PKCS#7 code to separate the digest location functionality
6231 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6232 New function PKCS7_set_digest() to set the digest type for PKCS#7
6233 digestedData type. Add additional code to correctly generate the
6234 digestedData type and add support for this type in PKCS7 initialization
6238 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6239 structure of type "other".
6242 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6243 sure the loop does correctly stop and breaking ("division by zero")
6244 modulus operations are not performed. The (pre-generated) prime
6245 table crypto/bn/bn_prime.h was already correct, but it could not be
6246 re-generated on some platforms because of the "division by zero"
6247 situation in the script.
6248 [Ralf S. Engelschall]
6250 *) Update support for ECC-based TLS ciphersuites according to
6251 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6252 SHA-1 now is only used for "small" curves (where the
6253 representation of a field element takes up to 24 bytes); for
6254 larger curves, the field element resulting from ECDH is directly
6255 used as premaster secret.
6256 [Douglas Stebila (Sun Microsystems Laboratories)]
6258 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6259 curve secp160r1 to the tests.
6260 [Douglas Stebila (Sun Microsystems Laboratories)]
6262 *) Add the possibility to load symbols globally with DSO.
6263 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6265 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6266 control of the error stack.
6269 *) Add support for STORE in ENGINE.
6272 *) Add the STORE type. The intention is to provide a common interface
6273 to certificate and key stores, be they simple file-based stores, or
6274 HSM-type store, or LDAP stores, or...
6275 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6278 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6279 pass a list of arguments to any function as well as provide a way
6280 for a function to pass data back to the caller.
6283 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6284 works like BUF_strdup() but can be used to duplicate a portion of
6285 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6289 *) Add the function sk_find_ex() which works like sk_find(), but will
6290 return an index to an element even if an exact match couldn't be
6291 found. The index is guaranteed to point at the element where the
6292 searched-for key would be inserted to preserve sorting order.
6295 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6296 takes an extra flags argument for optional functionality. Currently,
6297 the following flags are defined:
6299 OBJ_BSEARCH_VALUE_ON_NOMATCH
6300 This one gets OBJ_bsearch_ex() to return a pointer to the first
6301 element where the comparing function returns a negative or zero
6304 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6305 This one gets OBJ_bsearch_ex() to return a pointer to the first
6306 element where the comparing function returns zero. This is useful
6307 if there are more than one element where the comparing function
6311 *) Make it possible to create self-signed certificates with 'openssl ca'
6312 in such a way that the self-signed certificate becomes part of the
6313 CA database and uses the same mechanisms for serial number generation
6314 as all other certificate signing. The new flag '-selfsign' enables
6315 this functionality. Adapt CA.sh and CA.pl.in.
6318 *) Add functionality to check the public key of a certificate request
6319 against a given private. This is useful to check that a certificate
6320 request can be signed by that key (self-signing).
6323 *) Make it possible to have multiple active certificates with the same
6324 subject in the CA index file. This is done only if the keyword
6325 'unique_subject' is set to 'no' in the main CA section (default
6326 if 'CA_default') of the configuration file. The value is saved
6327 with the database itself in a separate index attribute file,
6328 named like the index file with '.attr' appended to the name.
6331 *) Generate multi-valued AVAs using '+' notation in config files for
6335 *) Support for nameConstraints certificate extension.
6338 *) Support for policyConstraints certificate extension.
6341 *) Support for policyMappings certificate extension.
6344 *) Make sure the default DSA_METHOD implementation only uses its
6345 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6346 and change its own handlers to be NULL so as to remove unnecessary
6347 indirection. This lets alternative implementations fallback to the
6348 default implementation more easily.
6351 *) Support for directoryName in GeneralName related extensions
6355 *) Make it possible to link applications using Makefile.shared.
6356 Make that possible even when linking against static libraries!
6359 *) Support for single pass processing for S/MIME signing. This now
6360 means that S/MIME signing can be done from a pipe, in addition
6361 cleartext signing (multipart/signed type) is effectively streaming
6362 and the signed data does not need to be all held in memory.
6364 This is done with a new flag PKCS7_STREAM. When this flag is set
6365 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6366 is done after the data is output (and digests calculated) in
6367 SMIME_write_PKCS7().
6370 *) Add full support for -rpath/-R, both in shared libraries and
6371 applications, at least on the platforms where it's known how
6375 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6376 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6377 will now compute a table of multiples of the generator that
6378 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6379 faster (notably in the case of a single point multiplication,
6380 scalar * generator).
6381 [Nils Larsch, Bodo Moeller]
6383 *) IPv6 support for certificate extensions. The various extensions
6384 which use the IP:a.b.c.d can now take IPv6 addresses using the
6385 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6389 *) Added an ENGINE that implements RSA by performing private key
6390 exponentiations with the GMP library. The conversions to and from
6391 GMP's mpz_t format aren't optimised nor are any montgomery forms
6392 cached, and on x86 it appears OpenSSL's own performance has caught up.
6393 However there are likely to be other architectures where GMP could
6394 provide a boost. This ENGINE is not built in by default, but it can be
6395 specified at Configure time and should be accompanied by the necessary
6396 linker additions, eg;
6397 ./config -DOPENSSL_USE_GMP -lgmp
6400 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6401 testing availability of engines with "-t" - the old behaviour is
6402 produced by increasing the feature's verbosity with "-tt".
6405 *) ECDSA routines: under certain error conditions uninitialized BN objects
6406 could be freed. Solution: make sure initialization is performed early
6407 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6411 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6412 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6413 software implementations. For DSA and DH, parameter generation can
6414 also be overridden by providing the appropriate method callbacks.
6417 *) Change the "progress" mechanism used in key-generation and
6418 primality testing to functions that take a new BN_GENCB pointer in
6419 place of callback/argument pairs. The new API functions have "_ex"
6420 postfixes and the older functions are reimplemented as wrappers for
6421 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6422 declarations of the old functions to help (graceful) attempts to
6423 migrate to the new functions. Also, the new key-generation API
6424 functions operate on a caller-supplied key-structure and return
6425 success/failure rather than returning a key or NULL - this is to
6426 help make "keygen" another member function of RSA_METHOD etc.
6428 Example for using the new callback interface:
6430 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6434 BN_GENCB_set(&my_cb, my_callback, my_arg);
6436 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6437 /* For the meaning of a, b in calls to my_callback(), see the
6438 * documentation of the function that calls the callback.
6439 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6440 * my_callback should return 1 if it wants BN_is_prime_ex()
6441 * to continue, or 0 to stop.
6446 *) Change the ZLIB compression method to be stateful, and make it
6447 available to TLS with the number defined in
6448 draft-ietf-tls-compression-04.txt.
6451 *) Add the ASN.1 structures and functions for CertificatePair, which
6452 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6454 CertificatePair ::= SEQUENCE {
6455 forward [0] Certificate OPTIONAL,
6456 reverse [1] Certificate OPTIONAL,
6457 -- at least one of the pair shall be present -- }
6459 Also implement the PEM functions to read and write certificate
6460 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6462 This needed to be defined, mostly for the sake of the LDAP
6463 attribute crossCertificatePair, but may prove useful elsewhere as
6467 *) Make it possible to inhibit symlinking of shared libraries in
6468 Makefile.shared, for Cygwin's sake.
6471 *) Extend the BIGNUM API by creating a function
6472 void BN_set_negative(BIGNUM *a, int neg);
6473 and a macro that behave like
6474 int BN_is_negative(const BIGNUM *a);
6476 to avoid the need to access 'a->neg' directly in applications.
6479 *) Implement fast modular reduction for pseudo-Mersenne primes
6480 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6481 EC_GROUP_new_curve_GFp() will now automatically use this
6483 [Nils Larsch <nla@trustcenter.de>]
6485 *) Add new lock type (CRYPTO_LOCK_BN).
6488 *) Change the ENGINE framework to automatically load engines
6489 dynamically from specific directories unless they could be
6490 found to already be built in or loaded. Move all the
6491 current engines except for the cryptodev one to a new
6493 The engines in engines/ are built as shared libraries if
6494 the "shared" options was given to ./Configure or ./config.
6495 Otherwise, they are inserted in libcrypto.a.
6496 /usr/local/ssl/engines is the default directory for dynamic
6497 engines, but that can be overridden at configure time through
6498 the usual use of --prefix and/or --openssldir, and at run
6499 time with the environment variable OPENSSL_ENGINES.
6500 [Geoff Thorpe and Richard Levitte]
6502 *) Add Makefile.shared, a helper makefile to build shared
6503 libraries. Adapt Makefile.org.
6506 *) Add version info to Win32 DLLs.
6507 [Peter 'Luna' Runestig" <peter@runestig.com>]
6509 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6510 can be added using this API to created arbitrary PKCS#12
6511 files while avoiding the low level API.
6513 New options to PKCS12_create(), key or cert can be NULL and
6514 will then be omitted from the output file. The encryption
6515 algorithm NIDs can be set to -1 for no encryption, the mac
6516 iteration count can be set to 0 to omit the mac.
6518 Enhance pkcs12 utility by making the -nokeys and -nocerts
6519 options work when creating a PKCS#12 file. New option -nomac
6520 to omit the mac, NONE can be set for an encryption algorithm.
6521 New code is modified to use the enhanced PKCS12_create()
6522 instead of the low level API.
6525 *) Extend ASN1 encoder to support indefinite length constructed
6526 encoding. This can output sequences tags and octet strings in
6527 this form. Modify pk7_asn1.c to support indefinite length
6528 encoding. This is experimental and needs additional code to
6529 be useful, such as an ASN1 bio and some enhanced streaming
6532 Extend template encode functionality so that tagging is passed
6533 down to the template encoder.
6536 *) Let 'openssl req' fail if an argument to '-newkey' is not
6537 recognized instead of using RSA as a default.
6540 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6541 As these are not official, they are not included in "ALL";
6542 the "ECCdraft" ciphersuite group alias can be used to select them.
6543 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6545 *) Add ECDH engine support.
6546 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6548 *) Add ECDH in new directory crypto/ecdh/.
6549 [Douglas Stebila (Sun Microsystems Laboratories)]
6551 *) Let BN_rand_range() abort with an error after 100 iterations
6552 without success (which indicates a broken PRNG).
6555 *) Change BN_mod_sqrt() so that it verifies that the input value
6556 is really the square of the return value. (Previously,
6557 BN_mod_sqrt would show GIGO behaviour.)
6560 *) Add named elliptic curves over binary fields from X9.62, SECG,
6561 and WAP/WTLS; add OIDs that were still missing.
6563 [Sheueling Chang Shantz and Douglas Stebila
6564 (Sun Microsystems Laboratories)]
6566 *) Extend the EC library for elliptic curves over binary fields
6567 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6570 EC_GF2m_simple_method
6574 EC_GROUP_new_curve_GF2m
6575 EC_GROUP_set_curve_GF2m
6576 EC_GROUP_get_curve_GF2m
6577 EC_POINT_set_affine_coordinates_GF2m
6578 EC_POINT_get_affine_coordinates_GF2m
6579 EC_POINT_set_compressed_coordinates_GF2m
6581 Point compression for binary fields is disabled by default for
6582 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6585 As binary polynomials are represented as BIGNUMs, various members
6586 of the EC_GROUP and EC_POINT data structures can be shared
6587 between the implementations for prime fields and binary fields;
6588 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6589 are essentially identical to their ..._GFp counterparts.
6590 (For simplicity, the '..._GFp' prefix has been dropped from
6591 various internal method names.)
6593 An internal 'field_div' method (similar to 'field_mul' and
6594 'field_sqr') has been added; this is used only for binary fields.
6596 [Sheueling Chang Shantz and Douglas Stebila
6597 (Sun Microsystems Laboratories)]
6599 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6600 through methods ('mul', 'precompute_mult').
6602 The generic implementations (now internally called 'ec_wNAF_mul'
6603 and 'ec_wNAF_precomputed_mult') remain the default if these
6604 methods are undefined.
6606 [Sheueling Chang Shantz and Douglas Stebila
6607 (Sun Microsystems Laboratories)]
6609 *) New function EC_GROUP_get_degree, which is defined through
6610 EC_METHOD. For curves over prime fields, this returns the bit
6611 length of the modulus.
6613 [Sheueling Chang Shantz and Douglas Stebila
6614 (Sun Microsystems Laboratories)]
6616 *) New functions EC_GROUP_dup, EC_POINT_dup.
6617 (These simply call ..._new and ..._copy).
6619 [Sheueling Chang Shantz and Douglas Stebila
6620 (Sun Microsystems Laboratories)]
6622 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6623 Polynomials are represented as BIGNUMs (where the sign bit is not
6624 used) in the following functions [macros]:
6627 BN_GF2m_sub [= BN_GF2m_add]
6628 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6629 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6630 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6632 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6633 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6634 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6635 BN_GF2m_cmp [= BN_ucmp]
6637 (Note that only the 'mod' functions are actually for fields GF(2^m).
6638 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6640 For some functions, an the irreducible polynomial defining a
6641 field can be given as an 'unsigned int[]' with strictly
6642 decreasing elements giving the indices of those bits that are set;
6643 i.e., p[] represents the polynomial
6644 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6646 p[0] > p[1] > ... > p[k] = 0.
6647 This applies to the following functions:
6652 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6653 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6655 BN_GF2m_mod_sqrt_arr
6656 BN_GF2m_mod_solve_quad_arr
6660 Conversion can be performed by the following functions:
6665 bntest.c has additional tests for binary polynomial arithmetic.
6667 Two implementations for BN_GF2m_mod_div() are available.
6668 The default algorithm simply uses BN_GF2m_mod_inv() and
6669 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6670 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6671 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6673 [Sheueling Chang Shantz and Douglas Stebila
6674 (Sun Microsystems Laboratories)]
6676 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6677 functionality is disabled at compile-time.
6678 [Douglas Stebila <douglas.stebila@sun.com>]
6680 *) Change default behaviour of 'openssl asn1parse' so that more
6681 information is visible when viewing, e.g., a certificate:
6683 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6684 mode the content of non-printable OCTET STRINGs is output in a
6685 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6686 avoid the appearance of a printable string.
6687 [Nils Larsch <nla@trustcenter.de>]
6689 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6691 EC_GROUP_set_asn1_flag()
6692 EC_GROUP_get_asn1_flag()
6693 EC_GROUP_set_point_conversion_form()
6694 EC_GROUP_get_point_conversion_form()
6695 These control ASN1 encoding details:
6696 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6697 has been set to OPENSSL_EC_NAMED_CURVE.
6698 - Points are encoded in uncompressed form by default; options for
6699 asn1_for are as for point2oct, namely
6700 POINT_CONVERSION_COMPRESSED
6701 POINT_CONVERSION_UNCOMPRESSED
6702 POINT_CONVERSION_HYBRID
6704 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6707 EC_GROUP_get0_seed()
6708 EC_GROUP_get_seed_len()
6709 This is used only for ASN1 purposes (so far).
6710 [Nils Larsch <nla@trustcenter.de>]
6712 *) Add 'field_type' member to EC_METHOD, which holds the NID
6713 of the appropriate field type OID. The new function
6714 EC_METHOD_get_field_type() returns this value.
6715 [Nils Larsch <nla@trustcenter.de>]
6720 EC_POINT_point2hex()
6721 EC_POINT_hex2point()
6722 providing useful interfaces to EC_POINT_point2oct() and
6723 EC_POINT_oct2point().
6724 [Nils Larsch <nla@trustcenter.de>]
6726 *) Change internals of the EC library so that the functions
6727 EC_GROUP_set_generator()
6728 EC_GROUP_get_generator()
6729 EC_GROUP_get_order()
6730 EC_GROUP_get_cofactor()
6731 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6732 to methods, which would lead to unnecessary code duplication when
6733 adding different types of curves.
6734 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6736 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6737 arithmetic, and such that modified wNAFs are generated
6738 (which avoid length expansion in many cases).
6741 *) Add a function EC_GROUP_check_discriminant() (defined via
6742 EC_METHOD) that verifies that the curve discriminant is non-zero.
6744 Add a function EC_GROUP_check() that makes some sanity tests
6745 on a EC_GROUP, its generator and order. This includes
6746 EC_GROUP_check_discriminant().
6747 [Nils Larsch <nla@trustcenter.de>]
6749 *) Add ECDSA in new directory crypto/ecdsa/.
6751 Add applications 'openssl ecparam' and 'openssl ecdsa'
6752 (these are based on 'openssl dsaparam' and 'openssl dsa').
6754 ECDSA support is also included in various other files across the
6755 library. Most notably,
6756 - 'openssl req' now has a '-newkey ecdsa:file' option;
6757 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6758 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6759 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6760 them suitable for ECDSA where domain parameters must be
6761 extracted before the specific public key;
6762 - ECDSA engine support has been added.
6763 [Nils Larsch <nla@trustcenter.de>]
6765 *) Include some named elliptic curves, and add OIDs from X9.62,
6766 SECG, and WAP/WTLS. Each curve can be obtained from the new
6768 EC_GROUP_new_by_curve_name(),
6769 and the list of available named curves can be obtained with
6770 EC_get_builtin_curves().
6771 Also add a 'curve_name' member to EC_GROUP objects, which can be
6773 EC_GROUP_set_curve_name()
6774 EC_GROUP_get_curve_name()
6775 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6777 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6778 was actually never needed) and in BN_mul(). The removal in BN_mul()
6779 required a small change in bn_mul_part_recursive() and the addition
6780 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6781 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6782 bn_sub_words() and bn_add_words() except they take arrays with
6786 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6788 *) Cleanse PEM buffers before freeing them since they may contain
6790 [Benjamin Bennett <ben@psc.edu>]
6792 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6793 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6794 authentication-only ciphersuites.
6797 *) Since AES128 and AES256 share a single mask bit in the logic of
6798 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6799 kludge to work properly if AES128 is available and AES256 isn't.
6802 *) Expand security boundary to match 1.1.1 module.
6805 *) Remove redundant features: hash file source, editing of test vectors
6806 modify fipsld to use external fips_premain.c signature.
6809 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6810 run algorithm test programs.
6813 *) Make algorithm test programs more tolerant of whitespace.
6816 *) Have SSL/TLS server implementation tolerate "mismatched" record
6817 protocol version while receiving ClientHello even if the
6818 ClientHello is fragmented. (The server can't insist on the
6819 particular protocol version it has chosen before the ServerHello
6820 message has informed the client about his choice.)
6823 *) Load error codes if they are not already present instead of using a
6824 static variable. This allows them to be cleanly unloaded and reloaded.
6827 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6829 *) Introduce limits to prevent malicious keys being able to
6830 cause a denial of service. (CVE-2006-2940)
6831 [Steve Henson, Bodo Moeller]
6833 *) Fix ASN.1 parsing of certain invalid structures that can result
6834 in a denial of service. (CVE-2006-2937) [Steve Henson]
6836 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6837 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6839 *) Fix SSL client code which could crash if connecting to a
6840 malicious SSLv2 server. (CVE-2006-4343)
6841 [Tavis Ormandy and Will Drewry, Google Security Team]
6843 *) Change ciphersuite string processing so that an explicit
6844 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6845 will no longer include "AES128-SHA"), and any other similar
6846 ciphersuite (same bitmap) from *other* protocol versions (so that
6847 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6848 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6849 changes from 0.9.8b and 0.9.8d.
6852 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6854 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6855 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6857 *) Change the Unix randomness entropy gathering to use poll() when
6858 possible instead of select(), since the latter has some
6859 undesirable limitations.
6860 [Darryl Miles via Richard Levitte and Bodo Moeller]
6862 *) Disable rogue ciphersuites:
6864 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6865 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6866 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6868 The latter two were purportedly from
6869 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6872 Also deactivate the remaining ciphersuites from
6873 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6874 unofficial, and the ID has long expired.
6877 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6878 dual-core machines) and other potential thread-safety issues.
6881 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6883 *) Adapt fipsld and the build system to link against the validated FIPS
6884 module in FIPS mode.
6887 *) Fixes for VC++ 2005 build under Windows.
6890 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6891 from a Windows bash shell such as MSYS. It is autodetected from the
6892 "config" script when run from a VC++ environment. Modify standard VC++
6893 build to use fipscanister.o from the GNU make build.
6896 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6898 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6899 The value now differs depending on if you build for FIPS or not.
6900 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6901 safely run with a non-FIPSed libcrypto, as it may crash because of
6902 the difference induced by this change.
6905 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6907 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6908 (part of SSL_OP_ALL). This option used to disable the
6909 countermeasure against man-in-the-middle protocol-version
6910 rollback in the SSL 2.0 server implementation, which is a bad
6911 idea. (CVE-2005-2969)
6913 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6914 for Information Security, National Institute of Advanced Industrial
6915 Science and Technology [AIST], Japan)]
6917 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6918 mainly for FIPS compliance and not fully integrated at this stage.
6921 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6922 the exponentiation using a fixed-length exponent. (Otherwise,
6923 the information leaked through timing could expose the secret key
6924 after many signatures; cf. Bleichenbacher's attack on DSA with
6928 *) Make a new fixed-window mod_exp implementation the default for
6929 RSA, DSA, and DH private-key operations so that the sequence of
6930 squares and multiplies and the memory access pattern are
6931 independent of the particular secret key. This will mitigate
6932 cache-timing and potential related attacks.
6934 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6935 and this is automatically used by BN_mod_exp_mont() if the new flag
6936 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6937 will use this BN flag for private exponents unless the flag
6938 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6939 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6941 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6943 *) Change the client implementation for SSLv23_method() and
6944 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6945 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6946 (Previously, the SSL 2.0 backwards compatible Client Hello
6947 message format would be used even with SSL_OP_NO_SSLv2.)
6950 *) Add support for smime-type MIME parameter in S/MIME messages which some
6954 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6955 a threadsafe manner. Modify rsa code to use new function and add calls
6956 to dsa and dh code (which had race conditions before).
6959 *) Include the fixed error library code in the C error file definitions
6960 instead of fixing them up at runtime. This keeps the error code
6961 structures constant.
6964 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6966 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6969 *) Fixes for newer kerberos headers. NB: the casts are needed because
6970 the 'length' field is signed on one version and unsigned on another
6971 with no (?) obvious way to tell the difference, without these VC++
6972 complains. Also the "definition" of FAR (blank) is no longer included
6973 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6974 some needed definitions.
6977 *) Undo Cygwin change.
6980 *) Added support for proxy certificates according to RFC 3820.
6981 Because they may be a security thread to unaware applications,
6982 they must be explicitly allowed in run-time. See
6983 docs/HOWTO/proxy_certificates.txt for further information.
6986 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6988 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6989 server and client random values. Previously
6990 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6991 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6993 This change has negligible security impact because:
6995 1. Server and client random values still have 24 bytes of pseudo random
6998 2. Server and client random values are sent in the clear in the initial
7001 3. The master secret is derived using the premaster secret (48 bytes in
7002 size for static RSA ciphersuites) as well as client server and random
7005 The OpenSSL team would like to thank the UK NISCC for bringing this issue
7008 [Stephen Henson, reported by UK NISCC]
7010 *) Use Windows randomness collection on Cygwin.
7013 *) Fix hang in EGD/PRNGD query when communication socket is closed
7014 prematurely by EGD/PRNGD.
7015 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
7017 *) Prompt for pass phrases when appropriate for PKCS12 input format.
7020 *) Back-port of selected performance improvements from development
7021 branch, as well as improved support for PowerPC platforms.
7024 *) Add lots of checks for memory allocation failure, error codes to indicate
7025 failure and freeing up memory if a failure occurs.
7026 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
7028 *) Add new -passin argument to dgst.
7031 *) Perform some character comparisons of different types in X509_NAME_cmp:
7032 this is needed for some certificates that re-encode DNs into UTF8Strings
7033 (in violation of RFC3280) and can't or won't issue name rollover
7037 *) Make an explicit check during certificate validation to see that
7038 the CA setting in each certificate on the chain is correct. As a
7039 side effect always do the following basic checks on extensions,
7040 not just when there's an associated purpose to the check:
7042 - if there is an unhandled critical extension (unless the user
7043 has chosen to ignore this fault)
7044 - if the path length has been exceeded (if one is set at all)
7045 - that certain extensions fit the associated purpose (if one has
7049 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
7051 *) Avoid a race condition when CRLs are checked in a multi threaded
7052 environment. This would happen due to the reordering of the revoked
7053 entries during signature checking and serial number lookup. Now the
7054 encoding is cached and the serial number sort performed under a lock.
7055 Add new STACK function sk_is_sorted().
7058 *) Add Delta CRL to the extension code.
7061 *) Various fixes to s3_pkt.c so alerts are sent properly.
7062 [David Holmes <d.holmes@f5.com>]
7064 *) Reduce the chances of duplicate issuer name and serial numbers (in
7065 violation of RFC3280) using the OpenSSL certificate creation utilities.
7066 This is done by creating a random 64 bit value for the initial serial
7067 number when a serial number file is created or when a self signed
7068 certificate is created using 'openssl req -x509'. The initial serial
7069 number file is created using 'openssl x509 -next_serial' in CA.pl
7070 rather than being initialized to 1.
7073 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
7075 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7076 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7077 [Joe Orton, Steve Henson]
7079 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
7081 [Joe Orton, Steve Henson]
7083 *) Make it possible to have multiple active certificates with the same
7084 subject in the CA index file. This is done only if the keyword
7085 'unique_subject' is set to 'no' in the main CA section (default
7086 if 'CA_default') of the configuration file. The value is saved
7087 with the database itself in a separate index attribute file,
7088 named like the index file with '.attr' appended to the name.
7091 *) X509 verify fixes. Disable broken certificate workarounds when
7092 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
7093 keyUsage extension present. Don't accept CRLs with unhandled critical
7094 extensions: since verify currently doesn't process CRL extensions this
7095 rejects a CRL with *any* critical extensions. Add new verify error codes
7099 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
7100 A clarification of RFC2560 will require the use of OCTET STRINGs and
7101 some implementations cannot handle the current raw format. Since OpenSSL
7102 copies and compares OCSP nonces as opaque blobs without any attempt at
7103 parsing them this should not create any compatibility issues.
7106 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
7107 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
7108 this HMAC (and other) operations are several times slower than OpenSSL
7112 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
7113 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7115 *) Use the correct content when signing type "other".
7118 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
7120 *) Fix various bugs revealed by running the NISCC test suite:
7122 Stop out of bounds reads in the ASN1 code when presented with
7123 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7125 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
7127 If verify callback ignores invalid public key errors don't try to check
7128 certificate signature with the NULL public key.
7132 *) New -ignore_err option in ocsp application to stop the server
7133 exiting on the first error in a request.
7136 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7137 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7141 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7142 extra data after the compression methods not only for TLS 1.0
7143 but also for SSL 3.0 (as required by the specification).
7144 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7146 *) Change X509_certificate_type() to mark the key as exported/exportable
7147 when it's 512 *bits* long, not 512 bytes.
7150 *) Change AES_cbc_encrypt() so it outputs exact multiple of
7151 blocks during encryption.
7154 *) Various fixes to base64 BIO and non blocking I/O. On write
7155 flushes were not handled properly if the BIO retried. On read
7156 data was not being buffered properly and had various logic bugs.
7157 This also affects blocking I/O when the data being decoded is a
7161 *) Various S/MIME bugfixes and compatibility changes:
7162 output correct application/pkcs7 MIME type if
7163 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
7164 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
7165 of files as .eml work). Correctly handle very long lines in MIME
7169 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
7171 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7172 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7173 a protocol version number mismatch like a decryption error
7174 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7177 *) Turn on RSA blinding by default in the default implementation
7178 to avoid a timing attack. Applications that don't want it can call
7179 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7180 They would be ill-advised to do so in most cases.
7181 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7183 *) Change RSA blinding code so that it works when the PRNG is not
7184 seeded (in this case, the secret RSA exponent is abused as
7185 an unpredictable seed -- if it is not unpredictable, there
7186 is no point in blinding anyway). Make RSA blinding thread-safe
7187 by remembering the creator's thread ID in rsa->blinding and
7188 having all other threads use local one-time blinding factors
7189 (this requires more computation than sharing rsa->blinding, but
7190 avoids excessive locking; and if an RSA object is not shared
7191 between threads, blinding will still be very fast).
7194 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
7195 ENGINE as defaults for all supported algorithms irrespective of
7196 the 'flags' parameter. 'flags' is now honoured, so applications
7197 should make sure they are passing it correctly.
7200 *) Target "mingw" now allows native Windows code to be generated in
7201 the Cygwin environment as well as with the MinGW compiler.
7204 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7206 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7207 via timing by performing a MAC computation even if incorrect
7208 block cipher padding has been found. This is a countermeasure
7209 against active attacks where the attacker has to distinguish
7210 between bad padding and a MAC verification error. (CVE-2003-0078)
7212 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7213 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7214 Martin Vuagnoux (EPFL, Ilion)]
7216 *) Make the no-err option work as intended. The intention with no-err
7217 is not to have the whole error stack handling routines removed from
7218 libcrypto, it's only intended to remove all the function name and
7219 reason texts, thereby removing some of the footprint that may not
7220 be interesting if those errors aren't displayed anyway.
7222 NOTE: it's still possible for any application or module to have its
7223 own set of error texts inserted. The routines are there, just not
7224 used by default when no-err is given.
7227 *) Add support for FreeBSD on IA64.
7228 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7230 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7231 Kerberos function mit_des_cbc_cksum(). Before this change,
7232 the value returned by DES_cbc_cksum() was like the one from
7233 mit_des_cbc_cksum(), except the bytes were swapped.
7234 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7236 *) Allow an application to disable the automatic SSL chain building.
7237 Before this a rather primitive chain build was always performed in
7238 ssl3_output_cert_chain(): an application had no way to send the
7239 correct chain if the automatic operation produced an incorrect result.
7241 Now the chain builder is disabled if either:
7243 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7245 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7247 The reasoning behind this is that an application would not want the
7248 auto chain building to take place if extra chain certificates are
7249 present and it might also want a means of sending no additional
7250 certificates (for example the chain has two certificates and the
7254 *) Add the possibility to build without the ENGINE framework.
7255 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7257 *) Under Win32 gmtime() can return NULL: check return value in
7258 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7261 *) DSA routines: under certain error conditions uninitialized BN objects
7262 could be freed. Solution: make sure initialization is performed early
7263 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7264 Nils Larsch <nla@trustcenter.de> via PR#459)
7267 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7268 checked on reconnect on the client side, therefore session resumption
7269 could still fail with a "ssl session id is different" error. This
7270 behaviour is masked when SSL_OP_ALL is used due to
7271 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7272 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7273 followup to PR #377.
7276 *) IA-32 assembler support enhancements: unified ELF targets, support
7277 for SCO/Caldera platforms, fix for Cygwin shared build.
7280 *) Add support for FreeBSD on sparc64. As a consequence, support for
7281 FreeBSD on non-x86 processors is separate from x86 processors on
7282 the config script, much like the NetBSD support.
7283 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7285 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7287 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7290 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7291 code (06) was taken as the first octet of the session ID and the last
7292 octet was ignored consequently. As a result SSLv2 client side session
7293 caching could not have worked due to the session ID mismatch between
7295 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7299 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7300 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7304 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7305 seems that in spite of existing for more than a year, many application
7306 author have done nothing to provide the necessary callbacks, which
7307 means that this particular engine will not work properly anywhere.
7308 This is a very unfortunate situation which forces us, in the name
7309 of usability, to give the hw_ncipher.c a static lock, which is part
7311 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7312 appear in 0.9.8 or later. We EXPECT application authors to have
7313 dealt properly with this when 0.9.8 is released (unless we actually
7314 make such changes in the libcrypto locking code that changes will
7315 have to be made anyway).
7318 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7319 octets have been read, EOF or an error occurs. Without this change
7320 some truncated ASN1 structures will not produce an error.
7323 *) Disable Heimdal support, since it hasn't been fully implemented.
7324 Still give the possibility to force the use of Heimdal, but with
7325 warnings and a request that patches get sent to openssl-dev.
7328 *) Add the VC-CE target, introduce the WINCE sysname, and add
7329 INSTALL.WCE and appropriate conditionals to make it build.
7330 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7332 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7333 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7334 edit numbers of the version.
7335 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7337 *) Introduce safe string copy and catenation functions
7338 (BUF_strlcpy() and BUF_strlcat()).
7339 [Ben Laurie (CHATS) and Richard Levitte]
7341 *) Avoid using fixed-size buffers for one-line DNs.
7342 [Ben Laurie (CHATS)]
7344 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7345 resizing buffers containing secrets, and use where appropriate.
7346 [Ben Laurie (CHATS)]
7348 *) Avoid using fixed size buffers for configuration file location.
7349 [Ben Laurie (CHATS)]
7351 *) Avoid filename truncation for various CA files.
7352 [Ben Laurie (CHATS)]
7354 *) Use sizeof in preference to magic numbers.
7355 [Ben Laurie (CHATS)]
7357 *) Avoid filename truncation in cert requests.
7358 [Ben Laurie (CHATS)]
7360 *) Add assertions to check for (supposedly impossible) buffer
7362 [Ben Laurie (CHATS)]
7364 *) Don't cache truncated DNS entries in the local cache (this could
7365 potentially lead to a spoofing attack).
7366 [Ben Laurie (CHATS)]
7368 *) Fix various buffers to be large enough for hex/decimal
7369 representations in a platform independent manner.
7370 [Ben Laurie (CHATS)]
7372 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7373 resizing buffers containing secrets, and use where appropriate.
7374 [Ben Laurie (CHATS)]
7376 *) Add BIO_indent() to avoid much slightly worrying code to do
7378 [Ben Laurie (CHATS)]
7380 *) Convert sprintf()/BIO_puts() to BIO_printf().
7381 [Ben Laurie (CHATS)]
7383 *) buffer_gets() could terminate with the buffer only half
7385 [Ben Laurie (CHATS)]
7387 *) Add assertions to prevent user-supplied crypto functions from
7388 overflowing internal buffers by having large block sizes, etc.
7389 [Ben Laurie (CHATS)]
7391 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7393 [Ben Laurie (CHATS)]
7395 *) Eliminate unused copy of key in RC4.
7396 [Ben Laurie (CHATS)]
7398 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7399 [Ben Laurie (CHATS)]
7401 *) Fix off-by-one error in EGD path.
7402 [Ben Laurie (CHATS)]
7404 *) If RANDFILE path is too long, ignore instead of truncating.
7405 [Ben Laurie (CHATS)]
7407 *) Eliminate unused and incorrectly sized X.509 structure
7409 [Ben Laurie (CHATS)]
7411 *) Eliminate unused and dangerous function knumber().
7412 [Ben Laurie (CHATS)]
7414 *) Eliminate unused and dangerous structure, KSSL_ERR.
7415 [Ben Laurie (CHATS)]
7417 *) Protect against overlong session ID context length in an encoded
7418 session object. Since these are local, this does not appear to be
7420 [Ben Laurie (CHATS)]
7422 *) Change from security patch (see 0.9.6e below) that did not affect
7423 the 0.9.6 release series:
7425 Remote buffer overflow in SSL3 protocol - an attacker could
7426 supply an oversized master key in Kerberos-enabled versions.
7428 [Ben Laurie (CHATS)]
7430 *) Change the SSL kerb5 codes to match RFC 2712.
7433 *) Make -nameopt work fully for req and add -reqopt switch.
7434 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7436 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7437 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7439 *) Make sure tests can be performed even if the corresponding algorithms
7440 have been removed entirely. This was also the last step to make
7441 OpenSSL compilable with DJGPP under all reasonable conditions.
7442 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7444 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7445 to allow version independent disabling of normally unselected ciphers,
7446 which may be activated as a side-effect of selecting a single cipher.
7448 (E.g., cipher list string "RSA" enables ciphersuites that are left
7449 out of "ALL" because they do not provide symmetric encryption.
7450 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7451 [Lutz Jaenicke, Bodo Moeller]
7453 *) Add appropriate support for separate platform-dependent build
7454 directories. The recommended way to make a platform-dependent
7455 build directory is the following (tested on Linux), maybe with
7458 # Place yourself outside of the OpenSSL source tree. In
7459 # this example, the environment variable OPENSSL_SOURCE
7460 # is assumed to contain the absolute OpenSSL source directory.
7461 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7462 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7463 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7464 mkdir -p `dirname $F`
7465 ln -s $OPENSSL_SOURCE/$F $F
7468 To be absolutely sure not to disturb the source tree, a "make clean"
7469 is a good thing. If it isn't successful, don't worry about it,
7470 it probably means the source directory is very clean.
7473 *) Make sure any ENGINE control commands make local copies of string
7474 pointers passed to them whenever necessary. Otherwise it is possible
7475 the caller may have overwritten (or deallocated) the original string
7476 data when a later ENGINE operation tries to use the stored values.
7477 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7479 *) Improve diagnostics in file reading and command-line digests.
7480 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7482 *) Add AES modes CFB and OFB to the object database. Correct an
7483 error in AES-CFB decryption.
7486 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7487 allows existing EVP_CIPHER_CTX structures to be reused after
7488 calling EVP_*Final(). This behaviour is used by encryption
7489 BIOs and some applications. This has the side effect that
7490 applications must explicitly clean up cipher contexts with
7491 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7494 *) Check the values of dna and dnb in bn_mul_recursive before calling
7495 bn_mul_comba (a non zero value means the a or b arrays do not contain
7496 n2 elements) and fallback to bn_mul_normal if either is not zero.
7499 *) Fix escaping of non-ASCII characters when using the -subj option
7500 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7503 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7504 form for "surname", serialNumber has no short form.
7505 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7506 therefore remove "mail" short name for "internet 7".
7507 The OID for unique identifiers in X509 certificates is
7508 x500UniqueIdentifier, not uniqueIdentifier.
7509 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7512 *) Add an "init" command to the ENGINE config module and auto initialize
7513 ENGINEs. Without any "init" command the ENGINE will be initialized
7514 after all ctrl commands have been executed on it. If init=1 the
7515 ENGINE is initialized at that point (ctrls before that point are run
7516 on the uninitialized ENGINE and after on the initialized one). If
7517 init=0 then the ENGINE will not be initialized at all.
7520 *) Fix the 'app_verify_callback' interface so that the user-defined
7521 argument is actually passed to the callback: In the
7522 SSL_CTX_set_cert_verify_callback() prototype, the callback
7523 declaration has been changed from
7526 int (*cb)(X509_STORE_CTX *,void *);
7527 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7528 i=s->ctx->app_verify_callback(&ctx)
7529 has been changed into
7530 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7532 To update applications using SSL_CTX_set_cert_verify_callback(),
7533 a dummy argument can be added to their callback functions.
7534 [D. K. Smetters <smetters@parc.xerox.com>]
7536 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7537 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7539 *) Add and OPENSSL_LOAD_CONF define which will cause
7540 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7541 This allows older applications to transparently support certain
7542 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7543 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7544 load the config file and OPENSSL_add_all_algorithms_conf() which will
7545 always load it have also been added.
7548 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7549 Adjust NIDs and EVP layer.
7550 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7552 *) Config modules support in openssl utility.
7554 Most commands now load modules from the config file,
7555 though in a few (such as version) this isn't done
7556 because it couldn't be used for anything.
7558 In the case of ca and req the config file used is
7559 the same as the utility itself: that is the -config
7560 command line option can be used to specify an
7564 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7565 use "openssl_conf" if filename is NULL use default openssl config file.
7568 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7569 config section name. Add a new flag to tolerate a missing config file
7570 and move code to CONF_modules_load_file().
7573 *) Support for crypto accelerator cards from Accelerated Encryption
7574 Processing, www.aep.ie. (Use engine 'aep')
7575 The support was copied from 0.9.6c [engine] and adapted/corrected
7576 to work with the new engine framework.
7577 [AEP Inc. and Richard Levitte]
7579 *) Support for SureWare crypto accelerator cards from Baltimore
7580 Technologies. (Use engine 'sureware')
7581 The support was copied from 0.9.6c [engine] and adapted
7582 to work with the new engine framework.
7585 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7586 make the newer ENGINE framework commands for the CHIL engine work.
7587 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7589 *) Make it possible to produce shared libraries on ReliantUNIX.
7590 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7592 *) Add the configuration target debug-linux-ppro.
7593 Make 'openssl rsa' use the general key loading routines
7594 implemented in apps.c, and make those routines able to
7595 handle the key format FORMAT_NETSCAPE and the variant
7597 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7599 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7600 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7602 *) Add -keyform to rsautl, and document -engine.
7603 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7605 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7606 BIO_R_NO_SUCH_FILE error code rather than the generic
7607 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7610 *) Add new functions
7612 ERR_peek_last_error_line
7613 ERR_peek_last_error_line_data.
7614 These are similar to
7617 ERR_peek_error_line_data,
7618 but report on the latest error recorded rather than the first one
7619 still in the error queue.
7620 [Ben Laurie, Bodo Moeller]
7622 *) default_algorithms option in ENGINE config module. This allows things
7624 default_algorithms = ALL
7625 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7628 *) Preliminary ENGINE config module.
7631 *) New experimental application configuration code.
7634 *) Change the AES code to follow the same name structure as all other
7635 symmetric ciphers, and behave the same way. Move everything to
7636 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7637 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7639 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7640 [Ben Laurie and Theo de Raadt]
7642 *) Add option to output public keys in req command.
7643 [Massimiliano Pala madwolf@openca.org]
7645 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7646 (up to about 10% better than before for P-192 and P-224).
7649 *) New functions/macros
7651 SSL_CTX_set_msg_callback(ctx, cb)
7652 SSL_CTX_set_msg_callback_arg(ctx, arg)
7653 SSL_set_msg_callback(ssl, cb)
7654 SSL_set_msg_callback_arg(ssl, arg)
7656 to request calling a callback function
7658 void cb(int write_p, int version, int content_type,
7659 const void *buf, size_t len, SSL *ssl, void *arg)
7661 whenever a protocol message has been completely received
7662 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7663 protocol version according to which the SSL library interprets
7664 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7665 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7666 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7667 specification (change_cipher_spec(20), alert(21), handshake(22)).
7668 'buf' and 'len' point to the actual message, 'ssl' to the
7669 SSL object, and 'arg' is the application-defined value set by
7670 SSL[_CTX]_set_msg_callback_arg().
7672 'openssl s_client' and 'openssl s_server' have new '-msg' options
7673 to enable a callback that displays all protocol messages.
7676 *) Change the shared library support so shared libraries are built as
7677 soon as the corresponding static library is finished, and thereby get
7678 openssl and the test programs linked against the shared library.
7679 This still only happens when the keyword "shard" has been given to
7680 the configuration scripts.
7682 NOTE: shared library support is still an experimental thing, and
7683 backward binary compatibility is still not guaranteed.
7684 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7686 *) Add support for Subject Information Access extension.
7687 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7689 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7690 additional bytes when new memory had to be allocated, not just
7691 when reusing an existing buffer.
7694 *) New command line and configuration option 'utf8' for the req command.
7695 This allows field values to be specified as UTF8 strings.
7698 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7699 runs for the former and machine-readable output for the latter.
7702 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7703 of the e-mail address in the DN (i.e., it will go into a certificate
7704 extension only). The new configuration file option 'email_in_dn = no'
7705 has the same effect.
7706 [Massimiliano Pala madwolf@openca.org]
7708 *) Change all functions with names starting with des_ to be starting
7709 with DES_ instead. Add wrappers that are compatible with libdes,
7710 but are named _ossl_old_des_*. Finally, add macros that map the
7711 des_* symbols to the corresponding _ossl_old_des_* if libdes
7712 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7713 desired, the des_* symbols will be mapped to DES_*, with one
7716 Since we provide two compatibility mappings, the user needs to
7717 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7718 compatibility is desired. The default (i.e., when that macro
7719 isn't defined) is OpenSSL 0.9.6c compatibility.
7721 There are also macros that enable and disable the support of old
7722 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7723 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7724 are defined, the default will apply: to support the old des routines.
7726 In either case, one must include openssl/des.h to get the correct
7727 definitions. Do not try to just include openssl/des_old.h, that
7730 NOTE: This is a major break of an old API into a new one. Software
7731 authors are encouraged to switch to the DES_ style functions. Some
7732 time in the future, des_old.h and the libdes compatibility functions
7733 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7734 default), and then completely removed.
7737 *) Test for certificates which contain unsupported critical extensions.
7738 If such a certificate is found during a verify operation it is
7739 rejected by default: this behaviour can be overridden by either
7740 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7741 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7742 X509_supported_extension() has also been added which returns 1 if a
7743 particular extension is supported.
7746 *) Modify the behaviour of EVP cipher functions in similar way to digests
7747 to retain compatibility with existing code.
7750 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7751 compatibility with existing code. In particular the 'ctx' parameter does
7752 not have to be to be initialized before the call to EVP_DigestInit() and
7753 it is tidied up after a call to EVP_DigestFinal(). New function
7754 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7755 EVP_MD_CTX_copy() changed to not require the destination to be
7756 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7757 requires the destination to be valid.
7759 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7760 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7763 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7764 so that complete 'Handshake' protocol structures are kept in memory
7765 instead of overwriting 'msg_type' and 'length' with 'body' data.
7768 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7769 [Massimo Santin via Richard Levitte]
7771 *) Major restructuring to the underlying ENGINE code. This includes
7772 reduction of linker bloat, separation of pure "ENGINE" manipulation
7773 (initialisation, etc) from functionality dealing with implementations
7774 of specific crypto interfaces. This change also introduces integrated
7775 support for symmetric ciphers and digest implementations - so ENGINEs
7776 can now accelerate these by providing EVP_CIPHER and EVP_MD
7777 implementations of their own. This is detailed in crypto/engine/README
7778 as it couldn't be adequately described here. However, there are a few
7779 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7780 were changed in the original introduction of ENGINE code have now
7781 reverted back - the hooking from this code to ENGINE is now a good
7782 deal more passive and at run-time, operations deal directly with
7783 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7784 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7785 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7786 they were not being used by the framework as there is no concept of a
7787 BIGNUM_METHOD and they could not be generalised to the new
7788 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7789 ENGINE_cpy() has been removed as it cannot be consistently defined in
7793 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7796 *) Change mkdef.pl to sort symbols that get the same entry number,
7797 and make sure the automatically generated functions ERR_load_*
7798 become part of libeay.num as well.
7801 *) New function SSL_renegotiate_pending(). This returns true once
7802 renegotiation has been requested (either SSL_renegotiate() call
7803 or HelloRequest/ClientHello received from the peer) and becomes
7804 false once a handshake has been completed.
7805 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7806 sends a HelloRequest, but does not ensure that a handshake takes
7807 place. SSL_renegotiate_pending() is useful for checking if the
7808 client has followed the request.)
7811 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7812 By default, clients may request session resumption even during
7813 renegotiation (if session ID contexts permit); with this option,
7814 session resumption is possible only in the first handshake.
7816 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7817 more bits available for options that should not be part of
7818 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7821 *) Add some demos for certificate and certificate request creation.
7824 *) Make maximum certificate chain size accepted from the peer application
7825 settable (SSL*_get/set_max_cert_list()), as proposed by
7826 "Douglas E. Engert" <deengert@anl.gov>.
7829 *) Add support for shared libraries for Unixware-7
7830 (Boyd Lynn Gerber <gerberb@zenez.com>).
7833 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7834 be done prior to destruction. Use this to unload error strings from
7835 ENGINEs that load their own error strings. NB: This adds two new API
7836 functions to "get" and "set" this destroy handler in an ENGINE.
7839 *) Alter all existing ENGINE implementations (except "openssl" and
7840 "openbsd") to dynamically instantiate their own error strings. This
7841 makes them more flexible to be built both as statically-linked ENGINEs
7842 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7843 Also, add stub code to each that makes building them as self-contained
7844 shared-libraries easier (see README.ENGINE).
7847 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7848 implementations into applications that are completely implemented in
7849 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7850 commands that can be used to configure what shared-library to load and
7851 to control aspects of the way it is handled. Also, made an update to
7852 the README.ENGINE file that brings its information up-to-date and
7853 provides some information and instructions on the "dynamic" ENGINE
7854 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7857 *) Make it possible to unload ranges of ERR strings with a new
7858 "ERR_unload_strings" function.
7861 *) Add a copy() function to EVP_MD.
7864 *) Make EVP_MD routines take a context pointer instead of just the
7865 md_data void pointer.
7868 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7869 that the digest can only process a single chunk of data
7870 (typically because it is provided by a piece of
7871 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7872 is only going to provide a single chunk of data, and hence the
7873 framework needn't accumulate the data for oneshot drivers.
7876 *) As with "ERR", make it possible to replace the underlying "ex_data"
7877 functions. This change also alters the storage and management of global
7878 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7879 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7880 index counters. The API functions that use this state have been changed
7881 to take a "class_index" rather than pointers to the class's local STACK
7882 and counter, and there is now an API function to dynamically create new
7883 classes. This centralisation allows us to (a) plug a lot of the
7884 thread-safety problems that existed, and (b) makes it possible to clean
7885 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7886 such data would previously have always leaked in application code and
7887 workarounds were in place to make the memory debugging turn a blind eye
7888 to it. Application code that doesn't use this new function will still
7889 leak as before, but their memory debugging output will announce it now
7890 rather than letting it slide.
7892 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7893 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7894 has a return value to indicate success or failure.
7897 *) Make it possible to replace the underlying "ERR" functions such that the
7898 global state (2 LHASH tables and 2 locks) is only used by the "default"
7899 implementation. This change also adds two functions to "get" and "set"
7900 the implementation prior to it being automatically set the first time
7901 any other ERR function takes place. Ie. an application can call "get",
7902 pass the return value to a module it has just loaded, and that module
7903 can call its own "set" function using that value. This means the
7904 module's "ERR" operations will use (and modify) the error state in the
7905 application and not in its own statically linked copy of OpenSSL code.
7908 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7909 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7910 the operation, and provides a more encapsulated way for external code
7911 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7912 to use these functions rather than manually incrementing the counts.
7914 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7917 *) Add EVP test program.
7920 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7923 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7924 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7925 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7926 These allow a CRL to be built without having to access X509_CRL fields
7927 directly. Modify 'ca' application to use new functions.
7930 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7931 bug workarounds. Rollback attack detection is a security feature.
7932 The problem will only arise on OpenSSL servers when TLSv1 is not
7933 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7934 Software authors not wanting to support TLSv1 will have special reasons
7935 for their choice and can explicitly enable this option.
7936 [Bodo Moeller, Lutz Jaenicke]
7938 *) Rationalise EVP so it can be extended: don't include a union of
7939 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7940 (similar to those existing for EVP_CIPHER_CTX).
7945 EVP_MD_CTX_init(&md); /* new function call */
7946 EVP_DigestInit(&md, EVP_sha1());
7947 EVP_DigestUpdate(&md, in, len);
7948 EVP_DigestFinal(&md, out, NULL);
7949 EVP_MD_CTX_cleanup(&md); /* new function call */
7953 *) Make DES key schedule conform to the usual scheme, as well as
7954 correcting its structure. This means that calls to DES functions
7955 now have to pass a pointer to a des_key_schedule instead of a
7956 plain des_key_schedule (which was actually always a pointer
7959 des_key_schedule ks;
7961 des_set_key_checked(..., &ks);
7962 des_ncbc_encrypt(..., &ks, ...);
7964 (Note that a later change renames 'des_...' into 'DES_...'.)
7967 *) Initial reduction of linker bloat: the use of some functions, such as
7968 PEM causes large amounts of unused functions to be linked in due to
7969 poor organisation. For example pem_all.c contains every PEM function
7970 which has a knock on effect of linking in large amounts of (unused)
7971 ASN1 code. Grouping together similar functions and splitting unrelated
7972 functions prevents this.
7975 *) Cleanup of EVP macros.
7978 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7979 correct _ecb suffix.
7982 *) Add initial OCSP responder support to ocsp application. The
7983 revocation information is handled using the text based index
7984 use by the ca application. The responder can either handle
7985 requests generated internally, supplied in files (for example
7986 via a CGI script) or using an internal minimal server.
7989 *) Add configuration choices to get zlib compression for TLS.
7992 *) Changes to Kerberos SSL for RFC 2712 compliance:
7993 1. Implemented real KerberosWrapper, instead of just using
7994 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7995 2. Implemented optional authenticator field of KerberosWrapper.
7997 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7998 and authenticator structs; see crypto/krb5/.
8000 Generalized Kerberos calls to support multiple Kerberos libraries.
8001 [Vern Staats <staatsvr@asc.hpc.mil>,
8002 Jeffrey Altman <jaltman@columbia.edu>
8003 via Richard Levitte]
8005 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
8006 already does with RSA. testdsa.h now has 'priv_key/pub_key'
8007 values for each of the key sizes rather than having just
8008 parameters (and 'speed' generating keys each time).
8011 *) Speed up EVP routines.
8014 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
8015 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
8016 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
8017 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
8019 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
8020 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
8021 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
8024 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
8026 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
8029 *) Added the OS2-EMX target.
8030 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
8032 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
8033 to support NCONF routines in extension code. New function CONF_set_nconf()
8034 to allow functions which take an NCONF to also handle the old LHASH
8035 structure: this means that the old CONF compatible routines can be
8036 retained (in particular wrt extensions) without having to duplicate the
8037 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
8040 *) Enhance the general user interface with mechanisms for inner control
8041 and with possibilities to have yes/no kind of prompts.
8044 *) Change all calls to low level digest routines in the library and
8045 applications to use EVP. Add missing calls to HMAC_cleanup() and
8046 don't assume HMAC_CTX can be copied using memcpy().
8047 [Verdon Walker <VWalker@novell.com>, Steve Henson]
8049 *) Add the possibility to control engines through control names but with
8050 arbitrary arguments instead of just a string.
8051 Change the key loaders to take a UI_METHOD instead of a callback
8052 function pointer. NOTE: this breaks binary compatibility with earlier
8053 versions of OpenSSL [engine].
8054 Adapt the nCipher code for these new conditions and add a card insertion
8058 *) Enhance the general user interface with mechanisms to better support
8059 dialog box interfaces, application-defined prompts, the possibility
8060 to use defaults (for example default passwords from somewhere else)
8061 and interrupts/cancellations.
8064 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
8065 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
8068 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
8069 tidy up some unnecessarily weird code in 'sk_new()').
8070 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
8072 *) Change the key loading routines for ENGINEs to use the same kind
8073 callback (pem_password_cb) as all other routines that need this
8077 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
8078 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
8079 than this minimum value is recommended.
8082 *) New random seeder for OpenVMS, using the system process statistics
8083 that are easily reachable.
8086 *) Windows apparently can't transparently handle global
8087 variables defined in DLLs. Initialisations such as:
8089 const ASN1_ITEM *it = &ASN1_INTEGER_it;
8091 won't compile. This is used by the any applications that need to
8092 declare their own ASN1 modules. This was fixed by adding the option
8093 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
8094 needed for static libraries under Win32.
8097 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
8098 setting of purpose and trust fields. New X509_STORE trust and
8099 purpose functions and tidy up setting in other SSL functions.
8102 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
8103 structure. These are inherited by X509_STORE_CTX when it is
8104 initialised. This allows various defaults to be set in the
8105 X509_STORE structure (such as flags for CRL checking and custom
8106 purpose or trust settings) for functions which only use X509_STORE_CTX
8107 internally such as S/MIME.
8109 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
8110 trust settings if they are not set in X509_STORE. This allows X509_STORE
8111 purposes and trust (in S/MIME for example) to override any set by default.
8113 Add command line options for CRL checking to smime, s_client and s_server
8117 *) Initial CRL based revocation checking. If the CRL checking flag(s)
8118 are set then the CRL is looked up in the X509_STORE structure and
8119 its validity and signature checked, then if the certificate is found
8120 in the CRL the verify fails with a revoked error.
8122 Various new CRL related callbacks added to X509_STORE_CTX structure.
8124 Command line options added to 'verify' application to support this.
8126 This needs some additional work, such as being able to handle multiple
8127 CRLs with different times, extension based lookup (rather than just
8128 by subject name) and ultimately more complete V2 CRL extension
8132 *) Add a general user interface API (crypto/ui/). This is designed
8133 to replace things like des_read_password and friends (backward
8134 compatibility functions using this new API are provided).
8135 The purpose is to remove prompting functions from the DES code
8136 section as well as provide for prompting through dialog boxes in
8137 a window system and the like.
8140 *) Add "ex_data" support to ENGINE so implementations can add state at a
8141 per-structure level rather than having to store it globally.
8144 *) Make it possible for ENGINE structures to be copied when retrieved by
8145 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
8146 This causes the "original" ENGINE structure to act like a template,
8147 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
8148 operational state can be localised to each ENGINE structure, despite the
8149 fact they all share the same "methods". New ENGINE structures returned in
8150 this case have no functional references and the return value is the single
8151 structural reference. This matches the single structural reference returned
8152 by ENGINE_by_id() normally, when it is incremented on the pre-existing
8156 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
8157 needs to match any other type at all we need to manually clear the
8161 *) Changes to the "openssl engine" utility to include;
8162 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
8163 about an ENGINE's available control commands.
8164 - executing control commands from command line arguments using the
8165 '-pre' and '-post' switches. '-post' is only used if '-t' is
8166 specified and the ENGINE is successfully initialised. The syntax for
8167 the individual commands are colon-separated, for example;
8168 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
8171 *) New dynamic control command support for ENGINEs. ENGINEs can now
8172 declare their own commands (numbers), names (strings), descriptions,
8173 and input types for run-time discovery by calling applications. A
8174 subset of these commands are implicitly classed as "executable"
8175 depending on their input type, and only these can be invoked through
8176 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
8177 can be based on user input, config files, etc). The distinction is
8178 that "executable" commands cannot return anything other than a boolean
8179 result and can only support numeric or string input, whereas some
8180 discoverable commands may only be for direct use through
8181 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
8182 pointers, or other custom uses. The "executable" commands are to
8183 support parameterisations of ENGINE behaviour that can be
8184 unambiguously defined by ENGINEs and used consistently across any
8185 OpenSSL-based application. Commands have been added to all the
8186 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
8187 control over shared-library paths without source code alterations.
8190 *) Changed all ENGINE implementations to dynamically allocate their
8191 ENGINEs rather than declaring them statically. Apart from this being
8192 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
8193 this also allows the implementations to compile without using the
8194 internal engine_int.h header.
8197 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8198 'const' value. Any code that should be able to modify a RAND_METHOD
8199 should already have non-const pointers to it (ie. they should only
8200 modify their own ones).
8203 *) Made a variety of little tweaks to the ENGINE code.
8204 - "atalla" and "ubsec" string definitions were moved from header files
8205 to C code. "nuron" string definitions were placed in variables
8206 rather than hard-coded - allowing parameterisation of these values
8207 later on via ctrl() commands.
8208 - Removed unused "#if 0"'d code.
8209 - Fixed engine list iteration code so it uses ENGINE_free() to release
8210 structural references.
8211 - Constified the RAND_METHOD element of ENGINE structures.
8212 - Constified various get/set functions as appropriate and added
8213 missing functions (including a catch-all ENGINE_cpy that duplicates
8214 all ENGINE values onto a new ENGINE except reference counts/state).
8215 - Removed NULL parameter checks in get/set functions. Setting a method
8216 or function to NULL is a way of cancelling out a previously set
8217 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8218 and doesn't justify the extra error symbols and code.
8219 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8220 flags from engine_int.h to engine.h.
8221 - Changed prototypes for ENGINE handler functions (init(), finish(),
8222 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8225 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8226 to the algorithm using long division. The binary algorithm can be
8227 used only if the modulus is odd. On 32-bit systems, it is faster
8228 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8229 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8230 up to 450 bits. In 64-bit environments, the binary algorithm
8231 appears to be advantageous for much longer moduli; here we use it
8232 for moduli up to 2048 bits.
8235 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8236 could not support the combine flag in choice fields.
8239 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8240 extensions from a certificate request to the certificate.
8243 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8244 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8245 file: this allows the display of the certificate about to be
8246 signed to be customised, to allow certain fields to be included
8247 or excluded and extension details. The old system didn't display
8248 multicharacter strings properly, omitted fields not in the policy
8249 and couldn't display additional details such as extensions.
8252 *) Function EC_POINTs_mul for multiple scalar multiplication
8253 of an arbitrary number of elliptic curve points
8254 \sum scalars[i]*points[i],
8255 optionally including the generator defined for the EC_GROUP:
8256 scalar*generator + \sum scalars[i]*points[i].
8258 EC_POINT_mul is a simple wrapper function for the typical case
8259 that the point list has just one item (besides the optional
8263 *) First EC_METHODs for curves over GF(p):
8265 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8266 operations and provides various method functions that can also
8267 operate with faster implementations of modular arithmetic.
8269 EC_GFp_mont_method() reuses most functions that are part of
8270 EC_GFp_simple_method, but uses Montgomery arithmetic.
8272 [Bodo Moeller; point addition and point doubling
8273 implementation directly derived from source code provided by
8274 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8276 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8277 crypto/ec/ec_lib.c):
8279 Curves are EC_GROUP objects (with an optional group generator)
8280 based on EC_METHODs that are built into the library.
8282 Points are EC_POINT objects based on EC_GROUP objects.
8284 Most of the framework would be able to handle curves over arbitrary
8285 finite fields, but as there are no obvious types for fields other
8286 than GF(p), some functions are limited to that for now.
8289 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8290 that the file contains a complete HTTP response.
8293 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8294 change the def and num file printf format specifier from "%-40sXXX"
8295 to "%-39s XXX". The latter will always guarantee a space after the
8296 field while the former will cause them to run together if the field
8297 is 40 of more characters long.
8300 *) Constify the cipher and digest 'method' functions and structures
8301 and modify related functions to take constant EVP_MD and EVP_CIPHER
8305 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8306 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8309 *) Modify EVP_Digest*() routines so they now return values. Although the
8310 internal software routines can never fail additional hardware versions
8314 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8316 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8317 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8320 ERR_R_NESTED_ASN1_ERROR
8322 ERR_R_MISSING_ASN1_EOS
8323 were 4 .. 9, conflicting with
8324 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8326 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8327 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8329 Add new error code 'ERR_R_INTERNAL_ERROR'.
8332 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8336 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8337 sets the subject name for a new request or supersedes the
8338 subject name in a given request. Formats that can be parsed are
8339 'CN=Some Name, OU=myOU, C=IT'
8341 'CN=Some Name/OU=myOU/C=IT'.
8343 Add options '-batch' and '-verbose' to 'openssl req'.
8344 [Massimiliano Pala <madwolf@hackmasters.net>]
8346 *) Introduce the possibility to access global variables through
8347 functions on platform were that's the best way to handle exporting
8348 global variables in shared libraries. To enable this functionality,
8349 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8350 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8351 is normally done by Configure or something similar).
8353 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8354 in the source file (foo.c) like this:
8356 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8357 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8359 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8360 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8362 OPENSSL_DECLARE_GLOBAL(int,foo);
8363 #define foo OPENSSL_GLOBAL_REF(foo)
8364 OPENSSL_DECLARE_GLOBAL(double,bar);
8365 #define bar OPENSSL_GLOBAL_REF(bar)
8367 The #defines are very important, and therefore so is including the
8368 header file everywhere where the defined globals are used.
8370 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8371 of ASN.1 items, but that structure is a bit different.
8373 The largest change is in util/mkdef.pl which has been enhanced with
8374 better and easier to understand logic to choose which symbols should
8375 go into the Windows .def files as well as a number of fixes and code
8376 cleanup (among others, algorithm keywords are now sorted
8377 lexicographically to avoid constant rewrites).
8380 *) In BN_div() keep a copy of the sign of 'num' before writing the
8381 result to 'rm' because if rm==num the value will be overwritten
8382 and produce the wrong result if 'num' is negative: this caused
8383 problems with BN_mod() and BN_nnmod().
8386 *) Function OCSP_request_verify(). This checks the signature on an
8387 OCSP request and verifies the signer certificate. The signer
8388 certificate is just checked for a generic purpose and OCSP request
8392 *) Add OCSP_check_validity() function to check the validity of OCSP
8393 responses. OCSP responses are prepared in real time and may only
8394 be a few seconds old. Simply checking that the current time lies
8395 between thisUpdate and nextUpdate max reject otherwise valid responses
8396 caused by either OCSP responder or client clock inaccuracy. Instead
8397 we allow thisUpdate and nextUpdate to fall within a certain period of
8398 the current time. The age of the response can also optionally be
8399 checked. Two new options -validity_period and -status_age added to
8403 *) If signature or public key algorithm is unrecognized print out its
8404 OID rather that just UNKNOWN.
8407 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8408 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8409 ID to be generated from the issuer certificate alone which can then be
8410 passed to OCSP_id_issuer_cmp().
8413 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8414 ASN1 modules to export functions returning ASN1_ITEM pointers
8415 instead of the ASN1_ITEM structures themselves. This adds several
8416 new macros which allow the underlying ASN1 function/structure to
8417 be accessed transparently. As a result code should not use ASN1_ITEM
8418 references directly (such as &X509_it) but instead use the relevant
8419 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8420 use of the new ASN1 code on platforms where exporting structures
8421 is problematical (for example in shared libraries) but exporting
8422 functions returning pointers to structures is not.
8425 *) Add support for overriding the generation of SSL/TLS session IDs.
8426 These callbacks can be registered either in an SSL_CTX or per SSL.
8427 The purpose of this is to allow applications to control, if they wish,
8428 the arbitrary values chosen for use as session IDs, particularly as it
8429 can be useful for session caching in multiple-server environments. A
8430 command-line switch for testing this (and any client code that wishes
8431 to use such a feature) has been added to "s_server".
8432 [Geoff Thorpe, Lutz Jaenicke]
8434 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8435 of the form '#if defined(...) || defined(...) || ...' and
8436 '#if !defined(...) && !defined(...) && ...'. This also avoids
8437 the growing number of special cases it was previously handling.
8440 *) Make all configuration macros available for application by making
8441 sure they are available in opensslconf.h, by giving them names starting
8442 with "OPENSSL_" to avoid conflicts with other packages and by making
8443 sure e_os2.h will cover all platform-specific cases together with
8445 Additionally, it is now possible to define configuration/platform-
8446 specific names (called "system identities"). In the C code, these
8447 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8448 macro with the name beginning with "OPENSSL_SYS_", which is determined
8449 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8453 *) New option -set_serial to 'req' and 'x509' this allows the serial
8454 number to use to be specified on the command line. Previously self
8455 signed certificates were hard coded with serial number 0 and the
8456 CA options of 'x509' had to use a serial number in a file which was
8460 *) New options to 'ca' utility to support V2 CRL entry extensions.
8461 Currently CRL reason, invalidity date and hold instruction are
8462 supported. Add new CRL extensions to V3 code and some new objects.
8465 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8466 disable standard block padding (aka PKCS#5 padding) in the EVP
8467 API, which was previously mandatory. This means that the data is
8468 not padded in any way and so the total length much be a multiple
8469 of the block size, otherwise an error occurs.
8472 *) Initial (incomplete) OCSP SSL support.
8475 *) New function OCSP_parse_url(). This splits up a URL into its host,
8476 port and path components: primarily to parse OCSP URLs. New -url
8477 option to ocsp utility.
8480 *) New nonce behavior. The return value of OCSP_check_nonce() now
8481 reflects the various checks performed. Applications can decide
8482 whether to tolerate certain situations such as an absent nonce
8483 in a response when one was present in a request: the ocsp application
8484 just prints out a warning. New function OCSP_add1_basic_nonce()
8485 this is to allow responders to include a nonce in a response even if
8486 the request is nonce-less.
8489 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8490 skipped when using openssl x509 multiple times on a single input file,
8491 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8494 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8495 set string type: to handle setting ASN1_TIME structures. Fix ca
8496 utility to correctly initialize revocation date of CRLs.
8499 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8500 the clients preferred ciphersuites and rather use its own preferences.
8501 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8502 Internet Explorer by ensuring unchanged hash method during stepup.
8503 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8506 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8507 to aes and add a new 'exist' option to print out symbols that don't
8511 *) Additional options to ocsp utility to allow flags to be set and
8512 additional certificates supplied.
8515 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8516 OCSP client a number of certificate to only verify the response
8520 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8521 handle the new API. Currently only ECB, CBC modes supported. Add new
8524 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8525 Encryption Standard (AES) Ciphersuites for Transport Layer
8526 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8527 not enabled by default and were not part of the "ALL" ciphersuite
8528 alias because they were not yet official; they could be
8529 explicitly requested by specifying the "AESdraft" ciphersuite
8530 group alias. In the final release of OpenSSL 0.9.7, the group
8531 alias is called "AES" and is part of "ALL".)
8532 [Ben Laurie, Steve Henson, Bodo Moeller]
8534 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8535 request to response.
8538 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8539 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8540 extract information from a certificate request. OCSP_response_create()
8541 creates a response and optionally adds a basic response structure.
8542 OCSP_basic_add1_status() adds a complete single response to a basic
8543 response and returns the OCSP_SINGLERESP structure just added (to allow
8544 extensions to be included for example). OCSP_basic_add1_cert() adds a
8545 certificate to a basic response and OCSP_basic_sign() signs a basic
8546 response with various flags. New helper functions ASN1_TIME_check()
8547 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8548 (converts ASN1_TIME to GeneralizedTime).
8551 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8552 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8553 structure from a certificate. X509_pubkey_digest() digests the public_key
8554 contents: this is used in various key identifiers.
8557 *) Make sk_sort() tolerate a NULL argument.
8558 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8560 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8561 passed by the function are trusted implicitly. If any of them signed the
8562 response then it is assumed to be valid and is not verified.
8565 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8566 to data. This was previously part of the PKCS7 ASN1 code. This
8567 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8568 [Steve Henson, reported by Kenneth R. Robinette
8569 <support@securenetterm.com>]
8571 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8572 routines: without these tracing memory leaks is very painful.
8573 Fix leaks in PKCS12 and PKCS7 routines.
8576 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8577 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8578 effectively meant GeneralizedTime would never be used. Now it
8579 is initialised to -1 but X509_time_adj() now has to check the value
8580 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8581 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8582 [Steve Henson, reported by Kenneth R. Robinette
8583 <support@securenetterm.com>]
8585 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8586 result in a zero length in the ASN1_INTEGER structure which was
8587 not consistent with the structure when d2i_ASN1_INTEGER() was used
8588 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8589 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8590 where it did not print out a minus for negative ASN1_INTEGER.
8593 *) Add summary printout to ocsp utility. The various functions which
8594 convert status values to strings have been renamed to:
8595 OCSP_response_status_str(), OCSP_cert_status_str() and
8596 OCSP_crl_reason_str() and are no longer static. New options
8597 to verify nonce values and to disable verification. OCSP response
8598 printout format cleaned up.
8601 *) Add additional OCSP certificate checks. These are those specified
8602 in RFC2560. This consists of two separate checks: the CA of the
8603 certificate being checked must either be the OCSP signer certificate
8604 or the issuer of the OCSP signer certificate. In the latter case the
8605 OCSP signer certificate must contain the OCSP signing extended key
8606 usage. This check is performed by attempting to match the OCSP
8607 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8608 in the OCSP_CERTID structures of the response.
8611 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8612 and related routines. This uses the standard OpenSSL certificate
8613 verify routines to perform initial checks (just CA validity) and
8614 to obtain the certificate chain. Then additional checks will be
8615 performed on the chain. Currently the root CA is checked to see
8616 if it is explicitly trusted for OCSP signing. This is used to set
8617 a root CA as a global signing root: that is any certificate that
8618 chains to that CA is an acceptable OCSP signing certificate.
8621 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8622 extensions from a separate configuration file.
8623 As when reading extensions from the main configuration file,
8624 the '-extensions ...' option may be used for specifying the
8626 [Massimiliano Pala <madwolf@comune.modena.it>]
8628 *) New OCSP utility. Allows OCSP requests to be generated or
8629 read. The request can be sent to a responder and the output
8630 parsed, outputted or printed in text form. Not complete yet:
8631 still needs to check the OCSP response validity.
8634 *) New subcommands for 'openssl ca':
8635 'openssl ca -status <serial>' prints the status of the cert with
8636 the given serial number (according to the index file).
8637 'openssl ca -updatedb' updates the expiry status of certificates
8639 [Massimiliano Pala <madwolf@comune.modena.it>]
8641 *) New '-newreq-nodes' command option to CA.pl. This is like
8642 '-newreq', but calls 'openssl req' with the '-nodes' option
8643 so that the resulting key is not encrypted.
8644 [Damien Miller <djm@mindrot.org>]
8646 *) New configuration for the GNU Hurd.
8647 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8649 *) Initial code to implement OCSP basic response verify. This
8650 is currently incomplete. Currently just finds the signer's
8651 certificate and verifies the signature on the response.
8654 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8655 value of OPENSSLDIR. This is available via the new '-d' option
8656 to 'openssl version', and is also included in 'openssl version -a'.
8659 *) Allowing defining memory allocation callbacks that will be given
8660 file name and line number information in additional arguments
8661 (a const char* and an int). The basic functionality remains, as
8662 well as the original possibility to just replace malloc(),
8663 realloc() and free() by functions that do not know about these
8664 additional arguments. To register and find out the current
8665 settings for extended allocation functions, the following
8666 functions are provided:
8668 CRYPTO_set_mem_ex_functions
8669 CRYPTO_set_locked_mem_ex_functions
8670 CRYPTO_get_mem_ex_functions
8671 CRYPTO_get_locked_mem_ex_functions
8673 These work the same way as CRYPTO_set_mem_functions and friends.
8674 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8675 extended allocation function is enabled.
8676 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8677 a conventional allocation function is enabled.
8678 [Richard Levitte, Bodo Moeller]
8680 *) Finish off removing the remaining LHASH function pointer casts.
8681 There should no longer be any prototype-casting required when using
8682 the LHASH abstraction, and any casts that remain are "bugs". See
8683 the callback types and macros at the head of lhash.h for details
8684 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8687 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8688 If /dev/[u]random devices are not available or do not return enough
8689 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8691 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8692 /etc/entropy will be queried once each in this sequence, querying stops
8693 when enough entropy was collected without querying more sockets.
8696 *) Change the Unix RAND_poll() variant to be able to poll several
8697 random devices, as specified by DEVRANDOM, until a sufficient amount
8698 of data has been collected. We spend at most 10 ms on each file
8699 (select timeout) and read in non-blocking mode. DEVRANDOM now
8700 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8701 (previously it was just the string "/dev/urandom"), so on typical
8702 platforms the 10 ms delay will never occur.
8703 Also separate out the Unix variant to its own file, rand_unix.c.
8704 For VMS, there's a currently-empty rand_vms.c.
8707 *) Move OCSP client related routines to ocsp_cl.c. These
8708 provide utility functions which an application needing
8709 to issue a request to an OCSP responder and analyse the
8710 response will typically need: as opposed to those which an
8711 OCSP responder itself would need which will be added later.
8713 OCSP_request_sign() signs an OCSP request with an API similar
8714 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8715 response. OCSP_response_get1_basic() extracts basic response
8716 from response. OCSP_resp_find_status(): finds and extracts status
8717 information from an OCSP_CERTID structure (which will be created
8718 when the request structure is built). These are built from lower
8719 level functions which work on OCSP_SINGLERESP structures but
8720 won't normally be used unless the application wishes to examine
8721 extensions in the OCSP response for example.
8723 Replace nonce routines with a pair of functions.
8724 OCSP_request_add1_nonce() adds a nonce value and optionally
8725 generates a random value. OCSP_check_nonce() checks the
8726 validity of the nonce in an OCSP response.
8729 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8730 This doesn't copy the supplied OCSP_CERTID and avoids the
8731 need to free up the newly created id. Change return type
8732 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8733 This can then be used to add extensions to the request.
8734 Deleted OCSP_request_new(), since most of its functionality
8735 is now in OCSP_REQUEST_new() (and the case insensitive name
8736 clash) apart from the ability to set the request name which
8737 will be added elsewhere.
8740 *) Update OCSP API. Remove obsolete extensions argument from
8741 various functions. Extensions are now handled using the new
8742 OCSP extension code. New simple OCSP HTTP function which
8743 can be used to send requests and parse the response.
8746 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8747 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8748 uses the special reorder version of SET OF to sort the attributes
8749 and reorder them to match the encoded order. This resolves a long
8750 standing problem: a verify on a PKCS7 structure just after signing
8751 it used to fail because the attribute order did not match the
8752 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8753 it uses the received order. This is necessary to tolerate some broken
8754 software that does not order SET OF. This is handled by encoding
8755 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8756 to produce the required SET OF.
8759 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8760 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8761 files to get correct declarations of the ASN.1 item variables.
8764 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8765 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8766 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8767 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8768 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8769 ASN1_ITEM and no wrapper functions.
8772 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8773 replace the old function pointer based I/O routines. Change most of
8774 the *_d2i_bio() and *_d2i_fp() functions to use these.
8777 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8778 lines, recognize more "algorithms" that can be deselected, and make
8779 it complain about algorithm deselection that isn't recognised.
8782 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8783 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8784 to use new functions. Add NO_ASN1_OLD which can be set to remove
8785 some old style ASN1 functions: this can be used to determine if old
8786 code will still work when these eventually go away.
8789 *) New extension functions for OCSP structures, these follow the
8790 same conventions as certificates and CRLs.
8793 *) New function X509V3_add1_i2d(). This automatically encodes and
8794 adds an extension. Its behaviour can be customised with various
8795 flags to append, replace or delete. Various wrappers added for
8796 certificates and CRLs.
8799 *) Fix to avoid calling the underlying ASN1 print routine when
8800 an extension cannot be parsed. Correct a typo in the
8801 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8804 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8805 entries for variables.
8808 *) Add functionality to apps/openssl.c for detecting locking
8809 problems: As the program is single-threaded, all we have
8810 to do is register a locking callback using an array for
8811 storing which locks are currently held by the program.
8814 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8815 SSL_get_ex_data_X509_STORE_idx(), which is used in
8816 ssl_verify_cert_chain() and thus can be called at any time
8817 during TLS/SSL handshakes so that thread-safety is essential.
8818 Unfortunately, the ex_data design is not at all suited
8819 for multi-threaded use, so it probably should be abolished.
8822 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8823 [Broadcom, tweaked and integrated by Geoff Thorpe]
8825 *) Move common extension printing code to new function
8826 X509V3_print_extensions(). Reorganise OCSP print routines and
8827 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8830 *) New function X509_signature_print() to remove duplication in some
8834 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8835 set (this was treated exactly the same as SET OF previously). This
8836 is used to reorder the STACK representing the structure to match the
8837 encoding. This will be used to get round a problem where a PKCS7
8838 structure which was signed could not be verified because the STACK
8839 order did not reflect the encoded order.
8842 *) Reimplement the OCSP ASN1 module using the new code.
8845 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8846 for its ASN1 operations. The old style function pointers still exist
8847 for now but they will eventually go away.
8850 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8851 completely replaces the old ASN1 functionality with a table driven
8852 encoder and decoder which interprets an ASN1_ITEM structure describing
8853 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8854 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8855 has also been converted to the new form.
8858 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8859 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8860 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8861 for negative moduli.
8864 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8865 of not touching the result's sign bit.
8868 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8872 *) Changed the LHASH code to use prototypes for callbacks, and created
8873 macros to declare and implement thin (optionally static) functions
8874 that provide type-safety and avoid function pointer casting for the
8875 type-specific callbacks.
8878 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8880 [Veers Staats <staatsvr@asc.hpc.mil>,
8881 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8883 *) Reformat the FAQ so the different questions and answers can be divided
8884 in sections depending on the subject.
8887 *) Have the zlib compression code load ZLIB.DLL dynamically under
8891 *) New function BN_mod_sqrt for computing square roots modulo a prime
8892 (using the probabilistic Tonelli-Shanks algorithm unless
8893 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8894 be handled deterministically).
8895 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8897 *) Make BN_mod_inverse faster by explicitly handling small quotients
8898 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8899 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8902 *) New function BN_kronecker.
8905 *) Fix BN_gcd so that it works on negative inputs; the result is
8906 positive unless both parameters are zero.
8907 Previously something reasonably close to an infinite loop was
8908 possible because numbers could be growing instead of shrinking
8909 in the implementation of Euclid's algorithm.
8912 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8913 sign of the number in question.
8915 Fix BN_is_word(a,w) to work correctly for w == 0.
8917 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8918 because its test if the absolute value of 'a' equals 'w'.
8919 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8920 it exists mostly for use in the implementations of BN_is_zero(),
8921 BN_is_one(), and BN_is_word().
8924 *) New function BN_swap.
8927 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8928 the exponentiation functions are more likely to produce reasonable
8929 results on negative inputs.
8932 *) Change BN_mod_mul so that the result is always non-negative.
8933 Previously, it could be negative if one of the factors was negative;
8934 I don't think anyone really wanted that behaviour.
8937 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8938 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8939 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8940 and add new functions:
8949 BN_mod_lshift1_quick
8953 These functions always generate non-negative results.
8955 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8956 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8958 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8959 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8960 be reduced modulo m.
8961 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8964 The following entry accidentally appeared in the CHANGES file
8965 distributed with OpenSSL 0.9.7. The modifications described in
8966 it do *not* apply to OpenSSL 0.9.7.
8968 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8969 was actually never needed) and in BN_mul(). The removal in BN_mul()
8970 required a small change in bn_mul_part_recursive() and the addition
8971 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8972 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8973 bn_sub_words() and bn_add_words() except they take arrays with
8978 *) In 'openssl passwd', verify passwords read from the terminal
8979 unless the '-salt' option is used (which usually means that
8980 verification would just waste user's time since the resulting
8981 hash is going to be compared with some given password hash)
8982 or the new '-noverify' option is used.
8984 This is an incompatible change, but it does not affect
8985 non-interactive use of 'openssl passwd' (passwords on the command
8986 line, '-stdin' option, '-in ...' option) and thus should not
8990 *) Remove all references to RSAref, since there's no more need for it.
8993 *) Make DSO load along a path given through an environment variable
8994 (SHLIB_PATH) with shl_load().
8997 *) Constify the ENGINE code as a result of BIGNUM constification.
8998 Also constify the RSA code and most things related to it. In a
8999 few places, most notable in the depth of the ASN.1 code, ugly
9000 casts back to non-const were required (to be solved at a later
9004 *) Make it so the openssl application has all engines loaded by default.
9007 *) Constify the BIGNUM routines a little more.
9010 *) Add the following functions:
9012 ENGINE_load_cswift()
9014 ENGINE_load_atalla()
9016 ENGINE_load_builtin_engines()
9018 That way, an application can itself choose if external engines that
9019 are built-in in OpenSSL shall ever be used or not. The benefit is
9020 that applications won't have to be linked with libdl or other dso
9021 libraries unless it's really needed.
9023 Changed 'openssl engine' to load all engines on demand.
9024 Changed the engine header files to avoid the duplication of some
9025 declarations (they differed!).
9028 *) 'openssl engine' can now list capabilities.
9031 *) Better error reporting in 'openssl engine'.
9034 *) Never call load_dh_param(NULL) in s_server.
9037 *) Add engine application. It can currently list engines by name and
9038 identity, and test if they are actually available.
9041 *) Improve RPM specification file by forcing symbolic linking and making
9042 sure the installed documentation is also owned by root.root.
9043 [Damien Miller <djm@mindrot.org>]
9045 *) Give the OpenSSL applications more possibilities to make use of
9046 keys (public as well as private) handled by engines.
9049 *) Add OCSP code that comes from CertCo.
9052 *) Add VMS support for the Rijndael code.
9055 *) Added untested support for Nuron crypto accelerator.
9058 *) Add support for external cryptographic devices. This code was
9059 previously distributed separately as the "engine" branch.
9060 [Geoff Thorpe, Richard Levitte]
9062 *) Rework the filename-translation in the DSO code. It is now possible to
9063 have far greater control over how a "name" is turned into a filename
9064 depending on the operating environment and any oddities about the
9065 different shared library filenames on each system.
9068 *) Support threads on FreeBSD-elf in Configure.
9071 *) Fix for SHA1 assembly problem with MASM: it produces
9072 warnings about corrupt line number information when assembling
9073 with debugging information. This is caused by the overlapping
9075 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
9078 NCONF_get_number() has no error checking at all. As a replacement,
9079 NCONF_get_number_e() is defined (_e for "error checking") and is
9080 promoted strongly. The old NCONF_get_number is kept around for
9081 binary backward compatibility.
9082 Make it possible for methods to load from something other than a BIO,
9083 by providing a function pointer that is given a name instead of a BIO.
9084 For example, this could be used to load configuration data from an
9088 *) Fix for non blocking accept BIOs. Added new I/O special reason
9089 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
9090 with non blocking I/O was not possible because no retry code was
9091 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
9095 *) Added the beginnings of Rijndael support.
9098 *) Fix for bug in DirectoryString mask setting. Add support for
9099 X509_NAME_print_ex() in 'req' and X509_print_ex() function
9100 to allow certificate printing to more controllable, additional
9101 'certopt' option to 'x509' to allow new printing options to be
9105 *) Clean old EAY MD5 hack from e_os.h.
9108 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
9110 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
9111 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
9112 [Joe Orton, Steve Henson]
9114 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
9116 *) Fix additional bug revealed by the NISCC test suite:
9118 Stop bug triggering large recursion when presented with
9119 certain ASN.1 tags (CVE-2003-0851)
9122 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
9124 *) Fix various bugs revealed by running the NISCC test suite:
9126 Stop out of bounds reads in the ASN1 code when presented with
9127 invalid tags (CVE-2003-0543 and CVE-2003-0544).
9129 If verify callback ignores invalid public key errors don't try to check
9130 certificate signature with the NULL public key.
9134 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
9135 if the server requested one: as stated in TLS 1.0 and SSL 3.0
9139 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
9140 extra data after the compression methods not only for TLS 1.0
9141 but also for SSL 3.0 (as required by the specification).
9142 [Bodo Moeller; problem pointed out by Matthias Loepfe]
9144 *) Change X509_certificate_type() to mark the key as exported/exportable
9145 when it's 512 *bits* long, not 512 bytes.
9148 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
9150 *) Countermeasure against the Klima-Pokorny-Rosa extension of
9151 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
9152 a protocol version number mismatch like a decryption error
9153 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
9156 *) Turn on RSA blinding by default in the default implementation
9157 to avoid a timing attack. Applications that don't want it can call
9158 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
9159 They would be ill-advised to do so in most cases.
9160 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
9162 *) Change RSA blinding code so that it works when the PRNG is not
9163 seeded (in this case, the secret RSA exponent is abused as
9164 an unpredictable seed -- if it is not unpredictable, there
9165 is no point in blinding anyway). Make RSA blinding thread-safe
9166 by remembering the creator's thread ID in rsa->blinding and
9167 having all other threads use local one-time blinding factors
9168 (this requires more computation than sharing rsa->blinding, but
9169 avoids excessive locking; and if an RSA object is not shared
9170 between threads, blinding will still be very fast).
9173 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
9175 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
9176 via timing by performing a MAC computation even if incorrect
9177 block cipher padding has been found. This is a countermeasure
9178 against active attacks where the attacker has to distinguish
9179 between bad padding and a MAC verification error. (CVE-2003-0078)
9181 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
9182 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
9183 Martin Vuagnoux (EPFL, Ilion)]
9185 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
9187 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
9188 memory from its contents. This is done with a counter that will
9189 place alternating values in each byte. This can be used to solve
9190 two issues: 1) the removal of calls to memset() by highly optimizing
9191 compilers, and 2) cleansing with other values than 0, since those can
9192 be read through on certain media, for example a swap space on disk.
9195 *) Bugfix: client side session caching did not work with external caching,
9196 because the session->cipher setting was not restored when reloading
9197 from the external cache. This problem was masked, when
9198 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9199 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9202 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9203 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9204 [Zeev Lieber <zeev-l@yahoo.com>]
9206 *) Undo an undocumented change introduced in 0.9.6e which caused
9207 repeated calls to OpenSSL_add_all_ciphers() and
9208 OpenSSL_add_all_digests() to be ignored, even after calling
9212 *) Change the default configuration reader to deal with last line not
9213 being properly terminated.
9216 *) Change X509_NAME_cmp() so it applies the special rules on handling
9217 DN values that are of type PrintableString, as well as RDNs of type
9218 emailAddress where the value has the type ia5String.
9219 [stefank@valicert.com via Richard Levitte]
9221 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9222 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9223 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9224 the bitwise-OR of the two for use by the majority of applications
9225 wanting this behaviour, and update the docs. The documented
9226 behaviour and actual behaviour were inconsistent and had been
9227 changing anyway, so this is more a bug-fix than a behavioural
9229 [Geoff Thorpe, diagnosed by Nadav Har'El]
9231 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9232 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9235 *) Fix initialization code race conditions in
9236 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9237 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9238 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9239 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9240 ssl2_get_cipher_by_char(),
9241 ssl3_get_cipher_by_char().
9242 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9244 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9245 the cached sessions are flushed, as the remove_cb() might use ex_data
9246 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9247 (see [openssl.org #212]).
9248 [Geoff Thorpe, Lutz Jaenicke]
9250 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9251 length, instead of the encoding length to d2i_ASN1_OBJECT.
9254 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9256 *) [In 0.9.6g-engine release:]
9257 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9258 [Lynn Gazis <lgazis@rainbow.com>]
9260 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9262 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9263 and get fix the header length calculation.
9264 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9265 Alon Kantor <alonk@checkpoint.com> (and others),
9268 *) Use proper error handling instead of 'assertions' in buffer
9269 overflow checks added in 0.9.6e. This prevents DoS (the
9270 assertions could call abort()).
9271 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9273 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9275 *) Add various sanity checks to asn1_get_length() to reject
9276 the ASN1 length bytes if they exceed sizeof(long), will appear
9277 negative or the content length exceeds the length of the
9279 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9281 *) Fix cipher selection routines: ciphers without encryption had no flags
9282 for the cipher strength set and where therefore not handled correctly
9283 by the selection routines (PR #130).
9286 *) Fix EVP_dsa_sha macro.
9290 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9291 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9292 that was added in OpenSSL 0.9.6d.
9294 As the countermeasure turned out to be incompatible with some
9295 broken SSL implementations, the new option is part of SSL_OP_ALL.
9296 SSL_OP_ALL is usually employed when compatibility with weird SSL
9297 implementations is desired (e.g. '-bugs' option to 's_client' and
9298 's_server'), so the new option is automatically set in many
9302 *) Changes in security patch:
9304 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9305 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9306 Air Force Materiel Command, USAF, under agreement number
9309 *) Add various sanity checks to asn1_get_length() to reject
9310 the ASN1 length bytes if they exceed sizeof(long), will appear
9311 negative or the content length exceeds the length of the
9312 supplied buffer. (CVE-2002-0659)
9313 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9315 *) Assertions for various potential buffer overflows, not known to
9317 [Ben Laurie (CHATS)]
9319 *) Various temporary buffers to hold ASCII versions of integers were
9320 too small for 64 bit platforms. (CVE-2002-0655)
9321 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9323 *) Remote buffer overflow in SSL3 protocol - an attacker could
9324 supply an oversized session ID to a client. (CVE-2002-0656)
9325 [Ben Laurie (CHATS)]
9327 *) Remote buffer overflow in SSL2 protocol - an attacker could
9328 supply an oversized client master key. (CVE-2002-0656)
9329 [Ben Laurie (CHATS)]
9331 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9333 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9334 encoded as NULL) with id-dsa-with-sha1.
9335 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9337 *) Check various X509_...() return values in apps/req.c.
9338 [Nils Larsch <nla@trustcenter.de>]
9340 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9341 an end-of-file condition would erroneously be flagged, when the CRLF
9342 was just at the end of a processed block. The bug was discovered when
9343 processing data through a buffering memory BIO handing the data to a
9344 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9345 <ptsekov@syntrex.com> and Nedelcho Stanev.
9348 *) Implement a countermeasure against a vulnerability recently found
9349 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9350 before application data chunks to avoid the use of known IVs
9351 with data potentially chosen by the attacker.
9354 *) Fix length checks in ssl3_get_client_hello().
9357 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9358 to prevent ssl3_read_internal() from incorrectly assuming that
9359 ssl3_read_bytes() found application data while handshake
9360 processing was enabled when in fact s->s3->in_read_app_data was
9361 merely automatically cleared during the initial handshake.
9362 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9364 *) Fix object definitions for Private and Enterprise: they were not
9365 recognized in their shortname (=lowercase) representation. Extend
9366 obj_dat.pl to issue an error when using undefined keywords instead
9367 of silently ignoring the problem (Svenning Sorensen
9368 <sss@sss.dnsalias.net>).
9371 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9372 generators, i.e. generators other than 2 and 5. (Previously, the
9373 code did not properly initialise the 'add' and 'rem' values to
9374 BN_generate_prime().)
9376 In the new general case, we do not insist that 'generator' is
9377 actually a primitive root: This requirement is rather pointless;
9378 a generator of the order-q subgroup is just as good, if not
9382 *) Map new X509 verification errors to alerts. Discovered and submitted by
9383 Tom Wu <tom@arcot.com>.
9386 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9387 returning non-zero before the data has been completely received
9388 when using non-blocking I/O.
9389 [Bodo Moeller; problem pointed out by John Hughes]
9391 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9392 [Ben Laurie, Lutz Jaenicke]
9394 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9395 Yoram Zahavi <YoramZ@gilian.com>).
9398 *) Add information about CygWin 1.3 and on, and preserve proper
9399 configuration for the versions before that.
9400 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9402 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9403 check whether we deal with a copy of a session and do not delete from
9404 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9405 <izhar@checkpoint.com>.
9408 *) Do not store session data into the internal session cache, if it
9409 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9410 flag is set). Proposed by Aslam <aslam@funk.com>.
9413 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9417 *) [In 0.9.6d-engine release:]
9418 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9419 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9421 *) Add the configuration target linux-s390x.
9422 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9424 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9425 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9426 variable as an indication that a ClientHello message has been
9427 received. As the flag value will be lost between multiple
9428 invocations of ssl3_accept when using non-blocking I/O, the
9429 function may not be aware that a handshake has actually taken
9430 place, thus preventing a new session from being added to the
9433 To avoid this problem, we now set s->new_session to 2 instead of
9434 using a local variable.
9435 [Lutz Jaenicke, Bodo Moeller]
9437 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9438 if the SSL_R_LENGTH_MISMATCH error is detected.
9439 [Geoff Thorpe, Bodo Moeller]
9441 *) New 'shared_ldflag' column in Configure platform table.
9444 *) Fix EVP_CIPHER_mode macro.
9445 ["Dan S. Camper" <dan@bti.net>]
9447 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9448 type, we must throw them away by setting rr->length to 0.
9449 [D P Chang <dpc@qualys.com>]
9451 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9453 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9454 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9455 worked incorrectly for those cases where range = 10..._2 and
9456 3*range is two bits longer than range.)
9459 *) Only add signing time to PKCS7 structures if it is not already
9463 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9464 OBJ_ld_ce should be OBJ_id_ce.
9465 Also some ip-pda OIDs in crypto/objects/objects.txt were
9466 incorrect (cf. RFC 3039).
9467 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9469 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9470 returns early because it has nothing to do.
9471 [Andy Schneider <andy.schneider@bjss.co.uk>]
9473 *) [In 0.9.6c-engine release:]
9474 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9475 [Andy Schneider <andy.schneider@bjss.co.uk>]
9477 *) [In 0.9.6c-engine release:]
9478 Add support for Cryptographic Appliance's keyserver technology.
9479 (Use engine 'keyclient')
9480 [Cryptographic Appliances and Geoff Thorpe]
9482 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9483 is called via tools/c89.sh because arguments have to be
9484 rearranged (all '-L' options must appear before the first object
9486 [Richard Shapiro <rshapiro@abinitio.com>]
9488 *) [In 0.9.6c-engine release:]
9489 Add support for Broadcom crypto accelerator cards, backported
9491 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9493 *) [In 0.9.6c-engine release:]
9494 Add support for SureWare crypto accelerator cards from
9495 Baltimore Technologies. (Use engine 'sureware')
9496 [Baltimore Technologies and Mark Cox]
9498 *) [In 0.9.6c-engine release:]
9499 Add support for crypto accelerator cards from Accelerated
9500 Encryption Processing, www.aep.ie. (Use engine 'aep')
9501 [AEP Inc. and Mark Cox]
9503 *) Add a configuration entry for gcc on UnixWare.
9504 [Gary Benson <gbenson@redhat.com>]
9506 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9507 messages are stored in a single piece (fixed-length part and
9508 variable-length part combined) and fix various bugs found on the way.
9511 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9512 instead. BIO_gethostbyname() does not know what timeouts are
9513 appropriate, so entries would stay in cache even when they have
9515 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9517 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9518 faced with a pathologically small ClientHello fragment that does
9519 not contain client_version: Instead of aborting with an error,
9520 simply choose the highest available protocol version (i.e.,
9521 TLS 1.0 unless it is disabled). In practice, ClientHello
9522 messages are never sent like this, but this change gives us
9523 strictly correct behaviour at least for TLS.
9526 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9527 never resets s->method to s->ctx->method when called from within
9528 one of the SSL handshake functions.
9529 [Bodo Moeller; problem pointed out by Niko Baric]
9531 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9532 (sent using the client's version number) if client_version is
9533 smaller than the protocol version in use. Also change
9534 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9535 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9536 the client will at least see that alert.
9539 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9543 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9544 client receives HelloRequest while in a handshake.
9545 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9547 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9548 should end in 'break', not 'goto end' which circumvents various
9549 cleanups done in state SSL_ST_OK. But session related stuff
9550 must be disabled for SSL_ST_OK in the case that we just sent a
9553 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9554 before just sending a HelloRequest.
9555 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9557 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9558 reveal whether illegal block cipher padding was found or a MAC
9559 verification error occurred. (Neither SSLerr() codes nor alerts
9560 are directly visible to potential attackers, but the information
9561 may leak via logfiles.)
9563 Similar changes are not required for the SSL 2.0 implementation
9564 because the number of padding bytes is sent in clear for SSL 2.0,
9565 and the extra bytes are just ignored. However ssl/s2_pkt.c
9566 failed to verify that the purported number of padding bytes is in
9570 *) Add OpenUNIX-8 support including shared libraries
9571 (Boyd Lynn Gerber <gerberb@zenez.com>).
9574 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9575 'wristwatch attack' using huge encoding parameters (cf.
9576 James H. Manger's CRYPTO 2001 paper). Note that the
9577 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9578 encoding parameters and hence was not vulnerable.
9581 *) BN_sqr() bug fix.
9582 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9584 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9585 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9586 followed by modular reduction.
9587 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9589 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9590 equivalent based on BN_pseudo_rand() instead of BN_rand().
9593 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9594 This function was broken, as the check for a new client hello message
9595 to handle SGC did not allow these large messages.
9596 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9599 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9602 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9603 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9606 *) Rework the configuration and shared library support for Tru64 Unix.
9607 The configuration part makes use of modern compiler features and
9608 still retains old compiler behavior for those that run older versions
9609 of the OS. The shared library support part includes a variant that
9610 uses the RPATH feature, and is available through the special
9611 configuration target "alpha-cc-rpath", which will never be selected
9613 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9615 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9616 with the same message size as in ssl3_get_certificate_request().
9617 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9618 messages might inadvertently be reject as too long.
9619 [Petr Lampa <lampa@fee.vutbr.cz>]
9621 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9624 *) Modified SSL library such that the verify_callback that has been set
9625 specifically for an SSL object with SSL_set_verify() is actually being
9626 used. Before the change, a verify_callback set with this function was
9627 ignored and the verify_callback() set in the SSL_CTX at the time of
9628 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9629 to allow the necessary settings.
9632 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9633 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9634 done automatically (in contradiction to the requirements of the C
9635 standard). This made problems when used from OpenSSH.
9638 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9639 dh->length and always used
9641 BN_rand_range(priv_key, dh->p).
9643 BN_rand_range() is not necessary for Diffie-Hellman, and this
9644 specific range makes Diffie-Hellman unnecessarily inefficient if
9645 dh->length (recommended exponent length) is much smaller than the
9646 length of dh->p. We could use BN_rand_range() if the order of
9647 the subgroup was stored in the DH structure, but we only have
9652 BN_rand(priv_key, l, ...)
9654 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9660 RSA_eay_public_encrypt
9661 RSA_eay_private_decrypt
9662 RSA_eay_private_encrypt (signing)
9663 RSA_eay_public_decrypt (signature verification)
9665 (default implementations for RSA_public_encrypt,
9666 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9667 always reject numbers >= n.
9670 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9671 to synchronize access to 'locking_thread'. This is necessary on
9672 systems where access to 'locking_thread' (an 'unsigned long'
9673 variable) is not atomic.
9676 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9677 *before* setting the 'crypto_lock_rand' flag. The previous code had
9678 a race condition if 0 is a valid thread ID.
9679 [Travis Vitek <vitek@roguewave.com>]
9681 *) Add support for shared libraries under Irix.
9682 [Albert Chin-A-Young <china@thewrittenword.com>]
9684 *) Add configuration option to build on Linux on both big-endian and
9686 [Ralf Baechle <ralf@uni-koblenz.de>]
9688 *) Add the possibility to create shared libraries on HP-UX.
9691 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9693 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9694 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9695 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9696 PRNG state recovery was possible based on the output of
9697 one PRNG request appropriately sized to gain knowledge on
9698 'md' followed by enough consecutive 1-byte PRNG requests
9699 to traverse all of 'state'.
9701 1. When updating 'md_local' (the current thread's copy of 'md')
9702 during PRNG output generation, hash all of the previous
9703 'md_local' value, not just the half used for PRNG output.
9705 2. Make the number of bytes from 'state' included into the hash
9706 independent from the number of PRNG bytes requested.
9708 The first measure alone would be sufficient to avoid
9709 Markku-Juhani's attack. (Actually it had never occurred
9710 to me that the half of 'md_local' used for chaining was the
9711 half from which PRNG output bytes were taken -- I had always
9712 assumed that the secret half would be used.) The second
9713 measure makes sure that additional data from 'state' is never
9714 mixed into 'md_local' in small portions; this heuristically
9715 further strengthens the PRNG.
9718 *) Fix crypto/bn/asm/mips3.s.
9721 *) When only the key is given to "enc", the IV is undefined. Print out
9722 an error message in this case.
9725 *) Handle special case when X509_NAME is empty in X509 printing routines.
9728 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9729 positive and less than q.
9732 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9733 used: it isn't thread safe and the add_lock_callback should handle
9735 [Paul Rose <Paul.Rose@bridge.com>]
9737 *) Verify that incoming data obeys the block size in
9738 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9742 [Ulf Möller, Bodo Möller]
9744 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9745 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9746 when fixing the server behaviour for backwards-compatible 'client
9747 hello' messages. (Note that the attack is impractical against
9748 SSL 3.0 and TLS 1.0 anyway because length and version checking
9749 means that the probability of guessing a valid ciphertext is
9750 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9753 Before 0.9.5, the countermeasure (hide the error by generating a
9754 random 'decryption result') did not work properly because
9755 ERR_clear_error() was missing, meaning that SSL_get_error() would
9756 detect the supposedly ignored error.
9758 Both problems are now fixed.
9761 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9762 (previously it was 1024).
9765 *) Fix for compatibility mode trust settings: ignore trust settings
9766 unless some valid trust or reject settings are present.
9769 *) Fix for blowfish EVP: its a variable length cipher.
9772 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9773 parameters in DSA public key structures and return an error in the
9774 DSA routines if parameters are absent.
9777 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9778 in the current directory if neither $RANDFILE nor $HOME was set.
9779 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9780 caused some confusion to Windows users who haven't defined $HOME.
9781 Thus RAND_file_name() is changed again: e_os.h can define a
9782 DEFAULT_HOME, which will be used if $HOME is not set.
9783 For Windows, we use "C:"; on other platforms, we still require
9784 environment variables.
9786 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9787 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9788 having multiple threads call RAND_poll() concurrently.
9791 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9792 combination of a flag and a thread ID variable.
9793 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9794 flag), *other* threads can enter ssleay_add_bytes without obeying
9795 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9796 that they do not hold after the first thread unsets add_do_not_lock).
9799 *) Change bctest again: '-x' expressions are not available in all
9803 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9805 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9806 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9808 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9809 the default extension for executables, if any. Also, make the perl
9810 scripts that use symlink() to test if it really exists and use "cp"
9811 if it doesn't. All this made OpenSSL compilable and installable in
9815 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9816 If SEQUENCE is length is indefinite just set c->slen to the total
9817 amount of data available.
9818 [Steve Henson, reported by shige@FreeBSD.org]
9819 [This change does not apply to 0.9.7.]
9821 *) Change bctest to avoid here-documents inside command substitution
9822 (workaround for FreeBSD /bin/sh bug).
9823 For compatibility with Ultrix, avoid shell functions (introduced
9824 in the bctest version that searches along $PATH).
9827 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9828 with des_encrypt() defined on some operating systems, like Solaris
9832 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9833 On the Importance of Eliminating Errors in Cryptographic
9834 Computations, J. Cryptology 14 (2001) 2, 101-119,
9835 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9838 *) MIPS assembler BIGNUM division bug fix.
9841 *) Disabled incorrect Alpha assembler code.
9844 *) Fix PKCS#7 decode routines so they correctly update the length
9845 after reading an EOC for the EXPLICIT tag.
9847 [This change does not apply to 0.9.7.]
9849 *) Fix bug in PKCS#12 key generation routines. This was triggered
9850 if a 3DES key was generated with a 0 initial byte. Include
9851 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9852 (but broken) behaviour.
9855 *) Enhance bctest to search for a working bc along $PATH and print
9857 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9859 *) Fix memory leaks in err.c: free err_data string if necessary;
9860 don't write to the wrong index in ERR_set_error_data.
9863 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9867 *) Replace rdtsc with _emit statements for VC++ version 5.
9868 [Jeremy Cooper <jeremy@baymoo.org>]
9870 *) Make it possible to reuse SSLv2 sessions.
9873 *) In copy_email() check for >= 0 as a return value for
9874 X509_NAME_get_index_by_NID() since 0 is a valid index.
9875 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9877 *) Avoid coredump with unsupported or invalid public keys by checking if
9878 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9879 PKCS7_verify() fails with non detached data.
9882 *) Don't use getenv in library functions when run as setuid/setgid.
9883 New function OPENSSL_issetugid().
9886 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9887 due to incorrect handling of multi-threading:
9889 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9891 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9893 3. Count how many times MemCheck_off() has been called so that
9894 nested use can be treated correctly. This also avoids
9895 inband-signalling in the previous code (which relied on the
9896 assumption that thread ID 0 is impossible).
9899 *) Add "-rand" option also to s_client and s_server.
9902 *) Fix CPU detection on Irix 6.x.
9903 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9904 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9906 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9909 [This change does not apply to 0.9.7.]
9911 *) Use the cached encoding of an X509_NAME structure rather than
9912 copying it. This is apparently the reason for the libsafe "errors"
9913 but the code is actually correct.
9916 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9917 Bleichenbacher's DSA attack.
9918 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9919 to be set and top=0 forces the highest bit to be set; top=-1 is new
9920 and leaves the highest bit random.
9921 [Ulf Moeller, Bodo Moeller]
9923 *) In the NCONF_...-based implementations for CONF_... queries
9924 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9925 a temporary CONF structure with the data component set to NULL
9926 (which gives segmentation faults in lh_retrieve).
9927 Instead, use NULL for the CONF pointer in CONF_get_string and
9928 CONF_get_number (which may use environment variables) and directly
9929 return NULL from CONF_get_section.
9932 *) Fix potential buffer overrun for EBCDIC.
9935 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9936 keyUsage if basicConstraints absent for a CA.
9939 *) Make SMIME_write_PKCS7() write mail header values with a format that
9940 is more generally accepted (no spaces before the semicolon), since
9941 some programs can't parse those values properly otherwise. Also make
9942 sure BIO's that break lines after each write do not create invalid
9946 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9947 macros previously used would not encode an empty SEQUENCE OF
9948 and break the signature.
9950 [This change does not apply to 0.9.7.]
9952 *) Zero the premaster secret after deriving the master secret in
9956 *) Add some EVP_add_digest_alias registrations (as found in
9957 OpenSSL_add_all_digests()) to SSL_library_init()
9958 aka OpenSSL_add_ssl_algorithms(). This provides improved
9959 compatibility with peers using X.509 certificates
9960 with unconventional AlgorithmIdentifier OIDs.
9963 *) Fix for Irix with NO_ASM.
9964 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9966 *) ./config script fixes.
9967 [Ulf Moeller, Richard Levitte]
9969 *) Fix 'openssl passwd -1'.
9972 *) Change PKCS12_key_gen_asc() so it can cope with non null
9973 terminated strings whose length is passed in the passlen
9974 parameter, for example from PEM callbacks. This was done
9975 by adding an extra length parameter to asc2uni().
9976 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9978 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9979 call failed, free the DSA structure.
9982 *) Fix to uni2asc() to cope with zero length Unicode strings.
9983 These are present in some PKCS#12 files.
9986 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9987 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9988 when writing a 32767 byte record.
9989 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9991 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9992 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9994 (RSA objects have a reference count access to which is protected
9995 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9996 so they are meant to be shared between threads.)
9997 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9998 "Reddie, Steven" <Steven.Reddie@ca.com>]
10000 *) Fix a deadlock in CRYPTO_mem_leaks().
10003 *) Use better test patterns in bntest.
10006 *) rand_win.c fix for Borland C.
10009 *) BN_rshift bugfix for n == 0.
10012 *) Add a 'bctest' script that checks for some known 'bc' bugs
10013 so that 'make test' does not abort just because 'bc' is broken.
10016 *) Store verify_result within SSL_SESSION also for client side to
10017 avoid potential security hole. (Re-used sessions on the client side
10018 always resulted in verify_result==X509_V_OK, not using the original
10019 result of the server certificate verification.)
10022 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
10023 SSL3_RT_APPLICATION_DATA, return 0.
10024 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
10028 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
10029 releases, have been re-implemented by renaming the previous
10030 implementations of ssl2_read and ssl3_read to ssl2_read_internal
10031 and ssl3_read_internal, respectively, and adding 'peek' parameters
10032 to them. The new ssl[23]_{read,peek} functions are calls to
10033 ssl[23]_read_internal with the 'peek' flag set appropriately.
10034 A 'peek' parameter has also been added to ssl3_read_bytes, which
10035 does the actual work for ssl3_read_internal.
10038 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
10039 the method-specific "init()" handler. Also clean up ex_data after
10040 calling the method-specific "finish()" handler. Previously, this was
10041 happening the other way round.
10044 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
10045 The previous value, 12, was not always sufficient for BN_mod_exp().
10048 *) Make sure that shared libraries get the internal name engine with
10049 the full version number and not just 0. This should mark the
10050 shared libraries as not backward compatible. Of course, this should
10051 be changed again when we can guarantee backward binary compatibility.
10054 *) Fix typo in get_cert_by_subject() in by_dir.c
10055 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
10057 *) Rework the system to generate shared libraries:
10059 - Make note of the expected extension for the shared libraries and
10060 if there is a need for symbolic links from for example libcrypto.so.0
10061 to libcrypto.so.0.9.7. There is extended info in Configure for
10064 - Make as few rebuilds of the shared libraries as possible.
10066 - Still avoid linking the OpenSSL programs with the shared libraries.
10068 - When installing, install the shared libraries separately from the
10072 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
10074 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
10075 and not in SSL_clear because the latter is also used by the
10076 accept/connect functions; previously, the settings made by
10077 SSL_set_read_ahead would be lost during the handshake.
10078 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
10080 *) Correct util/mkdef.pl to be selective about disabled algorithms.
10081 Previously, it would create entries for disabled algorithms no
10085 *) Added several new manual pages for SSL_* function.
10088 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
10090 *) In ssl23_get_client_hello, generate an error message when faced
10091 with an initial SSL 3.0/TLS record that is too small to contain the
10092 first two bytes of the ClientHello message, i.e. client_version.
10093 (Note that this is a pathologic case that probably has never happened
10094 in real life.) The previous approach was to use the version number
10095 from the record header as a substitute; but our protocol choice
10096 should not depend on that one because it is not authenticated
10097 by the Finished messages.
10100 *) More robust randomness gathering functions for Windows.
10101 [Jeffrey Altman <jaltman@columbia.edu>]
10103 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
10104 not set then we don't setup the error code for issuer check errors
10105 to avoid possibly overwriting other errors which the callback does
10106 handle. If an application does set the flag then we assume it knows
10107 what it is doing and can handle the new informational codes
10111 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
10112 a general "ANY" type, as such it should be able to decode anything
10113 including tagged types. However it didn't check the class so it would
10114 wrongly interpret tagged types in the same way as their universal
10115 counterpart and unknown types were just rejected. Changed so that the
10116 tagged and unknown types are handled in the same way as a SEQUENCE:
10117 that is the encoding is stored intact. There is also a new type
10118 "V_ASN1_OTHER" which is used when the class is not universal, in this
10119 case we have no idea what the actual type is so we just lump them all
10123 *) On VMS, stdout may very well lead to a file that is written to
10124 in a record-oriented fashion. That means that every write() will
10125 write a separate record, which will be read separately by the
10126 programs trying to read from it. This can be very confusing.
10128 The solution is to put a BIO filter in the way that will buffer
10129 text until a linefeed is reached, and then write everything a
10130 line at a time, so every record written will be an actual line,
10131 not chunks of lines and not (usually doesn't happen, but I've
10132 seen it once) several lines in one record. BIO_f_linebuffer() is
10135 Currently, it's a VMS-only method, because that's where it has
10136 been tested well enough.
10139 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
10140 it can return incorrect results.
10141 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
10142 but it was in 0.9.6-beta[12].)
10145 *) Disable the check for content being present when verifying detached
10146 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
10147 include zero length content when signing messages.
10150 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
10151 BIO_ctrl (for BIO pairs).
10154 *) Add DSO method for VMS.
10157 *) Bug fix: Montgomery multiplication could produce results with the
10161 *) Add RPM specification openssl.spec and modify it to build three
10162 packages. The default package contains applications, application
10163 documentation and run-time libraries. The devel package contains
10164 include files, static libraries and function documentation. The
10165 doc package contains the contents of the doc directory. The original
10166 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
10169 *) Add a large number of documentation files for many SSL routines.
10170 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10172 *) Add a configuration entry for Sony News 4.
10173 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
10175 *) Don't set the two most significant bits to one when generating a
10176 random number < q in the DSA library.
10179 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
10180 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
10181 the underlying transport is blocking) if a handshake took place.
10182 (The default behaviour is needed by applications such as s_client
10183 and s_server that use select() to determine when to use SSL_read;
10184 but for applications that know in advance when to expect data, it
10185 just makes things more complicated.)
10188 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
10192 *) Add a few more EBCDIC conditionals that make `req' and `x509'
10193 work better on such systems.
10194 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10196 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10197 Update PKCS12_parse() so it copies the friendlyName and the
10198 keyid to the certificates aux info.
10201 *) Fix bug in PKCS7_verify() which caused an infinite loop
10202 if there was more than one signature.
10203 [Sven Uszpelkat <su@celocom.de>]
10205 *) Major change in util/mkdef.pl to include extra information
10206 about each symbol, as well as presenting variables as well
10207 as functions. This change means that there's n more need
10208 to rebuild the .num files when some algorithms are excluded.
10211 *) Allow the verify time to be set by an application,
10212 rather than always using the current time.
10215 *) Phase 2 verify code reorganisation. The certificate
10216 verify code now looks up an issuer certificate by a
10217 number of criteria: subject name, authority key id
10218 and key usage. It also verifies self signed certificates
10219 by the same criteria. The main comparison function is
10220 X509_check_issued() which performs these checks.
10222 Lot of changes were necessary in order to support this
10223 without completely rewriting the lookup code.
10225 Authority and subject key identifier are now cached.
10227 The LHASH 'certs' is X509_STORE has now been replaced
10228 by a STACK_OF(X509_OBJECT). This is mainly because an
10229 LHASH can't store or retrieve multiple objects with
10230 the same hash value.
10232 As a result various functions (which were all internal
10233 use only) have changed to handle the new X509_STORE
10234 structure. This will break anything that messed round
10235 with X509_STORE internally.
10237 The functions X509_STORE_add_cert() now checks for an
10238 exact match, rather than just subject name.
10240 The X509_STORE API doesn't directly support the retrieval
10241 of multiple certificates matching a given criteria, however
10242 this can be worked round by performing a lookup first
10243 (which will fill the cache with candidate certificates)
10244 and then examining the cache for matches. This is probably
10245 the best we can do without throwing out X509_LOOKUP
10246 entirely (maybe later...).
10248 The X509_VERIFY_CTX structure has been enhanced considerably.
10250 All certificate lookup operations now go via a get_issuer()
10251 callback. Although this currently uses an X509_STORE it
10252 can be replaced by custom lookups. This is a simple way
10253 to bypass the X509_STORE hackery necessary to make this
10254 work and makes it possible to use more efficient techniques
10255 in future. A very simple version which uses a simple
10256 STACK for its trusted certificate store is also provided
10257 using X509_STORE_CTX_trusted_stack().
10259 The verify_cb() and verify() callbacks now have equivalents
10260 in the X509_STORE_CTX structure.
10262 X509_STORE_CTX also has a 'flags' field which can be used
10263 to customise the verify behaviour.
10266 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10267 excludes S/MIME capabilities.
10270 *) When a certificate request is read in keep a copy of the
10271 original encoding of the signed data and use it when outputting
10272 again. Signatures then use the original encoding rather than
10273 a decoded, encoded version which may cause problems if the
10274 request is improperly encoded.
10277 *) For consistency with other BIO_puts implementations, call
10278 buffer_write(b, ...) directly in buffer_puts instead of calling
10281 In BIO_puts, increment b->num_write as in BIO_write.
10282 [Peter.Sylvester@EdelWeb.fr]
10284 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10285 BN_zero, we may not return a BIGNUM with an array consisting of
10286 words set to zero.)
10289 *) Avoid calling abort() from within the library when problems are
10290 detected, except if preprocessor symbols have been defined
10291 (such as REF_CHECK, BN_DEBUG etc.).
10294 *) New openssl application 'rsautl'. This utility can be
10295 used for low level RSA operations. DER public key
10296 BIO/fp routines also added.
10299 *) New Configure entry and patches for compiling on QNX 4.
10300 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10302 *) A demo state-machine implementation was sponsored by
10303 Nuron (http://www.nuron.com/) and is now available in
10304 demos/state_machine.
10307 *) New options added to the 'dgst' utility for signature
10308 generation and verification.
10311 *) Unrecognized PKCS#7 content types are now handled via a
10312 catch all ASN1_TYPE structure. This allows unsupported
10313 types to be stored as a "blob" and an application can
10314 encode and decode it manually.
10317 *) Fix various signed/unsigned issues to make a_strex.c
10318 compile under VC++.
10319 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10321 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10322 length if passed a buffer. ASN1_INTEGER_to_BN failed
10323 if passed a NULL BN and its argument was negative.
10324 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10326 *) Modification to PKCS#7 encoding routines to output definite
10327 length encoding. Since currently the whole structures are in
10328 memory there's not real point in using indefinite length
10329 constructed encoding. However if OpenSSL is compiled with
10330 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10333 *) Added BIO_vprintf() and BIO_vsnprintf().
10336 *) Added more prefixes to parse for in the strings written
10337 through a logging bio, to cover all the levels that are available
10338 through syslog. The prefixes are now:
10340 PANIC, EMERG, EMR => LOG_EMERG
10341 ALERT, ALR => LOG_ALERT
10342 CRIT, CRI => LOG_CRIT
10343 ERROR, ERR => LOG_ERR
10344 WARNING, WARN, WAR => LOG_WARNING
10345 NOTICE, NOTE, NOT => LOG_NOTICE
10346 INFO, INF => LOG_INFO
10347 DEBUG, DBG => LOG_DEBUG
10349 and as before, if none of those prefixes are present at the
10350 beginning of the string, LOG_ERR is chosen.
10352 On Win32, the LOG_* levels are mapped according to this:
10354 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10355 LOG_WARNING => EVENTLOG_WARNING_TYPE
10356 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10360 *) Made it possible to reconfigure with just the configuration
10361 argument "reconf" or "reconfigure". The command line arguments
10362 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10363 and are retrieved from there when reconfiguring.
10366 *) MD4 implemented.
10367 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10369 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10372 *) The obj_dat.pl script was messing up the sorting of object
10373 names. The reason was that it compared the quoted version
10374 of strings as a result "OCSP" > "OCSP Signing" because
10375 " > SPACE. Changed script to store unquoted versions of
10376 names and add quotes on output. It was also omitting some
10377 names from the lookup table if they were given a default
10378 value (that is if SN is missing it is given the same
10379 value as LN and vice versa), these are now added on the
10380 grounds that if an object has a name we should be able to
10381 look it up. Finally added warning output when duplicate
10382 short or long names are found.
10385 *) Changes needed for Tandem NSK.
10386 [Scott Uroff <scott@xypro.com>]
10388 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10389 RSA_padding_check_SSLv23(), special padding was never detected
10390 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10391 version rollback attacks was not effective.
10393 In s23_clnt.c, don't use special rollback-attack detection padding
10394 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10395 client; similarly, in s23_srvr.c, don't do the rollback check if
10396 SSL 2.0 is the only protocol enabled in the server.
10399 *) Make it possible to get hexdumps of unprintable data with 'openssl
10400 asn1parse'. By implication, the functions ASN1_parse_dump() and
10401 BIO_dump_indent() are added.
10404 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10405 these print out strings and name structures based on various
10406 flags including RFC2253 support and proper handling of
10407 multibyte characters. Added options to the 'x509' utility
10408 to allow the various flags to be set.
10411 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10412 Also change the functions X509_cmp_current_time() and
10413 X509_gmtime_adj() work with an ASN1_TIME structure,
10414 this will enable certificates using GeneralizedTime in validity
10415 dates to be checked.
10418 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10419 negative public key encodings) on by default,
10420 NO_NEG_PUBKEY_BUG can be set to disable it.
10423 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10424 content octets. An i2c_ASN1_OBJECT is unnecessary because
10425 the encoding can be trivially obtained from the structure.
10428 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10429 not read locks (CRYPTO_r_[un]lock).
10432 *) A first attempt at creating official support for shared
10433 libraries through configuration. I've kept it so the
10434 default is static libraries only, and the OpenSSL programs
10435 are always statically linked for now, but there are
10436 preparations for dynamic linking in place.
10437 This has been tested on Linux and Tru64.
10440 *) Randomness polling function for Win9x, as described in:
10441 Peter Gutmann, Software Generation of Practically Strong
10445 *) Fix so PRNG is seeded in req if using an already existing
10449 *) New options to smime application. -inform and -outform
10450 allow alternative formats for the S/MIME message including
10451 PEM and DER. The -content option allows the content to be
10452 specified separately. This should allow things like Netscape
10453 form signing output easier to verify.
10456 *) Fix the ASN1 encoding of tags using the 'long form'.
10459 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10460 STRING types. These convert content octets to and from the
10461 underlying type. The actual tag and length octets are
10462 already assumed to have been read in and checked. These
10463 are needed because all other string types have virtually
10464 identical handling apart from the tag. By having versions
10465 of the ASN1 functions that just operate on content octets
10466 IMPLICIT tagging can be handled properly. It also allows
10467 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10468 and ASN1_INTEGER are identical apart from the tag.
10471 *) Change the handling of OID objects as follows:
10473 - New object identifiers are inserted in objects.txt, following
10474 the syntax given in objects.README.
10475 - objects.pl is used to process obj_mac.num and create a new
10477 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10480 This is currently kind of a hack, and the perl code in objects.pl
10481 isn't very elegant, but it works as I intended. The simplest way
10482 to check that it worked correctly is to look in obj_dat.h and
10483 check the array nid_objs and make sure the objects haven't moved
10484 around (this is important!). Additions are OK, as well as
10485 consistent name changes.
10488 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10491 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10492 The given file adds to whatever has already been seeded into the
10493 random pool through the RANDFILE configuration file option or
10494 environment variable, or the default random state file.
10497 *) mkstack.pl now sorts each macro group into lexical order.
10498 Previously the output order depended on the order the files
10499 appeared in the directory, resulting in needless rewriting
10503 *) Patches to make OpenSSL compile under Win32 again. Mostly
10504 work arounds for the VC++ problem that it treats func() as
10505 func(void). Also stripped out the parts of mkdef.pl that
10506 added extra typesafe functions: these no longer exist.
10509 *) Reorganisation of the stack code. The macros are now all
10510 collected in safestack.h . Each macro is defined in terms of
10511 a "stack macro" of the form SKM_<name>(type, a, b). The
10512 DEBUG_SAFESTACK is now handled in terms of function casts,
10513 this has the advantage of retaining type safety without the
10514 use of additional functions. If DEBUG_SAFESTACK is not defined
10515 then the non typesafe macros are used instead. Also modified the
10516 mkstack.pl script to handle the new form. Needs testing to see
10517 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10518 the default if no major problems. Similar behaviour for ASN1_SET_OF
10519 and PKCS12_STACK_OF.
10522 *) When some versions of IIS use the 'NET' form of private key the
10523 key derivation algorithm is different. Normally MD5(password) is
10524 used as a 128 bit RC4 key. In the modified case
10525 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10526 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10527 as the old Netscape_RSA functions except they have an additional
10528 'sgckey' parameter which uses the modified algorithm. Also added
10529 an -sgckey command line option to the rsa utility. Thanks to
10530 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10531 algorithm to openssl-dev.
10534 *) The evp_local.h macros were using 'c.##kname' which resulted in
10535 invalid expansion on some systems (SCO 5.0.5 for example).
10536 Corrected to 'c.kname'.
10537 [Phillip Porch <root@theporch.com>]
10539 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10540 a STACK of email addresses from a certificate or request, these look
10541 in the subject name and the subject alternative name extensions and
10542 omit any duplicate addresses.
10545 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10546 This makes DSA verification about 2 % faster.
10549 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10550 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10551 plus overhead for 1024 bit moduli).
10552 This makes exponentiations about 0.5 % faster for 1024 bit
10553 exponents (as measured by "openssl speed rsa2048").
10556 *) Rename memory handling macros to avoid conflicts with other
10558 Malloc => OPENSSL_malloc
10559 Malloc_locked => OPENSSL_malloc_locked
10560 Realloc => OPENSSL_realloc
10561 Free => OPENSSL_free
10564 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10565 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10568 *) CygWin32 support.
10569 [John Jarvie <jjarvie@newsguy.com>]
10571 *) The type-safe stack code has been rejigged. It is now only compiled
10572 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10573 by default all type-specific stack functions are "#define"d back to
10574 standard stack functions. This results in more streamlined output
10575 but retains the type-safety checking possibilities of the original
10579 *) The STACK code has been cleaned up, and certain type declarations
10580 that didn't make a lot of sense have been brought in line. This has
10581 also involved a cleanup of sorts in safestack.h to more correctly
10582 map type-safe stack functions onto their plain stack counterparts.
10583 This work has also resulted in a variety of "const"ifications of
10584 lots of the code, especially "_cmp" operations which should normally
10585 be prototyped with "const" parameters anyway.
10588 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10589 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10590 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10591 where all of 'md' is used each time the PRNG is used, but 'state'
10592 is used only indexed by a cyclic counter. As entropy may not be
10593 well distributed from the beginning, 'md' is important as a
10594 chaining variable. However, the output function chains only half
10595 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10596 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10597 in all of 'state' being rewritten, with the new values depending
10598 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10601 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10602 the handshake is continued after ssl_verify_cert_chain();
10603 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10604 can lead to 'unexplainable' connection aborts later.
10605 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10607 *) Major EVP API cipher revision.
10608 Add hooks for extra EVP features. This allows various cipher
10609 parameters to be set in the EVP interface. Support added for variable
10610 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10611 setting of RC2 and RC5 parameters.
10613 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10616 Remove lots of duplicated code from the EVP library. For example *every*
10617 cipher init() function handles the 'iv' in the same way according to the
10618 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10619 for CFB and OFB modes they zero ctx->num.
10621 New functionality allows removal of S/MIME code RC2 hack.
10623 Most of the routines have the same form and so can be declared in terms
10626 By shifting this to the top level EVP_CipherInit() it can be removed from
10627 all individual ciphers. If the cipher wants to handle IVs or keys
10628 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10631 Change lots of functions like EVP_EncryptUpdate() to now return a
10632 value: although software versions of the algorithms cannot fail
10633 any installed hardware versions can.
10636 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10637 this option is set, tolerate broken clients that send the negotiated
10638 protocol version number instead of the requested protocol version
10642 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10643 i.e. non-zero for export ciphersuites, zero otherwise.
10644 Previous versions had this flag inverted, inconsistent with
10645 rsa_tmp_cb (..._TMP_RSA_CB).
10646 [Bodo Moeller; problem reported by Amit Chopra]
10648 *) Add missing DSA library text string. Work around for some IIS
10649 key files with invalid SEQUENCE encoding.
10652 *) Add a document (doc/standards.txt) that list all kinds of standards
10653 and so on that are implemented in OpenSSL.
10656 *) Enhance c_rehash script. Old version would mishandle certificates
10657 with the same subject name hash and wouldn't handle CRLs at all.
10658 Added -fingerprint option to crl utility, to support new c_rehash
10662 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10665 *) Fix for SSL server purpose checking. Server checking was
10666 rejecting certificates which had extended key usage present
10667 but no ssl client purpose.
10668 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10670 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10671 is a little unclear about how a blank password is handled.
10672 Since the password in encoded as a BMPString with terminating
10673 double NULL a zero length password would end up as just the
10674 double NULL. However no password at all is different and is
10675 handled differently in the PKCS#12 key generation code. NS
10676 treats a blank password as zero length. MSIE treats it as no
10677 password on export: but it will try both on import. We now do
10678 the same: PKCS12_parse() tries zero length and no password if
10679 the password is set to "" or NULL (NULL is now a valid password:
10680 it wasn't before) as does the pkcs12 application.
10683 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10684 perror when PEM_read_bio_X509_REQ fails, the error message must
10685 be obtained from the error queue.
10688 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10689 it in ERR_remove_state if appropriate, and change ERR_get_state
10690 accordingly to avoid race conditions (this is necessary because
10691 thread_hash is no longer constant once set).
10694 *) Bugfix for linux-elf makefile.one.
10697 *) RSA_get_default_method() will now cause a default
10698 RSA_METHOD to be chosen if one doesn't exist already.
10699 Previously this was only set during a call to RSA_new()
10700 or RSA_new_method(NULL) meaning it was possible for
10701 RSA_get_default_method() to return NULL.
10704 *) Added native name translation to the existing DSO code
10705 that will convert (if the flag to do so is set) filenames
10706 that are sufficiently small and have no path information
10707 into a canonical native form. Eg. "blah" converted to
10708 "libblah.so" or "blah.dll" etc.
10711 *) New function ERR_error_string_n(e, buf, len) which is like
10712 ERR_error_string(e, buf), but writes at most 'len' bytes
10713 including the 0 terminator. For ERR_error_string_n, 'buf'
10715 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10717 *) CONF library reworked to become more general. A new CONF
10718 configuration file reader "class" is implemented as well as a
10719 new functions (NCONF_*, for "New CONF") to handle it. The now
10720 old CONF_* functions are still there, but are reimplemented to
10721 work in terms of the new functions. Also, a set of functions
10722 to handle the internal storage of the configuration data is
10723 provided to make it easier to write new configuration file
10724 reader "classes" (I can definitely see something reading a
10725 configuration file in XML format, for example), called _CONF_*,
10726 or "the configuration storage API"...
10728 The new configuration file reading functions are:
10730 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10731 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10733 NCONF_default, NCONF_WIN32
10735 NCONF_dump_fp, NCONF_dump_bio
10737 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10738 NCONF_new creates a new CONF object. This works in the same way
10739 as other interfaces in OpenSSL, like the BIO interface.
10740 NCONF_dump_* dump the internal storage of the configuration file,
10741 which is useful for debugging. All other functions take the same
10742 arguments as the old CONF_* functions with the exception of the
10743 first that must be a `CONF *' instead of a `LHASH *'.
10745 To make it easier to use the new classes with the old CONF_* functions,
10746 the function CONF_set_default_method is provided.
10749 *) Add '-tls1' option to 'openssl ciphers', which was already
10750 mentioned in the documentation but had not been implemented.
10751 (This option is not yet really useful because even the additional
10752 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10755 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10756 OpenSSL-based applications) load shared libraries and bind to
10757 them in a portable way.
10758 [Geoff Thorpe, with contributions from Richard Levitte]
10760 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10762 *) Make sure _lrotl and _lrotr are only used with MSVC.
10764 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10765 (the default implementation of RAND_status).
10767 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10768 to '-clrext' (= clear extensions), as intended and documented.
10769 [Bodo Moeller; inconsistency pointed out by Michael Attili
10770 <attili@amaxo.com>]
10772 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10773 was larger than the MD block size.
10774 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10776 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10777 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10778 using the passed key: if the passed key was a private key the result
10779 of X509_print(), for example, would be to print out all the private key
10783 *) des_quad_cksum() byte order bug fix.
10784 [Ulf Möller, using the problem description in krb4-0.9.7, where
10785 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10787 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10789 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10791 *) For easily testing in shell scripts whether some command
10792 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10793 returns with exit code 0 iff no command of the given name is available.
10794 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10795 the output goes to stdout and nothing is printed to stderr.
10796 Additional arguments are always ignored.
10798 Since for each cipher there is a command of the same name,
10799 the 'no-cipher' compilation switches can be tested this way.
10801 ('openssl no-XXX' is not able to detect pseudo-commands such
10802 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10805 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10808 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10809 is set; it will be thrown away anyway because each handshake creates
10811 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10812 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10813 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10814 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10817 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10818 'Q' and 'R' lose their special meanings (quit/renegotiate).
10819 This is part of what -quiet does; unlike -quiet, -ign_eof
10820 does not suppress any output.
10823 *) Add compatibility options to the purpose and trust code. The
10824 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10825 accepts a certificate or CA, this was the previous behaviour,
10826 with all the associated security issues.
10828 X509_TRUST_COMPAT is the old trust behaviour: only and
10829 automatically trust self signed roots in certificate store. A
10830 new trust setting X509_TRUST_DEFAULT is used to specify that
10831 a purpose has no associated trust setting and it should instead
10832 use the value in the default purpose.
10835 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10836 and fix a memory leak.
10839 *) In util/mkerr.pl (which implements 'make errors'), preserve
10840 reason strings from the previous version of the .c file, as
10841 the default to have only downcase letters (and digits) in
10842 automatically generated reasons codes is not always appropriate.
10845 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10846 using strerror. Previously, ERR_reason_error_string() returned
10847 library names as reason strings for SYSerr; but SYSerr is a special
10848 case where small numbers are errno values, not library numbers.
10851 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10852 converts DSA parameters into DH parameters. (When creating parameters,
10853 DSA_generate_parameters is used.)
10856 *) Include 'length' (recommended exponent length) in C code generated
10857 by 'openssl dhparam -C'.
10860 *) The second argument to set_label in perlasm was already being used
10861 so couldn't be used as a "file scope" flag. Moved to third argument
10865 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10866 instead of RAND_bytes for encryption IVs and salts.
10869 *) Include RAND_status() into RAND_METHOD instead of implementing
10870 it only for md_rand.c Otherwise replacing the PRNG by calling
10871 RAND_set_rand_method would be impossible.
10874 *) Don't let DSA_generate_key() enter an infinite loop if the random
10875 number generation fails.
10878 *) New 'rand' application for creating pseudo-random output.
10881 *) Added configuration support for Linux/IA64
10882 [Rolf Haberrecker <rolf@suse.de>]
10884 *) Assembler module support for Mingw32.
10887 *) Shared library support for HPUX (in shlib/).
10888 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10890 *) Shared library support for Solaris gcc.
10891 [Lutz Behnke <behnke@trustcenter.de>]
10893 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10895 *) PKCS7_encrypt() was adding text MIME headers twice because they
10896 were added manually and by SMIME_crlf_copy().
10899 *) In bntest.c don't call BN_rand with zero bits argument.
10900 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10902 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10903 case was implemented. This caused BN_div_recp() to fail occasionally.
10906 *) Add an optional second argument to the set_label() in the perl
10907 assembly language builder. If this argument exists and is set
10908 to 1 it signals that the assembler should use a symbol whose
10909 scope is the entire file, not just the current function. This
10910 is needed with MASM which uses the format label:: for this scope.
10911 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10913 *) Change the ASN1 types so they are typedefs by default. Before
10914 almost all types were #define'd to ASN1_STRING which was causing
10915 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10919 *) Change names of new functions to the new get1/get0 naming
10920 convention: After 'get1', the caller owns a reference count
10921 and has to call ..._free; 'get0' returns a pointer to some
10922 data structure without incrementing reference counters.
10923 (Some of the existing 'get' functions increment a reference
10924 counter, some don't.)
10925 Similarly, 'set1' and 'add1' functions increase reference
10926 counters or duplicate objects.
10929 *) Allow for the possibility of temp RSA key generation failure:
10930 the code used to assume it always worked and crashed on failure.
10933 *) Fix potential buffer overrun problem in BIO_printf().
10934 [Ulf Möller, using public domain code by Patrick Powell; problem
10935 pointed out by David Sacerdote <das33@cornell.edu>]
10937 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10938 RAND_egd() and RAND_status(). In the command line application,
10939 the EGD socket can be specified like a seed file using RANDFILE
10943 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10944 Some CAs (e.g. Verisign) distribute certificates in this form.
10947 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10948 list to exclude them. This means that no special compilation option
10949 is needed to use anonymous DH: it just needs to be included in the
10953 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10954 EVP_MD_type. The old functionality is available in a new macro called
10955 EVP_MD_md(). Change code that uses it and update docs.
10958 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10959 where the 'void *' argument is replaced by a function pointer argument.
10960 Previously 'void *' was abused to point to functions, which works on
10961 many platforms, but is not correct. As these functions are usually
10962 called by macros defined in OpenSSL header files, most source code
10963 should work without changes.
10966 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10967 sections with information on -D... compiler switches used for
10968 compiling the library so that applications can see them. To enable
10969 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10970 must be defined. E.g.,
10971 #define OPENSSL_ALGORITHM_DEFINES
10972 #include <openssl/opensslconf.h>
10973 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10974 [Richard Levitte, Ulf and Bodo Möller]
10976 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10980 *) Change the 'other' type in certificate aux info to a STACK_OF
10981 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10982 the required ASN1 format: arbitrary types determined by an OID.
10985 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10986 argument to 'req'. This is not because the function is newer or
10987 better than others it just uses the work 'NEW' in the certificate
10988 request header lines. Some software needs this.
10991 *) Reorganise password command line arguments: now passwords can be
10992 obtained from various sources. Delete the PEM_cb function and make
10993 it the default behaviour: i.e. if the callback is NULL and the
10994 usrdata argument is not NULL interpret it as a null terminated pass
10995 phrase. If usrdata and the callback are NULL then the pass phrase
10996 is prompted for as usual.
10999 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
11000 the support is automatically enabled. The resulting binaries will
11001 autodetect the card and use it if present.
11002 [Ben Laurie and Compaq Inc.]
11004 *) Work around for Netscape hang bug. This sends certificate request
11005 and server done in one record. Since this is perfectly legal in the
11006 SSL/TLS protocol it isn't a "bug" option and is on by default. See
11007 the bugs/SSLv3 entry for more info.
11010 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
11013 *) Add -rand argument to smime and pkcs12 applications and read/write
11017 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
11020 *) Add command line password options to the remaining applications.
11023 *) Bug fix for BN_div_recp() for numerators with an even number of
11027 *) More tests in bntest.c, and changed test_bn output.
11030 *) ./config recognizes MacOS X now.
11033 *) Bug fix for BN_div() when the first words of num and divisor are
11034 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
11037 *) Add support for various broken PKCS#8 formats, and command line
11038 options to produce them.
11041 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
11042 get temporary BIGNUMs from a BN_CTX.
11045 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
11049 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
11050 include a #define from the old name to the new. The original intent
11051 was that statically linked binaries could for example just call
11052 SSLeay_add_all_ciphers() to just add ciphers to the table and not
11053 link with digests. This never worked because SSLeay_add_all_digests()
11054 and SSLeay_add_all_ciphers() were in the same source file so calling
11055 one would link with the other. They are now in separate source files.
11058 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
11061 *) Use a less unusual form of the Miller-Rabin primality test (it used
11062 a binary algorithm for exponentiation integrated into the Miller-Rabin
11063 loop, our standard modexp algorithms are faster).
11066 *) Support for the EBCDIC character set completed.
11067 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
11069 *) Source code cleanups: use const where appropriate, eliminate casts,
11070 use void * instead of char * in lhash.
11073 *) Bugfix: ssl3_send_server_key_exchange was not restartable
11074 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
11075 this the server could overwrite ephemeral keys that the client
11079 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
11080 using 50 iterations of the Rabin-Miller test.
11082 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
11083 iterations of the Rabin-Miller test as required by the appendix
11084 to FIPS PUB 186[-1]) instead of DSA_is_prime.
11085 As BN_is_prime_fasttest includes trial division, DSA parameter
11086 generation becomes much faster.
11088 This implies a change for the callback functions in DSA_is_prime
11089 and DSA_generate_parameters: The callback function is called once
11090 for each positive witness in the Rabin-Miller test, not just
11091 occasionally in the inner loop; and the parameters to the
11092 callback function now provide an iteration count for the outer
11093 loop rather than for the current invocation of the inner loop.
11094 DSA_generate_parameters additionally can call the callback
11095 function with an 'iteration count' of -1, meaning that a
11096 candidate has passed the trial division test (when q is generated
11097 from an application-provided seed, trial division is skipped).
11100 *) New function BN_is_prime_fasttest that optionally does trial
11101 division before starting the Rabin-Miller test and has
11102 an additional BN_CTX * argument (whereas BN_is_prime always
11103 has to allocate at least one BN_CTX).
11104 'callback(1, -1, cb_arg)' is called when a number has passed the
11105 trial division stage.
11108 *) Fix for bug in CRL encoding. The validity dates weren't being handled
11112 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
11115 *) New function BN_pseudo_rand().
11118 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
11119 bignum version of BN_from_montgomery() with the working code from
11120 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
11124 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
11125 made it impossible to use the same SSL_SESSION data structure in
11126 SSL2 clients in multiple threads.
11129 *) The return value of RAND_load_file() no longer counts bytes obtained
11130 by stat(). RAND_load_file(..., -1) is new and uses the complete file
11131 to seed the PRNG (previously an explicit byte count was required).
11132 [Ulf Möller, Bodo Möller]
11134 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
11135 used (char *) instead of (void *) and had casts all over the place.
11138 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
11141 *) Retain source code compatibility for BN_prime_checks macro:
11142 BN_is_prime(..., BN_prime_checks, ...) now uses
11143 BN_prime_checks_for_size to determine the appropriate number of
11144 Rabin-Miller iterations.
11147 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
11148 DH_CHECK_P_NOT_SAFE_PRIME.
11149 (Check if this is true? OpenPGP calls them "strong".)
11152 *) Merge the functionality of "dh" and "gendh" programs into a new program
11153 "dhparam". The old programs are retained for now but will handle DH keys
11154 (instead of parameters) in future.
11157 *) Make the ciphers, s_server and s_client programs check the return values
11158 when a new cipher list is set.
11161 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
11162 ciphers. Before when the 56bit ciphers were enabled the sorting was
11165 The syntax for the cipher sorting has been extended to support sorting by
11166 cipher-strength (using the strength_bits hard coded in the tables).
11167 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
11169 Fix a bug in the cipher-command parser: when supplying a cipher command
11170 string with an "undefined" symbol (neither command nor alphanumeric
11171 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
11172 an error is flagged.
11174 Due to the strength-sorting extension, the code of the
11175 ssl_create_cipher_list() function was completely rearranged. I hope that
11176 the readability was also increased :-)
11177 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
11179 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
11180 for the first serial number and places 2 in the serial number file. This
11181 avoids problems when the root CA is created with serial number zero and
11182 the first user certificate has the same issuer name and serial number
11186 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
11187 the new code. Add documentation for this stuff.
11190 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
11191 X509_*() to X509at_*() on the grounds that they don't handle X509
11192 structures and behave in an analogous way to the X509v3 functions:
11193 they shouldn't be called directly but wrapper functions should be used
11196 So we also now have some wrapper functions that call the X509at functions
11197 when passed certificate requests. (TO DO: similar things can be done with
11198 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11199 things. Some of these need some d2i or i2d and print functionality
11200 because they handle more complex structures.)
11203 *) Add missing #ifndefs that caused missing symbols when building libssl
11204 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11205 NO_RSA in ssl/s2*.c.
11206 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11208 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11209 has a return value which indicates the quality of the random data
11210 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11211 error queue. New function RAND_pseudo_bytes() generates output that is
11212 guaranteed to be unique but not unpredictable. RAND_add is like
11213 RAND_seed, but takes an extra argument for an entropy estimate
11214 (RAND_seed always assumes full entropy).
11217 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11218 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11219 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11220 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11221 false-positive rate of at most 2^-80 for random input.
11224 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11227 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11228 in the 0.9.5 release), this returns the chain
11229 from an X509_CTX structure with a dup of the stack and all
11230 the X509 reference counts upped: so the stack will exist
11231 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11234 Also make SSL_SESSION_print() print out the verify return
11238 *) Add manpage for the pkcs12 command. Also change the default
11239 behaviour so MAC iteration counts are used unless the new
11240 -nomaciter option is used. This improves file security and
11241 only older versions of MSIE (4.0 for example) need it.
11244 *) Honor the no-xxx Configure options when creating .DEF files.
11247 *) Add PKCS#10 attributes to field table: challengePassword,
11248 unstructuredName and unstructuredAddress. These are taken from
11249 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11250 international characters are used.
11252 More changes to X509_ATTRIBUTE code: allow the setting of types
11253 based on strings. Remove the 'loc' parameter when adding
11254 attributes because these will be a SET OF encoding which is sorted
11258 *) Initial changes to the 'req' utility to allow request generation
11259 automation. This will allow an application to just generate a template
11260 file containing all the field values and have req construct the
11263 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11264 used all over the place including certificate requests and PKCS#7
11265 structures. They are currently handled manually where necessary with
11266 some primitive wrappers for PKCS#7. The new functions behave in a
11267 manner analogous to the X509 extension functions: they allow
11268 attributes to be looked up by NID and added.
11270 Later something similar to the X509V3 code would be desirable to
11271 automatically handle the encoding, decoding and printing of the
11272 more complex types. The string types like challengePassword can
11273 be handled by the string table functions.
11275 Also modified the multi byte string table handling. Now there is
11276 a 'global mask' which masks out certain types. The table itself
11277 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11278 is useful when for example there is only one permissible type
11279 (as in countryName) and using the mask might result in no valid
11283 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11284 SSL_get_peer_finished to allow applications to obtain the latest
11285 Finished messages sent to the peer or expected from the peer,
11286 respectively. (SSL_get_peer_finished is usually the Finished message
11287 actually received from the peer, otherwise the protocol will be aborted.)
11289 As the Finished message are message digests of the complete handshake
11290 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11291 be used for external authentication procedures when the authentication
11292 provided by SSL/TLS is not desired or is not enough.
11295 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11296 the host supports BWX extension and if Compaq C is present on the
11297 $PATH. Just exploiting of the BWX extension results in 20-30%
11298 performance kick for some algorithms, e.g. DES and RC4 to mention
11299 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11303 *) Add support for MS "fast SGC". This is arguably a violation of the
11304 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11305 weak crypto and after checking the certificate is SGC a second one
11306 with strong crypto. MS SGC stops the first handshake after receiving
11307 the server certificate message and sends a second client hello. Since
11308 a server will typically do all the time consuming operations before
11309 expecting any further messages from the client (server key exchange
11310 is the most expensive) there is little difference between the two.
11312 To get OpenSSL to support MS SGC we have to permit a second client
11313 hello message after we have sent server done. In addition we have to
11314 reset the MAC if we do get this second client hello.
11317 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11318 if a DER encoded private key is RSA or DSA traditional format. Changed
11319 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11320 format DER encoded private key. Newer code should use PKCS#8 format which
11321 has the key type encoded in the ASN1 structure. Added DER private key
11322 support to pkcs8 application.
11325 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11326 ciphersuites has been selected (as required by the SSL 3/TLS 1
11327 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11328 is set, we interpret this as a request to violate the specification
11329 (the worst that can happen is a handshake failure, and 'correct'
11330 behaviour would result in a handshake failure anyway).
11333 *) In SSL_CTX_add_session, take into account that there might be multiple
11334 SSL_SESSION structures with the same session ID (e.g. when two threads
11335 concurrently obtain them from an external cache).
11336 The internal cache can handle only one SSL_SESSION with a given ID,
11337 so if there's a conflict, we now throw out the old one to achieve
11341 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11342 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11343 some routines that use cipher OIDs: some ciphers do not have OIDs
11344 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11348 *) Simplify the trust setting structure and code. Now we just have
11349 two sequences of OIDs for trusted and rejected settings. These will
11350 typically have values the same as the extended key usage extension
11351 and any application specific purposes.
11353 The trust checking code now has a default behaviour: it will just
11354 check for an object with the same NID as the passed id. Functions can
11355 be provided to override either the default behaviour or the behaviour
11356 for a given id. SSL client, server and email already have functions
11357 in place for compatibility: they check the NID and also return "trusted"
11358 if the certificate is self signed.
11361 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11362 traditional format into an EVP_PKEY structure.
11365 *) Add a password callback function PEM_cb() which either prompts for
11366 a password if usr_data is NULL or otherwise assumes it is a null
11367 terminated password. Allow passwords to be passed on command line
11368 environment or config files in a few more utilities.
11371 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11372 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11373 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11374 Update documentation.
11377 *) Support for ASN1 "NULL" type. This could be handled before by using
11378 ASN1_TYPE but there wasn't any function that would try to read a NULL
11379 and produce an error if it couldn't. For compatibility we also have
11380 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11381 don't allocate anything because they don't need to.
11384 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11386 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11388 *) Rebuild of the memory allocation routines used by OpenSSL code and
11389 possibly others as well. The purpose is to make an interface that
11390 provide hooks so anyone can build a separate set of allocation and
11391 deallocation routines to be used by OpenSSL, for example memory
11392 pool implementations, or something else, which was previously hard
11393 since Malloc(), Realloc() and Free() were defined as macros having
11394 the values malloc, realloc and free, respectively (except for Win32
11395 compilations). The same is provided for memory debugging code.
11396 OpenSSL already comes with functionality to find memory leaks, but
11397 this gives people a chance to debug other memory problems.
11399 With these changes, a new set of functions and macros have appeared:
11401 CRYPTO_set_mem_debug_functions() [F]
11402 CRYPTO_get_mem_debug_functions() [F]
11403 CRYPTO_dbg_set_options() [F]
11404 CRYPTO_dbg_get_options() [F]
11405 CRYPTO_malloc_debug_init() [M]
11407 The memory debug functions are NULL by default, unless the library
11408 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11409 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11410 gives the standard debugging functions that come with OpenSSL) or
11411 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11412 provided by the library user) must be used. When the standard
11413 debugging functions are used, CRYPTO_dbg_set_options can be used to
11414 request additional information:
11415 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11416 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11418 Also, things like CRYPTO_set_mem_functions will always give the
11419 expected result (the new set of functions is used for allocation
11420 and deallocation) at all times, regardless of platform and compiler
11423 To finish it up, some functions that were never use in any other
11424 way than through macros have a new API and new semantic:
11426 CRYPTO_dbg_malloc()
11427 CRYPTO_dbg_realloc()
11430 All macros of value have retained their old syntax.
11431 [Richard Levitte and Bodo Moeller]
11433 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11434 ordering of SMIMECapabilities wasn't in "strength order" and there
11435 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11439 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11440 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11441 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11443 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11444 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11445 functionality to handle multipart/signed properly) and a utility
11446 called 'smime' to call all this stuff. This is based on code I
11447 originally wrote for Celo who have kindly allowed it to be
11448 included in OpenSSL.
11451 *) Add variants des_set_key_checked and des_set_key_unchecked of
11452 des_set_key (aka des_key_sched). Global variable des_check_key
11453 decides which of these is called by des_set_key; this way
11454 des_check_key behaves as it always did, but applications and
11455 the library itself, which was buggy for des_check_key == 1,
11456 have a cleaner way to pick the version they need.
11459 *) New function PKCS12_newpass() which changes the password of a
11463 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11464 dynamic mix. In both cases the ids can be used as an index into the
11465 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11466 functions so they accept a list of the field values and the
11467 application doesn't need to directly manipulate the X509_TRUST
11471 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11475 *) Modify the way the V3 extension code looks up extensions. This now
11476 works in a similar way to the object code: we have some "standard"
11477 extensions in a static table which is searched with OBJ_bsearch()
11478 and the application can add dynamic ones if needed. The file
11479 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11480 updated whenever a new extension is added to the core code and kept
11481 in ext_nid order. There is a simple program 'tabtest.c' which checks
11482 this. New extensions are not added too often so this file can readily
11483 be maintained manually.
11485 There are two big advantages in doing things this way. The extensions
11486 can be looked up immediately and no longer need to be "added" using
11487 X509V3_add_standard_extensions(): this function now does nothing.
11488 [Side note: I get *lots* of email saying the extension code doesn't
11489 work because people forget to call this function]
11490 Also no dynamic allocation is done unless new extensions are added:
11491 so if we don't add custom extensions there is no need to call
11492 X509V3_EXT_cleanup().
11495 *) Modify enc utility's salting as follows: make salting the default. Add a
11496 magic header, so unsalted files fail gracefully instead of just decrypting
11497 to garbage. This is because not salting is a big security hole, so people
11498 should be discouraged from doing it.
11501 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11502 digest to be passed on the command line but it only used this
11503 parameter when signing a certificate. Modified so all relevant
11504 operations are affected by the digest parameter including the
11505 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11506 DSA key was used because it didn't fix the digest.
11509 *) Initial certificate chain verify code. Currently tests the untrusted
11510 certificates for consistency with the verify purpose (which is set
11511 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11513 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11514 this is because it will reject chains with invalid extensions whereas
11515 every previous version of OpenSSL and SSLeay made no checks at all.
11517 Trust code: checks the root CA for the relevant trust settings. Trust
11518 settings have an initial value consistent with the verify purpose: e.g.
11519 if the verify purpose is for SSL client use it expects the CA to be
11520 trusted for SSL client use. However the default value can be changed to
11521 permit custom trust settings: one example of this would be to only trust
11522 certificates from a specific "secure" set of CAs.
11524 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11525 which should be used for version portability: especially since the
11526 verify structure is likely to change more often now.
11528 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11529 to set them. If not set then assume SSL clients will verify SSL servers
11532 Two new options to the verify program: -untrusted allows a set of
11533 untrusted certificates to be passed in and -purpose which sets the
11534 intended purpose of the certificate. If a purpose is set then the
11535 new chain verify code is used to check extension consistency.
11538 *) Support for the authority information access extension.
11541 *) Modify RSA and DSA PEM read routines to transparently handle
11542 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11543 public keys in a format compatible with certificate
11544 SubjectPublicKeyInfo structures. Unfortunately there were already
11545 functions called *_PublicKey_* which used various odd formats so
11546 these are retained for compatibility: however the DSA variants were
11547 never in a public release so they have been deleted. Changed dsa/rsa
11548 utilities to handle the new format: note no releases ever handled public
11549 keys so we should be OK.
11551 The primary motivation for this change is to avoid the same fiasco
11552 that dogs private keys: there are several incompatible private key
11553 formats some of which are standard and some OpenSSL specific and
11554 require various evil hacks to allow partial transparent handling and
11555 even then it doesn't work with DER formats. Given the option anything
11556 other than PKCS#8 should be dumped: but the other formats have to
11557 stay in the name of compatibility.
11559 With public keys and the benefit of hindsight one standard format
11560 is used which works with EVP_PKEY, RSA or DSA structures: though
11561 it clearly returns an error if you try to read the wrong kind of key.
11563 Added a -pubkey option to the 'x509' utility to output the public key.
11564 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11565 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11566 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11567 that do the same as the EVP_PKEY_assign_*() except they up the
11568 reference count of the added key (they don't "swallow" the
11572 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11573 CRLs would fail if the file contained no certificates or no CRLs:
11574 added a new function to read in both types and return the number
11575 read: this means that if none are read it will be an error. The
11576 DER versions of the certificate and CRL reader would always fail
11577 because it isn't possible to mix certificates and CRLs in DER format
11578 without choking one or the other routine. Changed this to just read
11579 a certificate: this is the best we can do. Also modified the code
11580 in apps/verify.c to take notice of return codes: it was previously
11581 attempting to read in certificates from NULL pointers and ignoring
11582 any errors: this is one reason why the cert and CRL reader seemed
11583 to work. It doesn't check return codes from the default certificate
11584 routines: these may well fail if the certificates aren't installed.
11587 *) Code to support otherName option in GeneralName.
11590 *) First update to verify code. Change the verify utility
11591 so it warns if it is passed a self signed certificate:
11592 for consistency with the normal behaviour. X509_verify
11593 has been modified to it will now verify a self signed
11594 certificate if *exactly* the same certificate appears
11595 in the store: it was previously impossible to trust a
11596 single self signed certificate. This means that:
11597 openssl verify ss.pem
11598 now gives a warning about a self signed certificate but
11599 openssl verify -CAfile ss.pem ss.pem
11603 *) For servers, store verify_result in SSL_SESSION data structure
11604 (and add it to external session representation).
11605 This is needed when client certificate verifications fails,
11606 but an application-provided verification callback (set by
11607 SSL_CTX_set_cert_verify_callback) allows accepting the session
11608 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11609 but returns 1): When the session is reused, we have to set
11610 ssl->verify_result to the appropriate error code to avoid
11612 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11614 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11615 case in PKCS7_dataInit() where the signed PKCS7 structure
11616 didn't contain any existing data because it was being created.
11617 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11619 *) Add a salt to the key derivation routines in enc.c. This
11620 forms the first 8 bytes of the encrypted file. Also add a
11621 -S option to allow a salt to be input on the command line.
11624 *) New function X509_cmp(). Oddly enough there wasn't a function
11625 to compare two certificates. We do this by working out the SHA1
11626 hash and comparing that. X509_cmp() will be needed by the trust
11630 *) SSL_get1_session() is like SSL_get_session(), but increments
11631 the reference count in the SSL_SESSION returned.
11632 [Geoff Thorpe <geoff@eu.c2.net>]
11634 *) Fix for 'req': it was adding a null to request attributes.
11635 Also change the X509_LOOKUP and X509_INFO code to handle
11636 certificate auxiliary information.
11639 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11643 *) Add the possibility to add extra information to the memory leak
11644 detecting output, to form tracebacks, showing from where each
11645 allocation was originated: CRYPTO_push_info("constant string") adds
11646 the string plus current file name and line number to a per-thread
11647 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11648 is like calling CYRPTO_pop_info() until the stack is empty.
11649 Also updated memory leak detection code to be multi-thread-safe.
11652 *) Add options -text and -noout to pkcs7 utility and delete the
11653 encryption options which never did anything. Update docs.
11656 *) Add options to some of the utilities to allow the pass phrase
11657 to be included on either the command line (not recommended on
11658 OSes like Unix) or read from the environment. Update the
11659 manpages and fix a few bugs.
11662 *) Add a few manpages for some of the openssl commands.
11665 *) Fix the -revoke option in ca. It was freeing up memory twice,
11666 leaking and not finding already revoked certificates.
11669 *) Extensive changes to support certificate auxiliary information.
11670 This involves the use of X509_CERT_AUX structure and X509_AUX
11671 functions. An X509_AUX function such as PEM_read_X509_AUX()
11672 can still read in a certificate file in the usual way but it
11673 will also read in any additional "auxiliary information". By
11674 doing things this way a fair degree of compatibility can be
11675 retained: existing certificates can have this information added
11676 using the new 'x509' options.
11678 Current auxiliary information includes an "alias" and some trust
11679 settings. The trust settings will ultimately be used in enhanced
11680 certificate chain verification routines: currently a certificate
11681 can only be trusted if it is self signed and then it is trusted
11685 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11686 The problem was that one of the replacement routines had not been working
11687 since SSLeay releases. For now the offending routine has been replaced
11688 with non-optimised assembler. Even so, this now gives around 95%
11689 performance improvement for 1024 bit RSA signs.
11692 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11693 handling. Most clients have the effective key size in bits equal to
11694 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11695 A few however don't do this and instead use the size of the decrypted key
11696 to determine the RC2 key length and the AlgorithmIdentifier to determine
11697 the effective key length. In this case the effective key length can still
11698 be 40 bits but the key length can be 168 bits for example. This is fixed
11699 by manually forcing an RC2 key into the EVP_PKEY structure because the
11700 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11701 the key length and effective key length are equal.
11704 *) Add a bunch of functions that should simplify the creation of
11705 X509_NAME structures. Now you should be able to do:
11706 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11707 and have it automatically work out the correct field type and fill in
11708 the structures. The more adventurous can try:
11709 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11710 and it will (hopefully) work out the correct multibyte encoding.
11713 *) Change the 'req' utility to use the new field handling and multibyte
11714 copy routines. Before the DN field creation was handled in an ad hoc
11715 way in req, ca, and x509 which was rather broken and didn't support
11716 BMPStrings or UTF8Strings. Since some software doesn't implement
11717 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11718 using the dirstring_type option. See the new comment in the default
11719 openssl.cnf for more info.
11722 *) Make crypto/rand/md_rand.c more robust:
11723 - Assure unique random numbers after fork().
11724 - Make sure that concurrent threads access the global counter and
11725 md serializably so that we never lose entropy in them
11726 or use exactly the same state in multiple threads.
11727 Access to the large state is not always serializable because
11728 the additional locking could be a performance killer, and
11729 md should be large enough anyway.
11732 *) New file apps/app_rand.c with commonly needed functionality
11733 for handling the random seed file.
11735 Use the random seed file in some applications that previously did not:
11737 dsaparam -genkey (which also ignored its '-rand' option),
11740 x509 (when signing).
11741 Except on systems with /dev/urandom, it is crucial to have a random
11742 seed file at least for key creation, DSA signing, and for DH exchanges;
11743 for RSA signatures we could do without one.
11745 gendh and gendsa (unlike genrsa) used to read only the first byte
11746 of each file listed in the '-rand' option. The function as previously
11747 found in genrsa is now in app_rand.c and is used by all programs
11748 that support '-rand'.
11751 *) In RAND_write_file, use mode 0600 for creating files;
11752 don't just chmod when it may be too late.
11755 *) Report an error from X509_STORE_load_locations
11756 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11759 *) New function ASN1_mbstring_copy() this copies a string in either
11760 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11761 into an ASN1_STRING type. A mask of permissible types is passed
11762 and it chooses the "minimal" type to use or an error if not type
11766 *) Add function equivalents to the various macros in asn1.h. The old
11767 macros are retained with an M_ prefix. Code inside the library can
11768 use the M_ macros. External code (including the openssl utility)
11769 should *NOT* in order to be "shared library friendly".
11772 *) Add various functions that can check a certificate's extensions
11773 to see if it usable for various purposes such as SSL client,
11774 server or S/MIME and CAs of these types. This is currently
11775 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11776 verification. Also added a -purpose flag to x509 utility to
11777 print out all the purposes.
11780 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11784 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11785 for, obtain and decode and extension and obtain its critical flag.
11786 This allows all the necessary extension code to be handled in a
11787 single function call.
11790 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11791 platforms. See crypto/rc4/rc4_enc.c for further details.
11794 *) New -noout option to asn1parse. This causes no output to be produced
11795 its main use is when combined with -strparse and -out to extract data
11796 from a file (which may not be in ASN.1 format).
11799 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11800 when producing the local key id.
11801 [Richard Levitte <levitte@stacken.kth.se>]
11803 *) New option -dhparam in s_server. This allows a DH parameter file to be
11804 stated explicitly. If it is not stated then it tries the first server
11805 certificate file. The previous behaviour hard coded the filename
11809 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11810 a public key to be input or output. For example:
11811 openssl rsa -in key.pem -pubout -out pubkey.pem
11812 Also added necessary DSA public key functions to handle this.
11815 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11816 in the message. This was handled by allowing
11817 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11818 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11820 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11821 to the end of the strings whereas this didn't. This would cause problems
11822 if strings read with d2i_ASN1_bytes() were later modified.
11823 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11825 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11826 data and it contains EOF it will end up returning an error. This is
11827 caused by input 46 bytes long. The cause is due to the way base64
11828 BIOs find the start of base64 encoded data. They do this by trying a
11829 trial decode on each line until they find one that works. When they
11830 do a flag is set and it starts again knowing it can pass all the
11831 data directly through the decoder. Unfortunately it doesn't reset
11832 the context it uses. This means that if EOF is reached an attempt
11833 is made to pass two EOFs through the context and this causes the
11834 resulting error. This can also cause other problems as well. As is
11835 usual with these problems it takes *ages* to find and the fix is
11836 trivial: move one line.
11837 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11839 *) Ugly workaround to get s_client and s_server working under Windows. The
11840 old code wouldn't work because it needed to select() on sockets and the
11841 tty (for keypresses and to see if data could be written). Win32 only
11842 supports select() on sockets so we select() with a 1s timeout on the
11843 sockets and then see if any characters are waiting to be read, if none
11844 are present then we retry, we also assume we can always write data to
11845 the tty. This isn't nice because the code then blocks until we've
11846 received a complete line of data and it is effectively polling the
11847 keyboard at 1s intervals: however it's quite a bit better than not
11848 working at all :-) A dedicated Windows application might handle this
11849 with an event loop for example.
11852 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11853 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11854 will be called when RSA_sign() and RSA_verify() are used. This is useful
11855 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11856 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11857 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11858 This necessitated the support of an extra signature type NID_md5_sha1
11859 for SSL signatures and modifications to the SSL library to use it instead
11860 of calling RSA_public_decrypt() and RSA_private_encrypt().
11863 *) Add new -verify -CAfile and -CApath options to the crl program, these
11864 will lookup a CRL issuers certificate and verify the signature in a
11865 similar way to the verify program. Tidy up the crl program so it
11866 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11867 less strict. It will now permit CRL extensions even if it is not
11868 a V2 CRL: this will allow it to tolerate some broken CRLs.
11871 *) Initialize all non-automatic variables each time one of the openssl
11872 sub-programs is started (this is necessary as they may be started
11873 multiple times from the "OpenSSL>" prompt).
11874 [Lennart Bang, Bodo Moeller]
11876 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11877 removing all other RSA functionality (this is what NO_RSA does). This
11878 is so (for example) those in the US can disable those operations covered
11879 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11883 *) Non-copying interface to BIO pairs.
11884 (still largely untested)
11887 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
11888 ASCII string. This was handled independently in various places before.
11891 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11892 UTF8 strings a character at a time.
11895 *) Use client_version from client hello to select the protocol
11896 (s23_srvr.c) and for RSA client key exchange verification
11897 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11900 *) Add various utility functions to handle SPKACs, these were previously
11901 handled by poking round in the structure internals. Added new function
11902 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11903 print, verify and generate SPKACs. Based on an original idea from
11904 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11907 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11910 *) Allow the config file extension section to be overwritten on the
11911 command line. Based on an original idea from Massimiliano Pala
11912 <madwolf@comune.modena.it>. The new option is called -extensions
11913 and can be applied to ca, req and x509. Also -reqexts to override
11914 the request extensions in req and -crlexts to override the crl extensions
11918 *) Add new feature to the SPKAC handling in ca. Now you can include
11919 the same field multiple times by preceding it by "XXXX." for example:
11922 this is the same syntax as used in the req config file.
11925 *) Allow certificate extensions to be added to certificate requests. These
11926 are specified in a 'req_extensions' option of the req section of the
11927 config file. They can be printed out with the -text option to req but
11928 are otherwise ignored at present.
11931 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11932 data read consists of only the final block it would not decrypted because
11933 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11934 A misplaced 'break' also meant the decrypted final block might not be
11935 copied until the next read.
11938 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11939 a few extra parameters to the DH structure: these will be useful if
11940 for example we want the value of 'q' or implement X9.42 DH.
11943 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11944 provides hooks that allow the default DSA functions or functions on a
11945 "per key" basis to be replaced. This allows hardware acceleration and
11946 hardware key storage to be handled without major modification to the
11947 library. Also added low level modexp hooks and CRYPTO_EX structure and
11948 associated functions.
11951 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11952 as "read only": it can't be written to and the buffer it points to will
11953 not be freed. Reading from a read only BIO is much more efficient than
11954 a normal memory BIO. This was added because there are several times when
11955 an area of memory needs to be read from a BIO. The previous method was
11956 to create a memory BIO and write the data to it, this results in two
11957 copies of the data and an O(n^2) reading algorithm. There is a new
11958 function BIO_new_mem_buf() which creates a read only memory BIO from
11959 an area of memory. Also modified the PKCS#7 routines to use read only
11963 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11964 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11965 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11966 but a retry condition occurred while trying to read the rest.
11969 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11970 NID_pkcs7_encrypted by default: this was wrong since this should almost
11971 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11972 the encrypted data type: this is a more sensible place to put it and it
11973 allows the PKCS#12 code to be tidied up that duplicated this
11977 *) Changed obj_dat.pl script so it takes its input and output files on
11978 the command line. This should avoid shell escape redirection problems
11982 *) Initial support for certificate extension requests, these are included
11983 in things like Xenroll certificate requests. Included functions to allow
11984 extensions to be obtained and added.
11987 *) -crlf option to s_client and s_server for sending newlines as
11988 CRLF (as required by many protocols).
11991 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11993 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11994 [Ralf S. Engelschall]
11996 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11997 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11999 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
12003 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
12004 DH parameters/keys (q is lost during that conversion, but the resulting
12005 DH parameters contain its length).
12007 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
12008 much faster than DH_generate_parameters (which creates parameters
12009 where p = 2*q + 1), and also the smaller q makes DH computations
12010 much more efficient (160-bit exponentiation instead of 1024-bit
12011 exponentiation); so this provides a convenient way to support DHE
12012 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
12013 utter importance to use
12014 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12016 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12017 when such DH parameters are used, because otherwise small subgroup
12018 attacks may become possible!
12021 *) Avoid memory leak in i2d_DHparams.
12024 *) Allow the -k option to be used more than once in the enc program:
12025 this allows the same encrypted message to be read by multiple recipients.
12028 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
12029 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
12030 it will always use the numerical form of the OID, even if it has a short
12034 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
12035 method only got called if p,q,dmp1,dmq1,iqmp components were present,
12036 otherwise bn_mod_exp was called. In the case of hardware keys for example
12037 no private key components need be present and it might store extra data
12038 in the RSA structure, which cannot be accessed from bn_mod_exp.
12039 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
12040 private key operations.
12043 *) Added support for SPARC Linux.
12046 *) pem_password_cb function type incompatibly changed from
12047 typedef int pem_password_cb(char *buf, int size, int rwflag);
12049 ....(char *buf, int size, int rwflag, void *userdata);
12050 so that applications can pass data to their callbacks:
12051 The PEM[_ASN1]_{read,write}... functions and macros now take an
12052 additional void * argument, which is just handed through whenever
12053 the password callback is called.
12054 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
12056 New function SSL_CTX_set_default_passwd_cb_userdata.
12058 Compatibility note: As many C implementations push function arguments
12059 onto the stack in reverse order, the new library version is likely to
12060 interoperate with programs that have been compiled with the old
12061 pem_password_cb definition (PEM_whatever takes some data that
12062 happens to be on the stack as its last argument, and the callback
12063 just ignores this garbage); but there is no guarantee whatsoever that
12066 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
12067 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
12068 problems not only on Windows, but also on some Unix platforms.
12069 To avoid problematic command lines, these definitions are now in an
12070 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
12071 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
12074 *) MIPS III/IV assembler module is reimplemented.
12077 *) More DES library cleanups: remove references to srand/rand and
12078 delete an unused file.
12081 *) Add support for the free Netwide assembler (NASM) under Win32,
12082 since not many people have MASM (ml) and it can be hard to obtain.
12083 This is currently experimental but it seems to work OK and pass all
12084 the tests. Check out INSTALL.W32 for info.
12087 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
12088 without temporary keys kept an extra copy of the server key,
12089 and connections with temporary keys did not free everything in case
12093 *) New function RSA_check_key and new openssl rsa option -check
12094 for verifying the consistency of RSA keys.
12095 [Ulf Moeller, Bodo Moeller]
12097 *) Various changes to make Win32 compile work:
12098 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
12099 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
12100 comparison" warnings.
12101 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
12104 *) Add a debugging option to PKCS#5 v2 key generation function: when
12105 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
12106 derived keys are printed to stderr.
12109 *) Copy the flags in ASN1_STRING_dup().
12110 [Roman E. Pavlov <pre@mo.msk.ru>]
12112 *) The x509 application mishandled signing requests containing DSA
12113 keys when the signing key was also DSA and the parameters didn't match.
12115 It was supposed to omit the parameters when they matched the signing key:
12116 the verifying software was then supposed to automatically use the CA's
12117 parameters if they were absent from the end user certificate.
12119 Omitting parameters is no longer recommended. The test was also
12120 the wrong way round! This was probably due to unusual behaviour in
12121 EVP_cmp_parameters() which returns 1 if the parameters match.
12122 This meant that parameters were omitted when they *didn't* match and
12123 the certificate was useless. Certificates signed with 'ca' didn't have
12125 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
12127 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
12128 The interface is as follows:
12129 Applications can use
12130 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
12131 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
12132 "off" is now the default.
12133 The library internally uses
12134 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
12135 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
12136 to disable memory-checking temporarily.
12138 Some inconsistent states that previously were possible (and were
12139 even the default) are now avoided.
12141 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
12142 with each memory chunk allocated; this is occasionally more helpful
12143 than just having a counter.
12145 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
12147 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
12151 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
12152 which largely parallels "options", but is for changing API behaviour,
12153 whereas "options" are about protocol behaviour.
12154 Initial "mode" flags are:
12156 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
12157 a single record has been written.
12158 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
12159 retries use the same buffer location.
12160 (But all of the contents must be
12164 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
12167 *) Fix problems with no-hmac etc.
12168 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
12170 *) New functions RSA_get_default_method(), RSA_set_method() and
12171 RSA_get_method(). These allows replacement of RSA_METHODs without having
12172 to mess around with the internals of an RSA structure.
12175 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
12176 Also really enable memory leak checks in openssl.c and in some
12178 [Chad C. Mulligan, Bodo Moeller]
12180 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
12181 up the length of negative integers. This has now been simplified to just
12182 store the length when it is first determined and use it later, rather
12183 than trying to keep track of where data is copied and updating it to
12185 [Steve Henson, reported by Brien Wheeler
12186 <bwheeler@authentica-security.com>]
12188 *) Add a new function PKCS7_signatureVerify. This allows the verification
12189 of a PKCS#7 signature but with the signing certificate passed to the
12190 function itself. This contrasts with PKCS7_dataVerify which assumes the
12191 certificate is present in the PKCS#7 structure. This isn't always the
12192 case: certificates can be omitted from a PKCS#7 structure and be
12193 distributed by "out of band" means (such as a certificate database).
12196 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12197 function prototypes in pem.h, also change util/mkdef.pl to add the
12198 necessary function names.
12201 *) mk1mf.pl (used by Windows builds) did not properly read the
12202 options set by Configure in the top level Makefile, and Configure
12203 was not even able to write more than one option correctly.
12204 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12207 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12208 file to be loaded from a BIO or FILE pointer. The BIO version will
12209 for example allow memory BIOs to contain config info.
12212 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12213 Whoever hopes to achieve shared-library compatibility across versions
12214 must use this, not the compile-time macro.
12215 (Exercise 0.9.4: Which is the minimum library version required by
12217 Note: All this applies only to multi-threaded programs, others don't
12221 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12222 through a BIO pair triggered the default case, i.e.
12223 SSLerr(...,SSL_R_UNKNOWN_STATE).
12226 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12227 can use the SSL library even if none of the specific BIOs is
12231 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12232 for the encoded length.
12233 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12235 *) Add initial documentation of the X509V3 functions.
12238 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12239 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12240 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12241 secure PKCS#8 private key format with a high iteration count.
12244 *) Fix determination of Perl interpreter: A perl or perl5
12245 _directory_ in $PATH was also accepted as the interpreter.
12246 [Ralf S. Engelschall]
12248 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12249 wrong with it but it was very old and did things like calling
12250 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12251 unusual formatting.
12254 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12255 to use the new extension code.
12258 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12259 with macros. This should make it easier to change their form, add extra
12260 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12264 *) Add to configuration table a new entry that can specify an alternative
12265 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12266 according to Mark Crispin <MRC@Panda.COM>.
12270 *) DES CBC did not update the IV. Weird.
12273 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12274 Changing the behaviour of the former might break existing programs --
12275 where IV updating is needed, des_ncbc_encrypt can be used.
12278 *) When bntest is run from "make test" it drives bc to check its
12279 calculations, as well as internally checking them. If an internal check
12280 fails, it needs to cause bc to give a non-zero result or make test carries
12281 on without noticing the failure. Fixed.
12284 *) DES library cleanups.
12287 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12288 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12289 ciphers. NOTE: although the key derivation function has been verified
12290 against some published test vectors it has not been extensively tested
12291 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12295 *) Instead of "mkdir -p", which is not fully portable, use new
12296 Perl script "util/mkdir-p.pl".
12299 *) Rewrite the way password based encryption (PBE) is handled. It used to
12300 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12301 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12302 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12303 the 'parameter' field of the AlgorithmIdentifier is passed to the
12304 underlying key generation function so it must do its own ASN1 parsing.
12305 This has also changed the EVP_PBE_CipherInit() function which now has a
12306 'parameter' argument instead of literal salt and iteration count values
12307 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12310 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12311 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12312 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12313 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12314 value was just used as a "magic string" and not used directly its
12315 value doesn't matter.
12318 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12322 *) "linux-sparc64" configuration (ultrapenguin).
12323 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12324 "linux-sparc" configuration.
12325 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12327 *) config now generates no-xxx options for missing ciphers.
12330 *) Support the EBCDIC character set (work in progress).
12331 File ebcdic.c not yet included because it has a different license.
12332 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12334 *) Support BS2000/OSD-POSIX.
12335 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12337 *) Make callbacks for key generation use void * instead of char *.
12340 *) Make S/MIME samples compile (not yet tested).
12343 *) Additional typesafe stacks.
12346 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12350 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12352 *) New configuration variant "sco5-gcc".
12354 *) Updated some demos.
12355 [Sean O Riordain, Wade Scholine]
12357 *) Add missing BIO_free at exit of pkcs12 application.
12360 *) Fix memory leak in conf.c.
12363 *) Updates for Win32 to assembler version of MD5.
12366 *) Set #! path to perl in apps/der_chop to where we found it
12367 instead of using a fixed path.
12370 *) SHA library changes for irix64-mips4-cc.
12373 *) Improvements for VMS support.
12377 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12379 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12380 This also avoids the problems with SC4.2 and unpatched SC5.
12381 [Andy Polyakov <appro@fy.chalmers.se>]
12383 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12384 These are required because of the typesafe stack would otherwise break
12385 existing code. If old code used a structure member which used to be STACK
12386 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12387 sk_num or sk_value it would produce an error because the num, data members
12388 are not present in STACK_OF. Now it just produces a warning. sk_set
12389 replaces the old method of assigning a value to sk_value
12390 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12391 that does this will no longer work (and should use sk_set instead) but
12392 this could be regarded as a "questionable" behaviour anyway.
12395 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12396 correctly handle encrypted S/MIME data.
12399 *) Change type of various DES function arguments from des_cblock
12400 (which means, in function argument declarations, pointer to char)
12401 to des_cblock * (meaning pointer to array with 8 char elements),
12402 which allows the compiler to do more typechecking; it was like
12403 that back in SSLeay, but with lots of ugly casts.
12405 Introduce new type const_des_cblock.
12408 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12409 problems: find RecipientInfo structure that matches recipient certificate
12410 and initialise the ASN1 structures properly based on passed cipher.
12413 *) Belatedly make the BN tests actually check the results.
12416 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12417 to and from BNs: it was completely broken. New compilation option
12418 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12419 key elements as negative integers.
12422 *) Reorganize and speed up MD5.
12423 [Andy Polyakov <appro@fy.chalmers.se>]
12426 [Richard Levitte <richard@levitte.org>]
12428 *) New option -out to asn1parse to allow the parsed structure to be
12429 output to a file. This is most useful when combined with the -strparse
12430 option to examine the output of things like OCTET STRINGS.
12433 *) Make SSL library a little more fool-proof by not requiring any longer
12434 that SSL_set_{accept,connect}_state be called before
12435 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12436 in many applications because usually everything *appeared* to work as
12437 intended anyway -- now it really works as intended).
12440 *) Move openssl.cnf out of lib/.
12443 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12444 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12445 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12446 [Ralf S. Engelschall]
12448 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12449 handle PKCS#7 enveloped data properly.
12450 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12452 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12453 copying pointers. The cert_st handling is changed by this in
12454 various ways (and thus what used to be known as ctx->default_cert
12455 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12456 any longer when s->cert does not give us what we need).
12457 ssl_cert_instantiate becomes obsolete by this change.
12458 As soon as we've got the new code right (possibly it already is?),
12459 we have solved a couple of bugs of the earlier code where s->cert
12460 was used as if it could not have been shared with other SSL structures.
12462 Note that using the SSL API in certain dirty ways now will result
12463 in different behaviour than observed with earlier library versions:
12464 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12465 does not influence s as it used to.
12467 In order to clean up things more thoroughly, inside SSL_SESSION
12468 we don't use CERT any longer, but a new structure SESS_CERT
12469 that holds per-session data (if available); currently, this is
12470 the peer's certificate chain and, for clients, the server's certificate
12471 and temporary key. CERT holds only those values that can have
12472 meaningful defaults in an SSL_CTX.
12475 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12476 from the internal representation. Various PKCS#7 fixes: remove some
12477 evil casts and set the enc_dig_alg field properly based on the signing
12481 *) Allow PKCS#12 password to be set from the command line or the
12482 environment. Let 'ca' get its config file name from the environment
12483 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12487 *) Allow certificate policies extension to use an IA5STRING for the
12488 organization field. This is contrary to the PKIX definition but
12489 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12493 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12494 without disallowing inline assembler and the like for non-pedantic builds.
12497 *) Support Borland C++ builder.
12498 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12500 *) Support Mingw32.
12503 *) SHA-1 cleanups and performance enhancements.
12504 [Andy Polyakov <appro@fy.chalmers.se>]
12506 *) Sparc v8plus assembler for the bignum library.
12507 [Andy Polyakov <appro@fy.chalmers.se>]
12509 *) Accept any -xxx and +xxx compiler options in Configure.
12512 *) Update HPUX configuration.
12515 *) Add missing sk_<type>_unshift() function to safestack.h
12516 [Ralf S. Engelschall]
12518 *) New function SSL_CTX_use_certificate_chain_file that sets the
12519 "extra_cert"s in addition to the certificate. (This makes sense
12520 only for "PEM" format files, as chains as a whole are not
12524 *) Support verify_depth from the SSL API.
12525 x509_vfy.c had what can be considered an off-by-one-error:
12526 Its depth (which was not part of the external interface)
12527 was actually counting the number of certificates in a chain;
12528 now it really counts the depth.
12531 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12532 instead of X509err, which often resulted in confusing error
12533 messages since the error codes are not globally unique
12534 (e.g. an alleged error in ssl3_accept when a certificate
12535 didn't match the private key).
12537 *) New function SSL_CTX_set_session_id_context that allows to set a default
12538 value (so that you don't need SSL_set_session_id_context for each
12539 connection using the SSL_CTX).
12542 *) OAEP decoding bug fix.
12545 *) Support INSTALL_PREFIX for package builders, as proposed by
12549 *) New Configure options "threads" and "no-threads". For systems
12550 where the proper compiler options are known (currently Solaris
12551 and Linux), "threads" is the default.
12554 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12557 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12558 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12559 such as /usr/local/bin.
12562 *) "make linux-shared" to build shared libraries.
12563 [Niels Poppe <niels@netbox.org>]
12565 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12568 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12569 extension adding in x509 utility.
12572 *) Remove NOPROTO sections and error code comments.
12575 *) Partial rewrite of the DEF file generator to now parse the ANSI
12579 *) New Configure options --prefix=DIR and --openssldir=DIR.
12582 *) Complete rewrite of the error code script(s). It is all now handled
12583 by one script at the top level which handles error code gathering,
12584 header rewriting and C source file generation. It should be much better
12585 than the old method: it now uses a modified version of Ulf's parser to
12586 read the ANSI prototypes in all header files (thus the old K&R definitions
12587 aren't needed for error creation any more) and do a better job of
12588 translating function codes into names. The old 'ASN1 error code embedded
12589 in a comment' is no longer necessary and it doesn't use .err files which
12590 have now been deleted. Also the error code call doesn't have to appear all
12591 on one line (which resulted in some large lines...).
12594 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12597 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12598 0 (which usually indicates a closed connection), but continue reading.
12601 *) Fix some race conditions.
12604 *) Add support for CRL distribution points extension. Add Certificate
12605 Policies and CRL distribution points documentation.
12608 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12611 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12612 8 of keying material. Merlin has also confirmed interop with this fix
12613 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12614 [Merlin Hughes <merlin@baltimore.ie>]
12616 *) Fix lots of warnings.
12617 [Richard Levitte <levitte@stacken.kth.se>]
12619 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12620 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12621 [Richard Levitte <levitte@stacken.kth.se>]
12623 *) Fix problems with sizeof(long) == 8.
12624 [Andy Polyakov <appro@fy.chalmers.se>]
12626 *) Change functions to ANSI C.
12629 *) Fix typos in error codes.
12630 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12632 *) Remove defunct assembler files from Configure.
12635 *) SPARC v8 assembler BIGNUM implementation.
12636 [Andy Polyakov <appro@fy.chalmers.se>]
12638 *) Support for Certificate Policies extension: both print and set.
12639 Various additions to support the r2i method this uses.
12642 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12643 return a const string when you are expecting an allocated buffer.
12646 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12647 types DirectoryString and DisplayText.
12650 *) Add code to allow r2i extensions to access the configuration database,
12651 add an LHASH database driver and add several ctx helper functions.
12654 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12655 fail when they extended the size of a BIGNUM.
12658 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12659 support typesafe stack.
12662 *) Fix typo in SSL_[gs]et_options().
12663 [Nils Frostberg <nils@medcom.se>]
12665 *) Delete various functions and files that belonged to the (now obsolete)
12666 old X509V3 handling code.
12669 *) New Configure option "rsaref".
12672 *) Don't auto-generate pem.h.
12675 *) Introduce type-safe ASN.1 SETs.
12678 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12679 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12681 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12682 that links with OpenSSL (well at least cause lots of warnings), but fear
12683 not: the conversion is trivial, and it eliminates loads of evil casts. A
12684 few STACKed things have been converted already. Feel free to convert more.
12685 In the fullness of time, I'll do away with the STACK type altogether.
12688 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12689 specified in <certfile> by updating the entry in the index.txt file.
12690 This way one no longer has to edit the index.txt file manually for
12691 revoking a certificate. The -revoke option does the gory details now.
12692 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12694 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12695 `-text' option at all and this way the `-noout -text' combination was
12696 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12697 [Ralf S. Engelschall]
12699 *) Make sure a corresponding plain text error message exists for the
12700 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12701 verify callback function determined that a certificate was revoked.
12702 [Ralf S. Engelschall]
12704 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12705 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12706 all available ciphers including rc5, which was forgotten until now.
12707 In order to let the testing shell script know which algorithms
12708 are available, a new (up to now undocumented) command
12709 "openssl list-cipher-commands" is used.
12712 *) Bugfix: s_client occasionally would sleep in select() when
12713 it should have checked SSL_pending() first.
12716 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12717 the raw DSA values prior to ASN.1 encoding.
12720 *) Tweaks to Configure
12721 [Niels Poppe <niels@netbox.org>]
12723 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12727 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12730 *) New config option to avoid instructions that are illegal on the 80386.
12731 The default code is faster, but requires at least a 486.
12734 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12735 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12736 same as SSL2_VERSION anyway.
12739 *) New "-showcerts" option for s_client.
12742 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12743 application. Various cleanups and fixes.
12746 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12747 modify error routines to work internally. Add error codes and PBE init
12748 to library startup routines.
12751 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12752 packing functions to asn1 and evp. Changed function names and error
12753 codes along the way.
12756 *) PKCS12 integration: and so it begins... First of several patches to
12757 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12758 objects to objects.h
12761 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12762 and display support for Thawte strong extranet extension.
12765 *) Add LinuxPPC support.
12766 [Jeff Dubrule <igor@pobox.org>]
12768 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12769 bn_div_words in alpha.s.
12770 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12772 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12773 OAEP isn't supported when OpenSSL is built with RSAref.
12774 [Ulf Moeller <ulf@fitug.de>]
12776 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12777 so they no longer are missing under -DNOPROTO.
12778 [Soren S. Jorvang <soren@t.dk>]
12781 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12783 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12784 doesn't work when the session is reused. Coming soon!
12787 *) Fix a security hole, that allows sessions to be reused in the wrong
12788 context thus bypassing client cert protection! All software that uses
12789 client certs and session caches in multiple contexts NEEDS PATCHING to
12790 allow session reuse! A fuller solution is in the works.
12791 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12793 *) Some more source tree cleanups (removed obsolete files
12794 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12795 permission on "config" script to be executable) and a fix for the INSTALL
12797 [Ulf Moeller <ulf@fitug.de>]
12799 *) Remove some legacy and erroneous uses of malloc, free instead of
12801 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12803 *) Make rsa_oaep_test return non-zero on error.
12804 [Ulf Moeller <ulf@fitug.de>]
12806 *) Add support for native Solaris shared libraries. Configure
12807 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12808 if someone would make that last step automatic.
12809 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12811 *) ctx_size was not built with the right compiler during "make links". Fixed.
12814 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12815 except NULL ciphers". This means the default cipher list will no longer
12816 enable NULL ciphers. They need to be specifically enabled e.g. with
12817 the string "DEFAULT:eNULL".
12820 *) Fix to RSA private encryption routines: if p < q then it would
12821 occasionally produce an invalid result. This will only happen with
12822 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12825 *) Be less restrictive and allow also `perl util/perlpath.pl
12826 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12827 because this way one can also use an interpreter named `perl5' (which is
12828 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12829 installed as `perl').
12830 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12832 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12833 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12835 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12836 advapi32.lib to Win32 build and change the pem test comparison
12837 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12838 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12839 and crypto/des/ede_cbcm_enc.c.
12842 *) DES quad checksum was broken on big-endian architectures. Fixed.
12845 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12846 Win32 test batch file so it (might) work again. The Win32 test batch file
12847 is horrible: I feel ill....
12850 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12851 in e_os.h. Audit of header files to check ANSI and non ANSI
12852 sections: 10 functions were absent from non ANSI section and not exported
12853 from Windows DLLs. Fixed up libeay.num for new functions.
12856 *) Make `openssl version' output lines consistent.
12857 [Ralf S. Engelschall]
12859 *) Fix Win32 symbol export lists for BIO functions: Added
12860 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12861 to ms/libeay{16,32}.def.
12862 [Ralf S. Engelschall]
12864 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12865 fine under Unix and passes some trivial tests I've now added. But the
12866 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12867 added to make sure no one expects that this stuff really works in the
12868 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12869 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12871 [Ralf S. Engelschall]
12873 *) Fix the generation of two part addresses in perl.
12874 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12876 *) Add config entry for Linux on MIPS.
12877 [John Tobey <jtobey@channel1.com>]
12879 *) Make links whenever Configure is run, unless we are on Windoze.
12882 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12883 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12887 *) Add a useful kludge to allow package maintainers to specify compiler and
12888 other platforms details on the command line without having to patch the
12889 Configure script every time: One now can use ``perl Configure
12890 <id>:<details>'', i.e. platform ids are allowed to have details appended
12891 to them (separated by colons). This is treated as there would be a static
12892 pre-configured entry in Configure's %table under key <id> with value
12893 <details> and ``perl Configure <id>'' is called. So, when you want to
12894 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12895 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12896 now, which overrides the FreeBSD-elf entry on-the-fly.
12897 [Ralf S. Engelschall]
12899 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12902 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12903 on the `perl Configure ...' command line. This way one can compile
12904 OpenSSL libraries with Position Independent Code (PIC) which is needed
12905 for linking it into DSOs.
12906 [Ralf S. Engelschall]
12908 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12912 *) Cleaned up the LICENSE document: The official contact for any license
12913 questions now is the OpenSSL core team under openssl-core@openssl.org.
12914 And add a paragraph about the dual-license situation to make sure people
12915 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12916 to the OpenSSL toolkit.
12917 [Ralf S. Engelschall]
12919 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12920 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12921 Additionally cleaned up the `make links' target: Remove unnecessary
12922 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12923 to speed processing and no longer clutter the display with confusing
12924 stuff. Instead only the actually done links are displayed.
12925 [Ralf S. Engelschall]
12927 *) Permit null encryption ciphersuites, used for authentication only. It used
12928 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12929 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12933 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12934 signed attributes when verifying signatures (this would break them),
12935 the detached data encoding was wrong and public keys obtained using
12936 X509_get_pubkey() weren't freed.
12939 *) Add text documentation for the BUFFER functions. Also added a work around
12940 to a Win95 console bug. This was triggered by the password read stuff: the
12941 last character typed gets carried over to the next fread(). If you were
12942 generating a new cert request using 'req' for example then the last
12943 character of the passphrase would be CR which would then enter the first
12947 *) Added the new `Includes OpenSSL Cryptography Software' button as
12948 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12949 button and can be used by applications based on OpenSSL to show the
12950 relationship to the OpenSSL project.
12951 [Ralf S. Engelschall]
12953 *) Remove confusing variables in function signatures in files
12954 ssl/ssl_lib.c and ssl/ssl.h.
12955 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12957 *) Don't install bss_file.c under PREFIX/include/
12958 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12960 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12961 functions that return function pointers and has support for NT specific
12962 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12963 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12964 unsigned to signed types: this was killing the Win32 compile.
12967 *) Add new certificate file to stack functions,
12968 SSL_add_dir_cert_subjects_to_stack() and
12969 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12970 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12971 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12972 This means that Apache-SSL and similar packages don't have to mess around
12973 to add as many CAs as they want to the preferred list.
12976 *) Experiment with doxygen documentation. Currently only partially applied to
12978 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12979 openssl.doxy as the configuration file.
12982 *) Get rid of remaining C++-style comments which strict C compilers hate.
12983 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12985 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12986 compiled in by default: it has problems with large keys.
12989 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12990 DH private keys and/or callback functions which directly correspond to
12991 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12992 is needed for applications which have to configure certificates on a
12993 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12995 For the RSA certificate situation is makes no difference, but
12996 for the DSA certificate situation this fixes the "no shared cipher"
12997 problem where the OpenSSL cipher selection procedure failed because the
12998 temporary keys were not overtaken from the context and the API provided
12999 no way to reconfigure them.
13000 The new functions now let applications reconfigure the stuff and they
13001 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
13002 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
13003 non-public-API function ssl_cert_instantiate() is used as a helper
13004 function and also to reduce code redundancy inside ssl_rsa.c.
13005 [Ralf S. Engelschall]
13007 *) Move s_server -dcert and -dkey options out of the undocumented feature
13008 area because they are useful for the DSA situation and should be
13009 recognized by the users.
13010 [Ralf S. Engelschall]
13012 *) Fix the cipher decision scheme for export ciphers: the export bits are
13013 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
13014 SSL_EXP_MASK. So, the original variable has to be used instead of the
13015 already masked variable.
13016 [Richard Levitte <levitte@stacken.kth.se>]
13018 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
13019 [Richard Levitte <levitte@stacken.kth.se>]
13021 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
13022 from `int' to `unsigned int' because it's a length and initialized by
13023 EVP_DigestFinal() which expects an `unsigned int *'.
13024 [Richard Levitte <levitte@stacken.kth.se>]
13026 *) Don't hard-code path to Perl interpreter on shebang line of Configure
13027 script. Instead use the usual Shell->Perl transition trick.
13028 [Ralf S. Engelschall]
13030 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
13031 (in addition to RSA certificates) to match the behaviour of `openssl dsa
13032 -noout -modulus' as it's already the case for `openssl rsa -noout
13033 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
13034 currently the public key is printed (a decision which was already done by
13035 `openssl dsa -modulus' in the past) which serves a similar purpose.
13036 Additionally the NO_RSA no longer completely removes the whole -modulus
13037 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
13039 [Ralf S. Engelschall]
13041 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
13042 BIO. See the source (crypto/evp/bio_ok.c) for more info.
13043 [Arne Ansper <arne@ats.cyber.ee>]
13045 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
13046 to be added. Now both 'req' and 'ca' can use new objects defined in the
13050 *) Add cool BIO that does syslog (or event log on NT).
13051 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
13053 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
13054 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
13055 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
13056 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
13059 *) Add preliminary config info for new extension code.
13062 *) Make RSA_NO_PADDING really use no padding.
13063 [Ulf Moeller <ulf@fitug.de>]
13065 *) Generate errors when private/public key check is done.
13068 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
13069 for some CRL extensions and new objects added.
13072 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
13073 key usage extension and fuller support for authority key id.
13076 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
13077 padding method for RSA, which is recommended for new applications in PKCS
13078 #1 v2.0 (RFC 2437, October 1998).
13079 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
13080 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
13081 against Bleichbacher's attack on RSA.
13082 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
13085 *) Updates to the new SSL compression code
13086 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13088 *) Fix so that the version number in the master secret, when passed
13089 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
13090 (because the server will not accept higher), that the version number
13091 is 0x03,0x01, not 0x03,0x00
13092 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13094 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
13095 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
13096 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
13099 *) Support for RAW extensions where an arbitrary extension can be
13100 created by including its DER encoding. See apps/openssl.cnf for
13104 *) Make sure latest Perl versions don't interpret some generated C array
13105 code as Perl array code in the crypto/err/err_genc.pl script.
13106 [Lars Weber <3weber@informatik.uni-hamburg.de>]
13108 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
13109 not many people have the assembler. Various Win32 compilation fixes and
13110 update to the INSTALL.W32 file with (hopefully) more accurate Win32
13111 build instructions.
13114 *) Modify configure script 'Configure' to automatically create crypto/date.h
13115 file under Win32 and also build pem.h from pem.org. New script
13116 util/mkfiles.pl to create the MINFO file on environments that can't do a
13117 'make files': perl util/mkfiles.pl >MINFO should work.
13120 *) Major rework of DES function declarations, in the pursuit of correctness
13121 and purity. As a result, many evil casts evaporated, and some weirdness,
13122 too. You may find this causes warnings in your code. Zapping your evil
13123 casts will probably fix them. Mostly.
13126 *) Fix for a typo in asn1.h. Bug fix to object creation script
13127 obj_dat.pl. It considered a zero in an object definition to mean
13128 "end of object": none of the objects in objects.h have any zeros
13129 so it wasn't spotted.
13130 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
13132 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
13133 Masking (CBCM). In the absence of test vectors, the best I have been able
13134 to do is check that the decrypt undoes the encrypt, so far. Send me test
13135 vectors if you have them.
13138 *) Correct calculation of key length for export ciphers (too much space was
13139 allocated for null ciphers). This has not been tested!
13142 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
13143 message is now correct (it understands "crypto" and "ssl" on its
13144 command line). There is also now an "update" option. This will update
13145 the util/ssleay.num and util/libeay.num files with any new functions.
13147 perl util/mkdef.pl crypto ssl update
13148 it will update them.
13151 *) Overhauled the Perl interface (perl/*):
13152 - ported BN stuff to OpenSSL's different BN library
13153 - made the perl/ source tree CVS-aware
13154 - renamed the package from SSLeay to OpenSSL (the files still contain
13155 their history because I've copied them in the repository)
13156 - removed obsolete files (the test scripts will be replaced
13157 by better Test::Harness variants in the future)
13158 [Ralf S. Engelschall]
13160 *) First cut for a very conservative source tree cleanup:
13161 1. merge various obsolete readme texts into doc/ssleay.txt
13162 where we collect the old documents and readme texts.
13163 2. remove the first part of files where I'm already sure that we no
13164 longer need them because of three reasons: either they are just temporary
13165 files which were left by Eric or they are preserved original files where
13166 I've verified that the diff is also available in the CVS via "cvs diff
13167 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
13168 the crypto/md/ stuff).
13169 [Ralf S. Engelschall]
13171 *) More extension code. Incomplete support for subject and issuer alt
13172 name, issuer and authority key id. Change the i2v function parameters
13173 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
13174 what that's for :-) Fix to ASN1 macro which messed up
13175 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
13178 *) Preliminary support for ENUMERATED type. This is largely copied from the
13182 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
13183 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13185 *) Make sure `make rehash' target really finds the `openssl' program.
13186 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13188 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
13189 like to hear about it if this slows down other processors.
13192 *) Add CygWin32 platform information to Configure script.
13193 [Alan Batie <batie@aahz.jf.intel.com>]
13195 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
13196 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13198 *) New program nseq to manipulate netscape certificate sequences
13201 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13205 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13206 but the BN code had some problems that would cause failures when
13207 doing certificate verification and some other functions.
13208 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13210 *) Add ASN1 and PEM code to support netscape certificate sequences.
13213 *) Add ASN1 and PEM code to support netscape certificate sequences.
13216 *) Add several PKIX and private extended key usage OIDs.
13219 *) Modify the 'ca' program to handle the new extension code. Modify
13220 openssl.cnf for new extension format, add comments.
13223 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13224 and add a sample to openssl.cnf so req -x509 now adds appropriate
13228 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13229 error code, add initial support to X509_print() and x509 application.
13232 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13233 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13234 stuff is currently isolated and isn't even compiled yet.
13237 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13238 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13239 Removed the versions check from X509 routines when loading extensions:
13240 this allows certain broken certificates that don't set the version
13241 properly to be processed.
13244 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13245 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13246 can still be regenerated with "make depend".
13249 *) Spelling mistake in C version of CAST-128.
13250 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13252 *) Changes to the error generation code. The perl script err-code.pl
13253 now reads in the old error codes and retains the old numbers, only
13254 adding new ones if necessary. It also only changes the .err files if new
13255 codes are added. The makefiles have been modified to only insert errors
13256 when needed (to avoid needlessly modifying header files). This is done
13257 by only inserting errors if the .err file is newer than the auto generated
13258 C file. To rebuild all the error codes from scratch (the old behaviour)
13259 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13260 or delete all the .err files.
13263 *) CAST-128 was incorrectly implemented for short keys. The C version has
13264 been fixed, but is untested. The assembler versions are also fixed, but
13265 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13266 to regenerate it if needed.
13267 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13268 Hagino <itojun@kame.net>]
13270 *) File was opened incorrectly in randfile.c.
13271 [Ulf Möller <ulf@fitug.de>]
13273 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13274 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13275 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13276 al: it's just almost always a UTCTime. Note this patch adds new error
13277 codes so do a "make errors" if there are problems.
13280 *) Correct Linux 1 recognition in config.
13281 [Ulf Möller <ulf@fitug.de>]
13283 *) Remove pointless MD5 hash when using DSA keys in ca.
13284 [Anonymous <nobody@replay.com>]
13286 *) Generate an error if given an empty string as a cert directory. Also
13287 generate an error if handed NULL (previously returned 0 to indicate an
13288 error, but didn't set one).
13289 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13291 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13294 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13295 parameters. This was causing a warning which killed off the Win32 compile.
13298 *) Remove C++ style comments from crypto/bn/bn_local.h.
13299 [Neil Costigan <neil.costigan@celocom.com>]
13301 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13302 based on a text string, looking up short and long names and finally
13303 "dot" format. The "dot" format stuff didn't work. Added new function
13304 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13305 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13306 OID is not part of the table.
13309 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13310 X509_LOOKUP_by_alias().
13313 *) Sort openssl functions by name.
13316 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13317 encryption from sample DSA keys (in case anyone is interested the password
13321 *) Make _all_ *_free functions accept a NULL pointer.
13322 [Frans Heymans <fheymans@isaserver.be>]
13324 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13326 [Anonymous <nobody@replay.com>]
13328 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13329 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13331 *) Don't blow it for numeric -newkey arguments to apps/req.
13332 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13334 *) Temp key "for export" tests were wrong in s3_srvr.c.
13335 [Anonymous <nobody@replay.com>]
13337 *) Add prototype for temp key callback functions
13338 SSL_CTX_set_tmp_{rsa,dh}_callback().
13341 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13342 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13345 *) X509_name_add_entry() freed the wrong thing after an error.
13346 [Arne Ansper <arne@ats.cyber.ee>]
13348 *) rsa_eay.c would attempt to free a NULL context.
13349 [Arne Ansper <arne@ats.cyber.ee>]
13351 *) BIO_s_socket() had a broken should_retry() on Windoze.
13352 [Arne Ansper <arne@ats.cyber.ee>]
13354 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13355 [Arne Ansper <arne@ats.cyber.ee>]
13357 *) Make sure the already existing X509_STORE->depth variable is initialized
13358 in X509_STORE_new(), but document the fact that this variable is still
13359 unused in the certificate verification process.
13360 [Ralf S. Engelschall]
13362 *) Fix the various library and apps files to free up pkeys obtained from
13363 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13366 *) Fix reference counting in X509_PUBKEY_get(). This makes
13367 demos/maurice/example2.c work, amongst others, probably.
13368 [Steve Henson and Ben Laurie]
13370 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13371 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13372 are no longer created. This way we have a single and consistent command
13373 line interface `openssl <command>', similar to `cvs <command>'.
13374 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13376 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13377 BIT STRING wrapper always have zero unused bits.
13380 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13383 *) Make the top-level INSTALL documentation easier to understand.
13386 *) Makefiles updated to exit if an error occurs in a sub-directory
13387 make (including if user presses ^C) [Paul Sutton]
13389 *) Make Montgomery context stuff explicit in RSA data structure.
13392 *) Fix build order of pem and err to allow for generated pem.h.
13395 *) Fix renumbering bug in X509_NAME_delete_entry().
13398 *) Enhanced the err-ins.pl script so it makes the error library number
13399 global and can add a library name. This is needed for external ASN1 and
13400 other error libraries.
13403 *) Fixed sk_insert which never worked properly.
13406 *) Fix ASN1 macros so they can handle indefinite length constructed
13407 EXPLICIT tags. Some non standard certificates use these: they can now
13411 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13412 into a single doc/ssleay.txt bundle. This way the information is still
13413 preserved but no longer messes up this directory. Now it's new room for
13414 the new set of documentation files.
13415 [Ralf S. Engelschall]
13417 *) SETs were incorrectly DER encoded. This was a major pain, because they
13418 shared code with SEQUENCEs, which aren't coded the same. This means that
13419 almost everything to do with SETs or SEQUENCEs has either changed name or
13420 number of arguments.
13421 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13423 *) Fix test data to work with the above.
13426 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13427 was already fixed by Eric for 0.9.1 it seems.
13428 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13430 *) Autodetect FreeBSD3.
13433 *) Fix various bugs in Configure. This affects the following platforms:
13437 unixware-2.0-pentium
13441 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13442 before they are needed.
13445 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13449 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13451 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13452 changed SSLeay to OpenSSL in version strings.
13453 [Ralf S. Engelschall]
13455 *) Some fixups to the top-level documents.
13458 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13459 because the symlink to include/ was missing.
13460 [Ralf S. Engelschall]
13462 *) Incorporated the popular no-RSA/DSA-only patches
13463 which allow to compile a RSA-free SSLeay.
13464 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13466 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13467 when "ssleay" is still not found.
13468 [Ralf S. Engelschall]
13470 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13471 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13473 *) Updated the README file.
13474 [Ralf S. Engelschall]
13476 *) Added various .cvsignore files in the CVS repository subdirs
13477 to make a "cvs update" really silent.
13478 [Ralf S. Engelschall]
13480 *) Recompiled the error-definition header files and added
13481 missing symbols to the Win32 linker tables.
13482 [Ralf S. Engelschall]
13484 *) Cleaned up the top-level documents;
13485 o new files: CHANGES and LICENSE
13486 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13487 o merged COPYRIGHT into LICENSE
13488 o removed obsolete TODO file
13489 o renamed MICROSOFT to INSTALL.W32
13490 [Ralf S. Engelschall]
13492 *) Removed dummy files from the 0.9.1b source tree:
13493 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13494 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13495 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13496 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13497 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13498 [Ralf S. Engelschall]
13500 *) Added various platform portability fixes.
13503 *) The Genesis of the OpenSSL rpject:
13504 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13505 Young and Tim J. Hudson created while they were working for C2Net until
13507 [The OpenSSL Project]
13510 Changes between 0.9.0b and 0.9.1b [not released]
13512 *) Updated a few CA certificates under certs/
13515 *) Changed some BIGNUM api stuff.
13518 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13519 DGUX x86, Linux Alpha, etc.
13522 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13523 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13527 *) Add -strparse option to asn1pars program which parses nested
13529 [Dr Stephen Henson <shenson@bigfoot.com>]
13531 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13534 *) DSA fix for "ca" program.
13537 *) Added "-genkey" option to "dsaparam" program.
13540 *) Added RIPE MD160 (rmd160) message digest.
13543 *) Added -a (all) option to "ssleay version" command.
13546 *) Added PLATFORM define which is the id given to Configure.
13549 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13552 *) Extended the ASN.1 parser routines.
13555 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13558 *) Added a BN_CTX to the BN library.
13561 *) Fixed the weak key values in DES library
13564 *) Changed API in EVP library for cipher aliases.
13567 *) Added support for RC2/64bit cipher.
13570 *) Converted the lhash library to the crypto/mem.c functions.
13573 *) Added more recognized ASN.1 object ids.
13576 *) Added more RSA padding checks for SSL/TLS.
13579 *) Added BIO proxy/filter functionality.
13582 *) Added extra_certs to SSL_CTX which can be used
13583 send extra CA certificates to the client in the CA cert chain sending
13584 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13587 *) Now Fortezza is denied in the authentication phase because
13588 this is key exchange mechanism is not supported by SSLeay at all.
13591 *) Additional PKCS1 checks.
13594 *) Support the string "TLSv1" for all TLS v1 ciphers.
13597 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13598 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13601 *) Fixed a few memory leaks.
13604 *) Fixed various code and comment typos.
13607 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13608 bytes sent in the client random.
13609 [Edward Bishop <ebishop@spyglass.com>]