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 1.1.1p and 1.1.1q [xx XXX xxxx]
14 Changes between 1.1.1o and 1.1.1p [21 Jun 2022]
16 *) In addition to the c_rehash shell command injection identified in
17 CVE-2022-1292, further bugs where the c_rehash script does not
18 properly sanitise shell metacharacters to prevent command injection have been
21 When the CVE-2022-1292 was fixed it was not discovered that there
22 are other places in the script where the file names of certificates
23 being hashed were possibly passed to a command executed through the shell.
25 This script is distributed by some operating systems in a manner where
26 it is automatically executed. On such operating systems, an attacker
27 could execute arbitrary commands with the privileges of the script.
29 Use of the c_rehash script is considered obsolete and should be replaced
30 by the OpenSSL rehash command line tool.
32 [Daniel Fiala, Tomáš Mráz]
34 *) When OpenSSL TLS client is connecting without any supported elliptic
35 curves and TLS-1.3 protocol is disabled the connection will no longer fail
36 if a ciphersuite that does not use a key exchange based on elliptic
37 curves can be negotiated.
40 Changes between 1.1.1n and 1.1.1o [3 May 2022]
42 *) Fixed a bug in the c_rehash script which was not properly sanitising shell
43 metacharacters to prevent command injection. This script is distributed
44 by some operating systems in a manner where it is automatically executed.
45 On such operating systems, an attacker could execute arbitrary commands
46 with the privileges of the script.
48 Use of the c_rehash script is considered obsolete and should be replaced
49 by the OpenSSL rehash command line tool.
53 Changes between 1.1.1m and 1.1.1n [15 Mar 2022]
55 *) Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
58 Internally this function is used when parsing certificates that contain
59 elliptic curve public keys in compressed form or explicit elliptic curve
60 parameters with a base point encoded in compressed form.
62 It is possible to trigger the infinite loop by crafting a certificate that
63 has invalid explicit curve parameters.
65 Since certificate parsing happens prior to verification of the certificate
66 signature, any process that parses an externally supplied certificate may
67 thus be subject to a denial of service attack. The infinite loop can also
68 be reached when parsing crafted private keys as they can contain explicit
69 elliptic curve parameters.
71 Thus vulnerable situations include:
73 - TLS clients consuming server certificates
74 - TLS servers consuming client certificates
75 - Hosting providers taking certificates or private keys from customers
76 - Certificate authorities parsing certification requests from subscribers
77 - Anything else which parses ASN.1 elliptic curve parameters
79 Also any other applications that use the BN_mod_sqrt() where the attacker
80 can control the parameter values are vulnerable to this DoS issue.
84 *) Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
85 to the list of ciphersuites providing Perfect Forward Secrecy as
86 required by SECLEVEL >= 3.
88 [Dmitry Belyavskiy, Nicola Tuveri]
90 Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
92 *) Avoid loading of a dynamic engine twice.
96 *) Fixed building on Debian with kfreebsd kernels
100 *) Prioritise DANE TLSA issuer certs over peer certs
104 *) Fixed random API for MacOS prior to 10.12
106 These MacOS versions don't support the CommonCrypto APIs
110 Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
112 *) Fixed an SM2 Decryption Buffer Overflow.
114 In order to decrypt SM2 encrypted data an application is expected to call the
115 API function EVP_PKEY_decrypt(). Typically an application will call this
116 function twice. The first time, on entry, the "out" parameter can be NULL and,
117 on exit, the "outlen" parameter is populated with the buffer size required to
118 hold the decrypted plaintext. The application can then allocate a sufficiently
119 sized buffer and call EVP_PKEY_decrypt() again, but this time passing a non-NULL
120 value for the "out" parameter.
122 A bug in the implementation of the SM2 decryption code means that the
123 calculation of the buffer size required to hold the plaintext returned by the
124 first call to EVP_PKEY_decrypt() can be smaller than the actual size required by
125 the second call. This can lead to a buffer overflow when EVP_PKEY_decrypt() is
126 called by the application a second time with a buffer that is too small.
128 A malicious attacker who is able present SM2 content for decryption to an
129 application could cause attacker chosen data to overflow the buffer by up to a
130 maximum of 62 bytes altering the contents of other data held after the
131 buffer, possibly changing application behaviour or causing the application to
132 crash. The location of the buffer is application dependent but is typically
137 *) Fixed various read buffer overruns processing ASN.1 strings
139 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
140 structure which contains a buffer holding the string data and a field holding
141 the buffer length. This contrasts with normal C strings which are repesented as
142 a buffer for the string data which is terminated with a NUL (0) byte.
144 Although not a strict requirement, ASN.1 strings that are parsed using OpenSSL's
145 own "d2i" functions (and other similar parsing functions) as well as any string
146 whose value has been set with the ASN1_STRING_set() function will additionally
147 NUL terminate the byte array in the ASN1_STRING structure.
149 However, it is possible for applications to directly construct valid ASN1_STRING
150 structures which do not NUL terminate the byte array by directly setting the
151 "data" and "length" fields in the ASN1_STRING array. This can also happen by
152 using the ASN1_STRING_set0() function.
154 Numerous OpenSSL functions that print ASN.1 data have been found to assume that
155 the ASN1_STRING byte array will be NUL terminated, even though this is not
156 guaranteed for strings that have been directly constructed. Where an application
157 requests an ASN.1 structure to be printed, and where that ASN.1 structure
158 contains ASN1_STRINGs that have been directly constructed by the application
159 without NUL terminating the "data" field, then a read buffer overrun can occur.
161 The same thing can also occur during name constraints processing of certificates
162 (for example if a certificate has been directly constructed by the application
163 instead of loading it via the OpenSSL parsing functions, and the certificate
164 contains non NUL terminated ASN1_STRING structures). It can also occur in the
165 X509_get1_email(), X509_REQ_get1_email() and X509_get1_ocsp() functions.
167 If a malicious actor can cause an application to directly construct an
168 ASN1_STRING and then process it through one of the affected OpenSSL functions
169 then this issue could be hit. This might result in a crash (causing a Denial of
170 Service attack). It could also result in the disclosure of private memory
171 contents (such as private keys, or sensitive plaintext).
175 Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
177 *) Fixed a problem with verifying a certificate chain when using the
178 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks
179 of the certificates present in a certificate chain. It is not set by
182 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
183 the chain that have explicitly encoded elliptic curve parameters was added
184 as an additional strict check.
186 An error in the implementation of this check meant that the result of a
187 previous check to confirm that certificates in the chain are valid CA
188 certificates was overwritten. This effectively bypasses the check
189 that non-CA certificates must not be able to issue other certificates.
191 If a "purpose" has been configured then there is a subsequent opportunity
192 for checks that the certificate is a valid CA. All of the named "purpose"
193 values implemented in libcrypto perform this check. Therefore, where
194 a purpose is set the certificate chain will still be rejected even when the
195 strict flag has been used. A purpose is set by default in libssl client and
196 server certificate verification routines, but it can be overridden or
197 removed by an application.
199 In order to be affected, an application must explicitly set the
200 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
201 for the certificate verification or, in the case of TLS client or server
202 applications, override the default purpose.
206 *) Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
207 crafted renegotiation ClientHello message from a client. If a TLSv1.2
208 renegotiation ClientHello omits the signature_algorithms extension (where
209 it was present in the initial ClientHello), but includes a
210 signature_algorithms_cert extension then a NULL pointer dereference will
211 result, leading to a crash and a denial of service attack.
213 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
214 (which is the default configuration). OpenSSL TLS clients are not impacted
217 [Peter Kästle and Samuel Sapalski]
219 Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
221 *) Fixed the X509_issuer_and_serial_hash() function. It attempts to
222 create a unique hash value based on the issuer and serial number data
223 contained within an X509 certificate. However it was failing to correctly
224 handle any errors that may occur while parsing the issuer field (which might
225 occur if the issuer field is maliciously constructed). This may subsequently
226 result in a NULL pointer deref and a crash leading to a potential denial of
231 *) Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
232 padding mode to correctly check for rollback attacks. This is considered a
233 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
237 *) Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
238 functions. Previously they could overflow the output length argument in some
239 cases where the input length is close to the maximum permissable length for
240 an integer on the platform. In such cases the return value from the function
241 call would be 1 (indicating success), but the output length value would be
242 negative. This could cause applications to behave incorrectly or crash.
246 *) Fixed SRP_Calc_client_key so that it runs in constant time. The previous
247 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
248 could be exploited in a side channel attack to recover the password. Since
249 the attack is local host only this is outside of the current OpenSSL
250 threat model and therefore no CVE is assigned.
252 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
256 Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
258 *) Fixed NULL pointer deref in the GENERAL_NAME_cmp function
259 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
260 If an attacker can control both items being compared then this could lead
261 to a possible denial of service attack. OpenSSL itself uses the
262 GENERAL_NAME_cmp function for two purposes:
263 1) Comparing CRL distribution point names between an available CRL and a
264 CRL distribution point embedded in an X509 certificate
265 2) When verifying that a timestamp response token signer matches the
266 timestamp authority name (exposed via the API functions
267 TS_RESP_verify_response and TS_RESP_verify_token)
271 *) Add support for Apple Silicon M1 Macs with the darwin64-arm64-cc target.
274 *) The security callback, which can be customised by application code, supports
275 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
276 in the "other" parameter. In most places this is what is passed. All these
277 places occur server side. However there was one client side call of this
278 security operation and it passed a DH object instead. This is incorrect
279 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
280 of the other locations. Therefore this client side call has been changed to
281 pass an EVP_PKEY instead.
284 *) In 1.1.1h, an expired trusted (root) certificate was not anymore rejected
285 when validating a certificate path. This check is restored in 1.1.1i.
288 Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
290 *) Certificates with explicit curve parameters are now disallowed in
291 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
294 *) The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
295 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
296 conversely, silently ignore DTLS protocol version bounds when configuring
297 TLS-based contexts. The commands can be repeated to set bounds of both
298 types. The same applies with the corresponding "min_protocol" and
299 "max_protocol" command-line switches, in case some application uses both TLS
302 SSL_CTX instances that are created for a fixed protocol version (e.g.
303 TLSv1_server_method()) also silently ignore version bounds. Previously
304 attempts to apply bounds to these protocol versions would result in an
305 error. Now only the "version-flexible" SSL_CTX instances are subject to
306 limits in configuration files in command-line options.
309 *) Handshake now fails if Extended Master Secret extension is dropped
313 *) Accidentally, an expired trusted (root) certificate is not anymore rejected
314 when validating a certificate path.
317 *) The Oracle Developer Studio compiler will start reporting deprecated APIs
319 Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
321 *) Fixed segmentation fault in SSL_check_chain()
322 Server or client applications that call the SSL_check_chain() function
323 during or after a TLS 1.3 handshake may crash due to a NULL pointer
324 dereference as a result of incorrect handling of the
325 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
326 or unrecognised signature algorithm is received from the peer. This could
327 be exploited by a malicious peer in a Denial of Service attack.
331 *) Added AES consttime code for no-asm configurations
332 an optional constant time support for AES was added
333 when building openssl for no-asm.
334 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
335 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
336 At this time this feature is by default disabled.
337 It will be enabled by default in 3.0.
340 Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
342 *) Revert the change of EOF detection while reading in libssl to avoid
343 regressions in applications depending on the current way of reporting
344 the EOF. As the existing method is not fully accurate the change to
345 reporting the EOF via SSL_ERROR_SSL is kept on the current development
346 branch and will be present in the 3.0 release.
349 *) Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
350 when primes for RSA keys are computed.
351 Since we previously always generated primes == 2 (mod 3) for RSA keys,
352 the 2-prime and 3-prime RSA modules were easy to distinguish, since
353 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
354 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
355 This avoids possible fingerprinting of newly generated RSA modules.
358 Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
359 *) Properly detect EOF while reading in libssl. Previously if we hit an EOF
360 while reading in libssl then we would report an error back to the
361 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
362 an error to the stack (which means we instead return SSL_ERROR_SSL) and
363 therefore give a hint as to what went wrong.
366 *) Check that ed25519 and ed448 are allowed by the security level. Previously
367 signature algorithms not using an MD were not being checked that they were
368 allowed by the security level.
371 *) Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
372 was not quite right. The behaviour was not consistent between resumption
373 and normal handshakes, and also not quite consistent with historical
374 behaviour. The behaviour in various scenarios has been clarified and
375 it has been updated to make it match historical behaviour as closely as
379 *) [VMS only] The header files that the VMS compilers include automatically,
380 __DECC_INCLUDE_PROLOGUE.H and __DECC_INCLUDE_EPILOGUE.H, use pragmas that
381 the C++ compiler doesn't understand. This is a shortcoming in the
382 compiler, but can be worked around with __cplusplus guards.
384 C++ applications that use OpenSSL libraries must be compiled using the
385 qualifier '/NAMES=(AS_IS,SHORTENED)' to be able to use all the OpenSSL
386 functions. Otherwise, only functions with symbols of less than 31
387 characters can be used, as the linker will not be able to successfully
388 resolve symbols with longer names.
391 *) Corrected the documentation of the return values from the EVP_DigestSign*
392 set of functions. The documentation mentioned negative values for some
393 errors, but this was never the case, so the mention of negative values
396 Code that followed the documentation and thereby check with something
397 like 'EVP_DigestSignInit(...) <= 0' will continue to work undisturbed.
400 *) Fixed an an overflow bug in the x64_64 Montgomery squaring procedure
401 used in exponentiation with 512-bit moduli. No EC algorithms are
402 affected. Analysis suggests that attacks against 2-prime RSA1024,
403 3-prime RSA1536, and DSA1024 as a result of this defect would be very
404 difficult to perform and are not believed likely. Attacks against DH512
405 are considered just feasible. However, for an attack the target would
406 have to re-use the DH512 private key, which is not recommended anyway.
407 Also applications directly using the low level API BN_mod_exp may be
408 affected if they use BN_FLG_CONSTTIME.
412 *) Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
413 The presence of this system service is determined at run-time.
416 *) Added newline escaping functionality to a filename when using openssl dgst.
417 This output format is to replicate the output format found in the '*sum'
418 checksum programs. This aims to preserve backward compatibility.
419 [Matt Eaton, Richard Levitte, and Paul Dale]
421 *) Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
425 Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
427 *) Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
428 number generator (RNG). This was intended to include protection in the
429 event of a fork() system call in order to ensure that the parent and child
430 processes did not share the same RNG state. However this protection was not
431 being used in the default case.
433 A partial mitigation for this issue is that the output from a high
434 precision timer is mixed into the RNG state so the likelihood of a parent
435 and child process sharing state is significantly reduced.
437 If an application already calls OPENSSL_init_crypto() explicitly using
438 OPENSSL_INIT_ATFORK then this problem does not occur at all.
440 [Matthias St. Pierre]
442 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is
443 used even when parsing explicit parameters, when loading a serialized key
444 or calling `EC_GROUP_new_from_ecpkparameters()`/
445 `EC_GROUP_new_from_ecparameters()`.
446 This prevents bypass of security hardening and performance gains,
447 especially for curves with specialized EC_METHODs.
448 By default, if a key encoded with explicit parameters is loaded and later
449 serialized, the output is still encoded with explicit parameters, even if
450 internally a "named" EC_GROUP is used for computation.
453 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before
454 this change, EC_GROUP_set_generator would accept order and/or cofactor as
455 NULL. After this change, only the cofactor parameter can be NULL. It also
456 does some minimal sanity checks on the passed order.
460 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
461 An attack is simple, if the first CMS_recipientInfo is valid but the
462 second CMS_recipientInfo is chosen ciphertext. If the second
463 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
464 encryption key will be replaced by garbage, and the message cannot be
465 decoded, but if the RSA decryption fails, the correct encryption key is
466 used and the recipient will not notice the attack.
467 As a work around for this potential attack the length of the decrypted
468 key must be equal to the cipher default key length, in case the
469 certifiate is not given and all recipientInfo are tried out.
470 The old behaviour can be re-enabled in the CMS code by setting the
471 CMS_DEBUG_DECRYPT flag.
475 *) Early start up entropy quality from the DEVRANDOM seed source has been
476 improved for older Linux systems. The RAND subsystem will wait for
477 /dev/random to be producing output before seeding from /dev/urandom.
478 The seeded state is stored for future library initialisations using
479 a system global shared memory segment. The shared memory identifier
480 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
481 the desired value. The default identifier is 114.
484 *) Correct the extended master secret constant on EBCDIC systems. Without this
485 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
486 negotiate EMS will fail. Unfortunately this also means that TLS connections
487 between EBCDIC systems with this fix, and EBCDIC systems without this
488 fix will fail if they negotiate EMS.
491 *) Use Windows installation paths in the mingw builds
493 Mingw isn't a POSIX environment per se, which means that Windows
494 paths should be used for installation.
498 *) Changed DH_check to accept parameters with order q and 2q subgroups.
499 With order 2q subgroups the bit 0 of the private key is not secret
500 but DH_generate_key works around that by clearing bit 0 of the
501 private key for those. This avoids leaking bit 0 of the private key.
504 *) Significantly reduce secure memory usage by the randomness pools.
507 *) Revert the DEVRANDOM_WAIT feature for Linux systems
509 The DEVRANDOM_WAIT feature added a select() call to wait for the
510 /dev/random device to become readable before reading from the
513 It turned out that this change had negative side effects on
514 performance which were not acceptable. After some discussion it
515 was decided to revert this feature and leave it up to the OS
516 resp. the platform maintainer to ensure a proper initialization
517 during early boot time.
518 [Matthias St. Pierre]
520 Changes between 1.1.1b and 1.1.1c [28 May 2019]
522 *) Add build tests for C++. These are generated files that only do one
523 thing, to include one public OpenSSL head file each. This tests that
524 the public header files can be usefully included in a C++ application.
526 This test isn't enabled by default. It can be enabled with the option
527 'enable-buildtest-c++'.
530 *) Enable SHA3 pre-hashing for ECDSA and DSA.
533 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
534 This changes the size when using the genpkey app when no size is given. It
535 fixes an omission in earlier changes that changed all RSA, DSA and DH
536 generation apps to use 2048 bits by default.
539 *) Reorganize the manual pages to consistently have RETURN VALUES,
540 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
541 util/fix-doc-nits accordingly.
542 [Paul Yang, Joshua Lock]
544 *) Add the missing accessor EVP_PKEY_get0_engine()
547 *) Have apps like 's_client' and 's_server' output the signature scheme
548 along with other cipher suite parameters when debugging.
551 *) Make OPENSSL_config() error agnostic again.
554 *) Do the error handling in RSA decryption constant time.
557 *) Prevent over long nonces in ChaCha20-Poly1305.
559 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
560 for every encryption operation. RFC 7539 specifies that the nonce value
561 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
562 and front pads the nonce with 0 bytes if it is less than 12
563 bytes. However it also incorrectly allows a nonce to be set of up to 16
564 bytes. In this case only the last 12 bytes are significant and any
565 additional leading bytes are ignored.
567 It is a requirement of using this cipher that nonce values are
568 unique. Messages encrypted using a reused nonce value are susceptible to
569 serious confidentiality and integrity attacks. If an application changes
570 the default nonce length to be longer than 12 bytes and then makes a
571 change to the leading bytes of the nonce expecting the new value to be a
572 new unique nonce then such an application could inadvertently encrypt
573 messages with a reused nonce.
575 Additionally the ignored bytes in a long nonce are not covered by the
576 integrity guarantee of this cipher. Any application that relies on the
577 integrity of these ignored leading bytes of a long nonce may be further
578 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
579 is safe because no such use sets such a long nonce value. However user
580 applications that use this cipher directly and set a non-default nonce
581 length to be longer than 12 bytes may be vulnerable.
583 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
588 *) Add DEVRANDOM_WAIT feature for Linux systems
590 On older Linux systems where the getrandom() system call is not available,
591 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
592 Contrary to getrandom(), the /dev/urandom device will not block during
593 early boot when the kernel CSPRNG has not been seeded yet.
595 To mitigate this known weakness, use select() to wait for /dev/random to
596 become readable before reading from /dev/urandom.
598 *) Ensure that SM2 only uses SM3 as digest algorithm
601 Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
603 *) Added SCA hardening for modular field inversion in EC_GROUP through
604 a new dedicated field_inv() pointer in EC_METHOD.
605 This also addresses a leakage affecting conversions from projective
606 to affine coordinates.
607 [Billy Bob Brumley, Nicola Tuveri]
609 *) Change the info callback signals for the start and end of a post-handshake
610 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
611 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
612 confused by this and assume that a TLSv1.2 renegotiation has started. This
613 can break KeyUpdate handling. Instead we no longer signal the start and end
614 of a post handshake message exchange (although the messages themselves are
615 still signalled). This could break some applications that were expecting
616 the old signals. However without this KeyUpdate is not usable for many
620 *) Fix a bug in the computation of the endpoint-pair shared secret used
621 by DTLS over SCTP. This breaks interoperability with older versions
622 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
623 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
624 interoperability with such broken implementations. However, enabling
625 this switch breaks interoperability with correct implementations.
627 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
628 re-used X509_PUBKEY object if the second PUBKEY is malformed.
631 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
634 *) Remove the 'dist' target and add a tarball building script. The
635 'dist' target has fallen out of use, and it shouldn't be
636 necessary to configure just to create a source distribution.
639 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
641 *) Timing vulnerability in DSA signature generation
643 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
644 timing side channel attack. An attacker could use variations in the signing
645 algorithm to recover the private key.
647 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
651 *) Timing vulnerability in ECDSA signature generation
653 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
654 timing side channel attack. An attacker could use variations in the signing
655 algorithm to recover the private key.
657 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
661 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
662 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
663 are retained for backwards compatibility.
666 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
667 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
668 of two gigabytes and the error handling improved.
670 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
671 categorized as a normal bug, not a security issue, because the DRBG reseeds
672 automatically and is fully functional even without additional randomness
673 provided by the application.
675 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
677 *) Add a new ClientHello callback. Provides a callback interface that gives
678 the application the ability to adjust the nascent SSL object at the
679 earliest stage of ClientHello processing, immediately after extensions have
680 been collected but before they have been processed. In particular, this
681 callback can adjust the supported TLS versions in response to the contents
685 *) Add SM2 base algorithm support.
688 *) s390x assembly pack: add (improved) hardware-support for the following
689 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
690 aes-cfb/cfb8, aes-ecb.
693 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
694 parameter is no longer accepted, as it leads to a corrupt table. NULL
695 pem_str is reserved for alias entries only.
698 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
699 step for prime curves. The new implementation is based on formulae from
700 differential addition-and-doubling in homogeneous projective coordinates
701 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
702 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
703 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
704 to work in projective coordinates.
705 [Billy Bob Brumley, Nicola Tuveri]
707 *) Change generating and checking of primes so that the error rate of not
708 being prime depends on the intended use based on the size of the input.
709 For larger primes this will result in more rounds of Miller-Rabin.
710 The maximal error rate for primes with more than 1080 bits is lowered
712 [Kurt Roeckx, Annie Yousar]
714 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
717 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
718 moving between systems, and to avoid confusion when a Windows build is
719 done with mingw vs with MSVC. For POSIX installs, there's still a
720 symlink or copy named 'tsget' to avoid that confusion as well.
723 *) Revert blinding in ECDSA sign and instead make problematic addition
724 length-invariant. Switch even to fixed-length Montgomery multiplication.
727 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
728 step for binary curves. The new implementation is based on formulae from
729 differential addition-and-doubling in mixed Lopez-Dahab projective
730 coordinates, modified to independently blind the operands.
731 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
733 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
734 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
735 EC_METHODs to implement their own specialized "ladder step", to take
736 advantage of more favorable coordinate systems or more efficient
737 differential addition-and-doubling algorithms.
738 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
740 *) Modified the random device based seed sources to keep the relevant
741 file descriptors open rather than reopening them on each access.
742 This allows such sources to operate in a chroot() jail without
743 the associated device nodes being available. This behaviour can be
744 controlled using RAND_keep_random_devices_open().
747 *) Numerous side-channel attack mitigations have been applied. This may have
748 performance impacts for some algorithms for the benefit of improved
749 security. Specific changes are noted in this change log by their respective
753 *) AIX shared library support overhaul. Switch to AIX "natural" way of
754 handling shared libraries, which means collecting shared objects of
755 different versions and bitnesses in one common archive. This allows to
756 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
757 doesn't affect the way 3rd party applications are linked, only how
758 multi-version installation is managed.
761 *) Make ec_group_do_inverse_ord() more robust and available to other
762 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
763 mitigations are applied to the fallback BN_mod_inverse().
764 When using this function rather than BN_mod_inverse() directly, new
765 EC cryptosystem implementations are then safer-by-default.
768 *) Add coordinate blinding for EC_POINT and implement projective
769 coordinate blinding for generic prime curves as a countermeasure to
770 chosen point SCA attacks.
771 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
773 *) Add blinding to ECDSA and DSA signatures to protect against side channel
774 attacks discovered by Keegan Ryan (NCC Group).
777 *) Enforce checking in the pkeyutl command line app to ensure that the input
778 length does not exceed the maximum supported digest length when performing
779 a sign, verify or verifyrecover operation.
782 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
783 I/O in combination with something like select() or poll() will hang. This
784 can be turned off again using SSL_CTX_clear_mode().
785 Many applications do not properly handle non-application data records, and
786 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
787 around the problems in those applications, but can also break some.
788 It's recommended to read the manpages about SSL_read(), SSL_write(),
789 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
790 SSL_CTX_set_read_ahead() again.
793 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
794 now allow empty (zero character) pass phrases.
797 *) Apply blinding to binary field modular inversion and remove patent
798 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
801 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
802 binary and prime elliptic curves.
805 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
806 constant time fixed point multiplication.
809 *) Revise elliptic curve scalar multiplication with timing attack
810 defenses: ec_wNAF_mul redirects to a constant time implementation
811 when computing fixed point and variable point multiplication (which
812 in OpenSSL are mostly used with secret scalars in keygen, sign,
813 ECDH derive operations).
814 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
817 *) Updated CONTRIBUTING
820 *) Updated DRBG / RAND to request nonce and additional low entropy
821 randomness from the system.
822 [Matthias St. Pierre]
824 *) Updated 'openssl rehash' to use OpenSSL consistent default.
827 *) Moved the load of the ssl_conf module to libcrypto, which helps
828 loading engines that libssl uses before libssl is initialised.
831 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
834 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
835 [Ingo Schwarze, Rich Salz]
837 *) Added output of accepting IP address and port for 'openssl s_server'
840 *) Added a new API for TLSv1.3 ciphersuites:
841 SSL_CTX_set_ciphersuites()
842 SSL_set_ciphersuites()
845 *) Memory allocation failures consistently add an error to the error
849 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
850 in libcrypto when run as setuid/setgid.
853 *) Load any config file by default when libssl is used.
856 *) Added new public header file <openssl/rand_drbg.h> and documentation
857 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
858 [Matthias St. Pierre]
860 *) QNX support removed (cannot find contributors to get their approval
861 for the license change).
864 *) TLSv1.3 replay protection for early data has been implemented. See the
865 SSL_read_early_data() man page for further details.
868 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
869 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
870 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
871 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
872 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
873 configuration has been separated out. See the ciphers man page or the
874 SSL_CTX_set_ciphersuites() man page for more information.
877 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
878 in responder mode now supports the new "-multi" option, which
879 spawns the specified number of child processes to handle OCSP
880 requests. The "-timeout" option now also limits the OCSP
881 responder's patience to wait to receive the full client request
882 on a newly accepted connection. Child processes are respawned
883 as needed, and the CA index file is automatically reloaded
884 when changed. This makes it possible to run the "ocsp" responder
885 as a long-running service, making the OpenSSL CA somewhat more
886 feature-complete. In this mode, most diagnostic messages logged
887 after entering the event loop are logged via syslog(3) rather than
891 *) Added support for X448 and Ed448. Heavily based on original work by
895 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
896 objects loaded. This adds the functions OSSL_STORE_expect() and
897 OSSL_STORE_find() as well as needed tools to construct searches and
898 get the search data out of them.
901 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
902 version of OpenSSL should review their configuration settings to ensure
903 that they are still appropriate for TLSv1.3. For further information see:
904 https://wiki.openssl.org/index.php/TLS1.3
907 *) Grand redesign of the OpenSSL random generator
909 The default RAND method now utilizes an AES-CTR DRBG according to
910 NIST standard SP 800-90Ar1. The new random generator is essentially
911 a port of the default random generator from the OpenSSL FIPS 2.0
912 object module. It is a hybrid deterministic random bit generator
913 using an AES-CTR bit stream and which seeds and reseeds itself
914 automatically using trusted system entropy sources.
916 Some of its new features are:
917 o Support for multiple DRBG instances with seed chaining.
918 o The default RAND method makes use of a DRBG.
919 o There is a public and private DRBG instance.
920 o The DRBG instances are fork-safe.
921 o Keep all global DRBG instances on the secure heap if it is enabled.
922 o The public and private DRBG instance are per thread for lock free
924 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
926 *) Changed Configure so it only says what it does and doesn't dump
927 so much data. Instead, ./configdata.pm should be used as a script
928 to display all sorts of configuration data.
931 *) Added processing of "make variables" to Configure.
934 *) Added SHA512/224 and SHA512/256 algorithm support.
937 *) The last traces of Netware support, first removed in 1.1.0, have
941 *) Get rid of Makefile.shared, and in the process, make the processing
942 of certain files (rc.obj, or the .def/.map/.opt files produced from
943 the ordinal files) more visible and hopefully easier to trace and
944 debug (or make silent).
947 *) Make it possible to have environment variable assignments as
948 arguments to config / Configure.
951 *) Add multi-prime RSA (RFC 8017) support.
954 *) Add SM3 implemented according to GB/T 32905-2016
955 [ Jack Lloyd <jack.lloyd@ribose.com>,
956 Ronald Tse <ronald.tse@ribose.com>,
957 Erick Borsboom <erick.borsboom@ribose.com> ]
959 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
960 as documented in RFC6066.
961 Based on a patch from Tomasz Moń
962 [Filipe Raimundo da Silva]
964 *) Add SM4 implemented according to GB/T 32907-2016.
965 [ Jack Lloyd <jack.lloyd@ribose.com>,
966 Ronald Tse <ronald.tse@ribose.com>,
967 Erick Borsboom <erick.borsboom@ribose.com> ]
969 *) Reimplement -newreq-nodes and ERR_error_string_n; the
970 original author does not agree with the license change.
973 *) Add ARIA AEAD TLS support.
976 *) Some macro definitions to support VS6 have been removed. Visual
977 Studio 6 has not worked since 1.1.0
980 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
981 without clearing the errors.
984 *) Add "atfork" functions. If building on a system that without
985 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
986 requirements. The RAND facility now uses/requires this.
992 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
993 not possible to disable entirely. However, it's still possible to
994 disable the console reading UI method, UI_OpenSSL() (use UI_null()
997 To disable, configure with 'no-ui-console'. 'no-ui' is still
998 possible to use as an alias. Check at compile time with the
999 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1000 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1003 *) Add a STORE module, which implements a uniform and URI based reader of
1004 stores that can contain keys, certificates, CRLs and numerous other
1005 objects. The main API is loosely based on a few stdio functions,
1006 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1007 OSSL_STORE_error and OSSL_STORE_close.
1008 The implementation uses backends called "loaders" to implement arbitrary
1009 URI schemes. There is one built in "loader" for the 'file' scheme.
1012 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
1013 then adjusted to work on FreeBSD 8.4 as well.
1014 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1015 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1018 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
1019 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1020 error code calls like this:
1022 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1024 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1025 that can be encoded in C. For the foreseeable future, this will only
1027 [Richard Levitte and Tim Hudson]
1029 *) Removed BSD cryptodev engine.
1032 *) Add a build target 'build_all_generated', to build all generated files
1033 and only that. This can be used to prepare everything that requires
1034 things like perl for a system that lacks perl and then move everything
1035 to that system and do the rest of the build there.
1038 *) In the UI interface, make it possible to duplicate the user data. This
1039 can be used by engines that need to retain the data for a longer time
1040 than just the call where this user data is passed.
1043 *) Ignore the '-named_curve auto' value for compatibility of applications
1045 [Tomas Mraz <tmraz@fedoraproject.org>]
1047 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1048 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1049 alerts across multiple records (some of which could be empty). In practice
1050 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1051 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1052 support this at all. Supporting it adds significant complexity to the
1053 record layer, and its removal is unlikely to cause interoperability
1057 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1058 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1059 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1063 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1064 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1065 [Richard Levitte, Andy Polyakov]
1067 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1071 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1072 platform rather than 'mingw'.
1075 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
1076 success if they are asked to add an object which already exists
1077 in the store. This change cascades to other functions which load
1078 certificates and CRLs.
1081 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
1082 facilitate stack unwinding even from assembly subroutines.
1085 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1086 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1089 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1090 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1091 which is the minimum version we support.
1094 *) Certificate time validation (X509_cmp_time) enforces stricter
1095 compliance with RFC 5280. Fractional seconds and timezone offsets
1096 are no longer allowed.
1099 *) Add support for ARIA
1102 *) s_client will now send the Server Name Indication (SNI) extension by
1103 default unless the new "-noservername" option is used. The server name is
1104 based on the host provided to the "-connect" option unless overridden by
1105 using "-servername".
1108 *) Add support for SipHash
1111 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1112 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1113 prevent issues where no progress is being made and the peer continually
1114 sends unrecognised record types, using up resources processing them.
1117 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
1118 using the algorithm defined in
1119 https://www.akkadia.org/drepper/SHA-crypt.txt
1122 *) Heartbeat support has been removed; the ABI is changed for now.
1123 [Richard Levitte, Rich Salz]
1125 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1128 *) The RSA "null" method, which was partially supported to avoid patent
1129 issues, has been replaced to always returns NULL.
1133 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
1135 *) Client DoS due to large DH parameter
1137 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1138 malicious server can send a very large prime value to the client. This will
1139 cause the client to spend an unreasonably long period of time generating a
1140 key for this prime resulting in a hang until the client has finished. This
1141 could be exploited in a Denial Of Service attack.
1143 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1147 *) Cache timing vulnerability in RSA Key Generation
1149 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1150 a cache timing side channel attack. An attacker with sufficient access to
1151 mount cache timing attacks during the RSA key generation process could
1152 recover the private key.
1154 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1155 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1159 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1160 parameter is no longer accepted, as it leads to a corrupt table. NULL
1161 pem_str is reserved for alias entries only.
1164 *) Revert blinding in ECDSA sign and instead make problematic addition
1165 length-invariant. Switch even to fixed-length Montgomery multiplication.
1168 *) Change generating and checking of primes so that the error rate of not
1169 being prime depends on the intended use based on the size of the input.
1170 For larger primes this will result in more rounds of Miller-Rabin.
1171 The maximal error rate for primes with more than 1080 bits is lowered
1173 [Kurt Roeckx, Annie Yousar]
1175 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1178 *) Add blinding to ECDSA and DSA signatures to protect against side channel
1179 attacks discovered by Keegan Ryan (NCC Group).
1182 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1183 now allow empty (zero character) pass phrases.
1186 *) Certificate time validation (X509_cmp_time) enforces stricter
1187 compliance with RFC 5280. Fractional seconds and timezone offsets
1188 are no longer allowed.
1191 *) Fixed a text canonicalisation bug in CMS
1193 Where a CMS detached signature is used with text content the text goes
1194 through a canonicalisation process first prior to signing or verifying a
1195 signature. This process strips trailing space at the end of lines, converts
1196 line terminators to CRLF and removes additional trailing line terminators
1197 at the end of a file. A bug in the canonicalisation process meant that
1198 some characters, such as form-feed, were incorrectly treated as whitespace
1199 and removed. This is contrary to the specification (RFC5485). This fix
1200 could mean that detached text data signed with an earlier version of
1201 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
1202 signed with a fixed OpenSSL may fail to verify with an earlier version of
1203 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
1204 and use the "-binary" flag (for the "cms" command line application) or set
1205 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
1208 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
1210 *) Constructed ASN.1 types with a recursive definition could exceed the stack
1212 Constructed ASN.1 types with a recursive definition (such as can be found
1213 in PKCS7) could eventually exceed the stack given malicious input with
1214 excessive recursion. This could result in a Denial Of Service attack. There
1215 are no such structures used within SSL/TLS that come from untrusted sources
1216 so this is considered safe.
1218 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
1223 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
1225 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
1226 effectively reduced to only comparing the least significant bit of each
1227 byte. This allows an attacker to forge messages that would be considered as
1228 authenticated in an amount of tries lower than that guaranteed by the
1229 security claims of the scheme. The module can only be compiled by the
1230 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
1232 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
1237 *) Add a build target 'build_all_generated', to build all generated files
1238 and only that. This can be used to prepare everything that requires
1239 things like perl for a system that lacks perl and then move everything
1240 to that system and do the rest of the build there.
1243 *) Backport SSL_OP_NO_RENGOTIATION
1245 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
1246 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
1247 changes this is no longer possible in 1.1.0. Therefore the new
1248 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
1249 1.1.0 to provide equivalent functionality.
1251 Note that if an application built against 1.1.0h headers (or above) is run
1252 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
1253 accepted but nothing will happen, i.e. renegotiation will not be prevented.
1256 *) Removed the OS390-Unix config target. It relied on a script that doesn't
1260 *) rsaz_1024_mul_avx2 overflow bug on x86_64
1262 There is an overflow bug in the AVX2 Montgomery multiplication procedure
1263 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
1264 Analysis suggests that attacks against RSA and DSA as a result of this
1265 defect would be very difficult to perform and are not believed likely.
1266 Attacks against DH1024 are considered just feasible, because most of the
1267 work necessary to deduce information about a private key may be performed
1268 offline. The amount of resources required for such an attack would be
1269 significant. However, for an attack on TLS to be meaningful, the server
1270 would have to share the DH1024 private key among multiple clients, which is
1271 no longer an option since CVE-2016-0701.
1273 This only affects processors that support the AVX2 but not ADX extensions
1274 like Intel Haswell (4th generation).
1276 This issue was reported to OpenSSL by David Benjamin (Google). The issue
1277 was originally found via the OSS-Fuzz project.
1281 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
1283 *) bn_sqrx8x_internal carry bug on x86_64
1285 There is a carry propagating bug in the x86_64 Montgomery squaring
1286 procedure. No EC algorithms are affected. Analysis suggests that attacks
1287 against RSA and DSA as a result of this defect would be very difficult to
1288 perform and are not believed likely. Attacks against DH are considered just
1289 feasible (although very difficult) because most of the work necessary to
1290 deduce information about a private key may be performed offline. The amount
1291 of resources required for such an attack would be very significant and
1292 likely only accessible to a limited number of attackers. An attacker would
1293 additionally need online access to an unpatched system using the target
1294 private key in a scenario with persistent DH parameters and a private
1295 key that is shared between multiple clients.
1297 This only affects processors that support the BMI1, BMI2 and ADX extensions
1298 like Intel Broadwell (5th generation) and later or AMD Ryzen.
1300 This issue was reported to OpenSSL by the OSS-Fuzz project.
1304 *) Malformed X.509 IPAddressFamily could cause OOB read
1306 If an X.509 certificate has a malformed IPAddressFamily extension,
1307 OpenSSL could do a one-byte buffer overread. The most likely result
1308 would be an erroneous display of the certificate in text format.
1310 This issue was reported to OpenSSL by the OSS-Fuzz project.
1314 Changes between 1.1.0e and 1.1.0f [25 May 2017]
1316 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1317 platform rather than 'mingw'.
1320 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1321 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1322 which is the minimum version we support.
1325 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
1327 *) Encrypt-Then-Mac renegotiation crash
1329 During a renegotiation handshake if the Encrypt-Then-Mac extension is
1330 negotiated where it was not in the original handshake (or vice-versa) then
1331 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
1332 and servers are affected.
1334 This issue was reported to OpenSSL by Joe Orton (Red Hat).
1338 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
1340 *) Truncated packet could crash via OOB read
1342 If one side of an SSL/TLS path is running on a 32-bit host and a specific
1343 cipher is being used, then a truncated packet can cause that host to
1344 perform an out-of-bounds read, usually resulting in a crash.
1346 This issue was reported to OpenSSL by Robert Święcki of Google.
1350 *) Bad (EC)DHE parameters cause a client crash
1352 If a malicious server supplies bad parameters for a DHE or ECDHE key
1353 exchange then this can result in the client attempting to dereference a
1354 NULL pointer leading to a client crash. This could be exploited in a Denial
1357 This issue was reported to OpenSSL by Guido Vranken.
1361 *) BN_mod_exp may produce incorrect results on x86_64
1363 There is a carry propagating bug in the x86_64 Montgomery squaring
1364 procedure. No EC algorithms are affected. Analysis suggests that attacks
1365 against RSA and DSA as a result of this defect would be very difficult to
1366 perform and are not believed likely. Attacks against DH are considered just
1367 feasible (although very difficult) because most of the work necessary to
1368 deduce information about a private key may be performed offline. The amount
1369 of resources required for such an attack would be very significant and
1370 likely only accessible to a limited number of attackers. An attacker would
1371 additionally need online access to an unpatched system using the target
1372 private key in a scenario with persistent DH parameters and a private
1373 key that is shared between multiple clients. For example this can occur by
1374 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1375 similar to CVE-2015-3193 but must be treated as a separate problem.
1377 This issue was reported to OpenSSL by the OSS-Fuzz project.
1381 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1383 *) ChaCha20/Poly1305 heap-buffer-overflow
1385 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1386 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1387 crash. This issue is not considered to be exploitable beyond a DoS.
1389 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1393 *) CMS Null dereference
1395 Applications parsing invalid CMS structures can crash with a NULL pointer
1396 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1397 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1398 structure callback if an attempt is made to free certain invalid encodings.
1399 Only CHOICE structures using a callback which do not handle NULL value are
1402 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1406 *) Montgomery multiplication may produce incorrect results
1408 There is a carry propagating bug in the Broadwell-specific Montgomery
1409 multiplication procedure that handles input lengths divisible by, but
1410 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1411 and DH private keys are impossible. This is because the subroutine in
1412 question is not used in operations with the private key itself and an input
1413 of the attacker's direct choice. Otherwise the bug can manifest itself as
1414 transient authentication and key negotiation failures or reproducible
1415 erroneous outcome of public-key operations with specially crafted input.
1416 Among EC algorithms only Brainpool P-512 curves are affected and one
1417 presumably can attack ECDH key negotiation. Impact was not analyzed in
1418 detail, because pre-requisites for attack are considered unlikely. Namely
1419 multiple clients have to choose the curve in question and the server has to
1420 share the private key among them, neither of which is default behaviour.
1421 Even then only clients that chose the curve will be affected.
1423 This issue was publicly reported as transient failures and was not
1424 initially recognized as a security issue. Thanks to Richard Morgan for
1425 providing reproducible case.
1429 *) Removed automatic addition of RPATH in shared libraries and executables,
1430 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1433 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1435 *) Fix Use After Free for large message sizes
1437 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1438 message larger than approx 16k is received then the underlying buffer to
1439 store the incoming message is reallocated and moved. Unfortunately a
1440 dangling pointer to the old location is left which results in an attempt to
1441 write to the previously freed location. This is likely to result in a
1442 crash, however it could potentially lead to execution of arbitrary code.
1444 This issue only affects OpenSSL 1.1.0a.
1446 This issue was reported to OpenSSL by Robert Święcki.
1450 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1452 *) OCSP Status Request extension unbounded memory growth
1454 A malicious client can send an excessively large OCSP Status Request
1455 extension. If that client continually requests renegotiation, sending a
1456 large OCSP Status Request extension each time, then there will be unbounded
1457 memory growth on the server. This will eventually lead to a Denial Of
1458 Service attack through memory exhaustion. Servers with a default
1459 configuration are vulnerable even if they do not support OCSP. Builds using
1460 the "no-ocsp" build time option are not affected.
1462 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1466 *) SSL_peek() hang on empty record
1468 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1469 sends an empty record. This could be exploited by a malicious peer in a
1470 Denial Of Service attack.
1472 This issue was reported to OpenSSL by Alex Gaynor.
1476 *) Excessive allocation of memory in tls_get_message_header() and
1477 dtls1_preprocess_fragment()
1479 A (D)TLS message includes 3 bytes for its length in the header for the
1480 message. This would allow for messages up to 16Mb in length. Messages of
1481 this length are excessive and OpenSSL includes a check to ensure that a
1482 peer is sending reasonably sized messages in order to avoid too much memory
1483 being consumed to service a connection. A flaw in the logic of version
1484 1.1.0 means that memory for the message is allocated too early, prior to
1485 the excessive message length check. Due to way memory is allocated in
1486 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1487 to service a connection. This could lead to a Denial of Service through
1488 memory exhaustion. However, the excessive message length check still takes
1489 place, and this would cause the connection to immediately fail. Assuming
1490 that the application calls SSL_free() on the failed connection in a timely
1491 manner then the 21Mb of allocated memory will then be immediately freed
1492 again. Therefore the excessive memory allocation will be transitory in
1493 nature. This then means that there is only a security impact if:
1495 1) The application does not call SSL_free() in a timely manner in the event
1496 that the connection fails
1498 2) The application is working in a constrained environment where there is
1499 very little free memory
1501 3) The attacker initiates multiple connection attempts such that there are
1502 multiple connections in a state where memory has been allocated for the
1503 connection; SSL_free() has not yet been called; and there is insufficient
1504 memory to service the multiple requests.
1506 Except in the instance of (1) above any Denial Of Service is likely to be
1507 transitory because as soon as the connection fails the memory is
1508 subsequently freed again in the SSL_free() call. However there is an
1509 increased risk during this period of application crashes due to the lack of
1510 memory - which would then mean a more serious Denial of Service.
1512 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1513 (CVE-2016-6307 and CVE-2016-6308)
1516 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1517 had to be removed. Primary reason is that vendor assembler can't
1518 assemble our modules with -KPIC flag. As result it, assembly
1519 support, was not even available as option. But its lack means
1520 lack of side-channel resistant code, which is incompatible with
1521 security by todays standards. Fortunately gcc is readily available
1522 prepackaged option, which we firmly point at...
1525 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1527 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1528 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1529 (to any value) allows Windows user to access PKCS#12 file generated
1530 with Windows CryptoAPI and protected with non-ASCII password, as well
1531 as files generated under UTF-8 locale on Linux also protected with
1535 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1536 have been disabled by default and removed from DEFAULT, just like RC4.
1537 See the RC4 item below to re-enable both.
1540 *) The method for finding the storage location for the Windows RAND seed file
1541 has changed. First we check %RANDFILE%. If that is not set then we check
1542 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1543 all else fails we fall back to C:\.
1546 *) The EVP_EncryptUpdate() function has had its return type changed from void
1547 to int. A return of 0 indicates and error while a return of 1 indicates
1551 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1552 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1553 off the constant time implementation for RSA, DSA and DH have been made
1554 no-ops and deprecated.
1557 *) Windows RAND implementation was simplified to only get entropy by
1558 calling CryptGenRandom(). Various other RAND-related tickets
1560 [Joseph Wylie Yandle, Rich Salz]
1562 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1563 and OPENSSL_LH_, respectively. The old names are available
1564 with API compatibility. They new names are now completely documented.
1567 *) Unify TYPE_up_ref(obj) methods signature.
1568 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1569 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1570 int (instead of void) like all others TYPE_up_ref() methods.
1571 So now these methods also check the return value of CRYPTO_atomic_add(),
1572 and the validity of object reference counter.
1573 [fdasilvayy@gmail.com]
1575 *) With Windows Visual Studio builds, the .pdb files are installed
1576 alongside the installed libraries and executables. For a static
1577 library installation, ossl_static.pdb is the associate compiler
1578 generated .pdb file to be used when linking programs.
1581 *) Remove openssl.spec. Packaging files belong with the packagers.
1584 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1585 recognise x86_64 architectures automatically. You can still decide
1586 to build for a different bitness with the environment variable
1587 KERNEL_BITS (can be 32 or 64), for example:
1589 KERNEL_BITS=32 ./config
1593 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1594 256 bit AES and HMAC with SHA256.
1597 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1600 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1603 *) To enable users to have their own config files and build file templates,
1604 Configure looks in the directory indicated by the environment variable
1605 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1606 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1607 name and is used as is.
1610 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1611 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1612 X509_CERT_FILE_CTX was removed.
1615 *) "shared" builds are now the default. To create only static libraries use
1616 the "no-shared" Configure option.
1619 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1620 All of these option have not worked for some while and are fundamental
1624 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1625 global cleanup functions are no longer required because they are handled
1626 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1627 Explicitly de-initing can cause problems (e.g. where a library that uses
1628 OpenSSL de-inits, but an application is still using it). The affected
1629 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1630 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1631 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1632 COMP_zlib_cleanup().
1635 *) --strict-warnings no longer enables runtime debugging options
1636 such as REF_DEBUG. Instead, debug options are automatically
1637 enabled with '--debug' builds.
1638 [Andy Polyakov, Emilia Käsper]
1640 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1641 have been moved out of the public header files. New functions for managing
1642 these have been added.
1645 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1646 objects have been moved out of the public header files. New
1647 functions for managing these have been added.
1650 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1651 have been moved out of the public header files. New functions for managing
1652 these have been added.
1655 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1656 moved out of the public header files. New functions for managing these
1660 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1663 *) Removed the mk1mf build scripts.
1666 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1667 it is always safe to #include a header now.
1670 *) Removed the aged BC-32 config and all its supporting scripts
1673 *) Removed support for Ultrix, Netware, and OS/2.
1676 *) Add support for HKDF.
1677 [Alessandro Ghedini]
1679 *) Add support for blake2b and blake2s
1682 *) Added support for "pipelining". Ciphers that have the
1683 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1684 encryptions/decryptions simultaneously. There are currently no built-in
1685 ciphers with this property but the expectation is that engines will be able
1686 to offer it to significantly improve throughput. Support has been extended
1687 into libssl so that multiple records for a single connection can be
1688 processed in one go (for >=TLS 1.1).
1691 *) Added the AFALG engine. This is an async capable engine which is able to
1692 offload work to the Linux kernel. In this initial version it only supports
1693 AES128-CBC. The kernel must be version 4.1.0 or greater.
1696 *) OpenSSL now uses a new threading API. It is no longer necessary to
1697 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1698 are two supported threading models: pthreads and windows threads. It is
1699 also possible to configure OpenSSL at compile time for "no-threads". The
1700 old threading API should no longer be used. The functions have been
1701 replaced with "no-op" compatibility macros.
1702 [Alessandro Ghedini, Matt Caswell]
1704 *) Modify behavior of ALPN to invoke callback after SNI/servername
1705 callback, such that updates to the SSL_CTX affect ALPN.
1708 *) Add SSL_CIPHER queries for authentication and key-exchange.
1711 *) Changes to the DEFAULT cipherlist:
1712 - Prefer (EC)DHE handshakes over plain RSA.
1713 - Prefer AEAD ciphers over legacy ciphers.
1714 - Prefer ECDSA over RSA when both certificates are available.
1715 - Prefer TLSv1.2 ciphers/PRF.
1716 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1720 *) Change the ECC default curve list to be this, in order: x25519,
1721 secp256r1, secp521r1, secp384r1.
1724 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1725 disabled by default. They can be re-enabled using the
1726 enable-weak-ssl-ciphers option to Configure.
1729 *) If the server has ALPN configured, but supports no protocols that the
1730 client advertises, send a fatal "no_application_protocol" alert.
1731 This behaviour is SHALL in RFC 7301, though it isn't universally
1732 implemented by other servers.
1735 *) Add X25519 support.
1736 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1737 for public and private key encoding using the format documented in
1738 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1739 key generation and key derivation.
1741 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1745 *) Deprecate SRP_VBASE_get_by_user.
1746 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1747 In order to fix an unavoidable memory leak (CVE-2016-0798),
1748 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1749 seed, even if the seed is configured.
1751 Users should use SRP_VBASE_get1_by_user instead. Note that in
1752 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1753 also that even though configuring the SRP seed attempts to hide
1754 invalid usernames by continuing the handshake with fake
1755 credentials, this behaviour is not constant time and no strong
1756 guarantees are made that the handshake is indistinguishable from
1757 that of a valid user.
1760 *) Configuration change; it's now possible to build dynamic engines
1761 without having to build shared libraries and vice versa. This
1762 only applies to the engines in engines/, those in crypto/engine/
1763 will always be built into libcrypto (i.e. "static").
1765 Building dynamic engines is enabled by default; to disable, use
1766 the configuration option "disable-dynamic-engine".
1768 The only requirements for building dynamic engines are the
1769 presence of the DSO module and building with position independent
1770 code, so they will also automatically be disabled if configuring
1771 with "disable-dso" or "disable-pic".
1773 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1774 are also taken away from openssl/opensslconf.h, as they are
1778 *) Configuration change; if there is a known flag to compile
1779 position independent code, it will always be applied on the
1780 libcrypto and libssl object files, and never on the application
1781 object files. This means other libraries that use routines from
1782 libcrypto / libssl can be made into shared libraries regardless
1783 of how OpenSSL was configured.
1785 If this isn't desirable, the configuration options "disable-pic"
1786 or "no-pic" can be used to disable the use of PIC. This will
1787 also disable building shared libraries and dynamic engines.
1790 *) Removed JPAKE code. It was experimental and has no wide use.
1793 *) The INSTALL_PREFIX Makefile variable has been renamed to
1794 DESTDIR. That makes for less confusion on what this variable
1795 is for. Also, the configuration option --install_prefix is
1799 *) Heartbeat for TLS has been removed and is disabled by default
1800 for DTLS; configure with enable-heartbeats. Code that uses the
1801 old #define's might need to be updated.
1802 [Emilia Käsper, Rich Salz]
1804 *) Rename REF_CHECK to REF_DEBUG.
1807 *) New "unified" build system
1809 The "unified" build system is aimed to be a common system for all
1810 platforms we support. With it comes new support for VMS.
1812 This system builds supports building in a different directory tree
1813 than the source tree. It produces one Makefile (for unix family
1814 or lookalikes), or one descrip.mms (for VMS).
1816 The source of information to make the Makefile / descrip.mms is
1817 small files called 'build.info', holding the necessary
1818 information for each directory with source to compile, and a
1819 template in Configurations, like unix-Makefile.tmpl or
1822 With this change, the library names were also renamed on Windows
1823 and on VMS. They now have names that are closer to the standard
1824 on Unix, and include the major version number, and in certain
1825 cases, the architecture they are built for. See "Notes on shared
1826 libraries" in INSTALL.
1828 We rely heavily on the perl module Text::Template.
1831 *) Added support for auto-initialisation and de-initialisation of the library.
1832 OpenSSL no longer requires explicit init or deinit routines to be called,
1833 except in certain circumstances. See the OPENSSL_init_crypto() and
1834 OPENSSL_init_ssl() man pages for further information.
1837 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1838 "peer" argument is now expected to be a BIO_ADDR object.
1840 *) Rewrite of BIO networking library. The BIO library lacked consistent
1841 support of IPv6, and adding it required some more extensive
1842 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1843 which hold all types of addresses and chains of address information.
1844 It also introduces a new API, with functions like BIO_socket,
1845 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1846 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1847 have been adapted accordingly.
1850 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1854 *) CRIME protection: disable compression by default, even if OpenSSL is
1855 compiled with zlib enabled. Applications can still enable compression
1856 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1857 using the SSL_CONF library to configure compression.
1860 *) The signature of the session callback configured with
1861 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1862 was explicitly marked as 'const unsigned char*' instead of
1866 *) Always DPURIFY. Remove the use of uninitialized memory in the
1867 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1870 *) Removed many obsolete configuration items, including
1871 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1872 MD2_CHAR, MD2_INT, MD2_LONG
1874 IDEA_SHORT, IDEA_LONG
1875 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1876 [Rich Salz, with advice from Andy Polyakov]
1878 *) Many BN internals have been moved to an internal header file.
1879 [Rich Salz with help from Andy Polyakov]
1881 *) Configuration and writing out the results from it has changed.
1882 Files such as Makefile include/openssl/opensslconf.h and are now
1883 produced through general templates, such as Makefile.in and
1884 crypto/opensslconf.h.in and some help from the perl module
1887 Also, the center of configuration information is no longer
1888 Makefile. Instead, Configure produces a perl module in
1889 configdata.pm which holds most of the config data (in the hash
1890 table %config), the target data that comes from the target
1891 configuration in one of the Configurations/*.conf files (in
1895 *) To clarify their intended purposes, the Configure options
1896 --prefix and --openssldir change their semantics, and become more
1897 straightforward and less interdependent.
1899 --prefix shall be used exclusively to give the location INSTALLTOP
1900 where programs, scripts, libraries, include files and manuals are
1901 going to be installed. The default is now /usr/local.
1903 --openssldir shall be used exclusively to give the default
1904 location OPENSSLDIR where certificates, private keys, CRLs are
1905 managed. This is also where the default openssl.cnf gets
1907 If the directory given with this option is a relative path, the
1908 values of both the --prefix value and the --openssldir value will
1909 be combined to become OPENSSLDIR.
1910 The default for --openssldir is INSTALLTOP/ssl.
1912 Anyone who uses --openssldir to specify where OpenSSL is to be
1913 installed MUST change to use --prefix instead.
1916 *) The GOST engine was out of date and therefore it has been removed. An up
1917 to date GOST engine is now being maintained in an external repository.
1918 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1919 support for GOST ciphersuites (these are only activated if a GOST engine
1923 *) EGD is no longer supported by default; use enable-egd when
1925 [Ben Kaduk and Rich Salz]
1927 *) The distribution now has Makefile.in files, which are used to
1928 create Makefile's when Configure is run. *Configure must be run
1929 before trying to build now.*
1932 *) The return value for SSL_CIPHER_description() for error conditions
1936 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1938 Obtaining and performing DNSSEC validation of TLSA records is
1939 the application's responsibility. The application provides
1940 the TLSA records of its choice to OpenSSL, and these are then
1941 used to authenticate the peer.
1943 The TLSA records need not even come from DNS. They can, for
1944 example, be used to implement local end-entity certificate or
1945 trust-anchor "pinning", where the "pin" data takes the form
1946 of TLSA records, which can augment or replace verification
1947 based on the usual WebPKI public certification authorities.
1950 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1951 continues to support deprecated interfaces in default builds.
1952 However, applications are strongly advised to compile their
1953 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1954 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1955 or the 1.1.0 releases.
1957 In environments in which all applications have been ported to
1958 not use any deprecated interfaces OpenSSL's Configure script
1959 should be used with the --api=1.1.0 option to entirely remove
1960 support for the deprecated features from the library and
1961 unconditionally disable them in the installed headers.
1962 Essentially the same effect can be achieved with the "no-deprecated"
1963 argument to Configure, except that this will always restrict
1964 the build to just the latest API, rather than a fixed API
1967 As applications are ported to future revisions of the API,
1968 they should update their compile-time OPENSSL_API_COMPAT define
1969 accordingly, but in most cases should be able to continue to
1970 compile with later releases.
1972 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1973 0x10000000L and 0x00908000L, respectively. However those
1974 versions did not support the OPENSSL_API_COMPAT feature, and
1975 so applications are not typically tested for explicit support
1976 of just the undeprecated features of either release.
1979 *) Add support for setting the minimum and maximum supported protocol.
1980 It can bet set via the SSL_set_min_proto_version() and
1981 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1982 MaxProtocol. It's recommended to use the new APIs to disable
1983 protocols instead of disabling individual protocols using
1984 SSL_set_options() or SSL_CONF's Protocol. This change also
1985 removes support for disabling TLS 1.2 in the OpenSSL TLS
1986 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1989 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1992 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1993 and integrates ECDSA and ECDH functionality into EC. Implementations can
1994 now redirect key generation and no longer need to convert to or from
1997 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1998 include the ec.h header file instead.
2001 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
2002 ciphers who are no longer supported and drops support the ephemeral RSA key
2003 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2006 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2007 opaque. For HMAC_CTX, the following constructors and destructors
2010 HMAC_CTX *HMAC_CTX_new(void);
2011 void HMAC_CTX_free(HMAC_CTX *ctx);
2013 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2014 destroy such methods has been added. See EVP_MD_meth_new(3) and
2015 EVP_CIPHER_meth_new(3) for documentation.
2018 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2019 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2020 EVP_MD_CTX_reset() should be called instead to reinitialise
2021 an already created structure.
2022 2) For consistency with the majority of our object creators and
2023 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2024 EVP_MD_CTX_(new|free). The old names are retained as macros
2025 for deprecated builds.
2028 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
2029 cryptographic operations to be performed asynchronously as long as an
2030 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2031 further details. Libssl has also had this capability integrated with the
2032 introduction of the new mode SSL_MODE_ASYNC and associated error
2033 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2034 pages. This work was developed in partnership with Intel Corp.
2037 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2038 always enabled now. If you want to disable the support you should
2039 exclude it using the list of supported ciphers. This also means that the
2040 "-no_ecdhe" option has been removed from s_server.
2043 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2044 SSL_{CTX_}set1_curves() which can set a list.
2047 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2048 curve you want to support using SSL_{CTX_}set1_curves().
2051 *) State machine rewrite. The state machine code has been significantly
2052 refactored in order to remove much duplication of code and solve issues
2053 with the old code (see ssl/statem/README for further details). This change
2054 does have some associated API changes. Notably the SSL_state() function
2055 has been removed and replaced by SSL_get_state which now returns an
2056 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
2057 altogether. The previous handshake states defined in ssl.h and ssl3.h have
2061 *) All instances of the string "ssleay" in the public API were replaced
2062 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
2063 Some error codes related to internal RSA_eay API's were renamed.
2066 *) The demo files in crypto/threads were moved to demo/threads.
2069 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
2071 [Matt Caswell, Rich Salz]
2073 *) New ASN.1 embed macro.
2075 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
2076 structure is not allocated: it is part of the parent. That is instead of
2084 This reduces memory fragmentation and make it impossible to accidentally
2085 set a mandatory field to NULL.
2087 This currently only works for some fields specifically a SEQUENCE, CHOICE,
2088 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
2089 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
2093 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
2096 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
2097 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
2098 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
2099 DES and RC4 ciphersuites.
2102 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2103 This changes the decoding behaviour for some invalid messages,
2104 though the change is mostly in the more lenient direction, and
2105 legacy behaviour is preserved as much as possible.
2108 *) Fix no-stdio build.
2109 [ David Woodhouse <David.Woodhouse@intel.com> and also
2110 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
2112 *) New testing framework
2113 The testing framework has been largely rewritten and is now using
2114 perl and the perl modules Test::Harness and an extended variant of
2115 Test::More called OpenSSL::Test to do its work. All test scripts in
2116 test/ have been rewritten into test recipes, and all direct calls to
2117 executables in test/Makefile have become individual recipes using the
2118 simplified testing OpenSSL::Test::Simple.
2120 For documentation on our testing modules, do:
2122 perldoc test/testlib/OpenSSL/Test/Simple.pm
2123 perldoc test/testlib/OpenSSL/Test.pm
2127 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
2128 are used; the latter aborts on memory leaks (usually checked on exit).
2129 Some undocumented "set malloc, etc., hooks" functions were removed
2130 and others were changed. All are now documented.
2133 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2135 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2137 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
2138 from RFC4279, RFC4785, RFC5487, RFC5489.
2140 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
2141 original RSA_PSK patch.
2144 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
2145 era flag was never set throughout the codebase (only read). Also removed
2146 SSL3_FLAGS_POP_BUFFER which was only used if
2147 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
2150 *) Changed the default name options in the "ca", "crl", "req" and "x509"
2151 to be "oneline" instead of "compat".
2154 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
2155 not aware of clients that still exhibit this bug, and the workaround
2156 hasn't been working properly for a while.
2159 *) The return type of BIO_number_read() and BIO_number_written() as well as
2160 the corresponding num_read and num_write members in the BIO structure has
2161 changed from unsigned long to uint64_t. On platforms where an unsigned
2162 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
2166 *) Given the pervasive nature of TLS extensions it is inadvisable to run
2167 OpenSSL without support for them. It also means that maintaining
2168 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
2169 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
2172 *) Removed support for the two export grade static DH ciphersuites
2173 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
2174 were newly added (along with a number of other static DH ciphersuites) to
2175 1.0.2. However the two export ones have *never* worked since they were
2176 introduced. It seems strange in any case to be adding new export
2177 ciphersuites, and given "logjam" it also does not seem correct to fix them.
2180 *) Version negotiation has been rewritten. In particular SSLv23_method(),
2181 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
2182 and turned into macros which simply call the new preferred function names
2183 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
2184 should use the new names instead. Also as part of this change the ssl23.h
2185 header file has been removed.
2188 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
2189 code and the associated standard is no longer considered fit-for-purpose.
2192 *) RT2547 was closed. When generating a private key, try to make the
2193 output file readable only by the owner. This behavior change might
2194 be noticeable when interacting with other software.
2196 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
2200 *) Added HTTP GET support to the ocsp command.
2203 *) Changed default digest for the dgst and enc commands from MD5 to
2207 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
2210 *) Added support for TLS extended master secret from
2211 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
2212 initial patch which was a great help during development.
2215 *) All libssl internal structures have been removed from the public header
2216 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
2217 now redundant). Users should not attempt to access internal structures
2218 directly. Instead they should use the provided API functions.
2221 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
2222 Access to deprecated functions can be re-enabled by running config with
2223 "enable-deprecated". In addition applications wishing to use deprecated
2224 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
2225 will, by default, disable some transitive includes that previously existed
2226 in the header files (e.g. ec.h will no longer, by default, include bn.h)
2229 *) Added support for OCB mode. OpenSSL has been granted a patent license
2230 compatible with the OpenSSL license for use of OCB. Details are available
2231 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
2232 for OCB can be removed by calling config with no-ocb.
2235 *) SSLv2 support has been removed. It still supports receiving a SSLv2
2236 compatible client hello.
2239 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
2240 done while fixing the error code for the key-too-small case.
2241 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
2243 *) CA.sh has been removed; use CA.pl instead.
2246 *) Removed old DES API.
2249 *) Remove various unsupported platforms:
2255 Sinix/ReliantUNIX RM400
2260 16-bit platforms such as WIN16
2263 *) Clean up OPENSSL_NO_xxx #define's
2264 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
2265 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
2266 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
2267 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
2268 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
2269 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
2270 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
2271 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
2272 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
2273 Remove MS_STATIC; it's a relic from platforms <32 bits.
2276 *) Cleaned up dead code
2277 Remove all but one '#ifdef undef' which is to be looked at.
2280 *) Clean up calling of xxx_free routines.
2281 Just like free(), fix most of the xxx_free routines to accept
2282 NULL. Remove the non-null checks from callers. Save much code.
2285 *) Add secure heap for storage of private keys (when possible).
2286 Add BIO_s_secmem(), CBIGNUM, etc.
2287 Contributed by Akamai Technologies under our Corporate CLA.
2290 *) Experimental support for a new, fast, unbiased prime candidate generator,
2291 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
2292 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
2294 *) New output format NSS in the sess_id command line tool. This allows
2295 exporting the session id and the master key in NSS keylog format.
2296 [Martin Kaiser <martin@kaiser.cx>]
2298 *) Harmonize version and its documentation. -f flag is used to display
2300 [mancha <mancha1@zoho.com>]
2302 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2303 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
2304 [mancha <mancha1@zoho.com>]
2306 *) Fix some double frees. These are not thought to be exploitable.
2307 [mancha <mancha1@zoho.com>]
2309 *) A missing bounds check in the handling of the TLS heartbeat extension
2310 can be used to reveal up to 64k of memory to a connected client or
2313 Thanks for Neel Mehta of Google Security for discovering this bug and to
2314 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2315 preparing the fix (CVE-2014-0160)
2316 [Adam Langley, Bodo Moeller]
2318 *) Fix for the attack described in the paper "Recovering OpenSSL
2319 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2320 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2321 http://eprint.iacr.org/2014/140
2323 Thanks to Yuval Yarom and Naomi Benger for discovering this
2324 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2325 [Yuval Yarom and Naomi Benger]
2327 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2328 this fixes a limitation in previous versions of OpenSSL.
2331 *) Experimental encrypt-then-mac support.
2333 Experimental support for encrypt then mac from
2334 draft-gutmann-tls-encrypt-then-mac-02.txt
2336 To enable it set the appropriate extension number (0x42 for the test
2337 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
2339 For non-compliant peers (i.e. just about everything) this should have no
2342 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
2346 *) Add EVP support for key wrapping algorithms, to avoid problems with
2347 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2348 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2349 algorithms and include tests cases.
2352 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2356 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2357 MGF1 digest and OAEP label.
2360 *) Make openssl verify return errors.
2361 [Chris Palmer <palmer@google.com> and Ben Laurie]
2363 *) New function ASN1_TIME_diff to calculate the difference between two
2364 ASN1_TIME structures or one structure and the current time.
2367 *) Update fips_test_suite to support multiple command line options. New
2368 test to induce all self test errors in sequence and check expected
2372 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2373 sign or verify all in one operation.
2376 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2377 test programs and fips_test_suite. Includes functionality to parse
2378 the minimal script output of fipsalgest.pl directly.
2381 *) Add authorisation parameter to FIPS_module_mode_set().
2384 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2387 *) Use separate DRBG fields for internal and external flags. New function
2388 FIPS_drbg_health_check() to perform on demand health checking. Add
2389 generation tests to fips_test_suite with reduced health check interval to
2390 demonstrate periodic health checking. Add "nodh" option to
2391 fips_test_suite to skip very slow DH test.
2394 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2398 *) More extensive health check for DRBG checking many more failure modes.
2399 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2400 combination: call this in fips_test_suite.
2403 *) Add support for canonical generation of DSA parameter 'g'. See
2406 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2407 POST to handle HMAC cases.
2410 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2411 to return numerical and string versions of the FIPS module number.
2414 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2415 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2416 outside the validated module in the FIPS capable OpenSSL.
2419 *) Minor change to DRBG entropy callback semantics. In some cases
2420 there is no multiple of the block length between min_len and
2421 max_len. Allow the callback to return more than max_len bytes
2422 of entropy but discard any extra: it is the callback's responsibility
2423 to ensure that the extra data discarded does not impact the
2424 requested amount of entropy.
2427 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2428 information in FIPS186-3, SP800-57 and SP800-131A.
2431 *) CCM support via EVP. Interface is very similar to GCM case except we
2432 must supply all data in one chunk (i.e. no update, final) and the
2433 message length must be supplied if AAD is used. Add algorithm test
2437 *) Initial version of POST overhaul. Add POST callback to allow the status
2438 of POST to be monitored and/or failures induced. Modify fips_test_suite
2439 to use callback. Always run all selftests even if one fails.
2442 *) XTS support including algorithm test driver in the fips_gcmtest program.
2443 Note: this does increase the maximum key length from 32 to 64 bytes but
2444 there should be no binary compatibility issues as existing applications
2445 will never use XTS mode.
2448 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2449 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2450 performs algorithm blocking for unapproved PRNG types. Also do not
2451 set PRNG type in FIPS_mode_set(): leave this to the application.
2452 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2453 the standard OpenSSL PRNG: set additional data to a date time vector.
2456 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2457 This shouldn't present any incompatibility problems because applications
2458 shouldn't be using these directly and any that are will need to rethink
2459 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2462 *) Extensive self tests and health checking required by SP800-90 DRBG.
2463 Remove strength parameter from FIPS_drbg_instantiate and always
2464 instantiate at maximum supported strength.
2467 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2470 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2473 *) New function DH_compute_key_padded() to compute a DH key and pad with
2474 leading zeroes if needed: this complies with SP800-56A et al.
2477 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2478 anything, incomplete, subject to change and largely untested at present.
2481 *) Modify fipscanisteronly build option to only build the necessary object
2482 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2485 *) Add experimental option FIPSSYMS to give all symbols in
2486 fipscanister.o and FIPS or fips prefix. This will avoid
2487 conflicts with future versions of OpenSSL. Add perl script
2488 util/fipsas.pl to preprocess assembly language source files
2489 and rename any affected symbols.
2492 *) Add selftest checks and algorithm block of non-fips algorithms in
2493 FIPS mode. Remove DES2 from selftests.
2496 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2497 return internal method without any ENGINE dependencies. Add new
2498 tiny fips sign and verify functions.
2501 *) New build option no-ec2m to disable characteristic 2 code.
2504 *) New build option "fipscanisteronly". This only builds fipscanister.o
2505 and (currently) associated fips utilities. Uses the file Makefile.fips
2506 instead of Makefile.org as the prototype.
2509 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2510 Update fips_gcmtest to use IV generator.
2513 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2514 setting output buffer to NULL. The *Final function must be
2515 called although it will not retrieve any additional data. The tag
2516 can be set or retrieved with a ctrl. The IV length is by default 12
2517 bytes (96 bits) but can be set to an alternative value. If the IV
2518 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2522 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2523 underlying do_cipher function handles all cipher semantics itself
2524 including padding and finalisation. This is useful if (for example)
2525 an ENGINE cipher handles block padding itself. The behaviour of
2526 do_cipher is subtly changed if this flag is set: the return value
2527 is the number of characters written to the output buffer (zero is
2528 no longer an error code) or a negative error code. Also if the
2529 input buffer is NULL and length 0 finalisation should be performed.
2532 *) If a candidate issuer certificate is already part of the constructed
2533 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2536 *) Improve forward-security support: add functions
2538 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2539 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2541 for use by SSL/TLS servers; the callback function will be called whenever a
2542 new session is created, and gets to decide whether the session may be
2543 cached to make it resumable (return 0) or not (return 1). (As by the
2544 SSL/TLS protocol specifications, the session_id sent by the server will be
2545 empty to indicate that the session is not resumable; also, the server will
2546 not generate RFC 4507 (RFC 5077) session tickets.)
2548 A simple reasonable callback implementation is to return is_forward_secure.
2549 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2550 by the SSL/TLS server library, indicating whether it can provide forward
2552 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2554 *) New -verify_name option in command line utilities to set verification
2558 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2559 Add CMAC pkey methods.
2562 *) Experimental renegotiation in s_server -www mode. If the client
2563 browses /reneg connection is renegotiated. If /renegcert it is
2564 renegotiated requesting a certificate.
2567 *) Add an "external" session cache for debugging purposes to s_server. This
2568 should help trace issues which normally are only apparent in deployed
2569 multi-process servers.
2572 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2573 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2574 BIO_set_cipher() and some obscure PEM functions were changed so they
2575 can now return an error. The RAND changes required a change to the
2576 RAND_METHOD structure.
2579 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2580 a gcc attribute to warn if the result of a function is ignored. This
2581 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2582 whose return value is often ignored.
2585 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2586 These allow SCTs (signed certificate timestamps) to be requested and
2587 validated when establishing a connection.
2588 [Rob Percival <robpercival@google.com>]
2590 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2592 *) Prevent padding oracle in AES-NI CBC MAC check
2594 A MITM attacker can use a padding oracle attack to decrypt traffic
2595 when the connection uses an AES CBC cipher and the server support
2598 This issue was introduced as part of the fix for Lucky 13 padding
2599 attack (CVE-2013-0169). The padding check was rewritten to be in
2600 constant time by making sure that always the same bytes are read and
2601 compared against either the MAC or padding bytes. But it no longer
2602 checked that there was enough data to have both the MAC and padding
2605 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2609 *) Fix EVP_EncodeUpdate overflow
2611 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2612 Base64 encoding of binary data. If an attacker is able to supply very large
2613 amounts of input data then a length check can overflow resulting in a heap
2616 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2617 the PEM_write_bio* family of functions. These are mainly used within the
2618 OpenSSL command line applications, so any application which processes data
2619 from an untrusted source and outputs it as a PEM file should be considered
2620 vulnerable to this issue. User applications that call these APIs directly
2621 with large amounts of untrusted data may also be vulnerable.
2623 This issue was reported by Guido Vranken.
2627 *) Fix EVP_EncryptUpdate overflow
2629 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2630 is able to supply very large amounts of input data after a previous call to
2631 EVP_EncryptUpdate() with a partial block then a length check can overflow
2632 resulting in a heap corruption. Following an analysis of all OpenSSL
2633 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2634 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2635 the first called function after an EVP_EncryptInit(), and therefore that
2636 specific call must be safe. The second form is where the length passed to
2637 EVP_EncryptUpdate() can be seen from the code to be some small value and
2638 therefore there is no possibility of an overflow. Since all instances are
2639 one of these two forms, it is believed that there can be no overflows in
2640 internal code due to this problem. It should be noted that
2641 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2642 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2643 of these calls have also been analysed too and it is believed there are no
2644 instances in internal usage where an overflow could occur.
2646 This issue was reported by Guido Vranken.
2650 *) Prevent ASN.1 BIO excessive memory allocation
2652 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2653 a short invalid encoding can cause allocation of large amounts of memory
2654 potentially consuming excessive resources or exhausting memory.
2656 Any application parsing untrusted data through d2i BIO functions is
2657 affected. The memory based functions such as d2i_X509() are *not* affected.
2658 Since the memory based functions are used by the TLS library, TLS
2659 applications are not affected.
2661 This issue was reported by Brian Carpenter.
2667 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2668 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2669 in arbitrary stack data being returned in the buffer.
2671 This issue was reported by Guido Vranken.
2675 *) Modify behavior of ALPN to invoke callback after SNI/servername
2676 callback, such that updates to the SSL_CTX affect ALPN.
2679 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2683 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2684 methods are enabled and ssl2 is disabled the methods return NULL.
2687 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2689 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2690 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2691 provide any "EXPORT" or "LOW" strength ciphers.
2694 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2695 is by default disabled at build-time. Builds that are not configured with
2696 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2697 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2698 will need to explicitly call either of:
2700 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2702 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2704 as appropriate. Even if either of those is used, or the application
2705 explicitly uses the version-specific SSLv2_method() or its client and
2706 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2707 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2708 ciphers, and SSLv2 56-bit DES are no longer available.
2712 *) Fix a double-free in DSA code
2714 A double free bug was discovered when OpenSSL parses malformed DSA private
2715 keys and could lead to a DoS attack or memory corruption for applications
2716 that receive DSA private keys from untrusted sources. This scenario is
2719 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2724 *) Disable SRP fake user seed to address a server memory leak.
2726 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2728 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2729 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2730 was changed to ignore the "fake user" SRP seed, even if the seed
2733 Users should use SRP_VBASE_get1_by_user instead. Note that in
2734 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2735 also that even though configuring the SRP seed attempts to hide
2736 invalid usernames by continuing the handshake with fake
2737 credentials, this behaviour is not constant time and no strong
2738 guarantees are made that the handshake is indistinguishable from
2739 that of a valid user.
2743 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2745 In the BN_hex2bn function the number of hex digits is calculated using an
2746 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2747 large values of |i| this can result in |bn_expand| not allocating any
2748 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2749 field as NULL leading to a subsequent NULL ptr deref. For very large values
2750 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2751 In this case memory is allocated to the internal BIGNUM data field, but it
2752 is insufficiently sized leading to heap corruption. A similar issue exists
2753 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2754 is ever called by user applications with very large untrusted hex/dec data.
2755 This is anticipated to be a rare occurrence.
2757 All OpenSSL internal usage of these functions use data that is not expected
2758 to be untrusted, e.g. config file data or application command line
2759 arguments. If user developed applications generate config file data based
2760 on untrusted data then it is possible that this could also lead to security
2761 consequences. This is also anticipated to be rare.
2763 This issue was reported to OpenSSL by Guido Vranken.
2767 *) Fix memory issues in BIO_*printf functions
2769 The internal |fmtstr| function used in processing a "%s" format string in
2770 the BIO_*printf functions could overflow while calculating the length of a
2771 string and cause an OOB read when printing very long strings.
2773 Additionally the internal |doapr_outch| function can attempt to write to an
2774 OOB memory location (at an offset from the NULL pointer) in the event of a
2775 memory allocation failure. In 1.0.2 and below this could be caused where
2776 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2777 could be in processing a very long "%s" format string. Memory leaks can
2780 The first issue may mask the second issue dependent on compiler behaviour.
2781 These problems could enable attacks where large amounts of untrusted data
2782 is passed to the BIO_*printf functions. If applications use these functions
2783 in this way then they could be vulnerable. OpenSSL itself uses these
2784 functions when printing out human-readable dumps of ASN.1 data. Therefore
2785 applications that print this data could be vulnerable if the data is from
2786 untrusted sources. OpenSSL command line applications could also be
2787 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2788 as command line arguments.
2790 Libssl is not considered directly vulnerable. Additionally certificates etc
2791 received via remote connections via libssl are also unlikely to be able to
2792 trigger these issues because of message size limits enforced within libssl.
2794 This issue was reported to OpenSSL Guido Vranken.
2798 *) Side channel attack on modular exponentiation
2800 A side-channel attack was found which makes use of cache-bank conflicts on
2801 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2802 of RSA keys. The ability to exploit this issue is limited as it relies on
2803 an attacker who has control of code in a thread running on the same
2804 hyper-threaded core as the victim thread which is performing decryptions.
2806 This issue was reported to OpenSSL by Yuval Yarom, The University of
2807 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2808 Nadia Heninger, University of Pennsylvania with more information at
2809 http://cachebleed.info.
2813 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2814 if no keysize is specified with default_bits. This fixes an
2815 omission in an earlier change that changed all RSA/DSA key generation
2816 apps to use 2048 bits by default.
2819 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2820 *) DH small subgroups
2822 Historically OpenSSL only ever generated DH parameters based on "safe"
2823 primes. More recently (in version 1.0.2) support was provided for
2824 generating X9.42 style parameter files such as those required for RFC 5114
2825 support. The primes used in such files may not be "safe". Where an
2826 application is using DH configured with parameters based on primes that are
2827 not "safe" then an attacker could use this fact to find a peer's private
2828 DH exponent. This attack requires that the attacker complete multiple
2829 handshakes in which the peer uses the same private DH exponent. For example
2830 this could be used to discover a TLS server's private DH exponent if it's
2831 reusing the private DH exponent or it's using a static DH ciphersuite.
2833 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2834 TLS. It is not on by default. If the option is not set then the server
2835 reuses the same private DH exponent for the life of the server process and
2836 would be vulnerable to this attack. It is believed that many popular
2837 applications do set this option and would therefore not be at risk.
2839 The fix for this issue adds an additional check where a "q" parameter is
2840 available (as is the case in X9.42 based parameters). This detects the
2841 only known attack, and is the only possible defense for static DH
2842 ciphersuites. This could have some performance impact.
2844 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2845 default and cannot be disabled. This could have some performance impact.
2847 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2851 *) SSLv2 doesn't block disabled ciphers
2853 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2854 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2855 been disabled, provided that the SSLv2 protocol was not also disabled via
2858 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2859 and Sebastian Schinzel.
2863 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2865 *) BN_mod_exp may produce incorrect results on x86_64
2867 There is a carry propagating bug in the x86_64 Montgomery squaring
2868 procedure. No EC algorithms are affected. Analysis suggests that attacks
2869 against RSA and DSA as a result of this defect would be very difficult to
2870 perform and are not believed likely. Attacks against DH are considered just
2871 feasible (although very difficult) because most of the work necessary to
2872 deduce information about a private key may be performed offline. The amount
2873 of resources required for such an attack would be very significant and
2874 likely only accessible to a limited number of attackers. An attacker would
2875 additionally need online access to an unpatched system using the target
2876 private key in a scenario with persistent DH parameters and a private
2877 key that is shared between multiple clients. For example this can occur by
2878 default in OpenSSL DHE based SSL/TLS ciphersuites.
2880 This issue was reported to OpenSSL by Hanno Böck.
2884 *) Certificate verify crash with missing PSS parameter
2886 The signature verification routines will crash with a NULL pointer
2887 dereference if presented with an ASN.1 signature using the RSA PSS
2888 algorithm and absent mask generation function parameter. Since these
2889 routines are used to verify certificate signature algorithms this can be
2890 used to crash any certificate verification operation and exploited in a
2891 DoS attack. Any application which performs certificate verification is
2892 vulnerable including OpenSSL clients and servers which enable client
2895 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2899 *) X509_ATTRIBUTE memory leak
2901 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2902 memory. This structure is used by the PKCS#7 and CMS routines so any
2903 application which reads PKCS#7 or CMS data from untrusted sources is
2904 affected. SSL/TLS is not affected.
2906 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2911 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2912 This changes the decoding behaviour for some invalid messages,
2913 though the change is mostly in the more lenient direction, and
2914 legacy behaviour is preserved as much as possible.
2917 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2919 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2921 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2923 *) Alternate chains certificate forgery
2925 During certificate verification, OpenSSL will attempt to find an
2926 alternative certificate chain if the first attempt to build such a chain
2927 fails. An error in the implementation of this logic can mean that an
2928 attacker could cause certain checks on untrusted certificates to be
2929 bypassed, such as the CA flag, enabling them to use a valid leaf
2930 certificate to act as a CA and "issue" an invalid certificate.
2932 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2936 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2938 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2939 incompatibility in the handling of HMAC. The previous ABI has now been
2943 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2945 *) Malformed ECParameters causes infinite loop
2947 When processing an ECParameters structure OpenSSL enters an infinite loop
2948 if the curve specified is over a specially malformed binary polynomial
2951 This can be used to perform denial of service against any
2952 system which processes public keys, certificate requests or
2953 certificates. This includes TLS clients and TLS servers with
2954 client authentication enabled.
2956 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2960 *) Exploitable out-of-bounds read in X509_cmp_time
2962 X509_cmp_time does not properly check the length of the ASN1_TIME
2963 string and can read a few bytes out of bounds. In addition,
2964 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2967 An attacker can use this to craft malformed certificates and CRLs of
2968 various sizes and potentially cause a segmentation fault, resulting in
2969 a DoS on applications that verify certificates or CRLs. TLS clients
2970 that verify CRLs are affected. TLS clients and servers with client
2971 authentication enabled may be affected if they use custom verification
2974 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2975 independently by Hanno Böck.
2979 *) PKCS7 crash with missing EnvelopedContent
2981 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2982 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2983 with missing content and trigger a NULL pointer dereference on parsing.
2985 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2986 structures from untrusted sources are affected. OpenSSL clients and
2987 servers are not affected.
2989 This issue was reported to OpenSSL by Michal Zalewski (Google).
2993 *) CMS verify infinite loop with unknown hash function
2995 When verifying a signedData message the CMS code can enter an infinite loop
2996 if presented with an unknown hash function OID. This can be used to perform
2997 denial of service against any system which verifies signedData messages using
2999 This issue was reported to OpenSSL by Johannes Bauer.
3003 *) Race condition handling NewSessionTicket
3005 If a NewSessionTicket is received by a multi-threaded client when attempting to
3006 reuse a previous ticket then a race condition can occur potentially leading to
3007 a double free of the ticket data.
3011 *) Only support 256-bit or stronger elliptic curves with the
3012 'ecdh_auto' setting (server) or by default (client). Of supported
3013 curves, prefer P-256 (both).
3016 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
3018 *) ClientHello sigalgs DoS fix
3020 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
3021 invalid signature algorithms extension a NULL pointer dereference will
3022 occur. This can be exploited in a DoS attack against the server.
3024 This issue was was reported to OpenSSL by David Ramos of Stanford
3027 [Stephen Henson and Matt Caswell]
3029 *) Multiblock corrupted pointer fix
3031 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
3032 feature only applies on 64 bit x86 architecture platforms that support AES
3033 NI instructions. A defect in the implementation of "multiblock" can cause
3034 OpenSSL's internal write buffer to become incorrectly set to NULL when
3035 using non-blocking IO. Typically, when the user application is using a
3036 socket BIO for writing, this will only result in a failed connection.
3037 However if some other BIO is used then it is likely that a segmentation
3038 fault will be triggered, thus enabling a potential DoS attack.
3040 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
3044 *) Segmentation fault in DTLSv1_listen fix
3046 The DTLSv1_listen function is intended to be stateless and processes the
3047 initial ClientHello from many peers. It is common for user code to loop
3048 over the call to DTLSv1_listen until a valid ClientHello is received with
3049 an associated cookie. A defect in the implementation of DTLSv1_listen means
3050 that state is preserved in the SSL object from one invocation to the next
3051 that can lead to a segmentation fault. Errors processing the initial
3052 ClientHello can trigger this scenario. An example of such an error could be
3053 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
3056 This issue was reported to OpenSSL by Per Allansson.
3060 *) Segmentation fault in ASN1_TYPE_cmp fix
3062 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
3063 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
3064 certificate signature algorithm consistency this can be used to crash any
3065 certificate verification operation and exploited in a DoS attack. Any
3066 application which performs certificate verification is vulnerable including
3067 OpenSSL clients and servers which enable client authentication.
3071 *) Segmentation fault for invalid PSS parameters fix
3073 The signature verification routines will crash with a NULL pointer
3074 dereference if presented with an ASN.1 signature using the RSA PSS
3075 algorithm and invalid parameters. Since these routines are used to verify
3076 certificate signature algorithms this can be used to crash any
3077 certificate verification operation and exploited in a DoS attack. Any
3078 application which performs certificate verification is vulnerable including
3079 OpenSSL clients and servers which enable client authentication.
3081 This issue was was reported to OpenSSL by Brian Carpenter.
3085 *) ASN.1 structure reuse memory corruption fix
3087 Reusing a structure in ASN.1 parsing may allow an attacker to cause
3088 memory corruption via an invalid write. Such reuse is and has been
3089 strongly discouraged and is believed to be rare.
3091 Applications that parse structures containing CHOICE or ANY DEFINED BY
3092 components may be affected. Certificate parsing (d2i_X509 and related
3093 functions) are however not affected. OpenSSL clients and servers are
3098 *) PKCS7 NULL pointer dereferences fix
3100 The PKCS#7 parsing code does not handle missing outer ContentInfo
3101 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
3102 missing content and trigger a NULL pointer dereference on parsing.
3104 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
3105 otherwise parse PKCS#7 structures from untrusted sources are
3106 affected. OpenSSL clients and servers are not affected.
3108 This issue was reported to OpenSSL by Michal Zalewski (Google).
3112 *) DoS via reachable assert in SSLv2 servers fix
3114 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
3115 servers that both support SSLv2 and enable export cipher suites by sending
3116 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
3118 This issue was discovered by Sean Burford (Google) and Emilia Käsper
3119 (OpenSSL development team).
3123 *) Empty CKE with client auth and DHE fix
3125 If client auth is used then a server can seg fault in the event of a DHE
3126 ciphersuite being selected and a zero length ClientKeyExchange message
3127 being sent by the client. This could be exploited in a DoS attack.
3131 *) Handshake with unseeded PRNG fix
3133 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
3134 with an unseeded PRNG. The conditions are:
3135 - The client is on a platform where the PRNG has not been seeded
3136 automatically, and the user has not seeded manually
3137 - A protocol specific client method version has been used (i.e. not
3138 SSL_client_methodv23)
3139 - A ciphersuite is used that does not require additional random data from
3140 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
3142 If the handshake succeeds then the client random that has been used will
3143 have been generated from a PRNG with insufficient entropy and therefore the
3144 output may be predictable.
3146 For example using the following command with an unseeded openssl will
3147 succeed on an unpatched platform:
3149 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
3153 *) Use After Free following d2i_ECPrivatekey error fix
3155 A malformed EC private key file consumed via the d2i_ECPrivateKey function
3156 could cause a use after free condition. This, in turn, could cause a double
3157 free in several private key parsing functions (such as d2i_PrivateKey
3158 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
3159 for applications that receive EC private keys from untrusted
3160 sources. This scenario is considered rare.
3162 This issue was discovered by the BoringSSL project and fixed in their
3167 *) X509_to_X509_REQ NULL pointer deref fix
3169 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
3170 the certificate key is invalid. This function is rarely used in practice.
3172 This issue was discovered by Brian Carpenter.
3176 *) Removed the export ciphers from the DEFAULT ciphers
3179 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
3181 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
3182 ARMv5 through ARMv8, as opposite to "locking" it to single one.
3183 So far those who have to target multiple platforms would compromise
3184 and argue that binary targeting say ARMv5 would still execute on
3185 ARMv8. "Universal" build resolves this compromise by providing
3186 near-optimal performance even on newer platforms.
3189 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
3190 (other platforms pending).
3191 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
3193 *) Add support for the SignedCertificateTimestampList certificate and
3194 OCSP response extensions from RFC6962.
3197 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3198 for corner cases. (Certain input points at infinity could lead to
3199 bogus results, with non-infinity inputs mapped to infinity too.)
3202 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
3203 This covers AES, SHA256/512 and GHASH. "Initial" means that most
3204 common cases are optimized and there still is room for further
3205 improvements. Vector Permutation AES for Altivec is also added.
3208 *) Add support for little-endian ppc64 Linux target.
3209 [Marcelo Cerri (IBM)]
3211 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
3212 SHA1, SHA256 and GHASH. "Initial" means that most common cases
3213 are optimized and there still is room for further improvements.
3214 Both 32- and 64-bit modes are supported.
3215 [Andy Polyakov, Ard Biesheuvel (Linaro)]
3217 *) Improved ARMv7 NEON support.
3220 *) Support for SPARC Architecture 2011 crypto extensions, first
3221 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
3222 SHA256/512, MD5, GHASH and modular exponentiation.
3223 [Andy Polyakov, David Miller]
3225 *) Accelerated modular exponentiation for Intel processors, a.k.a.
3227 [Shay Gueron & Vlad Krasnov (Intel Corp)]
3229 *) Support for new and upcoming Intel processors, including AVX2,
3230 BMI and SHA ISA extensions. This includes additional "stitched"
3231 implementations, AESNI-SHA256 and GCM, and multi-buffer support
3234 This work was sponsored by Intel Corp.
3237 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
3238 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
3239 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
3242 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3243 this fixes a limitation in previous versions of OpenSSL.
3246 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3247 MGF1 digest and OAEP label.
3250 *) Add EVP support for key wrapping algorithms, to avoid problems with
3251 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3252 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3253 algorithms and include tests cases.
3256 *) Add functions to allocate and set the fields of an ECDSA_METHOD
3258 [Douglas E. Engert, Steve Henson]
3260 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
3261 difference in days and seconds between two tm or ASN1_TIME structures.
3264 *) Add -rev test option to s_server to just reverse order of characters
3265 received by client and send back to server. Also prints an abbreviated
3266 summary of the connection parameters.
3269 *) New option -brief for s_client and s_server to print out a brief summary
3270 of connection parameters.
3273 *) Add callbacks for arbitrary TLS extensions.
3274 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
3276 *) New option -crl_download in several openssl utilities to download CRLs
3277 from CRLDP extension in certificates.
3280 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
3283 *) New function X509_CRL_diff to generate a delta CRL from the difference
3284 of two full CRLs. Add support to "crl" utility.
3287 *) New functions to set lookup_crls function and to retrieve
3288 X509_STORE from X509_STORE_CTX.
3291 *) Print out deprecated issuer and subject unique ID fields in
3295 *) Extend OCSP I/O functions so they can be used for simple general purpose
3296 HTTP as well as OCSP. New wrapper function which can be used to download
3297 CRLs using the OCSP API.
3300 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
3303 *) SSL_CONF* functions. These provide a common framework for application
3304 configuration using configuration files or command lines.
3307 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
3308 message callback and prints the results. Needs compile time option
3309 "enable-ssl-trace". New options to s_client and s_server to enable
3313 *) New ctrl and macro to retrieve supported points extensions.
3314 Print out extension in s_server and s_client.
3317 *) New functions to retrieve certificate signature and signature
3321 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
3325 *) New Suite B modes for TLS code. These use and enforce the requirements
3326 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
3327 only use Suite B curves. The Suite B modes can be set by using the
3328 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
3331 *) New chain verification flags for Suite B levels of security. Check
3332 algorithms are acceptable when flags are set in X509_verify_cert.
3335 *) Make tls1_check_chain return a set of flags indicating checks passed
3336 by a certificate chain. Add additional tests to handle client
3337 certificates: checks for matching certificate type and issuer name
3341 *) If an attempt is made to use a signature algorithm not in the peer
3342 preference list abort the handshake. If client has no suitable
3343 signature algorithms in response to a certificate request do not
3344 use the certificate.
3347 *) If server EC tmp key is not in client preference list abort handshake.
3350 *) Add support for certificate stores in CERT structure. This makes it
3351 possible to have different stores per SSL structure or one store in
3352 the parent SSL_CTX. Include distinct stores for certificate chain
3353 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3354 to build and store a certificate chain in CERT structure: returning
3355 an error if the chain cannot be built: this will allow applications
3356 to test if a chain is correctly configured.
3358 Note: if the CERT based stores are not set then the parent SSL_CTX
3359 store is used to retain compatibility with existing behaviour.
3363 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3364 mask based on the current session, check mask when sending client
3365 hello and checking the requested ciphersuite.
3368 *) New ctrls to retrieve and set certificate types in a certificate
3369 request message. Print out received values in s_client. If certificate
3370 types is not set with custom values set sensible values based on
3371 supported signature algorithms.
3374 *) Support for distinct client and server supported signature algorithms.
3377 *) Add certificate callback. If set this is called whenever a certificate
3378 is required by client or server. An application can decide which
3379 certificate chain to present based on arbitrary criteria: for example
3380 supported signature algorithms. Add very simple example to s_server.
3381 This fixes many of the problems and restrictions of the existing client
3382 certificate callback: for example you can now clear an existing
3383 certificate and specify the whole chain.
3386 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3387 the certificate can be used for (if anything). Set valid_flags field
3388 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3389 to have similar checks in it.
3391 Add new "cert_flags" field to CERT structure and include a "strict mode".
3392 This enforces some TLS certificate requirements (such as only permitting
3393 certificate signature algorithms contained in the supported algorithms
3394 extension) which some implementations ignore: this option should be used
3395 with caution as it could cause interoperability issues.
3398 *) Update and tidy signature algorithm extension processing. Work out
3399 shared signature algorithms based on preferences and peer algorithms
3400 and print them out in s_client and s_server. Abort handshake if no
3401 shared signature algorithms.
3404 *) Add new functions to allow customised supported signature algorithms
3405 for SSL and SSL_CTX structures. Add options to s_client and s_server
3409 *) New function SSL_certs_clear() to delete all references to certificates
3410 from an SSL structure. Before this once a certificate had been added
3411 it couldn't be removed.
3414 *) Integrate hostname, email address and IP address checking with certificate
3415 verification. New verify options supporting checking in openssl utility.
3418 *) Fixes and wildcard matching support to hostname and email checking
3419 functions. Add manual page.
3420 [Florian Weimer (Red Hat Product Security Team)]
3422 *) New functions to check a hostname email or IP address against a
3423 certificate. Add options x509 utility to print results of checks against
3427 *) Fix OCSP checking.
3428 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3430 *) Initial experimental support for explicitly trusted non-root CAs.
3431 OpenSSL still tries to build a complete chain to a root but if an
3432 intermediate CA has a trust setting included that is used. The first
3433 setting is used: whether to trust (e.g., -addtrust option to the x509
3437 *) Add -trusted_first option which attempts to find certificates in the
3438 trusted store even if an untrusted chain is also supplied.
3441 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3442 platform support for Linux and Android.
3445 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3448 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3449 When in FIPS mode the approved implementations are used as normal,
3450 when not in FIPS mode the internal unapproved versions are used instead.
3451 This means that the FIPS capable OpenSSL isn't forced to use the
3452 (often lower performance) FIPS implementations outside FIPS mode.
3455 *) Transparently support X9.42 DH parameters when calling
3456 PEM_read_bio_DHparameters. This means existing applications can handle
3457 the new parameter format automatically.
3460 *) Initial experimental support for X9.42 DH parameter format: mainly
3461 to support use of 'q' parameter for RFC5114 parameters.
3464 *) Add DH parameters from RFC5114 including test data to dhtest.
3467 *) Support for automatic EC temporary key parameter selection. If enabled
3468 the most preferred EC parameters are automatically used instead of
3469 hardcoded fixed parameters. Now a server just has to call:
3470 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3471 support ECDH and use the most appropriate parameters.
3474 *) Enhance and tidy EC curve and point format TLS extension code. Use
3475 static structures instead of allocation if default values are used.
3476 New ctrls to set curves we wish to support and to retrieve shared curves.
3477 Print out shared curves in s_server. New options to s_server and s_client
3478 to set list of supported curves.
3481 *) New ctrls to retrieve supported signature algorithms and
3482 supported curve values as an array of NIDs. Extend openssl utility
3483 to print out received values.
3486 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3487 between NIDs and the more common NIST names such as "P-256". Enhance
3488 ecparam utility and ECC method to recognise the NIST names for curves.
3491 *) Enhance SSL/TLS certificate chain handling to support different
3492 chains for each certificate instead of one chain in the parent SSL_CTX.
3495 *) Support for fixed DH ciphersuite client authentication: where both
3496 server and client use DH certificates with common parameters.
3499 *) Support for fixed DH ciphersuites: those requiring DH server
3503 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3505 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3506 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3507 X509_CINF_get_signature were reverted post internal team review.
3509 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3511 *) Build fixes for the Windows and OpenVMS platforms
3512 [Matt Caswell and Richard Levitte]
3514 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3516 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3517 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3518 dereference. This could lead to a Denial Of Service attack. Thanks to
3519 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3523 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3524 dtls1_buffer_record function under certain conditions. In particular this
3525 could occur if an attacker sent repeated DTLS records with the same
3526 sequence number but for the next epoch. The memory leak could be exploited
3527 by an attacker in a Denial of Service attack through memory exhaustion.
3528 Thanks to Chris Mueller for reporting this issue.
3532 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3533 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3534 method would be set to NULL which could later result in a NULL pointer
3535 dereference. Thanks to Frank Schmirler for reporting this issue.
3539 *) Abort handshake if server key exchange message is omitted for ephemeral
3542 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3543 reporting this issue.
3547 *) Remove non-export ephemeral RSA code on client and server. This code
3548 violated the TLS standard by allowing the use of temporary RSA keys in
3549 non-export ciphersuites and could be used by a server to effectively
3550 downgrade the RSA key length used to a value smaller than the server
3551 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3552 INRIA or reporting this issue.
3556 *) Fixed issue where DH client certificates are accepted without verification.
3557 An OpenSSL server will accept a DH certificate for client authentication
3558 without the certificate verify message. This effectively allows a client to
3559 authenticate without the use of a private key. This only affects servers
3560 which trust a client certificate authority which issues certificates
3561 containing DH keys: these are extremely rare and hardly ever encountered.
3562 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3567 *) Ensure that the session ID context of an SSL is updated when its
3568 SSL_CTX is updated via SSL_set_SSL_CTX.
3570 The session ID context is typically set from the parent SSL_CTX,
3571 and can vary with the CTX.
3574 *) Fix various certificate fingerprint issues.
3576 By using non-DER or invalid encodings outside the signed portion of a
3577 certificate the fingerprint can be changed without breaking the signature.
3578 Although no details of the signed portion of the certificate can be changed
3579 this can cause problems with some applications: e.g. those using the
3580 certificate fingerprint for blacklists.
3582 1. Reject signatures with non zero unused bits.
3584 If the BIT STRING containing the signature has non zero unused bits reject
3585 the signature. All current signature algorithms require zero unused bits.
3587 2. Check certificate algorithm consistency.
3589 Check the AlgorithmIdentifier inside TBS matches the one in the
3590 certificate signature. NB: this will result in signature failure
3591 errors for some broken certificates.
3593 Thanks to Konrad Kraszewski from Google for reporting this issue.
3595 3. Check DSA/ECDSA signatures use DER.
3597 Re-encode DSA/ECDSA signatures and compare with the original received
3598 signature. Return an error if there is a mismatch.
3600 This will reject various cases including garbage after signature
3601 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3602 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3603 (negative or with leading zeroes).
3605 Further analysis was conducted and fixes were developed by Stephen Henson
3606 of the OpenSSL core team.
3611 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3612 results on some platforms, including x86_64. This bug occurs at random
3613 with a very low probability, and is not known to be exploitable in any
3614 way, though its exact impact is difficult to determine. Thanks to Pieter
3615 Wuille (Blockstream) who reported this issue and also suggested an initial
3616 fix. Further analysis was conducted by the OpenSSL development team and
3617 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3618 the OpenSSL core team.
3622 *) Do not resume sessions on the server if the negotiated protocol
3623 version does not match the session's version. Resuming with a different
3624 version, while not strictly forbidden by the RFC, is of questionable
3625 sanity and breaks all known clients.
3626 [David Benjamin, Emilia Käsper]
3628 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3629 early CCS messages during renegotiation. (Note that because
3630 renegotiation is encrypted, this early CCS was not exploitable.)
3633 *) Tighten client-side session ticket handling during renegotiation:
3634 ensure that the client only accepts a session ticket if the server sends
3635 the extension anew in the ServerHello. Previously, a TLS client would
3636 reuse the old extension state and thus accept a session ticket if one was
3637 announced in the initial ServerHello.
3639 Similarly, ensure that the client requires a session ticket if one
3640 was advertised in the ServerHello. Previously, a TLS client would
3641 ignore a missing NewSessionTicket message.
3644 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3646 *) SRTP Memory Leak.
3648 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3649 sends a carefully crafted handshake message, to cause OpenSSL to fail
3650 to free up to 64k of memory causing a memory leak. This could be
3651 exploited in a Denial Of Service attack. This issue affects OpenSSL
3652 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3653 whether SRTP is used or configured. Implementations of OpenSSL that
3654 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3656 The fix was developed by the OpenSSL team.
3660 *) Session Ticket Memory Leak.
3662 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3663 integrity of that ticket is first verified. In the event of a session
3664 ticket integrity check failing, OpenSSL will fail to free memory
3665 causing a memory leak. By sending a large number of invalid session
3666 tickets an attacker could exploit this issue in a Denial Of Service
3671 *) Build option no-ssl3 is incomplete.
3673 When OpenSSL is configured with "no-ssl3" as a build option, servers
3674 could accept and complete a SSL 3.0 handshake, and clients could be
3675 configured to send them.
3677 [Akamai and the OpenSSL team]
3679 *) Add support for TLS_FALLBACK_SCSV.
3680 Client applications doing fallback retries should call
3681 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3683 [Adam Langley, Bodo Moeller]
3685 *) Add additional DigestInfo checks.
3687 Re-encode DigestInto in DER and check against the original when
3688 verifying RSA signature: this will reject any improperly encoded
3689 DigestInfo structures.
3691 Note: this is a precautionary measure and no attacks are currently known.
3695 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3697 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3698 SRP code can be overrun an internal buffer. Add sanity check that
3699 g, A, B < N to SRP code.
3701 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3702 Group for discovering this issue.
3706 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3707 TLS 1.0 instead of higher protocol versions when the ClientHello message
3708 is badly fragmented. This allows a man-in-the-middle attacker to force a
3709 downgrade to TLS 1.0 even if both the server and the client support a
3710 higher protocol version, by modifying the client's TLS records.
3712 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3713 researching this issue.
3717 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3718 to a denial of service attack. A malicious server can crash the client
3719 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3720 ciphersuite and sending carefully crafted handshake messages.
3722 Thanks to Felix Gröbert (Google) for discovering and researching this
3727 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3728 to leak memory. This can be exploited through a Denial of Service attack.
3729 Thanks to Adam Langley for discovering and researching this issue.
3733 *) An attacker can force openssl to consume large amounts of memory whilst
3734 processing DTLS handshake messages. This can be exploited through a
3735 Denial of Service attack.
3736 Thanks to Adam Langley for discovering and researching this issue.
3740 *) An attacker can force an error condition which causes openssl to crash
3741 whilst processing DTLS packets due to memory being freed twice. This
3742 can be exploited through a Denial of Service attack.
3743 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3748 *) If a multithreaded client connects to a malicious server using a resumed
3749 session and the server sends an ec point format extension it could write
3750 up to 255 bytes to freed memory.
3752 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3757 *) A malicious server can crash an OpenSSL client with a null pointer
3758 dereference (read) by specifying an SRP ciphersuite even though it was not
3759 properly negotiated with the client. This can be exploited through a
3760 Denial of Service attack.
3762 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3763 discovering and researching this issue.
3767 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3768 X509_name_oneline, X509_name_print_ex et al. to leak some information
3769 from the stack. Applications may be affected if they echo pretty printing
3770 output to the attacker.
3772 Thanks to Ivan Fratric (Google) for discovering this issue.
3774 [Emilia Käsper, and Steve Henson]
3776 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3777 for corner cases. (Certain input points at infinity could lead to
3778 bogus results, with non-infinity inputs mapped to infinity too.)
3781 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3783 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3784 handshake can force the use of weak keying material in OpenSSL
3785 SSL/TLS clients and servers.
3787 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3788 researching this issue. (CVE-2014-0224)
3789 [KIKUCHI Masashi, Steve Henson]
3791 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3792 OpenSSL DTLS client the code can be made to recurse eventually crashing
3795 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3797 [Imre Rad, Steve Henson]
3799 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3800 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3801 client or server. This is potentially exploitable to run arbitrary
3802 code on a vulnerable client or server.
3804 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3805 [Jüri Aedla, Steve Henson]
3807 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3808 are subject to a denial of service attack.
3810 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3811 this issue. (CVE-2014-3470)
3812 [Felix Gröbert, Ivan Fratric, Steve Henson]
3814 *) Harmonize version and its documentation. -f flag is used to display
3816 [mancha <mancha1@zoho.com>]
3818 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3819 in i2d_ECPrivateKey.
3820 [mancha <mancha1@zoho.com>]
3822 *) Fix some double frees. These are not thought to be exploitable.
3823 [mancha <mancha1@zoho.com>]
3825 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3827 *) A missing bounds check in the handling of the TLS heartbeat extension
3828 can be used to reveal up to 64k of memory to a connected client or
3831 Thanks for Neel Mehta of Google Security for discovering this bug and to
3832 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3833 preparing the fix (CVE-2014-0160)
3834 [Adam Langley, Bodo Moeller]
3836 *) Fix for the attack described in the paper "Recovering OpenSSL
3837 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3838 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3839 http://eprint.iacr.org/2014/140
3841 Thanks to Yuval Yarom and Naomi Benger for discovering this
3842 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3843 [Yuval Yarom and Naomi Benger]
3845 *) TLS pad extension: draft-agl-tls-padding-03
3847 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3848 TLS client Hello record length value would otherwise be > 255 and
3849 less that 512 pad with a dummy extension containing zeroes so it
3850 is at least 512 bytes long.
3852 [Adam Langley, Steve Henson]
3854 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3856 *) Fix for TLS record tampering bug. A carefully crafted invalid
3857 handshake could crash OpenSSL with a NULL pointer exception.
3858 Thanks to Anton Johansson for reporting this issues.
3861 *) Keep original DTLS digest and encryption contexts in retransmission
3862 structures so we can use the previous session parameters if they need
3863 to be resent. (CVE-2013-6450)
3866 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3867 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3868 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3869 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3870 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3871 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3872 [Rob Stradling, Adam Langley]
3874 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3876 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3877 supporting platforms or when small records were transferred.
3878 [Andy Polyakov, Steve Henson]
3880 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3882 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3884 This addresses the flaw in CBC record processing discovered by
3885 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3886 at: http://www.isg.rhul.ac.uk/tls/
3888 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3889 Security Group at Royal Holloway, University of London
3890 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3891 Emilia Käsper for the initial patch.
3893 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3895 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3896 ciphersuites which can be exploited in a denial of service attack.
3897 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3898 and detecting this bug and to Wolfgang Ettlinger
3899 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3903 *) Return an error when checking OCSP signatures when key is NULL.
3904 This fixes a DoS attack. (CVE-2013-0166)
3907 *) Make openssl verify return errors.
3908 [Chris Palmer <palmer@google.com> and Ben Laurie]
3910 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3911 the right response is stapled. Also change SSL_get_certificate()
3912 so it returns the certificate actually sent.
3913 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3914 [Rob Stradling <rob.stradling@comodo.com>]
3916 *) Fix possible deadlock when decoding public keys.
3919 *) Don't use TLS 1.0 record version number in initial client hello
3923 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3925 *) Sanity check record length before skipping explicit IV in TLS
3926 1.2, 1.1 and DTLS to fix DoS attack.
3928 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3929 fuzzing as a service testing platform.
3933 *) Initialise tkeylen properly when encrypting CMS messages.
3934 Thanks to Solar Designer of Openwall for reporting this issue.
3937 *) In FIPS mode don't try to use composite ciphers as they are not
3941 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3943 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3944 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3945 mean any application compiled against OpenSSL 1.0.0 headers setting
3946 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3947 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3948 0x10000000L Any application which was previously compiled against
3949 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3950 will need to be recompiled as a result. Letting be results in
3951 inability to disable specifically TLS 1.1 and in client context,
3952 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3955 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3956 disable just protocol X, but all protocols above X *if* there are
3957 protocols *below* X still enabled. In more practical terms it means
3958 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3959 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3960 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3964 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3966 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3967 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3968 in CRYPTO_realloc_clean.
3970 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3971 issue and to Adam Langley <agl@chromium.org> for fixing it.
3973 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3975 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3978 *) Workarounds for some broken servers that "hang" if a client hello
3979 record length exceeds 255 bytes.
3981 1. Do not use record version number > TLS 1.0 in initial client
3982 hello: some (but not all) hanging servers will now work.
3983 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3984 the number of ciphers sent in the client hello. This should be
3985 set to an even number, such as 50, for example by passing:
3986 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3987 Most broken servers should now work.
3988 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3989 TLS 1.2 client support entirely.
3992 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3995 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3997 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3998 STRING form instead of a DigestInfo.
4001 *) The format used for MDC2 RSA signatures is inconsistent between EVP
4002 and the RSA_sign/RSA_verify functions. This was made more apparent when
4003 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
4004 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
4005 the correct format in RSA_verify so both forms transparently work.
4008 *) Some servers which support TLS 1.0 can choke if we initially indicate
4009 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
4010 encrypted premaster secret. As a workaround use the maximum permitted
4011 client version in client hello, this should keep such servers happy
4012 and still work with previous versions of OpenSSL.
4015 *) Add support for TLS/DTLS heartbeats.
4016 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4018 *) Add support for SCTP.
4019 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4021 *) Improved PRNG seeding for VOS.
4022 [Paul Green <Paul.Green@stratus.com>]
4024 *) Extensive assembler packs updates, most notably:
4026 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
4027 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
4028 - x86_64: bit-sliced AES implementation;
4029 - ARM: NEON support, contemporary platforms optimizations;
4030 - s390x: z196 support;
4031 - *: GHASH and GF(2^m) multiplication implementations;
4035 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
4036 (removal of unnecessary code)
4037 [Peter Sylvester <peter.sylvester@edelweb.fr>]
4039 *) Add TLS key material exporter from RFC 5705.
4042 *) Add DTLS-SRTP negotiation from RFC 5764.
4045 *) Add Next Protocol Negotiation,
4046 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
4047 disabled with a no-npn flag to config or Configure. Code donated
4049 [Adam Langley <agl@google.com> and Ben Laurie]
4051 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
4052 NIST-P256, NIST-P521, with constant-time single point multiplication on
4053 typical inputs. Compiler support for the nonstandard type __uint128_t is
4054 required to use this (present in gcc 4.4 and later, for 64-bit builds).
4055 Code made available under Apache License version 2.0.
4057 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
4058 line to include this in your build of OpenSSL, and run "make depend" (or
4059 "make update"). This enables the following EC_METHODs:
4061 EC_GFp_nistp224_method()
4062 EC_GFp_nistp256_method()
4063 EC_GFp_nistp521_method()
4065 EC_GROUP_new_by_curve_name() will automatically use these (while
4066 EC_GROUP_new_curve_GFp() currently prefers the more flexible
4068 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
4070 *) Use type ossl_ssize_t instead of ssize_t which isn't available on
4071 all platforms. Move ssize_t definition from e_os.h to the public
4072 header file e_os2.h as it now appears in public header file cms.h
4075 *) New -sigopt option to the ca, req and x509 utilities. Additional
4076 signature parameters can be passed using this option and in
4080 *) Add RSA PSS signing function. This will generate and set the
4081 appropriate AlgorithmIdentifiers for PSS based on those in the
4082 corresponding EVP_MD_CTX structure. No application support yet.
4085 *) Support for companion algorithm specific ASN1 signing routines.
4086 New function ASN1_item_sign_ctx() signs a pre-initialised
4087 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
4088 the appropriate parameters.
4091 *) Add new algorithm specific ASN1 verification initialisation function
4092 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
4093 handling will be the same no matter what EVP_PKEY_METHOD is used.
4094 Add a PSS handler to support verification of PSS signatures: checked
4095 against a number of sample certificates.
4098 *) Add signature printing for PSS. Add PSS OIDs.
4099 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
4101 *) Add algorithm specific signature printing. An individual ASN1 method
4102 can now print out signatures instead of the standard hex dump.
4104 More complex signatures (e.g. PSS) can print out more meaningful
4105 information. Include DSA version that prints out the signature
4109 *) Password based recipient info support for CMS library: implementing
4113 *) Split password based encryption into PBES2 and PBKDF2 functions. This
4114 neatly separates the code into cipher and PBE sections and is required
4115 for some algorithms that split PBES2 into separate pieces (such as
4116 password based CMS).
4119 *) Session-handling fixes:
4120 - Fix handling of connections that are resuming with a session ID,
4121 but also support Session Tickets.
4122 - Fix a bug that suppressed issuing of a new ticket if the client
4123 presented a ticket with an expired session.
4124 - Try to set the ticket lifetime hint to something reasonable.
4125 - Make tickets shorter by excluding irrelevant information.
4126 - On the client side, don't ignore renewed tickets.
4127 [Adam Langley, Bodo Moeller (Google)]
4129 *) Fix PSK session representation.
4132 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
4134 This work was sponsored by Intel.
4137 *) Add GCM support to TLS library. Some custom code is needed to split
4138 the IV between the fixed (from PRF) and explicit (from TLS record)
4139 portions. This adds all GCM ciphersuites supported by RFC5288 and
4140 RFC5289. Generalise some AES* cipherstrings to include GCM and
4141 add a special AESGCM string for GCM only.
4144 *) Expand range of ctrls for AES GCM. Permit setting invocation
4145 field on decrypt and retrieval of invocation field only on encrypt.
4148 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
4149 As required by RFC5289 these ciphersuites cannot be used if for
4150 versions of TLS earlier than 1.2.
4153 *) For FIPS capable OpenSSL interpret a NULL default public key method
4154 as unset and return the appropriate default but do *not* set the default.
4155 This means we can return the appropriate method in applications that
4156 switch between FIPS and non-FIPS modes.
4159 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
4160 ENGINE is used then we cannot handle that in the FIPS module so we
4161 keep original code iff non-FIPS operations are allowed.
4164 *) Add -attime option to openssl utilities.
4165 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
4167 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
4170 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
4171 FIPS EC methods unconditionally for now.
4174 *) New build option no-ec2m to disable characteristic 2 code.
4177 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
4178 all cases can be covered as some introduce binary incompatibilities.
4181 *) Redirect RSA operations to FIPS module including keygen,
4182 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
4185 *) Add similar low level API blocking to ciphers.
4188 *) Low level digest APIs are not approved in FIPS mode: any attempt
4189 to use these will cause a fatal error. Applications that *really* want
4190 to use them can use the private_* version instead.
4193 *) Redirect cipher operations to FIPS module for FIPS builds.
4196 *) Redirect digest operations to FIPS module for FIPS builds.
4199 *) Update build system to add "fips" flag which will link in fipscanister.o
4200 for static and shared library builds embedding a signature if needed.
4203 *) Output TLS supported curves in preference order instead of numerical
4204 order. This is currently hardcoded for the highest order curves first.
4205 This should be configurable so applications can judge speed vs strength.
4208 *) Add TLS v1.2 server support for client authentication.
4211 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
4215 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
4216 FIPS modules versions.
4219 *) Add TLS v1.2 client side support for client authentication. Keep cache
4220 of handshake records longer as we don't know the hash algorithm to use
4221 until after the certificate request message is received.
4224 *) Initial TLS v1.2 client support. Add a default signature algorithms
4225 extension including all the algorithms we support. Parse new signature
4226 format in client key exchange. Relax some ECC signing restrictions for
4227 TLS v1.2 as indicated in RFC5246.
4230 *) Add server support for TLS v1.2 signature algorithms extension. Switch
4231 to new signature format when needed using client digest preference.
4232 All server ciphersuites should now work correctly in TLS v1.2. No client
4233 support yet and no support for client certificates.
4236 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
4237 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
4238 ciphersuites. At present only RSA key exchange ciphersuites work with
4239 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
4240 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
4241 and version checking.
4244 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
4245 with this defined it will not be affected by any changes to ssl internal
4246 structures. Add several utility functions to allow openssl application
4247 to work with OPENSSL_NO_SSL_INTERN defined.
4250 *) A long standing patch to add support for SRP from EdelWeb (Peter
4251 Sylvester and Christophe Renou) was integrated.
4252 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
4253 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
4256 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
4259 *) Permit abbreviated handshakes when renegotiating using the function
4260 SSL_renegotiate_abbreviated().
4261 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4263 *) Add call to ENGINE_register_all_complete() to
4264 ENGINE_load_builtin_engines(), so some implementations get used
4265 automatically instead of needing explicit application support.
4268 *) Add support for TLS key exporter as described in RFC5705.
4269 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
4271 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
4272 a few changes are required:
4274 Add SSL_OP_NO_TLSv1_1 flag.
4275 Add TLSv1_1 methods.
4276 Update version checking logic to handle version 1.1.
4277 Add explicit IV handling (ported from DTLS code).
4278 Add command line options to s_client/s_server.
4281 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
4283 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
4284 in CMS and PKCS7 code. When RSA decryption fails use a random key for
4285 content decryption and always return the same error. Note: this attack
4286 needs on average 2^20 messages so it only affects automated senders. The
4287 old behaviour can be re-enabled in the CMS code by setting the
4288 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
4289 an MMA defence is not necessary.
4290 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
4291 this issue. (CVE-2012-0884)
4294 *) Fix CVE-2011-4619: make sure we really are receiving a
4295 client hello before rejecting multiple SGC restarts. Thanks to
4296 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
4299 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
4301 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
4302 Thanks to Antonio Martin, Enterprise Secure Access Research and
4303 Development, Cisco Systems, Inc. for discovering this bug and
4304 preparing a fix. (CVE-2012-0050)
4307 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
4309 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
4310 of the Vaudenay padding oracle attack on CBC mode encryption
4311 which enables an efficient plaintext recovery attack against
4312 the OpenSSL implementation of DTLS. Their attack exploits timing
4313 differences arising during decryption processing. A research
4314 paper describing this attack can be found at:
4315 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
4316 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4317 Security Group at Royal Holloway, University of London
4318 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
4319 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
4320 for preparing the fix. (CVE-2011-4108)
4321 [Robin Seggelmann, Michael Tuexen]
4323 *) Clear bytes used for block padding of SSL 3.0 records.
4325 [Adam Langley (Google)]
4327 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
4328 Kadianakis <desnacked@gmail.com> for discovering this issue and
4329 Adam Langley for preparing the fix. (CVE-2011-4619)
4330 [Adam Langley (Google)]
4332 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
4333 [Andrey Kulikov <amdeich@gmail.com>]
4335 *) Prevent malformed RFC3779 data triggering an assertion failure.
4336 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
4337 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
4338 [Rob Austein <sra@hactrn.net>]
4340 *) Improved PRNG seeding for VOS.
4341 [Paul Green <Paul.Green@stratus.com>]
4343 *) Fix ssl_ciph.c set-up race.
4344 [Adam Langley (Google)]
4346 *) Fix spurious failures in ecdsatest.c.
4347 [Emilia Käsper (Google)]
4349 *) Fix the BIO_f_buffer() implementation (which was mixing different
4350 interpretations of the '..._len' fields).
4351 [Adam Langley (Google)]
4353 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4354 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4355 threads won't reuse the same blinding coefficients.
4357 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4358 lock to call BN_BLINDING_invert_ex, and avoids one use of
4359 BN_BLINDING_update for each BN_BLINDING structure (previously,
4360 the last update always remained unused).
4361 [Emilia Käsper (Google)]
4363 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4364 [Bob Buckholz (Google)]
4366 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4368 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4369 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4370 [Kaspar Brand <ossl@velox.ch>]
4372 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4373 for multi-threaded use of ECDH. (CVE-2011-3210)
4374 [Adam Langley (Google)]
4376 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4379 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4380 signature public key algorithm by using OID xref utilities instead.
4381 Before this you could only use some ECC ciphersuites with SHA1 only.
4384 *) Add protection against ECDSA timing attacks as mentioned in the paper
4385 by Billy Bob Brumley and Nicola Tuveri, see:
4387 http://eprint.iacr.org/2011/232.pdf
4389 [Billy Bob Brumley and Nicola Tuveri]
4391 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4393 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4394 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4396 *) Fix bug in string printing code: if *any* escaping is enabled we must
4397 escape the escape character (backslash) or the resulting string is
4401 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4403 *) Disable code workaround for ancient and obsolete Netscape browsers
4404 and servers: an attacker can use it in a ciphersuite downgrade attack.
4405 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4408 *) Fixed J-PAKE implementation error, originally discovered by
4409 Sebastien Martini, further info and confirmation from Stefan
4410 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4413 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4415 *) Fix extension code to avoid race conditions which can result in a buffer
4416 overrun vulnerability: resumed sessions must not be modified as they can
4417 be shared by multiple threads. CVE-2010-3864
4420 *) Fix WIN32 build system to correctly link an ENGINE directory into
4424 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4426 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4428 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4430 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4432 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4433 context. The operation can be customised via the ctrl mechanism in
4434 case ENGINEs want to include additional functionality.
4437 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4440 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4441 output hashes compatible with older versions of OpenSSL.
4442 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4444 *) Fix compression algorithm handling: if resuming a session use the
4445 compression algorithm of the resumed session instead of determining
4446 it from client hello again. Don't allow server to change algorithm.
4449 *) Add load_crls() function to apps tidying load_certs() too. Add option
4450 to verify utility to allow additional CRLs to be included.
4453 *) Update OCSP request code to permit adding custom headers to the request:
4454 some responders need this.
4457 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4459 [Julia Lawall <julia@diku.dk>]
4461 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4462 needlessly dereferenced structures, used obsolete functions and
4463 didn't handle all updated verify codes correctly.
4466 *) Disable MD2 in the default configuration.
4469 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4470 indicate the initial BIO being pushed or popped. This makes it possible
4471 to determine whether the BIO is the one explicitly called or as a result
4472 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4473 it handles reference counts correctly and doesn't zero out the I/O bio
4474 when it is not being explicitly popped. WARNING: applications which
4475 included workarounds for the old buggy behaviour will need to be modified
4476 or they could free up already freed BIOs.
4479 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4480 renaming to all platforms (within the 0.9.8 branch, this was
4481 done conditionally on Netware platforms to avoid a name clash).
4482 [Guenter <lists@gknw.net>]
4484 *) Add ECDHE and PSK support to DTLS.
4485 [Michael Tuexen <tuexen@fh-muenster.de>]
4487 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4491 *) Add "missing" function EVP_MD_flags() (without this the only way to
4492 retrieve a digest flags is by accessing the structure directly. Update
4493 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4494 or cipher is registered as in the "from" argument. Print out all
4495 registered digests in the dgst usage message instead of manually
4496 attempting to work them out.
4499 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4500 this allows the use of compression and extensions. Change default cipher
4501 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4502 by default unless an application cipher string requests it.
4505 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4506 key ids to find matching certificates and keys but some PKCS#12 files
4507 don't follow the (somewhat unwritten) rules and this strategy fails.
4508 Now just gather all certificates together and the first private key
4509 then look for the first certificate that matches the key.
4512 *) Support use of registered digest and cipher names for dgst and cipher
4513 commands instead of having to add each one as a special case. So now
4520 openssl dgst -sha256 foo
4522 and this works for ENGINE based algorithms too.
4526 *) Update Gost ENGINE to support parameter files.
4527 [Victor B. Wagner <vitus@cryptocom.ru>]
4529 *) Support GeneralizedTime in ca utility.
4530 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4532 *) Enhance the hash format used for certificate directory links. The new
4533 form uses the canonical encoding (meaning equivalent names will work
4534 even if they aren't identical) and uses SHA1 instead of MD5. This form
4535 is incompatible with the older format and as a result c_rehash should
4536 be used to rebuild symbolic links.
4539 *) Make PKCS#8 the default write format for private keys, replacing the
4540 traditional format. This form is standardised, more secure and doesn't
4541 include an implicit MD5 dependency.
4544 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4545 committed to OpenSSL should pass this lot as a minimum.
4548 *) Add session ticket override functionality for use by EAP-FAST.
4549 [Jouni Malinen <j@w1.fi>]
4551 *) Modify HMAC functions to return a value. Since these can be implemented
4552 in an ENGINE errors can occur.
4555 *) Type-checked OBJ_bsearch_ex.
4558 *) Type-checked OBJ_bsearch. Also some constification necessitated
4559 by type-checking. Still to come: TXT_DB, bsearch(?),
4560 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4564 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4565 seconds to a tm structure directly, instead of going through OS
4566 specific date routines. This avoids any issues with OS routines such
4567 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4568 and X509_time_adj_ex() to cover the extended range. The existing
4569 X509_time_adj() is still usable and will no longer have any date issues.
4572 *) Delta CRL support. New use deltas option which will attempt to locate
4573 and search any appropriate delta CRLs available.
4575 This work was sponsored by Google.
4578 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4579 code and add additional score elements. Validate alternate CRL paths
4580 as part of the CRL checking and indicate a new error "CRL path validation
4581 error" in this case. Applications wanting additional details can use
4582 the verify callback and check the new "parent" field. If this is not
4583 NULL CRL path validation is taking place. Existing applications won't
4584 see this because it requires extended CRL support which is off by
4587 This work was sponsored by Google.
4590 *) Support for freshest CRL extension.
4592 This work was sponsored by Google.
4595 *) Initial indirect CRL support. Currently only supported in the CRLs
4596 passed directly and not via lookup. Process certificate issuer
4597 CRL entry extension and lookup CRL entries by bother issuer name
4598 and serial number. Check and process CRL issuer entry in IDP extension.
4600 This work was sponsored by Google.
4603 *) Add support for distinct certificate and CRL paths. The CRL issuer
4604 certificate is validated separately in this case. Only enabled if
4605 an extended CRL support flag is set: this flag will enable additional
4606 CRL functionality in future.
4608 This work was sponsored by Google.
4611 *) Add support for policy mappings extension.
4613 This work was sponsored by Google.
4616 *) Fixes to pathlength constraint, self issued certificate handling,
4617 policy processing to align with RFC3280 and PKITS tests.
4619 This work was sponsored by Google.
4622 *) Support for name constraints certificate extension. DN, email, DNS
4623 and URI types are currently supported.
4625 This work was sponsored by Google.
4628 *) To cater for systems that provide a pointer-based thread ID rather
4629 than numeric, deprecate the current numeric thread ID mechanism and
4630 replace it with a structure and associated callback type. This
4631 mechanism allows a numeric "hash" to be extracted from a thread ID in
4632 either case, and on platforms where pointers are larger than 'long',
4633 mixing is done to help ensure the numeric 'hash' is usable even if it
4634 can't be guaranteed unique. The default mechanism is to use "&errno"
4635 as a pointer-based thread ID to distinguish between threads.
4637 Applications that want to provide their own thread IDs should now use
4638 CRYPTO_THREADID_set_callback() to register a callback that will call
4639 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4641 Note that ERR_remove_state() is now deprecated, because it is tied
4642 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4643 to free the current thread's error state should be replaced by
4644 ERR_remove_thread_state(NULL).
4646 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4647 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4648 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4649 application was previously providing a numeric thread callback that
4650 was inappropriate for distinguishing threads, then uniqueness might
4651 have been obtained with &errno that happened immediately in the
4652 intermediate development versions of OpenSSL; this is no longer the
4653 case, the numeric thread callback will now override the automatic use
4655 [Geoff Thorpe, with help from Bodo Moeller]
4657 *) Initial support for different CRL issuing certificates. This covers a
4658 simple case where the self issued certificates in the chain exist and
4659 the real CRL issuer is higher in the existing chain.
4661 This work was sponsored by Google.
4664 *) Removed effectively defunct crypto/store from the build.
4667 *) Revamp of STACK to provide stronger type-checking. Still to come:
4668 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4669 ASN1_STRING, CONF_VALUE.
4672 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4673 RAM on SSL connections. This option can save about 34k per idle SSL.
4676 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4677 STACK, TXT_DB, bsearch, qsort.
4680 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4681 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4682 support for data, signedData, compressedData, digestedData and
4683 encryptedData, envelopedData types included. Scripts to check against
4684 RFC4134 examples draft and interop and consistency checks of many
4685 content types and variants.
4688 *) Add options to enc utility to support use of zlib compression BIO.
4691 *) Extend mk1mf to support importing of options and assembly language
4692 files from Configure script, currently only included in VC-WIN32.
4693 The assembly language rules can now optionally generate the source
4694 files from the associated perl scripts.
4697 *) Implement remaining functionality needed to support GOST ciphersuites.
4698 Interop testing has been performed using CryptoPro implementations.
4699 [Victor B. Wagner <vitus@cryptocom.ru>]
4701 *) s390x assembler pack.
4704 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4708 *) Implement Opaque PRF Input TLS extension as specified in
4709 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4710 official specification yet and no extension type assignment by
4711 IANA exists, this extension (for now) will have to be explicitly
4712 enabled when building OpenSSL by providing the extension number
4713 to use. For example, specify an option
4715 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4717 to the "config" or "Configure" script to enable the extension,
4718 assuming extension number 0x9527 (which is a completely arbitrary
4719 and unofficial assignment based on the MD5 hash of the Internet
4720 Draft). Note that by doing so, you potentially lose
4721 interoperability with other TLS implementations since these might
4722 be using the same extension number for other purposes.
4724 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4725 opaque PRF input value to use in the handshake. This will create
4726 an internal copy of the length-'len' string at 'src', and will
4727 return non-zero for success.
4729 To get more control and flexibility, provide a callback function
4732 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4733 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4737 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4740 Callback function 'cb' will be called in handshakes, and is
4741 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4742 Argument 'arg' is for application purposes (the value as given to
4743 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4744 be provided to the callback function). The callback function
4745 has to return non-zero to report success: usually 1 to use opaque
4746 PRF input just if possible, or 2 to enforce use of the opaque PRF
4747 input. In the latter case, the library will abort the handshake
4748 if opaque PRF input is not successfully negotiated.
4750 Arguments 'peerinput' and 'len' given to the callback function
4751 will always be NULL and 0 in the case of a client. A server will
4752 see the client's opaque PRF input through these variables if
4753 available (NULL and 0 otherwise). Note that if the server
4754 provides an opaque PRF input, the length must be the same as the
4755 length of the client's opaque PRF input.
4757 Note that the callback function will only be called when creating
4758 a new session (session resumption can resume whatever was
4759 previously negotiated), and will not be called in SSL 2.0
4760 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4761 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4762 for applications that need to enforce opaque PRF input.
4766 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4769 [Victor B. Wagner <vitus@cryptocom.ru>]
4771 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4772 RFC4507bis. The encrypted ticket format is an encrypted encoded
4773 SSL_SESSION structure, that way new session features are automatically
4776 If a client application caches session in an SSL_SESSION structure
4777 support is transparent because tickets are now stored in the encoded
4780 The SSL_CTX structure automatically generates keys for ticket
4781 protection in servers so again support should be possible
4782 with no application modification.
4784 If a client or server wishes to disable RFC4507 support then the option
4785 SSL_OP_NO_TICKET can be set.
4787 Add a TLS extension debugging callback to allow the contents of any client
4788 or server extensions to be examined.
4790 This work was sponsored by Google.
4793 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4794 OpenSSL should now compile cleanly on gcc 4.2
4795 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4797 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4798 support including streaming MAC support: this is required for GOST
4799 ciphersuite support.
4800 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4802 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4803 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4804 to output in BER and PEM format.
4807 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4808 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4809 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4810 ENGINE support for HMAC keys which are unextractable. New -mac and
4811 -macopt options to dgst utility.
4814 *) New option -sigopt to dgst utility. Update dgst to use
4815 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4816 alternative signing parameters such as X9.31 or PSS in the dgst
4820 *) Change ssl_cipher_apply_rule(), the internal function that does
4821 the work each time a ciphersuite string requests enabling
4822 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4823 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4824 the order of disabled ciphersuites such that those ciphersuites
4825 that most recently went from enabled to disabled not only stay
4826 in order with respect to each other, but also have higher priority
4827 than other disabled ciphersuites the next time ciphersuites are
4830 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4831 the same ciphersuites as with "HIGH" alone, but in a specific
4832 order where the PSK ciphersuites come first (since they are the
4833 most recently disabled ciphersuites when "HIGH" is parsed).
4835 Also, change ssl_create_cipher_list() (using this new
4836 functionality) such that between otherwise identical
4837 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4841 *) Change ssl_create_cipher_list() so that it automatically
4842 arranges the ciphersuites in reasonable order before starting
4843 to process the rule string. Thus, the definition for "DEFAULT"
4844 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4845 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4846 This makes it much easier to arrive at a reasonable default order
4847 in applications for which anonymous ciphers are OK (meaning
4848 that you can't actually use DEFAULT).
4849 [Bodo Moeller; suggested by Victor Duchovni]
4851 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4852 processing) into multiple integers instead of setting
4853 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4854 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4855 (These masks as well as the individual bit definitions are hidden
4856 away into the non-exported interface ssl/ssl_locl.h, so this
4857 change to the definition of the SSL_CIPHER structure shouldn't
4858 affect applications.) This give us more bits for each of these
4859 categories, so there is no longer a need to coagulate AES128 and
4860 AES256 into a single algorithm bit, and to coagulate Camellia128
4861 and Camellia256 into a single algorithm bit, which has led to all
4864 Thus, among other things, the kludge introduced in 0.9.7m and
4865 0.9.8e for masking out AES256 independently of AES128 or masking
4866 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4868 With the change, we also introduce new ciphersuite aliases that
4869 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4873 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4874 Use the leftmost N bytes of the signature input if the input is
4875 larger than the prime q (with N being the size in bytes of q).
4878 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4879 it yet and it is largely untested.
4882 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4885 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4886 some compilers (gcc 4.2 and later) reject their use. Safestack is
4887 reimplemented. Update ASN1 to avoid use of legacy functions.
4890 *) Win32/64 targets are linked with Winsock2.
4893 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4894 to external functions. This can be used to increase CRL handling
4895 efficiency especially when CRLs are very large by (for example) storing
4896 the CRL revoked certificates in a database.
4899 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4900 new CRLs added to a directory can be used. New command line option
4901 -verify_return_error to s_client and s_server. This causes real errors
4902 to be returned by the verify callback instead of carrying on no matter
4903 what. This reflects the way a "real world" verify callback would behave.
4906 *) GOST engine, supporting several GOST algorithms and public key formats.
4907 Kindly donated by Cryptocom.
4910 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4911 partitioned by DP are handled but no indirect CRL or reason partitioning
4912 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4913 selected via a scoring technique which handles IDP and AKID in CRLs.
4916 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4917 will ultimately be used for all verify operations: this will remove the
4918 X509_STORE dependency on certificate verification and allow alternative
4919 lookup methods. X509_STORE based implementations of these two callbacks.
4922 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4923 Modify get_crl() to find a valid (unexpired) CRL if possible.
4926 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4927 this would be called X509_CRL_cmp() but that name is already used by
4928 a function that just compares CRL issuer names. Cache several CRL
4929 extensions in X509_CRL structure and cache CRLDP in X509.
4932 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4933 this maps equivalent X509_NAME structures into a consistent structure.
4934 Name comparison can then be performed rapidly using memcmp().
4937 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4941 *) Allow digests to supply their own micalg string for S/MIME type using
4942 the ctrl EVP_MD_CTRL_MICALG.
4945 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4946 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4947 ctrl. It can then customise the structure before and/or after signing
4951 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4952 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4953 to free up any added signature OIDs.
4956 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4957 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4958 digest and cipher tables. New options added to openssl utility:
4959 list-message-digest-algorithms and list-cipher-algorithms.
4962 *) Change the array representation of binary polynomials: the list
4963 of degrees of non-zero coefficients is now terminated with -1.
4964 Previously it was terminated with 0, which was also part of the
4965 value; thus, the array representation was not applicable to
4966 polynomials where t^0 has coefficient zero. This change makes
4967 the array representation useful in a more general context.
4970 *) Various modifications and fixes to SSL/TLS cipher string
4971 handling. For ECC, the code now distinguishes between fixed ECDH
4972 with RSA certificates on the one hand and with ECDSA certificates
4973 on the other hand, since these are separate ciphersuites. The
4974 unused code for Fortezza ciphersuites has been removed.
4976 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4977 (not "ECDHE"). For consistency with the code for DH
4978 certificates, use of ECDH certificates is now considered ECDH
4979 authentication, not RSA or ECDSA authentication (the latter is
4980 merely the CA's signing algorithm and not actively used in the
4983 The temporary ciphersuite alias "ECCdraft" is no longer
4984 available, and ECC ciphersuites are no longer excluded from "ALL"
4985 and "DEFAULT". The following aliases now exist for RFC 4492
4986 ciphersuites, most of these by analogy with the DH case:
4988 kECDHr - ECDH cert, signed with RSA
4989 kECDHe - ECDH cert, signed with ECDSA
4990 kECDH - ECDH cert (signed with either RSA or ECDSA)
4991 kEECDH - ephemeral ECDH
4992 ECDH - ECDH cert or ephemeral ECDH
4998 AECDH - anonymous ECDH
4999 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
5003 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
5004 Use correct micalg parameters depending on digest(s) in signed message.
5007 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
5008 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
5011 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
5012 an engine to register a method. Add ENGINE lookups for methods and
5013 functional reference processing.
5016 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
5017 EVP_{Sign,Verify}* which allow an application to customise the signature
5021 *) New -resign option to smime utility. This adds one or more signers
5022 to an existing PKCS#7 signedData structure. Also -md option to use an
5023 alternative message digest algorithm for signing.
5026 *) Tidy up PKCS#7 routines and add new functions to make it easier to
5027 create PKCS7 structures containing multiple signers. Update smime
5028 application to support multiple signers.
5031 *) New -macalg option to pkcs12 utility to allow setting of an alternative
5035 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
5036 Reorganize PBE internals to lookup from a static table using NIDs,
5037 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
5038 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
5039 PRF which will be automatically used with PBES2.
5042 *) Replace the algorithm specific calls to generate keys in "req" with the
5046 *) Update PKCS#7 enveloped data routines to use new API. This is now
5047 supported by any public key method supporting the encrypt operation. A
5048 ctrl is added to allow the public key algorithm to examine or modify
5049 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
5053 *) Add a ctrl to asn1 method to allow a public key algorithm to express
5054 a default digest type to use. In most cases this will be SHA1 but some
5055 algorithms (such as GOST) need to specify an alternative digest. The
5056 return value indicates how strong the preference is 1 means optional and
5057 2 is mandatory (that is it is the only supported type). Modify
5058 ASN1_item_sign() to accept a NULL digest argument to indicate it should
5059 use the default md. Update openssl utilities to use the default digest
5060 type for signing if it is not explicitly indicated.
5063 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
5064 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
5065 signing method from the key type. This effectively removes the link
5066 between digests and public key types.
5069 *) Add an OID cross reference table and utility functions. Its purpose is to
5070 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
5071 rsaEncryption. This will allow some of the algorithm specific hackery
5072 needed to use the correct OID to be removed.
5075 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
5076 structures for PKCS7_sign(). They are now set up by the relevant public
5080 *) Add provisional EC pkey method with support for ECDSA and ECDH.
5083 *) Add support for key derivation (agreement) in the API, DH method and
5087 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
5088 public and private key formats. As a side effect these add additional
5089 command line functionality not previously available: DSA signatures can be
5090 generated and verified using pkeyutl and DH key support and generation in
5095 [Oliver Tappe <zooey@hirschkaefer.de>]
5097 *) New make target "install_html_docs" installs HTML renditions of the
5099 [Oliver Tappe <zooey@hirschkaefer.de>]
5101 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
5102 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
5103 support key and parameter generation and add initial key generation
5104 functionality for RSA.
5107 *) Add functions for main EVP_PKEY_method operations. The undocumented
5108 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
5109 EVP_PKEY_{encrypt,decrypt}_old.
5112 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
5113 key API, doesn't do much yet.
5116 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
5117 public key algorithms. New option to openssl utility:
5118 "list-public-key-algorithms" to print out info.
5121 *) Implement the Supported Elliptic Curves Extension for
5122 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5125 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
5126 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
5129 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
5130 utilities such as rsa, dsa, dsaparam etc except they process any key
5134 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
5135 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
5136 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
5140 *) Initial support for pluggable public key ASN1.
5141 De-spaghettify the public key ASN1 handling. Move public and private
5142 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
5143 algorithm specific handling to a single module within the relevant
5144 algorithm directory. Add functions to allow (near) opaque processing
5145 of public and private key structures.
5148 *) Implement the Supported Point Formats Extension for
5149 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5152 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
5153 for the psk identity [hint] and the psk callback functions to the
5154 SSL_SESSION, SSL and SSL_CTX structure.
5157 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
5161 SSL_CTX_use_psk_identity_hint
5162 SSL_get_psk_identity_hint
5163 SSL_get_psk_identity
5164 SSL_use_psk_identity_hint
5166 [Mika Kousa and Pasi Eronen of Nokia Corporation]
5168 *) Add RFC 3161 compliant time stamp request creation, response generation
5169 and response verification functionality.
5170 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
5172 *) Add initial support for TLS extensions, specifically for the server_name
5173 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5174 have new members for a host name. The SSL data structure has an
5175 additional member SSL_CTX *initial_ctx so that new sessions can be
5176 stored in that context to allow for session resumption, even after the
5177 SSL has been switched to a new SSL_CTX in reaction to a client's
5178 server_name extension.
5180 New functions (subject to change):
5182 SSL_get_servername()
5183 SSL_get_servername_type()
5186 New CTRL codes and macros (subject to change):
5188 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5189 - SSL_CTX_set_tlsext_servername_callback()
5190 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5191 - SSL_CTX_set_tlsext_servername_arg()
5192 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5194 openssl s_client has a new '-servername ...' option.
5196 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5197 '-key2 ...', '-servername_fatal' (subject to change). This allows
5198 testing the HostName extension for a specific single host name ('-cert'
5199 and '-key' remain fallbacks for handshakes without HostName
5200 negotiation). If the unrecognized_name alert has to be sent, this by
5201 default is a warning; it becomes fatal with the '-servername_fatal'
5204 [Peter Sylvester, Remy Allais, Christophe Renou]
5206 *) Whirlpool hash implementation is added.
5209 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
5210 bn(64,32). Because of instruction set limitations it doesn't have
5211 any negative impact on performance. This was done mostly in order
5212 to make it possible to share assembler modules, such as bn_mul_mont
5213 implementations, between 32- and 64-bit builds without hassle.
5216 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
5217 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
5221 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
5222 dedicated Montgomery multiplication procedure, is introduced.
5223 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
5224 "64-bit" performance on certain 32-bit targets.
5227 *) New option SSL_OP_NO_COMP to disable use of compression selectively
5228 in SSL structures. New SSL ctrl to set maximum send fragment size.
5229 Save memory by setting the I/O buffer sizes dynamically instead of
5230 using the maximum available value.
5233 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
5234 in addition to the text details.
5237 *) Very, very preliminary EXPERIMENTAL support for printing of general
5238 ASN1 structures. This currently produces rather ugly output and doesn't
5239 handle several customised structures at all.
5242 *) Integrated support for PVK file format and some related formats such
5243 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
5244 these in the 'rsa' and 'dsa' utilities.
5247 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
5250 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
5251 place for the (very old) "NETSCAPE" format certificates which are now
5252 handled using new ASN1 code equivalents.
5255 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
5256 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
5257 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
5260 *) Modify CRL distribution points extension code to print out previously
5261 unsupported fields. Enhance extension setting code to allow setting of
5265 *) Add print and set support for Issuing Distribution Point CRL extension.
5268 *) Change 'Configure' script to enable Camellia by default.
5271 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
5273 *) When rejecting SSL/TLS records due to an incorrect version number, never
5274 update s->server with a new major version number. As of
5275 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
5276 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
5277 the previous behavior could result in a read attempt at NULL when
5278 receiving specific incorrect SSL/TLS records once record payload
5279 protection is active. (CVE-2010-0740)
5280 [Bodo Moeller, Adam Langley <agl@chromium.org>]
5282 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
5283 could be crashed if the relevant tables were not present (e.g. chrooted).
5284 [Tomas Hoger <thoger@redhat.com>]
5286 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
5288 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
5289 [Martin Olsson, Neel Mehta]
5291 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
5292 accommodate for stack sorting, always a write lock!).
5295 *) On some versions of WIN32 Heap32Next is very slow. This can cause
5296 excessive delays in the RAND_poll(): over a minute. As a workaround
5297 include a time check in the inner Heap32Next loop too.
5300 *) The code that handled flushing of data in SSL/TLS originally used the
5301 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
5302 the problem outlined in PR#1949. The fix suggested there however can
5303 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
5304 of Apache). So instead simplify the code to flush unconditionally.
5305 This should be fine since flushing with no data to flush is a no op.
5308 *) Handle TLS versions 2.0 and later properly and correctly use the
5309 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
5310 off ancient servers have a habit of sticking around for a while...
5313 *) Modify compression code so it frees up structures without using the
5314 ex_data callbacks. This works around a problem where some applications
5315 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
5316 restarting) then use compression (e.g. SSL with compression) later.
5317 This results in significant per-connection memory leaks and
5318 has caused some security issues including CVE-2008-1678 and
5322 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
5323 change when encrypting or decrypting.
5326 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
5327 connect and renegotiate with servers which do not support RI.
5328 Until RI is more widely deployed this option is enabled by default.
5331 *) Add "missing" ssl ctrls to clear options and mode.
5334 *) If client attempts to renegotiate and doesn't support RI respond with
5335 a no_renegotiation alert as required by RFC5746. Some renegotiating
5336 TLS clients will continue a connection gracefully when they receive
5337 the alert. Unfortunately OpenSSL mishandled this alert and would hang
5338 waiting for a server hello which it will never receive. Now we treat a
5339 received no_renegotiation alert as a fatal error. This is because
5340 applications requesting a renegotiation might well expect it to succeed
5341 and would have no code in place to handle the server denying it so the
5342 only safe thing to do is to terminate the connection.
5345 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
5346 peer supports secure renegotiation and 0 otherwise. Print out peer
5347 renegotiation support in s_client/s_server.
5350 *) Replace the highly broken and deprecated SPKAC certification method with
5351 the updated NID creation version. This should correctly handle UTF8.
5354 *) Implement RFC5746. Re-enable renegotiation but require the extension
5355 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
5356 turns out to be a bad idea. It has been replaced by
5357 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5358 SSL_CTX_set_options(). This is really not recommended unless you
5359 know what you are doing.
5360 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5362 *) Fixes to stateless session resumption handling. Use initial_ctx when
5363 issuing and attempting to decrypt tickets in case it has changed during
5364 servername handling. Use a non-zero length session ID when attempting
5365 stateless session resumption: this makes it possible to determine if
5366 a resumption has occurred immediately after receiving server hello
5367 (several places in OpenSSL subtly assume this) instead of later in
5371 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5372 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5373 fixes for a few places where the return code is not checked
5375 [Julia Lawall <julia@diku.dk>]
5377 *) Add --strict-warnings option to Configure script to include devteam
5378 warnings in other configurations.
5381 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5382 makes it possible to install openssl libraries in locations which
5383 have names other than "lib", for example "/usr/lib64" which some
5385 [Steve Henson, based on patch from Jeremy Utley]
5387 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5388 X690 8.9.12 and can produce some misleading textual output of OIDs.
5389 [Steve Henson, reported by Dan Kaminsky]
5391 *) Delete MD2 from algorithm tables. This follows the recommendation in
5392 several standards that it is not used in new applications due to
5393 several cryptographic weaknesses. For binary compatibility reasons
5394 the MD2 API is still compiled in by default.
5397 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5401 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5402 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5404 [Guenter <lists@gknw.net>]
5406 *) Fix the server certificate chain building code to use X509_verify_cert(),
5407 it used to have an ad-hoc builder which was unable to cope with anything
5408 other than a simple chain.
5409 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5411 *) Don't check self signed certificate signatures in X509_verify_cert()
5412 by default (a flag can override this): it just wastes time without
5413 adding any security. As a useful side effect self signed root CAs
5414 with non-FIPS digests are now usable in FIPS mode.
5417 *) In dtls1_process_out_of_seq_message() the check if the current message
5418 is already buffered was missing. For every new message was memory
5419 allocated, allowing an attacker to perform an denial of service attack
5420 with sending out of seq handshake messages until there is no memory
5421 left. Additionally every future message was buffered, even if the
5422 sequence number made no sense and would be part of another handshake.
5423 So only messages with sequence numbers less than 10 in advance will be
5424 buffered. (CVE-2009-1378)
5425 [Robin Seggelmann, discovered by Daniel Mentz]
5427 *) Records are buffered if they arrive with a future epoch to be
5428 processed after finishing the corresponding handshake. There is
5429 currently no limitation to this buffer allowing an attacker to perform
5430 a DOS attack with sending records with future epochs until there is no
5431 memory left. This patch adds the pqueue_size() function to determine
5432 the size of a buffer and limits the record buffer to 100 entries.
5434 [Robin Seggelmann, discovered by Daniel Mentz]
5436 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5437 parent structure is freed. (CVE-2009-1379)
5440 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5441 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5444 [Ilya O. <vrghost@gmail.com>]
5446 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5448 *) Disable renegotiation completely - this fixes a severe security
5449 problem (CVE-2009-3555) at the cost of breaking all
5450 renegotiation. Renegotiation can be re-enabled by setting
5451 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5452 run-time. This is really not recommended unless you know what
5456 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5458 *) Don't set val to NULL when freeing up structures, it is freed up by
5459 underlying code. If sizeof(void *) > sizeof(long) this can result in
5460 zeroing past the valid field. (CVE-2009-0789)
5461 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5463 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5464 checked correctly. This would allow some invalid signed attributes to
5465 appear to verify correctly. (CVE-2009-0591)
5466 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5468 *) Reject UniversalString and BMPString types with invalid lengths. This
5469 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5470 a legal length. (CVE-2009-0590)
5473 *) Set S/MIME signing as the default purpose rather than setting it
5474 unconditionally. This allows applications to override it at the store
5478 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5479 to handle some structures.
5482 *) Improve efficiency of mem_gets: don't search whole buffer each time
5484 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5486 *) New -hex option for openssl rand.
5489 *) Print out UTF8String and NumericString when parsing ASN1.
5492 *) Support NumericString type for name components.
5495 *) Allow CC in the environment to override the automatically chosen
5496 compiler. Note that nothing is done to ensure flags work with the
5500 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5502 *) Properly check EVP_VerifyFinal() and similar return values
5504 [Ben Laurie, Bodo Moeller, Google Security Team]
5506 *) Enable TLS extensions by default.
5509 *) Allow the CHIL engine to be loaded, whether the application is
5510 multithreaded or not. (This does not release the developer from the
5511 obligation to set up the dynamic locking callbacks.)
5512 [Sander Temme <sander@temme.net>]
5514 *) Use correct exit code if there is an error in dgst command.
5515 [Steve Henson; problem pointed out by Roland Dirlewanger]
5517 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5518 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5521 *) Add experimental JPAKE support, including demo authentication in
5522 s_client and s_server.
5525 *) Set the comparison function in v3_addr_canonize().
5526 [Rob Austein <sra@hactrn.net>]
5528 *) Add support for XMPP STARTTLS in s_client.
5529 [Philip Paeps <philip@freebsd.org>]
5531 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5532 to ensure that even with this option, only ciphersuites in the
5533 server's preference list will be accepted. (Note that the option
5534 applies only when resuming a session, so the earlier behavior was
5535 just about the algorithm choice for symmetric cryptography.)
5538 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5540 *) Fix NULL pointer dereference if a DTLS server received
5541 ChangeCipherSpec as first record (CVE-2009-1386).
5544 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5545 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5548 *) The fix in 0.9.8c that supposedly got rid of unsafe
5549 double-checked locking was incomplete for RSA blinding,
5550 addressing just one layer of what turns out to have been
5551 doubly unsafe triple-checked locking.
5553 So now fix this for real by retiring the MONT_HELPER macro
5554 in crypto/rsa/rsa_eay.c.
5556 [Bodo Moeller; problem pointed out by Marius Schilder]
5558 *) Various precautionary measures:
5560 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5562 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5563 (NB: This would require knowledge of the secret session ticket key
5564 to exploit, in which case you'd be SOL either way.)
5566 - Change bn_nist.c so that it will properly handle input BIGNUMs
5567 outside the expected range.
5569 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5572 [Neel Mehta, Bodo Moeller]
5574 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5575 the load fails. Useful for distros.
5576 [Ben Laurie and the FreeBSD team]
5578 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5581 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5584 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5586 This work was sponsored by Logica.
5589 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5590 keystores. Support for SSL/TLS client authentication too.
5591 Not compiled unless enable-capieng specified to Configure.
5593 This work was sponsored by Logica.
5596 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5597 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5598 attribute creation routines such as certificate requests and PKCS#12
5602 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5604 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5605 handshake which could lead to a client crash as found using the
5606 Codenomicon TLS test suite (CVE-2008-1672)
5607 [Steve Henson, Mark Cox]
5609 *) Fix double free in TLS server name extensions which could lead to
5610 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5613 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5615 Clear the error queue to ensure that error entries left from
5616 older function calls do not interfere with the correct operation.
5617 [Lutz Jaenicke, Erik de Castro Lopo]
5619 *) Remove root CA certificates of commercial CAs:
5621 The OpenSSL project does not recommend any specific CA and does not
5622 have any policy with respect to including or excluding any CA.
5623 Therefore it does not make any sense to ship an arbitrary selection
5624 of root CA certificates with the OpenSSL software.
5627 *) RSA OAEP patches to fix two separate invalid memory reads.
5628 The first one involves inputs when 'lzero' is greater than
5629 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5630 before the beginning of from). The second one involves inputs where
5631 the 'db' section contains nothing but zeroes (there is a one-byte
5632 invalid read after the end of 'db').
5633 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5635 *) Partial backport from 0.9.9-dev:
5637 Introduce bn_mul_mont (dedicated Montgomery multiplication
5638 procedure) as a candidate for BIGNUM assembler implementation.
5639 While 0.9.9-dev uses assembler for various architectures, only
5640 x86_64 is available by default here in the 0.9.8 branch, and
5641 32-bit x86 is available through a compile-time setting.
5643 To try the 32-bit x86 assembler implementation, use Configure
5644 option "enable-montasm" (which exists only for this backport).
5646 As "enable-montasm" for 32-bit x86 disclaims code stability
5647 anyway, in this constellation we activate additional code
5648 backported from 0.9.9-dev for further performance improvements,
5649 namely BN_from_montgomery_word. (To enable this otherwise,
5650 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5652 [Andy Polyakov (backport partially by Bodo Moeller)]
5654 *) Add TLS session ticket callback. This allows an application to set
5655 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5656 values. This is useful for key rollover for example where several key
5657 sets may exist with different names.
5660 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5661 This was broken until now in 0.9.8 releases, such that the only way
5662 a registered ENGINE could be used (assuming it initialises
5663 successfully on the host) was to explicitly set it as the default
5664 for the relevant algorithms. This is in contradiction with 0.9.7
5665 behaviour and the documentation. With this fix, when an ENGINE is
5666 registered into a given algorithm's table of implementations, the
5667 'uptodate' flag is reset so that auto-discovery will be used next
5668 time a new context for that algorithm attempts to select an
5670 [Ian Lister (tweaked by Geoff Thorpe)]
5672 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5673 implementation in the following ways:
5675 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5678 Lack of BER streaming support means one pass streaming processing is
5679 only supported if data is detached: setting the streaming flag is
5680 ignored for embedded content.
5682 CMS support is disabled by default and must be explicitly enabled
5683 with the enable-cms configuration option.
5686 *) Update the GMP engine glue to do direct copies between BIGNUM and
5687 mpz_t when openssl and GMP use the same limb size. Otherwise the
5688 existing "conversion via a text string export" trick is still used.
5689 [Paul Sheer <paulsheer@gmail.com>]
5691 *) Zlib compression BIO. This is a filter BIO which compressed and
5692 uncompresses any data passed through it.
5695 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5696 RFC3394 compatible AES key wrapping.
5699 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5700 sets string data without copying. X509_ALGOR_set0() and
5701 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5702 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5703 from an X509_ATTRIBUTE structure optionally checking it occurs only
5704 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5708 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5709 to get the expected BN_FLG_CONSTTIME behavior.
5710 [Bodo Moeller (Google)]
5714 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5715 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5716 - added some more tests to do_tests.pl
5717 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5718 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5719 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5720 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5721 - various changes to netware.pl to enable gcc-cross builds on Win32
5723 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5724 - various changes to fix missing prototype warnings
5725 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5726 - added AES, WHIRLPOOL and CPUID assembler code to build files
5727 - added missing AES assembler make rules to mk1mf.pl
5728 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5729 [Guenter Knauf <eflash@gmx.net>]
5731 *) Implement certificate status request TLS extension defined in RFC3546.
5732 A client can set the appropriate parameters and receive the encoded
5733 OCSP response via a callback. A server can query the supplied parameters
5734 and set the encoded OCSP response in the callback. Add simplified examples
5735 to s_client and s_server.
5738 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5740 *) Fix various bugs:
5741 + Binary incompatibility of ssl_ctx_st structure
5742 + DTLS interoperation with non-compliant servers
5743 + Don't call get_session_cb() without proposed session
5744 + Fix ia64 assembler code
5745 [Andy Polyakov, Steve Henson]
5747 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5749 *) DTLS Handshake overhaul. There were longstanding issues with
5750 OpenSSL DTLS implementation, which were making it impossible for
5751 RFC 4347 compliant client to communicate with OpenSSL server.
5752 Unfortunately just fixing these incompatibilities would "cut off"
5753 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5754 server keeps tolerating non RFC compliant syntax. The opposite is
5755 not true, 0.9.8f client can not communicate with earlier server.
5756 This update even addresses CVE-2007-4995.
5759 *) Changes to avoid need for function casts in OpenSSL: some compilers
5760 (gcc 4.2 and later) reject their use.
5761 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5764 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5765 RFC4507bis. The encrypted ticket format is an encrypted encoded
5766 SSL_SESSION structure, that way new session features are automatically
5769 If a client application caches session in an SSL_SESSION structure
5770 support is transparent because tickets are now stored in the encoded
5773 The SSL_CTX structure automatically generates keys for ticket
5774 protection in servers so again support should be possible
5775 with no application modification.
5777 If a client or server wishes to disable RFC4507 support then the option
5778 SSL_OP_NO_TICKET can be set.
5780 Add a TLS extension debugging callback to allow the contents of any client
5781 or server extensions to be examined.
5783 This work was sponsored by Google.
5786 *) Add initial support for TLS extensions, specifically for the server_name
5787 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5788 have new members for a host name. The SSL data structure has an
5789 additional member SSL_CTX *initial_ctx so that new sessions can be
5790 stored in that context to allow for session resumption, even after the
5791 SSL has been switched to a new SSL_CTX in reaction to a client's
5792 server_name extension.
5794 New functions (subject to change):
5796 SSL_get_servername()
5797 SSL_get_servername_type()
5800 New CTRL codes and macros (subject to change):
5802 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5803 - SSL_CTX_set_tlsext_servername_callback()
5804 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5805 - SSL_CTX_set_tlsext_servername_arg()
5806 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5808 openssl s_client has a new '-servername ...' option.
5810 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5811 '-key2 ...', '-servername_fatal' (subject to change). This allows
5812 testing the HostName extension for a specific single host name ('-cert'
5813 and '-key' remain fallbacks for handshakes without HostName
5814 negotiation). If the unrecognized_name alert has to be sent, this by
5815 default is a warning; it becomes fatal with the '-servername_fatal'
5818 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5820 *) Add AES and SSE2 assembly language support to VC++ build.
5823 *) Mitigate attack on final subtraction in Montgomery reduction.
5826 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5827 (which previously caused an internal error).
5830 *) Squeeze another 10% out of IGE mode when in != out.
5833 *) AES IGE mode speedup.
5834 [Dean Gaudet (Google)]
5836 *) Add the Korean symmetric 128-bit cipher SEED (see
5837 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5838 add SEED ciphersuites from RFC 4162:
5840 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5841 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5842 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5843 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5845 To minimize changes between patchlevels in the OpenSSL 0.9.8
5846 series, SEED remains excluded from compilation unless OpenSSL
5847 is configured with 'enable-seed'.
5848 [KISA, Bodo Moeller]
5850 *) Mitigate branch prediction attacks, which can be practical if a
5851 single processor is shared, allowing a spy process to extract
5852 information. For detailed background information, see
5853 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5854 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5855 and Necessary Software Countermeasures"). The core of the change
5856 are new versions BN_div_no_branch() and
5857 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5858 respectively, which are slower, but avoid the security-relevant
5859 conditional branches. These are automatically called by BN_div()
5860 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5861 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5862 remove a conditional branch.
5864 BN_FLG_CONSTTIME is the new name for the previous
5865 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5866 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5867 in the exponent causes BN_mod_exp_mont() to use the alternative
5868 implementation in BN_mod_exp_mont_consttime().) The old name
5869 remains as a deprecated alias.
5871 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5872 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5873 constant-time implementations for more than just exponentiation.
5874 Here too the old name is kept as a deprecated alias.
5876 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5877 the BN_BLINDING structure gets an independent copy of the
5878 modulus. This means that the previous "BIGNUM *m" argument to
5879 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5880 essentially becomes "const BIGNUM *m", although we can't actually
5881 change this in the header file before 0.9.9. It allows
5882 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5883 enable BN_FLG_CONSTTIME.
5885 [Matthew D Wood (Intel Corp)]
5887 *) In the SSL/TLS server implementation, be strict about session ID
5888 context matching (which matters if an application uses a single
5889 external cache for different purposes). Previously,
5890 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5891 set. This did ensure strict client verification, but meant that,
5892 with applications using a single external cache for quite
5893 different requirements, clients could circumvent ciphersuite
5894 restrictions for a given session ID context by starting a session
5895 in a different context.
5898 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5899 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5900 authentication-only ciphersuites.
5903 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5904 not complete and could lead to a possible single byte overflow
5905 (CVE-2007-5135) [Ben Laurie]
5907 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5909 *) Since AES128 and AES256 (and similarly Camellia128 and
5910 Camellia256) share a single mask bit in the logic of
5911 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5912 kludge to work properly if AES128 is available and AES256 isn't
5913 (or if Camellia128 is available and Camellia256 isn't).
5916 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5917 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5918 When a point or a seed is encoded in a BIT STRING, we need to
5919 prevent the removal of trailing zero bits to get the proper DER
5920 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5921 of a NamedBitList, for which trailing 0 bits need to be removed.)
5924 *) Have SSL/TLS server implementation tolerate "mismatched" record
5925 protocol version while receiving ClientHello even if the
5926 ClientHello is fragmented. (The server can't insist on the
5927 particular protocol version it has chosen before the ServerHello
5928 message has informed the client about his choice.)
5931 *) Add RFC 3779 support.
5932 [Rob Austein for ARIN, Ben Laurie]
5934 *) Load error codes if they are not already present instead of using a
5935 static variable. This allows them to be cleanly unloaded and reloaded.
5936 Improve header file function name parsing.
5939 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5940 or CAPABILITY handshake as required by RFCs.
5943 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5945 *) Introduce limits to prevent malicious keys being able to
5946 cause a denial of service. (CVE-2006-2940)
5947 [Steve Henson, Bodo Moeller]
5949 *) Fix ASN.1 parsing of certain invalid structures that can result
5950 in a denial of service. (CVE-2006-2937) [Steve Henson]
5952 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5953 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5955 *) Fix SSL client code which could crash if connecting to a
5956 malicious SSLv2 server. (CVE-2006-4343)
5957 [Tavis Ormandy and Will Drewry, Google Security Team]
5959 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5960 match only those. Before that, "AES256-SHA" would be interpreted
5961 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5962 the same strength classification in 0.9.7h) as we currently only
5963 have a single AES bit in the ciphersuite description bitmap.
5964 That change, however, also applied to ciphersuite strings such as
5965 "RC4-MD5" that intentionally matched multiple ciphersuites --
5966 namely, SSL 2.0 ciphersuites in addition to the more common ones
5967 from SSL 3.0/TLS 1.0.
5969 So we change the selection algorithm again: Naming an explicit
5970 ciphersuite selects this one ciphersuite, and any other similar
5971 ciphersuite (same bitmap) from *other* protocol versions.
5972 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5973 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5975 Since SSL 2.0 does not have any ciphersuites for which the
5976 128/256 bit distinction would be relevant, this works for now.
5977 The proper fix will be to use different bits for AES128 and
5978 AES256, which would have avoided the problems from the beginning;
5979 however, bits are scarce, so we can only do this in a new release
5980 (not just a patchlevel) when we can change the SSL_CIPHER
5981 definition to split the single 'unsigned long mask' bitmap into
5982 multiple values to extend the available space.
5986 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5988 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5989 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5991 *) Add AES IGE and biIGE modes.
5994 *) Change the Unix randomness entropy gathering to use poll() when
5995 possible instead of select(), since the latter has some
5996 undesirable limitations.
5997 [Darryl Miles via Richard Levitte and Bodo Moeller]
5999 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
6000 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
6001 cannot be implicitly activated as part of, e.g., the "AES" alias.
6002 However, please upgrade to OpenSSL 0.9.9[-dev] for
6003 non-experimental use of the ECC ciphersuites to get TLS extension
6004 support, which is required for curve and point format negotiation
6005 to avoid potential handshake problems.
6008 *) Disable rogue ciphersuites:
6010 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6011 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6012 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6014 The latter two were purportedly from
6015 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6018 Also deactivate the remaining ciphersuites from
6019 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6020 unofficial, and the ID has long expired.
6023 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6024 dual-core machines) and other potential thread-safety issues.
6027 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
6028 versions), which is now available for royalty-free use
6029 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
6030 Also, add Camellia TLS ciphersuites from RFC 4132.
6032 To minimize changes between patchlevels in the OpenSSL 0.9.8
6033 series, Camellia remains excluded from compilation unless OpenSSL
6034 is configured with 'enable-camellia'.
6037 *) Disable the padding bug check when compression is in use. The padding
6038 bug check assumes the first packet is of even length, this is not
6039 necessarily true if compression is enabled and can result in false
6040 positives causing handshake failure. The actual bug test is ancient
6041 code so it is hoped that implementations will either have fixed it by
6042 now or any which still have the bug do not support compression.
6045 Changes between 0.9.8a and 0.9.8b [04 May 2006]
6047 *) When applying a cipher rule check to see if string match is an explicit
6048 cipher suite and only match that one cipher suite if it is.
6051 *) Link in manifests for VC++ if needed.
6052 [Austin Ziegler <halostatue@gmail.com>]
6054 *) Update support for ECC-based TLS ciphersuites according to
6055 draft-ietf-tls-ecc-12.txt with proposed changes (but without
6056 TLS extensions, which are supported starting with the 0.9.9
6057 branch, not in the OpenSSL 0.9.8 branch).
6060 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
6061 opaque EVP_CIPHER_CTX handling.
6064 *) Fixes and enhancements to zlib compression code. We now only use
6065 "zlib1.dll" and use the default __cdecl calling convention on Win32
6066 to conform with the standards mentioned here:
6067 http://www.zlib.net/DLL_FAQ.txt
6068 Static zlib linking now works on Windows and the new --with-zlib-include
6069 --with-zlib-lib options to Configure can be used to supply the location
6070 of the headers and library. Gracefully handle case where zlib library
6074 *) Several fixes and enhancements to the OID generation code. The old code
6075 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
6076 handle numbers larger than ULONG_MAX, truncated printing and had a
6077 non standard OBJ_obj2txt() behaviour.
6080 *) Add support for building of engines under engine/ as shared libraries
6081 under VC++ build system.
6084 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
6085 Hopefully, we will not see any false combination of paths any more.
6088 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
6090 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6091 (part of SSL_OP_ALL). This option used to disable the
6092 countermeasure against man-in-the-middle protocol-version
6093 rollback in the SSL 2.0 server implementation, which is a bad
6094 idea. (CVE-2005-2969)
6096 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6097 for Information Security, National Institute of Advanced Industrial
6098 Science and Technology [AIST], Japan)]
6100 *) Add two function to clear and return the verify parameter flags.
6103 *) Keep cipherlists sorted in the source instead of sorting them at
6104 runtime, thus removing the need for a lock.
6107 *) Avoid some small subgroup attacks in Diffie-Hellman.
6108 [Nick Mathewson and Ben Laurie]
6110 *) Add functions for well-known primes.
6113 *) Extended Windows CE support.
6114 [Satoshi Nakamura and Andy Polyakov]
6116 *) Initialize SSL_METHOD structures at compile time instead of during
6117 runtime, thus removing the need for a lock.
6120 *) Make PKCS7_decrypt() work even if no certificate is supplied by
6121 attempting to decrypt each encrypted key in turn. Add support to
6125 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
6127 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
6130 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
6133 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
6134 key into the same file any more.
6137 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
6140 *) Add -utf8 command line and config file option to 'ca'.
6141 [Stefan <stf@udoma.org]
6143 *) Removed the macro des_crypt(), as it seems to conflict with some
6144 libraries. Use DES_crypt().
6147 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
6148 involves renaming the source and generated shared-libs for
6149 both. The engines will accept the corrected or legacy ids
6150 ('ncipher' and '4758_cca' respectively) when binding. NB,
6151 this only applies when building 'shared'.
6152 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
6154 *) Add attribute functions to EVP_PKEY structure. Modify
6155 PKCS12_create() to recognize a CSP name attribute and
6156 use it. Make -CSP option work again in pkcs12 utility.
6159 *) Add new functionality to the bn blinding code:
6160 - automatic re-creation of the BN_BLINDING parameters after
6161 a fixed number of uses (currently 32)
6162 - add new function for parameter creation
6163 - introduce flags to control the update behaviour of the
6164 BN_BLINDING parameters
6165 - hide BN_BLINDING structure
6166 Add a second BN_BLINDING slot to the RSA structure to improve
6167 performance when a single RSA object is shared among several
6171 *) Add support for DTLS.
6172 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
6174 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
6175 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
6178 *) Remove buggy and incomplete DH cert support from
6179 ssl/ssl_rsa.c and ssl/s3_both.c
6182 *) Use SHA-1 instead of MD5 as the default digest algorithm for
6183 the apps/openssl applications.
6186 *) Compile clean with "-Wall -Wmissing-prototypes
6187 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
6188 DEBUG_SAFESTACK must also be set.
6191 *) Change ./Configure so that certain algorithms can be disabled by default.
6192 The new counterpiece to "no-xxx" is "enable-xxx".
6194 The patented RC5 and MDC2 algorithms will now be disabled unless
6195 "enable-rc5" and "enable-mdc2", respectively, are specified.
6197 (IDEA remains enabled despite being patented. This is because IDEA
6198 is frequently required for interoperability, and there is no license
6199 fee for non-commercial use. As before, "no-idea" can be used to
6200 avoid this algorithm.)
6204 *) Add processing of proxy certificates (see RFC 3820). This work was
6205 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
6206 EGEE (Enabling Grids for E-science in Europe).
6209 *) RC4 performance overhaul on modern architectures/implementations, such
6210 as Intel P4, IA-64 and AMD64.
6213 *) New utility extract-section.pl. This can be used specify an alternative
6214 section number in a pod file instead of having to treat each file as
6215 a separate case in Makefile. This can be done by adding two lines to the
6218 =for comment openssl_section:XXX
6220 The blank line is mandatory.
6224 *) New arguments -certform, -keyform and -pass for s_client and s_server
6225 to allow alternative format key and certificate files and passphrase
6229 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
6230 update associated structures and add various utility functions.
6232 Add new policy related verify parameters, include policy checking in
6233 standard verify code. Enhance 'smime' application with extra parameters
6234 to support policy checking and print out.
6237 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
6238 Nehemiah processors. These extensions support AES encryption in hardware
6239 as well as RNG (though RNG support is currently disabled).
6240 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
6242 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
6245 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
6246 [Andy Polyakov and a number of other people]
6248 *) Improved PowerPC platform support. Most notably BIGNUM assembler
6249 implementation contributed by IBM.
6250 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
6252 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
6253 exponent rather than 'unsigned long'. There is a corresponding change to
6254 the new 'rsa_keygen' element of the RSA_METHOD structure.
6255 [Jelte Jansen, Geoff Thorpe]
6257 *) Functionality for creating the initial serial number file is now
6258 moved from CA.pl to the 'ca' utility with a new option -create_serial.
6260 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
6261 number file to 1, which is bound to cause problems. To avoid
6262 the problems while respecting compatibility between different 0.9.7
6263 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
6264 CA.pl for serial number initialization. With the new release 0.9.8,
6265 we can fix the problem directly in the 'ca' utility.)
6268 *) Reduced header interdependencies by declaring more opaque objects in
6269 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
6270 give fewer recursive includes, which could break lazy source code - so
6271 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
6272 developers should define this symbol when building and using openssl to
6273 ensure they track the recommended behaviour, interfaces, [etc], but
6274 backwards-compatible behaviour prevails when this isn't defined.
6277 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
6280 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
6281 This will generate a random key of the appropriate length based on the
6282 cipher context. The EVP_CIPHER can provide its own random key generation
6283 routine to support keys of a specific form. This is used in the des and
6284 3des routines to generate a key of the correct parity. Update S/MIME
6285 code to use new functions and hence generate correct parity DES keys.
6286 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
6287 valid (weak or incorrect parity).
6290 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
6291 as looking them up. This is useful when the verified structure may contain
6292 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
6293 present unless the new PKCS7_NO_CRL flag is asserted.
6296 *) Extend ASN1 oid configuration module. It now additionally accepts the
6299 shortName = some long name, 1.2.3.4
6302 *) Reimplemented the BN_CTX implementation. There is now no more static
6303 limitation on the number of variables it can handle nor the depth of the
6304 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
6305 information can now expand as required, and rather than having a single
6306 static array of bignums, BN_CTX now uses a linked-list of such arrays
6307 allowing it to expand on demand whilst maintaining the usefulness of
6308 BN_CTX's "bundling".
6311 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
6312 to allow all RSA operations to function using a single BN_CTX.
6315 *) Preliminary support for certificate policy evaluation and checking. This
6316 is initially intended to pass the tests outlined in "Conformance Testing
6317 of Relying Party Client Certificate Path Processing Logic" v1.07.
6320 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
6321 remained unused and not that useful. A variety of other little bignum
6322 tweaks and fixes have also been made continuing on from the audit (see
6326 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
6327 associated ASN1, EVP and SSL functions and old ASN1 macros.
6330 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
6331 and this should never fail. So the return value from the use of
6332 BN_set_word() (which can fail due to needless expansion) is now deprecated;
6333 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
6336 *) BN_CTX_get() should return zero-valued bignums, providing the same
6337 initialised value as BN_new().
6338 [Geoff Thorpe, suggested by Ulf Möller]
6340 *) Support for inhibitAnyPolicy certificate extension.
6343 *) An audit of the BIGNUM code is underway, for which debugging code is
6344 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
6345 is considered valid when processing BIGNUMs, and causes execution to
6346 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
6347 further steps are taken to deliberately pollute unused data in BIGNUM
6348 structures to try and expose faulty code further on. For now, openssl will
6349 (in its default mode of operation) continue to tolerate the inconsistent
6350 forms that it has tolerated in the past, but authors and packagers should
6351 consider trying openssl and their own applications when compiled with
6352 these debugging symbols defined. It will help highlight potential bugs in
6353 their own code, and will improve the test coverage for OpenSSL itself. At
6354 some point, these tighter rules will become openssl's default to improve
6355 maintainability, though the assert()s and other overheads will remain only
6356 in debugging configurations. See bn.h for more details.
6357 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6359 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6360 that can only be obtained through BN_CTX_new() (which implicitly
6361 initialises it). The presence of this function only made it possible
6362 to overwrite an existing structure (and cause memory leaks).
6365 *) Because of the callback-based approach for implementing LHASH as a
6366 template type, lh_insert() adds opaque objects to hash-tables and
6367 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6368 to clean up those corresponding objects before destroying the hash table
6369 (and losing the object pointers). So some over-zealous constifications in
6370 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6371 objects as "const" and the lh_doall[_arg] callback wrappers are not
6372 prototyped to have "const" restrictions on the object pointers they are
6373 given (and so aren't required to cast them away any more).
6376 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6377 (speed) prefers to use its own implementation. The two implementations
6378 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6379 its object type properly exposed (MS_TM) instead of casting to/from "char
6380 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6381 aren't necessarily the greatest nomenclatures - but this is what was used
6382 internally to the implementation so I've used that for now.
6385 *) Ensure that deprecated functions do not get compiled when
6386 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6387 the self-tests were still using deprecated key-generation functions so
6388 these have been updated also.
6391 *) Reorganise PKCS#7 code to separate the digest location functionality
6392 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6393 New function PKCS7_set_digest() to set the digest type for PKCS#7
6394 digestedData type. Add additional code to correctly generate the
6395 digestedData type and add support for this type in PKCS7 initialization
6399 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6400 structure of type "other".
6403 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6404 sure the loop does correctly stop and breaking ("division by zero")
6405 modulus operations are not performed. The (pre-generated) prime
6406 table crypto/bn/bn_prime.h was already correct, but it could not be
6407 re-generated on some platforms because of the "division by zero"
6408 situation in the script.
6409 [Ralf S. Engelschall]
6411 *) Update support for ECC-based TLS ciphersuites according to
6412 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6413 SHA-1 now is only used for "small" curves (where the
6414 representation of a field element takes up to 24 bytes); for
6415 larger curves, the field element resulting from ECDH is directly
6416 used as premaster secret.
6417 [Douglas Stebila (Sun Microsystems Laboratories)]
6419 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6420 curve secp160r1 to the tests.
6421 [Douglas Stebila (Sun Microsystems Laboratories)]
6423 *) Add the possibility to load symbols globally with DSO.
6424 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6426 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6427 control of the error stack.
6430 *) Add support for STORE in ENGINE.
6433 *) Add the STORE type. The intention is to provide a common interface
6434 to certificate and key stores, be they simple file-based stores, or
6435 HSM-type store, or LDAP stores, or...
6436 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6439 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6440 pass a list of arguments to any function as well as provide a way
6441 for a function to pass data back to the caller.
6444 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6445 works like BUF_strdup() but can be used to duplicate a portion of
6446 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6450 *) Add the function sk_find_ex() which works like sk_find(), but will
6451 return an index to an element even if an exact match couldn't be
6452 found. The index is guaranteed to point at the element where the
6453 searched-for key would be inserted to preserve sorting order.
6456 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6457 takes an extra flags argument for optional functionality. Currently,
6458 the following flags are defined:
6460 OBJ_BSEARCH_VALUE_ON_NOMATCH
6461 This one gets OBJ_bsearch_ex() to return a pointer to the first
6462 element where the comparing function returns a negative or zero
6465 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6466 This one gets OBJ_bsearch_ex() to return a pointer to the first
6467 element where the comparing function returns zero. This is useful
6468 if there are more than one element where the comparing function
6472 *) Make it possible to create self-signed certificates with 'openssl ca'
6473 in such a way that the self-signed certificate becomes part of the
6474 CA database and uses the same mechanisms for serial number generation
6475 as all other certificate signing. The new flag '-selfsign' enables
6476 this functionality. Adapt CA.sh and CA.pl.in.
6479 *) Add functionality to check the public key of a certificate request
6480 against a given private. This is useful to check that a certificate
6481 request can be signed by that key (self-signing).
6484 *) Make it possible to have multiple active certificates with the same
6485 subject in the CA index file. This is done only if the keyword
6486 'unique_subject' is set to 'no' in the main CA section (default
6487 if 'CA_default') of the configuration file. The value is saved
6488 with the database itself in a separate index attribute file,
6489 named like the index file with '.attr' appended to the name.
6492 *) Generate multi-valued AVAs using '+' notation in config files for
6496 *) Support for nameConstraints certificate extension.
6499 *) Support for policyConstraints certificate extension.
6502 *) Support for policyMappings certificate extension.
6505 *) Make sure the default DSA_METHOD implementation only uses its
6506 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6507 and change its own handlers to be NULL so as to remove unnecessary
6508 indirection. This lets alternative implementations fallback to the
6509 default implementation more easily.
6512 *) Support for directoryName in GeneralName related extensions
6516 *) Make it possible to link applications using Makefile.shared.
6517 Make that possible even when linking against static libraries!
6520 *) Support for single pass processing for S/MIME signing. This now
6521 means that S/MIME signing can be done from a pipe, in addition
6522 cleartext signing (multipart/signed type) is effectively streaming
6523 and the signed data does not need to be all held in memory.
6525 This is done with a new flag PKCS7_STREAM. When this flag is set
6526 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6527 is done after the data is output (and digests calculated) in
6528 SMIME_write_PKCS7().
6531 *) Add full support for -rpath/-R, both in shared libraries and
6532 applications, at least on the platforms where it's known how
6536 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6537 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6538 will now compute a table of multiples of the generator that
6539 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6540 faster (notably in the case of a single point multiplication,
6541 scalar * generator).
6542 [Nils Larsch, Bodo Moeller]
6544 *) IPv6 support for certificate extensions. The various extensions
6545 which use the IP:a.b.c.d can now take IPv6 addresses using the
6546 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6550 *) Added an ENGINE that implements RSA by performing private key
6551 exponentiations with the GMP library. The conversions to and from
6552 GMP's mpz_t format aren't optimised nor are any montgomery forms
6553 cached, and on x86 it appears OpenSSL's own performance has caught up.
6554 However there are likely to be other architectures where GMP could
6555 provide a boost. This ENGINE is not built in by default, but it can be
6556 specified at Configure time and should be accompanied by the necessary
6557 linker additions, eg;
6558 ./config -DOPENSSL_USE_GMP -lgmp
6561 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6562 testing availability of engines with "-t" - the old behaviour is
6563 produced by increasing the feature's verbosity with "-tt".
6566 *) ECDSA routines: under certain error conditions uninitialized BN objects
6567 could be freed. Solution: make sure initialization is performed early
6568 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6572 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6573 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6574 software implementations. For DSA and DH, parameter generation can
6575 also be overridden by providing the appropriate method callbacks.
6578 *) Change the "progress" mechanism used in key-generation and
6579 primality testing to functions that take a new BN_GENCB pointer in
6580 place of callback/argument pairs. The new API functions have "_ex"
6581 postfixes and the older functions are reimplemented as wrappers for
6582 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6583 declarations of the old functions to help (graceful) attempts to
6584 migrate to the new functions. Also, the new key-generation API
6585 functions operate on a caller-supplied key-structure and return
6586 success/failure rather than returning a key or NULL - this is to
6587 help make "keygen" another member function of RSA_METHOD etc.
6589 Example for using the new callback interface:
6591 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6595 BN_GENCB_set(&my_cb, my_callback, my_arg);
6597 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6598 /* For the meaning of a, b in calls to my_callback(), see the
6599 * documentation of the function that calls the callback.
6600 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6601 * my_callback should return 1 if it wants BN_is_prime_ex()
6602 * to continue, or 0 to stop.
6607 *) Change the ZLIB compression method to be stateful, and make it
6608 available to TLS with the number defined in
6609 draft-ietf-tls-compression-04.txt.
6612 *) Add the ASN.1 structures and functions for CertificatePair, which
6613 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6615 CertificatePair ::= SEQUENCE {
6616 forward [0] Certificate OPTIONAL,
6617 reverse [1] Certificate OPTIONAL,
6618 -- at least one of the pair shall be present -- }
6620 Also implement the PEM functions to read and write certificate
6621 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6623 This needed to be defined, mostly for the sake of the LDAP
6624 attribute crossCertificatePair, but may prove useful elsewhere as
6628 *) Make it possible to inhibit symlinking of shared libraries in
6629 Makefile.shared, for Cygwin's sake.
6632 *) Extend the BIGNUM API by creating a function
6633 void BN_set_negative(BIGNUM *a, int neg);
6634 and a macro that behave like
6635 int BN_is_negative(const BIGNUM *a);
6637 to avoid the need to access 'a->neg' directly in applications.
6640 *) Implement fast modular reduction for pseudo-Mersenne primes
6641 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6642 EC_GROUP_new_curve_GFp() will now automatically use this
6644 [Nils Larsch <nla@trustcenter.de>]
6646 *) Add new lock type (CRYPTO_LOCK_BN).
6649 *) Change the ENGINE framework to automatically load engines
6650 dynamically from specific directories unless they could be
6651 found to already be built in or loaded. Move all the
6652 current engines except for the cryptodev one to a new
6654 The engines in engines/ are built as shared libraries if
6655 the "shared" options was given to ./Configure or ./config.
6656 Otherwise, they are inserted in libcrypto.a.
6657 /usr/local/ssl/engines is the default directory for dynamic
6658 engines, but that can be overridden at configure time through
6659 the usual use of --prefix and/or --openssldir, and at run
6660 time with the environment variable OPENSSL_ENGINES.
6661 [Geoff Thorpe and Richard Levitte]
6663 *) Add Makefile.shared, a helper makefile to build shared
6664 libraries. Adapt Makefile.org.
6667 *) Add version info to Win32 DLLs.
6668 [Peter 'Luna' Runestig" <peter@runestig.com>]
6670 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6671 can be added using this API to created arbitrary PKCS#12
6672 files while avoiding the low level API.
6674 New options to PKCS12_create(), key or cert can be NULL and
6675 will then be omitted from the output file. The encryption
6676 algorithm NIDs can be set to -1 for no encryption, the mac
6677 iteration count can be set to 0 to omit the mac.
6679 Enhance pkcs12 utility by making the -nokeys and -nocerts
6680 options work when creating a PKCS#12 file. New option -nomac
6681 to omit the mac, NONE can be set for an encryption algorithm.
6682 New code is modified to use the enhanced PKCS12_create()
6683 instead of the low level API.
6686 *) Extend ASN1 encoder to support indefinite length constructed
6687 encoding. This can output sequences tags and octet strings in
6688 this form. Modify pk7_asn1.c to support indefinite length
6689 encoding. This is experimental and needs additional code to
6690 be useful, such as an ASN1 bio and some enhanced streaming
6693 Extend template encode functionality so that tagging is passed
6694 down to the template encoder.
6697 *) Let 'openssl req' fail if an argument to '-newkey' is not
6698 recognized instead of using RSA as a default.
6701 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6702 As these are not official, they are not included in "ALL";
6703 the "ECCdraft" ciphersuite group alias can be used to select them.
6704 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6706 *) Add ECDH engine support.
6707 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6709 *) Add ECDH in new directory crypto/ecdh/.
6710 [Douglas Stebila (Sun Microsystems Laboratories)]
6712 *) Let BN_rand_range() abort with an error after 100 iterations
6713 without success (which indicates a broken PRNG).
6716 *) Change BN_mod_sqrt() so that it verifies that the input value
6717 is really the square of the return value. (Previously,
6718 BN_mod_sqrt would show GIGO behaviour.)
6721 *) Add named elliptic curves over binary fields from X9.62, SECG,
6722 and WAP/WTLS; add OIDs that were still missing.
6724 [Sheueling Chang Shantz and Douglas Stebila
6725 (Sun Microsystems Laboratories)]
6727 *) Extend the EC library for elliptic curves over binary fields
6728 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6731 EC_GF2m_simple_method
6735 EC_GROUP_new_curve_GF2m
6736 EC_GROUP_set_curve_GF2m
6737 EC_GROUP_get_curve_GF2m
6738 EC_POINT_set_affine_coordinates_GF2m
6739 EC_POINT_get_affine_coordinates_GF2m
6740 EC_POINT_set_compressed_coordinates_GF2m
6742 Point compression for binary fields is disabled by default for
6743 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6746 As binary polynomials are represented as BIGNUMs, various members
6747 of the EC_GROUP and EC_POINT data structures can be shared
6748 between the implementations for prime fields and binary fields;
6749 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6750 are essentially identical to their ..._GFp counterparts.
6751 (For simplicity, the '..._GFp' prefix has been dropped from
6752 various internal method names.)
6754 An internal 'field_div' method (similar to 'field_mul' and
6755 'field_sqr') has been added; this is used only for binary fields.
6757 [Sheueling Chang Shantz and Douglas Stebila
6758 (Sun Microsystems Laboratories)]
6760 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6761 through methods ('mul', 'precompute_mult').
6763 The generic implementations (now internally called 'ec_wNAF_mul'
6764 and 'ec_wNAF_precomputed_mult') remain the default if these
6765 methods are undefined.
6767 [Sheueling Chang Shantz and Douglas Stebila
6768 (Sun Microsystems Laboratories)]
6770 *) New function EC_GROUP_get_degree, which is defined through
6771 EC_METHOD. For curves over prime fields, this returns the bit
6772 length of the modulus.
6774 [Sheueling Chang Shantz and Douglas Stebila
6775 (Sun Microsystems Laboratories)]
6777 *) New functions EC_GROUP_dup, EC_POINT_dup.
6778 (These simply call ..._new and ..._copy).
6780 [Sheueling Chang Shantz and Douglas Stebila
6781 (Sun Microsystems Laboratories)]
6783 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6784 Polynomials are represented as BIGNUMs (where the sign bit is not
6785 used) in the following functions [macros]:
6788 BN_GF2m_sub [= BN_GF2m_add]
6789 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6790 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6791 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6793 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6794 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6795 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6796 BN_GF2m_cmp [= BN_ucmp]
6798 (Note that only the 'mod' functions are actually for fields GF(2^m).
6799 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6801 For some functions, an the irreducible polynomial defining a
6802 field can be given as an 'unsigned int[]' with strictly
6803 decreasing elements giving the indices of those bits that are set;
6804 i.e., p[] represents the polynomial
6805 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6807 p[0] > p[1] > ... > p[k] = 0.
6808 This applies to the following functions:
6813 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6814 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6816 BN_GF2m_mod_sqrt_arr
6817 BN_GF2m_mod_solve_quad_arr
6821 Conversion can be performed by the following functions:
6826 bntest.c has additional tests for binary polynomial arithmetic.
6828 Two implementations for BN_GF2m_mod_div() are available.
6829 The default algorithm simply uses BN_GF2m_mod_inv() and
6830 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6831 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6832 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6834 [Sheueling Chang Shantz and Douglas Stebila
6835 (Sun Microsystems Laboratories)]
6837 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6838 functionality is disabled at compile-time.
6839 [Douglas Stebila <douglas.stebila@sun.com>]
6841 *) Change default behaviour of 'openssl asn1parse' so that more
6842 information is visible when viewing, e.g., a certificate:
6844 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6845 mode the content of non-printable OCTET STRINGs is output in a
6846 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6847 avoid the appearance of a printable string.
6848 [Nils Larsch <nla@trustcenter.de>]
6850 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6852 EC_GROUP_set_asn1_flag()
6853 EC_GROUP_get_asn1_flag()
6854 EC_GROUP_set_point_conversion_form()
6855 EC_GROUP_get_point_conversion_form()
6856 These control ASN1 encoding details:
6857 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6858 has been set to OPENSSL_EC_NAMED_CURVE.
6859 - Points are encoded in uncompressed form by default; options for
6860 asn1_for are as for point2oct, namely
6861 POINT_CONVERSION_COMPRESSED
6862 POINT_CONVERSION_UNCOMPRESSED
6863 POINT_CONVERSION_HYBRID
6865 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6868 EC_GROUP_get0_seed()
6869 EC_GROUP_get_seed_len()
6870 This is used only for ASN1 purposes (so far).
6871 [Nils Larsch <nla@trustcenter.de>]
6873 *) Add 'field_type' member to EC_METHOD, which holds the NID
6874 of the appropriate field type OID. The new function
6875 EC_METHOD_get_field_type() returns this value.
6876 [Nils Larsch <nla@trustcenter.de>]
6881 EC_POINT_point2hex()
6882 EC_POINT_hex2point()
6883 providing useful interfaces to EC_POINT_point2oct() and
6884 EC_POINT_oct2point().
6885 [Nils Larsch <nla@trustcenter.de>]
6887 *) Change internals of the EC library so that the functions
6888 EC_GROUP_set_generator()
6889 EC_GROUP_get_generator()
6890 EC_GROUP_get_order()
6891 EC_GROUP_get_cofactor()
6892 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6893 to methods, which would lead to unnecessary code duplication when
6894 adding different types of curves.
6895 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6897 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6898 arithmetic, and such that modified wNAFs are generated
6899 (which avoid length expansion in many cases).
6902 *) Add a function EC_GROUP_check_discriminant() (defined via
6903 EC_METHOD) that verifies that the curve discriminant is non-zero.
6905 Add a function EC_GROUP_check() that makes some sanity tests
6906 on a EC_GROUP, its generator and order. This includes
6907 EC_GROUP_check_discriminant().
6908 [Nils Larsch <nla@trustcenter.de>]
6910 *) Add ECDSA in new directory crypto/ecdsa/.
6912 Add applications 'openssl ecparam' and 'openssl ecdsa'
6913 (these are based on 'openssl dsaparam' and 'openssl dsa').
6915 ECDSA support is also included in various other files across the
6916 library. Most notably,
6917 - 'openssl req' now has a '-newkey ecdsa:file' option;
6918 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6919 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6920 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6921 them suitable for ECDSA where domain parameters must be
6922 extracted before the specific public key;
6923 - ECDSA engine support has been added.
6924 [Nils Larsch <nla@trustcenter.de>]
6926 *) Include some named elliptic curves, and add OIDs from X9.62,
6927 SECG, and WAP/WTLS. Each curve can be obtained from the new
6929 EC_GROUP_new_by_curve_name(),
6930 and the list of available named curves can be obtained with
6931 EC_get_builtin_curves().
6932 Also add a 'curve_name' member to EC_GROUP objects, which can be
6934 EC_GROUP_set_curve_name()
6935 EC_GROUP_get_curve_name()
6936 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6938 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6939 was actually never needed) and in BN_mul(). The removal in BN_mul()
6940 required a small change in bn_mul_part_recursive() and the addition
6941 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6942 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6943 bn_sub_words() and bn_add_words() except they take arrays with
6947 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6949 *) Cleanse PEM buffers before freeing them since they may contain
6951 [Benjamin Bennett <ben@psc.edu>]
6953 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6954 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6955 authentication-only ciphersuites.
6958 *) Since AES128 and AES256 share a single mask bit in the logic of
6959 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6960 kludge to work properly if AES128 is available and AES256 isn't.
6963 *) Expand security boundary to match 1.1.1 module.
6966 *) Remove redundant features: hash file source, editing of test vectors
6967 modify fipsld to use external fips_premain.c signature.
6970 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6971 run algorithm test programs.
6974 *) Make algorithm test programs more tolerant of whitespace.
6977 *) Have SSL/TLS server implementation tolerate "mismatched" record
6978 protocol version while receiving ClientHello even if the
6979 ClientHello is fragmented. (The server can't insist on the
6980 particular protocol version it has chosen before the ServerHello
6981 message has informed the client about his choice.)
6984 *) Load error codes if they are not already present instead of using a
6985 static variable. This allows them to be cleanly unloaded and reloaded.
6988 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6990 *) Introduce limits to prevent malicious keys being able to
6991 cause a denial of service. (CVE-2006-2940)
6992 [Steve Henson, Bodo Moeller]
6994 *) Fix ASN.1 parsing of certain invalid structures that can result
6995 in a denial of service. (CVE-2006-2937) [Steve Henson]
6997 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6998 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
7000 *) Fix SSL client code which could crash if connecting to a
7001 malicious SSLv2 server. (CVE-2006-4343)
7002 [Tavis Ormandy and Will Drewry, Google Security Team]
7004 *) Change ciphersuite string processing so that an explicit
7005 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
7006 will no longer include "AES128-SHA"), and any other similar
7007 ciphersuite (same bitmap) from *other* protocol versions (so that
7008 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
7009 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
7010 changes from 0.9.8b and 0.9.8d.
7013 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
7015 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
7016 (CVE-2006-4339) [Ben Laurie and Google Security Team]
7018 *) Change the Unix randomness entropy gathering to use poll() when
7019 possible instead of select(), since the latter has some
7020 undesirable limitations.
7021 [Darryl Miles via Richard Levitte and Bodo Moeller]
7023 *) Disable rogue ciphersuites:
7025 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
7026 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
7027 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
7029 The latter two were purportedly from
7030 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
7033 Also deactivate the remaining ciphersuites from
7034 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
7035 unofficial, and the ID has long expired.
7038 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
7039 dual-core machines) and other potential thread-safety issues.
7042 Changes between 0.9.7i and 0.9.7j [04 May 2006]
7044 *) Adapt fipsld and the build system to link against the validated FIPS
7045 module in FIPS mode.
7048 *) Fixes for VC++ 2005 build under Windows.
7051 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
7052 from a Windows bash shell such as MSYS. It is autodetected from the
7053 "config" script when run from a VC++ environment. Modify standard VC++
7054 build to use fipscanister.o from the GNU make build.
7057 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
7059 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
7060 The value now differs depending on if you build for FIPS or not.
7061 BEWARE! A program linked with a shared FIPSed libcrypto can't be
7062 safely run with a non-FIPSed libcrypto, as it may crash because of
7063 the difference induced by this change.
7066 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
7068 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
7069 (part of SSL_OP_ALL). This option used to disable the
7070 countermeasure against man-in-the-middle protocol-version
7071 rollback in the SSL 2.0 server implementation, which is a bad
7072 idea. (CVE-2005-2969)
7074 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
7075 for Information Security, National Institute of Advanced Industrial
7076 Science and Technology [AIST], Japan)]
7078 *) Minimal support for X9.31 signatures and PSS padding modes. This is
7079 mainly for FIPS compliance and not fully integrated at this stage.
7082 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
7083 the exponentiation using a fixed-length exponent. (Otherwise,
7084 the information leaked through timing could expose the secret key
7085 after many signatures; cf. Bleichenbacher's attack on DSA with
7089 *) Make a new fixed-window mod_exp implementation the default for
7090 RSA, DSA, and DH private-key operations so that the sequence of
7091 squares and multiplies and the memory access pattern are
7092 independent of the particular secret key. This will mitigate
7093 cache-timing and potential related attacks.
7095 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
7096 and this is automatically used by BN_mod_exp_mont() if the new flag
7097 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
7098 will use this BN flag for private exponents unless the flag
7099 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
7100 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
7102 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
7104 *) Change the client implementation for SSLv23_method() and
7105 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
7106 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
7107 (Previously, the SSL 2.0 backwards compatible Client Hello
7108 message format would be used even with SSL_OP_NO_SSLv2.)
7111 *) Add support for smime-type MIME parameter in S/MIME messages which some
7115 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
7116 a threadsafe manner. Modify rsa code to use new function and add calls
7117 to dsa and dh code (which had race conditions before).
7120 *) Include the fixed error library code in the C error file definitions
7121 instead of fixing them up at runtime. This keeps the error code
7122 structures constant.
7125 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
7127 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
7130 *) Fixes for newer kerberos headers. NB: the casts are needed because
7131 the 'length' field is signed on one version and unsigned on another
7132 with no (?) obvious way to tell the difference, without these VC++
7133 complains. Also the "definition" of FAR (blank) is no longer included
7134 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
7135 some needed definitions.
7138 *) Undo Cygwin change.
7141 *) Added support for proxy certificates according to RFC 3820.
7142 Because they may be a security thread to unaware applications,
7143 they must be explicitly allowed in run-time. See
7144 docs/HOWTO/proxy_certificates.txt for further information.
7147 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
7149 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
7150 server and client random values. Previously
7151 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
7152 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
7154 This change has negligible security impact because:
7156 1. Server and client random values still have 24 bytes of pseudo random
7159 2. Server and client random values are sent in the clear in the initial
7162 3. The master secret is derived using the premaster secret (48 bytes in
7163 size for static RSA ciphersuites) as well as client server and random
7166 The OpenSSL team would like to thank the UK NISCC for bringing this issue
7169 [Stephen Henson, reported by UK NISCC]
7171 *) Use Windows randomness collection on Cygwin.
7174 *) Fix hang in EGD/PRNGD query when communication socket is closed
7175 prematurely by EGD/PRNGD.
7176 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
7178 *) Prompt for pass phrases when appropriate for PKCS12 input format.
7181 *) Back-port of selected performance improvements from development
7182 branch, as well as improved support for PowerPC platforms.
7185 *) Add lots of checks for memory allocation failure, error codes to indicate
7186 failure and freeing up memory if a failure occurs.
7187 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
7189 *) Add new -passin argument to dgst.
7192 *) Perform some character comparisons of different types in X509_NAME_cmp:
7193 this is needed for some certificates that re-encode DNs into UTF8Strings
7194 (in violation of RFC3280) and can't or won't issue name rollover
7198 *) Make an explicit check during certificate validation to see that
7199 the CA setting in each certificate on the chain is correct. As a
7200 side effect always do the following basic checks on extensions,
7201 not just when there's an associated purpose to the check:
7203 - if there is an unhandled critical extension (unless the user
7204 has chosen to ignore this fault)
7205 - if the path length has been exceeded (if one is set at all)
7206 - that certain extensions fit the associated purpose (if one has
7210 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
7212 *) Avoid a race condition when CRLs are checked in a multi threaded
7213 environment. This would happen due to the reordering of the revoked
7214 entries during signature checking and serial number lookup. Now the
7215 encoding is cached and the serial number sort performed under a lock.
7216 Add new STACK function sk_is_sorted().
7219 *) Add Delta CRL to the extension code.
7222 *) Various fixes to s3_pkt.c so alerts are sent properly.
7223 [David Holmes <d.holmes@f5.com>]
7225 *) Reduce the chances of duplicate issuer name and serial numbers (in
7226 violation of RFC3280) using the OpenSSL certificate creation utilities.
7227 This is done by creating a random 64 bit value for the initial serial
7228 number when a serial number file is created or when a self signed
7229 certificate is created using 'openssl req -x509'. The initial serial
7230 number file is created using 'openssl x509 -next_serial' in CA.pl
7231 rather than being initialized to 1.
7234 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
7236 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7237 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7238 [Joe Orton, Steve Henson]
7240 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
7242 [Joe Orton, Steve Henson]
7244 *) Make it possible to have multiple active certificates with the same
7245 subject in the CA index file. This is done only if the keyword
7246 'unique_subject' is set to 'no' in the main CA section (default
7247 if 'CA_default') of the configuration file. The value is saved
7248 with the database itself in a separate index attribute file,
7249 named like the index file with '.attr' appended to the name.
7252 *) X509 verify fixes. Disable broken certificate workarounds when
7253 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
7254 keyUsage extension present. Don't accept CRLs with unhandled critical
7255 extensions: since verify currently doesn't process CRL extensions this
7256 rejects a CRL with *any* critical extensions. Add new verify error codes
7260 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
7261 A clarification of RFC2560 will require the use of OCTET STRINGs and
7262 some implementations cannot handle the current raw format. Since OpenSSL
7263 copies and compares OCSP nonces as opaque blobs without any attempt at
7264 parsing them this should not create any compatibility issues.
7267 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
7268 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
7269 this HMAC (and other) operations are several times slower than OpenSSL
7273 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
7274 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7276 *) Use the correct content when signing type "other".
7279 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
7281 *) Fix various bugs revealed by running the NISCC test suite:
7283 Stop out of bounds reads in the ASN1 code when presented with
7284 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7286 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
7288 If verify callback ignores invalid public key errors don't try to check
7289 certificate signature with the NULL public key.
7293 *) New -ignore_err option in ocsp application to stop the server
7294 exiting on the first error in a request.
7297 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7298 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7302 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7303 extra data after the compression methods not only for TLS 1.0
7304 but also for SSL 3.0 (as required by the specification).
7305 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7307 *) Change X509_certificate_type() to mark the key as exported/exportable
7308 when it's 512 *bits* long, not 512 bytes.
7311 *) Change AES_cbc_encrypt() so it outputs exact multiple of
7312 blocks during encryption.
7315 *) Various fixes to base64 BIO and non blocking I/O. On write
7316 flushes were not handled properly if the BIO retried. On read
7317 data was not being buffered properly and had various logic bugs.
7318 This also affects blocking I/O when the data being decoded is a
7322 *) Various S/MIME bugfixes and compatibility changes:
7323 output correct application/pkcs7 MIME type if
7324 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
7325 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
7326 of files as .eml work). Correctly handle very long lines in MIME
7330 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
7332 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7333 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7334 a protocol version number mismatch like a decryption error
7335 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7338 *) Turn on RSA blinding by default in the default implementation
7339 to avoid a timing attack. Applications that don't want it can call
7340 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7341 They would be ill-advised to do so in most cases.
7342 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7344 *) Change RSA blinding code so that it works when the PRNG is not
7345 seeded (in this case, the secret RSA exponent is abused as
7346 an unpredictable seed -- if it is not unpredictable, there
7347 is no point in blinding anyway). Make RSA blinding thread-safe
7348 by remembering the creator's thread ID in rsa->blinding and
7349 having all other threads use local one-time blinding factors
7350 (this requires more computation than sharing rsa->blinding, but
7351 avoids excessive locking; and if an RSA object is not shared
7352 between threads, blinding will still be very fast).
7355 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
7356 ENGINE as defaults for all supported algorithms irrespective of
7357 the 'flags' parameter. 'flags' is now honoured, so applications
7358 should make sure they are passing it correctly.
7361 *) Target "mingw" now allows native Windows code to be generated in
7362 the Cygwin environment as well as with the MinGW compiler.
7365 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7367 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7368 via timing by performing a MAC computation even if incorrect
7369 block cipher padding has been found. This is a countermeasure
7370 against active attacks where the attacker has to distinguish
7371 between bad padding and a MAC verification error. (CVE-2003-0078)
7373 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7374 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7375 Martin Vuagnoux (EPFL, Ilion)]
7377 *) Make the no-err option work as intended. The intention with no-err
7378 is not to have the whole error stack handling routines removed from
7379 libcrypto, it's only intended to remove all the function name and
7380 reason texts, thereby removing some of the footprint that may not
7381 be interesting if those errors aren't displayed anyway.
7383 NOTE: it's still possible for any application or module to have its
7384 own set of error texts inserted. The routines are there, just not
7385 used by default when no-err is given.
7388 *) Add support for FreeBSD on IA64.
7389 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7391 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7392 Kerberos function mit_des_cbc_cksum(). Before this change,
7393 the value returned by DES_cbc_cksum() was like the one from
7394 mit_des_cbc_cksum(), except the bytes were swapped.
7395 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7397 *) Allow an application to disable the automatic SSL chain building.
7398 Before this a rather primitive chain build was always performed in
7399 ssl3_output_cert_chain(): an application had no way to send the
7400 correct chain if the automatic operation produced an incorrect result.
7402 Now the chain builder is disabled if either:
7404 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7406 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7408 The reasoning behind this is that an application would not want the
7409 auto chain building to take place if extra chain certificates are
7410 present and it might also want a means of sending no additional
7411 certificates (for example the chain has two certificates and the
7415 *) Add the possibility to build without the ENGINE framework.
7416 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7418 *) Under Win32 gmtime() can return NULL: check return value in
7419 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7422 *) DSA routines: under certain error conditions uninitialized BN objects
7423 could be freed. Solution: make sure initialization is performed early
7424 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7425 Nils Larsch <nla@trustcenter.de> via PR#459)
7428 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7429 checked on reconnect on the client side, therefore session resumption
7430 could still fail with a "ssl session id is different" error. This
7431 behaviour is masked when SSL_OP_ALL is used due to
7432 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7433 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7434 followup to PR #377.
7437 *) IA-32 assembler support enhancements: unified ELF targets, support
7438 for SCO/Caldera platforms, fix for Cygwin shared build.
7441 *) Add support for FreeBSD on sparc64. As a consequence, support for
7442 FreeBSD on non-x86 processors is separate from x86 processors on
7443 the config script, much like the NetBSD support.
7444 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7446 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7448 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7451 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7452 code (06) was taken as the first octet of the session ID and the last
7453 octet was ignored consequently. As a result SSLv2 client side session
7454 caching could not have worked due to the session ID mismatch between
7456 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7460 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7461 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7465 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7466 seems that in spite of existing for more than a year, many application
7467 author have done nothing to provide the necessary callbacks, which
7468 means that this particular engine will not work properly anywhere.
7469 This is a very unfortunate situation which forces us, in the name
7470 of usability, to give the hw_ncipher.c a static lock, which is part
7472 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7473 appear in 0.9.8 or later. We EXPECT application authors to have
7474 dealt properly with this when 0.9.8 is released (unless we actually
7475 make such changes in the libcrypto locking code that changes will
7476 have to be made anyway).
7479 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7480 octets have been read, EOF or an error occurs. Without this change
7481 some truncated ASN1 structures will not produce an error.
7484 *) Disable Heimdal support, since it hasn't been fully implemented.
7485 Still give the possibility to force the use of Heimdal, but with
7486 warnings and a request that patches get sent to openssl-dev.
7489 *) Add the VC-CE target, introduce the WINCE sysname, and add
7490 INSTALL.WCE and appropriate conditionals to make it build.
7491 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7493 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7494 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7495 edit numbers of the version.
7496 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7498 *) Introduce safe string copy and catenation functions
7499 (BUF_strlcpy() and BUF_strlcat()).
7500 [Ben Laurie (CHATS) and Richard Levitte]
7502 *) Avoid using fixed-size buffers for one-line DNs.
7503 [Ben Laurie (CHATS)]
7505 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7506 resizing buffers containing secrets, and use where appropriate.
7507 [Ben Laurie (CHATS)]
7509 *) Avoid using fixed size buffers for configuration file location.
7510 [Ben Laurie (CHATS)]
7512 *) Avoid filename truncation for various CA files.
7513 [Ben Laurie (CHATS)]
7515 *) Use sizeof in preference to magic numbers.
7516 [Ben Laurie (CHATS)]
7518 *) Avoid filename truncation in cert requests.
7519 [Ben Laurie (CHATS)]
7521 *) Add assertions to check for (supposedly impossible) buffer
7523 [Ben Laurie (CHATS)]
7525 *) Don't cache truncated DNS entries in the local cache (this could
7526 potentially lead to a spoofing attack).
7527 [Ben Laurie (CHATS)]
7529 *) Fix various buffers to be large enough for hex/decimal
7530 representations in a platform independent manner.
7531 [Ben Laurie (CHATS)]
7533 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7534 resizing buffers containing secrets, and use where appropriate.
7535 [Ben Laurie (CHATS)]
7537 *) Add BIO_indent() to avoid much slightly worrying code to do
7539 [Ben Laurie (CHATS)]
7541 *) Convert sprintf()/BIO_puts() to BIO_printf().
7542 [Ben Laurie (CHATS)]
7544 *) buffer_gets() could terminate with the buffer only half
7546 [Ben Laurie (CHATS)]
7548 *) Add assertions to prevent user-supplied crypto functions from
7549 overflowing internal buffers by having large block sizes, etc.
7550 [Ben Laurie (CHATS)]
7552 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7554 [Ben Laurie (CHATS)]
7556 *) Eliminate unused copy of key in RC4.
7557 [Ben Laurie (CHATS)]
7559 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7560 [Ben Laurie (CHATS)]
7562 *) Fix off-by-one error in EGD path.
7563 [Ben Laurie (CHATS)]
7565 *) If RANDFILE path is too long, ignore instead of truncating.
7566 [Ben Laurie (CHATS)]
7568 *) Eliminate unused and incorrectly sized X.509 structure
7570 [Ben Laurie (CHATS)]
7572 *) Eliminate unused and dangerous function knumber().
7573 [Ben Laurie (CHATS)]
7575 *) Eliminate unused and dangerous structure, KSSL_ERR.
7576 [Ben Laurie (CHATS)]
7578 *) Protect against overlong session ID context length in an encoded
7579 session object. Since these are local, this does not appear to be
7581 [Ben Laurie (CHATS)]
7583 *) Change from security patch (see 0.9.6e below) that did not affect
7584 the 0.9.6 release series:
7586 Remote buffer overflow in SSL3 protocol - an attacker could
7587 supply an oversized master key in Kerberos-enabled versions.
7589 [Ben Laurie (CHATS)]
7591 *) Change the SSL kerb5 codes to match RFC 2712.
7594 *) Make -nameopt work fully for req and add -reqopt switch.
7595 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7597 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7598 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7600 *) Make sure tests can be performed even if the corresponding algorithms
7601 have been removed entirely. This was also the last step to make
7602 OpenSSL compilable with DJGPP under all reasonable conditions.
7603 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7605 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7606 to allow version independent disabling of normally unselected ciphers,
7607 which may be activated as a side-effect of selecting a single cipher.
7609 (E.g., cipher list string "RSA" enables ciphersuites that are left
7610 out of "ALL" because they do not provide symmetric encryption.
7611 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7612 [Lutz Jaenicke, Bodo Moeller]
7614 *) Add appropriate support for separate platform-dependent build
7615 directories. The recommended way to make a platform-dependent
7616 build directory is the following (tested on Linux), maybe with
7619 # Place yourself outside of the OpenSSL source tree. In
7620 # this example, the environment variable OPENSSL_SOURCE
7621 # is assumed to contain the absolute OpenSSL source directory.
7622 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7623 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7624 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7625 mkdir -p `dirname $F`
7626 ln -s $OPENSSL_SOURCE/$F $F
7629 To be absolutely sure not to disturb the source tree, a "make clean"
7630 is a good thing. If it isn't successful, don't worry about it,
7631 it probably means the source directory is very clean.
7634 *) Make sure any ENGINE control commands make local copies of string
7635 pointers passed to them whenever necessary. Otherwise it is possible
7636 the caller may have overwritten (or deallocated) the original string
7637 data when a later ENGINE operation tries to use the stored values.
7638 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7640 *) Improve diagnostics in file reading and command-line digests.
7641 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7643 *) Add AES modes CFB and OFB to the object database. Correct an
7644 error in AES-CFB decryption.
7647 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7648 allows existing EVP_CIPHER_CTX structures to be reused after
7649 calling EVP_*Final(). This behaviour is used by encryption
7650 BIOs and some applications. This has the side effect that
7651 applications must explicitly clean up cipher contexts with
7652 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7655 *) Check the values of dna and dnb in bn_mul_recursive before calling
7656 bn_mul_comba (a non zero value means the a or b arrays do not contain
7657 n2 elements) and fallback to bn_mul_normal if either is not zero.
7660 *) Fix escaping of non-ASCII characters when using the -subj option
7661 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7664 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7665 form for "surname", serialNumber has no short form.
7666 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7667 therefore remove "mail" short name for "internet 7".
7668 The OID for unique identifiers in X509 certificates is
7669 x500UniqueIdentifier, not uniqueIdentifier.
7670 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7673 *) Add an "init" command to the ENGINE config module and auto initialize
7674 ENGINEs. Without any "init" command the ENGINE will be initialized
7675 after all ctrl commands have been executed on it. If init=1 the
7676 ENGINE is initialized at that point (ctrls before that point are run
7677 on the uninitialized ENGINE and after on the initialized one). If
7678 init=0 then the ENGINE will not be initialized at all.
7681 *) Fix the 'app_verify_callback' interface so that the user-defined
7682 argument is actually passed to the callback: In the
7683 SSL_CTX_set_cert_verify_callback() prototype, the callback
7684 declaration has been changed from
7687 int (*cb)(X509_STORE_CTX *,void *);
7688 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7689 i=s->ctx->app_verify_callback(&ctx)
7690 has been changed into
7691 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7693 To update applications using SSL_CTX_set_cert_verify_callback(),
7694 a dummy argument can be added to their callback functions.
7695 [D. K. Smetters <smetters@parc.xerox.com>]
7697 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7698 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7700 *) Add and OPENSSL_LOAD_CONF define which will cause
7701 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7702 This allows older applications to transparently support certain
7703 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7704 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7705 load the config file and OPENSSL_add_all_algorithms_conf() which will
7706 always load it have also been added.
7709 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7710 Adjust NIDs and EVP layer.
7711 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7713 *) Config modules support in openssl utility.
7715 Most commands now load modules from the config file,
7716 though in a few (such as version) this isn't done
7717 because it couldn't be used for anything.
7719 In the case of ca and req the config file used is
7720 the same as the utility itself: that is the -config
7721 command line option can be used to specify an
7725 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7726 use "openssl_conf" if filename is NULL use default openssl config file.
7729 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7730 config section name. Add a new flag to tolerate a missing config file
7731 and move code to CONF_modules_load_file().
7734 *) Support for crypto accelerator cards from Accelerated Encryption
7735 Processing, www.aep.ie. (Use engine 'aep')
7736 The support was copied from 0.9.6c [engine] and adapted/corrected
7737 to work with the new engine framework.
7738 [AEP Inc. and Richard Levitte]
7740 *) Support for SureWare crypto accelerator cards from Baltimore
7741 Technologies. (Use engine 'sureware')
7742 The support was copied from 0.9.6c [engine] and adapted
7743 to work with the new engine framework.
7746 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7747 make the newer ENGINE framework commands for the CHIL engine work.
7748 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7750 *) Make it possible to produce shared libraries on ReliantUNIX.
7751 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7753 *) Add the configuration target debug-linux-ppro.
7754 Make 'openssl rsa' use the general key loading routines
7755 implemented in apps.c, and make those routines able to
7756 handle the key format FORMAT_NETSCAPE and the variant
7758 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7760 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7761 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7763 *) Add -keyform to rsautl, and document -engine.
7764 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7766 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7767 BIO_R_NO_SUCH_FILE error code rather than the generic
7768 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7771 *) Add new functions
7773 ERR_peek_last_error_line
7774 ERR_peek_last_error_line_data.
7775 These are similar to
7778 ERR_peek_error_line_data,
7779 but report on the latest error recorded rather than the first one
7780 still in the error queue.
7781 [Ben Laurie, Bodo Moeller]
7783 *) default_algorithms option in ENGINE config module. This allows things
7785 default_algorithms = ALL
7786 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7789 *) Preliminary ENGINE config module.
7792 *) New experimental application configuration code.
7795 *) Change the AES code to follow the same name structure as all other
7796 symmetric ciphers, and behave the same way. Move everything to
7797 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7798 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7800 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7801 [Ben Laurie and Theo de Raadt]
7803 *) Add option to output public keys in req command.
7804 [Massimiliano Pala madwolf@openca.org]
7806 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7807 (up to about 10% better than before for P-192 and P-224).
7810 *) New functions/macros
7812 SSL_CTX_set_msg_callback(ctx, cb)
7813 SSL_CTX_set_msg_callback_arg(ctx, arg)
7814 SSL_set_msg_callback(ssl, cb)
7815 SSL_set_msg_callback_arg(ssl, arg)
7817 to request calling a callback function
7819 void cb(int write_p, int version, int content_type,
7820 const void *buf, size_t len, SSL *ssl, void *arg)
7822 whenever a protocol message has been completely received
7823 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7824 protocol version according to which the SSL library interprets
7825 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7826 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7827 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7828 specification (change_cipher_spec(20), alert(21), handshake(22)).
7829 'buf' and 'len' point to the actual message, 'ssl' to the
7830 SSL object, and 'arg' is the application-defined value set by
7831 SSL[_CTX]_set_msg_callback_arg().
7833 'openssl s_client' and 'openssl s_server' have new '-msg' options
7834 to enable a callback that displays all protocol messages.
7837 *) Change the shared library support so shared libraries are built as
7838 soon as the corresponding static library is finished, and thereby get
7839 openssl and the test programs linked against the shared library.
7840 This still only happens when the keyword "shard" has been given to
7841 the configuration scripts.
7843 NOTE: shared library support is still an experimental thing, and
7844 backward binary compatibility is still not guaranteed.
7845 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7847 *) Add support for Subject Information Access extension.
7848 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7850 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7851 additional bytes when new memory had to be allocated, not just
7852 when reusing an existing buffer.
7855 *) New command line and configuration option 'utf8' for the req command.
7856 This allows field values to be specified as UTF8 strings.
7859 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7860 runs for the former and machine-readable output for the latter.
7863 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7864 of the e-mail address in the DN (i.e., it will go into a certificate
7865 extension only). The new configuration file option 'email_in_dn = no'
7866 has the same effect.
7867 [Massimiliano Pala madwolf@openca.org]
7869 *) Change all functions with names starting with des_ to be starting
7870 with DES_ instead. Add wrappers that are compatible with libdes,
7871 but are named _ossl_old_des_*. Finally, add macros that map the
7872 des_* symbols to the corresponding _ossl_old_des_* if libdes
7873 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7874 desired, the des_* symbols will be mapped to DES_*, with one
7877 Since we provide two compatibility mappings, the user needs to
7878 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7879 compatibility is desired. The default (i.e., when that macro
7880 isn't defined) is OpenSSL 0.9.6c compatibility.
7882 There are also macros that enable and disable the support of old
7883 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7884 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7885 are defined, the default will apply: to support the old des routines.
7887 In either case, one must include openssl/des.h to get the correct
7888 definitions. Do not try to just include openssl/des_old.h, that
7891 NOTE: This is a major break of an old API into a new one. Software
7892 authors are encouraged to switch to the DES_ style functions. Some
7893 time in the future, des_old.h and the libdes compatibility functions
7894 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7895 default), and then completely removed.
7898 *) Test for certificates which contain unsupported critical extensions.
7899 If such a certificate is found during a verify operation it is
7900 rejected by default: this behaviour can be overridden by either
7901 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7902 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7903 X509_supported_extension() has also been added which returns 1 if a
7904 particular extension is supported.
7907 *) Modify the behaviour of EVP cipher functions in similar way to digests
7908 to retain compatibility with existing code.
7911 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7912 compatibility with existing code. In particular the 'ctx' parameter does
7913 not have to be to be initialized before the call to EVP_DigestInit() and
7914 it is tidied up after a call to EVP_DigestFinal(). New function
7915 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7916 EVP_MD_CTX_copy() changed to not require the destination to be
7917 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7918 requires the destination to be valid.
7920 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7921 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7924 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7925 so that complete 'Handshake' protocol structures are kept in memory
7926 instead of overwriting 'msg_type' and 'length' with 'body' data.
7929 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7930 [Massimo Santin via Richard Levitte]
7932 *) Major restructuring to the underlying ENGINE code. This includes
7933 reduction of linker bloat, separation of pure "ENGINE" manipulation
7934 (initialisation, etc) from functionality dealing with implementations
7935 of specific crypto interfaces. This change also introduces integrated
7936 support for symmetric ciphers and digest implementations - so ENGINEs
7937 can now accelerate these by providing EVP_CIPHER and EVP_MD
7938 implementations of their own. This is detailed in crypto/engine/README
7939 as it couldn't be adequately described here. However, there are a few
7940 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7941 were changed in the original introduction of ENGINE code have now
7942 reverted back - the hooking from this code to ENGINE is now a good
7943 deal more passive and at run-time, operations deal directly with
7944 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7945 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7946 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7947 they were not being used by the framework as there is no concept of a
7948 BIGNUM_METHOD and they could not be generalised to the new
7949 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7950 ENGINE_cpy() has been removed as it cannot be consistently defined in
7954 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7957 *) Change mkdef.pl to sort symbols that get the same entry number,
7958 and make sure the automatically generated functions ERR_load_*
7959 become part of libeay.num as well.
7962 *) New function SSL_renegotiate_pending(). This returns true once
7963 renegotiation has been requested (either SSL_renegotiate() call
7964 or HelloRequest/ClientHello received from the peer) and becomes
7965 false once a handshake has been completed.
7966 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7967 sends a HelloRequest, but does not ensure that a handshake takes
7968 place. SSL_renegotiate_pending() is useful for checking if the
7969 client has followed the request.)
7972 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7973 By default, clients may request session resumption even during
7974 renegotiation (if session ID contexts permit); with this option,
7975 session resumption is possible only in the first handshake.
7977 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7978 more bits available for options that should not be part of
7979 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7982 *) Add some demos for certificate and certificate request creation.
7985 *) Make maximum certificate chain size accepted from the peer application
7986 settable (SSL*_get/set_max_cert_list()), as proposed by
7987 "Douglas E. Engert" <deengert@anl.gov>.
7990 *) Add support for shared libraries for Unixware-7
7991 (Boyd Lynn Gerber <gerberb@zenez.com>).
7994 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7995 be done prior to destruction. Use this to unload error strings from
7996 ENGINEs that load their own error strings. NB: This adds two new API
7997 functions to "get" and "set" this destroy handler in an ENGINE.
8000 *) Alter all existing ENGINE implementations (except "openssl" and
8001 "openbsd") to dynamically instantiate their own error strings. This
8002 makes them more flexible to be built both as statically-linked ENGINEs
8003 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
8004 Also, add stub code to each that makes building them as self-contained
8005 shared-libraries easier (see README.ENGINE).
8008 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
8009 implementations into applications that are completely implemented in
8010 self-contained shared-libraries. The "dynamic" ENGINE exposes control
8011 commands that can be used to configure what shared-library to load and
8012 to control aspects of the way it is handled. Also, made an update to
8013 the README.ENGINE file that brings its information up-to-date and
8014 provides some information and instructions on the "dynamic" ENGINE
8015 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
8018 *) Make it possible to unload ranges of ERR strings with a new
8019 "ERR_unload_strings" function.
8022 *) Add a copy() function to EVP_MD.
8025 *) Make EVP_MD routines take a context pointer instead of just the
8026 md_data void pointer.
8029 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
8030 that the digest can only process a single chunk of data
8031 (typically because it is provided by a piece of
8032 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
8033 is only going to provide a single chunk of data, and hence the
8034 framework needn't accumulate the data for oneshot drivers.
8037 *) As with "ERR", make it possible to replace the underlying "ex_data"
8038 functions. This change also alters the storage and management of global
8039 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
8040 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
8041 index counters. The API functions that use this state have been changed
8042 to take a "class_index" rather than pointers to the class's local STACK
8043 and counter, and there is now an API function to dynamically create new
8044 classes. This centralisation allows us to (a) plug a lot of the
8045 thread-safety problems that existed, and (b) makes it possible to clean
8046 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
8047 such data would previously have always leaked in application code and
8048 workarounds were in place to make the memory debugging turn a blind eye
8049 to it. Application code that doesn't use this new function will still
8050 leak as before, but their memory debugging output will announce it now
8051 rather than letting it slide.
8053 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
8054 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
8055 has a return value to indicate success or failure.
8058 *) Make it possible to replace the underlying "ERR" functions such that the
8059 global state (2 LHASH tables and 2 locks) is only used by the "default"
8060 implementation. This change also adds two functions to "get" and "set"
8061 the implementation prior to it being automatically set the first time
8062 any other ERR function takes place. Ie. an application can call "get",
8063 pass the return value to a module it has just loaded, and that module
8064 can call its own "set" function using that value. This means the
8065 module's "ERR" operations will use (and modify) the error state in the
8066 application and not in its own statically linked copy of OpenSSL code.
8069 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
8070 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
8071 the operation, and provides a more encapsulated way for external code
8072 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
8073 to use these functions rather than manually incrementing the counts.
8075 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
8078 *) Add EVP test program.
8081 *) Add symmetric cipher support to ENGINE. Expect the API to change!
8084 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
8085 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
8086 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
8087 These allow a CRL to be built without having to access X509_CRL fields
8088 directly. Modify 'ca' application to use new functions.
8091 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
8092 bug workarounds. Rollback attack detection is a security feature.
8093 The problem will only arise on OpenSSL servers when TLSv1 is not
8094 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
8095 Software authors not wanting to support TLSv1 will have special reasons
8096 for their choice and can explicitly enable this option.
8097 [Bodo Moeller, Lutz Jaenicke]
8099 *) Rationalise EVP so it can be extended: don't include a union of
8100 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
8101 (similar to those existing for EVP_CIPHER_CTX).
8106 EVP_MD_CTX_init(&md); /* new function call */
8107 EVP_DigestInit(&md, EVP_sha1());
8108 EVP_DigestUpdate(&md, in, len);
8109 EVP_DigestFinal(&md, out, NULL);
8110 EVP_MD_CTX_cleanup(&md); /* new function call */
8114 *) Make DES key schedule conform to the usual scheme, as well as
8115 correcting its structure. This means that calls to DES functions
8116 now have to pass a pointer to a des_key_schedule instead of a
8117 plain des_key_schedule (which was actually always a pointer
8120 des_key_schedule ks;
8122 des_set_key_checked(..., &ks);
8123 des_ncbc_encrypt(..., &ks, ...);
8125 (Note that a later change renames 'des_...' into 'DES_...'.)
8128 *) Initial reduction of linker bloat: the use of some functions, such as
8129 PEM causes large amounts of unused functions to be linked in due to
8130 poor organisation. For example pem_all.c contains every PEM function
8131 which has a knock on effect of linking in large amounts of (unused)
8132 ASN1 code. Grouping together similar functions and splitting unrelated
8133 functions prevents this.
8136 *) Cleanup of EVP macros.
8139 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
8140 correct _ecb suffix.
8143 *) Add initial OCSP responder support to ocsp application. The
8144 revocation information is handled using the text based index
8145 use by the ca application. The responder can either handle
8146 requests generated internally, supplied in files (for example
8147 via a CGI script) or using an internal minimal server.
8150 *) Add configuration choices to get zlib compression for TLS.
8153 *) Changes to Kerberos SSL for RFC 2712 compliance:
8154 1. Implemented real KerberosWrapper, instead of just using
8155 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
8156 2. Implemented optional authenticator field of KerberosWrapper.
8158 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
8159 and authenticator structs; see crypto/krb5/.
8161 Generalized Kerberos calls to support multiple Kerberos libraries.
8162 [Vern Staats <staatsvr@asc.hpc.mil>,
8163 Jeffrey Altman <jaltman@columbia.edu>
8164 via Richard Levitte]
8166 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
8167 already does with RSA. testdsa.h now has 'priv_key/pub_key'
8168 values for each of the key sizes rather than having just
8169 parameters (and 'speed' generating keys each time).
8172 *) Speed up EVP routines.
8175 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
8176 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
8177 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
8178 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
8180 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
8181 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
8182 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
8185 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
8187 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
8190 *) Added the OS2-EMX target.
8191 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
8193 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
8194 to support NCONF routines in extension code. New function CONF_set_nconf()
8195 to allow functions which take an NCONF to also handle the old LHASH
8196 structure: this means that the old CONF compatible routines can be
8197 retained (in particular wrt extensions) without having to duplicate the
8198 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
8201 *) Enhance the general user interface with mechanisms for inner control
8202 and with possibilities to have yes/no kind of prompts.
8205 *) Change all calls to low level digest routines in the library and
8206 applications to use EVP. Add missing calls to HMAC_cleanup() and
8207 don't assume HMAC_CTX can be copied using memcpy().
8208 [Verdon Walker <VWalker@novell.com>, Steve Henson]
8210 *) Add the possibility to control engines through control names but with
8211 arbitrary arguments instead of just a string.
8212 Change the key loaders to take a UI_METHOD instead of a callback
8213 function pointer. NOTE: this breaks binary compatibility with earlier
8214 versions of OpenSSL [engine].
8215 Adapt the nCipher code for these new conditions and add a card insertion
8219 *) Enhance the general user interface with mechanisms to better support
8220 dialog box interfaces, application-defined prompts, the possibility
8221 to use defaults (for example default passwords from somewhere else)
8222 and interrupts/cancellations.
8225 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
8226 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
8229 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
8230 tidy up some unnecessarily weird code in 'sk_new()').
8231 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
8233 *) Change the key loading routines for ENGINEs to use the same kind
8234 callback (pem_password_cb) as all other routines that need this
8238 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
8239 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
8240 than this minimum value is recommended.
8243 *) New random seeder for OpenVMS, using the system process statistics
8244 that are easily reachable.
8247 *) Windows apparently can't transparently handle global
8248 variables defined in DLLs. Initialisations such as:
8250 const ASN1_ITEM *it = &ASN1_INTEGER_it;
8252 won't compile. This is used by the any applications that need to
8253 declare their own ASN1 modules. This was fixed by adding the option
8254 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
8255 needed for static libraries under Win32.
8258 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
8259 setting of purpose and trust fields. New X509_STORE trust and
8260 purpose functions and tidy up setting in other SSL functions.
8263 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
8264 structure. These are inherited by X509_STORE_CTX when it is
8265 initialised. This allows various defaults to be set in the
8266 X509_STORE structure (such as flags for CRL checking and custom
8267 purpose or trust settings) for functions which only use X509_STORE_CTX
8268 internally such as S/MIME.
8270 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
8271 trust settings if they are not set in X509_STORE. This allows X509_STORE
8272 purposes and trust (in S/MIME for example) to override any set by default.
8274 Add command line options for CRL checking to smime, s_client and s_server
8278 *) Initial CRL based revocation checking. If the CRL checking flag(s)
8279 are set then the CRL is looked up in the X509_STORE structure and
8280 its validity and signature checked, then if the certificate is found
8281 in the CRL the verify fails with a revoked error.
8283 Various new CRL related callbacks added to X509_STORE_CTX structure.
8285 Command line options added to 'verify' application to support this.
8287 This needs some additional work, such as being able to handle multiple
8288 CRLs with different times, extension based lookup (rather than just
8289 by subject name) and ultimately more complete V2 CRL extension
8293 *) Add a general user interface API (crypto/ui/). This is designed
8294 to replace things like des_read_password and friends (backward
8295 compatibility functions using this new API are provided).
8296 The purpose is to remove prompting functions from the DES code
8297 section as well as provide for prompting through dialog boxes in
8298 a window system and the like.
8301 *) Add "ex_data" support to ENGINE so implementations can add state at a
8302 per-structure level rather than having to store it globally.
8305 *) Make it possible for ENGINE structures to be copied when retrieved by
8306 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
8307 This causes the "original" ENGINE structure to act like a template,
8308 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
8309 operational state can be localised to each ENGINE structure, despite the
8310 fact they all share the same "methods". New ENGINE structures returned in
8311 this case have no functional references and the return value is the single
8312 structural reference. This matches the single structural reference returned
8313 by ENGINE_by_id() normally, when it is incremented on the pre-existing
8317 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
8318 needs to match any other type at all we need to manually clear the
8322 *) Changes to the "openssl engine" utility to include;
8323 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
8324 about an ENGINE's available control commands.
8325 - executing control commands from command line arguments using the
8326 '-pre' and '-post' switches. '-post' is only used if '-t' is
8327 specified and the ENGINE is successfully initialised. The syntax for
8328 the individual commands are colon-separated, for example;
8329 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
8332 *) New dynamic control command support for ENGINEs. ENGINEs can now
8333 declare their own commands (numbers), names (strings), descriptions,
8334 and input types for run-time discovery by calling applications. A
8335 subset of these commands are implicitly classed as "executable"
8336 depending on their input type, and only these can be invoked through
8337 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
8338 can be based on user input, config files, etc). The distinction is
8339 that "executable" commands cannot return anything other than a boolean
8340 result and can only support numeric or string input, whereas some
8341 discoverable commands may only be for direct use through
8342 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
8343 pointers, or other custom uses. The "executable" commands are to
8344 support parameterisations of ENGINE behaviour that can be
8345 unambiguously defined by ENGINEs and used consistently across any
8346 OpenSSL-based application. Commands have been added to all the
8347 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
8348 control over shared-library paths without source code alterations.
8351 *) Changed all ENGINE implementations to dynamically allocate their
8352 ENGINEs rather than declaring them statically. Apart from this being
8353 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
8354 this also allows the implementations to compile without using the
8355 internal engine_int.h header.
8358 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8359 'const' value. Any code that should be able to modify a RAND_METHOD
8360 should already have non-const pointers to it (ie. they should only
8361 modify their own ones).
8364 *) Made a variety of little tweaks to the ENGINE code.
8365 - "atalla" and "ubsec" string definitions were moved from header files
8366 to C code. "nuron" string definitions were placed in variables
8367 rather than hard-coded - allowing parameterisation of these values
8368 later on via ctrl() commands.
8369 - Removed unused "#if 0"'d code.
8370 - Fixed engine list iteration code so it uses ENGINE_free() to release
8371 structural references.
8372 - Constified the RAND_METHOD element of ENGINE structures.
8373 - Constified various get/set functions as appropriate and added
8374 missing functions (including a catch-all ENGINE_cpy that duplicates
8375 all ENGINE values onto a new ENGINE except reference counts/state).
8376 - Removed NULL parameter checks in get/set functions. Setting a method
8377 or function to NULL is a way of cancelling out a previously set
8378 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8379 and doesn't justify the extra error symbols and code.
8380 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8381 flags from engine_int.h to engine.h.
8382 - Changed prototypes for ENGINE handler functions (init(), finish(),
8383 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8386 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8387 to the algorithm using long division. The binary algorithm can be
8388 used only if the modulus is odd. On 32-bit systems, it is faster
8389 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8390 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8391 up to 450 bits. In 64-bit environments, the binary algorithm
8392 appears to be advantageous for much longer moduli; here we use it
8393 for moduli up to 2048 bits.
8396 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8397 could not support the combine flag in choice fields.
8400 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8401 extensions from a certificate request to the certificate.
8404 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8405 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8406 file: this allows the display of the certificate about to be
8407 signed to be customised, to allow certain fields to be included
8408 or excluded and extension details. The old system didn't display
8409 multicharacter strings properly, omitted fields not in the policy
8410 and couldn't display additional details such as extensions.
8413 *) Function EC_POINTs_mul for multiple scalar multiplication
8414 of an arbitrary number of elliptic curve points
8415 \sum scalars[i]*points[i],
8416 optionally including the generator defined for the EC_GROUP:
8417 scalar*generator + \sum scalars[i]*points[i].
8419 EC_POINT_mul is a simple wrapper function for the typical case
8420 that the point list has just one item (besides the optional
8424 *) First EC_METHODs for curves over GF(p):
8426 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8427 operations and provides various method functions that can also
8428 operate with faster implementations of modular arithmetic.
8430 EC_GFp_mont_method() reuses most functions that are part of
8431 EC_GFp_simple_method, but uses Montgomery arithmetic.
8433 [Bodo Moeller; point addition and point doubling
8434 implementation directly derived from source code provided by
8435 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8437 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8438 crypto/ec/ec_lib.c):
8440 Curves are EC_GROUP objects (with an optional group generator)
8441 based on EC_METHODs that are built into the library.
8443 Points are EC_POINT objects based on EC_GROUP objects.
8445 Most of the framework would be able to handle curves over arbitrary
8446 finite fields, but as there are no obvious types for fields other
8447 than GF(p), some functions are limited to that for now.
8450 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8451 that the file contains a complete HTTP response.
8454 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8455 change the def and num file printf format specifier from "%-40sXXX"
8456 to "%-39s XXX". The latter will always guarantee a space after the
8457 field while the former will cause them to run together if the field
8458 is 40 of more characters long.
8461 *) Constify the cipher and digest 'method' functions and structures
8462 and modify related functions to take constant EVP_MD and EVP_CIPHER
8466 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8467 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8470 *) Modify EVP_Digest*() routines so they now return values. Although the
8471 internal software routines can never fail additional hardware versions
8475 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8477 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8478 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8481 ERR_R_NESTED_ASN1_ERROR
8483 ERR_R_MISSING_ASN1_EOS
8484 were 4 .. 9, conflicting with
8485 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8487 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8488 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8490 Add new error code 'ERR_R_INTERNAL_ERROR'.
8493 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8497 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8498 sets the subject name for a new request or supersedes the
8499 subject name in a given request. Formats that can be parsed are
8500 'CN=Some Name, OU=myOU, C=IT'
8502 'CN=Some Name/OU=myOU/C=IT'.
8504 Add options '-batch' and '-verbose' to 'openssl req'.
8505 [Massimiliano Pala <madwolf@hackmasters.net>]
8507 *) Introduce the possibility to access global variables through
8508 functions on platform were that's the best way to handle exporting
8509 global variables in shared libraries. To enable this functionality,
8510 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8511 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8512 is normally done by Configure or something similar).
8514 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8515 in the source file (foo.c) like this:
8517 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8518 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8520 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8521 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8523 OPENSSL_DECLARE_GLOBAL(int,foo);
8524 #define foo OPENSSL_GLOBAL_REF(foo)
8525 OPENSSL_DECLARE_GLOBAL(double,bar);
8526 #define bar OPENSSL_GLOBAL_REF(bar)
8528 The #defines are very important, and therefore so is including the
8529 header file everywhere where the defined globals are used.
8531 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8532 of ASN.1 items, but that structure is a bit different.
8534 The largest change is in util/mkdef.pl which has been enhanced with
8535 better and easier to understand logic to choose which symbols should
8536 go into the Windows .def files as well as a number of fixes and code
8537 cleanup (among others, algorithm keywords are now sorted
8538 lexicographically to avoid constant rewrites).
8541 *) In BN_div() keep a copy of the sign of 'num' before writing the
8542 result to 'rm' because if rm==num the value will be overwritten
8543 and produce the wrong result if 'num' is negative: this caused
8544 problems with BN_mod() and BN_nnmod().
8547 *) Function OCSP_request_verify(). This checks the signature on an
8548 OCSP request and verifies the signer certificate. The signer
8549 certificate is just checked for a generic purpose and OCSP request
8553 *) Add OCSP_check_validity() function to check the validity of OCSP
8554 responses. OCSP responses are prepared in real time and may only
8555 be a few seconds old. Simply checking that the current time lies
8556 between thisUpdate and nextUpdate max reject otherwise valid responses
8557 caused by either OCSP responder or client clock inaccuracy. Instead
8558 we allow thisUpdate and nextUpdate to fall within a certain period of
8559 the current time. The age of the response can also optionally be
8560 checked. Two new options -validity_period and -status_age added to
8564 *) If signature or public key algorithm is unrecognized print out its
8565 OID rather that just UNKNOWN.
8568 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8569 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8570 ID to be generated from the issuer certificate alone which can then be
8571 passed to OCSP_id_issuer_cmp().
8574 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8575 ASN1 modules to export functions returning ASN1_ITEM pointers
8576 instead of the ASN1_ITEM structures themselves. This adds several
8577 new macros which allow the underlying ASN1 function/structure to
8578 be accessed transparently. As a result code should not use ASN1_ITEM
8579 references directly (such as &X509_it) but instead use the relevant
8580 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8581 use of the new ASN1 code on platforms where exporting structures
8582 is problematical (for example in shared libraries) but exporting
8583 functions returning pointers to structures is not.
8586 *) Add support for overriding the generation of SSL/TLS session IDs.
8587 These callbacks can be registered either in an SSL_CTX or per SSL.
8588 The purpose of this is to allow applications to control, if they wish,
8589 the arbitrary values chosen for use as session IDs, particularly as it
8590 can be useful for session caching in multiple-server environments. A
8591 command-line switch for testing this (and any client code that wishes
8592 to use such a feature) has been added to "s_server".
8593 [Geoff Thorpe, Lutz Jaenicke]
8595 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8596 of the form '#if defined(...) || defined(...) || ...' and
8597 '#if !defined(...) && !defined(...) && ...'. This also avoids
8598 the growing number of special cases it was previously handling.
8601 *) Make all configuration macros available for application by making
8602 sure they are available in opensslconf.h, by giving them names starting
8603 with "OPENSSL_" to avoid conflicts with other packages and by making
8604 sure e_os2.h will cover all platform-specific cases together with
8606 Additionally, it is now possible to define configuration/platform-
8607 specific names (called "system identities"). In the C code, these
8608 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8609 macro with the name beginning with "OPENSSL_SYS_", which is determined
8610 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8614 *) New option -set_serial to 'req' and 'x509' this allows the serial
8615 number to use to be specified on the command line. Previously self
8616 signed certificates were hard coded with serial number 0 and the
8617 CA options of 'x509' had to use a serial number in a file which was
8621 *) New options to 'ca' utility to support V2 CRL entry extensions.
8622 Currently CRL reason, invalidity date and hold instruction are
8623 supported. Add new CRL extensions to V3 code and some new objects.
8626 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8627 disable standard block padding (aka PKCS#5 padding) in the EVP
8628 API, which was previously mandatory. This means that the data is
8629 not padded in any way and so the total length much be a multiple
8630 of the block size, otherwise an error occurs.
8633 *) Initial (incomplete) OCSP SSL support.
8636 *) New function OCSP_parse_url(). This splits up a URL into its host,
8637 port and path components: primarily to parse OCSP URLs. New -url
8638 option to ocsp utility.
8641 *) New nonce behavior. The return value of OCSP_check_nonce() now
8642 reflects the various checks performed. Applications can decide
8643 whether to tolerate certain situations such as an absent nonce
8644 in a response when one was present in a request: the ocsp application
8645 just prints out a warning. New function OCSP_add1_basic_nonce()
8646 this is to allow responders to include a nonce in a response even if
8647 the request is nonce-less.
8650 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8651 skipped when using openssl x509 multiple times on a single input file,
8652 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8655 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8656 set string type: to handle setting ASN1_TIME structures. Fix ca
8657 utility to correctly initialize revocation date of CRLs.
8660 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8661 the clients preferred ciphersuites and rather use its own preferences.
8662 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8663 Internet Explorer by ensuring unchanged hash method during stepup.
8664 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8667 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8668 to aes and add a new 'exist' option to print out symbols that don't
8672 *) Additional options to ocsp utility to allow flags to be set and
8673 additional certificates supplied.
8676 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8677 OCSP client a number of certificate to only verify the response
8681 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8682 handle the new API. Currently only ECB, CBC modes supported. Add new
8685 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8686 Encryption Standard (AES) Ciphersuites for Transport Layer
8687 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8688 not enabled by default and were not part of the "ALL" ciphersuite
8689 alias because they were not yet official; they could be
8690 explicitly requested by specifying the "AESdraft" ciphersuite
8691 group alias. In the final release of OpenSSL 0.9.7, the group
8692 alias is called "AES" and is part of "ALL".)
8693 [Ben Laurie, Steve Henson, Bodo Moeller]
8695 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8696 request to response.
8699 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8700 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8701 extract information from a certificate request. OCSP_response_create()
8702 creates a response and optionally adds a basic response structure.
8703 OCSP_basic_add1_status() adds a complete single response to a basic
8704 response and returns the OCSP_SINGLERESP structure just added (to allow
8705 extensions to be included for example). OCSP_basic_add1_cert() adds a
8706 certificate to a basic response and OCSP_basic_sign() signs a basic
8707 response with various flags. New helper functions ASN1_TIME_check()
8708 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8709 (converts ASN1_TIME to GeneralizedTime).
8712 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8713 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8714 structure from a certificate. X509_pubkey_digest() digests the public_key
8715 contents: this is used in various key identifiers.
8718 *) Make sk_sort() tolerate a NULL argument.
8719 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8721 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8722 passed by the function are trusted implicitly. If any of them signed the
8723 response then it is assumed to be valid and is not verified.
8726 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8727 to data. This was previously part of the PKCS7 ASN1 code. This
8728 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8729 [Steve Henson, reported by Kenneth R. Robinette
8730 <support@securenetterm.com>]
8732 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8733 routines: without these tracing memory leaks is very painful.
8734 Fix leaks in PKCS12 and PKCS7 routines.
8737 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8738 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8739 effectively meant GeneralizedTime would never be used. Now it
8740 is initialised to -1 but X509_time_adj() now has to check the value
8741 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8742 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8743 [Steve Henson, reported by Kenneth R. Robinette
8744 <support@securenetterm.com>]
8746 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8747 result in a zero length in the ASN1_INTEGER structure which was
8748 not consistent with the structure when d2i_ASN1_INTEGER() was used
8749 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8750 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8751 where it did not print out a minus for negative ASN1_INTEGER.
8754 *) Add summary printout to ocsp utility. The various functions which
8755 convert status values to strings have been renamed to:
8756 OCSP_response_status_str(), OCSP_cert_status_str() and
8757 OCSP_crl_reason_str() and are no longer static. New options
8758 to verify nonce values and to disable verification. OCSP response
8759 printout format cleaned up.
8762 *) Add additional OCSP certificate checks. These are those specified
8763 in RFC2560. This consists of two separate checks: the CA of the
8764 certificate being checked must either be the OCSP signer certificate
8765 or the issuer of the OCSP signer certificate. In the latter case the
8766 OCSP signer certificate must contain the OCSP signing extended key
8767 usage. This check is performed by attempting to match the OCSP
8768 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8769 in the OCSP_CERTID structures of the response.
8772 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8773 and related routines. This uses the standard OpenSSL certificate
8774 verify routines to perform initial checks (just CA validity) and
8775 to obtain the certificate chain. Then additional checks will be
8776 performed on the chain. Currently the root CA is checked to see
8777 if it is explicitly trusted for OCSP signing. This is used to set
8778 a root CA as a global signing root: that is any certificate that
8779 chains to that CA is an acceptable OCSP signing certificate.
8782 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8783 extensions from a separate configuration file.
8784 As when reading extensions from the main configuration file,
8785 the '-extensions ...' option may be used for specifying the
8787 [Massimiliano Pala <madwolf@comune.modena.it>]
8789 *) New OCSP utility. Allows OCSP requests to be generated or
8790 read. The request can be sent to a responder and the output
8791 parsed, outputted or printed in text form. Not complete yet:
8792 still needs to check the OCSP response validity.
8795 *) New subcommands for 'openssl ca':
8796 'openssl ca -status <serial>' prints the status of the cert with
8797 the given serial number (according to the index file).
8798 'openssl ca -updatedb' updates the expiry status of certificates
8800 [Massimiliano Pala <madwolf@comune.modena.it>]
8802 *) New '-newreq-nodes' command option to CA.pl. This is like
8803 '-newreq', but calls 'openssl req' with the '-nodes' option
8804 so that the resulting key is not encrypted.
8805 [Damien Miller <djm@mindrot.org>]
8807 *) New configuration for the GNU Hurd.
8808 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8810 *) Initial code to implement OCSP basic response verify. This
8811 is currently incomplete. Currently just finds the signer's
8812 certificate and verifies the signature on the response.
8815 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8816 value of OPENSSLDIR. This is available via the new '-d' option
8817 to 'openssl version', and is also included in 'openssl version -a'.
8820 *) Allowing defining memory allocation callbacks that will be given
8821 file name and line number information in additional arguments
8822 (a const char* and an int). The basic functionality remains, as
8823 well as the original possibility to just replace malloc(),
8824 realloc() and free() by functions that do not know about these
8825 additional arguments. To register and find out the current
8826 settings for extended allocation functions, the following
8827 functions are provided:
8829 CRYPTO_set_mem_ex_functions
8830 CRYPTO_set_locked_mem_ex_functions
8831 CRYPTO_get_mem_ex_functions
8832 CRYPTO_get_locked_mem_ex_functions
8834 These work the same way as CRYPTO_set_mem_functions and friends.
8835 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8836 extended allocation function is enabled.
8837 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8838 a conventional allocation function is enabled.
8839 [Richard Levitte, Bodo Moeller]
8841 *) Finish off removing the remaining LHASH function pointer casts.
8842 There should no longer be any prototype-casting required when using
8843 the LHASH abstraction, and any casts that remain are "bugs". See
8844 the callback types and macros at the head of lhash.h for details
8845 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8848 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8849 If /dev/[u]random devices are not available or do not return enough
8850 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8852 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8853 /etc/entropy will be queried once each in this sequence, querying stops
8854 when enough entropy was collected without querying more sockets.
8857 *) Change the Unix RAND_poll() variant to be able to poll several
8858 random devices, as specified by DEVRANDOM, until a sufficient amount
8859 of data has been collected. We spend at most 10 ms on each file
8860 (select timeout) and read in non-blocking mode. DEVRANDOM now
8861 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8862 (previously it was just the string "/dev/urandom"), so on typical
8863 platforms the 10 ms delay will never occur.
8864 Also separate out the Unix variant to its own file, rand_unix.c.
8865 For VMS, there's a currently-empty rand_vms.c.
8868 *) Move OCSP client related routines to ocsp_cl.c. These
8869 provide utility functions which an application needing
8870 to issue a request to an OCSP responder and analyse the
8871 response will typically need: as opposed to those which an
8872 OCSP responder itself would need which will be added later.
8874 OCSP_request_sign() signs an OCSP request with an API similar
8875 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8876 response. OCSP_response_get1_basic() extracts basic response
8877 from response. OCSP_resp_find_status(): finds and extracts status
8878 information from an OCSP_CERTID structure (which will be created
8879 when the request structure is built). These are built from lower
8880 level functions which work on OCSP_SINGLERESP structures but
8881 won't normally be used unless the application wishes to examine
8882 extensions in the OCSP response for example.
8884 Replace nonce routines with a pair of functions.
8885 OCSP_request_add1_nonce() adds a nonce value and optionally
8886 generates a random value. OCSP_check_nonce() checks the
8887 validity of the nonce in an OCSP response.
8890 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8891 This doesn't copy the supplied OCSP_CERTID and avoids the
8892 need to free up the newly created id. Change return type
8893 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8894 This can then be used to add extensions to the request.
8895 Deleted OCSP_request_new(), since most of its functionality
8896 is now in OCSP_REQUEST_new() (and the case insensitive name
8897 clash) apart from the ability to set the request name which
8898 will be added elsewhere.
8901 *) Update OCSP API. Remove obsolete extensions argument from
8902 various functions. Extensions are now handled using the new
8903 OCSP extension code. New simple OCSP HTTP function which
8904 can be used to send requests and parse the response.
8907 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8908 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8909 uses the special reorder version of SET OF to sort the attributes
8910 and reorder them to match the encoded order. This resolves a long
8911 standing problem: a verify on a PKCS7 structure just after signing
8912 it used to fail because the attribute order did not match the
8913 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8914 it uses the received order. This is necessary to tolerate some broken
8915 software that does not order SET OF. This is handled by encoding
8916 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8917 to produce the required SET OF.
8920 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8921 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8922 files to get correct declarations of the ASN.1 item variables.
8925 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8926 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8927 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8928 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8929 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8930 ASN1_ITEM and no wrapper functions.
8933 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8934 replace the old function pointer based I/O routines. Change most of
8935 the *_d2i_bio() and *_d2i_fp() functions to use these.
8938 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8939 lines, recognize more "algorithms" that can be deselected, and make
8940 it complain about algorithm deselection that isn't recognised.
8943 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8944 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8945 to use new functions. Add NO_ASN1_OLD which can be set to remove
8946 some old style ASN1 functions: this can be used to determine if old
8947 code will still work when these eventually go away.
8950 *) New extension functions for OCSP structures, these follow the
8951 same conventions as certificates and CRLs.
8954 *) New function X509V3_add1_i2d(). This automatically encodes and
8955 adds an extension. Its behaviour can be customised with various
8956 flags to append, replace or delete. Various wrappers added for
8957 certificates and CRLs.
8960 *) Fix to avoid calling the underlying ASN1 print routine when
8961 an extension cannot be parsed. Correct a typo in the
8962 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8965 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8966 entries for variables.
8969 *) Add functionality to apps/openssl.c for detecting locking
8970 problems: As the program is single-threaded, all we have
8971 to do is register a locking callback using an array for
8972 storing which locks are currently held by the program.
8975 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8976 SSL_get_ex_data_X509_STORE_idx(), which is used in
8977 ssl_verify_cert_chain() and thus can be called at any time
8978 during TLS/SSL handshakes so that thread-safety is essential.
8979 Unfortunately, the ex_data design is not at all suited
8980 for multi-threaded use, so it probably should be abolished.
8983 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8984 [Broadcom, tweaked and integrated by Geoff Thorpe]
8986 *) Move common extension printing code to new function
8987 X509V3_print_extensions(). Reorganise OCSP print routines and
8988 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8991 *) New function X509_signature_print() to remove duplication in some
8995 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8996 set (this was treated exactly the same as SET OF previously). This
8997 is used to reorder the STACK representing the structure to match the
8998 encoding. This will be used to get round a problem where a PKCS7
8999 structure which was signed could not be verified because the STACK
9000 order did not reflect the encoded order.
9003 *) Reimplement the OCSP ASN1 module using the new code.
9006 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
9007 for its ASN1 operations. The old style function pointers still exist
9008 for now but they will eventually go away.
9011 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
9012 completely replaces the old ASN1 functionality with a table driven
9013 encoder and decoder which interprets an ASN1_ITEM structure describing
9014 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
9015 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
9016 has also been converted to the new form.
9019 *) Change BN_mod_exp_recp so that negative moduli are tolerated
9020 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
9021 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
9022 for negative moduli.
9025 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
9026 of not touching the result's sign bit.
9029 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
9033 *) Changed the LHASH code to use prototypes for callbacks, and created
9034 macros to declare and implement thin (optionally static) functions
9035 that provide type-safety and avoid function pointer casting for the
9036 type-specific callbacks.
9039 *) Added Kerberos Cipher Suites to be used with TLS, as written in
9041 [Veers Staats <staatsvr@asc.hpc.mil>,
9042 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
9044 *) Reformat the FAQ so the different questions and answers can be divided
9045 in sections depending on the subject.
9048 *) Have the zlib compression code load ZLIB.DLL dynamically under
9052 *) New function BN_mod_sqrt for computing square roots modulo a prime
9053 (using the probabilistic Tonelli-Shanks algorithm unless
9054 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
9055 be handled deterministically).
9056 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9058 *) Make BN_mod_inverse faster by explicitly handling small quotients
9059 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
9060 512 bits], about 30% for larger ones [1024 or 2048 bits].)
9063 *) New function BN_kronecker.
9066 *) Fix BN_gcd so that it works on negative inputs; the result is
9067 positive unless both parameters are zero.
9068 Previously something reasonably close to an infinite loop was
9069 possible because numbers could be growing instead of shrinking
9070 in the implementation of Euclid's algorithm.
9073 *) Fix BN_is_word() and BN_is_one() macros to take into account the
9074 sign of the number in question.
9076 Fix BN_is_word(a,w) to work correctly for w == 0.
9078 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
9079 because its test if the absolute value of 'a' equals 'w'.
9080 Note that BN_abs_is_word does *not* handle w == 0 reliably;
9081 it exists mostly for use in the implementations of BN_is_zero(),
9082 BN_is_one(), and BN_is_word().
9085 *) New function BN_swap.
9088 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
9089 the exponentiation functions are more likely to produce reasonable
9090 results on negative inputs.
9093 *) Change BN_mod_mul so that the result is always non-negative.
9094 Previously, it could be negative if one of the factors was negative;
9095 I don't think anyone really wanted that behaviour.
9098 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
9099 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
9100 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
9101 and add new functions:
9110 BN_mod_lshift1_quick
9114 These functions always generate non-negative results.
9116 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
9117 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
9119 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
9120 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
9121 be reduced modulo m.
9122 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9125 The following entry accidentally appeared in the CHANGES file
9126 distributed with OpenSSL 0.9.7. The modifications described in
9127 it do *not* apply to OpenSSL 0.9.7.
9129 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
9130 was actually never needed) and in BN_mul(). The removal in BN_mul()
9131 required a small change in bn_mul_part_recursive() and the addition
9132 of the functions bn_cmp_part_words(), bn_sub_part_words() and
9133 bn_add_part_words(), which do the same thing as bn_cmp_words(),
9134 bn_sub_words() and bn_add_words() except they take arrays with
9139 *) In 'openssl passwd', verify passwords read from the terminal
9140 unless the '-salt' option is used (which usually means that
9141 verification would just waste user's time since the resulting
9142 hash is going to be compared with some given password hash)
9143 or the new '-noverify' option is used.
9145 This is an incompatible change, but it does not affect
9146 non-interactive use of 'openssl passwd' (passwords on the command
9147 line, '-stdin' option, '-in ...' option) and thus should not
9151 *) Remove all references to RSAref, since there's no more need for it.
9154 *) Make DSO load along a path given through an environment variable
9155 (SHLIB_PATH) with shl_load().
9158 *) Constify the ENGINE code as a result of BIGNUM constification.
9159 Also constify the RSA code and most things related to it. In a
9160 few places, most notable in the depth of the ASN.1 code, ugly
9161 casts back to non-const were required (to be solved at a later
9165 *) Make it so the openssl application has all engines loaded by default.
9168 *) Constify the BIGNUM routines a little more.
9171 *) Add the following functions:
9173 ENGINE_load_cswift()
9175 ENGINE_load_atalla()
9177 ENGINE_load_builtin_engines()
9179 That way, an application can itself choose if external engines that
9180 are built-in in OpenSSL shall ever be used or not. The benefit is
9181 that applications won't have to be linked with libdl or other dso
9182 libraries unless it's really needed.
9184 Changed 'openssl engine' to load all engines on demand.
9185 Changed the engine header files to avoid the duplication of some
9186 declarations (they differed!).
9189 *) 'openssl engine' can now list capabilities.
9192 *) Better error reporting in 'openssl engine'.
9195 *) Never call load_dh_param(NULL) in s_server.
9198 *) Add engine application. It can currently list engines by name and
9199 identity, and test if they are actually available.
9202 *) Improve RPM specification file by forcing symbolic linking and making
9203 sure the installed documentation is also owned by root.root.
9204 [Damien Miller <djm@mindrot.org>]
9206 *) Give the OpenSSL applications more possibilities to make use of
9207 keys (public as well as private) handled by engines.
9210 *) Add OCSP code that comes from CertCo.
9213 *) Add VMS support for the Rijndael code.
9216 *) Added untested support for Nuron crypto accelerator.
9219 *) Add support for external cryptographic devices. This code was
9220 previously distributed separately as the "engine" branch.
9221 [Geoff Thorpe, Richard Levitte]
9223 *) Rework the filename-translation in the DSO code. It is now possible to
9224 have far greater control over how a "name" is turned into a filename
9225 depending on the operating environment and any oddities about the
9226 different shared library filenames on each system.
9229 *) Support threads on FreeBSD-elf in Configure.
9232 *) Fix for SHA1 assembly problem with MASM: it produces
9233 warnings about corrupt line number information when assembling
9234 with debugging information. This is caused by the overlapping
9236 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
9239 NCONF_get_number() has no error checking at all. As a replacement,
9240 NCONF_get_number_e() is defined (_e for "error checking") and is
9241 promoted strongly. The old NCONF_get_number is kept around for
9242 binary backward compatibility.
9243 Make it possible for methods to load from something other than a BIO,
9244 by providing a function pointer that is given a name instead of a BIO.
9245 For example, this could be used to load configuration data from an
9249 *) Fix for non blocking accept BIOs. Added new I/O special reason
9250 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
9251 with non blocking I/O was not possible because no retry code was
9252 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
9256 *) Added the beginnings of Rijndael support.
9259 *) Fix for bug in DirectoryString mask setting. Add support for
9260 X509_NAME_print_ex() in 'req' and X509_print_ex() function
9261 to allow certificate printing to more controllable, additional
9262 'certopt' option to 'x509' to allow new printing options to be
9266 *) Clean old EAY MD5 hack from e_os.h.
9269 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
9271 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
9272 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
9273 [Joe Orton, Steve Henson]
9275 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
9277 *) Fix additional bug revealed by the NISCC test suite:
9279 Stop bug triggering large recursion when presented with
9280 certain ASN.1 tags (CVE-2003-0851)
9283 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
9285 *) Fix various bugs revealed by running the NISCC test suite:
9287 Stop out of bounds reads in the ASN1 code when presented with
9288 invalid tags (CVE-2003-0543 and CVE-2003-0544).
9290 If verify callback ignores invalid public key errors don't try to check
9291 certificate signature with the NULL public key.
9295 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
9296 if the server requested one: as stated in TLS 1.0 and SSL 3.0
9300 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
9301 extra data after the compression methods not only for TLS 1.0
9302 but also for SSL 3.0 (as required by the specification).
9303 [Bodo Moeller; problem pointed out by Matthias Loepfe]
9305 *) Change X509_certificate_type() to mark the key as exported/exportable
9306 when it's 512 *bits* long, not 512 bytes.
9309 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
9311 *) Countermeasure against the Klima-Pokorny-Rosa extension of
9312 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
9313 a protocol version number mismatch like a decryption error
9314 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
9317 *) Turn on RSA blinding by default in the default implementation
9318 to avoid a timing attack. Applications that don't want it can call
9319 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
9320 They would be ill-advised to do so in most cases.
9321 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
9323 *) Change RSA blinding code so that it works when the PRNG is not
9324 seeded (in this case, the secret RSA exponent is abused as
9325 an unpredictable seed -- if it is not unpredictable, there
9326 is no point in blinding anyway). Make RSA blinding thread-safe
9327 by remembering the creator's thread ID in rsa->blinding and
9328 having all other threads use local one-time blinding factors
9329 (this requires more computation than sharing rsa->blinding, but
9330 avoids excessive locking; and if an RSA object is not shared
9331 between threads, blinding will still be very fast).
9334 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
9336 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
9337 via timing by performing a MAC computation even if incorrect
9338 block cipher padding has been found. This is a countermeasure
9339 against active attacks where the attacker has to distinguish
9340 between bad padding and a MAC verification error. (CVE-2003-0078)
9342 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
9343 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
9344 Martin Vuagnoux (EPFL, Ilion)]
9346 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
9348 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
9349 memory from its contents. This is done with a counter that will
9350 place alternating values in each byte. This can be used to solve
9351 two issues: 1) the removal of calls to memset() by highly optimizing
9352 compilers, and 2) cleansing with other values than 0, since those can
9353 be read through on certain media, for example a swap space on disk.
9356 *) Bugfix: client side session caching did not work with external caching,
9357 because the session->cipher setting was not restored when reloading
9358 from the external cache. This problem was masked, when
9359 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9360 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9363 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9364 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9365 [Zeev Lieber <zeev-l@yahoo.com>]
9367 *) Undo an undocumented change introduced in 0.9.6e which caused
9368 repeated calls to OpenSSL_add_all_ciphers() and
9369 OpenSSL_add_all_digests() to be ignored, even after calling
9373 *) Change the default configuration reader to deal with last line not
9374 being properly terminated.
9377 *) Change X509_NAME_cmp() so it applies the special rules on handling
9378 DN values that are of type PrintableString, as well as RDNs of type
9379 emailAddress where the value has the type ia5String.
9380 [stefank@valicert.com via Richard Levitte]
9382 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9383 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9384 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9385 the bitwise-OR of the two for use by the majority of applications
9386 wanting this behaviour, and update the docs. The documented
9387 behaviour and actual behaviour were inconsistent and had been
9388 changing anyway, so this is more a bug-fix than a behavioural
9390 [Geoff Thorpe, diagnosed by Nadav Har'El]
9392 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9393 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9396 *) Fix initialization code race conditions in
9397 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9398 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9399 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9400 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9401 ssl2_get_cipher_by_char(),
9402 ssl3_get_cipher_by_char().
9403 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9405 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9406 the cached sessions are flushed, as the remove_cb() might use ex_data
9407 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9408 (see [openssl.org #212]).
9409 [Geoff Thorpe, Lutz Jaenicke]
9411 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9412 length, instead of the encoding length to d2i_ASN1_OBJECT.
9415 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9417 *) [In 0.9.6g-engine release:]
9418 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9419 [Lynn Gazis <lgazis@rainbow.com>]
9421 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9423 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9424 and get fix the header length calculation.
9425 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9426 Alon Kantor <alonk@checkpoint.com> (and others),
9429 *) Use proper error handling instead of 'assertions' in buffer
9430 overflow checks added in 0.9.6e. This prevents DoS (the
9431 assertions could call abort()).
9432 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9434 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9436 *) Add various sanity checks to asn1_get_length() to reject
9437 the ASN1 length bytes if they exceed sizeof(long), will appear
9438 negative or the content length exceeds the length of the
9440 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9442 *) Fix cipher selection routines: ciphers without encryption had no flags
9443 for the cipher strength set and where therefore not handled correctly
9444 by the selection routines (PR #130).
9447 *) Fix EVP_dsa_sha macro.
9451 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9452 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9453 that was added in OpenSSL 0.9.6d.
9455 As the countermeasure turned out to be incompatible with some
9456 broken SSL implementations, the new option is part of SSL_OP_ALL.
9457 SSL_OP_ALL is usually employed when compatibility with weird SSL
9458 implementations is desired (e.g. '-bugs' option to 's_client' and
9459 's_server'), so the new option is automatically set in many
9463 *) Changes in security patch:
9465 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9466 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9467 Air Force Materiel Command, USAF, under agreement number
9470 *) Add various sanity checks to asn1_get_length() to reject
9471 the ASN1 length bytes if they exceed sizeof(long), will appear
9472 negative or the content length exceeds the length of the
9473 supplied buffer. (CVE-2002-0659)
9474 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9476 *) Assertions for various potential buffer overflows, not known to
9478 [Ben Laurie (CHATS)]
9480 *) Various temporary buffers to hold ASCII versions of integers were
9481 too small for 64 bit platforms. (CVE-2002-0655)
9482 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9484 *) Remote buffer overflow in SSL3 protocol - an attacker could
9485 supply an oversized session ID to a client. (CVE-2002-0656)
9486 [Ben Laurie (CHATS)]
9488 *) Remote buffer overflow in SSL2 protocol - an attacker could
9489 supply an oversized client master key. (CVE-2002-0656)
9490 [Ben Laurie (CHATS)]
9492 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9494 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9495 encoded as NULL) with id-dsa-with-sha1.
9496 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9498 *) Check various X509_...() return values in apps/req.c.
9499 [Nils Larsch <nla@trustcenter.de>]
9501 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9502 an end-of-file condition would erroneously be flagged, when the CRLF
9503 was just at the end of a processed block. The bug was discovered when
9504 processing data through a buffering memory BIO handing the data to a
9505 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9506 <ptsekov@syntrex.com> and Nedelcho Stanev.
9509 *) Implement a countermeasure against a vulnerability recently found
9510 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9511 before application data chunks to avoid the use of known IVs
9512 with data potentially chosen by the attacker.
9515 *) Fix length checks in ssl3_get_client_hello().
9518 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9519 to prevent ssl3_read_internal() from incorrectly assuming that
9520 ssl3_read_bytes() found application data while handshake
9521 processing was enabled when in fact s->s3->in_read_app_data was
9522 merely automatically cleared during the initial handshake.
9523 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9525 *) Fix object definitions for Private and Enterprise: they were not
9526 recognized in their shortname (=lowercase) representation. Extend
9527 obj_dat.pl to issue an error when using undefined keywords instead
9528 of silently ignoring the problem (Svenning Sorensen
9529 <sss@sss.dnsalias.net>).
9532 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9533 generators, i.e. generators other than 2 and 5. (Previously, the
9534 code did not properly initialise the 'add' and 'rem' values to
9535 BN_generate_prime().)
9537 In the new general case, we do not insist that 'generator' is
9538 actually a primitive root: This requirement is rather pointless;
9539 a generator of the order-q subgroup is just as good, if not
9543 *) Map new X509 verification errors to alerts. Discovered and submitted by
9544 Tom Wu <tom@arcot.com>.
9547 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9548 returning non-zero before the data has been completely received
9549 when using non-blocking I/O.
9550 [Bodo Moeller; problem pointed out by John Hughes]
9552 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9553 [Ben Laurie, Lutz Jaenicke]
9555 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9556 Yoram Zahavi <YoramZ@gilian.com>).
9559 *) Add information about CygWin 1.3 and on, and preserve proper
9560 configuration for the versions before that.
9561 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9563 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9564 check whether we deal with a copy of a session and do not delete from
9565 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9566 <izhar@checkpoint.com>.
9569 *) Do not store session data into the internal session cache, if it
9570 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9571 flag is set). Proposed by Aslam <aslam@funk.com>.
9574 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9578 *) [In 0.9.6d-engine release:]
9579 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9580 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9582 *) Add the configuration target linux-s390x.
9583 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9585 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9586 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9587 variable as an indication that a ClientHello message has been
9588 received. As the flag value will be lost between multiple
9589 invocations of ssl3_accept when using non-blocking I/O, the
9590 function may not be aware that a handshake has actually taken
9591 place, thus preventing a new session from being added to the
9594 To avoid this problem, we now set s->new_session to 2 instead of
9595 using a local variable.
9596 [Lutz Jaenicke, Bodo Moeller]
9598 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9599 if the SSL_R_LENGTH_MISMATCH error is detected.
9600 [Geoff Thorpe, Bodo Moeller]
9602 *) New 'shared_ldflag' column in Configure platform table.
9605 *) Fix EVP_CIPHER_mode macro.
9606 ["Dan S. Camper" <dan@bti.net>]
9608 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9609 type, we must throw them away by setting rr->length to 0.
9610 [D P Chang <dpc@qualys.com>]
9612 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9614 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9615 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9616 worked incorrectly for those cases where range = 10..._2 and
9617 3*range is two bits longer than range.)
9620 *) Only add signing time to PKCS7 structures if it is not already
9624 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9625 OBJ_ld_ce should be OBJ_id_ce.
9626 Also some ip-pda OIDs in crypto/objects/objects.txt were
9627 incorrect (cf. RFC 3039).
9628 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9630 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9631 returns early because it has nothing to do.
9632 [Andy Schneider <andy.schneider@bjss.co.uk>]
9634 *) [In 0.9.6c-engine release:]
9635 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9636 [Andy Schneider <andy.schneider@bjss.co.uk>]
9638 *) [In 0.9.6c-engine release:]
9639 Add support for Cryptographic Appliance's keyserver technology.
9640 (Use engine 'keyclient')
9641 [Cryptographic Appliances and Geoff Thorpe]
9643 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9644 is called via tools/c89.sh because arguments have to be
9645 rearranged (all '-L' options must appear before the first object
9647 [Richard Shapiro <rshapiro@abinitio.com>]
9649 *) [In 0.9.6c-engine release:]
9650 Add support for Broadcom crypto accelerator cards, backported
9652 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9654 *) [In 0.9.6c-engine release:]
9655 Add support for SureWare crypto accelerator cards from
9656 Baltimore Technologies. (Use engine 'sureware')
9657 [Baltimore Technologies and Mark Cox]
9659 *) [In 0.9.6c-engine release:]
9660 Add support for crypto accelerator cards from Accelerated
9661 Encryption Processing, www.aep.ie. (Use engine 'aep')
9662 [AEP Inc. and Mark Cox]
9664 *) Add a configuration entry for gcc on UnixWare.
9665 [Gary Benson <gbenson@redhat.com>]
9667 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9668 messages are stored in a single piece (fixed-length part and
9669 variable-length part combined) and fix various bugs found on the way.
9672 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9673 instead. BIO_gethostbyname() does not know what timeouts are
9674 appropriate, so entries would stay in cache even when they have
9676 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9678 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9679 faced with a pathologically small ClientHello fragment that does
9680 not contain client_version: Instead of aborting with an error,
9681 simply choose the highest available protocol version (i.e.,
9682 TLS 1.0 unless it is disabled). In practice, ClientHello
9683 messages are never sent like this, but this change gives us
9684 strictly correct behaviour at least for TLS.
9687 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9688 never resets s->method to s->ctx->method when called from within
9689 one of the SSL handshake functions.
9690 [Bodo Moeller; problem pointed out by Niko Baric]
9692 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9693 (sent using the client's version number) if client_version is
9694 smaller than the protocol version in use. Also change
9695 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9696 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9697 the client will at least see that alert.
9700 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9704 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9705 client receives HelloRequest while in a handshake.
9706 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9708 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9709 should end in 'break', not 'goto end' which circumvents various
9710 cleanups done in state SSL_ST_OK. But session related stuff
9711 must be disabled for SSL_ST_OK in the case that we just sent a
9714 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9715 before just sending a HelloRequest.
9716 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9718 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9719 reveal whether illegal block cipher padding was found or a MAC
9720 verification error occurred. (Neither SSLerr() codes nor alerts
9721 are directly visible to potential attackers, but the information
9722 may leak via logfiles.)
9724 Similar changes are not required for the SSL 2.0 implementation
9725 because the number of padding bytes is sent in clear for SSL 2.0,
9726 and the extra bytes are just ignored. However ssl/s2_pkt.c
9727 failed to verify that the purported number of padding bytes is in
9731 *) Add OpenUNIX-8 support including shared libraries
9732 (Boyd Lynn Gerber <gerberb@zenez.com>).
9735 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9736 'wristwatch attack' using huge encoding parameters (cf.
9737 James H. Manger's CRYPTO 2001 paper). Note that the
9738 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9739 encoding parameters and hence was not vulnerable.
9742 *) BN_sqr() bug fix.
9743 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9745 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9746 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9747 followed by modular reduction.
9748 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9750 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9751 equivalent based on BN_pseudo_rand() instead of BN_rand().
9754 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9755 This function was broken, as the check for a new client hello message
9756 to handle SGC did not allow these large messages.
9757 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9760 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9763 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9764 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9767 *) Rework the configuration and shared library support for Tru64 Unix.
9768 The configuration part makes use of modern compiler features and
9769 still retains old compiler behavior for those that run older versions
9770 of the OS. The shared library support part includes a variant that
9771 uses the RPATH feature, and is available through the special
9772 configuration target "alpha-cc-rpath", which will never be selected
9774 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9776 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9777 with the same message size as in ssl3_get_certificate_request().
9778 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9779 messages might inadvertently be reject as too long.
9780 [Petr Lampa <lampa@fee.vutbr.cz>]
9782 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9785 *) Modified SSL library such that the verify_callback that has been set
9786 specifically for an SSL object with SSL_set_verify() is actually being
9787 used. Before the change, a verify_callback set with this function was
9788 ignored and the verify_callback() set in the SSL_CTX at the time of
9789 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9790 to allow the necessary settings.
9793 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9794 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9795 done automatically (in contradiction to the requirements of the C
9796 standard). This made problems when used from OpenSSH.
9799 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9800 dh->length and always used
9802 BN_rand_range(priv_key, dh->p).
9804 BN_rand_range() is not necessary for Diffie-Hellman, and this
9805 specific range makes Diffie-Hellman unnecessarily inefficient if
9806 dh->length (recommended exponent length) is much smaller than the
9807 length of dh->p. We could use BN_rand_range() if the order of
9808 the subgroup was stored in the DH structure, but we only have
9813 BN_rand(priv_key, l, ...)
9815 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9821 RSA_eay_public_encrypt
9822 RSA_eay_private_decrypt
9823 RSA_eay_private_encrypt (signing)
9824 RSA_eay_public_decrypt (signature verification)
9826 (default implementations for RSA_public_encrypt,
9827 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9828 always reject numbers >= n.
9831 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9832 to synchronize access to 'locking_thread'. This is necessary on
9833 systems where access to 'locking_thread' (an 'unsigned long'
9834 variable) is not atomic.
9837 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9838 *before* setting the 'crypto_lock_rand' flag. The previous code had
9839 a race condition if 0 is a valid thread ID.
9840 [Travis Vitek <vitek@roguewave.com>]
9842 *) Add support for shared libraries under Irix.
9843 [Albert Chin-A-Young <china@thewrittenword.com>]
9845 *) Add configuration option to build on Linux on both big-endian and
9847 [Ralf Baechle <ralf@uni-koblenz.de>]
9849 *) Add the possibility to create shared libraries on HP-UX.
9852 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9854 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9855 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9856 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9857 PRNG state recovery was possible based on the output of
9858 one PRNG request appropriately sized to gain knowledge on
9859 'md' followed by enough consecutive 1-byte PRNG requests
9860 to traverse all of 'state'.
9862 1. When updating 'md_local' (the current thread's copy of 'md')
9863 during PRNG output generation, hash all of the previous
9864 'md_local' value, not just the half used for PRNG output.
9866 2. Make the number of bytes from 'state' included into the hash
9867 independent from the number of PRNG bytes requested.
9869 The first measure alone would be sufficient to avoid
9870 Markku-Juhani's attack. (Actually it had never occurred
9871 to me that the half of 'md_local' used for chaining was the
9872 half from which PRNG output bytes were taken -- I had always
9873 assumed that the secret half would be used.) The second
9874 measure makes sure that additional data from 'state' is never
9875 mixed into 'md_local' in small portions; this heuristically
9876 further strengthens the PRNG.
9879 *) Fix crypto/bn/asm/mips3.s.
9882 *) When only the key is given to "enc", the IV is undefined. Print out
9883 an error message in this case.
9886 *) Handle special case when X509_NAME is empty in X509 printing routines.
9889 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9890 positive and less than q.
9893 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9894 used: it isn't thread safe and the add_lock_callback should handle
9896 [Paul Rose <Paul.Rose@bridge.com>]
9898 *) Verify that incoming data obeys the block size in
9899 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9903 [Ulf Möller, Bodo Möller]
9905 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9906 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9907 when fixing the server behaviour for backwards-compatible 'client
9908 hello' messages. (Note that the attack is impractical against
9909 SSL 3.0 and TLS 1.0 anyway because length and version checking
9910 means that the probability of guessing a valid ciphertext is
9911 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9914 Before 0.9.5, the countermeasure (hide the error by generating a
9915 random 'decryption result') did not work properly because
9916 ERR_clear_error() was missing, meaning that SSL_get_error() would
9917 detect the supposedly ignored error.
9919 Both problems are now fixed.
9922 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9923 (previously it was 1024).
9926 *) Fix for compatibility mode trust settings: ignore trust settings
9927 unless some valid trust or reject settings are present.
9930 *) Fix for blowfish EVP: its a variable length cipher.
9933 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9934 parameters in DSA public key structures and return an error in the
9935 DSA routines if parameters are absent.
9938 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9939 in the current directory if neither $RANDFILE nor $HOME was set.
9940 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9941 caused some confusion to Windows users who haven't defined $HOME.
9942 Thus RAND_file_name() is changed again: e_os.h can define a
9943 DEFAULT_HOME, which will be used if $HOME is not set.
9944 For Windows, we use "C:"; on other platforms, we still require
9945 environment variables.
9947 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9948 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9949 having multiple threads call RAND_poll() concurrently.
9952 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9953 combination of a flag and a thread ID variable.
9954 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9955 flag), *other* threads can enter ssleay_add_bytes without obeying
9956 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9957 that they do not hold after the first thread unsets add_do_not_lock).
9960 *) Change bctest again: '-x' expressions are not available in all
9964 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9966 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9967 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9969 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9970 the default extension for executables, if any. Also, make the perl
9971 scripts that use symlink() to test if it really exists and use "cp"
9972 if it doesn't. All this made OpenSSL compilable and installable in
9976 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9977 If SEQUENCE is length is indefinite just set c->slen to the total
9978 amount of data available.
9979 [Steve Henson, reported by shige@FreeBSD.org]
9980 [This change does not apply to 0.9.7.]
9982 *) Change bctest to avoid here-documents inside command substitution
9983 (workaround for FreeBSD /bin/sh bug).
9984 For compatibility with Ultrix, avoid shell functions (introduced
9985 in the bctest version that searches along $PATH).
9988 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9989 with des_encrypt() defined on some operating systems, like Solaris
9993 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9994 On the Importance of Eliminating Errors in Cryptographic
9995 Computations, J. Cryptology 14 (2001) 2, 101-119,
9996 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9999 *) MIPS assembler BIGNUM division bug fix.
10002 *) Disabled incorrect Alpha assembler code.
10005 *) Fix PKCS#7 decode routines so they correctly update the length
10006 after reading an EOC for the EXPLICIT tag.
10008 [This change does not apply to 0.9.7.]
10010 *) Fix bug in PKCS#12 key generation routines. This was triggered
10011 if a 3DES key was generated with a 0 initial byte. Include
10012 PKCS12_BROKEN_KEYGEN compilation option to retain the old
10013 (but broken) behaviour.
10016 *) Enhance bctest to search for a working bc along $PATH and print
10018 [Tim Rice <tim@multitalents.net> via Richard Levitte]
10020 *) Fix memory leaks in err.c: free err_data string if necessary;
10021 don't write to the wrong index in ERR_set_error_data.
10024 *) Implement ssl23_peek (analogous to ssl23_read), which previously
10028 *) Replace rdtsc with _emit statements for VC++ version 5.
10029 [Jeremy Cooper <jeremy@baymoo.org>]
10031 *) Make it possible to reuse SSLv2 sessions.
10034 *) In copy_email() check for >= 0 as a return value for
10035 X509_NAME_get_index_by_NID() since 0 is a valid index.
10036 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
10038 *) Avoid coredump with unsupported or invalid public keys by checking if
10039 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
10040 PKCS7_verify() fails with non detached data.
10043 *) Don't use getenv in library functions when run as setuid/setgid.
10044 New function OPENSSL_issetugid().
10047 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
10048 due to incorrect handling of multi-threading:
10050 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
10052 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
10054 3. Count how many times MemCheck_off() has been called so that
10055 nested use can be treated correctly. This also avoids
10056 inband-signalling in the previous code (which relied on the
10057 assumption that thread ID 0 is impossible).
10060 *) Add "-rand" option also to s_client and s_server.
10063 *) Fix CPU detection on Irix 6.x.
10064 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
10065 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10067 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
10070 [This change does not apply to 0.9.7.]
10072 *) Use the cached encoding of an X509_NAME structure rather than
10073 copying it. This is apparently the reason for the libsafe "errors"
10074 but the code is actually correct.
10077 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
10078 Bleichenbacher's DSA attack.
10079 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
10080 to be set and top=0 forces the highest bit to be set; top=-1 is new
10081 and leaves the highest bit random.
10082 [Ulf Moeller, Bodo Moeller]
10084 *) In the NCONF_...-based implementations for CONF_... queries
10085 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
10086 a temporary CONF structure with the data component set to NULL
10087 (which gives segmentation faults in lh_retrieve).
10088 Instead, use NULL for the CONF pointer in CONF_get_string and
10089 CONF_get_number (which may use environment variables) and directly
10090 return NULL from CONF_get_section.
10093 *) Fix potential buffer overrun for EBCDIC.
10096 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
10097 keyUsage if basicConstraints absent for a CA.
10100 *) Make SMIME_write_PKCS7() write mail header values with a format that
10101 is more generally accepted (no spaces before the semicolon), since
10102 some programs can't parse those values properly otherwise. Also make
10103 sure BIO's that break lines after each write do not create invalid
10107 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
10108 macros previously used would not encode an empty SEQUENCE OF
10109 and break the signature.
10111 [This change does not apply to 0.9.7.]
10113 *) Zero the premaster secret after deriving the master secret in
10117 *) Add some EVP_add_digest_alias registrations (as found in
10118 OpenSSL_add_all_digests()) to SSL_library_init()
10119 aka OpenSSL_add_ssl_algorithms(). This provides improved
10120 compatibility with peers using X.509 certificates
10121 with unconventional AlgorithmIdentifier OIDs.
10124 *) Fix for Irix with NO_ASM.
10125 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10127 *) ./config script fixes.
10128 [Ulf Moeller, Richard Levitte]
10130 *) Fix 'openssl passwd -1'.
10133 *) Change PKCS12_key_gen_asc() so it can cope with non null
10134 terminated strings whose length is passed in the passlen
10135 parameter, for example from PEM callbacks. This was done
10136 by adding an extra length parameter to asc2uni().
10137 [Steve Henson, reported by <oddissey@samsung.co.kr>]
10139 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
10140 call failed, free the DSA structure.
10143 *) Fix to uni2asc() to cope with zero length Unicode strings.
10144 These are present in some PKCS#12 files.
10147 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
10148 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
10149 when writing a 32767 byte record.
10150 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
10152 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
10153 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
10155 (RSA objects have a reference count access to which is protected
10156 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
10157 so they are meant to be shared between threads.)
10158 [Bodo Moeller, Geoff Thorpe; original patch submitted by
10159 "Reddie, Steven" <Steven.Reddie@ca.com>]
10161 *) Fix a deadlock in CRYPTO_mem_leaks().
10164 *) Use better test patterns in bntest.
10167 *) rand_win.c fix for Borland C.
10170 *) BN_rshift bugfix for n == 0.
10173 *) Add a 'bctest' script that checks for some known 'bc' bugs
10174 so that 'make test' does not abort just because 'bc' is broken.
10177 *) Store verify_result within SSL_SESSION also for client side to
10178 avoid potential security hole. (Re-used sessions on the client side
10179 always resulted in verify_result==X509_V_OK, not using the original
10180 result of the server certificate verification.)
10183 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
10184 SSL3_RT_APPLICATION_DATA, return 0.
10185 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
10189 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
10190 releases, have been re-implemented by renaming the previous
10191 implementations of ssl2_read and ssl3_read to ssl2_read_internal
10192 and ssl3_read_internal, respectively, and adding 'peek' parameters
10193 to them. The new ssl[23]_{read,peek} functions are calls to
10194 ssl[23]_read_internal with the 'peek' flag set appropriately.
10195 A 'peek' parameter has also been added to ssl3_read_bytes, which
10196 does the actual work for ssl3_read_internal.
10199 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
10200 the method-specific "init()" handler. Also clean up ex_data after
10201 calling the method-specific "finish()" handler. Previously, this was
10202 happening the other way round.
10205 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
10206 The previous value, 12, was not always sufficient for BN_mod_exp().
10209 *) Make sure that shared libraries get the internal name engine with
10210 the full version number and not just 0. This should mark the
10211 shared libraries as not backward compatible. Of course, this should
10212 be changed again when we can guarantee backward binary compatibility.
10215 *) Fix typo in get_cert_by_subject() in by_dir.c
10216 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
10218 *) Rework the system to generate shared libraries:
10220 - Make note of the expected extension for the shared libraries and
10221 if there is a need for symbolic links from for example libcrypto.so.0
10222 to libcrypto.so.0.9.7. There is extended info in Configure for
10225 - Make as few rebuilds of the shared libraries as possible.
10227 - Still avoid linking the OpenSSL programs with the shared libraries.
10229 - When installing, install the shared libraries separately from the
10233 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
10235 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
10236 and not in SSL_clear because the latter is also used by the
10237 accept/connect functions; previously, the settings made by
10238 SSL_set_read_ahead would be lost during the handshake.
10239 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
10241 *) Correct util/mkdef.pl to be selective about disabled algorithms.
10242 Previously, it would create entries for disabled algorithms no
10246 *) Added several new manual pages for SSL_* function.
10249 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
10251 *) In ssl23_get_client_hello, generate an error message when faced
10252 with an initial SSL 3.0/TLS record that is too small to contain the
10253 first two bytes of the ClientHello message, i.e. client_version.
10254 (Note that this is a pathologic case that probably has never happened
10255 in real life.) The previous approach was to use the version number
10256 from the record header as a substitute; but our protocol choice
10257 should not depend on that one because it is not authenticated
10258 by the Finished messages.
10261 *) More robust randomness gathering functions for Windows.
10262 [Jeffrey Altman <jaltman@columbia.edu>]
10264 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
10265 not set then we don't setup the error code for issuer check errors
10266 to avoid possibly overwriting other errors which the callback does
10267 handle. If an application does set the flag then we assume it knows
10268 what it is doing and can handle the new informational codes
10272 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
10273 a general "ANY" type, as such it should be able to decode anything
10274 including tagged types. However it didn't check the class so it would
10275 wrongly interpret tagged types in the same way as their universal
10276 counterpart and unknown types were just rejected. Changed so that the
10277 tagged and unknown types are handled in the same way as a SEQUENCE:
10278 that is the encoding is stored intact. There is also a new type
10279 "V_ASN1_OTHER" which is used when the class is not universal, in this
10280 case we have no idea what the actual type is so we just lump them all
10284 *) On VMS, stdout may very well lead to a file that is written to
10285 in a record-oriented fashion. That means that every write() will
10286 write a separate record, which will be read separately by the
10287 programs trying to read from it. This can be very confusing.
10289 The solution is to put a BIO filter in the way that will buffer
10290 text until a linefeed is reached, and then write everything a
10291 line at a time, so every record written will be an actual line,
10292 not chunks of lines and not (usually doesn't happen, but I've
10293 seen it once) several lines in one record. BIO_f_linebuffer() is
10296 Currently, it's a VMS-only method, because that's where it has
10297 been tested well enough.
10300 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
10301 it can return incorrect results.
10302 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
10303 but it was in 0.9.6-beta[12].)
10306 *) Disable the check for content being present when verifying detached
10307 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
10308 include zero length content when signing messages.
10311 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
10312 BIO_ctrl (for BIO pairs).
10315 *) Add DSO method for VMS.
10318 *) Bug fix: Montgomery multiplication could produce results with the
10322 *) Add RPM specification openssl.spec and modify it to build three
10323 packages. The default package contains applications, application
10324 documentation and run-time libraries. The devel package contains
10325 include files, static libraries and function documentation. The
10326 doc package contains the contents of the doc directory. The original
10327 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
10330 *) Add a large number of documentation files for many SSL routines.
10331 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10333 *) Add a configuration entry for Sony News 4.
10334 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
10336 *) Don't set the two most significant bits to one when generating a
10337 random number < q in the DSA library.
10340 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
10341 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
10342 the underlying transport is blocking) if a handshake took place.
10343 (The default behaviour is needed by applications such as s_client
10344 and s_server that use select() to determine when to use SSL_read;
10345 but for applications that know in advance when to expect data, it
10346 just makes things more complicated.)
10349 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
10353 *) Add a few more EBCDIC conditionals that make `req' and `x509'
10354 work better on such systems.
10355 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10357 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10358 Update PKCS12_parse() so it copies the friendlyName and the
10359 keyid to the certificates aux info.
10362 *) Fix bug in PKCS7_verify() which caused an infinite loop
10363 if there was more than one signature.
10364 [Sven Uszpelkat <su@celocom.de>]
10366 *) Major change in util/mkdef.pl to include extra information
10367 about each symbol, as well as presenting variables as well
10368 as functions. This change means that there's n more need
10369 to rebuild the .num files when some algorithms are excluded.
10372 *) Allow the verify time to be set by an application,
10373 rather than always using the current time.
10376 *) Phase 2 verify code reorganisation. The certificate
10377 verify code now looks up an issuer certificate by a
10378 number of criteria: subject name, authority key id
10379 and key usage. It also verifies self signed certificates
10380 by the same criteria. The main comparison function is
10381 X509_check_issued() which performs these checks.
10383 Lot of changes were necessary in order to support this
10384 without completely rewriting the lookup code.
10386 Authority and subject key identifier are now cached.
10388 The LHASH 'certs' is X509_STORE has now been replaced
10389 by a STACK_OF(X509_OBJECT). This is mainly because an
10390 LHASH can't store or retrieve multiple objects with
10391 the same hash value.
10393 As a result various functions (which were all internal
10394 use only) have changed to handle the new X509_STORE
10395 structure. This will break anything that messed round
10396 with X509_STORE internally.
10398 The functions X509_STORE_add_cert() now checks for an
10399 exact match, rather than just subject name.
10401 The X509_STORE API doesn't directly support the retrieval
10402 of multiple certificates matching a given criteria, however
10403 this can be worked round by performing a lookup first
10404 (which will fill the cache with candidate certificates)
10405 and then examining the cache for matches. This is probably
10406 the best we can do without throwing out X509_LOOKUP
10407 entirely (maybe later...).
10409 The X509_VERIFY_CTX structure has been enhanced considerably.
10411 All certificate lookup operations now go via a get_issuer()
10412 callback. Although this currently uses an X509_STORE it
10413 can be replaced by custom lookups. This is a simple way
10414 to bypass the X509_STORE hackery necessary to make this
10415 work and makes it possible to use more efficient techniques
10416 in future. A very simple version which uses a simple
10417 STACK for its trusted certificate store is also provided
10418 using X509_STORE_CTX_trusted_stack().
10420 The verify_cb() and verify() callbacks now have equivalents
10421 in the X509_STORE_CTX structure.
10423 X509_STORE_CTX also has a 'flags' field which can be used
10424 to customise the verify behaviour.
10427 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10428 excludes S/MIME capabilities.
10431 *) When a certificate request is read in keep a copy of the
10432 original encoding of the signed data and use it when outputting
10433 again. Signatures then use the original encoding rather than
10434 a decoded, encoded version which may cause problems if the
10435 request is improperly encoded.
10438 *) For consistency with other BIO_puts implementations, call
10439 buffer_write(b, ...) directly in buffer_puts instead of calling
10442 In BIO_puts, increment b->num_write as in BIO_write.
10443 [Peter.Sylvester@EdelWeb.fr]
10445 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10446 BN_zero, we may not return a BIGNUM with an array consisting of
10447 words set to zero.)
10450 *) Avoid calling abort() from within the library when problems are
10451 detected, except if preprocessor symbols have been defined
10452 (such as REF_CHECK, BN_DEBUG etc.).
10455 *) New openssl application 'rsautl'. This utility can be
10456 used for low level RSA operations. DER public key
10457 BIO/fp routines also added.
10460 *) New Configure entry and patches for compiling on QNX 4.
10461 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10463 *) A demo state-machine implementation was sponsored by
10464 Nuron (http://www.nuron.com/) and is now available in
10465 demos/state_machine.
10468 *) New options added to the 'dgst' utility for signature
10469 generation and verification.
10472 *) Unrecognized PKCS#7 content types are now handled via a
10473 catch all ASN1_TYPE structure. This allows unsupported
10474 types to be stored as a "blob" and an application can
10475 encode and decode it manually.
10478 *) Fix various signed/unsigned issues to make a_strex.c
10479 compile under VC++.
10480 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10482 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10483 length if passed a buffer. ASN1_INTEGER_to_BN failed
10484 if passed a NULL BN and its argument was negative.
10485 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10487 *) Modification to PKCS#7 encoding routines to output definite
10488 length encoding. Since currently the whole structures are in
10489 memory there's not real point in using indefinite length
10490 constructed encoding. However if OpenSSL is compiled with
10491 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10494 *) Added BIO_vprintf() and BIO_vsnprintf().
10497 *) Added more prefixes to parse for in the strings written
10498 through a logging bio, to cover all the levels that are available
10499 through syslog. The prefixes are now:
10501 PANIC, EMERG, EMR => LOG_EMERG
10502 ALERT, ALR => LOG_ALERT
10503 CRIT, CRI => LOG_CRIT
10504 ERROR, ERR => LOG_ERR
10505 WARNING, WARN, WAR => LOG_WARNING
10506 NOTICE, NOTE, NOT => LOG_NOTICE
10507 INFO, INF => LOG_INFO
10508 DEBUG, DBG => LOG_DEBUG
10510 and as before, if none of those prefixes are present at the
10511 beginning of the string, LOG_ERR is chosen.
10513 On Win32, the LOG_* levels are mapped according to this:
10515 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10516 LOG_WARNING => EVENTLOG_WARNING_TYPE
10517 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10521 *) Made it possible to reconfigure with just the configuration
10522 argument "reconf" or "reconfigure". The command line arguments
10523 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10524 and are retrieved from there when reconfiguring.
10527 *) MD4 implemented.
10528 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10530 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10533 *) The obj_dat.pl script was messing up the sorting of object
10534 names. The reason was that it compared the quoted version
10535 of strings as a result "OCSP" > "OCSP Signing" because
10536 " > SPACE. Changed script to store unquoted versions of
10537 names and add quotes on output. It was also omitting some
10538 names from the lookup table if they were given a default
10539 value (that is if SN is missing it is given the same
10540 value as LN and vice versa), these are now added on the
10541 grounds that if an object has a name we should be able to
10542 look it up. Finally added warning output when duplicate
10543 short or long names are found.
10546 *) Changes needed for Tandem NSK.
10547 [Scott Uroff <scott@xypro.com>]
10549 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10550 RSA_padding_check_SSLv23(), special padding was never detected
10551 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10552 version rollback attacks was not effective.
10554 In s23_clnt.c, don't use special rollback-attack detection padding
10555 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10556 client; similarly, in s23_srvr.c, don't do the rollback check if
10557 SSL 2.0 is the only protocol enabled in the server.
10560 *) Make it possible to get hexdumps of unprintable data with 'openssl
10561 asn1parse'. By implication, the functions ASN1_parse_dump() and
10562 BIO_dump_indent() are added.
10565 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10566 these print out strings and name structures based on various
10567 flags including RFC2253 support and proper handling of
10568 multibyte characters. Added options to the 'x509' utility
10569 to allow the various flags to be set.
10572 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10573 Also change the functions X509_cmp_current_time() and
10574 X509_gmtime_adj() work with an ASN1_TIME structure,
10575 this will enable certificates using GeneralizedTime in validity
10576 dates to be checked.
10579 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10580 negative public key encodings) on by default,
10581 NO_NEG_PUBKEY_BUG can be set to disable it.
10584 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10585 content octets. An i2c_ASN1_OBJECT is unnecessary because
10586 the encoding can be trivially obtained from the structure.
10589 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10590 not read locks (CRYPTO_r_[un]lock).
10593 *) A first attempt at creating official support for shared
10594 libraries through configuration. I've kept it so the
10595 default is static libraries only, and the OpenSSL programs
10596 are always statically linked for now, but there are
10597 preparations for dynamic linking in place.
10598 This has been tested on Linux and Tru64.
10601 *) Randomness polling function for Win9x, as described in:
10602 Peter Gutmann, Software Generation of Practically Strong
10606 *) Fix so PRNG is seeded in req if using an already existing
10610 *) New options to smime application. -inform and -outform
10611 allow alternative formats for the S/MIME message including
10612 PEM and DER. The -content option allows the content to be
10613 specified separately. This should allow things like Netscape
10614 form signing output easier to verify.
10617 *) Fix the ASN1 encoding of tags using the 'long form'.
10620 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10621 STRING types. These convert content octets to and from the
10622 underlying type. The actual tag and length octets are
10623 already assumed to have been read in and checked. These
10624 are needed because all other string types have virtually
10625 identical handling apart from the tag. By having versions
10626 of the ASN1 functions that just operate on content octets
10627 IMPLICIT tagging can be handled properly. It also allows
10628 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10629 and ASN1_INTEGER are identical apart from the tag.
10632 *) Change the handling of OID objects as follows:
10634 - New object identifiers are inserted in objects.txt, following
10635 the syntax given in objects.README.
10636 - objects.pl is used to process obj_mac.num and create a new
10638 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10641 This is currently kind of a hack, and the perl code in objects.pl
10642 isn't very elegant, but it works as I intended. The simplest way
10643 to check that it worked correctly is to look in obj_dat.h and
10644 check the array nid_objs and make sure the objects haven't moved
10645 around (this is important!). Additions are OK, as well as
10646 consistent name changes.
10649 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10652 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10653 The given file adds to whatever has already been seeded into the
10654 random pool through the RANDFILE configuration file option or
10655 environment variable, or the default random state file.
10658 *) mkstack.pl now sorts each macro group into lexical order.
10659 Previously the output order depended on the order the files
10660 appeared in the directory, resulting in needless rewriting
10664 *) Patches to make OpenSSL compile under Win32 again. Mostly
10665 work arounds for the VC++ problem that it treats func() as
10666 func(void). Also stripped out the parts of mkdef.pl that
10667 added extra typesafe functions: these no longer exist.
10670 *) Reorganisation of the stack code. The macros are now all
10671 collected in safestack.h . Each macro is defined in terms of
10672 a "stack macro" of the form SKM_<name>(type, a, b). The
10673 DEBUG_SAFESTACK is now handled in terms of function casts,
10674 this has the advantage of retaining type safety without the
10675 use of additional functions. If DEBUG_SAFESTACK is not defined
10676 then the non typesafe macros are used instead. Also modified the
10677 mkstack.pl script to handle the new form. Needs testing to see
10678 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10679 the default if no major problems. Similar behaviour for ASN1_SET_OF
10680 and PKCS12_STACK_OF.
10683 *) When some versions of IIS use the 'NET' form of private key the
10684 key derivation algorithm is different. Normally MD5(password) is
10685 used as a 128 bit RC4 key. In the modified case
10686 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10687 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10688 as the old Netscape_RSA functions except they have an additional
10689 'sgckey' parameter which uses the modified algorithm. Also added
10690 an -sgckey command line option to the rsa utility. Thanks to
10691 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10692 algorithm to openssl-dev.
10695 *) The evp_local.h macros were using 'c.##kname' which resulted in
10696 invalid expansion on some systems (SCO 5.0.5 for example).
10697 Corrected to 'c.kname'.
10698 [Phillip Porch <root@theporch.com>]
10700 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10701 a STACK of email addresses from a certificate or request, these look
10702 in the subject name and the subject alternative name extensions and
10703 omit any duplicate addresses.
10706 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10707 This makes DSA verification about 2 % faster.
10710 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10711 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10712 plus overhead for 1024 bit moduli).
10713 This makes exponentiations about 0.5 % faster for 1024 bit
10714 exponents (as measured by "openssl speed rsa2048").
10717 *) Rename memory handling macros to avoid conflicts with other
10719 Malloc => OPENSSL_malloc
10720 Malloc_locked => OPENSSL_malloc_locked
10721 Realloc => OPENSSL_realloc
10722 Free => OPENSSL_free
10725 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10726 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10729 *) CygWin32 support.
10730 [John Jarvie <jjarvie@newsguy.com>]
10732 *) The type-safe stack code has been rejigged. It is now only compiled
10733 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10734 by default all type-specific stack functions are "#define"d back to
10735 standard stack functions. This results in more streamlined output
10736 but retains the type-safety checking possibilities of the original
10740 *) The STACK code has been cleaned up, and certain type declarations
10741 that didn't make a lot of sense have been brought in line. This has
10742 also involved a cleanup of sorts in safestack.h to more correctly
10743 map type-safe stack functions onto their plain stack counterparts.
10744 This work has also resulted in a variety of "const"ifications of
10745 lots of the code, especially "_cmp" operations which should normally
10746 be prototyped with "const" parameters anyway.
10749 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10750 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10751 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10752 where all of 'md' is used each time the PRNG is used, but 'state'
10753 is used only indexed by a cyclic counter. As entropy may not be
10754 well distributed from the beginning, 'md' is important as a
10755 chaining variable. However, the output function chains only half
10756 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10757 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10758 in all of 'state' being rewritten, with the new values depending
10759 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10762 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10763 the handshake is continued after ssl_verify_cert_chain();
10764 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10765 can lead to 'unexplainable' connection aborts later.
10766 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10768 *) Major EVP API cipher revision.
10769 Add hooks for extra EVP features. This allows various cipher
10770 parameters to be set in the EVP interface. Support added for variable
10771 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10772 setting of RC2 and RC5 parameters.
10774 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10777 Remove lots of duplicated code from the EVP library. For example *every*
10778 cipher init() function handles the 'iv' in the same way according to the
10779 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10780 for CFB and OFB modes they zero ctx->num.
10782 New functionality allows removal of S/MIME code RC2 hack.
10784 Most of the routines have the same form and so can be declared in terms
10787 By shifting this to the top level EVP_CipherInit() it can be removed from
10788 all individual ciphers. If the cipher wants to handle IVs or keys
10789 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10792 Change lots of functions like EVP_EncryptUpdate() to now return a
10793 value: although software versions of the algorithms cannot fail
10794 any installed hardware versions can.
10797 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10798 this option is set, tolerate broken clients that send the negotiated
10799 protocol version number instead of the requested protocol version
10803 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10804 i.e. non-zero for export ciphersuites, zero otherwise.
10805 Previous versions had this flag inverted, inconsistent with
10806 rsa_tmp_cb (..._TMP_RSA_CB).
10807 [Bodo Moeller; problem reported by Amit Chopra]
10809 *) Add missing DSA library text string. Work around for some IIS
10810 key files with invalid SEQUENCE encoding.
10813 *) Add a document (doc/standards.txt) that list all kinds of standards
10814 and so on that are implemented in OpenSSL.
10817 *) Enhance c_rehash script. Old version would mishandle certificates
10818 with the same subject name hash and wouldn't handle CRLs at all.
10819 Added -fingerprint option to crl utility, to support new c_rehash
10823 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10826 *) Fix for SSL server purpose checking. Server checking was
10827 rejecting certificates which had extended key usage present
10828 but no ssl client purpose.
10829 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10831 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10832 is a little unclear about how a blank password is handled.
10833 Since the password in encoded as a BMPString with terminating
10834 double NULL a zero length password would end up as just the
10835 double NULL. However no password at all is different and is
10836 handled differently in the PKCS#12 key generation code. NS
10837 treats a blank password as zero length. MSIE treats it as no
10838 password on export: but it will try both on import. We now do
10839 the same: PKCS12_parse() tries zero length and no password if
10840 the password is set to "" or NULL (NULL is now a valid password:
10841 it wasn't before) as does the pkcs12 application.
10844 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10845 perror when PEM_read_bio_X509_REQ fails, the error message must
10846 be obtained from the error queue.
10849 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10850 it in ERR_remove_state if appropriate, and change ERR_get_state
10851 accordingly to avoid race conditions (this is necessary because
10852 thread_hash is no longer constant once set).
10855 *) Bugfix for linux-elf makefile.one.
10858 *) RSA_get_default_method() will now cause a default
10859 RSA_METHOD to be chosen if one doesn't exist already.
10860 Previously this was only set during a call to RSA_new()
10861 or RSA_new_method(NULL) meaning it was possible for
10862 RSA_get_default_method() to return NULL.
10865 *) Added native name translation to the existing DSO code
10866 that will convert (if the flag to do so is set) filenames
10867 that are sufficiently small and have no path information
10868 into a canonical native form. Eg. "blah" converted to
10869 "libblah.so" or "blah.dll" etc.
10872 *) New function ERR_error_string_n(e, buf, len) which is like
10873 ERR_error_string(e, buf), but writes at most 'len' bytes
10874 including the 0 terminator. For ERR_error_string_n, 'buf'
10876 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10878 *) CONF library reworked to become more general. A new CONF
10879 configuration file reader "class" is implemented as well as a
10880 new functions (NCONF_*, for "New CONF") to handle it. The now
10881 old CONF_* functions are still there, but are reimplemented to
10882 work in terms of the new functions. Also, a set of functions
10883 to handle the internal storage of the configuration data is
10884 provided to make it easier to write new configuration file
10885 reader "classes" (I can definitely see something reading a
10886 configuration file in XML format, for example), called _CONF_*,
10887 or "the configuration storage API"...
10889 The new configuration file reading functions are:
10891 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10892 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10894 NCONF_default, NCONF_WIN32
10896 NCONF_dump_fp, NCONF_dump_bio
10898 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10899 NCONF_new creates a new CONF object. This works in the same way
10900 as other interfaces in OpenSSL, like the BIO interface.
10901 NCONF_dump_* dump the internal storage of the configuration file,
10902 which is useful for debugging. All other functions take the same
10903 arguments as the old CONF_* functions with the exception of the
10904 first that must be a `CONF *' instead of a `LHASH *'.
10906 To make it easier to use the new classes with the old CONF_* functions,
10907 the function CONF_set_default_method is provided.
10910 *) Add '-tls1' option to 'openssl ciphers', which was already
10911 mentioned in the documentation but had not been implemented.
10912 (This option is not yet really useful because even the additional
10913 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10916 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10917 OpenSSL-based applications) load shared libraries and bind to
10918 them in a portable way.
10919 [Geoff Thorpe, with contributions from Richard Levitte]
10921 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10923 *) Make sure _lrotl and _lrotr are only used with MSVC.
10925 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10926 (the default implementation of RAND_status).
10928 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10929 to '-clrext' (= clear extensions), as intended and documented.
10930 [Bodo Moeller; inconsistency pointed out by Michael Attili
10931 <attili@amaxo.com>]
10933 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10934 was larger than the MD block size.
10935 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10937 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10938 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10939 using the passed key: if the passed key was a private key the result
10940 of X509_print(), for example, would be to print out all the private key
10944 *) des_quad_cksum() byte order bug fix.
10945 [Ulf Möller, using the problem description in krb4-0.9.7, where
10946 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10948 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10950 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10952 *) For easily testing in shell scripts whether some command
10953 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10954 returns with exit code 0 iff no command of the given name is available.
10955 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10956 the output goes to stdout and nothing is printed to stderr.
10957 Additional arguments are always ignored.
10959 Since for each cipher there is a command of the same name,
10960 the 'no-cipher' compilation switches can be tested this way.
10962 ('openssl no-XXX' is not able to detect pseudo-commands such
10963 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10966 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10969 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10970 is set; it will be thrown away anyway because each handshake creates
10972 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10973 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10974 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10975 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10978 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10979 'Q' and 'R' lose their special meanings (quit/renegotiate).
10980 This is part of what -quiet does; unlike -quiet, -ign_eof
10981 does not suppress any output.
10984 *) Add compatibility options to the purpose and trust code. The
10985 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10986 accepts a certificate or CA, this was the previous behaviour,
10987 with all the associated security issues.
10989 X509_TRUST_COMPAT is the old trust behaviour: only and
10990 automatically trust self signed roots in certificate store. A
10991 new trust setting X509_TRUST_DEFAULT is used to specify that
10992 a purpose has no associated trust setting and it should instead
10993 use the value in the default purpose.
10996 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10997 and fix a memory leak.
11000 *) In util/mkerr.pl (which implements 'make errors'), preserve
11001 reason strings from the previous version of the .c file, as
11002 the default to have only downcase letters (and digits) in
11003 automatically generated reasons codes is not always appropriate.
11006 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
11007 using strerror. Previously, ERR_reason_error_string() returned
11008 library names as reason strings for SYSerr; but SYSerr is a special
11009 case where small numbers are errno values, not library numbers.
11012 *) Add '-dsaparam' option to 'openssl dhparam' application. This
11013 converts DSA parameters into DH parameters. (When creating parameters,
11014 DSA_generate_parameters is used.)
11017 *) Include 'length' (recommended exponent length) in C code generated
11018 by 'openssl dhparam -C'.
11021 *) The second argument to set_label in perlasm was already being used
11022 so couldn't be used as a "file scope" flag. Moved to third argument
11026 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
11027 instead of RAND_bytes for encryption IVs and salts.
11030 *) Include RAND_status() into RAND_METHOD instead of implementing
11031 it only for md_rand.c Otherwise replacing the PRNG by calling
11032 RAND_set_rand_method would be impossible.
11035 *) Don't let DSA_generate_key() enter an infinite loop if the random
11036 number generation fails.
11039 *) New 'rand' application for creating pseudo-random output.
11042 *) Added configuration support for Linux/IA64
11043 [Rolf Haberrecker <rolf@suse.de>]
11045 *) Assembler module support for Mingw32.
11048 *) Shared library support for HPUX (in shlib/).
11049 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
11051 *) Shared library support for Solaris gcc.
11052 [Lutz Behnke <behnke@trustcenter.de>]
11054 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
11056 *) PKCS7_encrypt() was adding text MIME headers twice because they
11057 were added manually and by SMIME_crlf_copy().
11060 *) In bntest.c don't call BN_rand with zero bits argument.
11061 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
11063 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
11064 case was implemented. This caused BN_div_recp() to fail occasionally.
11067 *) Add an optional second argument to the set_label() in the perl
11068 assembly language builder. If this argument exists and is set
11069 to 1 it signals that the assembler should use a symbol whose
11070 scope is the entire file, not just the current function. This
11071 is needed with MASM which uses the format label:: for this scope.
11072 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
11074 *) Change the ASN1 types so they are typedefs by default. Before
11075 almost all types were #define'd to ASN1_STRING which was causing
11076 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
11080 *) Change names of new functions to the new get1/get0 naming
11081 convention: After 'get1', the caller owns a reference count
11082 and has to call ..._free; 'get0' returns a pointer to some
11083 data structure without incrementing reference counters.
11084 (Some of the existing 'get' functions increment a reference
11085 counter, some don't.)
11086 Similarly, 'set1' and 'add1' functions increase reference
11087 counters or duplicate objects.
11090 *) Allow for the possibility of temp RSA key generation failure:
11091 the code used to assume it always worked and crashed on failure.
11094 *) Fix potential buffer overrun problem in BIO_printf().
11095 [Ulf Möller, using public domain code by Patrick Powell; problem
11096 pointed out by David Sacerdote <das33@cornell.edu>]
11098 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
11099 RAND_egd() and RAND_status(). In the command line application,
11100 the EGD socket can be specified like a seed file using RANDFILE
11104 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
11105 Some CAs (e.g. Verisign) distribute certificates in this form.
11108 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
11109 list to exclude them. This means that no special compilation option
11110 is needed to use anonymous DH: it just needs to be included in the
11114 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
11115 EVP_MD_type. The old functionality is available in a new macro called
11116 EVP_MD_md(). Change code that uses it and update docs.
11119 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
11120 where the 'void *' argument is replaced by a function pointer argument.
11121 Previously 'void *' was abused to point to functions, which works on
11122 many platforms, but is not correct. As these functions are usually
11123 called by macros defined in OpenSSL header files, most source code
11124 should work without changes.
11127 *) <openssl/opensslconf.h> (which is created by Configure) now contains
11128 sections with information on -D... compiler switches used for
11129 compiling the library so that applications can see them. To enable
11130 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
11131 must be defined. E.g.,
11132 #define OPENSSL_ALGORITHM_DEFINES
11133 #include <openssl/opensslconf.h>
11134 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
11135 [Richard Levitte, Ulf and Bodo Möller]
11137 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
11141 *) Change the 'other' type in certificate aux info to a STACK_OF
11142 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
11143 the required ASN1 format: arbitrary types determined by an OID.
11146 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
11147 argument to 'req'. This is not because the function is newer or
11148 better than others it just uses the work 'NEW' in the certificate
11149 request header lines. Some software needs this.
11152 *) Reorganise password command line arguments: now passwords can be
11153 obtained from various sources. Delete the PEM_cb function and make
11154 it the default behaviour: i.e. if the callback is NULL and the
11155 usrdata argument is not NULL interpret it as a null terminated pass
11156 phrase. If usrdata and the callback are NULL then the pass phrase
11157 is prompted for as usual.
11160 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
11161 the support is automatically enabled. The resulting binaries will
11162 autodetect the card and use it if present.
11163 [Ben Laurie and Compaq Inc.]
11165 *) Work around for Netscape hang bug. This sends certificate request
11166 and server done in one record. Since this is perfectly legal in the
11167 SSL/TLS protocol it isn't a "bug" option and is on by default. See
11168 the bugs/SSLv3 entry for more info.
11171 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
11174 *) Add -rand argument to smime and pkcs12 applications and read/write
11178 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
11181 *) Add command line password options to the remaining applications.
11184 *) Bug fix for BN_div_recp() for numerators with an even number of
11188 *) More tests in bntest.c, and changed test_bn output.
11191 *) ./config recognizes MacOS X now.
11194 *) Bug fix for BN_div() when the first words of num and divisor are
11195 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
11198 *) Add support for various broken PKCS#8 formats, and command line
11199 options to produce them.
11202 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
11203 get temporary BIGNUMs from a BN_CTX.
11206 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
11210 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
11211 include a #define from the old name to the new. The original intent
11212 was that statically linked binaries could for example just call
11213 SSLeay_add_all_ciphers() to just add ciphers to the table and not
11214 link with digests. This never worked because SSLeay_add_all_digests()
11215 and SSLeay_add_all_ciphers() were in the same source file so calling
11216 one would link with the other. They are now in separate source files.
11219 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
11222 *) Use a less unusual form of the Miller-Rabin primality test (it used
11223 a binary algorithm for exponentiation integrated into the Miller-Rabin
11224 loop, our standard modexp algorithms are faster).
11227 *) Support for the EBCDIC character set completed.
11228 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
11230 *) Source code cleanups: use const where appropriate, eliminate casts,
11231 use void * instead of char * in lhash.
11234 *) Bugfix: ssl3_send_server_key_exchange was not restartable
11235 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
11236 this the server could overwrite ephemeral keys that the client
11240 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
11241 using 50 iterations of the Rabin-Miller test.
11243 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
11244 iterations of the Rabin-Miller test as required by the appendix
11245 to FIPS PUB 186[-1]) instead of DSA_is_prime.
11246 As BN_is_prime_fasttest includes trial division, DSA parameter
11247 generation becomes much faster.
11249 This implies a change for the callback functions in DSA_is_prime
11250 and DSA_generate_parameters: The callback function is called once
11251 for each positive witness in the Rabin-Miller test, not just
11252 occasionally in the inner loop; and the parameters to the
11253 callback function now provide an iteration count for the outer
11254 loop rather than for the current invocation of the inner loop.
11255 DSA_generate_parameters additionally can call the callback
11256 function with an 'iteration count' of -1, meaning that a
11257 candidate has passed the trial division test (when q is generated
11258 from an application-provided seed, trial division is skipped).
11261 *) New function BN_is_prime_fasttest that optionally does trial
11262 division before starting the Rabin-Miller test and has
11263 an additional BN_CTX * argument (whereas BN_is_prime always
11264 has to allocate at least one BN_CTX).
11265 'callback(1, -1, cb_arg)' is called when a number has passed the
11266 trial division stage.
11269 *) Fix for bug in CRL encoding. The validity dates weren't being handled
11273 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
11276 *) New function BN_pseudo_rand().
11279 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
11280 bignum version of BN_from_montgomery() with the working code from
11281 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
11285 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
11286 made it impossible to use the same SSL_SESSION data structure in
11287 SSL2 clients in multiple threads.
11290 *) The return value of RAND_load_file() no longer counts bytes obtained
11291 by stat(). RAND_load_file(..., -1) is new and uses the complete file
11292 to seed the PRNG (previously an explicit byte count was required).
11293 [Ulf Möller, Bodo Möller]
11295 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
11296 used (char *) instead of (void *) and had casts all over the place.
11299 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
11302 *) Retain source code compatibility for BN_prime_checks macro:
11303 BN_is_prime(..., BN_prime_checks, ...) now uses
11304 BN_prime_checks_for_size to determine the appropriate number of
11305 Rabin-Miller iterations.
11308 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
11309 DH_CHECK_P_NOT_SAFE_PRIME.
11310 (Check if this is true? OpenPGP calls them "strong".)
11313 *) Merge the functionality of "dh" and "gendh" programs into a new program
11314 "dhparam". The old programs are retained for now but will handle DH keys
11315 (instead of parameters) in future.
11318 *) Make the ciphers, s_server and s_client programs check the return values
11319 when a new cipher list is set.
11322 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
11323 ciphers. Before when the 56bit ciphers were enabled the sorting was
11326 The syntax for the cipher sorting has been extended to support sorting by
11327 cipher-strength (using the strength_bits hard coded in the tables).
11328 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
11330 Fix a bug in the cipher-command parser: when supplying a cipher command
11331 string with an "undefined" symbol (neither command nor alphanumeric
11332 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
11333 an error is flagged.
11335 Due to the strength-sorting extension, the code of the
11336 ssl_create_cipher_list() function was completely rearranged. I hope that
11337 the readability was also increased :-)
11338 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
11340 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
11341 for the first serial number and places 2 in the serial number file. This
11342 avoids problems when the root CA is created with serial number zero and
11343 the first user certificate has the same issuer name and serial number
11347 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
11348 the new code. Add documentation for this stuff.
11351 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
11352 X509_*() to X509at_*() on the grounds that they don't handle X509
11353 structures and behave in an analogous way to the X509v3 functions:
11354 they shouldn't be called directly but wrapper functions should be used
11357 So we also now have some wrapper functions that call the X509at functions
11358 when passed certificate requests. (TO DO: similar things can be done with
11359 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11360 things. Some of these need some d2i or i2d and print functionality
11361 because they handle more complex structures.)
11364 *) Add missing #ifndefs that caused missing symbols when building libssl
11365 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11366 NO_RSA in ssl/s2*.c.
11367 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11369 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11370 has a return value which indicates the quality of the random data
11371 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11372 error queue. New function RAND_pseudo_bytes() generates output that is
11373 guaranteed to be unique but not unpredictable. RAND_add is like
11374 RAND_seed, but takes an extra argument for an entropy estimate
11375 (RAND_seed always assumes full entropy).
11378 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11379 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11380 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11381 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11382 false-positive rate of at most 2^-80 for random input.
11385 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11388 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11389 in the 0.9.5 release), this returns the chain
11390 from an X509_CTX structure with a dup of the stack and all
11391 the X509 reference counts upped: so the stack will exist
11392 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11395 Also make SSL_SESSION_print() print out the verify return
11399 *) Add manpage for the pkcs12 command. Also change the default
11400 behaviour so MAC iteration counts are used unless the new
11401 -nomaciter option is used. This improves file security and
11402 only older versions of MSIE (4.0 for example) need it.
11405 *) Honor the no-xxx Configure options when creating .DEF files.
11408 *) Add PKCS#10 attributes to field table: challengePassword,
11409 unstructuredName and unstructuredAddress. These are taken from
11410 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11411 international characters are used.
11413 More changes to X509_ATTRIBUTE code: allow the setting of types
11414 based on strings. Remove the 'loc' parameter when adding
11415 attributes because these will be a SET OF encoding which is sorted
11419 *) Initial changes to the 'req' utility to allow request generation
11420 automation. This will allow an application to just generate a template
11421 file containing all the field values and have req construct the
11424 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11425 used all over the place including certificate requests and PKCS#7
11426 structures. They are currently handled manually where necessary with
11427 some primitive wrappers for PKCS#7. The new functions behave in a
11428 manner analogous to the X509 extension functions: they allow
11429 attributes to be looked up by NID and added.
11431 Later something similar to the X509V3 code would be desirable to
11432 automatically handle the encoding, decoding and printing of the
11433 more complex types. The string types like challengePassword can
11434 be handled by the string table functions.
11436 Also modified the multi byte string table handling. Now there is
11437 a 'global mask' which masks out certain types. The table itself
11438 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11439 is useful when for example there is only one permissible type
11440 (as in countryName) and using the mask might result in no valid
11444 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11445 SSL_get_peer_finished to allow applications to obtain the latest
11446 Finished messages sent to the peer or expected from the peer,
11447 respectively. (SSL_get_peer_finished is usually the Finished message
11448 actually received from the peer, otherwise the protocol will be aborted.)
11450 As the Finished message are message digests of the complete handshake
11451 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11452 be used for external authentication procedures when the authentication
11453 provided by SSL/TLS is not desired or is not enough.
11456 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11457 the host supports BWX extension and if Compaq C is present on the
11458 $PATH. Just exploiting of the BWX extension results in 20-30%
11459 performance kick for some algorithms, e.g. DES and RC4 to mention
11460 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11464 *) Add support for MS "fast SGC". This is arguably a violation of the
11465 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11466 weak crypto and after checking the certificate is SGC a second one
11467 with strong crypto. MS SGC stops the first handshake after receiving
11468 the server certificate message and sends a second client hello. Since
11469 a server will typically do all the time consuming operations before
11470 expecting any further messages from the client (server key exchange
11471 is the most expensive) there is little difference between the two.
11473 To get OpenSSL to support MS SGC we have to permit a second client
11474 hello message after we have sent server done. In addition we have to
11475 reset the MAC if we do get this second client hello.
11478 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11479 if a DER encoded private key is RSA or DSA traditional format. Changed
11480 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11481 format DER encoded private key. Newer code should use PKCS#8 format which
11482 has the key type encoded in the ASN1 structure. Added DER private key
11483 support to pkcs8 application.
11486 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11487 ciphersuites has been selected (as required by the SSL 3/TLS 1
11488 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11489 is set, we interpret this as a request to violate the specification
11490 (the worst that can happen is a handshake failure, and 'correct'
11491 behaviour would result in a handshake failure anyway).
11494 *) In SSL_CTX_add_session, take into account that there might be multiple
11495 SSL_SESSION structures with the same session ID (e.g. when two threads
11496 concurrently obtain them from an external cache).
11497 The internal cache can handle only one SSL_SESSION with a given ID,
11498 so if there's a conflict, we now throw out the old one to achieve
11502 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11503 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11504 some routines that use cipher OIDs: some ciphers do not have OIDs
11505 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11509 *) Simplify the trust setting structure and code. Now we just have
11510 two sequences of OIDs for trusted and rejected settings. These will
11511 typically have values the same as the extended key usage extension
11512 and any application specific purposes.
11514 The trust checking code now has a default behaviour: it will just
11515 check for an object with the same NID as the passed id. Functions can
11516 be provided to override either the default behaviour or the behaviour
11517 for a given id. SSL client, server and email already have functions
11518 in place for compatibility: they check the NID and also return "trusted"
11519 if the certificate is self signed.
11522 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11523 traditional format into an EVP_PKEY structure.
11526 *) Add a password callback function PEM_cb() which either prompts for
11527 a password if usr_data is NULL or otherwise assumes it is a null
11528 terminated password. Allow passwords to be passed on command line
11529 environment or config files in a few more utilities.
11532 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11533 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11534 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11535 Update documentation.
11538 *) Support for ASN1 "NULL" type. This could be handled before by using
11539 ASN1_TYPE but there wasn't any function that would try to read a NULL
11540 and produce an error if it couldn't. For compatibility we also have
11541 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11542 don't allocate anything because they don't need to.
11545 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11547 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11549 *) Rebuild of the memory allocation routines used by OpenSSL code and
11550 possibly others as well. The purpose is to make an interface that
11551 provide hooks so anyone can build a separate set of allocation and
11552 deallocation routines to be used by OpenSSL, for example memory
11553 pool implementations, or something else, which was previously hard
11554 since Malloc(), Realloc() and Free() were defined as macros having
11555 the values malloc, realloc and free, respectively (except for Win32
11556 compilations). The same is provided for memory debugging code.
11557 OpenSSL already comes with functionality to find memory leaks, but
11558 this gives people a chance to debug other memory problems.
11560 With these changes, a new set of functions and macros have appeared:
11562 CRYPTO_set_mem_debug_functions() [F]
11563 CRYPTO_get_mem_debug_functions() [F]
11564 CRYPTO_dbg_set_options() [F]
11565 CRYPTO_dbg_get_options() [F]
11566 CRYPTO_malloc_debug_init() [M]
11568 The memory debug functions are NULL by default, unless the library
11569 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11570 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11571 gives the standard debugging functions that come with OpenSSL) or
11572 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11573 provided by the library user) must be used. When the standard
11574 debugging functions are used, CRYPTO_dbg_set_options can be used to
11575 request additional information:
11576 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11577 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11579 Also, things like CRYPTO_set_mem_functions will always give the
11580 expected result (the new set of functions is used for allocation
11581 and deallocation) at all times, regardless of platform and compiler
11584 To finish it up, some functions that were never use in any other
11585 way than through macros have a new API and new semantic:
11587 CRYPTO_dbg_malloc()
11588 CRYPTO_dbg_realloc()
11591 All macros of value have retained their old syntax.
11592 [Richard Levitte and Bodo Moeller]
11594 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11595 ordering of SMIMECapabilities wasn't in "strength order" and there
11596 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11600 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11601 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11602 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11604 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11605 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11606 functionality to handle multipart/signed properly) and a utility
11607 called 'smime' to call all this stuff. This is based on code I
11608 originally wrote for Celo who have kindly allowed it to be
11609 included in OpenSSL.
11612 *) Add variants des_set_key_checked and des_set_key_unchecked of
11613 des_set_key (aka des_key_sched). Global variable des_check_key
11614 decides which of these is called by des_set_key; this way
11615 des_check_key behaves as it always did, but applications and
11616 the library itself, which was buggy for des_check_key == 1,
11617 have a cleaner way to pick the version they need.
11620 *) New function PKCS12_newpass() which changes the password of a
11624 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11625 dynamic mix. In both cases the ids can be used as an index into the
11626 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11627 functions so they accept a list of the field values and the
11628 application doesn't need to directly manipulate the X509_TRUST
11632 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11636 *) Modify the way the V3 extension code looks up extensions. This now
11637 works in a similar way to the object code: we have some "standard"
11638 extensions in a static table which is searched with OBJ_bsearch()
11639 and the application can add dynamic ones if needed. The file
11640 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11641 updated whenever a new extension is added to the core code and kept
11642 in ext_nid order. There is a simple program 'tabtest.c' which checks
11643 this. New extensions are not added too often so this file can readily
11644 be maintained manually.
11646 There are two big advantages in doing things this way. The extensions
11647 can be looked up immediately and no longer need to be "added" using
11648 X509V3_add_standard_extensions(): this function now does nothing.
11649 [Side note: I get *lots* of email saying the extension code doesn't
11650 work because people forget to call this function]
11651 Also no dynamic allocation is done unless new extensions are added:
11652 so if we don't add custom extensions there is no need to call
11653 X509V3_EXT_cleanup().
11656 *) Modify enc utility's salting as follows: make salting the default. Add a
11657 magic header, so unsalted files fail gracefully instead of just decrypting
11658 to garbage. This is because not salting is a big security hole, so people
11659 should be discouraged from doing it.
11662 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11663 digest to be passed on the command line but it only used this
11664 parameter when signing a certificate. Modified so all relevant
11665 operations are affected by the digest parameter including the
11666 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11667 DSA key was used because it didn't fix the digest.
11670 *) Initial certificate chain verify code. Currently tests the untrusted
11671 certificates for consistency with the verify purpose (which is set
11672 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11674 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11675 this is because it will reject chains with invalid extensions whereas
11676 every previous version of OpenSSL and SSLeay made no checks at all.
11678 Trust code: checks the root CA for the relevant trust settings. Trust
11679 settings have an initial value consistent with the verify purpose: e.g.
11680 if the verify purpose is for SSL client use it expects the CA to be
11681 trusted for SSL client use. However the default value can be changed to
11682 permit custom trust settings: one example of this would be to only trust
11683 certificates from a specific "secure" set of CAs.
11685 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11686 which should be used for version portability: especially since the
11687 verify structure is likely to change more often now.
11689 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11690 to set them. If not set then assume SSL clients will verify SSL servers
11693 Two new options to the verify program: -untrusted allows a set of
11694 untrusted certificates to be passed in and -purpose which sets the
11695 intended purpose of the certificate. If a purpose is set then the
11696 new chain verify code is used to check extension consistency.
11699 *) Support for the authority information access extension.
11702 *) Modify RSA and DSA PEM read routines to transparently handle
11703 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11704 public keys in a format compatible with certificate
11705 SubjectPublicKeyInfo structures. Unfortunately there were already
11706 functions called *_PublicKey_* which used various odd formats so
11707 these are retained for compatibility: however the DSA variants were
11708 never in a public release so they have been deleted. Changed dsa/rsa
11709 utilities to handle the new format: note no releases ever handled public
11710 keys so we should be OK.
11712 The primary motivation for this change is to avoid the same fiasco
11713 that dogs private keys: there are several incompatible private key
11714 formats some of which are standard and some OpenSSL specific and
11715 require various evil hacks to allow partial transparent handling and
11716 even then it doesn't work with DER formats. Given the option anything
11717 other than PKCS#8 should be dumped: but the other formats have to
11718 stay in the name of compatibility.
11720 With public keys and the benefit of hindsight one standard format
11721 is used which works with EVP_PKEY, RSA or DSA structures: though
11722 it clearly returns an error if you try to read the wrong kind of key.
11724 Added a -pubkey option to the 'x509' utility to output the public key.
11725 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11726 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11727 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11728 that do the same as the EVP_PKEY_assign_*() except they up the
11729 reference count of the added key (they don't "swallow" the
11733 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11734 CRLs would fail if the file contained no certificates or no CRLs:
11735 added a new function to read in both types and return the number
11736 read: this means that if none are read it will be an error. The
11737 DER versions of the certificate and CRL reader would always fail
11738 because it isn't possible to mix certificates and CRLs in DER format
11739 without choking one or the other routine. Changed this to just read
11740 a certificate: this is the best we can do. Also modified the code
11741 in apps/verify.c to take notice of return codes: it was previously
11742 attempting to read in certificates from NULL pointers and ignoring
11743 any errors: this is one reason why the cert and CRL reader seemed
11744 to work. It doesn't check return codes from the default certificate
11745 routines: these may well fail if the certificates aren't installed.
11748 *) Code to support otherName option in GeneralName.
11751 *) First update to verify code. Change the verify utility
11752 so it warns if it is passed a self signed certificate:
11753 for consistency with the normal behaviour. X509_verify
11754 has been modified to it will now verify a self signed
11755 certificate if *exactly* the same certificate appears
11756 in the store: it was previously impossible to trust a
11757 single self signed certificate. This means that:
11758 openssl verify ss.pem
11759 now gives a warning about a self signed certificate but
11760 openssl verify -CAfile ss.pem ss.pem
11764 *) For servers, store verify_result in SSL_SESSION data structure
11765 (and add it to external session representation).
11766 This is needed when client certificate verifications fails,
11767 but an application-provided verification callback (set by
11768 SSL_CTX_set_cert_verify_callback) allows accepting the session
11769 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11770 but returns 1): When the session is reused, we have to set
11771 ssl->verify_result to the appropriate error code to avoid
11773 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11775 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11776 case in PKCS7_dataInit() where the signed PKCS7 structure
11777 didn't contain any existing data because it was being created.
11778 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11780 *) Add a salt to the key derivation routines in enc.c. This
11781 forms the first 8 bytes of the encrypted file. Also add a
11782 -S option to allow a salt to be input on the command line.
11785 *) New function X509_cmp(). Oddly enough there wasn't a function
11786 to compare two certificates. We do this by working out the SHA1
11787 hash and comparing that. X509_cmp() will be needed by the trust
11791 *) SSL_get1_session() is like SSL_get_session(), but increments
11792 the reference count in the SSL_SESSION returned.
11793 [Geoff Thorpe <geoff@eu.c2.net>]
11795 *) Fix for 'req': it was adding a null to request attributes.
11796 Also change the X509_LOOKUP and X509_INFO code to handle
11797 certificate auxiliary information.
11800 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11804 *) Add the possibility to add extra information to the memory leak
11805 detecting output, to form tracebacks, showing from where each
11806 allocation was originated: CRYPTO_push_info("constant string") adds
11807 the string plus current file name and line number to a per-thread
11808 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11809 is like calling CYRPTO_pop_info() until the stack is empty.
11810 Also updated memory leak detection code to be multi-thread-safe.
11813 *) Add options -text and -noout to pkcs7 utility and delete the
11814 encryption options which never did anything. Update docs.
11817 *) Add options to some of the utilities to allow the pass phrase
11818 to be included on either the command line (not recommended on
11819 OSes like Unix) or read from the environment. Update the
11820 manpages and fix a few bugs.
11823 *) Add a few manpages for some of the openssl commands.
11826 *) Fix the -revoke option in ca. It was freeing up memory twice,
11827 leaking and not finding already revoked certificates.
11830 *) Extensive changes to support certificate auxiliary information.
11831 This involves the use of X509_CERT_AUX structure and X509_AUX
11832 functions. An X509_AUX function such as PEM_read_X509_AUX()
11833 can still read in a certificate file in the usual way but it
11834 will also read in any additional "auxiliary information". By
11835 doing things this way a fair degree of compatibility can be
11836 retained: existing certificates can have this information added
11837 using the new 'x509' options.
11839 Current auxiliary information includes an "alias" and some trust
11840 settings. The trust settings will ultimately be used in enhanced
11841 certificate chain verification routines: currently a certificate
11842 can only be trusted if it is self signed and then it is trusted
11846 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11847 The problem was that one of the replacement routines had not been working
11848 since SSLeay releases. For now the offending routine has been replaced
11849 with non-optimised assembler. Even so, this now gives around 95%
11850 performance improvement for 1024 bit RSA signs.
11853 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11854 handling. Most clients have the effective key size in bits equal to
11855 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11856 A few however don't do this and instead use the size of the decrypted key
11857 to determine the RC2 key length and the AlgorithmIdentifier to determine
11858 the effective key length. In this case the effective key length can still
11859 be 40 bits but the key length can be 168 bits for example. This is fixed
11860 by manually forcing an RC2 key into the EVP_PKEY structure because the
11861 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11862 the key length and effective key length are equal.
11865 *) Add a bunch of functions that should simplify the creation of
11866 X509_NAME structures. Now you should be able to do:
11867 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11868 and have it automatically work out the correct field type and fill in
11869 the structures. The more adventurous can try:
11870 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11871 and it will (hopefully) work out the correct multibyte encoding.
11874 *) Change the 'req' utility to use the new field handling and multibyte
11875 copy routines. Before the DN field creation was handled in an ad hoc
11876 way in req, ca, and x509 which was rather broken and didn't support
11877 BMPStrings or UTF8Strings. Since some software doesn't implement
11878 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11879 using the dirstring_type option. See the new comment in the default
11880 openssl.cnf for more info.
11883 *) Make crypto/rand/md_rand.c more robust:
11884 - Assure unique random numbers after fork().
11885 - Make sure that concurrent threads access the global counter and
11886 md serializably so that we never lose entropy in them
11887 or use exactly the same state in multiple threads.
11888 Access to the large state is not always serializable because
11889 the additional locking could be a performance killer, and
11890 md should be large enough anyway.
11893 *) New file apps/app_rand.c with commonly needed functionality
11894 for handling the random seed file.
11896 Use the random seed file in some applications that previously did not:
11898 dsaparam -genkey (which also ignored its '-rand' option),
11901 x509 (when signing).
11902 Except on systems with /dev/urandom, it is crucial to have a random
11903 seed file at least for key creation, DSA signing, and for DH exchanges;
11904 for RSA signatures we could do without one.
11906 gendh and gendsa (unlike genrsa) used to read only the first byte
11907 of each file listed in the '-rand' option. The function as previously
11908 found in genrsa is now in app_rand.c and is used by all programs
11909 that support '-rand'.
11912 *) In RAND_write_file, use mode 0600 for creating files;
11913 don't just chmod when it may be too late.
11916 *) Report an error from X509_STORE_load_locations
11917 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11920 *) New function ASN1_mbstring_copy() this copies a string in either
11921 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11922 into an ASN1_STRING type. A mask of permissible types is passed
11923 and it chooses the "minimal" type to use or an error if not type
11927 *) Add function equivalents to the various macros in asn1.h. The old
11928 macros are retained with an M_ prefix. Code inside the library can
11929 use the M_ macros. External code (including the openssl utility)
11930 should *NOT* in order to be "shared library friendly".
11933 *) Add various functions that can check a certificate's extensions
11934 to see if it usable for various purposes such as SSL client,
11935 server or S/MIME and CAs of these types. This is currently
11936 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11937 verification. Also added a -purpose flag to x509 utility to
11938 print out all the purposes.
11941 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11945 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11946 for, obtain and decode and extension and obtain its critical flag.
11947 This allows all the necessary extension code to be handled in a
11948 single function call.
11951 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11952 platforms. See crypto/rc4/rc4_enc.c for further details.
11955 *) New -noout option to asn1parse. This causes no output to be produced
11956 its main use is when combined with -strparse and -out to extract data
11957 from a file (which may not be in ASN.1 format).
11960 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11961 when producing the local key id.
11962 [Richard Levitte <levitte@stacken.kth.se>]
11964 *) New option -dhparam in s_server. This allows a DH parameter file to be
11965 stated explicitly. If it is not stated then it tries the first server
11966 certificate file. The previous behaviour hard coded the filename
11970 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11971 a public key to be input or output. For example:
11972 openssl rsa -in key.pem -pubout -out pubkey.pem
11973 Also added necessary DSA public key functions to handle this.
11976 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11977 in the message. This was handled by allowing
11978 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11979 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11981 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11982 to the end of the strings whereas this didn't. This would cause problems
11983 if strings read with d2i_ASN1_bytes() were later modified.
11984 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11986 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11987 data and it contains EOF it will end up returning an error. This is
11988 caused by input 46 bytes long. The cause is due to the way base64
11989 BIOs find the start of base64 encoded data. They do this by trying a
11990 trial decode on each line until they find one that works. When they
11991 do a flag is set and it starts again knowing it can pass all the
11992 data directly through the decoder. Unfortunately it doesn't reset
11993 the context it uses. This means that if EOF is reached an attempt
11994 is made to pass two EOFs through the context and this causes the
11995 resulting error. This can also cause other problems as well. As is
11996 usual with these problems it takes *ages* to find and the fix is
11997 trivial: move one line.
11998 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
12000 *) Ugly workaround to get s_client and s_server working under Windows. The
12001 old code wouldn't work because it needed to select() on sockets and the
12002 tty (for keypresses and to see if data could be written). Win32 only
12003 supports select() on sockets so we select() with a 1s timeout on the
12004 sockets and then see if any characters are waiting to be read, if none
12005 are present then we retry, we also assume we can always write data to
12006 the tty. This isn't nice because the code then blocks until we've
12007 received a complete line of data and it is effectively polling the
12008 keyboard at 1s intervals: however it's quite a bit better than not
12009 working at all :-) A dedicated Windows application might handle this
12010 with an event loop for example.
12013 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
12014 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
12015 will be called when RSA_sign() and RSA_verify() are used. This is useful
12016 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
12017 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
12018 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
12019 This necessitated the support of an extra signature type NID_md5_sha1
12020 for SSL signatures and modifications to the SSL library to use it instead
12021 of calling RSA_public_decrypt() and RSA_private_encrypt().
12024 *) Add new -verify -CAfile and -CApath options to the crl program, these
12025 will lookup a CRL issuers certificate and verify the signature in a
12026 similar way to the verify program. Tidy up the crl program so it
12027 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
12028 less strict. It will now permit CRL extensions even if it is not
12029 a V2 CRL: this will allow it to tolerate some broken CRLs.
12032 *) Initialize all non-automatic variables each time one of the openssl
12033 sub-programs is started (this is necessary as they may be started
12034 multiple times from the "OpenSSL>" prompt).
12035 [Lennart Bang, Bodo Moeller]
12037 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
12038 removing all other RSA functionality (this is what NO_RSA does). This
12039 is so (for example) those in the US can disable those operations covered
12040 by the RSA patent while allowing storage and parsing of RSA keys and RSA
12044 *) Non-copying interface to BIO pairs.
12045 (still largely untested)
12048 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
12049 ASCII string. This was handled independently in various places before.
12052 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
12053 UTF8 strings a character at a time.
12056 *) Use client_version from client hello to select the protocol
12057 (s23_srvr.c) and for RSA client key exchange verification
12058 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
12061 *) Add various utility functions to handle SPKACs, these were previously
12062 handled by poking round in the structure internals. Added new function
12063 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
12064 print, verify and generate SPKACs. Based on an original idea from
12065 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
12068 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
12071 *) Allow the config file extension section to be overwritten on the
12072 command line. Based on an original idea from Massimiliano Pala
12073 <madwolf@comune.modena.it>. The new option is called -extensions
12074 and can be applied to ca, req and x509. Also -reqexts to override
12075 the request extensions in req and -crlexts to override the crl extensions
12079 *) Add new feature to the SPKAC handling in ca. Now you can include
12080 the same field multiple times by preceding it by "XXXX." for example:
12083 this is the same syntax as used in the req config file.
12086 *) Allow certificate extensions to be added to certificate requests. These
12087 are specified in a 'req_extensions' option of the req section of the
12088 config file. They can be printed out with the -text option to req but
12089 are otherwise ignored at present.
12092 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
12093 data read consists of only the final block it would not decrypted because
12094 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
12095 A misplaced 'break' also meant the decrypted final block might not be
12096 copied until the next read.
12099 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
12100 a few extra parameters to the DH structure: these will be useful if
12101 for example we want the value of 'q' or implement X9.42 DH.
12104 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
12105 provides hooks that allow the default DSA functions or functions on a
12106 "per key" basis to be replaced. This allows hardware acceleration and
12107 hardware key storage to be handled without major modification to the
12108 library. Also added low level modexp hooks and CRYPTO_EX structure and
12109 associated functions.
12112 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
12113 as "read only": it can't be written to and the buffer it points to will
12114 not be freed. Reading from a read only BIO is much more efficient than
12115 a normal memory BIO. This was added because there are several times when
12116 an area of memory needs to be read from a BIO. The previous method was
12117 to create a memory BIO and write the data to it, this results in two
12118 copies of the data and an O(n^2) reading algorithm. There is a new
12119 function BIO_new_mem_buf() which creates a read only memory BIO from
12120 an area of memory. Also modified the PKCS#7 routines to use read only
12124 *) Bugfix: ssl23_get_client_hello did not work properly when called in
12125 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
12126 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
12127 but a retry condition occurred while trying to read the rest.
12130 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
12131 NID_pkcs7_encrypted by default: this was wrong since this should almost
12132 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
12133 the encrypted data type: this is a more sensible place to put it and it
12134 allows the PKCS#12 code to be tidied up that duplicated this
12138 *) Changed obj_dat.pl script so it takes its input and output files on
12139 the command line. This should avoid shell escape redirection problems
12143 *) Initial support for certificate extension requests, these are included
12144 in things like Xenroll certificate requests. Included functions to allow
12145 extensions to be obtained and added.
12148 *) -crlf option to s_client and s_server for sending newlines as
12149 CRLF (as required by many protocols).
12152 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
12154 *) Install libRSAglue.a when OpenSSL is built with RSAref.
12155 [Ralf S. Engelschall]
12157 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
12158 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
12160 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
12164 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
12165 DH parameters/keys (q is lost during that conversion, but the resulting
12166 DH parameters contain its length).
12168 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
12169 much faster than DH_generate_parameters (which creates parameters
12170 where p = 2*q + 1), and also the smaller q makes DH computations
12171 much more efficient (160-bit exponentiation instead of 1024-bit
12172 exponentiation); so this provides a convenient way to support DHE
12173 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
12174 utter importance to use
12175 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12177 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12178 when such DH parameters are used, because otherwise small subgroup
12179 attacks may become possible!
12182 *) Avoid memory leak in i2d_DHparams.
12185 *) Allow the -k option to be used more than once in the enc program:
12186 this allows the same encrypted message to be read by multiple recipients.
12189 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
12190 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
12191 it will always use the numerical form of the OID, even if it has a short
12195 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
12196 method only got called if p,q,dmp1,dmq1,iqmp components were present,
12197 otherwise bn_mod_exp was called. In the case of hardware keys for example
12198 no private key components need be present and it might store extra data
12199 in the RSA structure, which cannot be accessed from bn_mod_exp.
12200 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
12201 private key operations.
12204 *) Added support for SPARC Linux.
12207 *) pem_password_cb function type incompatibly changed from
12208 typedef int pem_password_cb(char *buf, int size, int rwflag);
12210 ....(char *buf, int size, int rwflag, void *userdata);
12211 so that applications can pass data to their callbacks:
12212 The PEM[_ASN1]_{read,write}... functions and macros now take an
12213 additional void * argument, which is just handed through whenever
12214 the password callback is called.
12215 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
12217 New function SSL_CTX_set_default_passwd_cb_userdata.
12219 Compatibility note: As many C implementations push function arguments
12220 onto the stack in reverse order, the new library version is likely to
12221 interoperate with programs that have been compiled with the old
12222 pem_password_cb definition (PEM_whatever takes some data that
12223 happens to be on the stack as its last argument, and the callback
12224 just ignores this garbage); but there is no guarantee whatsoever that
12227 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
12228 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
12229 problems not only on Windows, but also on some Unix platforms.
12230 To avoid problematic command lines, these definitions are now in an
12231 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
12232 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
12235 *) MIPS III/IV assembler module is reimplemented.
12238 *) More DES library cleanups: remove references to srand/rand and
12239 delete an unused file.
12242 *) Add support for the free Netwide assembler (NASM) under Win32,
12243 since not many people have MASM (ml) and it can be hard to obtain.
12244 This is currently experimental but it seems to work OK and pass all
12245 the tests. Check out INSTALL.W32 for info.
12248 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
12249 without temporary keys kept an extra copy of the server key,
12250 and connections with temporary keys did not free everything in case
12254 *) New function RSA_check_key and new openssl rsa option -check
12255 for verifying the consistency of RSA keys.
12256 [Ulf Moeller, Bodo Moeller]
12258 *) Various changes to make Win32 compile work:
12259 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
12260 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
12261 comparison" warnings.
12262 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
12265 *) Add a debugging option to PKCS#5 v2 key generation function: when
12266 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
12267 derived keys are printed to stderr.
12270 *) Copy the flags in ASN1_STRING_dup().
12271 [Roman E. Pavlov <pre@mo.msk.ru>]
12273 *) The x509 application mishandled signing requests containing DSA
12274 keys when the signing key was also DSA and the parameters didn't match.
12276 It was supposed to omit the parameters when they matched the signing key:
12277 the verifying software was then supposed to automatically use the CA's
12278 parameters if they were absent from the end user certificate.
12280 Omitting parameters is no longer recommended. The test was also
12281 the wrong way round! This was probably due to unusual behaviour in
12282 EVP_cmp_parameters() which returns 1 if the parameters match.
12283 This meant that parameters were omitted when they *didn't* match and
12284 the certificate was useless. Certificates signed with 'ca' didn't have
12286 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
12288 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
12289 The interface is as follows:
12290 Applications can use
12291 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
12292 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
12293 "off" is now the default.
12294 The library internally uses
12295 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
12296 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
12297 to disable memory-checking temporarily.
12299 Some inconsistent states that previously were possible (and were
12300 even the default) are now avoided.
12302 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
12303 with each memory chunk allocated; this is occasionally more helpful
12304 than just having a counter.
12306 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
12308 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
12312 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
12313 which largely parallels "options", but is for changing API behaviour,
12314 whereas "options" are about protocol behaviour.
12315 Initial "mode" flags are:
12317 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
12318 a single record has been written.
12319 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
12320 retries use the same buffer location.
12321 (But all of the contents must be
12325 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
12328 *) Fix problems with no-hmac etc.
12329 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
12331 *) New functions RSA_get_default_method(), RSA_set_method() and
12332 RSA_get_method(). These allows replacement of RSA_METHODs without having
12333 to mess around with the internals of an RSA structure.
12336 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
12337 Also really enable memory leak checks in openssl.c and in some
12339 [Chad C. Mulligan, Bodo Moeller]
12341 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
12342 up the length of negative integers. This has now been simplified to just
12343 store the length when it is first determined and use it later, rather
12344 than trying to keep track of where data is copied and updating it to
12346 [Steve Henson, reported by Brien Wheeler
12347 <bwheeler@authentica-security.com>]
12349 *) Add a new function PKCS7_signatureVerify. This allows the verification
12350 of a PKCS#7 signature but with the signing certificate passed to the
12351 function itself. This contrasts with PKCS7_dataVerify which assumes the
12352 certificate is present in the PKCS#7 structure. This isn't always the
12353 case: certificates can be omitted from a PKCS#7 structure and be
12354 distributed by "out of band" means (such as a certificate database).
12357 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12358 function prototypes in pem.h, also change util/mkdef.pl to add the
12359 necessary function names.
12362 *) mk1mf.pl (used by Windows builds) did not properly read the
12363 options set by Configure in the top level Makefile, and Configure
12364 was not even able to write more than one option correctly.
12365 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12368 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12369 file to be loaded from a BIO or FILE pointer. The BIO version will
12370 for example allow memory BIOs to contain config info.
12373 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12374 Whoever hopes to achieve shared-library compatibility across versions
12375 must use this, not the compile-time macro.
12376 (Exercise 0.9.4: Which is the minimum library version required by
12378 Note: All this applies only to multi-threaded programs, others don't
12382 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12383 through a BIO pair triggered the default case, i.e.
12384 SSLerr(...,SSL_R_UNKNOWN_STATE).
12387 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12388 can use the SSL library even if none of the specific BIOs is
12392 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12393 for the encoded length.
12394 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12396 *) Add initial documentation of the X509V3 functions.
12399 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12400 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12401 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12402 secure PKCS#8 private key format with a high iteration count.
12405 *) Fix determination of Perl interpreter: A perl or perl5
12406 _directory_ in $PATH was also accepted as the interpreter.
12407 [Ralf S. Engelschall]
12409 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12410 wrong with it but it was very old and did things like calling
12411 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12412 unusual formatting.
12415 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12416 to use the new extension code.
12419 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12420 with macros. This should make it easier to change their form, add extra
12421 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12425 *) Add to configuration table a new entry that can specify an alternative
12426 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12427 according to Mark Crispin <MRC@Panda.COM>.
12431 *) DES CBC did not update the IV. Weird.
12434 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12435 Changing the behaviour of the former might break existing programs --
12436 where IV updating is needed, des_ncbc_encrypt can be used.
12439 *) When bntest is run from "make test" it drives bc to check its
12440 calculations, as well as internally checking them. If an internal check
12441 fails, it needs to cause bc to give a non-zero result or make test carries
12442 on without noticing the failure. Fixed.
12445 *) DES library cleanups.
12448 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12449 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12450 ciphers. NOTE: although the key derivation function has been verified
12451 against some published test vectors it has not been extensively tested
12452 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12456 *) Instead of "mkdir -p", which is not fully portable, use new
12457 Perl script "util/mkdir-p.pl".
12460 *) Rewrite the way password based encryption (PBE) is handled. It used to
12461 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12462 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12463 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12464 the 'parameter' field of the AlgorithmIdentifier is passed to the
12465 underlying key generation function so it must do its own ASN1 parsing.
12466 This has also changed the EVP_PBE_CipherInit() function which now has a
12467 'parameter' argument instead of literal salt and iteration count values
12468 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12471 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12472 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12473 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12474 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12475 value was just used as a "magic string" and not used directly its
12476 value doesn't matter.
12479 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12483 *) "linux-sparc64" configuration (ultrapenguin).
12484 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12485 "linux-sparc" configuration.
12486 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12488 *) config now generates no-xxx options for missing ciphers.
12491 *) Support the EBCDIC character set (work in progress).
12492 File ebcdic.c not yet included because it has a different license.
12493 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12495 *) Support BS2000/OSD-POSIX.
12496 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12498 *) Make callbacks for key generation use void * instead of char *.
12501 *) Make S/MIME samples compile (not yet tested).
12504 *) Additional typesafe stacks.
12507 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12511 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12513 *) New configuration variant "sco5-gcc".
12515 *) Updated some demos.
12516 [Sean O Riordain, Wade Scholine]
12518 *) Add missing BIO_free at exit of pkcs12 application.
12521 *) Fix memory leak in conf.c.
12524 *) Updates for Win32 to assembler version of MD5.
12527 *) Set #! path to perl in apps/der_chop to where we found it
12528 instead of using a fixed path.
12531 *) SHA library changes for irix64-mips4-cc.
12534 *) Improvements for VMS support.
12538 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12540 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12541 This also avoids the problems with SC4.2 and unpatched SC5.
12542 [Andy Polyakov <appro@fy.chalmers.se>]
12544 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12545 These are required because of the typesafe stack would otherwise break
12546 existing code. If old code used a structure member which used to be STACK
12547 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12548 sk_num or sk_value it would produce an error because the num, data members
12549 are not present in STACK_OF. Now it just produces a warning. sk_set
12550 replaces the old method of assigning a value to sk_value
12551 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12552 that does this will no longer work (and should use sk_set instead) but
12553 this could be regarded as a "questionable" behaviour anyway.
12556 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12557 correctly handle encrypted S/MIME data.
12560 *) Change type of various DES function arguments from des_cblock
12561 (which means, in function argument declarations, pointer to char)
12562 to des_cblock * (meaning pointer to array with 8 char elements),
12563 which allows the compiler to do more typechecking; it was like
12564 that back in SSLeay, but with lots of ugly casts.
12566 Introduce new type const_des_cblock.
12569 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12570 problems: find RecipientInfo structure that matches recipient certificate
12571 and initialise the ASN1 structures properly based on passed cipher.
12574 *) Belatedly make the BN tests actually check the results.
12577 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12578 to and from BNs: it was completely broken. New compilation option
12579 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12580 key elements as negative integers.
12583 *) Reorganize and speed up MD5.
12584 [Andy Polyakov <appro@fy.chalmers.se>]
12587 [Richard Levitte <richard@levitte.org>]
12589 *) New option -out to asn1parse to allow the parsed structure to be
12590 output to a file. This is most useful when combined with the -strparse
12591 option to examine the output of things like OCTET STRINGS.
12594 *) Make SSL library a little more fool-proof by not requiring any longer
12595 that SSL_set_{accept,connect}_state be called before
12596 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12597 in many applications because usually everything *appeared* to work as
12598 intended anyway -- now it really works as intended).
12601 *) Move openssl.cnf out of lib/.
12604 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12605 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12606 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12607 [Ralf S. Engelschall]
12609 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12610 handle PKCS#7 enveloped data properly.
12611 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12613 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12614 copying pointers. The cert_st handling is changed by this in
12615 various ways (and thus what used to be known as ctx->default_cert
12616 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12617 any longer when s->cert does not give us what we need).
12618 ssl_cert_instantiate becomes obsolete by this change.
12619 As soon as we've got the new code right (possibly it already is?),
12620 we have solved a couple of bugs of the earlier code where s->cert
12621 was used as if it could not have been shared with other SSL structures.
12623 Note that using the SSL API in certain dirty ways now will result
12624 in different behaviour than observed with earlier library versions:
12625 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12626 does not influence s as it used to.
12628 In order to clean up things more thoroughly, inside SSL_SESSION
12629 we don't use CERT any longer, but a new structure SESS_CERT
12630 that holds per-session data (if available); currently, this is
12631 the peer's certificate chain and, for clients, the server's certificate
12632 and temporary key. CERT holds only those values that can have
12633 meaningful defaults in an SSL_CTX.
12636 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12637 from the internal representation. Various PKCS#7 fixes: remove some
12638 evil casts and set the enc_dig_alg field properly based on the signing
12642 *) Allow PKCS#12 password to be set from the command line or the
12643 environment. Let 'ca' get its config file name from the environment
12644 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12648 *) Allow certificate policies extension to use an IA5STRING for the
12649 organization field. This is contrary to the PKIX definition but
12650 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12654 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12655 without disallowing inline assembler and the like for non-pedantic builds.
12658 *) Support Borland C++ builder.
12659 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12661 *) Support Mingw32.
12664 *) SHA-1 cleanups and performance enhancements.
12665 [Andy Polyakov <appro@fy.chalmers.se>]
12667 *) Sparc v8plus assembler for the bignum library.
12668 [Andy Polyakov <appro@fy.chalmers.se>]
12670 *) Accept any -xxx and +xxx compiler options in Configure.
12673 *) Update HPUX configuration.
12676 *) Add missing sk_<type>_unshift() function to safestack.h
12677 [Ralf S. Engelschall]
12679 *) New function SSL_CTX_use_certificate_chain_file that sets the
12680 "extra_cert"s in addition to the certificate. (This makes sense
12681 only for "PEM" format files, as chains as a whole are not
12685 *) Support verify_depth from the SSL API.
12686 x509_vfy.c had what can be considered an off-by-one-error:
12687 Its depth (which was not part of the external interface)
12688 was actually counting the number of certificates in a chain;
12689 now it really counts the depth.
12692 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12693 instead of X509err, which often resulted in confusing error
12694 messages since the error codes are not globally unique
12695 (e.g. an alleged error in ssl3_accept when a certificate
12696 didn't match the private key).
12698 *) New function SSL_CTX_set_session_id_context that allows to set a default
12699 value (so that you don't need SSL_set_session_id_context for each
12700 connection using the SSL_CTX).
12703 *) OAEP decoding bug fix.
12706 *) Support INSTALL_PREFIX for package builders, as proposed by
12710 *) New Configure options "threads" and "no-threads". For systems
12711 where the proper compiler options are known (currently Solaris
12712 and Linux), "threads" is the default.
12715 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12718 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12719 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12720 such as /usr/local/bin.
12723 *) "make linux-shared" to build shared libraries.
12724 [Niels Poppe <niels@netbox.org>]
12726 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12729 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12730 extension adding in x509 utility.
12733 *) Remove NOPROTO sections and error code comments.
12736 *) Partial rewrite of the DEF file generator to now parse the ANSI
12740 *) New Configure options --prefix=DIR and --openssldir=DIR.
12743 *) Complete rewrite of the error code script(s). It is all now handled
12744 by one script at the top level which handles error code gathering,
12745 header rewriting and C source file generation. It should be much better
12746 than the old method: it now uses a modified version of Ulf's parser to
12747 read the ANSI prototypes in all header files (thus the old K&R definitions
12748 aren't needed for error creation any more) and do a better job of
12749 translating function codes into names. The old 'ASN1 error code embedded
12750 in a comment' is no longer necessary and it doesn't use .err files which
12751 have now been deleted. Also the error code call doesn't have to appear all
12752 on one line (which resulted in some large lines...).
12755 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12758 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12759 0 (which usually indicates a closed connection), but continue reading.
12762 *) Fix some race conditions.
12765 *) Add support for CRL distribution points extension. Add Certificate
12766 Policies and CRL distribution points documentation.
12769 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12772 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12773 8 of keying material. Merlin has also confirmed interop with this fix
12774 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12775 [Merlin Hughes <merlin@baltimore.ie>]
12777 *) Fix lots of warnings.
12778 [Richard Levitte <levitte@stacken.kth.se>]
12780 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12781 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12782 [Richard Levitte <levitte@stacken.kth.se>]
12784 *) Fix problems with sizeof(long) == 8.
12785 [Andy Polyakov <appro@fy.chalmers.se>]
12787 *) Change functions to ANSI C.
12790 *) Fix typos in error codes.
12791 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12793 *) Remove defunct assembler files from Configure.
12796 *) SPARC v8 assembler BIGNUM implementation.
12797 [Andy Polyakov <appro@fy.chalmers.se>]
12799 *) Support for Certificate Policies extension: both print and set.
12800 Various additions to support the r2i method this uses.
12803 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12804 return a const string when you are expecting an allocated buffer.
12807 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12808 types DirectoryString and DisplayText.
12811 *) Add code to allow r2i extensions to access the configuration database,
12812 add an LHASH database driver and add several ctx helper functions.
12815 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12816 fail when they extended the size of a BIGNUM.
12819 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12820 support typesafe stack.
12823 *) Fix typo in SSL_[gs]et_options().
12824 [Nils Frostberg <nils@medcom.se>]
12826 *) Delete various functions and files that belonged to the (now obsolete)
12827 old X509V3 handling code.
12830 *) New Configure option "rsaref".
12833 *) Don't auto-generate pem.h.
12836 *) Introduce type-safe ASN.1 SETs.
12839 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12840 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12842 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12843 that links with OpenSSL (well at least cause lots of warnings), but fear
12844 not: the conversion is trivial, and it eliminates loads of evil casts. A
12845 few STACKed things have been converted already. Feel free to convert more.
12846 In the fullness of time, I'll do away with the STACK type altogether.
12849 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12850 specified in <certfile> by updating the entry in the index.txt file.
12851 This way one no longer has to edit the index.txt file manually for
12852 revoking a certificate. The -revoke option does the gory details now.
12853 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12855 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12856 `-text' option at all and this way the `-noout -text' combination was
12857 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12858 [Ralf S. Engelschall]
12860 *) Make sure a corresponding plain text error message exists for the
12861 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12862 verify callback function determined that a certificate was revoked.
12863 [Ralf S. Engelschall]
12865 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12866 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12867 all available ciphers including rc5, which was forgotten until now.
12868 In order to let the testing shell script know which algorithms
12869 are available, a new (up to now undocumented) command
12870 "openssl list-cipher-commands" is used.
12873 *) Bugfix: s_client occasionally would sleep in select() when
12874 it should have checked SSL_pending() first.
12877 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12878 the raw DSA values prior to ASN.1 encoding.
12881 *) Tweaks to Configure
12882 [Niels Poppe <niels@netbox.org>]
12884 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12888 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12891 *) New config option to avoid instructions that are illegal on the 80386.
12892 The default code is faster, but requires at least a 486.
12895 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12896 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12897 same as SSL2_VERSION anyway.
12900 *) New "-showcerts" option for s_client.
12903 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12904 application. Various cleanups and fixes.
12907 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12908 modify error routines to work internally. Add error codes and PBE init
12909 to library startup routines.
12912 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12913 packing functions to asn1 and evp. Changed function names and error
12914 codes along the way.
12917 *) PKCS12 integration: and so it begins... First of several patches to
12918 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12919 objects to objects.h
12922 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12923 and display support for Thawte strong extranet extension.
12926 *) Add LinuxPPC support.
12927 [Jeff Dubrule <igor@pobox.org>]
12929 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12930 bn_div_words in alpha.s.
12931 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12933 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12934 OAEP isn't supported when OpenSSL is built with RSAref.
12935 [Ulf Moeller <ulf@fitug.de>]
12937 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12938 so they no longer are missing under -DNOPROTO.
12939 [Soren S. Jorvang <soren@t.dk>]
12942 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12944 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12945 doesn't work when the session is reused. Coming soon!
12948 *) Fix a security hole, that allows sessions to be reused in the wrong
12949 context thus bypassing client cert protection! All software that uses
12950 client certs and session caches in multiple contexts NEEDS PATCHING to
12951 allow session reuse! A fuller solution is in the works.
12952 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12954 *) Some more source tree cleanups (removed obsolete files
12955 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12956 permission on "config" script to be executable) and a fix for the INSTALL
12958 [Ulf Moeller <ulf@fitug.de>]
12960 *) Remove some legacy and erroneous uses of malloc, free instead of
12962 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12964 *) Make rsa_oaep_test return non-zero on error.
12965 [Ulf Moeller <ulf@fitug.de>]
12967 *) Add support for native Solaris shared libraries. Configure
12968 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12969 if someone would make that last step automatic.
12970 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12972 *) ctx_size was not built with the right compiler during "make links". Fixed.
12975 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12976 except NULL ciphers". This means the default cipher list will no longer
12977 enable NULL ciphers. They need to be specifically enabled e.g. with
12978 the string "DEFAULT:eNULL".
12981 *) Fix to RSA private encryption routines: if p < q then it would
12982 occasionally produce an invalid result. This will only happen with
12983 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12986 *) Be less restrictive and allow also `perl util/perlpath.pl
12987 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12988 because this way one can also use an interpreter named `perl5' (which is
12989 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12990 installed as `perl').
12991 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12993 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12994 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12996 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12997 advapi32.lib to Win32 build and change the pem test comparison
12998 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12999 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
13000 and crypto/des/ede_cbcm_enc.c.
13003 *) DES quad checksum was broken on big-endian architectures. Fixed.
13006 *) Comment out two functions in bio.h that aren't implemented. Fix up the
13007 Win32 test batch file so it (might) work again. The Win32 test batch file
13008 is horrible: I feel ill....
13011 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
13012 in e_os.h. Audit of header files to check ANSI and non ANSI
13013 sections: 10 functions were absent from non ANSI section and not exported
13014 from Windows DLLs. Fixed up libeay.num for new functions.
13017 *) Make `openssl version' output lines consistent.
13018 [Ralf S. Engelschall]
13020 *) Fix Win32 symbol export lists for BIO functions: Added
13021 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
13022 to ms/libeay{16,32}.def.
13023 [Ralf S. Engelschall]
13025 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
13026 fine under Unix and passes some trivial tests I've now added. But the
13027 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
13028 added to make sure no one expects that this stuff really works in the
13029 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
13030 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
13032 [Ralf S. Engelschall]
13034 *) Fix the generation of two part addresses in perl.
13035 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
13037 *) Add config entry for Linux on MIPS.
13038 [John Tobey <jtobey@channel1.com>]
13040 *) Make links whenever Configure is run, unless we are on Windoze.
13043 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
13044 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
13048 *) Add a useful kludge to allow package maintainers to specify compiler and
13049 other platforms details on the command line without having to patch the
13050 Configure script every time: One now can use ``perl Configure
13051 <id>:<details>'', i.e. platform ids are allowed to have details appended
13052 to them (separated by colons). This is treated as there would be a static
13053 pre-configured entry in Configure's %table under key <id> with value
13054 <details> and ``perl Configure <id>'' is called. So, when you want to
13055 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
13056 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
13057 now, which overrides the FreeBSD-elf entry on-the-fly.
13058 [Ralf S. Engelschall]
13060 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
13063 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
13064 on the `perl Configure ...' command line. This way one can compile
13065 OpenSSL libraries with Position Independent Code (PIC) which is needed
13066 for linking it into DSOs.
13067 [Ralf S. Engelschall]
13069 *) Remarkably, export ciphers were totally broken and no-one had noticed!
13073 *) Cleaned up the LICENSE document: The official contact for any license
13074 questions now is the OpenSSL core team under openssl-core@openssl.org.
13075 And add a paragraph about the dual-license situation to make sure people
13076 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
13077 to the OpenSSL toolkit.
13078 [Ralf S. Engelschall]
13080 *) General source tree makefile cleanups: Made `making xxx in yyy...'
13081 display consistent in the source tree and replaced `/bin/rm' by `rm'.
13082 Additionally cleaned up the `make links' target: Remove unnecessary
13083 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
13084 to speed processing and no longer clutter the display with confusing
13085 stuff. Instead only the actually done links are displayed.
13086 [Ralf S. Engelschall]
13088 *) Permit null encryption ciphersuites, used for authentication only. It used
13089 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
13090 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
13094 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
13095 signed attributes when verifying signatures (this would break them),
13096 the detached data encoding was wrong and public keys obtained using
13097 X509_get_pubkey() weren't freed.
13100 *) Add text documentation for the BUFFER functions. Also added a work around
13101 to a Win95 console bug. This was triggered by the password read stuff: the
13102 last character typed gets carried over to the next fread(). If you were
13103 generating a new cert request using 'req' for example then the last
13104 character of the passphrase would be CR which would then enter the first
13108 *) Added the new `Includes OpenSSL Cryptography Software' button as
13109 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
13110 button and can be used by applications based on OpenSSL to show the
13111 relationship to the OpenSSL project.
13112 [Ralf S. Engelschall]
13114 *) Remove confusing variables in function signatures in files
13115 ssl/ssl_lib.c and ssl/ssl.h.
13116 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13118 *) Don't install bss_file.c under PREFIX/include/
13119 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13121 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
13122 functions that return function pointers and has support for NT specific
13123 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
13124 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
13125 unsigned to signed types: this was killing the Win32 compile.
13128 *) Add new certificate file to stack functions,
13129 SSL_add_dir_cert_subjects_to_stack() and
13130 SSL_add_file_cert_subjects_to_stack(). These largely supplant
13131 SSL_load_client_CA_file(), and can be used to add multiple certs easily
13132 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
13133 This means that Apache-SSL and similar packages don't have to mess around
13134 to add as many CAs as they want to the preferred list.
13137 *) Experiment with doxygen documentation. Currently only partially applied to
13139 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
13140 openssl.doxy as the configuration file.
13143 *) Get rid of remaining C++-style comments which strict C compilers hate.
13144 [Ralf S. Engelschall, pointed out by Carlos Amengual]
13146 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
13147 compiled in by default: it has problems with large keys.
13150 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
13151 DH private keys and/or callback functions which directly correspond to
13152 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
13153 is needed for applications which have to configure certificates on a
13154 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
13156 For the RSA certificate situation is makes no difference, but
13157 for the DSA certificate situation this fixes the "no shared cipher"
13158 problem where the OpenSSL cipher selection procedure failed because the
13159 temporary keys were not overtaken from the context and the API provided
13160 no way to reconfigure them.
13161 The new functions now let applications reconfigure the stuff and they
13162 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
13163 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
13164 non-public-API function ssl_cert_instantiate() is used as a helper
13165 function and also to reduce code redundancy inside ssl_rsa.c.
13166 [Ralf S. Engelschall]
13168 *) Move s_server -dcert and -dkey options out of the undocumented feature
13169 area because they are useful for the DSA situation and should be
13170 recognized by the users.
13171 [Ralf S. Engelschall]
13173 *) Fix the cipher decision scheme for export ciphers: the export bits are
13174 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
13175 SSL_EXP_MASK. So, the original variable has to be used instead of the
13176 already masked variable.
13177 [Richard Levitte <levitte@stacken.kth.se>]
13179 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
13180 [Richard Levitte <levitte@stacken.kth.se>]
13182 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
13183 from `int' to `unsigned int' because it's a length and initialized by
13184 EVP_DigestFinal() which expects an `unsigned int *'.
13185 [Richard Levitte <levitte@stacken.kth.se>]
13187 *) Don't hard-code path to Perl interpreter on shebang line of Configure
13188 script. Instead use the usual Shell->Perl transition trick.
13189 [Ralf S. Engelschall]
13191 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
13192 (in addition to RSA certificates) to match the behaviour of `openssl dsa
13193 -noout -modulus' as it's already the case for `openssl rsa -noout
13194 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
13195 currently the public key is printed (a decision which was already done by
13196 `openssl dsa -modulus' in the past) which serves a similar purpose.
13197 Additionally the NO_RSA no longer completely removes the whole -modulus
13198 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
13200 [Ralf S. Engelschall]
13202 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
13203 BIO. See the source (crypto/evp/bio_ok.c) for more info.
13204 [Arne Ansper <arne@ats.cyber.ee>]
13206 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
13207 to be added. Now both 'req' and 'ca' can use new objects defined in the
13211 *) Add cool BIO that does syslog (or event log on NT).
13212 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
13214 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
13215 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
13216 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
13217 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
13220 *) Add preliminary config info for new extension code.
13223 *) Make RSA_NO_PADDING really use no padding.
13224 [Ulf Moeller <ulf@fitug.de>]
13226 *) Generate errors when private/public key check is done.
13229 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
13230 for some CRL extensions and new objects added.
13233 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
13234 key usage extension and fuller support for authority key id.
13237 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
13238 padding method for RSA, which is recommended for new applications in PKCS
13239 #1 v2.0 (RFC 2437, October 1998).
13240 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
13241 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
13242 against Bleichbacher's attack on RSA.
13243 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
13246 *) Updates to the new SSL compression code
13247 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13249 *) Fix so that the version number in the master secret, when passed
13250 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
13251 (because the server will not accept higher), that the version number
13252 is 0x03,0x01, not 0x03,0x00
13253 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13255 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
13256 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
13257 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
13260 *) Support for RAW extensions where an arbitrary extension can be
13261 created by including its DER encoding. See apps/openssl.cnf for
13265 *) Make sure latest Perl versions don't interpret some generated C array
13266 code as Perl array code in the crypto/err/err_genc.pl script.
13267 [Lars Weber <3weber@informatik.uni-hamburg.de>]
13269 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
13270 not many people have the assembler. Various Win32 compilation fixes and
13271 update to the INSTALL.W32 file with (hopefully) more accurate Win32
13272 build instructions.
13275 *) Modify configure script 'Configure' to automatically create crypto/date.h
13276 file under Win32 and also build pem.h from pem.org. New script
13277 util/mkfiles.pl to create the MINFO file on environments that can't do a
13278 'make files': perl util/mkfiles.pl >MINFO should work.
13281 *) Major rework of DES function declarations, in the pursuit of correctness
13282 and purity. As a result, many evil casts evaporated, and some weirdness,
13283 too. You may find this causes warnings in your code. Zapping your evil
13284 casts will probably fix them. Mostly.
13287 *) Fix for a typo in asn1.h. Bug fix to object creation script
13288 obj_dat.pl. It considered a zero in an object definition to mean
13289 "end of object": none of the objects in objects.h have any zeros
13290 so it wasn't spotted.
13291 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
13293 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
13294 Masking (CBCM). In the absence of test vectors, the best I have been able
13295 to do is check that the decrypt undoes the encrypt, so far. Send me test
13296 vectors if you have them.
13299 *) Correct calculation of key length for export ciphers (too much space was
13300 allocated for null ciphers). This has not been tested!
13303 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
13304 message is now correct (it understands "crypto" and "ssl" on its
13305 command line). There is also now an "update" option. This will update
13306 the util/ssleay.num and util/libeay.num files with any new functions.
13308 perl util/mkdef.pl crypto ssl update
13309 it will update them.
13312 *) Overhauled the Perl interface (perl/*):
13313 - ported BN stuff to OpenSSL's different BN library
13314 - made the perl/ source tree CVS-aware
13315 - renamed the package from SSLeay to OpenSSL (the files still contain
13316 their history because I've copied them in the repository)
13317 - removed obsolete files (the test scripts will be replaced
13318 by better Test::Harness variants in the future)
13319 [Ralf S. Engelschall]
13321 *) First cut for a very conservative source tree cleanup:
13322 1. merge various obsolete readme texts into doc/ssleay.txt
13323 where we collect the old documents and readme texts.
13324 2. remove the first part of files where I'm already sure that we no
13325 longer need them because of three reasons: either they are just temporary
13326 files which were left by Eric or they are preserved original files where
13327 I've verified that the diff is also available in the CVS via "cvs diff
13328 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
13329 the crypto/md/ stuff).
13330 [Ralf S. Engelschall]
13332 *) More extension code. Incomplete support for subject and issuer alt
13333 name, issuer and authority key id. Change the i2v function parameters
13334 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
13335 what that's for :-) Fix to ASN1 macro which messed up
13336 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
13339 *) Preliminary support for ENUMERATED type. This is largely copied from the
13343 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
13344 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13346 *) Make sure `make rehash' target really finds the `openssl' program.
13347 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13349 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
13350 like to hear about it if this slows down other processors.
13353 *) Add CygWin32 platform information to Configure script.
13354 [Alan Batie <batie@aahz.jf.intel.com>]
13356 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
13357 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13359 *) New program nseq to manipulate netscape certificate sequences
13362 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13366 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13367 but the BN code had some problems that would cause failures when
13368 doing certificate verification and some other functions.
13369 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13371 *) Add ASN1 and PEM code to support netscape certificate sequences.
13374 *) Add ASN1 and PEM code to support netscape certificate sequences.
13377 *) Add several PKIX and private extended key usage OIDs.
13380 *) Modify the 'ca' program to handle the new extension code. Modify
13381 openssl.cnf for new extension format, add comments.
13384 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13385 and add a sample to openssl.cnf so req -x509 now adds appropriate
13389 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13390 error code, add initial support to X509_print() and x509 application.
13393 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13394 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13395 stuff is currently isolated and isn't even compiled yet.
13398 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13399 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13400 Removed the versions check from X509 routines when loading extensions:
13401 this allows certain broken certificates that don't set the version
13402 properly to be processed.
13405 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13406 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13407 can still be regenerated with "make depend".
13410 *) Spelling mistake in C version of CAST-128.
13411 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13413 *) Changes to the error generation code. The perl script err-code.pl
13414 now reads in the old error codes and retains the old numbers, only
13415 adding new ones if necessary. It also only changes the .err files if new
13416 codes are added. The makefiles have been modified to only insert errors
13417 when needed (to avoid needlessly modifying header files). This is done
13418 by only inserting errors if the .err file is newer than the auto generated
13419 C file. To rebuild all the error codes from scratch (the old behaviour)
13420 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13421 or delete all the .err files.
13424 *) CAST-128 was incorrectly implemented for short keys. The C version has
13425 been fixed, but is untested. The assembler versions are also fixed, but
13426 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13427 to regenerate it if needed.
13428 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13429 Hagino <itojun@kame.net>]
13431 *) File was opened incorrectly in randfile.c.
13432 [Ulf Möller <ulf@fitug.de>]
13434 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13435 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13436 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13437 al: it's just almost always a UTCTime. Note this patch adds new error
13438 codes so do a "make errors" if there are problems.
13441 *) Correct Linux 1 recognition in config.
13442 [Ulf Möller <ulf@fitug.de>]
13444 *) Remove pointless MD5 hash when using DSA keys in ca.
13445 [Anonymous <nobody@replay.com>]
13447 *) Generate an error if given an empty string as a cert directory. Also
13448 generate an error if handed NULL (previously returned 0 to indicate an
13449 error, but didn't set one).
13450 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13452 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13455 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13456 parameters. This was causing a warning which killed off the Win32 compile.
13459 *) Remove C++ style comments from crypto/bn/bn_local.h.
13460 [Neil Costigan <neil.costigan@celocom.com>]
13462 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13463 based on a text string, looking up short and long names and finally
13464 "dot" format. The "dot" format stuff didn't work. Added new function
13465 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13466 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13467 OID is not part of the table.
13470 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13471 X509_LOOKUP_by_alias().
13474 *) Sort openssl functions by name.
13477 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13478 encryption from sample DSA keys (in case anyone is interested the password
13482 *) Make _all_ *_free functions accept a NULL pointer.
13483 [Frans Heymans <fheymans@isaserver.be>]
13485 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13487 [Anonymous <nobody@replay.com>]
13489 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13490 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13492 *) Don't blow it for numeric -newkey arguments to apps/req.
13493 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13495 *) Temp key "for export" tests were wrong in s3_srvr.c.
13496 [Anonymous <nobody@replay.com>]
13498 *) Add prototype for temp key callback functions
13499 SSL_CTX_set_tmp_{rsa,dh}_callback().
13502 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13503 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13506 *) X509_name_add_entry() freed the wrong thing after an error.
13507 [Arne Ansper <arne@ats.cyber.ee>]
13509 *) rsa_eay.c would attempt to free a NULL context.
13510 [Arne Ansper <arne@ats.cyber.ee>]
13512 *) BIO_s_socket() had a broken should_retry() on Windoze.
13513 [Arne Ansper <arne@ats.cyber.ee>]
13515 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13516 [Arne Ansper <arne@ats.cyber.ee>]
13518 *) Make sure the already existing X509_STORE->depth variable is initialized
13519 in X509_STORE_new(), but document the fact that this variable is still
13520 unused in the certificate verification process.
13521 [Ralf S. Engelschall]
13523 *) Fix the various library and apps files to free up pkeys obtained from
13524 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13527 *) Fix reference counting in X509_PUBKEY_get(). This makes
13528 demos/maurice/example2.c work, amongst others, probably.
13529 [Steve Henson and Ben Laurie]
13531 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13532 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13533 are no longer created. This way we have a single and consistent command
13534 line interface `openssl <command>', similar to `cvs <command>'.
13535 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13537 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13538 BIT STRING wrapper always have zero unused bits.
13541 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13544 *) Make the top-level INSTALL documentation easier to understand.
13547 *) Makefiles updated to exit if an error occurs in a sub-directory
13548 make (including if user presses ^C) [Paul Sutton]
13550 *) Make Montgomery context stuff explicit in RSA data structure.
13553 *) Fix build order of pem and err to allow for generated pem.h.
13556 *) Fix renumbering bug in X509_NAME_delete_entry().
13559 *) Enhanced the err-ins.pl script so it makes the error library number
13560 global and can add a library name. This is needed for external ASN1 and
13561 other error libraries.
13564 *) Fixed sk_insert which never worked properly.
13567 *) Fix ASN1 macros so they can handle indefinite length constructed
13568 EXPLICIT tags. Some non standard certificates use these: they can now
13572 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13573 into a single doc/ssleay.txt bundle. This way the information is still
13574 preserved but no longer messes up this directory. Now it's new room for
13575 the new set of documentation files.
13576 [Ralf S. Engelschall]
13578 *) SETs were incorrectly DER encoded. This was a major pain, because they
13579 shared code with SEQUENCEs, which aren't coded the same. This means that
13580 almost everything to do with SETs or SEQUENCEs has either changed name or
13581 number of arguments.
13582 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13584 *) Fix test data to work with the above.
13587 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13588 was already fixed by Eric for 0.9.1 it seems.
13589 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13591 *) Autodetect FreeBSD3.
13594 *) Fix various bugs in Configure. This affects the following platforms:
13598 unixware-2.0-pentium
13602 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13603 before they are needed.
13606 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13610 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13612 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13613 changed SSLeay to OpenSSL in version strings.
13614 [Ralf S. Engelschall]
13616 *) Some fixups to the top-level documents.
13619 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13620 because the symlink to include/ was missing.
13621 [Ralf S. Engelschall]
13623 *) Incorporated the popular no-RSA/DSA-only patches
13624 which allow to compile a RSA-free SSLeay.
13625 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13627 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13628 when "ssleay" is still not found.
13629 [Ralf S. Engelschall]
13631 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13632 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13634 *) Updated the README file.
13635 [Ralf S. Engelschall]
13637 *) Added various .cvsignore files in the CVS repository subdirs
13638 to make a "cvs update" really silent.
13639 [Ralf S. Engelschall]
13641 *) Recompiled the error-definition header files and added
13642 missing symbols to the Win32 linker tables.
13643 [Ralf S. Engelschall]
13645 *) Cleaned up the top-level documents;
13646 o new files: CHANGES and LICENSE
13647 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13648 o merged COPYRIGHT into LICENSE
13649 o removed obsolete TODO file
13650 o renamed MICROSOFT to INSTALL.W32
13651 [Ralf S. Engelschall]
13653 *) Removed dummy files from the 0.9.1b source tree:
13654 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13655 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13656 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13657 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13658 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13659 [Ralf S. Engelschall]
13661 *) Added various platform portability fixes.
13664 *) The Genesis of the OpenSSL rpject:
13665 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13666 Young and Tim J. Hudson created while they were working for C2Net until
13668 [The OpenSSL Project]
13671 Changes between 0.9.0b and 0.9.1b [not released]
13673 *) Updated a few CA certificates under certs/
13676 *) Changed some BIGNUM api stuff.
13679 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13680 DGUX x86, Linux Alpha, etc.
13683 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13684 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13688 *) Add -strparse option to asn1pars program which parses nested
13690 [Dr Stephen Henson <shenson@bigfoot.com>]
13692 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13695 *) DSA fix for "ca" program.
13698 *) Added "-genkey" option to "dsaparam" program.
13701 *) Added RIPE MD160 (rmd160) message digest.
13704 *) Added -a (all) option to "ssleay version" command.
13707 *) Added PLATFORM define which is the id given to Configure.
13710 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13713 *) Extended the ASN.1 parser routines.
13716 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13719 *) Added a BN_CTX to the BN library.
13722 *) Fixed the weak key values in DES library
13725 *) Changed API in EVP library for cipher aliases.
13728 *) Added support for RC2/64bit cipher.
13731 *) Converted the lhash library to the crypto/mem.c functions.
13734 *) Added more recognized ASN.1 object ids.
13737 *) Added more RSA padding checks for SSL/TLS.
13740 *) Added BIO proxy/filter functionality.
13743 *) Added extra_certs to SSL_CTX which can be used
13744 send extra CA certificates to the client in the CA cert chain sending
13745 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13748 *) Now Fortezza is denied in the authentication phase because
13749 this is key exchange mechanism is not supported by SSLeay at all.
13752 *) Additional PKCS1 checks.
13755 *) Support the string "TLSv1" for all TLS v1 ciphers.
13758 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13759 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13762 *) Fixed a few memory leaks.
13765 *) Fixed various code and comment typos.
13768 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13769 bytes sent in the client random.
13770 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob