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 1.1.1o and 1.1.1p [21 Jun 2022]
12 *) In addition to the c_rehash shell command injection identified in
13 CVE-2022-1292, further bugs where the c_rehash script does not
14 properly sanitise shell metacharacters to prevent command injection have been
17 When the CVE-2022-1292 was fixed it was not discovered that there
18 are other places in the script where the file names of certificates
19 being hashed were possibly passed to a command executed through the shell.
21 This script is distributed by some operating systems in a manner where
22 it is automatically executed. On such operating systems, an attacker
23 could execute arbitrary commands with the privileges of the script.
25 Use of the c_rehash script is considered obsolete and should be replaced
26 by the OpenSSL rehash command line tool.
28 [Daniel Fiala, Tomáš Mráz]
30 *) When OpenSSL TLS client is connecting without any supported elliptic
31 curves and TLS-1.3 protocol is disabled the connection will no longer fail
32 if a ciphersuite that does not use a key exchange based on elliptic
33 curves can be negotiated.
36 Changes between 1.1.1n and 1.1.1o [3 May 2022]
38 *) Fixed a bug in the c_rehash script which was not properly sanitising shell
39 metacharacters to prevent command injection. This script is distributed
40 by some operating systems in a manner where it is automatically executed.
41 On such operating systems, an attacker could execute arbitrary commands
42 with the privileges of the script.
44 Use of the c_rehash script is considered obsolete and should be replaced
45 by the OpenSSL rehash command line tool.
49 Changes between 1.1.1m and 1.1.1n [15 Mar 2022]
51 *) Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
54 Internally this function is used when parsing certificates that contain
55 elliptic curve public keys in compressed form or explicit elliptic curve
56 parameters with a base point encoded in compressed form.
58 It is possible to trigger the infinite loop by crafting a certificate that
59 has invalid explicit curve parameters.
61 Since certificate parsing happens prior to verification of the certificate
62 signature, any process that parses an externally supplied certificate may
63 thus be subject to a denial of service attack. The infinite loop can also
64 be reached when parsing crafted private keys as they can contain explicit
65 elliptic curve parameters.
67 Thus vulnerable situations include:
69 - TLS clients consuming server certificates
70 - TLS servers consuming client certificates
71 - Hosting providers taking certificates or private keys from customers
72 - Certificate authorities parsing certification requests from subscribers
73 - Anything else which parses ASN.1 elliptic curve parameters
75 Also any other applications that use the BN_mod_sqrt() where the attacker
76 can control the parameter values are vulnerable to this DoS issue.
80 *) Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
81 to the list of ciphersuites providing Perfect Forward Secrecy as
82 required by SECLEVEL >= 3.
84 [Dmitry Belyavskiy, Nicola Tuveri]
86 Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
88 *) Avoid loading of a dynamic engine twice.
92 *) Fixed building on Debian with kfreebsd kernels
96 *) Prioritise DANE TLSA issuer certs over peer certs
100 *) Fixed random API for MacOS prior to 10.12
102 These MacOS versions don't support the CommonCrypto APIs
106 Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
108 *) Fixed an SM2 Decryption Buffer Overflow.
110 In order to decrypt SM2 encrypted data an application is expected to call the
111 API function EVP_PKEY_decrypt(). Typically an application will call this
112 function twice. The first time, on entry, the "out" parameter can be NULL and,
113 on exit, the "outlen" parameter is populated with the buffer size required to
114 hold the decrypted plaintext. The application can then allocate a sufficiently
115 sized buffer and call EVP_PKEY_decrypt() again, but this time passing a non-NULL
116 value for the "out" parameter.
118 A bug in the implementation of the SM2 decryption code means that the
119 calculation of the buffer size required to hold the plaintext returned by the
120 first call to EVP_PKEY_decrypt() can be smaller than the actual size required by
121 the second call. This can lead to a buffer overflow when EVP_PKEY_decrypt() is
122 called by the application a second time with a buffer that is too small.
124 A malicious attacker who is able present SM2 content for decryption to an
125 application could cause attacker chosen data to overflow the buffer by up to a
126 maximum of 62 bytes altering the contents of other data held after the
127 buffer, possibly changing application behaviour or causing the application to
128 crash. The location of the buffer is application dependent but is typically
133 *) Fixed various read buffer overruns processing ASN.1 strings
135 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
136 structure which contains a buffer holding the string data and a field holding
137 the buffer length. This contrasts with normal C strings which are repesented as
138 a buffer for the string data which is terminated with a NUL (0) byte.
140 Although not a strict requirement, ASN.1 strings that are parsed using OpenSSL's
141 own "d2i" functions (and other similar parsing functions) as well as any string
142 whose value has been set with the ASN1_STRING_set() function will additionally
143 NUL terminate the byte array in the ASN1_STRING structure.
145 However, it is possible for applications to directly construct valid ASN1_STRING
146 structures which do not NUL terminate the byte array by directly setting the
147 "data" and "length" fields in the ASN1_STRING array. This can also happen by
148 using the ASN1_STRING_set0() function.
150 Numerous OpenSSL functions that print ASN.1 data have been found to assume that
151 the ASN1_STRING byte array will be NUL terminated, even though this is not
152 guaranteed for strings that have been directly constructed. Where an application
153 requests an ASN.1 structure to be printed, and where that ASN.1 structure
154 contains ASN1_STRINGs that have been directly constructed by the application
155 without NUL terminating the "data" field, then a read buffer overrun can occur.
157 The same thing can also occur during name constraints processing of certificates
158 (for example if a certificate has been directly constructed by the application
159 instead of loading it via the OpenSSL parsing functions, and the certificate
160 contains non NUL terminated ASN1_STRING structures). It can also occur in the
161 X509_get1_email(), X509_REQ_get1_email() and X509_get1_ocsp() functions.
163 If a malicious actor can cause an application to directly construct an
164 ASN1_STRING and then process it through one of the affected OpenSSL functions
165 then this issue could be hit. This might result in a crash (causing a Denial of
166 Service attack). It could also result in the disclosure of private memory
167 contents (such as private keys, or sensitive plaintext).
171 Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
173 *) Fixed a problem with verifying a certificate chain when using the
174 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks
175 of the certificates present in a certificate chain. It is not set by
178 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
179 the chain that have explicitly encoded elliptic curve parameters was added
180 as an additional strict check.
182 An error in the implementation of this check meant that the result of a
183 previous check to confirm that certificates in the chain are valid CA
184 certificates was overwritten. This effectively bypasses the check
185 that non-CA certificates must not be able to issue other certificates.
187 If a "purpose" has been configured then there is a subsequent opportunity
188 for checks that the certificate is a valid CA. All of the named "purpose"
189 values implemented in libcrypto perform this check. Therefore, where
190 a purpose is set the certificate chain will still be rejected even when the
191 strict flag has been used. A purpose is set by default in libssl client and
192 server certificate verification routines, but it can be overridden or
193 removed by an application.
195 In order to be affected, an application must explicitly set the
196 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
197 for the certificate verification or, in the case of TLS client or server
198 applications, override the default purpose.
202 *) Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
203 crafted renegotiation ClientHello message from a client. If a TLSv1.2
204 renegotiation ClientHello omits the signature_algorithms extension (where
205 it was present in the initial ClientHello), but includes a
206 signature_algorithms_cert extension then a NULL pointer dereference will
207 result, leading to a crash and a denial of service attack.
209 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
210 (which is the default configuration). OpenSSL TLS clients are not impacted
213 [Peter Kästle and Samuel Sapalski]
215 Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
217 *) Fixed the X509_issuer_and_serial_hash() function. It attempts to
218 create a unique hash value based on the issuer and serial number data
219 contained within an X509 certificate. However it was failing to correctly
220 handle any errors that may occur while parsing the issuer field (which might
221 occur if the issuer field is maliciously constructed). This may subsequently
222 result in a NULL pointer deref and a crash leading to a potential denial of
227 *) Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
228 padding mode to correctly check for rollback attacks. This is considered a
229 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
233 *) Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
234 functions. Previously they could overflow the output length argument in some
235 cases where the input length is close to the maximum permissable length for
236 an integer on the platform. In such cases the return value from the function
237 call would be 1 (indicating success), but the output length value would be
238 negative. This could cause applications to behave incorrectly or crash.
242 *) Fixed SRP_Calc_client_key so that it runs in constant time. The previous
243 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
244 could be exploited in a side channel attack to recover the password. Since
245 the attack is local host only this is outside of the current OpenSSL
246 threat model and therefore no CVE is assigned.
248 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
252 Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
254 *) Fixed NULL pointer deref in the GENERAL_NAME_cmp function
255 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
256 If an attacker can control both items being compared then this could lead
257 to a possible denial of service attack. OpenSSL itself uses the
258 GENERAL_NAME_cmp function for two purposes:
259 1) Comparing CRL distribution point names between an available CRL and a
260 CRL distribution point embedded in an X509 certificate
261 2) When verifying that a timestamp response token signer matches the
262 timestamp authority name (exposed via the API functions
263 TS_RESP_verify_response and TS_RESP_verify_token)
267 *) Add support for Apple Silicon M1 Macs with the darwin64-arm64-cc target.
270 *) The security callback, which can be customised by application code, supports
271 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
272 in the "other" parameter. In most places this is what is passed. All these
273 places occur server side. However there was one client side call of this
274 security operation and it passed a DH object instead. This is incorrect
275 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
276 of the other locations. Therefore this client side call has been changed to
277 pass an EVP_PKEY instead.
280 *) In 1.1.1h, an expired trusted (root) certificate was not anymore rejected
281 when validating a certificate path. This check is restored in 1.1.1i.
284 Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
286 *) Certificates with explicit curve parameters are now disallowed in
287 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
290 *) The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
291 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
292 conversely, silently ignore DTLS protocol version bounds when configuring
293 TLS-based contexts. The commands can be repeated to set bounds of both
294 types. The same applies with the corresponding "min_protocol" and
295 "max_protocol" command-line switches, in case some application uses both TLS
298 SSL_CTX instances that are created for a fixed protocol version (e.g.
299 TLSv1_server_method()) also silently ignore version bounds. Previously
300 attempts to apply bounds to these protocol versions would result in an
301 error. Now only the "version-flexible" SSL_CTX instances are subject to
302 limits in configuration files in command-line options.
305 *) Handshake now fails if Extended Master Secret extension is dropped
309 *) Accidentally, an expired trusted (root) certificate is not anymore rejected
310 when validating a certificate path.
313 *) The Oracle Developer Studio compiler will start reporting deprecated APIs
315 Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
317 *) Fixed segmentation fault in SSL_check_chain()
318 Server or client applications that call the SSL_check_chain() function
319 during or after a TLS 1.3 handshake may crash due to a NULL pointer
320 dereference as a result of incorrect handling of the
321 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
322 or unrecognised signature algorithm is received from the peer. This could
323 be exploited by a malicious peer in a Denial of Service attack.
327 *) Added AES consttime code for no-asm configurations
328 an optional constant time support for AES was added
329 when building openssl for no-asm.
330 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
331 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
332 At this time this feature is by default disabled.
333 It will be enabled by default in 3.0.
336 Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
338 *) Revert the change of EOF detection while reading in libssl to avoid
339 regressions in applications depending on the current way of reporting
340 the EOF. As the existing method is not fully accurate the change to
341 reporting the EOF via SSL_ERROR_SSL is kept on the current development
342 branch and will be present in the 3.0 release.
345 *) Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
346 when primes for RSA keys are computed.
347 Since we previously always generated primes == 2 (mod 3) for RSA keys,
348 the 2-prime and 3-prime RSA modules were easy to distinguish, since
349 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
350 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
351 This avoids possible fingerprinting of newly generated RSA modules.
354 Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
355 *) Properly detect EOF while reading in libssl. Previously if we hit an EOF
356 while reading in libssl then we would report an error back to the
357 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
358 an error to the stack (which means we instead return SSL_ERROR_SSL) and
359 therefore give a hint as to what went wrong.
362 *) Check that ed25519 and ed448 are allowed by the security level. Previously
363 signature algorithms not using an MD were not being checked that they were
364 allowed by the security level.
367 *) Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
368 was not quite right. The behaviour was not consistent between resumption
369 and normal handshakes, and also not quite consistent with historical
370 behaviour. The behaviour in various scenarios has been clarified and
371 it has been updated to make it match historical behaviour as closely as
375 *) [VMS only] The header files that the VMS compilers include automatically,
376 __DECC_INCLUDE_PROLOGUE.H and __DECC_INCLUDE_EPILOGUE.H, use pragmas that
377 the C++ compiler doesn't understand. This is a shortcoming in the
378 compiler, but can be worked around with __cplusplus guards.
380 C++ applications that use OpenSSL libraries must be compiled using the
381 qualifier '/NAMES=(AS_IS,SHORTENED)' to be able to use all the OpenSSL
382 functions. Otherwise, only functions with symbols of less than 31
383 characters can be used, as the linker will not be able to successfully
384 resolve symbols with longer names.
387 *) Corrected the documentation of the return values from the EVP_DigestSign*
388 set of functions. The documentation mentioned negative values for some
389 errors, but this was never the case, so the mention of negative values
392 Code that followed the documentation and thereby check with something
393 like 'EVP_DigestSignInit(...) <= 0' will continue to work undisturbed.
396 *) Fixed an an overflow bug in the x64_64 Montgomery squaring procedure
397 used in exponentiation with 512-bit moduli. No EC algorithms are
398 affected. Analysis suggests that attacks against 2-prime RSA1024,
399 3-prime RSA1536, and DSA1024 as a result of this defect would be very
400 difficult to perform and are not believed likely. Attacks against DH512
401 are considered just feasible. However, for an attack the target would
402 have to re-use the DH512 private key, which is not recommended anyway.
403 Also applications directly using the low level API BN_mod_exp may be
404 affected if they use BN_FLG_CONSTTIME.
408 *) Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
409 The presence of this system service is determined at run-time.
412 *) Added newline escaping functionality to a filename when using openssl dgst.
413 This output format is to replicate the output format found in the '*sum'
414 checksum programs. This aims to preserve backward compatibility.
415 [Matt Eaton, Richard Levitte, and Paul Dale]
417 *) Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
421 Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
423 *) Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
424 number generator (RNG). This was intended to include protection in the
425 event of a fork() system call in order to ensure that the parent and child
426 processes did not share the same RNG state. However this protection was not
427 being used in the default case.
429 A partial mitigation for this issue is that the output from a high
430 precision timer is mixed into the RNG state so the likelihood of a parent
431 and child process sharing state is significantly reduced.
433 If an application already calls OPENSSL_init_crypto() explicitly using
434 OPENSSL_INIT_ATFORK then this problem does not occur at all.
436 [Matthias St. Pierre]
438 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is
439 used even when parsing explicit parameters, when loading a serialized key
440 or calling `EC_GROUP_new_from_ecpkparameters()`/
441 `EC_GROUP_new_from_ecparameters()`.
442 This prevents bypass of security hardening and performance gains,
443 especially for curves with specialized EC_METHODs.
444 By default, if a key encoded with explicit parameters is loaded and later
445 serialized, the output is still encoded with explicit parameters, even if
446 internally a "named" EC_GROUP is used for computation.
449 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before
450 this change, EC_GROUP_set_generator would accept order and/or cofactor as
451 NULL. After this change, only the cofactor parameter can be NULL. It also
452 does some minimal sanity checks on the passed order.
456 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
457 An attack is simple, if the first CMS_recipientInfo is valid but the
458 second CMS_recipientInfo is chosen ciphertext. If the second
459 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
460 encryption key will be replaced by garbage, and the message cannot be
461 decoded, but if the RSA decryption fails, the correct encryption key is
462 used and the recipient will not notice the attack.
463 As a work around for this potential attack the length of the decrypted
464 key must be equal to the cipher default key length, in case the
465 certifiate is not given and all recipientInfo are tried out.
466 The old behaviour can be re-enabled in the CMS code by setting the
467 CMS_DEBUG_DECRYPT flag.
471 *) Early start up entropy quality from the DEVRANDOM seed source has been
472 improved for older Linux systems. The RAND subsystem will wait for
473 /dev/random to be producing output before seeding from /dev/urandom.
474 The seeded state is stored for future library initialisations using
475 a system global shared memory segment. The shared memory identifier
476 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
477 the desired value. The default identifier is 114.
480 *) Correct the extended master secret constant on EBCDIC systems. Without this
481 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
482 negotiate EMS will fail. Unfortunately this also means that TLS connections
483 between EBCDIC systems with this fix, and EBCDIC systems without this
484 fix will fail if they negotiate EMS.
487 *) Use Windows installation paths in the mingw builds
489 Mingw isn't a POSIX environment per se, which means that Windows
490 paths should be used for installation.
494 *) Changed DH_check to accept parameters with order q and 2q subgroups.
495 With order 2q subgroups the bit 0 of the private key is not secret
496 but DH_generate_key works around that by clearing bit 0 of the
497 private key for those. This avoids leaking bit 0 of the private key.
500 *) Significantly reduce secure memory usage by the randomness pools.
503 *) Revert the DEVRANDOM_WAIT feature for Linux systems
505 The DEVRANDOM_WAIT feature added a select() call to wait for the
506 /dev/random device to become readable before reading from the
509 It turned out that this change had negative side effects on
510 performance which were not acceptable. After some discussion it
511 was decided to revert this feature and leave it up to the OS
512 resp. the platform maintainer to ensure a proper initialization
513 during early boot time.
514 [Matthias St. Pierre]
516 Changes between 1.1.1b and 1.1.1c [28 May 2019]
518 *) Add build tests for C++. These are generated files that only do one
519 thing, to include one public OpenSSL head file each. This tests that
520 the public header files can be usefully included in a C++ application.
522 This test isn't enabled by default. It can be enabled with the option
523 'enable-buildtest-c++'.
526 *) Enable SHA3 pre-hashing for ECDSA and DSA.
529 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
530 This changes the size when using the genpkey app when no size is given. It
531 fixes an omission in earlier changes that changed all RSA, DSA and DH
532 generation apps to use 2048 bits by default.
535 *) Reorganize the manual pages to consistently have RETURN VALUES,
536 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
537 util/fix-doc-nits accordingly.
538 [Paul Yang, Joshua Lock]
540 *) Add the missing accessor EVP_PKEY_get0_engine()
543 *) Have apps like 's_client' and 's_server' output the signature scheme
544 along with other cipher suite parameters when debugging.
547 *) Make OPENSSL_config() error agnostic again.
550 *) Do the error handling in RSA decryption constant time.
553 *) Prevent over long nonces in ChaCha20-Poly1305.
555 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
556 for every encryption operation. RFC 7539 specifies that the nonce value
557 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
558 and front pads the nonce with 0 bytes if it is less than 12
559 bytes. However it also incorrectly allows a nonce to be set of up to 16
560 bytes. In this case only the last 12 bytes are significant and any
561 additional leading bytes are ignored.
563 It is a requirement of using this cipher that nonce values are
564 unique. Messages encrypted using a reused nonce value are susceptible to
565 serious confidentiality and integrity attacks. If an application changes
566 the default nonce length to be longer than 12 bytes and then makes a
567 change to the leading bytes of the nonce expecting the new value to be a
568 new unique nonce then such an application could inadvertently encrypt
569 messages with a reused nonce.
571 Additionally the ignored bytes in a long nonce are not covered by the
572 integrity guarantee of this cipher. Any application that relies on the
573 integrity of these ignored leading bytes of a long nonce may be further
574 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
575 is safe because no such use sets such a long nonce value. However user
576 applications that use this cipher directly and set a non-default nonce
577 length to be longer than 12 bytes may be vulnerable.
579 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
584 *) Add DEVRANDOM_WAIT feature for Linux systems
586 On older Linux systems where the getrandom() system call is not available,
587 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
588 Contrary to getrandom(), the /dev/urandom device will not block during
589 early boot when the kernel CSPRNG has not been seeded yet.
591 To mitigate this known weakness, use select() to wait for /dev/random to
592 become readable before reading from /dev/urandom.
594 *) Ensure that SM2 only uses SM3 as digest algorithm
597 Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
599 *) Added SCA hardening for modular field inversion in EC_GROUP through
600 a new dedicated field_inv() pointer in EC_METHOD.
601 This also addresses a leakage affecting conversions from projective
602 to affine coordinates.
603 [Billy Bob Brumley, Nicola Tuveri]
605 *) Change the info callback signals for the start and end of a post-handshake
606 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
607 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
608 confused by this and assume that a TLSv1.2 renegotiation has started. This
609 can break KeyUpdate handling. Instead we no longer signal the start and end
610 of a post handshake message exchange (although the messages themselves are
611 still signalled). This could break some applications that were expecting
612 the old signals. However without this KeyUpdate is not usable for many
616 *) Fix a bug in the computation of the endpoint-pair shared secret used
617 by DTLS over SCTP. This breaks interoperability with older versions
618 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
619 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
620 interoperability with such broken implementations. However, enabling
621 this switch breaks interoperability with correct implementations.
623 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
624 re-used X509_PUBKEY object if the second PUBKEY is malformed.
627 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
630 *) Remove the 'dist' target and add a tarball building script. The
631 'dist' target has fallen out of use, and it shouldn't be
632 necessary to configure just to create a source distribution.
635 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
637 *) Timing vulnerability in DSA signature generation
639 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
640 timing side channel attack. An attacker could use variations in the signing
641 algorithm to recover the private key.
643 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
647 *) Timing vulnerability in ECDSA signature generation
649 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
650 timing side channel attack. An attacker could use variations in the signing
651 algorithm to recover the private key.
653 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
657 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
658 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
659 are retained for backwards compatibility.
662 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
663 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
664 of two gigabytes and the error handling improved.
666 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
667 categorized as a normal bug, not a security issue, because the DRBG reseeds
668 automatically and is fully functional even without additional randomness
669 provided by the application.
671 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
673 *) Add a new ClientHello callback. Provides a callback interface that gives
674 the application the ability to adjust the nascent SSL object at the
675 earliest stage of ClientHello processing, immediately after extensions have
676 been collected but before they have been processed. In particular, this
677 callback can adjust the supported TLS versions in response to the contents
681 *) Add SM2 base algorithm support.
684 *) s390x assembly pack: add (improved) hardware-support for the following
685 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
686 aes-cfb/cfb8, aes-ecb.
689 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
690 parameter is no longer accepted, as it leads to a corrupt table. NULL
691 pem_str is reserved for alias entries only.
694 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
695 step for prime curves. The new implementation is based on formulae from
696 differential addition-and-doubling in homogeneous projective coordinates
697 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
698 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
699 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
700 to work in projective coordinates.
701 [Billy Bob Brumley, Nicola Tuveri]
703 *) Change generating and checking of primes so that the error rate of not
704 being prime depends on the intended use based on the size of the input.
705 For larger primes this will result in more rounds of Miller-Rabin.
706 The maximal error rate for primes with more than 1080 bits is lowered
708 [Kurt Roeckx, Annie Yousar]
710 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
713 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
714 moving between systems, and to avoid confusion when a Windows build is
715 done with mingw vs with MSVC. For POSIX installs, there's still a
716 symlink or copy named 'tsget' to avoid that confusion as well.
719 *) Revert blinding in ECDSA sign and instead make problematic addition
720 length-invariant. Switch even to fixed-length Montgomery multiplication.
723 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
724 step for binary curves. The new implementation is based on formulae from
725 differential addition-and-doubling in mixed Lopez-Dahab projective
726 coordinates, modified to independently blind the operands.
727 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
729 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
730 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
731 EC_METHODs to implement their own specialized "ladder step", to take
732 advantage of more favorable coordinate systems or more efficient
733 differential addition-and-doubling algorithms.
734 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
736 *) Modified the random device based seed sources to keep the relevant
737 file descriptors open rather than reopening them on each access.
738 This allows such sources to operate in a chroot() jail without
739 the associated device nodes being available. This behaviour can be
740 controlled using RAND_keep_random_devices_open().
743 *) Numerous side-channel attack mitigations have been applied. This may have
744 performance impacts for some algorithms for the benefit of improved
745 security. Specific changes are noted in this change log by their respective
749 *) AIX shared library support overhaul. Switch to AIX "natural" way of
750 handling shared libraries, which means collecting shared objects of
751 different versions and bitnesses in one common archive. This allows to
752 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
753 doesn't affect the way 3rd party applications are linked, only how
754 multi-version installation is managed.
757 *) Make ec_group_do_inverse_ord() more robust and available to other
758 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
759 mitigations are applied to the fallback BN_mod_inverse().
760 When using this function rather than BN_mod_inverse() directly, new
761 EC cryptosystem implementations are then safer-by-default.
764 *) Add coordinate blinding for EC_POINT and implement projective
765 coordinate blinding for generic prime curves as a countermeasure to
766 chosen point SCA attacks.
767 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
769 *) Add blinding to ECDSA and DSA signatures to protect against side channel
770 attacks discovered by Keegan Ryan (NCC Group).
773 *) Enforce checking in the pkeyutl command line app to ensure that the input
774 length does not exceed the maximum supported digest length when performing
775 a sign, verify or verifyrecover operation.
778 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
779 I/O in combination with something like select() or poll() will hang. This
780 can be turned off again using SSL_CTX_clear_mode().
781 Many applications do not properly handle non-application data records, and
782 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
783 around the problems in those applications, but can also break some.
784 It's recommended to read the manpages about SSL_read(), SSL_write(),
785 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
786 SSL_CTX_set_read_ahead() again.
789 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
790 now allow empty (zero character) pass phrases.
793 *) Apply blinding to binary field modular inversion and remove patent
794 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
797 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
798 binary and prime elliptic curves.
801 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
802 constant time fixed point multiplication.
805 *) Revise elliptic curve scalar multiplication with timing attack
806 defenses: ec_wNAF_mul redirects to a constant time implementation
807 when computing fixed point and variable point multiplication (which
808 in OpenSSL are mostly used with secret scalars in keygen, sign,
809 ECDH derive operations).
810 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
813 *) Updated CONTRIBUTING
816 *) Updated DRBG / RAND to request nonce and additional low entropy
817 randomness from the system.
818 [Matthias St. Pierre]
820 *) Updated 'openssl rehash' to use OpenSSL consistent default.
823 *) Moved the load of the ssl_conf module to libcrypto, which helps
824 loading engines that libssl uses before libssl is initialised.
827 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
830 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
831 [Ingo Schwarze, Rich Salz]
833 *) Added output of accepting IP address and port for 'openssl s_server'
836 *) Added a new API for TLSv1.3 ciphersuites:
837 SSL_CTX_set_ciphersuites()
838 SSL_set_ciphersuites()
841 *) Memory allocation failures consistently add an error to the error
845 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
846 in libcrypto when run as setuid/setgid.
849 *) Load any config file by default when libssl is used.
852 *) Added new public header file <openssl/rand_drbg.h> and documentation
853 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
854 [Matthias St. Pierre]
856 *) QNX support removed (cannot find contributors to get their approval
857 for the license change).
860 *) TLSv1.3 replay protection for early data has been implemented. See the
861 SSL_read_early_data() man page for further details.
864 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
865 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
866 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
867 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
868 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
869 configuration has been separated out. See the ciphers man page or the
870 SSL_CTX_set_ciphersuites() man page for more information.
873 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
874 in responder mode now supports the new "-multi" option, which
875 spawns the specified number of child processes to handle OCSP
876 requests. The "-timeout" option now also limits the OCSP
877 responder's patience to wait to receive the full client request
878 on a newly accepted connection. Child processes are respawned
879 as needed, and the CA index file is automatically reloaded
880 when changed. This makes it possible to run the "ocsp" responder
881 as a long-running service, making the OpenSSL CA somewhat more
882 feature-complete. In this mode, most diagnostic messages logged
883 after entering the event loop are logged via syslog(3) rather than
887 *) Added support for X448 and Ed448. Heavily based on original work by
891 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
892 objects loaded. This adds the functions OSSL_STORE_expect() and
893 OSSL_STORE_find() as well as needed tools to construct searches and
894 get the search data out of them.
897 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
898 version of OpenSSL should review their configuration settings to ensure
899 that they are still appropriate for TLSv1.3. For further information see:
900 https://wiki.openssl.org/index.php/TLS1.3
903 *) Grand redesign of the OpenSSL random generator
905 The default RAND method now utilizes an AES-CTR DRBG according to
906 NIST standard SP 800-90Ar1. The new random generator is essentially
907 a port of the default random generator from the OpenSSL FIPS 2.0
908 object module. It is a hybrid deterministic random bit generator
909 using an AES-CTR bit stream and which seeds and reseeds itself
910 automatically using trusted system entropy sources.
912 Some of its new features are:
913 o Support for multiple DRBG instances with seed chaining.
914 o The default RAND method makes use of a DRBG.
915 o There is a public and private DRBG instance.
916 o The DRBG instances are fork-safe.
917 o Keep all global DRBG instances on the secure heap if it is enabled.
918 o The public and private DRBG instance are per thread for lock free
920 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
922 *) Changed Configure so it only says what it does and doesn't dump
923 so much data. Instead, ./configdata.pm should be used as a script
924 to display all sorts of configuration data.
927 *) Added processing of "make variables" to Configure.
930 *) Added SHA512/224 and SHA512/256 algorithm support.
933 *) The last traces of Netware support, first removed in 1.1.0, have
937 *) Get rid of Makefile.shared, and in the process, make the processing
938 of certain files (rc.obj, or the .def/.map/.opt files produced from
939 the ordinal files) more visible and hopefully easier to trace and
940 debug (or make silent).
943 *) Make it possible to have environment variable assignments as
944 arguments to config / Configure.
947 *) Add multi-prime RSA (RFC 8017) support.
950 *) Add SM3 implemented according to GB/T 32905-2016
951 [ Jack Lloyd <jack.lloyd@ribose.com>,
952 Ronald Tse <ronald.tse@ribose.com>,
953 Erick Borsboom <erick.borsboom@ribose.com> ]
955 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
956 as documented in RFC6066.
957 Based on a patch from Tomasz Moń
958 [Filipe Raimundo da Silva]
960 *) Add SM4 implemented according to GB/T 32907-2016.
961 [ Jack Lloyd <jack.lloyd@ribose.com>,
962 Ronald Tse <ronald.tse@ribose.com>,
963 Erick Borsboom <erick.borsboom@ribose.com> ]
965 *) Reimplement -newreq-nodes and ERR_error_string_n; the
966 original author does not agree with the license change.
969 *) Add ARIA AEAD TLS support.
972 *) Some macro definitions to support VS6 have been removed. Visual
973 Studio 6 has not worked since 1.1.0
976 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
977 without clearing the errors.
980 *) Add "atfork" functions. If building on a system that without
981 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
982 requirements. The RAND facility now uses/requires this.
988 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
989 not possible to disable entirely. However, it's still possible to
990 disable the console reading UI method, UI_OpenSSL() (use UI_null()
993 To disable, configure with 'no-ui-console'. 'no-ui' is still
994 possible to use as an alias. Check at compile time with the
995 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
996 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
999 *) Add a STORE module, which implements a uniform and URI based reader of
1000 stores that can contain keys, certificates, CRLs and numerous other
1001 objects. The main API is loosely based on a few stdio functions,
1002 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1003 OSSL_STORE_error and OSSL_STORE_close.
1004 The implementation uses backends called "loaders" to implement arbitrary
1005 URI schemes. There is one built in "loader" for the 'file' scheme.
1008 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
1009 then adjusted to work on FreeBSD 8.4 as well.
1010 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1011 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1014 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
1015 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1016 error code calls like this:
1018 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1020 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1021 that can be encoded in C. For the foreseeable future, this will only
1023 [Richard Levitte and Tim Hudson]
1025 *) Removed BSD cryptodev engine.
1028 *) Add a build target 'build_all_generated', to build all generated files
1029 and only that. This can be used to prepare everything that requires
1030 things like perl for a system that lacks perl and then move everything
1031 to that system and do the rest of the build there.
1034 *) In the UI interface, make it possible to duplicate the user data. This
1035 can be used by engines that need to retain the data for a longer time
1036 than just the call where this user data is passed.
1039 *) Ignore the '-named_curve auto' value for compatibility of applications
1041 [Tomas Mraz <tmraz@fedoraproject.org>]
1043 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1044 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1045 alerts across multiple records (some of which could be empty). In practice
1046 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1047 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1048 support this at all. Supporting it adds significant complexity to the
1049 record layer, and its removal is unlikely to cause interoperability
1053 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1054 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1055 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1059 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1060 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1061 [Richard Levitte, Andy Polyakov]
1063 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1067 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1068 platform rather than 'mingw'.
1071 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
1072 success if they are asked to add an object which already exists
1073 in the store. This change cascades to other functions which load
1074 certificates and CRLs.
1077 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
1078 facilitate stack unwinding even from assembly subroutines.
1081 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1082 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1085 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1086 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1087 which is the minimum version we support.
1090 *) Certificate time validation (X509_cmp_time) enforces stricter
1091 compliance with RFC 5280. Fractional seconds and timezone offsets
1092 are no longer allowed.
1095 *) Add support for ARIA
1098 *) s_client will now send the Server Name Indication (SNI) extension by
1099 default unless the new "-noservername" option is used. The server name is
1100 based on the host provided to the "-connect" option unless overridden by
1101 using "-servername".
1104 *) Add support for SipHash
1107 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1108 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1109 prevent issues where no progress is being made and the peer continually
1110 sends unrecognised record types, using up resources processing them.
1113 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
1114 using the algorithm defined in
1115 https://www.akkadia.org/drepper/SHA-crypt.txt
1118 *) Heartbeat support has been removed; the ABI is changed for now.
1119 [Richard Levitte, Rich Salz]
1121 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1124 *) The RSA "null" method, which was partially supported to avoid patent
1125 issues, has been replaced to always returns NULL.
1129 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
1131 *) Client DoS due to large DH parameter
1133 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1134 malicious server can send a very large prime value to the client. This will
1135 cause the client to spend an unreasonably long period of time generating a
1136 key for this prime resulting in a hang until the client has finished. This
1137 could be exploited in a Denial Of Service attack.
1139 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1143 *) Cache timing vulnerability in RSA Key Generation
1145 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1146 a cache timing side channel attack. An attacker with sufficient access to
1147 mount cache timing attacks during the RSA key generation process could
1148 recover the private key.
1150 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1151 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1155 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1156 parameter is no longer accepted, as it leads to a corrupt table. NULL
1157 pem_str is reserved for alias entries only.
1160 *) Revert blinding in ECDSA sign and instead make problematic addition
1161 length-invariant. Switch even to fixed-length Montgomery multiplication.
1164 *) Change generating and checking of primes so that the error rate of not
1165 being prime depends on the intended use based on the size of the input.
1166 For larger primes this will result in more rounds of Miller-Rabin.
1167 The maximal error rate for primes with more than 1080 bits is lowered
1169 [Kurt Roeckx, Annie Yousar]
1171 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1174 *) Add blinding to ECDSA and DSA signatures to protect against side channel
1175 attacks discovered by Keegan Ryan (NCC Group).
1178 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1179 now allow empty (zero character) pass phrases.
1182 *) Certificate time validation (X509_cmp_time) enforces stricter
1183 compliance with RFC 5280. Fractional seconds and timezone offsets
1184 are no longer allowed.
1187 *) Fixed a text canonicalisation bug in CMS
1189 Where a CMS detached signature is used with text content the text goes
1190 through a canonicalisation process first prior to signing or verifying a
1191 signature. This process strips trailing space at the end of lines, converts
1192 line terminators to CRLF and removes additional trailing line terminators
1193 at the end of a file. A bug in the canonicalisation process meant that
1194 some characters, such as form-feed, were incorrectly treated as whitespace
1195 and removed. This is contrary to the specification (RFC5485). This fix
1196 could mean that detached text data signed with an earlier version of
1197 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
1198 signed with a fixed OpenSSL may fail to verify with an earlier version of
1199 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
1200 and use the "-binary" flag (for the "cms" command line application) or set
1201 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
1204 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
1206 *) Constructed ASN.1 types with a recursive definition could exceed the stack
1208 Constructed ASN.1 types with a recursive definition (such as can be found
1209 in PKCS7) could eventually exceed the stack given malicious input with
1210 excessive recursion. This could result in a Denial Of Service attack. There
1211 are no such structures used within SSL/TLS that come from untrusted sources
1212 so this is considered safe.
1214 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
1219 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
1221 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
1222 effectively reduced to only comparing the least significant bit of each
1223 byte. This allows an attacker to forge messages that would be considered as
1224 authenticated in an amount of tries lower than that guaranteed by the
1225 security claims of the scheme. The module can only be compiled by the
1226 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
1228 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
1233 *) Add a build target 'build_all_generated', to build all generated files
1234 and only that. This can be used to prepare everything that requires
1235 things like perl for a system that lacks perl and then move everything
1236 to that system and do the rest of the build there.
1239 *) Backport SSL_OP_NO_RENGOTIATION
1241 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
1242 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
1243 changes this is no longer possible in 1.1.0. Therefore the new
1244 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
1245 1.1.0 to provide equivalent functionality.
1247 Note that if an application built against 1.1.0h headers (or above) is run
1248 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
1249 accepted but nothing will happen, i.e. renegotiation will not be prevented.
1252 *) Removed the OS390-Unix config target. It relied on a script that doesn't
1256 *) rsaz_1024_mul_avx2 overflow bug on x86_64
1258 There is an overflow bug in the AVX2 Montgomery multiplication procedure
1259 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
1260 Analysis suggests that attacks against RSA and DSA as a result of this
1261 defect would be very difficult to perform and are not believed likely.
1262 Attacks against DH1024 are considered just feasible, because most of the
1263 work necessary to deduce information about a private key may be performed
1264 offline. The amount of resources required for such an attack would be
1265 significant. However, for an attack on TLS to be meaningful, the server
1266 would have to share the DH1024 private key among multiple clients, which is
1267 no longer an option since CVE-2016-0701.
1269 This only affects processors that support the AVX2 but not ADX extensions
1270 like Intel Haswell (4th generation).
1272 This issue was reported to OpenSSL by David Benjamin (Google). The issue
1273 was originally found via the OSS-Fuzz project.
1277 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
1279 *) bn_sqrx8x_internal carry bug on x86_64
1281 There is a carry propagating bug in the x86_64 Montgomery squaring
1282 procedure. No EC algorithms are affected. Analysis suggests that attacks
1283 against RSA and DSA as a result of this defect would be very difficult to
1284 perform and are not believed likely. Attacks against DH are considered just
1285 feasible (although very difficult) because most of the work necessary to
1286 deduce information about a private key may be performed offline. The amount
1287 of resources required for such an attack would be very significant and
1288 likely only accessible to a limited number of attackers. An attacker would
1289 additionally need online access to an unpatched system using the target
1290 private key in a scenario with persistent DH parameters and a private
1291 key that is shared between multiple clients.
1293 This only affects processors that support the BMI1, BMI2 and ADX extensions
1294 like Intel Broadwell (5th generation) and later or AMD Ryzen.
1296 This issue was reported to OpenSSL by the OSS-Fuzz project.
1300 *) Malformed X.509 IPAddressFamily could cause OOB read
1302 If an X.509 certificate has a malformed IPAddressFamily extension,
1303 OpenSSL could do a one-byte buffer overread. The most likely result
1304 would be an erroneous display of the certificate in text format.
1306 This issue was reported to OpenSSL by the OSS-Fuzz project.
1310 Changes between 1.1.0e and 1.1.0f [25 May 2017]
1312 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1313 platform rather than 'mingw'.
1316 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1317 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1318 which is the minimum version we support.
1321 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
1323 *) Encrypt-Then-Mac renegotiation crash
1325 During a renegotiation handshake if the Encrypt-Then-Mac extension is
1326 negotiated where it was not in the original handshake (or vice-versa) then
1327 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
1328 and servers are affected.
1330 This issue was reported to OpenSSL by Joe Orton (Red Hat).
1334 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
1336 *) Truncated packet could crash via OOB read
1338 If one side of an SSL/TLS path is running on a 32-bit host and a specific
1339 cipher is being used, then a truncated packet can cause that host to
1340 perform an out-of-bounds read, usually resulting in a crash.
1342 This issue was reported to OpenSSL by Robert Święcki of Google.
1346 *) Bad (EC)DHE parameters cause a client crash
1348 If a malicious server supplies bad parameters for a DHE or ECDHE key
1349 exchange then this can result in the client attempting to dereference a
1350 NULL pointer leading to a client crash. This could be exploited in a Denial
1353 This issue was reported to OpenSSL by Guido Vranken.
1357 *) BN_mod_exp may produce incorrect results on x86_64
1359 There is a carry propagating bug in the x86_64 Montgomery squaring
1360 procedure. No EC algorithms are affected. Analysis suggests that attacks
1361 against RSA and DSA as a result of this defect would be very difficult to
1362 perform and are not believed likely. Attacks against DH are considered just
1363 feasible (although very difficult) because most of the work necessary to
1364 deduce information about a private key may be performed offline. The amount
1365 of resources required for such an attack would be very significant and
1366 likely only accessible to a limited number of attackers. An attacker would
1367 additionally need online access to an unpatched system using the target
1368 private key in a scenario with persistent DH parameters and a private
1369 key that is shared between multiple clients. For example this can occur by
1370 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1371 similar to CVE-2015-3193 but must be treated as a separate problem.
1373 This issue was reported to OpenSSL by the OSS-Fuzz project.
1377 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1379 *) ChaCha20/Poly1305 heap-buffer-overflow
1381 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1382 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1383 crash. This issue is not considered to be exploitable beyond a DoS.
1385 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1389 *) CMS Null dereference
1391 Applications parsing invalid CMS structures can crash with a NULL pointer
1392 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1393 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1394 structure callback if an attempt is made to free certain invalid encodings.
1395 Only CHOICE structures using a callback which do not handle NULL value are
1398 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1402 *) Montgomery multiplication may produce incorrect results
1404 There is a carry propagating bug in the Broadwell-specific Montgomery
1405 multiplication procedure that handles input lengths divisible by, but
1406 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1407 and DH private keys are impossible. This is because the subroutine in
1408 question is not used in operations with the private key itself and an input
1409 of the attacker's direct choice. Otherwise the bug can manifest itself as
1410 transient authentication and key negotiation failures or reproducible
1411 erroneous outcome of public-key operations with specially crafted input.
1412 Among EC algorithms only Brainpool P-512 curves are affected and one
1413 presumably can attack ECDH key negotiation. Impact was not analyzed in
1414 detail, because pre-requisites for attack are considered unlikely. Namely
1415 multiple clients have to choose the curve in question and the server has to
1416 share the private key among them, neither of which is default behaviour.
1417 Even then only clients that chose the curve will be affected.
1419 This issue was publicly reported as transient failures and was not
1420 initially recognized as a security issue. Thanks to Richard Morgan for
1421 providing reproducible case.
1425 *) Removed automatic addition of RPATH in shared libraries and executables,
1426 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1429 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1431 *) Fix Use After Free for large message sizes
1433 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1434 message larger than approx 16k is received then the underlying buffer to
1435 store the incoming message is reallocated and moved. Unfortunately a
1436 dangling pointer to the old location is left which results in an attempt to
1437 write to the previously freed location. This is likely to result in a
1438 crash, however it could potentially lead to execution of arbitrary code.
1440 This issue only affects OpenSSL 1.1.0a.
1442 This issue was reported to OpenSSL by Robert Święcki.
1446 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1448 *) OCSP Status Request extension unbounded memory growth
1450 A malicious client can send an excessively large OCSP Status Request
1451 extension. If that client continually requests renegotiation, sending a
1452 large OCSP Status Request extension each time, then there will be unbounded
1453 memory growth on the server. This will eventually lead to a Denial Of
1454 Service attack through memory exhaustion. Servers with a default
1455 configuration are vulnerable even if they do not support OCSP. Builds using
1456 the "no-ocsp" build time option are not affected.
1458 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1462 *) SSL_peek() hang on empty record
1464 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1465 sends an empty record. This could be exploited by a malicious peer in a
1466 Denial Of Service attack.
1468 This issue was reported to OpenSSL by Alex Gaynor.
1472 *) Excessive allocation of memory in tls_get_message_header() and
1473 dtls1_preprocess_fragment()
1475 A (D)TLS message includes 3 bytes for its length in the header for the
1476 message. This would allow for messages up to 16Mb in length. Messages of
1477 this length are excessive and OpenSSL includes a check to ensure that a
1478 peer is sending reasonably sized messages in order to avoid too much memory
1479 being consumed to service a connection. A flaw in the logic of version
1480 1.1.0 means that memory for the message is allocated too early, prior to
1481 the excessive message length check. Due to way memory is allocated in
1482 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1483 to service a connection. This could lead to a Denial of Service through
1484 memory exhaustion. However, the excessive message length check still takes
1485 place, and this would cause the connection to immediately fail. Assuming
1486 that the application calls SSL_free() on the failed connection in a timely
1487 manner then the 21Mb of allocated memory will then be immediately freed
1488 again. Therefore the excessive memory allocation will be transitory in
1489 nature. This then means that there is only a security impact if:
1491 1) The application does not call SSL_free() in a timely manner in the event
1492 that the connection fails
1494 2) The application is working in a constrained environment where there is
1495 very little free memory
1497 3) The attacker initiates multiple connection attempts such that there are
1498 multiple connections in a state where memory has been allocated for the
1499 connection; SSL_free() has not yet been called; and there is insufficient
1500 memory to service the multiple requests.
1502 Except in the instance of (1) above any Denial Of Service is likely to be
1503 transitory because as soon as the connection fails the memory is
1504 subsequently freed again in the SSL_free() call. However there is an
1505 increased risk during this period of application crashes due to the lack of
1506 memory - which would then mean a more serious Denial of Service.
1508 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1509 (CVE-2016-6307 and CVE-2016-6308)
1512 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1513 had to be removed. Primary reason is that vendor assembler can't
1514 assemble our modules with -KPIC flag. As result it, assembly
1515 support, was not even available as option. But its lack means
1516 lack of side-channel resistant code, which is incompatible with
1517 security by todays standards. Fortunately gcc is readily available
1518 prepackaged option, which we firmly point at...
1521 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1523 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1524 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1525 (to any value) allows Windows user to access PKCS#12 file generated
1526 with Windows CryptoAPI and protected with non-ASCII password, as well
1527 as files generated under UTF-8 locale on Linux also protected with
1531 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1532 have been disabled by default and removed from DEFAULT, just like RC4.
1533 See the RC4 item below to re-enable both.
1536 *) The method for finding the storage location for the Windows RAND seed file
1537 has changed. First we check %RANDFILE%. If that is not set then we check
1538 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1539 all else fails we fall back to C:\.
1542 *) The EVP_EncryptUpdate() function has had its return type changed from void
1543 to int. A return of 0 indicates and error while a return of 1 indicates
1547 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1548 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1549 off the constant time implementation for RSA, DSA and DH have been made
1550 no-ops and deprecated.
1553 *) Windows RAND implementation was simplified to only get entropy by
1554 calling CryptGenRandom(). Various other RAND-related tickets
1556 [Joseph Wylie Yandle, Rich Salz]
1558 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1559 and OPENSSL_LH_, respectively. The old names are available
1560 with API compatibility. They new names are now completely documented.
1563 *) Unify TYPE_up_ref(obj) methods signature.
1564 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1565 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1566 int (instead of void) like all others TYPE_up_ref() methods.
1567 So now these methods also check the return value of CRYPTO_atomic_add(),
1568 and the validity of object reference counter.
1569 [fdasilvayy@gmail.com]
1571 *) With Windows Visual Studio builds, the .pdb files are installed
1572 alongside the installed libraries and executables. For a static
1573 library installation, ossl_static.pdb is the associate compiler
1574 generated .pdb file to be used when linking programs.
1577 *) Remove openssl.spec. Packaging files belong with the packagers.
1580 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1581 recognise x86_64 architectures automatically. You can still decide
1582 to build for a different bitness with the environment variable
1583 KERNEL_BITS (can be 32 or 64), for example:
1585 KERNEL_BITS=32 ./config
1589 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1590 256 bit AES and HMAC with SHA256.
1593 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1596 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1599 *) To enable users to have their own config files and build file templates,
1600 Configure looks in the directory indicated by the environment variable
1601 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1602 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1603 name and is used as is.
1606 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1607 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1608 X509_CERT_FILE_CTX was removed.
1611 *) "shared" builds are now the default. To create only static libraries use
1612 the "no-shared" Configure option.
1615 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1616 All of these option have not worked for some while and are fundamental
1620 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1621 global cleanup functions are no longer required because they are handled
1622 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1623 Explicitly de-initing can cause problems (e.g. where a library that uses
1624 OpenSSL de-inits, but an application is still using it). The affected
1625 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1626 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1627 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1628 COMP_zlib_cleanup().
1631 *) --strict-warnings no longer enables runtime debugging options
1632 such as REF_DEBUG. Instead, debug options are automatically
1633 enabled with '--debug' builds.
1634 [Andy Polyakov, Emilia Käsper]
1636 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1637 have been moved out of the public header files. New functions for managing
1638 these have been added.
1641 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1642 objects have been moved out of the public header files. New
1643 functions for managing these have been added.
1646 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1647 have been moved out of the public header files. New functions for managing
1648 these have been added.
1651 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1652 moved out of the public header files. New functions for managing these
1656 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1659 *) Removed the mk1mf build scripts.
1662 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1663 it is always safe to #include a header now.
1666 *) Removed the aged BC-32 config and all its supporting scripts
1669 *) Removed support for Ultrix, Netware, and OS/2.
1672 *) Add support for HKDF.
1673 [Alessandro Ghedini]
1675 *) Add support for blake2b and blake2s
1678 *) Added support for "pipelining". Ciphers that have the
1679 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1680 encryptions/decryptions simultaneously. There are currently no built-in
1681 ciphers with this property but the expectation is that engines will be able
1682 to offer it to significantly improve throughput. Support has been extended
1683 into libssl so that multiple records for a single connection can be
1684 processed in one go (for >=TLS 1.1).
1687 *) Added the AFALG engine. This is an async capable engine which is able to
1688 offload work to the Linux kernel. In this initial version it only supports
1689 AES128-CBC. The kernel must be version 4.1.0 or greater.
1692 *) OpenSSL now uses a new threading API. It is no longer necessary to
1693 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1694 are two supported threading models: pthreads and windows threads. It is
1695 also possible to configure OpenSSL at compile time for "no-threads". The
1696 old threading API should no longer be used. The functions have been
1697 replaced with "no-op" compatibility macros.
1698 [Alessandro Ghedini, Matt Caswell]
1700 *) Modify behavior of ALPN to invoke callback after SNI/servername
1701 callback, such that updates to the SSL_CTX affect ALPN.
1704 *) Add SSL_CIPHER queries for authentication and key-exchange.
1707 *) Changes to the DEFAULT cipherlist:
1708 - Prefer (EC)DHE handshakes over plain RSA.
1709 - Prefer AEAD ciphers over legacy ciphers.
1710 - Prefer ECDSA over RSA when both certificates are available.
1711 - Prefer TLSv1.2 ciphers/PRF.
1712 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1716 *) Change the ECC default curve list to be this, in order: x25519,
1717 secp256r1, secp521r1, secp384r1.
1720 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1721 disabled by default. They can be re-enabled using the
1722 enable-weak-ssl-ciphers option to Configure.
1725 *) If the server has ALPN configured, but supports no protocols that the
1726 client advertises, send a fatal "no_application_protocol" alert.
1727 This behaviour is SHALL in RFC 7301, though it isn't universally
1728 implemented by other servers.
1731 *) Add X25519 support.
1732 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1733 for public and private key encoding using the format documented in
1734 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1735 key generation and key derivation.
1737 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1741 *) Deprecate SRP_VBASE_get_by_user.
1742 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1743 In order to fix an unavoidable memory leak (CVE-2016-0798),
1744 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1745 seed, even if the seed is configured.
1747 Users should use SRP_VBASE_get1_by_user instead. Note that in
1748 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1749 also that even though configuring the SRP seed attempts to hide
1750 invalid usernames by continuing the handshake with fake
1751 credentials, this behaviour is not constant time and no strong
1752 guarantees are made that the handshake is indistinguishable from
1753 that of a valid user.
1756 *) Configuration change; it's now possible to build dynamic engines
1757 without having to build shared libraries and vice versa. This
1758 only applies to the engines in engines/, those in crypto/engine/
1759 will always be built into libcrypto (i.e. "static").
1761 Building dynamic engines is enabled by default; to disable, use
1762 the configuration option "disable-dynamic-engine".
1764 The only requirements for building dynamic engines are the
1765 presence of the DSO module and building with position independent
1766 code, so they will also automatically be disabled if configuring
1767 with "disable-dso" or "disable-pic".
1769 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1770 are also taken away from openssl/opensslconf.h, as they are
1774 *) Configuration change; if there is a known flag to compile
1775 position independent code, it will always be applied on the
1776 libcrypto and libssl object files, and never on the application
1777 object files. This means other libraries that use routines from
1778 libcrypto / libssl can be made into shared libraries regardless
1779 of how OpenSSL was configured.
1781 If this isn't desirable, the configuration options "disable-pic"
1782 or "no-pic" can be used to disable the use of PIC. This will
1783 also disable building shared libraries and dynamic engines.
1786 *) Removed JPAKE code. It was experimental and has no wide use.
1789 *) The INSTALL_PREFIX Makefile variable has been renamed to
1790 DESTDIR. That makes for less confusion on what this variable
1791 is for. Also, the configuration option --install_prefix is
1795 *) Heartbeat for TLS has been removed and is disabled by default
1796 for DTLS; configure with enable-heartbeats. Code that uses the
1797 old #define's might need to be updated.
1798 [Emilia Käsper, Rich Salz]
1800 *) Rename REF_CHECK to REF_DEBUG.
1803 *) New "unified" build system
1805 The "unified" build system is aimed to be a common system for all
1806 platforms we support. With it comes new support for VMS.
1808 This system builds supports building in a different directory tree
1809 than the source tree. It produces one Makefile (for unix family
1810 or lookalikes), or one descrip.mms (for VMS).
1812 The source of information to make the Makefile / descrip.mms is
1813 small files called 'build.info', holding the necessary
1814 information for each directory with source to compile, and a
1815 template in Configurations, like unix-Makefile.tmpl or
1818 With this change, the library names were also renamed on Windows
1819 and on VMS. They now have names that are closer to the standard
1820 on Unix, and include the major version number, and in certain
1821 cases, the architecture they are built for. See "Notes on shared
1822 libraries" in INSTALL.
1824 We rely heavily on the perl module Text::Template.
1827 *) Added support for auto-initialisation and de-initialisation of the library.
1828 OpenSSL no longer requires explicit init or deinit routines to be called,
1829 except in certain circumstances. See the OPENSSL_init_crypto() and
1830 OPENSSL_init_ssl() man pages for further information.
1833 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1834 "peer" argument is now expected to be a BIO_ADDR object.
1836 *) Rewrite of BIO networking library. The BIO library lacked consistent
1837 support of IPv6, and adding it required some more extensive
1838 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1839 which hold all types of addresses and chains of address information.
1840 It also introduces a new API, with functions like BIO_socket,
1841 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1842 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1843 have been adapted accordingly.
1846 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1850 *) CRIME protection: disable compression by default, even if OpenSSL is
1851 compiled with zlib enabled. Applications can still enable compression
1852 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1853 using the SSL_CONF library to configure compression.
1856 *) The signature of the session callback configured with
1857 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1858 was explicitly marked as 'const unsigned char*' instead of
1862 *) Always DPURIFY. Remove the use of uninitialized memory in the
1863 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1866 *) Removed many obsolete configuration items, including
1867 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1868 MD2_CHAR, MD2_INT, MD2_LONG
1870 IDEA_SHORT, IDEA_LONG
1871 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1872 [Rich Salz, with advice from Andy Polyakov]
1874 *) Many BN internals have been moved to an internal header file.
1875 [Rich Salz with help from Andy Polyakov]
1877 *) Configuration and writing out the results from it has changed.
1878 Files such as Makefile include/openssl/opensslconf.h and are now
1879 produced through general templates, such as Makefile.in and
1880 crypto/opensslconf.h.in and some help from the perl module
1883 Also, the center of configuration information is no longer
1884 Makefile. Instead, Configure produces a perl module in
1885 configdata.pm which holds most of the config data (in the hash
1886 table %config), the target data that comes from the target
1887 configuration in one of the Configurations/*.conf files (in
1891 *) To clarify their intended purposes, the Configure options
1892 --prefix and --openssldir change their semantics, and become more
1893 straightforward and less interdependent.
1895 --prefix shall be used exclusively to give the location INSTALLTOP
1896 where programs, scripts, libraries, include files and manuals are
1897 going to be installed. The default is now /usr/local.
1899 --openssldir shall be used exclusively to give the default
1900 location OPENSSLDIR where certificates, private keys, CRLs are
1901 managed. This is also where the default openssl.cnf gets
1903 If the directory given with this option is a relative path, the
1904 values of both the --prefix value and the --openssldir value will
1905 be combined to become OPENSSLDIR.
1906 The default for --openssldir is INSTALLTOP/ssl.
1908 Anyone who uses --openssldir to specify where OpenSSL is to be
1909 installed MUST change to use --prefix instead.
1912 *) The GOST engine was out of date and therefore it has been removed. An up
1913 to date GOST engine is now being maintained in an external repository.
1914 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1915 support for GOST ciphersuites (these are only activated if a GOST engine
1919 *) EGD is no longer supported by default; use enable-egd when
1921 [Ben Kaduk and Rich Salz]
1923 *) The distribution now has Makefile.in files, which are used to
1924 create Makefile's when Configure is run. *Configure must be run
1925 before trying to build now.*
1928 *) The return value for SSL_CIPHER_description() for error conditions
1932 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1934 Obtaining and performing DNSSEC validation of TLSA records is
1935 the application's responsibility. The application provides
1936 the TLSA records of its choice to OpenSSL, and these are then
1937 used to authenticate the peer.
1939 The TLSA records need not even come from DNS. They can, for
1940 example, be used to implement local end-entity certificate or
1941 trust-anchor "pinning", where the "pin" data takes the form
1942 of TLSA records, which can augment or replace verification
1943 based on the usual WebPKI public certification authorities.
1946 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1947 continues to support deprecated interfaces in default builds.
1948 However, applications are strongly advised to compile their
1949 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1950 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1951 or the 1.1.0 releases.
1953 In environments in which all applications have been ported to
1954 not use any deprecated interfaces OpenSSL's Configure script
1955 should be used with the --api=1.1.0 option to entirely remove
1956 support for the deprecated features from the library and
1957 unconditionally disable them in the installed headers.
1958 Essentially the same effect can be achieved with the "no-deprecated"
1959 argument to Configure, except that this will always restrict
1960 the build to just the latest API, rather than a fixed API
1963 As applications are ported to future revisions of the API,
1964 they should update their compile-time OPENSSL_API_COMPAT define
1965 accordingly, but in most cases should be able to continue to
1966 compile with later releases.
1968 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1969 0x10000000L and 0x00908000L, respectively. However those
1970 versions did not support the OPENSSL_API_COMPAT feature, and
1971 so applications are not typically tested for explicit support
1972 of just the undeprecated features of either release.
1975 *) Add support for setting the minimum and maximum supported protocol.
1976 It can bet set via the SSL_set_min_proto_version() and
1977 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1978 MaxProtocol. It's recommended to use the new APIs to disable
1979 protocols instead of disabling individual protocols using
1980 SSL_set_options() or SSL_CONF's Protocol. This change also
1981 removes support for disabling TLS 1.2 in the OpenSSL TLS
1982 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1985 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1988 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1989 and integrates ECDSA and ECDH functionality into EC. Implementations can
1990 now redirect key generation and no longer need to convert to or from
1993 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1994 include the ec.h header file instead.
1997 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1998 ciphers who are no longer supported and drops support the ephemeral RSA key
1999 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2002 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2003 opaque. For HMAC_CTX, the following constructors and destructors
2006 HMAC_CTX *HMAC_CTX_new(void);
2007 void HMAC_CTX_free(HMAC_CTX *ctx);
2009 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2010 destroy such methods has been added. See EVP_MD_meth_new(3) and
2011 EVP_CIPHER_meth_new(3) for documentation.
2014 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2015 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2016 EVP_MD_CTX_reset() should be called instead to reinitialise
2017 an already created structure.
2018 2) For consistency with the majority of our object creators and
2019 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2020 EVP_MD_CTX_(new|free). The old names are retained as macros
2021 for deprecated builds.
2024 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
2025 cryptographic operations to be performed asynchronously as long as an
2026 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2027 further details. Libssl has also had this capability integrated with the
2028 introduction of the new mode SSL_MODE_ASYNC and associated error
2029 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2030 pages. This work was developed in partnership with Intel Corp.
2033 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2034 always enabled now. If you want to disable the support you should
2035 exclude it using the list of supported ciphers. This also means that the
2036 "-no_ecdhe" option has been removed from s_server.
2039 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2040 SSL_{CTX_}set1_curves() which can set a list.
2043 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2044 curve you want to support using SSL_{CTX_}set1_curves().
2047 *) State machine rewrite. The state machine code has been significantly
2048 refactored in order to remove much duplication of code and solve issues
2049 with the old code (see ssl/statem/README for further details). This change
2050 does have some associated API changes. Notably the SSL_state() function
2051 has been removed and replaced by SSL_get_state which now returns an
2052 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
2053 altogether. The previous handshake states defined in ssl.h and ssl3.h have
2057 *) All instances of the string "ssleay" in the public API were replaced
2058 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
2059 Some error codes related to internal RSA_eay API's were renamed.
2062 *) The demo files in crypto/threads were moved to demo/threads.
2065 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
2067 [Matt Caswell, Rich Salz]
2069 *) New ASN.1 embed macro.
2071 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
2072 structure is not allocated: it is part of the parent. That is instead of
2080 This reduces memory fragmentation and make it impossible to accidentally
2081 set a mandatory field to NULL.
2083 This currently only works for some fields specifically a SEQUENCE, CHOICE,
2084 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
2085 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
2089 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
2092 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
2093 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
2094 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
2095 DES and RC4 ciphersuites.
2098 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2099 This changes the decoding behaviour for some invalid messages,
2100 though the change is mostly in the more lenient direction, and
2101 legacy behaviour is preserved as much as possible.
2104 *) Fix no-stdio build.
2105 [ David Woodhouse <David.Woodhouse@intel.com> and also
2106 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
2108 *) New testing framework
2109 The testing framework has been largely rewritten and is now using
2110 perl and the perl modules Test::Harness and an extended variant of
2111 Test::More called OpenSSL::Test to do its work. All test scripts in
2112 test/ have been rewritten into test recipes, and all direct calls to
2113 executables in test/Makefile have become individual recipes using the
2114 simplified testing OpenSSL::Test::Simple.
2116 For documentation on our testing modules, do:
2118 perldoc test/testlib/OpenSSL/Test/Simple.pm
2119 perldoc test/testlib/OpenSSL/Test.pm
2123 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
2124 are used; the latter aborts on memory leaks (usually checked on exit).
2125 Some undocumented "set malloc, etc., hooks" functions were removed
2126 and others were changed. All are now documented.
2129 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2131 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2133 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
2134 from RFC4279, RFC4785, RFC5487, RFC5489.
2136 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
2137 original RSA_PSK patch.
2140 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
2141 era flag was never set throughout the codebase (only read). Also removed
2142 SSL3_FLAGS_POP_BUFFER which was only used if
2143 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
2146 *) Changed the default name options in the "ca", "crl", "req" and "x509"
2147 to be "oneline" instead of "compat".
2150 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
2151 not aware of clients that still exhibit this bug, and the workaround
2152 hasn't been working properly for a while.
2155 *) The return type of BIO_number_read() and BIO_number_written() as well as
2156 the corresponding num_read and num_write members in the BIO structure has
2157 changed from unsigned long to uint64_t. On platforms where an unsigned
2158 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
2162 *) Given the pervasive nature of TLS extensions it is inadvisable to run
2163 OpenSSL without support for them. It also means that maintaining
2164 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
2165 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
2168 *) Removed support for the two export grade static DH ciphersuites
2169 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
2170 were newly added (along with a number of other static DH ciphersuites) to
2171 1.0.2. However the two export ones have *never* worked since they were
2172 introduced. It seems strange in any case to be adding new export
2173 ciphersuites, and given "logjam" it also does not seem correct to fix them.
2176 *) Version negotiation has been rewritten. In particular SSLv23_method(),
2177 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
2178 and turned into macros which simply call the new preferred function names
2179 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
2180 should use the new names instead. Also as part of this change the ssl23.h
2181 header file has been removed.
2184 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
2185 code and the associated standard is no longer considered fit-for-purpose.
2188 *) RT2547 was closed. When generating a private key, try to make the
2189 output file readable only by the owner. This behavior change might
2190 be noticeable when interacting with other software.
2192 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
2196 *) Added HTTP GET support to the ocsp command.
2199 *) Changed default digest for the dgst and enc commands from MD5 to
2203 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
2206 *) Added support for TLS extended master secret from
2207 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
2208 initial patch which was a great help during development.
2211 *) All libssl internal structures have been removed from the public header
2212 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
2213 now redundant). Users should not attempt to access internal structures
2214 directly. Instead they should use the provided API functions.
2217 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
2218 Access to deprecated functions can be re-enabled by running config with
2219 "enable-deprecated". In addition applications wishing to use deprecated
2220 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
2221 will, by default, disable some transitive includes that previously existed
2222 in the header files (e.g. ec.h will no longer, by default, include bn.h)
2225 *) Added support for OCB mode. OpenSSL has been granted a patent license
2226 compatible with the OpenSSL license for use of OCB. Details are available
2227 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
2228 for OCB can be removed by calling config with no-ocb.
2231 *) SSLv2 support has been removed. It still supports receiving a SSLv2
2232 compatible client hello.
2235 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
2236 done while fixing the error code for the key-too-small case.
2237 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
2239 *) CA.sh has been removed; use CA.pl instead.
2242 *) Removed old DES API.
2245 *) Remove various unsupported platforms:
2251 Sinix/ReliantUNIX RM400
2256 16-bit platforms such as WIN16
2259 *) Clean up OPENSSL_NO_xxx #define's
2260 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
2261 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
2262 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
2263 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
2264 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
2265 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
2266 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
2267 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
2268 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
2269 Remove MS_STATIC; it's a relic from platforms <32 bits.
2272 *) Cleaned up dead code
2273 Remove all but one '#ifdef undef' which is to be looked at.
2276 *) Clean up calling of xxx_free routines.
2277 Just like free(), fix most of the xxx_free routines to accept
2278 NULL. Remove the non-null checks from callers. Save much code.
2281 *) Add secure heap for storage of private keys (when possible).
2282 Add BIO_s_secmem(), CBIGNUM, etc.
2283 Contributed by Akamai Technologies under our Corporate CLA.
2286 *) Experimental support for a new, fast, unbiased prime candidate generator,
2287 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
2288 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
2290 *) New output format NSS in the sess_id command line tool. This allows
2291 exporting the session id and the master key in NSS keylog format.
2292 [Martin Kaiser <martin@kaiser.cx>]
2294 *) Harmonize version and its documentation. -f flag is used to display
2296 [mancha <mancha1@zoho.com>]
2298 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2299 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
2300 [mancha <mancha1@zoho.com>]
2302 *) Fix some double frees. These are not thought to be exploitable.
2303 [mancha <mancha1@zoho.com>]
2305 *) A missing bounds check in the handling of the TLS heartbeat extension
2306 can be used to reveal up to 64k of memory to a connected client or
2309 Thanks for Neel Mehta of Google Security for discovering this bug and to
2310 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2311 preparing the fix (CVE-2014-0160)
2312 [Adam Langley, Bodo Moeller]
2314 *) Fix for the attack described in the paper "Recovering OpenSSL
2315 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2316 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2317 http://eprint.iacr.org/2014/140
2319 Thanks to Yuval Yarom and Naomi Benger for discovering this
2320 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2321 [Yuval Yarom and Naomi Benger]
2323 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2324 this fixes a limitation in previous versions of OpenSSL.
2327 *) Experimental encrypt-then-mac support.
2329 Experimental support for encrypt then mac from
2330 draft-gutmann-tls-encrypt-then-mac-02.txt
2332 To enable it set the appropriate extension number (0x42 for the test
2333 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
2335 For non-compliant peers (i.e. just about everything) this should have no
2338 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
2342 *) Add EVP support for key wrapping algorithms, to avoid problems with
2343 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2344 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2345 algorithms and include tests cases.
2348 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2352 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2353 MGF1 digest and OAEP label.
2356 *) Make openssl verify return errors.
2357 [Chris Palmer <palmer@google.com> and Ben Laurie]
2359 *) New function ASN1_TIME_diff to calculate the difference between two
2360 ASN1_TIME structures or one structure and the current time.
2363 *) Update fips_test_suite to support multiple command line options. New
2364 test to induce all self test errors in sequence and check expected
2368 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2369 sign or verify all in one operation.
2372 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2373 test programs and fips_test_suite. Includes functionality to parse
2374 the minimal script output of fipsalgest.pl directly.
2377 *) Add authorisation parameter to FIPS_module_mode_set().
2380 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2383 *) Use separate DRBG fields for internal and external flags. New function
2384 FIPS_drbg_health_check() to perform on demand health checking. Add
2385 generation tests to fips_test_suite with reduced health check interval to
2386 demonstrate periodic health checking. Add "nodh" option to
2387 fips_test_suite to skip very slow DH test.
2390 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2394 *) More extensive health check for DRBG checking many more failure modes.
2395 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2396 combination: call this in fips_test_suite.
2399 *) Add support for canonical generation of DSA parameter 'g'. See
2402 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2403 POST to handle HMAC cases.
2406 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2407 to return numerical and string versions of the FIPS module number.
2410 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2411 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2412 outside the validated module in the FIPS capable OpenSSL.
2415 *) Minor change to DRBG entropy callback semantics. In some cases
2416 there is no multiple of the block length between min_len and
2417 max_len. Allow the callback to return more than max_len bytes
2418 of entropy but discard any extra: it is the callback's responsibility
2419 to ensure that the extra data discarded does not impact the
2420 requested amount of entropy.
2423 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2424 information in FIPS186-3, SP800-57 and SP800-131A.
2427 *) CCM support via EVP. Interface is very similar to GCM case except we
2428 must supply all data in one chunk (i.e. no update, final) and the
2429 message length must be supplied if AAD is used. Add algorithm test
2433 *) Initial version of POST overhaul. Add POST callback to allow the status
2434 of POST to be monitored and/or failures induced. Modify fips_test_suite
2435 to use callback. Always run all selftests even if one fails.
2438 *) XTS support including algorithm test driver in the fips_gcmtest program.
2439 Note: this does increase the maximum key length from 32 to 64 bytes but
2440 there should be no binary compatibility issues as existing applications
2441 will never use XTS mode.
2444 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2445 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2446 performs algorithm blocking for unapproved PRNG types. Also do not
2447 set PRNG type in FIPS_mode_set(): leave this to the application.
2448 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2449 the standard OpenSSL PRNG: set additional data to a date time vector.
2452 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2453 This shouldn't present any incompatibility problems because applications
2454 shouldn't be using these directly and any that are will need to rethink
2455 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2458 *) Extensive self tests and health checking required by SP800-90 DRBG.
2459 Remove strength parameter from FIPS_drbg_instantiate and always
2460 instantiate at maximum supported strength.
2463 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2466 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2469 *) New function DH_compute_key_padded() to compute a DH key and pad with
2470 leading zeroes if needed: this complies with SP800-56A et al.
2473 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2474 anything, incomplete, subject to change and largely untested at present.
2477 *) Modify fipscanisteronly build option to only build the necessary object
2478 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2481 *) Add experimental option FIPSSYMS to give all symbols in
2482 fipscanister.o and FIPS or fips prefix. This will avoid
2483 conflicts with future versions of OpenSSL. Add perl script
2484 util/fipsas.pl to preprocess assembly language source files
2485 and rename any affected symbols.
2488 *) Add selftest checks and algorithm block of non-fips algorithms in
2489 FIPS mode. Remove DES2 from selftests.
2492 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2493 return internal method without any ENGINE dependencies. Add new
2494 tiny fips sign and verify functions.
2497 *) New build option no-ec2m to disable characteristic 2 code.
2500 *) New build option "fipscanisteronly". This only builds fipscanister.o
2501 and (currently) associated fips utilities. Uses the file Makefile.fips
2502 instead of Makefile.org as the prototype.
2505 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2506 Update fips_gcmtest to use IV generator.
2509 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2510 setting output buffer to NULL. The *Final function must be
2511 called although it will not retrieve any additional data. The tag
2512 can be set or retrieved with a ctrl. The IV length is by default 12
2513 bytes (96 bits) but can be set to an alternative value. If the IV
2514 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2518 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2519 underlying do_cipher function handles all cipher semantics itself
2520 including padding and finalisation. This is useful if (for example)
2521 an ENGINE cipher handles block padding itself. The behaviour of
2522 do_cipher is subtly changed if this flag is set: the return value
2523 is the number of characters written to the output buffer (zero is
2524 no longer an error code) or a negative error code. Also if the
2525 input buffer is NULL and length 0 finalisation should be performed.
2528 *) If a candidate issuer certificate is already part of the constructed
2529 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2532 *) Improve forward-security support: add functions
2534 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2535 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2537 for use by SSL/TLS servers; the callback function will be called whenever a
2538 new session is created, and gets to decide whether the session may be
2539 cached to make it resumable (return 0) or not (return 1). (As by the
2540 SSL/TLS protocol specifications, the session_id sent by the server will be
2541 empty to indicate that the session is not resumable; also, the server will
2542 not generate RFC 4507 (RFC 5077) session tickets.)
2544 A simple reasonable callback implementation is to return is_forward_secure.
2545 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2546 by the SSL/TLS server library, indicating whether it can provide forward
2548 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2550 *) New -verify_name option in command line utilities to set verification
2554 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2555 Add CMAC pkey methods.
2558 *) Experimental renegotiation in s_server -www mode. If the client
2559 browses /reneg connection is renegotiated. If /renegcert it is
2560 renegotiated requesting a certificate.
2563 *) Add an "external" session cache for debugging purposes to s_server. This
2564 should help trace issues which normally are only apparent in deployed
2565 multi-process servers.
2568 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2569 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2570 BIO_set_cipher() and some obscure PEM functions were changed so they
2571 can now return an error. The RAND changes required a change to the
2572 RAND_METHOD structure.
2575 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2576 a gcc attribute to warn if the result of a function is ignored. This
2577 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2578 whose return value is often ignored.
2581 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2582 These allow SCTs (signed certificate timestamps) to be requested and
2583 validated when establishing a connection.
2584 [Rob Percival <robpercival@google.com>]
2586 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2588 *) Prevent padding oracle in AES-NI CBC MAC check
2590 A MITM attacker can use a padding oracle attack to decrypt traffic
2591 when the connection uses an AES CBC cipher and the server support
2594 This issue was introduced as part of the fix for Lucky 13 padding
2595 attack (CVE-2013-0169). The padding check was rewritten to be in
2596 constant time by making sure that always the same bytes are read and
2597 compared against either the MAC or padding bytes. But it no longer
2598 checked that there was enough data to have both the MAC and padding
2601 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2605 *) Fix EVP_EncodeUpdate overflow
2607 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2608 Base64 encoding of binary data. If an attacker is able to supply very large
2609 amounts of input data then a length check can overflow resulting in a heap
2612 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2613 the PEM_write_bio* family of functions. These are mainly used within the
2614 OpenSSL command line applications, so any application which processes data
2615 from an untrusted source and outputs it as a PEM file should be considered
2616 vulnerable to this issue. User applications that call these APIs directly
2617 with large amounts of untrusted data may also be vulnerable.
2619 This issue was reported by Guido Vranken.
2623 *) Fix EVP_EncryptUpdate overflow
2625 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2626 is able to supply very large amounts of input data after a previous call to
2627 EVP_EncryptUpdate() with a partial block then a length check can overflow
2628 resulting in a heap corruption. Following an analysis of all OpenSSL
2629 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2630 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2631 the first called function after an EVP_EncryptInit(), and therefore that
2632 specific call must be safe. The second form is where the length passed to
2633 EVP_EncryptUpdate() can be seen from the code to be some small value and
2634 therefore there is no possibility of an overflow. Since all instances are
2635 one of these two forms, it is believed that there can be no overflows in
2636 internal code due to this problem. It should be noted that
2637 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2638 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2639 of these calls have also been analysed too and it is believed there are no
2640 instances in internal usage where an overflow could occur.
2642 This issue was reported by Guido Vranken.
2646 *) Prevent ASN.1 BIO excessive memory allocation
2648 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2649 a short invalid encoding can cause allocation of large amounts of memory
2650 potentially consuming excessive resources or exhausting memory.
2652 Any application parsing untrusted data through d2i BIO functions is
2653 affected. The memory based functions such as d2i_X509() are *not* affected.
2654 Since the memory based functions are used by the TLS library, TLS
2655 applications are not affected.
2657 This issue was reported by Brian Carpenter.
2663 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2664 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2665 in arbitrary stack data being returned in the buffer.
2667 This issue was reported by Guido Vranken.
2671 *) Modify behavior of ALPN to invoke callback after SNI/servername
2672 callback, such that updates to the SSL_CTX affect ALPN.
2675 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2679 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2680 methods are enabled and ssl2 is disabled the methods return NULL.
2683 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2685 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2686 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2687 provide any "EXPORT" or "LOW" strength ciphers.
2690 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2691 is by default disabled at build-time. Builds that are not configured with
2692 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2693 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2694 will need to explicitly call either of:
2696 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2698 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2700 as appropriate. Even if either of those is used, or the application
2701 explicitly uses the version-specific SSLv2_method() or its client and
2702 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2703 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2704 ciphers, and SSLv2 56-bit DES are no longer available.
2708 *) Fix a double-free in DSA code
2710 A double free bug was discovered when OpenSSL parses malformed DSA private
2711 keys and could lead to a DoS attack or memory corruption for applications
2712 that receive DSA private keys from untrusted sources. This scenario is
2715 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2720 *) Disable SRP fake user seed to address a server memory leak.
2722 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2724 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2725 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2726 was changed to ignore the "fake user" SRP seed, even if the seed
2729 Users should use SRP_VBASE_get1_by_user instead. Note that in
2730 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2731 also that even though configuring the SRP seed attempts to hide
2732 invalid usernames by continuing the handshake with fake
2733 credentials, this behaviour is not constant time and no strong
2734 guarantees are made that the handshake is indistinguishable from
2735 that of a valid user.
2739 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2741 In the BN_hex2bn function the number of hex digits is calculated using an
2742 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2743 large values of |i| this can result in |bn_expand| not allocating any
2744 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2745 field as NULL leading to a subsequent NULL ptr deref. For very large values
2746 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2747 In this case memory is allocated to the internal BIGNUM data field, but it
2748 is insufficiently sized leading to heap corruption. A similar issue exists
2749 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2750 is ever called by user applications with very large untrusted hex/dec data.
2751 This is anticipated to be a rare occurrence.
2753 All OpenSSL internal usage of these functions use data that is not expected
2754 to be untrusted, e.g. config file data or application command line
2755 arguments. If user developed applications generate config file data based
2756 on untrusted data then it is possible that this could also lead to security
2757 consequences. This is also anticipated to be rare.
2759 This issue was reported to OpenSSL by Guido Vranken.
2763 *) Fix memory issues in BIO_*printf functions
2765 The internal |fmtstr| function used in processing a "%s" format string in
2766 the BIO_*printf functions could overflow while calculating the length of a
2767 string and cause an OOB read when printing very long strings.
2769 Additionally the internal |doapr_outch| function can attempt to write to an
2770 OOB memory location (at an offset from the NULL pointer) in the event of a
2771 memory allocation failure. In 1.0.2 and below this could be caused where
2772 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2773 could be in processing a very long "%s" format string. Memory leaks can
2776 The first issue may mask the second issue dependent on compiler behaviour.
2777 These problems could enable attacks where large amounts of untrusted data
2778 is passed to the BIO_*printf functions. If applications use these functions
2779 in this way then they could be vulnerable. OpenSSL itself uses these
2780 functions when printing out human-readable dumps of ASN.1 data. Therefore
2781 applications that print this data could be vulnerable if the data is from
2782 untrusted sources. OpenSSL command line applications could also be
2783 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2784 as command line arguments.
2786 Libssl is not considered directly vulnerable. Additionally certificates etc
2787 received via remote connections via libssl are also unlikely to be able to
2788 trigger these issues because of message size limits enforced within libssl.
2790 This issue was reported to OpenSSL Guido Vranken.
2794 *) Side channel attack on modular exponentiation
2796 A side-channel attack was found which makes use of cache-bank conflicts on
2797 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2798 of RSA keys. The ability to exploit this issue is limited as it relies on
2799 an attacker who has control of code in a thread running on the same
2800 hyper-threaded core as the victim thread which is performing decryptions.
2802 This issue was reported to OpenSSL by Yuval Yarom, The University of
2803 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2804 Nadia Heninger, University of Pennsylvania with more information at
2805 http://cachebleed.info.
2809 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2810 if no keysize is specified with default_bits. This fixes an
2811 omission in an earlier change that changed all RSA/DSA key generation
2812 apps to use 2048 bits by default.
2815 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2816 *) DH small subgroups
2818 Historically OpenSSL only ever generated DH parameters based on "safe"
2819 primes. More recently (in version 1.0.2) support was provided for
2820 generating X9.42 style parameter files such as those required for RFC 5114
2821 support. The primes used in such files may not be "safe". Where an
2822 application is using DH configured with parameters based on primes that are
2823 not "safe" then an attacker could use this fact to find a peer's private
2824 DH exponent. This attack requires that the attacker complete multiple
2825 handshakes in which the peer uses the same private DH exponent. For example
2826 this could be used to discover a TLS server's private DH exponent if it's
2827 reusing the private DH exponent or it's using a static DH ciphersuite.
2829 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2830 TLS. It is not on by default. If the option is not set then the server
2831 reuses the same private DH exponent for the life of the server process and
2832 would be vulnerable to this attack. It is believed that many popular
2833 applications do set this option and would therefore not be at risk.
2835 The fix for this issue adds an additional check where a "q" parameter is
2836 available (as is the case in X9.42 based parameters). This detects the
2837 only known attack, and is the only possible defense for static DH
2838 ciphersuites. This could have some performance impact.
2840 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2841 default and cannot be disabled. This could have some performance impact.
2843 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2847 *) SSLv2 doesn't block disabled ciphers
2849 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2850 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2851 been disabled, provided that the SSLv2 protocol was not also disabled via
2854 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2855 and Sebastian Schinzel.
2859 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2861 *) BN_mod_exp may produce incorrect results on x86_64
2863 There is a carry propagating bug in the x86_64 Montgomery squaring
2864 procedure. No EC algorithms are affected. Analysis suggests that attacks
2865 against RSA and DSA as a result of this defect would be very difficult to
2866 perform and are not believed likely. Attacks against DH are considered just
2867 feasible (although very difficult) because most of the work necessary to
2868 deduce information about a private key may be performed offline. The amount
2869 of resources required for such an attack would be very significant and
2870 likely only accessible to a limited number of attackers. An attacker would
2871 additionally need online access to an unpatched system using the target
2872 private key in a scenario with persistent DH parameters and a private
2873 key that is shared between multiple clients. For example this can occur by
2874 default in OpenSSL DHE based SSL/TLS ciphersuites.
2876 This issue was reported to OpenSSL by Hanno Böck.
2880 *) Certificate verify crash with missing PSS parameter
2882 The signature verification routines will crash with a NULL pointer
2883 dereference if presented with an ASN.1 signature using the RSA PSS
2884 algorithm and absent mask generation function parameter. Since these
2885 routines are used to verify certificate signature algorithms this can be
2886 used to crash any certificate verification operation and exploited in a
2887 DoS attack. Any application which performs certificate verification is
2888 vulnerable including OpenSSL clients and servers which enable client
2891 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2895 *) X509_ATTRIBUTE memory leak
2897 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2898 memory. This structure is used by the PKCS#7 and CMS routines so any
2899 application which reads PKCS#7 or CMS data from untrusted sources is
2900 affected. SSL/TLS is not affected.
2902 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2907 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2908 This changes the decoding behaviour for some invalid messages,
2909 though the change is mostly in the more lenient direction, and
2910 legacy behaviour is preserved as much as possible.
2913 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2915 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2917 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2919 *) Alternate chains certificate forgery
2921 During certificate verification, OpenSSL will attempt to find an
2922 alternative certificate chain if the first attempt to build such a chain
2923 fails. An error in the implementation of this logic can mean that an
2924 attacker could cause certain checks on untrusted certificates to be
2925 bypassed, such as the CA flag, enabling them to use a valid leaf
2926 certificate to act as a CA and "issue" an invalid certificate.
2928 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2932 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2934 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2935 incompatibility in the handling of HMAC. The previous ABI has now been
2939 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2941 *) Malformed ECParameters causes infinite loop
2943 When processing an ECParameters structure OpenSSL enters an infinite loop
2944 if the curve specified is over a specially malformed binary polynomial
2947 This can be used to perform denial of service against any
2948 system which processes public keys, certificate requests or
2949 certificates. This includes TLS clients and TLS servers with
2950 client authentication enabled.
2952 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2956 *) Exploitable out-of-bounds read in X509_cmp_time
2958 X509_cmp_time does not properly check the length of the ASN1_TIME
2959 string and can read a few bytes out of bounds. In addition,
2960 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2963 An attacker can use this to craft malformed certificates and CRLs of
2964 various sizes and potentially cause a segmentation fault, resulting in
2965 a DoS on applications that verify certificates or CRLs. TLS clients
2966 that verify CRLs are affected. TLS clients and servers with client
2967 authentication enabled may be affected if they use custom verification
2970 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2971 independently by Hanno Böck.
2975 *) PKCS7 crash with missing EnvelopedContent
2977 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2978 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2979 with missing content and trigger a NULL pointer dereference on parsing.
2981 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2982 structures from untrusted sources are affected. OpenSSL clients and
2983 servers are not affected.
2985 This issue was reported to OpenSSL by Michal Zalewski (Google).
2989 *) CMS verify infinite loop with unknown hash function
2991 When verifying a signedData message the CMS code can enter an infinite loop
2992 if presented with an unknown hash function OID. This can be used to perform
2993 denial of service against any system which verifies signedData messages using
2995 This issue was reported to OpenSSL by Johannes Bauer.
2999 *) Race condition handling NewSessionTicket
3001 If a NewSessionTicket is received by a multi-threaded client when attempting to
3002 reuse a previous ticket then a race condition can occur potentially leading to
3003 a double free of the ticket data.
3007 *) Only support 256-bit or stronger elliptic curves with the
3008 'ecdh_auto' setting (server) or by default (client). Of supported
3009 curves, prefer P-256 (both).
3012 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
3014 *) ClientHello sigalgs DoS fix
3016 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
3017 invalid signature algorithms extension a NULL pointer dereference will
3018 occur. This can be exploited in a DoS attack against the server.
3020 This issue was was reported to OpenSSL by David Ramos of Stanford
3023 [Stephen Henson and Matt Caswell]
3025 *) Multiblock corrupted pointer fix
3027 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
3028 feature only applies on 64 bit x86 architecture platforms that support AES
3029 NI instructions. A defect in the implementation of "multiblock" can cause
3030 OpenSSL's internal write buffer to become incorrectly set to NULL when
3031 using non-blocking IO. Typically, when the user application is using a
3032 socket BIO for writing, this will only result in a failed connection.
3033 However if some other BIO is used then it is likely that a segmentation
3034 fault will be triggered, thus enabling a potential DoS attack.
3036 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
3040 *) Segmentation fault in DTLSv1_listen fix
3042 The DTLSv1_listen function is intended to be stateless and processes the
3043 initial ClientHello from many peers. It is common for user code to loop
3044 over the call to DTLSv1_listen until a valid ClientHello is received with
3045 an associated cookie. A defect in the implementation of DTLSv1_listen means
3046 that state is preserved in the SSL object from one invocation to the next
3047 that can lead to a segmentation fault. Errors processing the initial
3048 ClientHello can trigger this scenario. An example of such an error could be
3049 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
3052 This issue was reported to OpenSSL by Per Allansson.
3056 *) Segmentation fault in ASN1_TYPE_cmp fix
3058 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
3059 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
3060 certificate signature algorithm consistency this can be used to crash any
3061 certificate verification operation and exploited in a DoS attack. Any
3062 application which performs certificate verification is vulnerable including
3063 OpenSSL clients and servers which enable client authentication.
3067 *) Segmentation fault for invalid PSS parameters fix
3069 The signature verification routines will crash with a NULL pointer
3070 dereference if presented with an ASN.1 signature using the RSA PSS
3071 algorithm and invalid parameters. Since these routines are used to verify
3072 certificate signature algorithms this can be used to crash any
3073 certificate verification operation and exploited in a DoS attack. Any
3074 application which performs certificate verification is vulnerable including
3075 OpenSSL clients and servers which enable client authentication.
3077 This issue was was reported to OpenSSL by Brian Carpenter.
3081 *) ASN.1 structure reuse memory corruption fix
3083 Reusing a structure in ASN.1 parsing may allow an attacker to cause
3084 memory corruption via an invalid write. Such reuse is and has been
3085 strongly discouraged and is believed to be rare.
3087 Applications that parse structures containing CHOICE or ANY DEFINED BY
3088 components may be affected. Certificate parsing (d2i_X509 and related
3089 functions) are however not affected. OpenSSL clients and servers are
3094 *) PKCS7 NULL pointer dereferences fix
3096 The PKCS#7 parsing code does not handle missing outer ContentInfo
3097 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
3098 missing content and trigger a NULL pointer dereference on parsing.
3100 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
3101 otherwise parse PKCS#7 structures from untrusted sources are
3102 affected. OpenSSL clients and servers are not affected.
3104 This issue was reported to OpenSSL by Michal Zalewski (Google).
3108 *) DoS via reachable assert in SSLv2 servers fix
3110 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
3111 servers that both support SSLv2 and enable export cipher suites by sending
3112 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
3114 This issue was discovered by Sean Burford (Google) and Emilia Käsper
3115 (OpenSSL development team).
3119 *) Empty CKE with client auth and DHE fix
3121 If client auth is used then a server can seg fault in the event of a DHE
3122 ciphersuite being selected and a zero length ClientKeyExchange message
3123 being sent by the client. This could be exploited in a DoS attack.
3127 *) Handshake with unseeded PRNG fix
3129 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
3130 with an unseeded PRNG. The conditions are:
3131 - The client is on a platform where the PRNG has not been seeded
3132 automatically, and the user has not seeded manually
3133 - A protocol specific client method version has been used (i.e. not
3134 SSL_client_methodv23)
3135 - A ciphersuite is used that does not require additional random data from
3136 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
3138 If the handshake succeeds then the client random that has been used will
3139 have been generated from a PRNG with insufficient entropy and therefore the
3140 output may be predictable.
3142 For example using the following command with an unseeded openssl will
3143 succeed on an unpatched platform:
3145 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
3149 *) Use After Free following d2i_ECPrivatekey error fix
3151 A malformed EC private key file consumed via the d2i_ECPrivateKey function
3152 could cause a use after free condition. This, in turn, could cause a double
3153 free in several private key parsing functions (such as d2i_PrivateKey
3154 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
3155 for applications that receive EC private keys from untrusted
3156 sources. This scenario is considered rare.
3158 This issue was discovered by the BoringSSL project and fixed in their
3163 *) X509_to_X509_REQ NULL pointer deref fix
3165 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
3166 the certificate key is invalid. This function is rarely used in practice.
3168 This issue was discovered by Brian Carpenter.
3172 *) Removed the export ciphers from the DEFAULT ciphers
3175 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
3177 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
3178 ARMv5 through ARMv8, as opposite to "locking" it to single one.
3179 So far those who have to target multiple platforms would compromise
3180 and argue that binary targeting say ARMv5 would still execute on
3181 ARMv8. "Universal" build resolves this compromise by providing
3182 near-optimal performance even on newer platforms.
3185 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
3186 (other platforms pending).
3187 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
3189 *) Add support for the SignedCertificateTimestampList certificate and
3190 OCSP response extensions from RFC6962.
3193 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3194 for corner cases. (Certain input points at infinity could lead to
3195 bogus results, with non-infinity inputs mapped to infinity too.)
3198 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
3199 This covers AES, SHA256/512 and GHASH. "Initial" means that most
3200 common cases are optimized and there still is room for further
3201 improvements. Vector Permutation AES for Altivec is also added.
3204 *) Add support for little-endian ppc64 Linux target.
3205 [Marcelo Cerri (IBM)]
3207 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
3208 SHA1, SHA256 and GHASH. "Initial" means that most common cases
3209 are optimized and there still is room for further improvements.
3210 Both 32- and 64-bit modes are supported.
3211 [Andy Polyakov, Ard Biesheuvel (Linaro)]
3213 *) Improved ARMv7 NEON support.
3216 *) Support for SPARC Architecture 2011 crypto extensions, first
3217 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
3218 SHA256/512, MD5, GHASH and modular exponentiation.
3219 [Andy Polyakov, David Miller]
3221 *) Accelerated modular exponentiation for Intel processors, a.k.a.
3223 [Shay Gueron & Vlad Krasnov (Intel Corp)]
3225 *) Support for new and upcoming Intel processors, including AVX2,
3226 BMI and SHA ISA extensions. This includes additional "stitched"
3227 implementations, AESNI-SHA256 and GCM, and multi-buffer support
3230 This work was sponsored by Intel Corp.
3233 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
3234 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
3235 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
3238 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3239 this fixes a limitation in previous versions of OpenSSL.
3242 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3243 MGF1 digest and OAEP label.
3246 *) Add EVP support for key wrapping algorithms, to avoid problems with
3247 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3248 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3249 algorithms and include tests cases.
3252 *) Add functions to allocate and set the fields of an ECDSA_METHOD
3254 [Douglas E. Engert, Steve Henson]
3256 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
3257 difference in days and seconds between two tm or ASN1_TIME structures.
3260 *) Add -rev test option to s_server to just reverse order of characters
3261 received by client and send back to server. Also prints an abbreviated
3262 summary of the connection parameters.
3265 *) New option -brief for s_client and s_server to print out a brief summary
3266 of connection parameters.
3269 *) Add callbacks for arbitrary TLS extensions.
3270 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
3272 *) New option -crl_download in several openssl utilities to download CRLs
3273 from CRLDP extension in certificates.
3276 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
3279 *) New function X509_CRL_diff to generate a delta CRL from the difference
3280 of two full CRLs. Add support to "crl" utility.
3283 *) New functions to set lookup_crls function and to retrieve
3284 X509_STORE from X509_STORE_CTX.
3287 *) Print out deprecated issuer and subject unique ID fields in
3291 *) Extend OCSP I/O functions so they can be used for simple general purpose
3292 HTTP as well as OCSP. New wrapper function which can be used to download
3293 CRLs using the OCSP API.
3296 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
3299 *) SSL_CONF* functions. These provide a common framework for application
3300 configuration using configuration files or command lines.
3303 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
3304 message callback and prints the results. Needs compile time option
3305 "enable-ssl-trace". New options to s_client and s_server to enable
3309 *) New ctrl and macro to retrieve supported points extensions.
3310 Print out extension in s_server and s_client.
3313 *) New functions to retrieve certificate signature and signature
3317 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
3321 *) New Suite B modes for TLS code. These use and enforce the requirements
3322 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
3323 only use Suite B curves. The Suite B modes can be set by using the
3324 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
3327 *) New chain verification flags for Suite B levels of security. Check
3328 algorithms are acceptable when flags are set in X509_verify_cert.
3331 *) Make tls1_check_chain return a set of flags indicating checks passed
3332 by a certificate chain. Add additional tests to handle client
3333 certificates: checks for matching certificate type and issuer name
3337 *) If an attempt is made to use a signature algorithm not in the peer
3338 preference list abort the handshake. If client has no suitable
3339 signature algorithms in response to a certificate request do not
3340 use the certificate.
3343 *) If server EC tmp key is not in client preference list abort handshake.
3346 *) Add support for certificate stores in CERT structure. This makes it
3347 possible to have different stores per SSL structure or one store in
3348 the parent SSL_CTX. Include distinct stores for certificate chain
3349 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3350 to build and store a certificate chain in CERT structure: returning
3351 an error if the chain cannot be built: this will allow applications
3352 to test if a chain is correctly configured.
3354 Note: if the CERT based stores are not set then the parent SSL_CTX
3355 store is used to retain compatibility with existing behaviour.
3359 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3360 mask based on the current session, check mask when sending client
3361 hello and checking the requested ciphersuite.
3364 *) New ctrls to retrieve and set certificate types in a certificate
3365 request message. Print out received values in s_client. If certificate
3366 types is not set with custom values set sensible values based on
3367 supported signature algorithms.
3370 *) Support for distinct client and server supported signature algorithms.
3373 *) Add certificate callback. If set this is called whenever a certificate
3374 is required by client or server. An application can decide which
3375 certificate chain to present based on arbitrary criteria: for example
3376 supported signature algorithms. Add very simple example to s_server.
3377 This fixes many of the problems and restrictions of the existing client
3378 certificate callback: for example you can now clear an existing
3379 certificate and specify the whole chain.
3382 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3383 the certificate can be used for (if anything). Set valid_flags field
3384 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3385 to have similar checks in it.
3387 Add new "cert_flags" field to CERT structure and include a "strict mode".
3388 This enforces some TLS certificate requirements (such as only permitting
3389 certificate signature algorithms contained in the supported algorithms
3390 extension) which some implementations ignore: this option should be used
3391 with caution as it could cause interoperability issues.
3394 *) Update and tidy signature algorithm extension processing. Work out
3395 shared signature algorithms based on preferences and peer algorithms
3396 and print them out in s_client and s_server. Abort handshake if no
3397 shared signature algorithms.
3400 *) Add new functions to allow customised supported signature algorithms
3401 for SSL and SSL_CTX structures. Add options to s_client and s_server
3405 *) New function SSL_certs_clear() to delete all references to certificates
3406 from an SSL structure. Before this once a certificate had been added
3407 it couldn't be removed.
3410 *) Integrate hostname, email address and IP address checking with certificate
3411 verification. New verify options supporting checking in openssl utility.
3414 *) Fixes and wildcard matching support to hostname and email checking
3415 functions. Add manual page.
3416 [Florian Weimer (Red Hat Product Security Team)]
3418 *) New functions to check a hostname email or IP address against a
3419 certificate. Add options x509 utility to print results of checks against
3423 *) Fix OCSP checking.
3424 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3426 *) Initial experimental support for explicitly trusted non-root CAs.
3427 OpenSSL still tries to build a complete chain to a root but if an
3428 intermediate CA has a trust setting included that is used. The first
3429 setting is used: whether to trust (e.g., -addtrust option to the x509
3433 *) Add -trusted_first option which attempts to find certificates in the
3434 trusted store even if an untrusted chain is also supplied.
3437 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3438 platform support for Linux and Android.
3441 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3444 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3445 When in FIPS mode the approved implementations are used as normal,
3446 when not in FIPS mode the internal unapproved versions are used instead.
3447 This means that the FIPS capable OpenSSL isn't forced to use the
3448 (often lower performance) FIPS implementations outside FIPS mode.
3451 *) Transparently support X9.42 DH parameters when calling
3452 PEM_read_bio_DHparameters. This means existing applications can handle
3453 the new parameter format automatically.
3456 *) Initial experimental support for X9.42 DH parameter format: mainly
3457 to support use of 'q' parameter for RFC5114 parameters.
3460 *) Add DH parameters from RFC5114 including test data to dhtest.
3463 *) Support for automatic EC temporary key parameter selection. If enabled
3464 the most preferred EC parameters are automatically used instead of
3465 hardcoded fixed parameters. Now a server just has to call:
3466 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3467 support ECDH and use the most appropriate parameters.
3470 *) Enhance and tidy EC curve and point format TLS extension code. Use
3471 static structures instead of allocation if default values are used.
3472 New ctrls to set curves we wish to support and to retrieve shared curves.
3473 Print out shared curves in s_server. New options to s_server and s_client
3474 to set list of supported curves.
3477 *) New ctrls to retrieve supported signature algorithms and
3478 supported curve values as an array of NIDs. Extend openssl utility
3479 to print out received values.
3482 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3483 between NIDs and the more common NIST names such as "P-256". Enhance
3484 ecparam utility and ECC method to recognise the NIST names for curves.
3487 *) Enhance SSL/TLS certificate chain handling to support different
3488 chains for each certificate instead of one chain in the parent SSL_CTX.
3491 *) Support for fixed DH ciphersuite client authentication: where both
3492 server and client use DH certificates with common parameters.
3495 *) Support for fixed DH ciphersuites: those requiring DH server
3499 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3501 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3502 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3503 X509_CINF_get_signature were reverted post internal team review.
3505 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3507 *) Build fixes for the Windows and OpenVMS platforms
3508 [Matt Caswell and Richard Levitte]
3510 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3512 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3513 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3514 dereference. This could lead to a Denial Of Service attack. Thanks to
3515 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3519 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3520 dtls1_buffer_record function under certain conditions. In particular this
3521 could occur if an attacker sent repeated DTLS records with the same
3522 sequence number but for the next epoch. The memory leak could be exploited
3523 by an attacker in a Denial of Service attack through memory exhaustion.
3524 Thanks to Chris Mueller for reporting this issue.
3528 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3529 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3530 method would be set to NULL which could later result in a NULL pointer
3531 dereference. Thanks to Frank Schmirler for reporting this issue.
3535 *) Abort handshake if server key exchange message is omitted for ephemeral
3538 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3539 reporting this issue.
3543 *) Remove non-export ephemeral RSA code on client and server. This code
3544 violated the TLS standard by allowing the use of temporary RSA keys in
3545 non-export ciphersuites and could be used by a server to effectively
3546 downgrade the RSA key length used to a value smaller than the server
3547 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3548 INRIA or reporting this issue.
3552 *) Fixed issue where DH client certificates are accepted without verification.
3553 An OpenSSL server will accept a DH certificate for client authentication
3554 without the certificate verify message. This effectively allows a client to
3555 authenticate without the use of a private key. This only affects servers
3556 which trust a client certificate authority which issues certificates
3557 containing DH keys: these are extremely rare and hardly ever encountered.
3558 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3563 *) Ensure that the session ID context of an SSL is updated when its
3564 SSL_CTX is updated via SSL_set_SSL_CTX.
3566 The session ID context is typically set from the parent SSL_CTX,
3567 and can vary with the CTX.
3570 *) Fix various certificate fingerprint issues.
3572 By using non-DER or invalid encodings outside the signed portion of a
3573 certificate the fingerprint can be changed without breaking the signature.
3574 Although no details of the signed portion of the certificate can be changed
3575 this can cause problems with some applications: e.g. those using the
3576 certificate fingerprint for blacklists.
3578 1. Reject signatures with non zero unused bits.
3580 If the BIT STRING containing the signature has non zero unused bits reject
3581 the signature. All current signature algorithms require zero unused bits.
3583 2. Check certificate algorithm consistency.
3585 Check the AlgorithmIdentifier inside TBS matches the one in the
3586 certificate signature. NB: this will result in signature failure
3587 errors for some broken certificates.
3589 Thanks to Konrad Kraszewski from Google for reporting this issue.
3591 3. Check DSA/ECDSA signatures use DER.
3593 Re-encode DSA/ECDSA signatures and compare with the original received
3594 signature. Return an error if there is a mismatch.
3596 This will reject various cases including garbage after signature
3597 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3598 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3599 (negative or with leading zeroes).
3601 Further analysis was conducted and fixes were developed by Stephen Henson
3602 of the OpenSSL core team.
3607 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3608 results on some platforms, including x86_64. This bug occurs at random
3609 with a very low probability, and is not known to be exploitable in any
3610 way, though its exact impact is difficult to determine. Thanks to Pieter
3611 Wuille (Blockstream) who reported this issue and also suggested an initial
3612 fix. Further analysis was conducted by the OpenSSL development team and
3613 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3614 the OpenSSL core team.
3618 *) Do not resume sessions on the server if the negotiated protocol
3619 version does not match the session's version. Resuming with a different
3620 version, while not strictly forbidden by the RFC, is of questionable
3621 sanity and breaks all known clients.
3622 [David Benjamin, Emilia Käsper]
3624 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3625 early CCS messages during renegotiation. (Note that because
3626 renegotiation is encrypted, this early CCS was not exploitable.)
3629 *) Tighten client-side session ticket handling during renegotiation:
3630 ensure that the client only accepts a session ticket if the server sends
3631 the extension anew in the ServerHello. Previously, a TLS client would
3632 reuse the old extension state and thus accept a session ticket if one was
3633 announced in the initial ServerHello.
3635 Similarly, ensure that the client requires a session ticket if one
3636 was advertised in the ServerHello. Previously, a TLS client would
3637 ignore a missing NewSessionTicket message.
3640 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3642 *) SRTP Memory Leak.
3644 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3645 sends a carefully crafted handshake message, to cause OpenSSL to fail
3646 to free up to 64k of memory causing a memory leak. This could be
3647 exploited in a Denial Of Service attack. This issue affects OpenSSL
3648 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3649 whether SRTP is used or configured. Implementations of OpenSSL that
3650 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3652 The fix was developed by the OpenSSL team.
3656 *) Session Ticket Memory Leak.
3658 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3659 integrity of that ticket is first verified. In the event of a session
3660 ticket integrity check failing, OpenSSL will fail to free memory
3661 causing a memory leak. By sending a large number of invalid session
3662 tickets an attacker could exploit this issue in a Denial Of Service
3667 *) Build option no-ssl3 is incomplete.
3669 When OpenSSL is configured with "no-ssl3" as a build option, servers
3670 could accept and complete a SSL 3.0 handshake, and clients could be
3671 configured to send them.
3673 [Akamai and the OpenSSL team]
3675 *) Add support for TLS_FALLBACK_SCSV.
3676 Client applications doing fallback retries should call
3677 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3679 [Adam Langley, Bodo Moeller]
3681 *) Add additional DigestInfo checks.
3683 Re-encode DigestInto in DER and check against the original when
3684 verifying RSA signature: this will reject any improperly encoded
3685 DigestInfo structures.
3687 Note: this is a precautionary measure and no attacks are currently known.
3691 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3693 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3694 SRP code can be overrun an internal buffer. Add sanity check that
3695 g, A, B < N to SRP code.
3697 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3698 Group for discovering this issue.
3702 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3703 TLS 1.0 instead of higher protocol versions when the ClientHello message
3704 is badly fragmented. This allows a man-in-the-middle attacker to force a
3705 downgrade to TLS 1.0 even if both the server and the client support a
3706 higher protocol version, by modifying the client's TLS records.
3708 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3709 researching this issue.
3713 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3714 to a denial of service attack. A malicious server can crash the client
3715 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3716 ciphersuite and sending carefully crafted handshake messages.
3718 Thanks to Felix Gröbert (Google) for discovering and researching this
3723 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3724 to leak memory. This can be exploited through a Denial of Service attack.
3725 Thanks to Adam Langley for discovering and researching this issue.
3729 *) An attacker can force openssl to consume large amounts of memory whilst
3730 processing DTLS handshake messages. This can be exploited through a
3731 Denial of Service attack.
3732 Thanks to Adam Langley for discovering and researching this issue.
3736 *) An attacker can force an error condition which causes openssl to crash
3737 whilst processing DTLS packets due to memory being freed twice. This
3738 can be exploited through a Denial of Service attack.
3739 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3744 *) If a multithreaded client connects to a malicious server using a resumed
3745 session and the server sends an ec point format extension it could write
3746 up to 255 bytes to freed memory.
3748 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3753 *) A malicious server can crash an OpenSSL client with a null pointer
3754 dereference (read) by specifying an SRP ciphersuite even though it was not
3755 properly negotiated with the client. This can be exploited through a
3756 Denial of Service attack.
3758 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3759 discovering and researching this issue.
3763 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3764 X509_name_oneline, X509_name_print_ex et al. to leak some information
3765 from the stack. Applications may be affected if they echo pretty printing
3766 output to the attacker.
3768 Thanks to Ivan Fratric (Google) for discovering this issue.
3770 [Emilia Käsper, and Steve Henson]
3772 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3773 for corner cases. (Certain input points at infinity could lead to
3774 bogus results, with non-infinity inputs mapped to infinity too.)
3777 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3779 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3780 handshake can force the use of weak keying material in OpenSSL
3781 SSL/TLS clients and servers.
3783 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3784 researching this issue. (CVE-2014-0224)
3785 [KIKUCHI Masashi, Steve Henson]
3787 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3788 OpenSSL DTLS client the code can be made to recurse eventually crashing
3791 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3793 [Imre Rad, Steve Henson]
3795 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3796 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3797 client or server. This is potentially exploitable to run arbitrary
3798 code on a vulnerable client or server.
3800 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3801 [Jüri Aedla, Steve Henson]
3803 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3804 are subject to a denial of service attack.
3806 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3807 this issue. (CVE-2014-3470)
3808 [Felix Gröbert, Ivan Fratric, Steve Henson]
3810 *) Harmonize version and its documentation. -f flag is used to display
3812 [mancha <mancha1@zoho.com>]
3814 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3815 in i2d_ECPrivateKey.
3816 [mancha <mancha1@zoho.com>]
3818 *) Fix some double frees. These are not thought to be exploitable.
3819 [mancha <mancha1@zoho.com>]
3821 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3823 *) A missing bounds check in the handling of the TLS heartbeat extension
3824 can be used to reveal up to 64k of memory to a connected client or
3827 Thanks for Neel Mehta of Google Security for discovering this bug and to
3828 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3829 preparing the fix (CVE-2014-0160)
3830 [Adam Langley, Bodo Moeller]
3832 *) Fix for the attack described in the paper "Recovering OpenSSL
3833 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3834 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3835 http://eprint.iacr.org/2014/140
3837 Thanks to Yuval Yarom and Naomi Benger for discovering this
3838 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3839 [Yuval Yarom and Naomi Benger]
3841 *) TLS pad extension: draft-agl-tls-padding-03
3843 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3844 TLS client Hello record length value would otherwise be > 255 and
3845 less that 512 pad with a dummy extension containing zeroes so it
3846 is at least 512 bytes long.
3848 [Adam Langley, Steve Henson]
3850 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3852 *) Fix for TLS record tampering bug. A carefully crafted invalid
3853 handshake could crash OpenSSL with a NULL pointer exception.
3854 Thanks to Anton Johansson for reporting this issues.
3857 *) Keep original DTLS digest and encryption contexts in retransmission
3858 structures so we can use the previous session parameters if they need
3859 to be resent. (CVE-2013-6450)
3862 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3863 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3864 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3865 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3866 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3867 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3868 [Rob Stradling, Adam Langley]
3870 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3872 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3873 supporting platforms or when small records were transferred.
3874 [Andy Polyakov, Steve Henson]
3876 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3878 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3880 This addresses the flaw in CBC record processing discovered by
3881 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3882 at: http://www.isg.rhul.ac.uk/tls/
3884 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3885 Security Group at Royal Holloway, University of London
3886 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3887 Emilia Käsper for the initial patch.
3889 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3891 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3892 ciphersuites which can be exploited in a denial of service attack.
3893 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3894 and detecting this bug and to Wolfgang Ettlinger
3895 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3899 *) Return an error when checking OCSP signatures when key is NULL.
3900 This fixes a DoS attack. (CVE-2013-0166)
3903 *) Make openssl verify return errors.
3904 [Chris Palmer <palmer@google.com> and Ben Laurie]
3906 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3907 the right response is stapled. Also change SSL_get_certificate()
3908 so it returns the certificate actually sent.
3909 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3910 [Rob Stradling <rob.stradling@comodo.com>]
3912 *) Fix possible deadlock when decoding public keys.
3915 *) Don't use TLS 1.0 record version number in initial client hello
3919 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3921 *) Sanity check record length before skipping explicit IV in TLS
3922 1.2, 1.1 and DTLS to fix DoS attack.
3924 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3925 fuzzing as a service testing platform.
3929 *) Initialise tkeylen properly when encrypting CMS messages.
3930 Thanks to Solar Designer of Openwall for reporting this issue.
3933 *) In FIPS mode don't try to use composite ciphers as they are not
3937 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3939 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3940 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3941 mean any application compiled against OpenSSL 1.0.0 headers setting
3942 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3943 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3944 0x10000000L Any application which was previously compiled against
3945 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3946 will need to be recompiled as a result. Letting be results in
3947 inability to disable specifically TLS 1.1 and in client context,
3948 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3951 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3952 disable just protocol X, but all protocols above X *if* there are
3953 protocols *below* X still enabled. In more practical terms it means
3954 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3955 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3956 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3960 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3962 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3963 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3964 in CRYPTO_realloc_clean.
3966 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3967 issue and to Adam Langley <agl@chromium.org> for fixing it.
3969 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3971 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3974 *) Workarounds for some broken servers that "hang" if a client hello
3975 record length exceeds 255 bytes.
3977 1. Do not use record version number > TLS 1.0 in initial client
3978 hello: some (but not all) hanging servers will now work.
3979 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3980 the number of ciphers sent in the client hello. This should be
3981 set to an even number, such as 50, for example by passing:
3982 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3983 Most broken servers should now work.
3984 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3985 TLS 1.2 client support entirely.
3988 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3991 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3993 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3994 STRING form instead of a DigestInfo.
3997 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3998 and the RSA_sign/RSA_verify functions. This was made more apparent when
3999 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
4000 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
4001 the correct format in RSA_verify so both forms transparently work.
4004 *) Some servers which support TLS 1.0 can choke if we initially indicate
4005 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
4006 encrypted premaster secret. As a workaround use the maximum permitted
4007 client version in client hello, this should keep such servers happy
4008 and still work with previous versions of OpenSSL.
4011 *) Add support for TLS/DTLS heartbeats.
4012 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4014 *) Add support for SCTP.
4015 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4017 *) Improved PRNG seeding for VOS.
4018 [Paul Green <Paul.Green@stratus.com>]
4020 *) Extensive assembler packs updates, most notably:
4022 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
4023 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
4024 - x86_64: bit-sliced AES implementation;
4025 - ARM: NEON support, contemporary platforms optimizations;
4026 - s390x: z196 support;
4027 - *: GHASH and GF(2^m) multiplication implementations;
4031 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
4032 (removal of unnecessary code)
4033 [Peter Sylvester <peter.sylvester@edelweb.fr>]
4035 *) Add TLS key material exporter from RFC 5705.
4038 *) Add DTLS-SRTP negotiation from RFC 5764.
4041 *) Add Next Protocol Negotiation,
4042 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
4043 disabled with a no-npn flag to config or Configure. Code donated
4045 [Adam Langley <agl@google.com> and Ben Laurie]
4047 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
4048 NIST-P256, NIST-P521, with constant-time single point multiplication on
4049 typical inputs. Compiler support for the nonstandard type __uint128_t is
4050 required to use this (present in gcc 4.4 and later, for 64-bit builds).
4051 Code made available under Apache License version 2.0.
4053 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
4054 line to include this in your build of OpenSSL, and run "make depend" (or
4055 "make update"). This enables the following EC_METHODs:
4057 EC_GFp_nistp224_method()
4058 EC_GFp_nistp256_method()
4059 EC_GFp_nistp521_method()
4061 EC_GROUP_new_by_curve_name() will automatically use these (while
4062 EC_GROUP_new_curve_GFp() currently prefers the more flexible
4064 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
4066 *) Use type ossl_ssize_t instead of ssize_t which isn't available on
4067 all platforms. Move ssize_t definition from e_os.h to the public
4068 header file e_os2.h as it now appears in public header file cms.h
4071 *) New -sigopt option to the ca, req and x509 utilities. Additional
4072 signature parameters can be passed using this option and in
4076 *) Add RSA PSS signing function. This will generate and set the
4077 appropriate AlgorithmIdentifiers for PSS based on those in the
4078 corresponding EVP_MD_CTX structure. No application support yet.
4081 *) Support for companion algorithm specific ASN1 signing routines.
4082 New function ASN1_item_sign_ctx() signs a pre-initialised
4083 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
4084 the appropriate parameters.
4087 *) Add new algorithm specific ASN1 verification initialisation function
4088 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
4089 handling will be the same no matter what EVP_PKEY_METHOD is used.
4090 Add a PSS handler to support verification of PSS signatures: checked
4091 against a number of sample certificates.
4094 *) Add signature printing for PSS. Add PSS OIDs.
4095 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
4097 *) Add algorithm specific signature printing. An individual ASN1 method
4098 can now print out signatures instead of the standard hex dump.
4100 More complex signatures (e.g. PSS) can print out more meaningful
4101 information. Include DSA version that prints out the signature
4105 *) Password based recipient info support for CMS library: implementing
4109 *) Split password based encryption into PBES2 and PBKDF2 functions. This
4110 neatly separates the code into cipher and PBE sections and is required
4111 for some algorithms that split PBES2 into separate pieces (such as
4112 password based CMS).
4115 *) Session-handling fixes:
4116 - Fix handling of connections that are resuming with a session ID,
4117 but also support Session Tickets.
4118 - Fix a bug that suppressed issuing of a new ticket if the client
4119 presented a ticket with an expired session.
4120 - Try to set the ticket lifetime hint to something reasonable.
4121 - Make tickets shorter by excluding irrelevant information.
4122 - On the client side, don't ignore renewed tickets.
4123 [Adam Langley, Bodo Moeller (Google)]
4125 *) Fix PSK session representation.
4128 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
4130 This work was sponsored by Intel.
4133 *) Add GCM support to TLS library. Some custom code is needed to split
4134 the IV between the fixed (from PRF) and explicit (from TLS record)
4135 portions. This adds all GCM ciphersuites supported by RFC5288 and
4136 RFC5289. Generalise some AES* cipherstrings to include GCM and
4137 add a special AESGCM string for GCM only.
4140 *) Expand range of ctrls for AES GCM. Permit setting invocation
4141 field on decrypt and retrieval of invocation field only on encrypt.
4144 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
4145 As required by RFC5289 these ciphersuites cannot be used if for
4146 versions of TLS earlier than 1.2.
4149 *) For FIPS capable OpenSSL interpret a NULL default public key method
4150 as unset and return the appropriate default but do *not* set the default.
4151 This means we can return the appropriate method in applications that
4152 switch between FIPS and non-FIPS modes.
4155 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
4156 ENGINE is used then we cannot handle that in the FIPS module so we
4157 keep original code iff non-FIPS operations are allowed.
4160 *) Add -attime option to openssl utilities.
4161 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
4163 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
4166 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
4167 FIPS EC methods unconditionally for now.
4170 *) New build option no-ec2m to disable characteristic 2 code.
4173 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
4174 all cases can be covered as some introduce binary incompatibilities.
4177 *) Redirect RSA operations to FIPS module including keygen,
4178 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
4181 *) Add similar low level API blocking to ciphers.
4184 *) Low level digest APIs are not approved in FIPS mode: any attempt
4185 to use these will cause a fatal error. Applications that *really* want
4186 to use them can use the private_* version instead.
4189 *) Redirect cipher operations to FIPS module for FIPS builds.
4192 *) Redirect digest operations to FIPS module for FIPS builds.
4195 *) Update build system to add "fips" flag which will link in fipscanister.o
4196 for static and shared library builds embedding a signature if needed.
4199 *) Output TLS supported curves in preference order instead of numerical
4200 order. This is currently hardcoded for the highest order curves first.
4201 This should be configurable so applications can judge speed vs strength.
4204 *) Add TLS v1.2 server support for client authentication.
4207 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
4211 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
4212 FIPS modules versions.
4215 *) Add TLS v1.2 client side support for client authentication. Keep cache
4216 of handshake records longer as we don't know the hash algorithm to use
4217 until after the certificate request message is received.
4220 *) Initial TLS v1.2 client support. Add a default signature algorithms
4221 extension including all the algorithms we support. Parse new signature
4222 format in client key exchange. Relax some ECC signing restrictions for
4223 TLS v1.2 as indicated in RFC5246.
4226 *) Add server support for TLS v1.2 signature algorithms extension. Switch
4227 to new signature format when needed using client digest preference.
4228 All server ciphersuites should now work correctly in TLS v1.2. No client
4229 support yet and no support for client certificates.
4232 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
4233 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
4234 ciphersuites. At present only RSA key exchange ciphersuites work with
4235 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
4236 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
4237 and version checking.
4240 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
4241 with this defined it will not be affected by any changes to ssl internal
4242 structures. Add several utility functions to allow openssl application
4243 to work with OPENSSL_NO_SSL_INTERN defined.
4246 *) A long standing patch to add support for SRP from EdelWeb (Peter
4247 Sylvester and Christophe Renou) was integrated.
4248 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
4249 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
4252 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
4255 *) Permit abbreviated handshakes when renegotiating using the function
4256 SSL_renegotiate_abbreviated().
4257 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4259 *) Add call to ENGINE_register_all_complete() to
4260 ENGINE_load_builtin_engines(), so some implementations get used
4261 automatically instead of needing explicit application support.
4264 *) Add support for TLS key exporter as described in RFC5705.
4265 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
4267 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
4268 a few changes are required:
4270 Add SSL_OP_NO_TLSv1_1 flag.
4271 Add TLSv1_1 methods.
4272 Update version checking logic to handle version 1.1.
4273 Add explicit IV handling (ported from DTLS code).
4274 Add command line options to s_client/s_server.
4277 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
4279 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
4280 in CMS and PKCS7 code. When RSA decryption fails use a random key for
4281 content decryption and always return the same error. Note: this attack
4282 needs on average 2^20 messages so it only affects automated senders. The
4283 old behaviour can be re-enabled in the CMS code by setting the
4284 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
4285 an MMA defence is not necessary.
4286 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
4287 this issue. (CVE-2012-0884)
4290 *) Fix CVE-2011-4619: make sure we really are receiving a
4291 client hello before rejecting multiple SGC restarts. Thanks to
4292 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
4295 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
4297 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
4298 Thanks to Antonio Martin, Enterprise Secure Access Research and
4299 Development, Cisco Systems, Inc. for discovering this bug and
4300 preparing a fix. (CVE-2012-0050)
4303 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
4305 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
4306 of the Vaudenay padding oracle attack on CBC mode encryption
4307 which enables an efficient plaintext recovery attack against
4308 the OpenSSL implementation of DTLS. Their attack exploits timing
4309 differences arising during decryption processing. A research
4310 paper describing this attack can be found at:
4311 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
4312 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4313 Security Group at Royal Holloway, University of London
4314 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
4315 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
4316 for preparing the fix. (CVE-2011-4108)
4317 [Robin Seggelmann, Michael Tuexen]
4319 *) Clear bytes used for block padding of SSL 3.0 records.
4321 [Adam Langley (Google)]
4323 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
4324 Kadianakis <desnacked@gmail.com> for discovering this issue and
4325 Adam Langley for preparing the fix. (CVE-2011-4619)
4326 [Adam Langley (Google)]
4328 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
4329 [Andrey Kulikov <amdeich@gmail.com>]
4331 *) Prevent malformed RFC3779 data triggering an assertion failure.
4332 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
4333 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
4334 [Rob Austein <sra@hactrn.net>]
4336 *) Improved PRNG seeding for VOS.
4337 [Paul Green <Paul.Green@stratus.com>]
4339 *) Fix ssl_ciph.c set-up race.
4340 [Adam Langley (Google)]
4342 *) Fix spurious failures in ecdsatest.c.
4343 [Emilia Käsper (Google)]
4345 *) Fix the BIO_f_buffer() implementation (which was mixing different
4346 interpretations of the '..._len' fields).
4347 [Adam Langley (Google)]
4349 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4350 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4351 threads won't reuse the same blinding coefficients.
4353 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4354 lock to call BN_BLINDING_invert_ex, and avoids one use of
4355 BN_BLINDING_update for each BN_BLINDING structure (previously,
4356 the last update always remained unused).
4357 [Emilia Käsper (Google)]
4359 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4360 [Bob Buckholz (Google)]
4362 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4364 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4365 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4366 [Kaspar Brand <ossl@velox.ch>]
4368 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4369 for multi-threaded use of ECDH. (CVE-2011-3210)
4370 [Adam Langley (Google)]
4372 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4375 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4376 signature public key algorithm by using OID xref utilities instead.
4377 Before this you could only use some ECC ciphersuites with SHA1 only.
4380 *) Add protection against ECDSA timing attacks as mentioned in the paper
4381 by Billy Bob Brumley and Nicola Tuveri, see:
4383 http://eprint.iacr.org/2011/232.pdf
4385 [Billy Bob Brumley and Nicola Tuveri]
4387 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4389 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4390 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4392 *) Fix bug in string printing code: if *any* escaping is enabled we must
4393 escape the escape character (backslash) or the resulting string is
4397 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4399 *) Disable code workaround for ancient and obsolete Netscape browsers
4400 and servers: an attacker can use it in a ciphersuite downgrade attack.
4401 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4404 *) Fixed J-PAKE implementation error, originally discovered by
4405 Sebastien Martini, further info and confirmation from Stefan
4406 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4409 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4411 *) Fix extension code to avoid race conditions which can result in a buffer
4412 overrun vulnerability: resumed sessions must not be modified as they can
4413 be shared by multiple threads. CVE-2010-3864
4416 *) Fix WIN32 build system to correctly link an ENGINE directory into
4420 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4422 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4424 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4426 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4428 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4429 context. The operation can be customised via the ctrl mechanism in
4430 case ENGINEs want to include additional functionality.
4433 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4436 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4437 output hashes compatible with older versions of OpenSSL.
4438 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4440 *) Fix compression algorithm handling: if resuming a session use the
4441 compression algorithm of the resumed session instead of determining
4442 it from client hello again. Don't allow server to change algorithm.
4445 *) Add load_crls() function to apps tidying load_certs() too. Add option
4446 to verify utility to allow additional CRLs to be included.
4449 *) Update OCSP request code to permit adding custom headers to the request:
4450 some responders need this.
4453 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4455 [Julia Lawall <julia@diku.dk>]
4457 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4458 needlessly dereferenced structures, used obsolete functions and
4459 didn't handle all updated verify codes correctly.
4462 *) Disable MD2 in the default configuration.
4465 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4466 indicate the initial BIO being pushed or popped. This makes it possible
4467 to determine whether the BIO is the one explicitly called or as a result
4468 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4469 it handles reference counts correctly and doesn't zero out the I/O bio
4470 when it is not being explicitly popped. WARNING: applications which
4471 included workarounds for the old buggy behaviour will need to be modified
4472 or they could free up already freed BIOs.
4475 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4476 renaming to all platforms (within the 0.9.8 branch, this was
4477 done conditionally on Netware platforms to avoid a name clash).
4478 [Guenter <lists@gknw.net>]
4480 *) Add ECDHE and PSK support to DTLS.
4481 [Michael Tuexen <tuexen@fh-muenster.de>]
4483 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4487 *) Add "missing" function EVP_MD_flags() (without this the only way to
4488 retrieve a digest flags is by accessing the structure directly. Update
4489 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4490 or cipher is registered as in the "from" argument. Print out all
4491 registered digests in the dgst usage message instead of manually
4492 attempting to work them out.
4495 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4496 this allows the use of compression and extensions. Change default cipher
4497 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4498 by default unless an application cipher string requests it.
4501 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4502 key ids to find matching certificates and keys but some PKCS#12 files
4503 don't follow the (somewhat unwritten) rules and this strategy fails.
4504 Now just gather all certificates together and the first private key
4505 then look for the first certificate that matches the key.
4508 *) Support use of registered digest and cipher names for dgst and cipher
4509 commands instead of having to add each one as a special case. So now
4516 openssl dgst -sha256 foo
4518 and this works for ENGINE based algorithms too.
4522 *) Update Gost ENGINE to support parameter files.
4523 [Victor B. Wagner <vitus@cryptocom.ru>]
4525 *) Support GeneralizedTime in ca utility.
4526 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4528 *) Enhance the hash format used for certificate directory links. The new
4529 form uses the canonical encoding (meaning equivalent names will work
4530 even if they aren't identical) and uses SHA1 instead of MD5. This form
4531 is incompatible with the older format and as a result c_rehash should
4532 be used to rebuild symbolic links.
4535 *) Make PKCS#8 the default write format for private keys, replacing the
4536 traditional format. This form is standardised, more secure and doesn't
4537 include an implicit MD5 dependency.
4540 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4541 committed to OpenSSL should pass this lot as a minimum.
4544 *) Add session ticket override functionality for use by EAP-FAST.
4545 [Jouni Malinen <j@w1.fi>]
4547 *) Modify HMAC functions to return a value. Since these can be implemented
4548 in an ENGINE errors can occur.
4551 *) Type-checked OBJ_bsearch_ex.
4554 *) Type-checked OBJ_bsearch. Also some constification necessitated
4555 by type-checking. Still to come: TXT_DB, bsearch(?),
4556 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4560 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4561 seconds to a tm structure directly, instead of going through OS
4562 specific date routines. This avoids any issues with OS routines such
4563 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4564 and X509_time_adj_ex() to cover the extended range. The existing
4565 X509_time_adj() is still usable and will no longer have any date issues.
4568 *) Delta CRL support. New use deltas option which will attempt to locate
4569 and search any appropriate delta CRLs available.
4571 This work was sponsored by Google.
4574 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4575 code and add additional score elements. Validate alternate CRL paths
4576 as part of the CRL checking and indicate a new error "CRL path validation
4577 error" in this case. Applications wanting additional details can use
4578 the verify callback and check the new "parent" field. If this is not
4579 NULL CRL path validation is taking place. Existing applications won't
4580 see this because it requires extended CRL support which is off by
4583 This work was sponsored by Google.
4586 *) Support for freshest CRL extension.
4588 This work was sponsored by Google.
4591 *) Initial indirect CRL support. Currently only supported in the CRLs
4592 passed directly and not via lookup. Process certificate issuer
4593 CRL entry extension and lookup CRL entries by bother issuer name
4594 and serial number. Check and process CRL issuer entry in IDP extension.
4596 This work was sponsored by Google.
4599 *) Add support for distinct certificate and CRL paths. The CRL issuer
4600 certificate is validated separately in this case. Only enabled if
4601 an extended CRL support flag is set: this flag will enable additional
4602 CRL functionality in future.
4604 This work was sponsored by Google.
4607 *) Add support for policy mappings extension.
4609 This work was sponsored by Google.
4612 *) Fixes to pathlength constraint, self issued certificate handling,
4613 policy processing to align with RFC3280 and PKITS tests.
4615 This work was sponsored by Google.
4618 *) Support for name constraints certificate extension. DN, email, DNS
4619 and URI types are currently supported.
4621 This work was sponsored by Google.
4624 *) To cater for systems that provide a pointer-based thread ID rather
4625 than numeric, deprecate the current numeric thread ID mechanism and
4626 replace it with a structure and associated callback type. This
4627 mechanism allows a numeric "hash" to be extracted from a thread ID in
4628 either case, and on platforms where pointers are larger than 'long',
4629 mixing is done to help ensure the numeric 'hash' is usable even if it
4630 can't be guaranteed unique. The default mechanism is to use "&errno"
4631 as a pointer-based thread ID to distinguish between threads.
4633 Applications that want to provide their own thread IDs should now use
4634 CRYPTO_THREADID_set_callback() to register a callback that will call
4635 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4637 Note that ERR_remove_state() is now deprecated, because it is tied
4638 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4639 to free the current thread's error state should be replaced by
4640 ERR_remove_thread_state(NULL).
4642 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4643 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4644 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4645 application was previously providing a numeric thread callback that
4646 was inappropriate for distinguishing threads, then uniqueness might
4647 have been obtained with &errno that happened immediately in the
4648 intermediate development versions of OpenSSL; this is no longer the
4649 case, the numeric thread callback will now override the automatic use
4651 [Geoff Thorpe, with help from Bodo Moeller]
4653 *) Initial support for different CRL issuing certificates. This covers a
4654 simple case where the self issued certificates in the chain exist and
4655 the real CRL issuer is higher in the existing chain.
4657 This work was sponsored by Google.
4660 *) Removed effectively defunct crypto/store from the build.
4663 *) Revamp of STACK to provide stronger type-checking. Still to come:
4664 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4665 ASN1_STRING, CONF_VALUE.
4668 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4669 RAM on SSL connections. This option can save about 34k per idle SSL.
4672 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4673 STACK, TXT_DB, bsearch, qsort.
4676 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4677 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4678 support for data, signedData, compressedData, digestedData and
4679 encryptedData, envelopedData types included. Scripts to check against
4680 RFC4134 examples draft and interop and consistency checks of many
4681 content types and variants.
4684 *) Add options to enc utility to support use of zlib compression BIO.
4687 *) Extend mk1mf to support importing of options and assembly language
4688 files from Configure script, currently only included in VC-WIN32.
4689 The assembly language rules can now optionally generate the source
4690 files from the associated perl scripts.
4693 *) Implement remaining functionality needed to support GOST ciphersuites.
4694 Interop testing has been performed using CryptoPro implementations.
4695 [Victor B. Wagner <vitus@cryptocom.ru>]
4697 *) s390x assembler pack.
4700 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4704 *) Implement Opaque PRF Input TLS extension as specified in
4705 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4706 official specification yet and no extension type assignment by
4707 IANA exists, this extension (for now) will have to be explicitly
4708 enabled when building OpenSSL by providing the extension number
4709 to use. For example, specify an option
4711 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4713 to the "config" or "Configure" script to enable the extension,
4714 assuming extension number 0x9527 (which is a completely arbitrary
4715 and unofficial assignment based on the MD5 hash of the Internet
4716 Draft). Note that by doing so, you potentially lose
4717 interoperability with other TLS implementations since these might
4718 be using the same extension number for other purposes.
4720 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4721 opaque PRF input value to use in the handshake. This will create
4722 an internal copy of the length-'len' string at 'src', and will
4723 return non-zero for success.
4725 To get more control and flexibility, provide a callback function
4728 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4729 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4733 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4736 Callback function 'cb' will be called in handshakes, and is
4737 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4738 Argument 'arg' is for application purposes (the value as given to
4739 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4740 be provided to the callback function). The callback function
4741 has to return non-zero to report success: usually 1 to use opaque
4742 PRF input just if possible, or 2 to enforce use of the opaque PRF
4743 input. In the latter case, the library will abort the handshake
4744 if opaque PRF input is not successfully negotiated.
4746 Arguments 'peerinput' and 'len' given to the callback function
4747 will always be NULL and 0 in the case of a client. A server will
4748 see the client's opaque PRF input through these variables if
4749 available (NULL and 0 otherwise). Note that if the server
4750 provides an opaque PRF input, the length must be the same as the
4751 length of the client's opaque PRF input.
4753 Note that the callback function will only be called when creating
4754 a new session (session resumption can resume whatever was
4755 previously negotiated), and will not be called in SSL 2.0
4756 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4757 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4758 for applications that need to enforce opaque PRF input.
4762 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4765 [Victor B. Wagner <vitus@cryptocom.ru>]
4767 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4768 RFC4507bis. The encrypted ticket format is an encrypted encoded
4769 SSL_SESSION structure, that way new session features are automatically
4772 If a client application caches session in an SSL_SESSION structure
4773 support is transparent because tickets are now stored in the encoded
4776 The SSL_CTX structure automatically generates keys for ticket
4777 protection in servers so again support should be possible
4778 with no application modification.
4780 If a client or server wishes to disable RFC4507 support then the option
4781 SSL_OP_NO_TICKET can be set.
4783 Add a TLS extension debugging callback to allow the contents of any client
4784 or server extensions to be examined.
4786 This work was sponsored by Google.
4789 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4790 OpenSSL should now compile cleanly on gcc 4.2
4791 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4793 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4794 support including streaming MAC support: this is required for GOST
4795 ciphersuite support.
4796 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4798 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4799 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4800 to output in BER and PEM format.
4803 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4804 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4805 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4806 ENGINE support for HMAC keys which are unextractable. New -mac and
4807 -macopt options to dgst utility.
4810 *) New option -sigopt to dgst utility. Update dgst to use
4811 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4812 alternative signing parameters such as X9.31 or PSS in the dgst
4816 *) Change ssl_cipher_apply_rule(), the internal function that does
4817 the work each time a ciphersuite string requests enabling
4818 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4819 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4820 the order of disabled ciphersuites such that those ciphersuites
4821 that most recently went from enabled to disabled not only stay
4822 in order with respect to each other, but also have higher priority
4823 than other disabled ciphersuites the next time ciphersuites are
4826 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4827 the same ciphersuites as with "HIGH" alone, but in a specific
4828 order where the PSK ciphersuites come first (since they are the
4829 most recently disabled ciphersuites when "HIGH" is parsed).
4831 Also, change ssl_create_cipher_list() (using this new
4832 functionality) such that between otherwise identical
4833 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4837 *) Change ssl_create_cipher_list() so that it automatically
4838 arranges the ciphersuites in reasonable order before starting
4839 to process the rule string. Thus, the definition for "DEFAULT"
4840 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4841 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4842 This makes it much easier to arrive at a reasonable default order
4843 in applications for which anonymous ciphers are OK (meaning
4844 that you can't actually use DEFAULT).
4845 [Bodo Moeller; suggested by Victor Duchovni]
4847 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4848 processing) into multiple integers instead of setting
4849 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4850 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4851 (These masks as well as the individual bit definitions are hidden
4852 away into the non-exported interface ssl/ssl_locl.h, so this
4853 change to the definition of the SSL_CIPHER structure shouldn't
4854 affect applications.) This give us more bits for each of these
4855 categories, so there is no longer a need to coagulate AES128 and
4856 AES256 into a single algorithm bit, and to coagulate Camellia128
4857 and Camellia256 into a single algorithm bit, which has led to all
4860 Thus, among other things, the kludge introduced in 0.9.7m and
4861 0.9.8e for masking out AES256 independently of AES128 or masking
4862 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4864 With the change, we also introduce new ciphersuite aliases that
4865 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4869 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4870 Use the leftmost N bytes of the signature input if the input is
4871 larger than the prime q (with N being the size in bytes of q).
4874 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4875 it yet and it is largely untested.
4878 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4881 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4882 some compilers (gcc 4.2 and later) reject their use. Safestack is
4883 reimplemented. Update ASN1 to avoid use of legacy functions.
4886 *) Win32/64 targets are linked with Winsock2.
4889 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4890 to external functions. This can be used to increase CRL handling
4891 efficiency especially when CRLs are very large by (for example) storing
4892 the CRL revoked certificates in a database.
4895 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4896 new CRLs added to a directory can be used. New command line option
4897 -verify_return_error to s_client and s_server. This causes real errors
4898 to be returned by the verify callback instead of carrying on no matter
4899 what. This reflects the way a "real world" verify callback would behave.
4902 *) GOST engine, supporting several GOST algorithms and public key formats.
4903 Kindly donated by Cryptocom.
4906 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4907 partitioned by DP are handled but no indirect CRL or reason partitioning
4908 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4909 selected via a scoring technique which handles IDP and AKID in CRLs.
4912 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4913 will ultimately be used for all verify operations: this will remove the
4914 X509_STORE dependency on certificate verification and allow alternative
4915 lookup methods. X509_STORE based implementations of these two callbacks.
4918 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4919 Modify get_crl() to find a valid (unexpired) CRL if possible.
4922 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4923 this would be called X509_CRL_cmp() but that name is already used by
4924 a function that just compares CRL issuer names. Cache several CRL
4925 extensions in X509_CRL structure and cache CRLDP in X509.
4928 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4929 this maps equivalent X509_NAME structures into a consistent structure.
4930 Name comparison can then be performed rapidly using memcmp().
4933 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4937 *) Allow digests to supply their own micalg string for S/MIME type using
4938 the ctrl EVP_MD_CTRL_MICALG.
4941 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4942 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4943 ctrl. It can then customise the structure before and/or after signing
4947 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4948 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4949 to free up any added signature OIDs.
4952 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4953 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4954 digest and cipher tables. New options added to openssl utility:
4955 list-message-digest-algorithms and list-cipher-algorithms.
4958 *) Change the array representation of binary polynomials: the list
4959 of degrees of non-zero coefficients is now terminated with -1.
4960 Previously it was terminated with 0, which was also part of the
4961 value; thus, the array representation was not applicable to
4962 polynomials where t^0 has coefficient zero. This change makes
4963 the array representation useful in a more general context.
4966 *) Various modifications and fixes to SSL/TLS cipher string
4967 handling. For ECC, the code now distinguishes between fixed ECDH
4968 with RSA certificates on the one hand and with ECDSA certificates
4969 on the other hand, since these are separate ciphersuites. The
4970 unused code for Fortezza ciphersuites has been removed.
4972 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4973 (not "ECDHE"). For consistency with the code for DH
4974 certificates, use of ECDH certificates is now considered ECDH
4975 authentication, not RSA or ECDSA authentication (the latter is
4976 merely the CA's signing algorithm and not actively used in the
4979 The temporary ciphersuite alias "ECCdraft" is no longer
4980 available, and ECC ciphersuites are no longer excluded from "ALL"
4981 and "DEFAULT". The following aliases now exist for RFC 4492
4982 ciphersuites, most of these by analogy with the DH case:
4984 kECDHr - ECDH cert, signed with RSA
4985 kECDHe - ECDH cert, signed with ECDSA
4986 kECDH - ECDH cert (signed with either RSA or ECDSA)
4987 kEECDH - ephemeral ECDH
4988 ECDH - ECDH cert or ephemeral ECDH
4994 AECDH - anonymous ECDH
4995 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4999 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
5000 Use correct micalg parameters depending on digest(s) in signed message.
5003 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
5004 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
5007 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
5008 an engine to register a method. Add ENGINE lookups for methods and
5009 functional reference processing.
5012 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
5013 EVP_{Sign,Verify}* which allow an application to customise the signature
5017 *) New -resign option to smime utility. This adds one or more signers
5018 to an existing PKCS#7 signedData structure. Also -md option to use an
5019 alternative message digest algorithm for signing.
5022 *) Tidy up PKCS#7 routines and add new functions to make it easier to
5023 create PKCS7 structures containing multiple signers. Update smime
5024 application to support multiple signers.
5027 *) New -macalg option to pkcs12 utility to allow setting of an alternative
5031 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
5032 Reorganize PBE internals to lookup from a static table using NIDs,
5033 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
5034 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
5035 PRF which will be automatically used with PBES2.
5038 *) Replace the algorithm specific calls to generate keys in "req" with the
5042 *) Update PKCS#7 enveloped data routines to use new API. This is now
5043 supported by any public key method supporting the encrypt operation. A
5044 ctrl is added to allow the public key algorithm to examine or modify
5045 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
5049 *) Add a ctrl to asn1 method to allow a public key algorithm to express
5050 a default digest type to use. In most cases this will be SHA1 but some
5051 algorithms (such as GOST) need to specify an alternative digest. The
5052 return value indicates how strong the preference is 1 means optional and
5053 2 is mandatory (that is it is the only supported type). Modify
5054 ASN1_item_sign() to accept a NULL digest argument to indicate it should
5055 use the default md. Update openssl utilities to use the default digest
5056 type for signing if it is not explicitly indicated.
5059 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
5060 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
5061 signing method from the key type. This effectively removes the link
5062 between digests and public key types.
5065 *) Add an OID cross reference table and utility functions. Its purpose is to
5066 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
5067 rsaEncryption. This will allow some of the algorithm specific hackery
5068 needed to use the correct OID to be removed.
5071 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
5072 structures for PKCS7_sign(). They are now set up by the relevant public
5076 *) Add provisional EC pkey method with support for ECDSA and ECDH.
5079 *) Add support for key derivation (agreement) in the API, DH method and
5083 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
5084 public and private key formats. As a side effect these add additional
5085 command line functionality not previously available: DSA signatures can be
5086 generated and verified using pkeyutl and DH key support and generation in
5091 [Oliver Tappe <zooey@hirschkaefer.de>]
5093 *) New make target "install_html_docs" installs HTML renditions of the
5095 [Oliver Tappe <zooey@hirschkaefer.de>]
5097 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
5098 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
5099 support key and parameter generation and add initial key generation
5100 functionality for RSA.
5103 *) Add functions for main EVP_PKEY_method operations. The undocumented
5104 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
5105 EVP_PKEY_{encrypt,decrypt}_old.
5108 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
5109 key API, doesn't do much yet.
5112 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
5113 public key algorithms. New option to openssl utility:
5114 "list-public-key-algorithms" to print out info.
5117 *) Implement the Supported Elliptic Curves Extension for
5118 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5121 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
5122 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
5125 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
5126 utilities such as rsa, dsa, dsaparam etc except they process any key
5130 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
5131 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
5132 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
5136 *) Initial support for pluggable public key ASN1.
5137 De-spaghettify the public key ASN1 handling. Move public and private
5138 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
5139 algorithm specific handling to a single module within the relevant
5140 algorithm directory. Add functions to allow (near) opaque processing
5141 of public and private key structures.
5144 *) Implement the Supported Point Formats Extension for
5145 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5148 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
5149 for the psk identity [hint] and the psk callback functions to the
5150 SSL_SESSION, SSL and SSL_CTX structure.
5153 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
5157 SSL_CTX_use_psk_identity_hint
5158 SSL_get_psk_identity_hint
5159 SSL_get_psk_identity
5160 SSL_use_psk_identity_hint
5162 [Mika Kousa and Pasi Eronen of Nokia Corporation]
5164 *) Add RFC 3161 compliant time stamp request creation, response generation
5165 and response verification functionality.
5166 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
5168 *) Add initial support for TLS extensions, specifically for the server_name
5169 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5170 have new members for a host name. The SSL data structure has an
5171 additional member SSL_CTX *initial_ctx so that new sessions can be
5172 stored in that context to allow for session resumption, even after the
5173 SSL has been switched to a new SSL_CTX in reaction to a client's
5174 server_name extension.
5176 New functions (subject to change):
5178 SSL_get_servername()
5179 SSL_get_servername_type()
5182 New CTRL codes and macros (subject to change):
5184 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5185 - SSL_CTX_set_tlsext_servername_callback()
5186 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5187 - SSL_CTX_set_tlsext_servername_arg()
5188 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5190 openssl s_client has a new '-servername ...' option.
5192 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5193 '-key2 ...', '-servername_fatal' (subject to change). This allows
5194 testing the HostName extension for a specific single host name ('-cert'
5195 and '-key' remain fallbacks for handshakes without HostName
5196 negotiation). If the unrecognized_name alert has to be sent, this by
5197 default is a warning; it becomes fatal with the '-servername_fatal'
5200 [Peter Sylvester, Remy Allais, Christophe Renou]
5202 *) Whirlpool hash implementation is added.
5205 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
5206 bn(64,32). Because of instruction set limitations it doesn't have
5207 any negative impact on performance. This was done mostly in order
5208 to make it possible to share assembler modules, such as bn_mul_mont
5209 implementations, between 32- and 64-bit builds without hassle.
5212 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
5213 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
5217 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
5218 dedicated Montgomery multiplication procedure, is introduced.
5219 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
5220 "64-bit" performance on certain 32-bit targets.
5223 *) New option SSL_OP_NO_COMP to disable use of compression selectively
5224 in SSL structures. New SSL ctrl to set maximum send fragment size.
5225 Save memory by setting the I/O buffer sizes dynamically instead of
5226 using the maximum available value.
5229 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
5230 in addition to the text details.
5233 *) Very, very preliminary EXPERIMENTAL support for printing of general
5234 ASN1 structures. This currently produces rather ugly output and doesn't
5235 handle several customised structures at all.
5238 *) Integrated support for PVK file format and some related formats such
5239 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
5240 these in the 'rsa' and 'dsa' utilities.
5243 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
5246 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
5247 place for the (very old) "NETSCAPE" format certificates which are now
5248 handled using new ASN1 code equivalents.
5251 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
5252 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
5253 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
5256 *) Modify CRL distribution points extension code to print out previously
5257 unsupported fields. Enhance extension setting code to allow setting of
5261 *) Add print and set support for Issuing Distribution Point CRL extension.
5264 *) Change 'Configure' script to enable Camellia by default.
5267 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
5269 *) When rejecting SSL/TLS records due to an incorrect version number, never
5270 update s->server with a new major version number. As of
5271 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
5272 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
5273 the previous behavior could result in a read attempt at NULL when
5274 receiving specific incorrect SSL/TLS records once record payload
5275 protection is active. (CVE-2010-0740)
5276 [Bodo Moeller, Adam Langley <agl@chromium.org>]
5278 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
5279 could be crashed if the relevant tables were not present (e.g. chrooted).
5280 [Tomas Hoger <thoger@redhat.com>]
5282 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
5284 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
5285 [Martin Olsson, Neel Mehta]
5287 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
5288 accommodate for stack sorting, always a write lock!).
5291 *) On some versions of WIN32 Heap32Next is very slow. This can cause
5292 excessive delays in the RAND_poll(): over a minute. As a workaround
5293 include a time check in the inner Heap32Next loop too.
5296 *) The code that handled flushing of data in SSL/TLS originally used the
5297 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
5298 the problem outlined in PR#1949. The fix suggested there however can
5299 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
5300 of Apache). So instead simplify the code to flush unconditionally.
5301 This should be fine since flushing with no data to flush is a no op.
5304 *) Handle TLS versions 2.0 and later properly and correctly use the
5305 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
5306 off ancient servers have a habit of sticking around for a while...
5309 *) Modify compression code so it frees up structures without using the
5310 ex_data callbacks. This works around a problem where some applications
5311 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
5312 restarting) then use compression (e.g. SSL with compression) later.
5313 This results in significant per-connection memory leaks and
5314 has caused some security issues including CVE-2008-1678 and
5318 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
5319 change when encrypting or decrypting.
5322 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
5323 connect and renegotiate with servers which do not support RI.
5324 Until RI is more widely deployed this option is enabled by default.
5327 *) Add "missing" ssl ctrls to clear options and mode.
5330 *) If client attempts to renegotiate and doesn't support RI respond with
5331 a no_renegotiation alert as required by RFC5746. Some renegotiating
5332 TLS clients will continue a connection gracefully when they receive
5333 the alert. Unfortunately OpenSSL mishandled this alert and would hang
5334 waiting for a server hello which it will never receive. Now we treat a
5335 received no_renegotiation alert as a fatal error. This is because
5336 applications requesting a renegotiation might well expect it to succeed
5337 and would have no code in place to handle the server denying it so the
5338 only safe thing to do is to terminate the connection.
5341 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
5342 peer supports secure renegotiation and 0 otherwise. Print out peer
5343 renegotiation support in s_client/s_server.
5346 *) Replace the highly broken and deprecated SPKAC certification method with
5347 the updated NID creation version. This should correctly handle UTF8.
5350 *) Implement RFC5746. Re-enable renegotiation but require the extension
5351 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
5352 turns out to be a bad idea. It has been replaced by
5353 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5354 SSL_CTX_set_options(). This is really not recommended unless you
5355 know what you are doing.
5356 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5358 *) Fixes to stateless session resumption handling. Use initial_ctx when
5359 issuing and attempting to decrypt tickets in case it has changed during
5360 servername handling. Use a non-zero length session ID when attempting
5361 stateless session resumption: this makes it possible to determine if
5362 a resumption has occurred immediately after receiving server hello
5363 (several places in OpenSSL subtly assume this) instead of later in
5367 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5368 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5369 fixes for a few places where the return code is not checked
5371 [Julia Lawall <julia@diku.dk>]
5373 *) Add --strict-warnings option to Configure script to include devteam
5374 warnings in other configurations.
5377 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5378 makes it possible to install openssl libraries in locations which
5379 have names other than "lib", for example "/usr/lib64" which some
5381 [Steve Henson, based on patch from Jeremy Utley]
5383 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5384 X690 8.9.12 and can produce some misleading textual output of OIDs.
5385 [Steve Henson, reported by Dan Kaminsky]
5387 *) Delete MD2 from algorithm tables. This follows the recommendation in
5388 several standards that it is not used in new applications due to
5389 several cryptographic weaknesses. For binary compatibility reasons
5390 the MD2 API is still compiled in by default.
5393 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5397 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5398 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5400 [Guenter <lists@gknw.net>]
5402 *) Fix the server certificate chain building code to use X509_verify_cert(),
5403 it used to have an ad-hoc builder which was unable to cope with anything
5404 other than a simple chain.
5405 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5407 *) Don't check self signed certificate signatures in X509_verify_cert()
5408 by default (a flag can override this): it just wastes time without
5409 adding any security. As a useful side effect self signed root CAs
5410 with non-FIPS digests are now usable in FIPS mode.
5413 *) In dtls1_process_out_of_seq_message() the check if the current message
5414 is already buffered was missing. For every new message was memory
5415 allocated, allowing an attacker to perform an denial of service attack
5416 with sending out of seq handshake messages until there is no memory
5417 left. Additionally every future message was buffered, even if the
5418 sequence number made no sense and would be part of another handshake.
5419 So only messages with sequence numbers less than 10 in advance will be
5420 buffered. (CVE-2009-1378)
5421 [Robin Seggelmann, discovered by Daniel Mentz]
5423 *) Records are buffered if they arrive with a future epoch to be
5424 processed after finishing the corresponding handshake. There is
5425 currently no limitation to this buffer allowing an attacker to perform
5426 a DOS attack with sending records with future epochs until there is no
5427 memory left. This patch adds the pqueue_size() function to determine
5428 the size of a buffer and limits the record buffer to 100 entries.
5430 [Robin Seggelmann, discovered by Daniel Mentz]
5432 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5433 parent structure is freed. (CVE-2009-1379)
5436 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5437 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5440 [Ilya O. <vrghost@gmail.com>]
5442 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5444 *) Disable renegotiation completely - this fixes a severe security
5445 problem (CVE-2009-3555) at the cost of breaking all
5446 renegotiation. Renegotiation can be re-enabled by setting
5447 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5448 run-time. This is really not recommended unless you know what
5452 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5454 *) Don't set val to NULL when freeing up structures, it is freed up by
5455 underlying code. If sizeof(void *) > sizeof(long) this can result in
5456 zeroing past the valid field. (CVE-2009-0789)
5457 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5459 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5460 checked correctly. This would allow some invalid signed attributes to
5461 appear to verify correctly. (CVE-2009-0591)
5462 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5464 *) Reject UniversalString and BMPString types with invalid lengths. This
5465 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5466 a legal length. (CVE-2009-0590)
5469 *) Set S/MIME signing as the default purpose rather than setting it
5470 unconditionally. This allows applications to override it at the store
5474 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5475 to handle some structures.
5478 *) Improve efficiency of mem_gets: don't search whole buffer each time
5480 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5482 *) New -hex option for openssl rand.
5485 *) Print out UTF8String and NumericString when parsing ASN1.
5488 *) Support NumericString type for name components.
5491 *) Allow CC in the environment to override the automatically chosen
5492 compiler. Note that nothing is done to ensure flags work with the
5496 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5498 *) Properly check EVP_VerifyFinal() and similar return values
5500 [Ben Laurie, Bodo Moeller, Google Security Team]
5502 *) Enable TLS extensions by default.
5505 *) Allow the CHIL engine to be loaded, whether the application is
5506 multithreaded or not. (This does not release the developer from the
5507 obligation to set up the dynamic locking callbacks.)
5508 [Sander Temme <sander@temme.net>]
5510 *) Use correct exit code if there is an error in dgst command.
5511 [Steve Henson; problem pointed out by Roland Dirlewanger]
5513 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5514 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5517 *) Add experimental JPAKE support, including demo authentication in
5518 s_client and s_server.
5521 *) Set the comparison function in v3_addr_canonize().
5522 [Rob Austein <sra@hactrn.net>]
5524 *) Add support for XMPP STARTTLS in s_client.
5525 [Philip Paeps <philip@freebsd.org>]
5527 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5528 to ensure that even with this option, only ciphersuites in the
5529 server's preference list will be accepted. (Note that the option
5530 applies only when resuming a session, so the earlier behavior was
5531 just about the algorithm choice for symmetric cryptography.)
5534 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5536 *) Fix NULL pointer dereference if a DTLS server received
5537 ChangeCipherSpec as first record (CVE-2009-1386).
5540 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5541 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5544 *) The fix in 0.9.8c that supposedly got rid of unsafe
5545 double-checked locking was incomplete for RSA blinding,
5546 addressing just one layer of what turns out to have been
5547 doubly unsafe triple-checked locking.
5549 So now fix this for real by retiring the MONT_HELPER macro
5550 in crypto/rsa/rsa_eay.c.
5552 [Bodo Moeller; problem pointed out by Marius Schilder]
5554 *) Various precautionary measures:
5556 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5558 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5559 (NB: This would require knowledge of the secret session ticket key
5560 to exploit, in which case you'd be SOL either way.)
5562 - Change bn_nist.c so that it will properly handle input BIGNUMs
5563 outside the expected range.
5565 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5568 [Neel Mehta, Bodo Moeller]
5570 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5571 the load fails. Useful for distros.
5572 [Ben Laurie and the FreeBSD team]
5574 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5577 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5580 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5582 This work was sponsored by Logica.
5585 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5586 keystores. Support for SSL/TLS client authentication too.
5587 Not compiled unless enable-capieng specified to Configure.
5589 This work was sponsored by Logica.
5592 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5593 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5594 attribute creation routines such as certificate requests and PKCS#12
5598 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5600 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5601 handshake which could lead to a client crash as found using the
5602 Codenomicon TLS test suite (CVE-2008-1672)
5603 [Steve Henson, Mark Cox]
5605 *) Fix double free in TLS server name extensions which could lead to
5606 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5609 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5611 Clear the error queue to ensure that error entries left from
5612 older function calls do not interfere with the correct operation.
5613 [Lutz Jaenicke, Erik de Castro Lopo]
5615 *) Remove root CA certificates of commercial CAs:
5617 The OpenSSL project does not recommend any specific CA and does not
5618 have any policy with respect to including or excluding any CA.
5619 Therefore it does not make any sense to ship an arbitrary selection
5620 of root CA certificates with the OpenSSL software.
5623 *) RSA OAEP patches to fix two separate invalid memory reads.
5624 The first one involves inputs when 'lzero' is greater than
5625 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5626 before the beginning of from). The second one involves inputs where
5627 the 'db' section contains nothing but zeroes (there is a one-byte
5628 invalid read after the end of 'db').
5629 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5631 *) Partial backport from 0.9.9-dev:
5633 Introduce bn_mul_mont (dedicated Montgomery multiplication
5634 procedure) as a candidate for BIGNUM assembler implementation.
5635 While 0.9.9-dev uses assembler for various architectures, only
5636 x86_64 is available by default here in the 0.9.8 branch, and
5637 32-bit x86 is available through a compile-time setting.
5639 To try the 32-bit x86 assembler implementation, use Configure
5640 option "enable-montasm" (which exists only for this backport).
5642 As "enable-montasm" for 32-bit x86 disclaims code stability
5643 anyway, in this constellation we activate additional code
5644 backported from 0.9.9-dev for further performance improvements,
5645 namely BN_from_montgomery_word. (To enable this otherwise,
5646 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5648 [Andy Polyakov (backport partially by Bodo Moeller)]
5650 *) Add TLS session ticket callback. This allows an application to set
5651 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5652 values. This is useful for key rollover for example where several key
5653 sets may exist with different names.
5656 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5657 This was broken until now in 0.9.8 releases, such that the only way
5658 a registered ENGINE could be used (assuming it initialises
5659 successfully on the host) was to explicitly set it as the default
5660 for the relevant algorithms. This is in contradiction with 0.9.7
5661 behaviour and the documentation. With this fix, when an ENGINE is
5662 registered into a given algorithm's table of implementations, the
5663 'uptodate' flag is reset so that auto-discovery will be used next
5664 time a new context for that algorithm attempts to select an
5666 [Ian Lister (tweaked by Geoff Thorpe)]
5668 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5669 implementation in the following ways:
5671 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5674 Lack of BER streaming support means one pass streaming processing is
5675 only supported if data is detached: setting the streaming flag is
5676 ignored for embedded content.
5678 CMS support is disabled by default and must be explicitly enabled
5679 with the enable-cms configuration option.
5682 *) Update the GMP engine glue to do direct copies between BIGNUM and
5683 mpz_t when openssl and GMP use the same limb size. Otherwise the
5684 existing "conversion via a text string export" trick is still used.
5685 [Paul Sheer <paulsheer@gmail.com>]
5687 *) Zlib compression BIO. This is a filter BIO which compressed and
5688 uncompresses any data passed through it.
5691 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5692 RFC3394 compatible AES key wrapping.
5695 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5696 sets string data without copying. X509_ALGOR_set0() and
5697 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5698 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5699 from an X509_ATTRIBUTE structure optionally checking it occurs only
5700 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5704 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5705 to get the expected BN_FLG_CONSTTIME behavior.
5706 [Bodo Moeller (Google)]
5710 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5711 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5712 - added some more tests to do_tests.pl
5713 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5714 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5715 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5716 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5717 - various changes to netware.pl to enable gcc-cross builds on Win32
5719 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5720 - various changes to fix missing prototype warnings
5721 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5722 - added AES, WHIRLPOOL and CPUID assembler code to build files
5723 - added missing AES assembler make rules to mk1mf.pl
5724 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5725 [Guenter Knauf <eflash@gmx.net>]
5727 *) Implement certificate status request TLS extension defined in RFC3546.
5728 A client can set the appropriate parameters and receive the encoded
5729 OCSP response via a callback. A server can query the supplied parameters
5730 and set the encoded OCSP response in the callback. Add simplified examples
5731 to s_client and s_server.
5734 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5736 *) Fix various bugs:
5737 + Binary incompatibility of ssl_ctx_st structure
5738 + DTLS interoperation with non-compliant servers
5739 + Don't call get_session_cb() without proposed session
5740 + Fix ia64 assembler code
5741 [Andy Polyakov, Steve Henson]
5743 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5745 *) DTLS Handshake overhaul. There were longstanding issues with
5746 OpenSSL DTLS implementation, which were making it impossible for
5747 RFC 4347 compliant client to communicate with OpenSSL server.
5748 Unfortunately just fixing these incompatibilities would "cut off"
5749 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5750 server keeps tolerating non RFC compliant syntax. The opposite is
5751 not true, 0.9.8f client can not communicate with earlier server.
5752 This update even addresses CVE-2007-4995.
5755 *) Changes to avoid need for function casts in OpenSSL: some compilers
5756 (gcc 4.2 and later) reject their use.
5757 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5760 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5761 RFC4507bis. The encrypted ticket format is an encrypted encoded
5762 SSL_SESSION structure, that way new session features are automatically
5765 If a client application caches session in an SSL_SESSION structure
5766 support is transparent because tickets are now stored in the encoded
5769 The SSL_CTX structure automatically generates keys for ticket
5770 protection in servers so again support should be possible
5771 with no application modification.
5773 If a client or server wishes to disable RFC4507 support then the option
5774 SSL_OP_NO_TICKET can be set.
5776 Add a TLS extension debugging callback to allow the contents of any client
5777 or server extensions to be examined.
5779 This work was sponsored by Google.
5782 *) Add initial support for TLS extensions, specifically for the server_name
5783 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5784 have new members for a host name. The SSL data structure has an
5785 additional member SSL_CTX *initial_ctx so that new sessions can be
5786 stored in that context to allow for session resumption, even after the
5787 SSL has been switched to a new SSL_CTX in reaction to a client's
5788 server_name extension.
5790 New functions (subject to change):
5792 SSL_get_servername()
5793 SSL_get_servername_type()
5796 New CTRL codes and macros (subject to change):
5798 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5799 - SSL_CTX_set_tlsext_servername_callback()
5800 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5801 - SSL_CTX_set_tlsext_servername_arg()
5802 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5804 openssl s_client has a new '-servername ...' option.
5806 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5807 '-key2 ...', '-servername_fatal' (subject to change). This allows
5808 testing the HostName extension for a specific single host name ('-cert'
5809 and '-key' remain fallbacks for handshakes without HostName
5810 negotiation). If the unrecognized_name alert has to be sent, this by
5811 default is a warning; it becomes fatal with the '-servername_fatal'
5814 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5816 *) Add AES and SSE2 assembly language support to VC++ build.
5819 *) Mitigate attack on final subtraction in Montgomery reduction.
5822 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5823 (which previously caused an internal error).
5826 *) Squeeze another 10% out of IGE mode when in != out.
5829 *) AES IGE mode speedup.
5830 [Dean Gaudet (Google)]
5832 *) Add the Korean symmetric 128-bit cipher SEED (see
5833 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5834 add SEED ciphersuites from RFC 4162:
5836 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5837 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5838 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5839 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5841 To minimize changes between patchlevels in the OpenSSL 0.9.8
5842 series, SEED remains excluded from compilation unless OpenSSL
5843 is configured with 'enable-seed'.
5844 [KISA, Bodo Moeller]
5846 *) Mitigate branch prediction attacks, which can be practical if a
5847 single processor is shared, allowing a spy process to extract
5848 information. For detailed background information, see
5849 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5850 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5851 and Necessary Software Countermeasures"). The core of the change
5852 are new versions BN_div_no_branch() and
5853 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5854 respectively, which are slower, but avoid the security-relevant
5855 conditional branches. These are automatically called by BN_div()
5856 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5857 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5858 remove a conditional branch.
5860 BN_FLG_CONSTTIME is the new name for the previous
5861 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5862 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5863 in the exponent causes BN_mod_exp_mont() to use the alternative
5864 implementation in BN_mod_exp_mont_consttime().) The old name
5865 remains as a deprecated alias.
5867 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5868 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5869 constant-time implementations for more than just exponentiation.
5870 Here too the old name is kept as a deprecated alias.
5872 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5873 the BN_BLINDING structure gets an independent copy of the
5874 modulus. This means that the previous "BIGNUM *m" argument to
5875 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5876 essentially becomes "const BIGNUM *m", although we can't actually
5877 change this in the header file before 0.9.9. It allows
5878 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5879 enable BN_FLG_CONSTTIME.
5881 [Matthew D Wood (Intel Corp)]
5883 *) In the SSL/TLS server implementation, be strict about session ID
5884 context matching (which matters if an application uses a single
5885 external cache for different purposes). Previously,
5886 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5887 set. This did ensure strict client verification, but meant that,
5888 with applications using a single external cache for quite
5889 different requirements, clients could circumvent ciphersuite
5890 restrictions for a given session ID context by starting a session
5891 in a different context.
5894 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5895 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5896 authentication-only ciphersuites.
5899 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5900 not complete and could lead to a possible single byte overflow
5901 (CVE-2007-5135) [Ben Laurie]
5903 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5905 *) Since AES128 and AES256 (and similarly Camellia128 and
5906 Camellia256) share a single mask bit in the logic of
5907 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5908 kludge to work properly if AES128 is available and AES256 isn't
5909 (or if Camellia128 is available and Camellia256 isn't).
5912 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5913 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5914 When a point or a seed is encoded in a BIT STRING, we need to
5915 prevent the removal of trailing zero bits to get the proper DER
5916 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5917 of a NamedBitList, for which trailing 0 bits need to be removed.)
5920 *) Have SSL/TLS server implementation tolerate "mismatched" record
5921 protocol version while receiving ClientHello even if the
5922 ClientHello is fragmented. (The server can't insist on the
5923 particular protocol version it has chosen before the ServerHello
5924 message has informed the client about his choice.)
5927 *) Add RFC 3779 support.
5928 [Rob Austein for ARIN, Ben Laurie]
5930 *) Load error codes if they are not already present instead of using a
5931 static variable. This allows them to be cleanly unloaded and reloaded.
5932 Improve header file function name parsing.
5935 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5936 or CAPABILITY handshake as required by RFCs.
5939 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5941 *) Introduce limits to prevent malicious keys being able to
5942 cause a denial of service. (CVE-2006-2940)
5943 [Steve Henson, Bodo Moeller]
5945 *) Fix ASN.1 parsing of certain invalid structures that can result
5946 in a denial of service. (CVE-2006-2937) [Steve Henson]
5948 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5949 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5951 *) Fix SSL client code which could crash if connecting to a
5952 malicious SSLv2 server. (CVE-2006-4343)
5953 [Tavis Ormandy and Will Drewry, Google Security Team]
5955 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5956 match only those. Before that, "AES256-SHA" would be interpreted
5957 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5958 the same strength classification in 0.9.7h) as we currently only
5959 have a single AES bit in the ciphersuite description bitmap.
5960 That change, however, also applied to ciphersuite strings such as
5961 "RC4-MD5" that intentionally matched multiple ciphersuites --
5962 namely, SSL 2.0 ciphersuites in addition to the more common ones
5963 from SSL 3.0/TLS 1.0.
5965 So we change the selection algorithm again: Naming an explicit
5966 ciphersuite selects this one ciphersuite, and any other similar
5967 ciphersuite (same bitmap) from *other* protocol versions.
5968 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5969 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5971 Since SSL 2.0 does not have any ciphersuites for which the
5972 128/256 bit distinction would be relevant, this works for now.
5973 The proper fix will be to use different bits for AES128 and
5974 AES256, which would have avoided the problems from the beginning;
5975 however, bits are scarce, so we can only do this in a new release
5976 (not just a patchlevel) when we can change the SSL_CIPHER
5977 definition to split the single 'unsigned long mask' bitmap into
5978 multiple values to extend the available space.
5982 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5984 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5985 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5987 *) Add AES IGE and biIGE modes.
5990 *) Change the Unix randomness entropy gathering to use poll() when
5991 possible instead of select(), since the latter has some
5992 undesirable limitations.
5993 [Darryl Miles via Richard Levitte and Bodo Moeller]
5995 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5996 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5997 cannot be implicitly activated as part of, e.g., the "AES" alias.
5998 However, please upgrade to OpenSSL 0.9.9[-dev] for
5999 non-experimental use of the ECC ciphersuites to get TLS extension
6000 support, which is required for curve and point format negotiation
6001 to avoid potential handshake problems.
6004 *) Disable rogue ciphersuites:
6006 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6007 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6008 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6010 The latter two were purportedly from
6011 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6014 Also deactivate the remaining ciphersuites from
6015 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6016 unofficial, and the ID has long expired.
6019 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6020 dual-core machines) and other potential thread-safety issues.
6023 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
6024 versions), which is now available for royalty-free use
6025 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
6026 Also, add Camellia TLS ciphersuites from RFC 4132.
6028 To minimize changes between patchlevels in the OpenSSL 0.9.8
6029 series, Camellia remains excluded from compilation unless OpenSSL
6030 is configured with 'enable-camellia'.
6033 *) Disable the padding bug check when compression is in use. The padding
6034 bug check assumes the first packet is of even length, this is not
6035 necessarily true if compression is enabled and can result in false
6036 positives causing handshake failure. The actual bug test is ancient
6037 code so it is hoped that implementations will either have fixed it by
6038 now or any which still have the bug do not support compression.
6041 Changes between 0.9.8a and 0.9.8b [04 May 2006]
6043 *) When applying a cipher rule check to see if string match is an explicit
6044 cipher suite and only match that one cipher suite if it is.
6047 *) Link in manifests for VC++ if needed.
6048 [Austin Ziegler <halostatue@gmail.com>]
6050 *) Update support for ECC-based TLS ciphersuites according to
6051 draft-ietf-tls-ecc-12.txt with proposed changes (but without
6052 TLS extensions, which are supported starting with the 0.9.9
6053 branch, not in the OpenSSL 0.9.8 branch).
6056 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
6057 opaque EVP_CIPHER_CTX handling.
6060 *) Fixes and enhancements to zlib compression code. We now only use
6061 "zlib1.dll" and use the default __cdecl calling convention on Win32
6062 to conform with the standards mentioned here:
6063 http://www.zlib.net/DLL_FAQ.txt
6064 Static zlib linking now works on Windows and the new --with-zlib-include
6065 --with-zlib-lib options to Configure can be used to supply the location
6066 of the headers and library. Gracefully handle case where zlib library
6070 *) Several fixes and enhancements to the OID generation code. The old code
6071 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
6072 handle numbers larger than ULONG_MAX, truncated printing and had a
6073 non standard OBJ_obj2txt() behaviour.
6076 *) Add support for building of engines under engine/ as shared libraries
6077 under VC++ build system.
6080 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
6081 Hopefully, we will not see any false combination of paths any more.
6084 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
6086 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6087 (part of SSL_OP_ALL). This option used to disable the
6088 countermeasure against man-in-the-middle protocol-version
6089 rollback in the SSL 2.0 server implementation, which is a bad
6090 idea. (CVE-2005-2969)
6092 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6093 for Information Security, National Institute of Advanced Industrial
6094 Science and Technology [AIST], Japan)]
6096 *) Add two function to clear and return the verify parameter flags.
6099 *) Keep cipherlists sorted in the source instead of sorting them at
6100 runtime, thus removing the need for a lock.
6103 *) Avoid some small subgroup attacks in Diffie-Hellman.
6104 [Nick Mathewson and Ben Laurie]
6106 *) Add functions for well-known primes.
6109 *) Extended Windows CE support.
6110 [Satoshi Nakamura and Andy Polyakov]
6112 *) Initialize SSL_METHOD structures at compile time instead of during
6113 runtime, thus removing the need for a lock.
6116 *) Make PKCS7_decrypt() work even if no certificate is supplied by
6117 attempting to decrypt each encrypted key in turn. Add support to
6121 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
6123 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
6126 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
6129 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
6130 key into the same file any more.
6133 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
6136 *) Add -utf8 command line and config file option to 'ca'.
6137 [Stefan <stf@udoma.org]
6139 *) Removed the macro des_crypt(), as it seems to conflict with some
6140 libraries. Use DES_crypt().
6143 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
6144 involves renaming the source and generated shared-libs for
6145 both. The engines will accept the corrected or legacy ids
6146 ('ncipher' and '4758_cca' respectively) when binding. NB,
6147 this only applies when building 'shared'.
6148 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
6150 *) Add attribute functions to EVP_PKEY structure. Modify
6151 PKCS12_create() to recognize a CSP name attribute and
6152 use it. Make -CSP option work again in pkcs12 utility.
6155 *) Add new functionality to the bn blinding code:
6156 - automatic re-creation of the BN_BLINDING parameters after
6157 a fixed number of uses (currently 32)
6158 - add new function for parameter creation
6159 - introduce flags to control the update behaviour of the
6160 BN_BLINDING parameters
6161 - hide BN_BLINDING structure
6162 Add a second BN_BLINDING slot to the RSA structure to improve
6163 performance when a single RSA object is shared among several
6167 *) Add support for DTLS.
6168 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
6170 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
6171 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
6174 *) Remove buggy and incomplete DH cert support from
6175 ssl/ssl_rsa.c and ssl/s3_both.c
6178 *) Use SHA-1 instead of MD5 as the default digest algorithm for
6179 the apps/openssl applications.
6182 *) Compile clean with "-Wall -Wmissing-prototypes
6183 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
6184 DEBUG_SAFESTACK must also be set.
6187 *) Change ./Configure so that certain algorithms can be disabled by default.
6188 The new counterpiece to "no-xxx" is "enable-xxx".
6190 The patented RC5 and MDC2 algorithms will now be disabled unless
6191 "enable-rc5" and "enable-mdc2", respectively, are specified.
6193 (IDEA remains enabled despite being patented. This is because IDEA
6194 is frequently required for interoperability, and there is no license
6195 fee for non-commercial use. As before, "no-idea" can be used to
6196 avoid this algorithm.)
6200 *) Add processing of proxy certificates (see RFC 3820). This work was
6201 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
6202 EGEE (Enabling Grids for E-science in Europe).
6205 *) RC4 performance overhaul on modern architectures/implementations, such
6206 as Intel P4, IA-64 and AMD64.
6209 *) New utility extract-section.pl. This can be used specify an alternative
6210 section number in a pod file instead of having to treat each file as
6211 a separate case in Makefile. This can be done by adding two lines to the
6214 =for comment openssl_section:XXX
6216 The blank line is mandatory.
6220 *) New arguments -certform, -keyform and -pass for s_client and s_server
6221 to allow alternative format key and certificate files and passphrase
6225 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
6226 update associated structures and add various utility functions.
6228 Add new policy related verify parameters, include policy checking in
6229 standard verify code. Enhance 'smime' application with extra parameters
6230 to support policy checking and print out.
6233 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
6234 Nehemiah processors. These extensions support AES encryption in hardware
6235 as well as RNG (though RNG support is currently disabled).
6236 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
6238 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
6241 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
6242 [Andy Polyakov and a number of other people]
6244 *) Improved PowerPC platform support. Most notably BIGNUM assembler
6245 implementation contributed by IBM.
6246 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
6248 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
6249 exponent rather than 'unsigned long'. There is a corresponding change to
6250 the new 'rsa_keygen' element of the RSA_METHOD structure.
6251 [Jelte Jansen, Geoff Thorpe]
6253 *) Functionality for creating the initial serial number file is now
6254 moved from CA.pl to the 'ca' utility with a new option -create_serial.
6256 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
6257 number file to 1, which is bound to cause problems. To avoid
6258 the problems while respecting compatibility between different 0.9.7
6259 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
6260 CA.pl for serial number initialization. With the new release 0.9.8,
6261 we can fix the problem directly in the 'ca' utility.)
6264 *) Reduced header interdependencies by declaring more opaque objects in
6265 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
6266 give fewer recursive includes, which could break lazy source code - so
6267 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
6268 developers should define this symbol when building and using openssl to
6269 ensure they track the recommended behaviour, interfaces, [etc], but
6270 backwards-compatible behaviour prevails when this isn't defined.
6273 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
6276 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
6277 This will generate a random key of the appropriate length based on the
6278 cipher context. The EVP_CIPHER can provide its own random key generation
6279 routine to support keys of a specific form. This is used in the des and
6280 3des routines to generate a key of the correct parity. Update S/MIME
6281 code to use new functions and hence generate correct parity DES keys.
6282 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
6283 valid (weak or incorrect parity).
6286 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
6287 as looking them up. This is useful when the verified structure may contain
6288 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
6289 present unless the new PKCS7_NO_CRL flag is asserted.
6292 *) Extend ASN1 oid configuration module. It now additionally accepts the
6295 shortName = some long name, 1.2.3.4
6298 *) Reimplemented the BN_CTX implementation. There is now no more static
6299 limitation on the number of variables it can handle nor the depth of the
6300 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
6301 information can now expand as required, and rather than having a single
6302 static array of bignums, BN_CTX now uses a linked-list of such arrays
6303 allowing it to expand on demand whilst maintaining the usefulness of
6304 BN_CTX's "bundling".
6307 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
6308 to allow all RSA operations to function using a single BN_CTX.
6311 *) Preliminary support for certificate policy evaluation and checking. This
6312 is initially intended to pass the tests outlined in "Conformance Testing
6313 of Relying Party Client Certificate Path Processing Logic" v1.07.
6316 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
6317 remained unused and not that useful. A variety of other little bignum
6318 tweaks and fixes have also been made continuing on from the audit (see
6322 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
6323 associated ASN1, EVP and SSL functions and old ASN1 macros.
6326 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
6327 and this should never fail. So the return value from the use of
6328 BN_set_word() (which can fail due to needless expansion) is now deprecated;
6329 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
6332 *) BN_CTX_get() should return zero-valued bignums, providing the same
6333 initialised value as BN_new().
6334 [Geoff Thorpe, suggested by Ulf Möller]
6336 *) Support for inhibitAnyPolicy certificate extension.
6339 *) An audit of the BIGNUM code is underway, for which debugging code is
6340 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
6341 is considered valid when processing BIGNUMs, and causes execution to
6342 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
6343 further steps are taken to deliberately pollute unused data in BIGNUM
6344 structures to try and expose faulty code further on. For now, openssl will
6345 (in its default mode of operation) continue to tolerate the inconsistent
6346 forms that it has tolerated in the past, but authors and packagers should
6347 consider trying openssl and their own applications when compiled with
6348 these debugging symbols defined. It will help highlight potential bugs in
6349 their own code, and will improve the test coverage for OpenSSL itself. At
6350 some point, these tighter rules will become openssl's default to improve
6351 maintainability, though the assert()s and other overheads will remain only
6352 in debugging configurations. See bn.h for more details.
6353 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6355 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6356 that can only be obtained through BN_CTX_new() (which implicitly
6357 initialises it). The presence of this function only made it possible
6358 to overwrite an existing structure (and cause memory leaks).
6361 *) Because of the callback-based approach for implementing LHASH as a
6362 template type, lh_insert() adds opaque objects to hash-tables and
6363 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6364 to clean up those corresponding objects before destroying the hash table
6365 (and losing the object pointers). So some over-zealous constifications in
6366 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6367 objects as "const" and the lh_doall[_arg] callback wrappers are not
6368 prototyped to have "const" restrictions on the object pointers they are
6369 given (and so aren't required to cast them away any more).
6372 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6373 (speed) prefers to use its own implementation. The two implementations
6374 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6375 its object type properly exposed (MS_TM) instead of casting to/from "char
6376 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6377 aren't necessarily the greatest nomenclatures - but this is what was used
6378 internally to the implementation so I've used that for now.
6381 *) Ensure that deprecated functions do not get compiled when
6382 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6383 the self-tests were still using deprecated key-generation functions so
6384 these have been updated also.
6387 *) Reorganise PKCS#7 code to separate the digest location functionality
6388 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6389 New function PKCS7_set_digest() to set the digest type for PKCS#7
6390 digestedData type. Add additional code to correctly generate the
6391 digestedData type and add support for this type in PKCS7 initialization
6395 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6396 structure of type "other".
6399 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6400 sure the loop does correctly stop and breaking ("division by zero")
6401 modulus operations are not performed. The (pre-generated) prime
6402 table crypto/bn/bn_prime.h was already correct, but it could not be
6403 re-generated on some platforms because of the "division by zero"
6404 situation in the script.
6405 [Ralf S. Engelschall]
6407 *) Update support for ECC-based TLS ciphersuites according to
6408 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6409 SHA-1 now is only used for "small" curves (where the
6410 representation of a field element takes up to 24 bytes); for
6411 larger curves, the field element resulting from ECDH is directly
6412 used as premaster secret.
6413 [Douglas Stebila (Sun Microsystems Laboratories)]
6415 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6416 curve secp160r1 to the tests.
6417 [Douglas Stebila (Sun Microsystems Laboratories)]
6419 *) Add the possibility to load symbols globally with DSO.
6420 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6422 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6423 control of the error stack.
6426 *) Add support for STORE in ENGINE.
6429 *) Add the STORE type. The intention is to provide a common interface
6430 to certificate and key stores, be they simple file-based stores, or
6431 HSM-type store, or LDAP stores, or...
6432 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6435 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6436 pass a list of arguments to any function as well as provide a way
6437 for a function to pass data back to the caller.
6440 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6441 works like BUF_strdup() but can be used to duplicate a portion of
6442 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6446 *) Add the function sk_find_ex() which works like sk_find(), but will
6447 return an index to an element even if an exact match couldn't be
6448 found. The index is guaranteed to point at the element where the
6449 searched-for key would be inserted to preserve sorting order.
6452 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6453 takes an extra flags argument for optional functionality. Currently,
6454 the following flags are defined:
6456 OBJ_BSEARCH_VALUE_ON_NOMATCH
6457 This one gets OBJ_bsearch_ex() to return a pointer to the first
6458 element where the comparing function returns a negative or zero
6461 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6462 This one gets OBJ_bsearch_ex() to return a pointer to the first
6463 element where the comparing function returns zero. This is useful
6464 if there are more than one element where the comparing function
6468 *) Make it possible to create self-signed certificates with 'openssl ca'
6469 in such a way that the self-signed certificate becomes part of the
6470 CA database and uses the same mechanisms for serial number generation
6471 as all other certificate signing. The new flag '-selfsign' enables
6472 this functionality. Adapt CA.sh and CA.pl.in.
6475 *) Add functionality to check the public key of a certificate request
6476 against a given private. This is useful to check that a certificate
6477 request can be signed by that key (self-signing).
6480 *) Make it possible to have multiple active certificates with the same
6481 subject in the CA index file. This is done only if the keyword
6482 'unique_subject' is set to 'no' in the main CA section (default
6483 if 'CA_default') of the configuration file. The value is saved
6484 with the database itself in a separate index attribute file,
6485 named like the index file with '.attr' appended to the name.
6488 *) Generate multi-valued AVAs using '+' notation in config files for
6492 *) Support for nameConstraints certificate extension.
6495 *) Support for policyConstraints certificate extension.
6498 *) Support for policyMappings certificate extension.
6501 *) Make sure the default DSA_METHOD implementation only uses its
6502 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6503 and change its own handlers to be NULL so as to remove unnecessary
6504 indirection. This lets alternative implementations fallback to the
6505 default implementation more easily.
6508 *) Support for directoryName in GeneralName related extensions
6512 *) Make it possible to link applications using Makefile.shared.
6513 Make that possible even when linking against static libraries!
6516 *) Support for single pass processing for S/MIME signing. This now
6517 means that S/MIME signing can be done from a pipe, in addition
6518 cleartext signing (multipart/signed type) is effectively streaming
6519 and the signed data does not need to be all held in memory.
6521 This is done with a new flag PKCS7_STREAM. When this flag is set
6522 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6523 is done after the data is output (and digests calculated) in
6524 SMIME_write_PKCS7().
6527 *) Add full support for -rpath/-R, both in shared libraries and
6528 applications, at least on the platforms where it's known how
6532 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6533 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6534 will now compute a table of multiples of the generator that
6535 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6536 faster (notably in the case of a single point multiplication,
6537 scalar * generator).
6538 [Nils Larsch, Bodo Moeller]
6540 *) IPv6 support for certificate extensions. The various extensions
6541 which use the IP:a.b.c.d can now take IPv6 addresses using the
6542 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6546 *) Added an ENGINE that implements RSA by performing private key
6547 exponentiations with the GMP library. The conversions to and from
6548 GMP's mpz_t format aren't optimised nor are any montgomery forms
6549 cached, and on x86 it appears OpenSSL's own performance has caught up.
6550 However there are likely to be other architectures where GMP could
6551 provide a boost. This ENGINE is not built in by default, but it can be
6552 specified at Configure time and should be accompanied by the necessary
6553 linker additions, eg;
6554 ./config -DOPENSSL_USE_GMP -lgmp
6557 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6558 testing availability of engines with "-t" - the old behaviour is
6559 produced by increasing the feature's verbosity with "-tt".
6562 *) ECDSA routines: under certain error conditions uninitialized BN objects
6563 could be freed. Solution: make sure initialization is performed early
6564 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6568 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6569 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6570 software implementations. For DSA and DH, parameter generation can
6571 also be overridden by providing the appropriate method callbacks.
6574 *) Change the "progress" mechanism used in key-generation and
6575 primality testing to functions that take a new BN_GENCB pointer in
6576 place of callback/argument pairs. The new API functions have "_ex"
6577 postfixes and the older functions are reimplemented as wrappers for
6578 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6579 declarations of the old functions to help (graceful) attempts to
6580 migrate to the new functions. Also, the new key-generation API
6581 functions operate on a caller-supplied key-structure and return
6582 success/failure rather than returning a key or NULL - this is to
6583 help make "keygen" another member function of RSA_METHOD etc.
6585 Example for using the new callback interface:
6587 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6591 BN_GENCB_set(&my_cb, my_callback, my_arg);
6593 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6594 /* For the meaning of a, b in calls to my_callback(), see the
6595 * documentation of the function that calls the callback.
6596 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6597 * my_callback should return 1 if it wants BN_is_prime_ex()
6598 * to continue, or 0 to stop.
6603 *) Change the ZLIB compression method to be stateful, and make it
6604 available to TLS with the number defined in
6605 draft-ietf-tls-compression-04.txt.
6608 *) Add the ASN.1 structures and functions for CertificatePair, which
6609 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6611 CertificatePair ::= SEQUENCE {
6612 forward [0] Certificate OPTIONAL,
6613 reverse [1] Certificate OPTIONAL,
6614 -- at least one of the pair shall be present -- }
6616 Also implement the PEM functions to read and write certificate
6617 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6619 This needed to be defined, mostly for the sake of the LDAP
6620 attribute crossCertificatePair, but may prove useful elsewhere as
6624 *) Make it possible to inhibit symlinking of shared libraries in
6625 Makefile.shared, for Cygwin's sake.
6628 *) Extend the BIGNUM API by creating a function
6629 void BN_set_negative(BIGNUM *a, int neg);
6630 and a macro that behave like
6631 int BN_is_negative(const BIGNUM *a);
6633 to avoid the need to access 'a->neg' directly in applications.
6636 *) Implement fast modular reduction for pseudo-Mersenne primes
6637 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6638 EC_GROUP_new_curve_GFp() will now automatically use this
6640 [Nils Larsch <nla@trustcenter.de>]
6642 *) Add new lock type (CRYPTO_LOCK_BN).
6645 *) Change the ENGINE framework to automatically load engines
6646 dynamically from specific directories unless they could be
6647 found to already be built in or loaded. Move all the
6648 current engines except for the cryptodev one to a new
6650 The engines in engines/ are built as shared libraries if
6651 the "shared" options was given to ./Configure or ./config.
6652 Otherwise, they are inserted in libcrypto.a.
6653 /usr/local/ssl/engines is the default directory for dynamic
6654 engines, but that can be overridden at configure time through
6655 the usual use of --prefix and/or --openssldir, and at run
6656 time with the environment variable OPENSSL_ENGINES.
6657 [Geoff Thorpe and Richard Levitte]
6659 *) Add Makefile.shared, a helper makefile to build shared
6660 libraries. Adapt Makefile.org.
6663 *) Add version info to Win32 DLLs.
6664 [Peter 'Luna' Runestig" <peter@runestig.com>]
6666 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6667 can be added using this API to created arbitrary PKCS#12
6668 files while avoiding the low level API.
6670 New options to PKCS12_create(), key or cert can be NULL and
6671 will then be omitted from the output file. The encryption
6672 algorithm NIDs can be set to -1 for no encryption, the mac
6673 iteration count can be set to 0 to omit the mac.
6675 Enhance pkcs12 utility by making the -nokeys and -nocerts
6676 options work when creating a PKCS#12 file. New option -nomac
6677 to omit the mac, NONE can be set for an encryption algorithm.
6678 New code is modified to use the enhanced PKCS12_create()
6679 instead of the low level API.
6682 *) Extend ASN1 encoder to support indefinite length constructed
6683 encoding. This can output sequences tags and octet strings in
6684 this form. Modify pk7_asn1.c to support indefinite length
6685 encoding. This is experimental and needs additional code to
6686 be useful, such as an ASN1 bio and some enhanced streaming
6689 Extend template encode functionality so that tagging is passed
6690 down to the template encoder.
6693 *) Let 'openssl req' fail if an argument to '-newkey' is not
6694 recognized instead of using RSA as a default.
6697 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6698 As these are not official, they are not included in "ALL";
6699 the "ECCdraft" ciphersuite group alias can be used to select them.
6700 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6702 *) Add ECDH engine support.
6703 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6705 *) Add ECDH in new directory crypto/ecdh/.
6706 [Douglas Stebila (Sun Microsystems Laboratories)]
6708 *) Let BN_rand_range() abort with an error after 100 iterations
6709 without success (which indicates a broken PRNG).
6712 *) Change BN_mod_sqrt() so that it verifies that the input value
6713 is really the square of the return value. (Previously,
6714 BN_mod_sqrt would show GIGO behaviour.)
6717 *) Add named elliptic curves over binary fields from X9.62, SECG,
6718 and WAP/WTLS; add OIDs that were still missing.
6720 [Sheueling Chang Shantz and Douglas Stebila
6721 (Sun Microsystems Laboratories)]
6723 *) Extend the EC library for elliptic curves over binary fields
6724 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6727 EC_GF2m_simple_method
6731 EC_GROUP_new_curve_GF2m
6732 EC_GROUP_set_curve_GF2m
6733 EC_GROUP_get_curve_GF2m
6734 EC_POINT_set_affine_coordinates_GF2m
6735 EC_POINT_get_affine_coordinates_GF2m
6736 EC_POINT_set_compressed_coordinates_GF2m
6738 Point compression for binary fields is disabled by default for
6739 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6742 As binary polynomials are represented as BIGNUMs, various members
6743 of the EC_GROUP and EC_POINT data structures can be shared
6744 between the implementations for prime fields and binary fields;
6745 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6746 are essentially identical to their ..._GFp counterparts.
6747 (For simplicity, the '..._GFp' prefix has been dropped from
6748 various internal method names.)
6750 An internal 'field_div' method (similar to 'field_mul' and
6751 'field_sqr') has been added; this is used only for binary fields.
6753 [Sheueling Chang Shantz and Douglas Stebila
6754 (Sun Microsystems Laboratories)]
6756 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6757 through methods ('mul', 'precompute_mult').
6759 The generic implementations (now internally called 'ec_wNAF_mul'
6760 and 'ec_wNAF_precomputed_mult') remain the default if these
6761 methods are undefined.
6763 [Sheueling Chang Shantz and Douglas Stebila
6764 (Sun Microsystems Laboratories)]
6766 *) New function EC_GROUP_get_degree, which is defined through
6767 EC_METHOD. For curves over prime fields, this returns the bit
6768 length of the modulus.
6770 [Sheueling Chang Shantz and Douglas Stebila
6771 (Sun Microsystems Laboratories)]
6773 *) New functions EC_GROUP_dup, EC_POINT_dup.
6774 (These simply call ..._new and ..._copy).
6776 [Sheueling Chang Shantz and Douglas Stebila
6777 (Sun Microsystems Laboratories)]
6779 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6780 Polynomials are represented as BIGNUMs (where the sign bit is not
6781 used) in the following functions [macros]:
6784 BN_GF2m_sub [= BN_GF2m_add]
6785 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6786 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6787 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6789 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6790 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6791 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6792 BN_GF2m_cmp [= BN_ucmp]
6794 (Note that only the 'mod' functions are actually for fields GF(2^m).
6795 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6797 For some functions, an the irreducible polynomial defining a
6798 field can be given as an 'unsigned int[]' with strictly
6799 decreasing elements giving the indices of those bits that are set;
6800 i.e., p[] represents the polynomial
6801 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6803 p[0] > p[1] > ... > p[k] = 0.
6804 This applies to the following functions:
6809 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6810 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6812 BN_GF2m_mod_sqrt_arr
6813 BN_GF2m_mod_solve_quad_arr
6817 Conversion can be performed by the following functions:
6822 bntest.c has additional tests for binary polynomial arithmetic.
6824 Two implementations for BN_GF2m_mod_div() are available.
6825 The default algorithm simply uses BN_GF2m_mod_inv() and
6826 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6827 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6828 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6830 [Sheueling Chang Shantz and Douglas Stebila
6831 (Sun Microsystems Laboratories)]
6833 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6834 functionality is disabled at compile-time.
6835 [Douglas Stebila <douglas.stebila@sun.com>]
6837 *) Change default behaviour of 'openssl asn1parse' so that more
6838 information is visible when viewing, e.g., a certificate:
6840 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6841 mode the content of non-printable OCTET STRINGs is output in a
6842 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6843 avoid the appearance of a printable string.
6844 [Nils Larsch <nla@trustcenter.de>]
6846 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6848 EC_GROUP_set_asn1_flag()
6849 EC_GROUP_get_asn1_flag()
6850 EC_GROUP_set_point_conversion_form()
6851 EC_GROUP_get_point_conversion_form()
6852 These control ASN1 encoding details:
6853 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6854 has been set to OPENSSL_EC_NAMED_CURVE.
6855 - Points are encoded in uncompressed form by default; options for
6856 asn1_for are as for point2oct, namely
6857 POINT_CONVERSION_COMPRESSED
6858 POINT_CONVERSION_UNCOMPRESSED
6859 POINT_CONVERSION_HYBRID
6861 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6864 EC_GROUP_get0_seed()
6865 EC_GROUP_get_seed_len()
6866 This is used only for ASN1 purposes (so far).
6867 [Nils Larsch <nla@trustcenter.de>]
6869 *) Add 'field_type' member to EC_METHOD, which holds the NID
6870 of the appropriate field type OID. The new function
6871 EC_METHOD_get_field_type() returns this value.
6872 [Nils Larsch <nla@trustcenter.de>]
6877 EC_POINT_point2hex()
6878 EC_POINT_hex2point()
6879 providing useful interfaces to EC_POINT_point2oct() and
6880 EC_POINT_oct2point().
6881 [Nils Larsch <nla@trustcenter.de>]
6883 *) Change internals of the EC library so that the functions
6884 EC_GROUP_set_generator()
6885 EC_GROUP_get_generator()
6886 EC_GROUP_get_order()
6887 EC_GROUP_get_cofactor()
6888 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6889 to methods, which would lead to unnecessary code duplication when
6890 adding different types of curves.
6891 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6893 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6894 arithmetic, and such that modified wNAFs are generated
6895 (which avoid length expansion in many cases).
6898 *) Add a function EC_GROUP_check_discriminant() (defined via
6899 EC_METHOD) that verifies that the curve discriminant is non-zero.
6901 Add a function EC_GROUP_check() that makes some sanity tests
6902 on a EC_GROUP, its generator and order. This includes
6903 EC_GROUP_check_discriminant().
6904 [Nils Larsch <nla@trustcenter.de>]
6906 *) Add ECDSA in new directory crypto/ecdsa/.
6908 Add applications 'openssl ecparam' and 'openssl ecdsa'
6909 (these are based on 'openssl dsaparam' and 'openssl dsa').
6911 ECDSA support is also included in various other files across the
6912 library. Most notably,
6913 - 'openssl req' now has a '-newkey ecdsa:file' option;
6914 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6915 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6916 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6917 them suitable for ECDSA where domain parameters must be
6918 extracted before the specific public key;
6919 - ECDSA engine support has been added.
6920 [Nils Larsch <nla@trustcenter.de>]
6922 *) Include some named elliptic curves, and add OIDs from X9.62,
6923 SECG, and WAP/WTLS. Each curve can be obtained from the new
6925 EC_GROUP_new_by_curve_name(),
6926 and the list of available named curves can be obtained with
6927 EC_get_builtin_curves().
6928 Also add a 'curve_name' member to EC_GROUP objects, which can be
6930 EC_GROUP_set_curve_name()
6931 EC_GROUP_get_curve_name()
6932 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6934 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6935 was actually never needed) and in BN_mul(). The removal in BN_mul()
6936 required a small change in bn_mul_part_recursive() and the addition
6937 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6938 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6939 bn_sub_words() and bn_add_words() except they take arrays with
6943 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6945 *) Cleanse PEM buffers before freeing them since they may contain
6947 [Benjamin Bennett <ben@psc.edu>]
6949 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6950 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6951 authentication-only ciphersuites.
6954 *) Since AES128 and AES256 share a single mask bit in the logic of
6955 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6956 kludge to work properly if AES128 is available and AES256 isn't.
6959 *) Expand security boundary to match 1.1.1 module.
6962 *) Remove redundant features: hash file source, editing of test vectors
6963 modify fipsld to use external fips_premain.c signature.
6966 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6967 run algorithm test programs.
6970 *) Make algorithm test programs more tolerant of whitespace.
6973 *) Have SSL/TLS server implementation tolerate "mismatched" record
6974 protocol version while receiving ClientHello even if the
6975 ClientHello is fragmented. (The server can't insist on the
6976 particular protocol version it has chosen before the ServerHello
6977 message has informed the client about his choice.)
6980 *) Load error codes if they are not already present instead of using a
6981 static variable. This allows them to be cleanly unloaded and reloaded.
6984 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6986 *) Introduce limits to prevent malicious keys being able to
6987 cause a denial of service. (CVE-2006-2940)
6988 [Steve Henson, Bodo Moeller]
6990 *) Fix ASN.1 parsing of certain invalid structures that can result
6991 in a denial of service. (CVE-2006-2937) [Steve Henson]
6993 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6994 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6996 *) Fix SSL client code which could crash if connecting to a
6997 malicious SSLv2 server. (CVE-2006-4343)
6998 [Tavis Ormandy and Will Drewry, Google Security Team]
7000 *) Change ciphersuite string processing so that an explicit
7001 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
7002 will no longer include "AES128-SHA"), and any other similar
7003 ciphersuite (same bitmap) from *other* protocol versions (so that
7004 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
7005 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
7006 changes from 0.9.8b and 0.9.8d.
7009 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
7011 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
7012 (CVE-2006-4339) [Ben Laurie and Google Security Team]
7014 *) Change the Unix randomness entropy gathering to use poll() when
7015 possible instead of select(), since the latter has some
7016 undesirable limitations.
7017 [Darryl Miles via Richard Levitte and Bodo Moeller]
7019 *) Disable rogue ciphersuites:
7021 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
7022 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
7023 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
7025 The latter two were purportedly from
7026 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
7029 Also deactivate the remaining ciphersuites from
7030 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
7031 unofficial, and the ID has long expired.
7034 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
7035 dual-core machines) and other potential thread-safety issues.
7038 Changes between 0.9.7i and 0.9.7j [04 May 2006]
7040 *) Adapt fipsld and the build system to link against the validated FIPS
7041 module in FIPS mode.
7044 *) Fixes for VC++ 2005 build under Windows.
7047 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
7048 from a Windows bash shell such as MSYS. It is autodetected from the
7049 "config" script when run from a VC++ environment. Modify standard VC++
7050 build to use fipscanister.o from the GNU make build.
7053 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
7055 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
7056 The value now differs depending on if you build for FIPS or not.
7057 BEWARE! A program linked with a shared FIPSed libcrypto can't be
7058 safely run with a non-FIPSed libcrypto, as it may crash because of
7059 the difference induced by this change.
7062 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
7064 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
7065 (part of SSL_OP_ALL). This option used to disable the
7066 countermeasure against man-in-the-middle protocol-version
7067 rollback in the SSL 2.0 server implementation, which is a bad
7068 idea. (CVE-2005-2969)
7070 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
7071 for Information Security, National Institute of Advanced Industrial
7072 Science and Technology [AIST], Japan)]
7074 *) Minimal support for X9.31 signatures and PSS padding modes. This is
7075 mainly for FIPS compliance and not fully integrated at this stage.
7078 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
7079 the exponentiation using a fixed-length exponent. (Otherwise,
7080 the information leaked through timing could expose the secret key
7081 after many signatures; cf. Bleichenbacher's attack on DSA with
7085 *) Make a new fixed-window mod_exp implementation the default for
7086 RSA, DSA, and DH private-key operations so that the sequence of
7087 squares and multiplies and the memory access pattern are
7088 independent of the particular secret key. This will mitigate
7089 cache-timing and potential related attacks.
7091 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
7092 and this is automatically used by BN_mod_exp_mont() if the new flag
7093 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
7094 will use this BN flag for private exponents unless the flag
7095 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
7096 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
7098 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
7100 *) Change the client implementation for SSLv23_method() and
7101 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
7102 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
7103 (Previously, the SSL 2.0 backwards compatible Client Hello
7104 message format would be used even with SSL_OP_NO_SSLv2.)
7107 *) Add support for smime-type MIME parameter in S/MIME messages which some
7111 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
7112 a threadsafe manner. Modify rsa code to use new function and add calls
7113 to dsa and dh code (which had race conditions before).
7116 *) Include the fixed error library code in the C error file definitions
7117 instead of fixing them up at runtime. This keeps the error code
7118 structures constant.
7121 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
7123 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
7126 *) Fixes for newer kerberos headers. NB: the casts are needed because
7127 the 'length' field is signed on one version and unsigned on another
7128 with no (?) obvious way to tell the difference, without these VC++
7129 complains. Also the "definition" of FAR (blank) is no longer included
7130 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
7131 some needed definitions.
7134 *) Undo Cygwin change.
7137 *) Added support for proxy certificates according to RFC 3820.
7138 Because they may be a security thread to unaware applications,
7139 they must be explicitly allowed in run-time. See
7140 docs/HOWTO/proxy_certificates.txt for further information.
7143 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
7145 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
7146 server and client random values. Previously
7147 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
7148 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
7150 This change has negligible security impact because:
7152 1. Server and client random values still have 24 bytes of pseudo random
7155 2. Server and client random values are sent in the clear in the initial
7158 3. The master secret is derived using the premaster secret (48 bytes in
7159 size for static RSA ciphersuites) as well as client server and random
7162 The OpenSSL team would like to thank the UK NISCC for bringing this issue
7165 [Stephen Henson, reported by UK NISCC]
7167 *) Use Windows randomness collection on Cygwin.
7170 *) Fix hang in EGD/PRNGD query when communication socket is closed
7171 prematurely by EGD/PRNGD.
7172 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
7174 *) Prompt for pass phrases when appropriate for PKCS12 input format.
7177 *) Back-port of selected performance improvements from development
7178 branch, as well as improved support for PowerPC platforms.
7181 *) Add lots of checks for memory allocation failure, error codes to indicate
7182 failure and freeing up memory if a failure occurs.
7183 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
7185 *) Add new -passin argument to dgst.
7188 *) Perform some character comparisons of different types in X509_NAME_cmp:
7189 this is needed for some certificates that re-encode DNs into UTF8Strings
7190 (in violation of RFC3280) and can't or won't issue name rollover
7194 *) Make an explicit check during certificate validation to see that
7195 the CA setting in each certificate on the chain is correct. As a
7196 side effect always do the following basic checks on extensions,
7197 not just when there's an associated purpose to the check:
7199 - if there is an unhandled critical extension (unless the user
7200 has chosen to ignore this fault)
7201 - if the path length has been exceeded (if one is set at all)
7202 - that certain extensions fit the associated purpose (if one has
7206 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
7208 *) Avoid a race condition when CRLs are checked in a multi threaded
7209 environment. This would happen due to the reordering of the revoked
7210 entries during signature checking and serial number lookup. Now the
7211 encoding is cached and the serial number sort performed under a lock.
7212 Add new STACK function sk_is_sorted().
7215 *) Add Delta CRL to the extension code.
7218 *) Various fixes to s3_pkt.c so alerts are sent properly.
7219 [David Holmes <d.holmes@f5.com>]
7221 *) Reduce the chances of duplicate issuer name and serial numbers (in
7222 violation of RFC3280) using the OpenSSL certificate creation utilities.
7223 This is done by creating a random 64 bit value for the initial serial
7224 number when a serial number file is created or when a self signed
7225 certificate is created using 'openssl req -x509'. The initial serial
7226 number file is created using 'openssl x509 -next_serial' in CA.pl
7227 rather than being initialized to 1.
7230 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
7232 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7233 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7234 [Joe Orton, Steve Henson]
7236 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
7238 [Joe Orton, Steve Henson]
7240 *) Make it possible to have multiple active certificates with the same
7241 subject in the CA index file. This is done only if the keyword
7242 'unique_subject' is set to 'no' in the main CA section (default
7243 if 'CA_default') of the configuration file. The value is saved
7244 with the database itself in a separate index attribute file,
7245 named like the index file with '.attr' appended to the name.
7248 *) X509 verify fixes. Disable broken certificate workarounds when
7249 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
7250 keyUsage extension present. Don't accept CRLs with unhandled critical
7251 extensions: since verify currently doesn't process CRL extensions this
7252 rejects a CRL with *any* critical extensions. Add new verify error codes
7256 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
7257 A clarification of RFC2560 will require the use of OCTET STRINGs and
7258 some implementations cannot handle the current raw format. Since OpenSSL
7259 copies and compares OCSP nonces as opaque blobs without any attempt at
7260 parsing them this should not create any compatibility issues.
7263 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
7264 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
7265 this HMAC (and other) operations are several times slower than OpenSSL
7269 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
7270 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7272 *) Use the correct content when signing type "other".
7275 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
7277 *) Fix various bugs revealed by running the NISCC test suite:
7279 Stop out of bounds reads in the ASN1 code when presented with
7280 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7282 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
7284 If verify callback ignores invalid public key errors don't try to check
7285 certificate signature with the NULL public key.
7289 *) New -ignore_err option in ocsp application to stop the server
7290 exiting on the first error in a request.
7293 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7294 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7298 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7299 extra data after the compression methods not only for TLS 1.0
7300 but also for SSL 3.0 (as required by the specification).
7301 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7303 *) Change X509_certificate_type() to mark the key as exported/exportable
7304 when it's 512 *bits* long, not 512 bytes.
7307 *) Change AES_cbc_encrypt() so it outputs exact multiple of
7308 blocks during encryption.
7311 *) Various fixes to base64 BIO and non blocking I/O. On write
7312 flushes were not handled properly if the BIO retried. On read
7313 data was not being buffered properly and had various logic bugs.
7314 This also affects blocking I/O when the data being decoded is a
7318 *) Various S/MIME bugfixes and compatibility changes:
7319 output correct application/pkcs7 MIME type if
7320 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
7321 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
7322 of files as .eml work). Correctly handle very long lines in MIME
7326 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
7328 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7329 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7330 a protocol version number mismatch like a decryption error
7331 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7334 *) Turn on RSA blinding by default in the default implementation
7335 to avoid a timing attack. Applications that don't want it can call
7336 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7337 They would be ill-advised to do so in most cases.
7338 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7340 *) Change RSA blinding code so that it works when the PRNG is not
7341 seeded (in this case, the secret RSA exponent is abused as
7342 an unpredictable seed -- if it is not unpredictable, there
7343 is no point in blinding anyway). Make RSA blinding thread-safe
7344 by remembering the creator's thread ID in rsa->blinding and
7345 having all other threads use local one-time blinding factors
7346 (this requires more computation than sharing rsa->blinding, but
7347 avoids excessive locking; and if an RSA object is not shared
7348 between threads, blinding will still be very fast).
7351 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
7352 ENGINE as defaults for all supported algorithms irrespective of
7353 the 'flags' parameter. 'flags' is now honoured, so applications
7354 should make sure they are passing it correctly.
7357 *) Target "mingw" now allows native Windows code to be generated in
7358 the Cygwin environment as well as with the MinGW compiler.
7361 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7363 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7364 via timing by performing a MAC computation even if incorrect
7365 block cipher padding has been found. This is a countermeasure
7366 against active attacks where the attacker has to distinguish
7367 between bad padding and a MAC verification error. (CVE-2003-0078)
7369 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7370 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7371 Martin Vuagnoux (EPFL, Ilion)]
7373 *) Make the no-err option work as intended. The intention with no-err
7374 is not to have the whole error stack handling routines removed from
7375 libcrypto, it's only intended to remove all the function name and
7376 reason texts, thereby removing some of the footprint that may not
7377 be interesting if those errors aren't displayed anyway.
7379 NOTE: it's still possible for any application or module to have its
7380 own set of error texts inserted. The routines are there, just not
7381 used by default when no-err is given.
7384 *) Add support for FreeBSD on IA64.
7385 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7387 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7388 Kerberos function mit_des_cbc_cksum(). Before this change,
7389 the value returned by DES_cbc_cksum() was like the one from
7390 mit_des_cbc_cksum(), except the bytes were swapped.
7391 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7393 *) Allow an application to disable the automatic SSL chain building.
7394 Before this a rather primitive chain build was always performed in
7395 ssl3_output_cert_chain(): an application had no way to send the
7396 correct chain if the automatic operation produced an incorrect result.
7398 Now the chain builder is disabled if either:
7400 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7402 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7404 The reasoning behind this is that an application would not want the
7405 auto chain building to take place if extra chain certificates are
7406 present and it might also want a means of sending no additional
7407 certificates (for example the chain has two certificates and the
7411 *) Add the possibility to build without the ENGINE framework.
7412 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7414 *) Under Win32 gmtime() can return NULL: check return value in
7415 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7418 *) DSA routines: under certain error conditions uninitialized BN objects
7419 could be freed. Solution: make sure initialization is performed early
7420 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7421 Nils Larsch <nla@trustcenter.de> via PR#459)
7424 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7425 checked on reconnect on the client side, therefore session resumption
7426 could still fail with a "ssl session id is different" error. This
7427 behaviour is masked when SSL_OP_ALL is used due to
7428 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7429 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7430 followup to PR #377.
7433 *) IA-32 assembler support enhancements: unified ELF targets, support
7434 for SCO/Caldera platforms, fix for Cygwin shared build.
7437 *) Add support for FreeBSD on sparc64. As a consequence, support for
7438 FreeBSD on non-x86 processors is separate from x86 processors on
7439 the config script, much like the NetBSD support.
7440 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7442 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7444 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7447 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7448 code (06) was taken as the first octet of the session ID and the last
7449 octet was ignored consequently. As a result SSLv2 client side session
7450 caching could not have worked due to the session ID mismatch between
7452 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7456 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7457 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7461 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7462 seems that in spite of existing for more than a year, many application
7463 author have done nothing to provide the necessary callbacks, which
7464 means that this particular engine will not work properly anywhere.
7465 This is a very unfortunate situation which forces us, in the name
7466 of usability, to give the hw_ncipher.c a static lock, which is part
7468 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7469 appear in 0.9.8 or later. We EXPECT application authors to have
7470 dealt properly with this when 0.9.8 is released (unless we actually
7471 make such changes in the libcrypto locking code that changes will
7472 have to be made anyway).
7475 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7476 octets have been read, EOF or an error occurs. Without this change
7477 some truncated ASN1 structures will not produce an error.
7480 *) Disable Heimdal support, since it hasn't been fully implemented.
7481 Still give the possibility to force the use of Heimdal, but with
7482 warnings and a request that patches get sent to openssl-dev.
7485 *) Add the VC-CE target, introduce the WINCE sysname, and add
7486 INSTALL.WCE and appropriate conditionals to make it build.
7487 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7489 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7490 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7491 edit numbers of the version.
7492 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7494 *) Introduce safe string copy and catenation functions
7495 (BUF_strlcpy() and BUF_strlcat()).
7496 [Ben Laurie (CHATS) and Richard Levitte]
7498 *) Avoid using fixed-size buffers for one-line DNs.
7499 [Ben Laurie (CHATS)]
7501 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7502 resizing buffers containing secrets, and use where appropriate.
7503 [Ben Laurie (CHATS)]
7505 *) Avoid using fixed size buffers for configuration file location.
7506 [Ben Laurie (CHATS)]
7508 *) Avoid filename truncation for various CA files.
7509 [Ben Laurie (CHATS)]
7511 *) Use sizeof in preference to magic numbers.
7512 [Ben Laurie (CHATS)]
7514 *) Avoid filename truncation in cert requests.
7515 [Ben Laurie (CHATS)]
7517 *) Add assertions to check for (supposedly impossible) buffer
7519 [Ben Laurie (CHATS)]
7521 *) Don't cache truncated DNS entries in the local cache (this could
7522 potentially lead to a spoofing attack).
7523 [Ben Laurie (CHATS)]
7525 *) Fix various buffers to be large enough for hex/decimal
7526 representations in a platform independent manner.
7527 [Ben Laurie (CHATS)]
7529 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7530 resizing buffers containing secrets, and use where appropriate.
7531 [Ben Laurie (CHATS)]
7533 *) Add BIO_indent() to avoid much slightly worrying code to do
7535 [Ben Laurie (CHATS)]
7537 *) Convert sprintf()/BIO_puts() to BIO_printf().
7538 [Ben Laurie (CHATS)]
7540 *) buffer_gets() could terminate with the buffer only half
7542 [Ben Laurie (CHATS)]
7544 *) Add assertions to prevent user-supplied crypto functions from
7545 overflowing internal buffers by having large block sizes, etc.
7546 [Ben Laurie (CHATS)]
7548 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7550 [Ben Laurie (CHATS)]
7552 *) Eliminate unused copy of key in RC4.
7553 [Ben Laurie (CHATS)]
7555 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7556 [Ben Laurie (CHATS)]
7558 *) Fix off-by-one error in EGD path.
7559 [Ben Laurie (CHATS)]
7561 *) If RANDFILE path is too long, ignore instead of truncating.
7562 [Ben Laurie (CHATS)]
7564 *) Eliminate unused and incorrectly sized X.509 structure
7566 [Ben Laurie (CHATS)]
7568 *) Eliminate unused and dangerous function knumber().
7569 [Ben Laurie (CHATS)]
7571 *) Eliminate unused and dangerous structure, KSSL_ERR.
7572 [Ben Laurie (CHATS)]
7574 *) Protect against overlong session ID context length in an encoded
7575 session object. Since these are local, this does not appear to be
7577 [Ben Laurie (CHATS)]
7579 *) Change from security patch (see 0.9.6e below) that did not affect
7580 the 0.9.6 release series:
7582 Remote buffer overflow in SSL3 protocol - an attacker could
7583 supply an oversized master key in Kerberos-enabled versions.
7585 [Ben Laurie (CHATS)]
7587 *) Change the SSL kerb5 codes to match RFC 2712.
7590 *) Make -nameopt work fully for req and add -reqopt switch.
7591 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7593 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7594 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7596 *) Make sure tests can be performed even if the corresponding algorithms
7597 have been removed entirely. This was also the last step to make
7598 OpenSSL compilable with DJGPP under all reasonable conditions.
7599 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7601 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7602 to allow version independent disabling of normally unselected ciphers,
7603 which may be activated as a side-effect of selecting a single cipher.
7605 (E.g., cipher list string "RSA" enables ciphersuites that are left
7606 out of "ALL" because they do not provide symmetric encryption.
7607 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7608 [Lutz Jaenicke, Bodo Moeller]
7610 *) Add appropriate support for separate platform-dependent build
7611 directories. The recommended way to make a platform-dependent
7612 build directory is the following (tested on Linux), maybe with
7615 # Place yourself outside of the OpenSSL source tree. In
7616 # this example, the environment variable OPENSSL_SOURCE
7617 # is assumed to contain the absolute OpenSSL source directory.
7618 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7619 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7620 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7621 mkdir -p `dirname $F`
7622 ln -s $OPENSSL_SOURCE/$F $F
7625 To be absolutely sure not to disturb the source tree, a "make clean"
7626 is a good thing. If it isn't successful, don't worry about it,
7627 it probably means the source directory is very clean.
7630 *) Make sure any ENGINE control commands make local copies of string
7631 pointers passed to them whenever necessary. Otherwise it is possible
7632 the caller may have overwritten (or deallocated) the original string
7633 data when a later ENGINE operation tries to use the stored values.
7634 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7636 *) Improve diagnostics in file reading and command-line digests.
7637 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7639 *) Add AES modes CFB and OFB to the object database. Correct an
7640 error in AES-CFB decryption.
7643 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7644 allows existing EVP_CIPHER_CTX structures to be reused after
7645 calling EVP_*Final(). This behaviour is used by encryption
7646 BIOs and some applications. This has the side effect that
7647 applications must explicitly clean up cipher contexts with
7648 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7651 *) Check the values of dna and dnb in bn_mul_recursive before calling
7652 bn_mul_comba (a non zero value means the a or b arrays do not contain
7653 n2 elements) and fallback to bn_mul_normal if either is not zero.
7656 *) Fix escaping of non-ASCII characters when using the -subj option
7657 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7660 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7661 form for "surname", serialNumber has no short form.
7662 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7663 therefore remove "mail" short name for "internet 7".
7664 The OID for unique identifiers in X509 certificates is
7665 x500UniqueIdentifier, not uniqueIdentifier.
7666 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7669 *) Add an "init" command to the ENGINE config module and auto initialize
7670 ENGINEs. Without any "init" command the ENGINE will be initialized
7671 after all ctrl commands have been executed on it. If init=1 the
7672 ENGINE is initialized at that point (ctrls before that point are run
7673 on the uninitialized ENGINE and after on the initialized one). If
7674 init=0 then the ENGINE will not be initialized at all.
7677 *) Fix the 'app_verify_callback' interface so that the user-defined
7678 argument is actually passed to the callback: In the
7679 SSL_CTX_set_cert_verify_callback() prototype, the callback
7680 declaration has been changed from
7683 int (*cb)(X509_STORE_CTX *,void *);
7684 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7685 i=s->ctx->app_verify_callback(&ctx)
7686 has been changed into
7687 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7689 To update applications using SSL_CTX_set_cert_verify_callback(),
7690 a dummy argument can be added to their callback functions.
7691 [D. K. Smetters <smetters@parc.xerox.com>]
7693 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7694 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7696 *) Add and OPENSSL_LOAD_CONF define which will cause
7697 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7698 This allows older applications to transparently support certain
7699 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7700 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7701 load the config file and OPENSSL_add_all_algorithms_conf() which will
7702 always load it have also been added.
7705 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7706 Adjust NIDs and EVP layer.
7707 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7709 *) Config modules support in openssl utility.
7711 Most commands now load modules from the config file,
7712 though in a few (such as version) this isn't done
7713 because it couldn't be used for anything.
7715 In the case of ca and req the config file used is
7716 the same as the utility itself: that is the -config
7717 command line option can be used to specify an
7721 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7722 use "openssl_conf" if filename is NULL use default openssl config file.
7725 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7726 config section name. Add a new flag to tolerate a missing config file
7727 and move code to CONF_modules_load_file().
7730 *) Support for crypto accelerator cards from Accelerated Encryption
7731 Processing, www.aep.ie. (Use engine 'aep')
7732 The support was copied from 0.9.6c [engine] and adapted/corrected
7733 to work with the new engine framework.
7734 [AEP Inc. and Richard Levitte]
7736 *) Support for SureWare crypto accelerator cards from Baltimore
7737 Technologies. (Use engine 'sureware')
7738 The support was copied from 0.9.6c [engine] and adapted
7739 to work with the new engine framework.
7742 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7743 make the newer ENGINE framework commands for the CHIL engine work.
7744 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7746 *) Make it possible to produce shared libraries on ReliantUNIX.
7747 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7749 *) Add the configuration target debug-linux-ppro.
7750 Make 'openssl rsa' use the general key loading routines
7751 implemented in apps.c, and make those routines able to
7752 handle the key format FORMAT_NETSCAPE and the variant
7754 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7756 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7757 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7759 *) Add -keyform to rsautl, and document -engine.
7760 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7762 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7763 BIO_R_NO_SUCH_FILE error code rather than the generic
7764 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7767 *) Add new functions
7769 ERR_peek_last_error_line
7770 ERR_peek_last_error_line_data.
7771 These are similar to
7774 ERR_peek_error_line_data,
7775 but report on the latest error recorded rather than the first one
7776 still in the error queue.
7777 [Ben Laurie, Bodo Moeller]
7779 *) default_algorithms option in ENGINE config module. This allows things
7781 default_algorithms = ALL
7782 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7785 *) Preliminary ENGINE config module.
7788 *) New experimental application configuration code.
7791 *) Change the AES code to follow the same name structure as all other
7792 symmetric ciphers, and behave the same way. Move everything to
7793 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7794 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7796 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7797 [Ben Laurie and Theo de Raadt]
7799 *) Add option to output public keys in req command.
7800 [Massimiliano Pala madwolf@openca.org]
7802 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7803 (up to about 10% better than before for P-192 and P-224).
7806 *) New functions/macros
7808 SSL_CTX_set_msg_callback(ctx, cb)
7809 SSL_CTX_set_msg_callback_arg(ctx, arg)
7810 SSL_set_msg_callback(ssl, cb)
7811 SSL_set_msg_callback_arg(ssl, arg)
7813 to request calling a callback function
7815 void cb(int write_p, int version, int content_type,
7816 const void *buf, size_t len, SSL *ssl, void *arg)
7818 whenever a protocol message has been completely received
7819 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7820 protocol version according to which the SSL library interprets
7821 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7822 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7823 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7824 specification (change_cipher_spec(20), alert(21), handshake(22)).
7825 'buf' and 'len' point to the actual message, 'ssl' to the
7826 SSL object, and 'arg' is the application-defined value set by
7827 SSL[_CTX]_set_msg_callback_arg().
7829 'openssl s_client' and 'openssl s_server' have new '-msg' options
7830 to enable a callback that displays all protocol messages.
7833 *) Change the shared library support so shared libraries are built as
7834 soon as the corresponding static library is finished, and thereby get
7835 openssl and the test programs linked against the shared library.
7836 This still only happens when the keyword "shard" has been given to
7837 the configuration scripts.
7839 NOTE: shared library support is still an experimental thing, and
7840 backward binary compatibility is still not guaranteed.
7841 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7843 *) Add support for Subject Information Access extension.
7844 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7846 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7847 additional bytes when new memory had to be allocated, not just
7848 when reusing an existing buffer.
7851 *) New command line and configuration option 'utf8' for the req command.
7852 This allows field values to be specified as UTF8 strings.
7855 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7856 runs for the former and machine-readable output for the latter.
7859 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7860 of the e-mail address in the DN (i.e., it will go into a certificate
7861 extension only). The new configuration file option 'email_in_dn = no'
7862 has the same effect.
7863 [Massimiliano Pala madwolf@openca.org]
7865 *) Change all functions with names starting with des_ to be starting
7866 with DES_ instead. Add wrappers that are compatible with libdes,
7867 but are named _ossl_old_des_*. Finally, add macros that map the
7868 des_* symbols to the corresponding _ossl_old_des_* if libdes
7869 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7870 desired, the des_* symbols will be mapped to DES_*, with one
7873 Since we provide two compatibility mappings, the user needs to
7874 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7875 compatibility is desired. The default (i.e., when that macro
7876 isn't defined) is OpenSSL 0.9.6c compatibility.
7878 There are also macros that enable and disable the support of old
7879 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7880 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7881 are defined, the default will apply: to support the old des routines.
7883 In either case, one must include openssl/des.h to get the correct
7884 definitions. Do not try to just include openssl/des_old.h, that
7887 NOTE: This is a major break of an old API into a new one. Software
7888 authors are encouraged to switch to the DES_ style functions. Some
7889 time in the future, des_old.h and the libdes compatibility functions
7890 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7891 default), and then completely removed.
7894 *) Test for certificates which contain unsupported critical extensions.
7895 If such a certificate is found during a verify operation it is
7896 rejected by default: this behaviour can be overridden by either
7897 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7898 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7899 X509_supported_extension() has also been added which returns 1 if a
7900 particular extension is supported.
7903 *) Modify the behaviour of EVP cipher functions in similar way to digests
7904 to retain compatibility with existing code.
7907 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7908 compatibility with existing code. In particular the 'ctx' parameter does
7909 not have to be to be initialized before the call to EVP_DigestInit() and
7910 it is tidied up after a call to EVP_DigestFinal(). New function
7911 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7912 EVP_MD_CTX_copy() changed to not require the destination to be
7913 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7914 requires the destination to be valid.
7916 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7917 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7920 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7921 so that complete 'Handshake' protocol structures are kept in memory
7922 instead of overwriting 'msg_type' and 'length' with 'body' data.
7925 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7926 [Massimo Santin via Richard Levitte]
7928 *) Major restructuring to the underlying ENGINE code. This includes
7929 reduction of linker bloat, separation of pure "ENGINE" manipulation
7930 (initialisation, etc) from functionality dealing with implementations
7931 of specific crypto interfaces. This change also introduces integrated
7932 support for symmetric ciphers and digest implementations - so ENGINEs
7933 can now accelerate these by providing EVP_CIPHER and EVP_MD
7934 implementations of their own. This is detailed in crypto/engine/README
7935 as it couldn't be adequately described here. However, there are a few
7936 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7937 were changed in the original introduction of ENGINE code have now
7938 reverted back - the hooking from this code to ENGINE is now a good
7939 deal more passive and at run-time, operations deal directly with
7940 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7941 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7942 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7943 they were not being used by the framework as there is no concept of a
7944 BIGNUM_METHOD and they could not be generalised to the new
7945 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7946 ENGINE_cpy() has been removed as it cannot be consistently defined in
7950 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7953 *) Change mkdef.pl to sort symbols that get the same entry number,
7954 and make sure the automatically generated functions ERR_load_*
7955 become part of libeay.num as well.
7958 *) New function SSL_renegotiate_pending(). This returns true once
7959 renegotiation has been requested (either SSL_renegotiate() call
7960 or HelloRequest/ClientHello received from the peer) and becomes
7961 false once a handshake has been completed.
7962 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7963 sends a HelloRequest, but does not ensure that a handshake takes
7964 place. SSL_renegotiate_pending() is useful for checking if the
7965 client has followed the request.)
7968 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7969 By default, clients may request session resumption even during
7970 renegotiation (if session ID contexts permit); with this option,
7971 session resumption is possible only in the first handshake.
7973 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7974 more bits available for options that should not be part of
7975 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7978 *) Add some demos for certificate and certificate request creation.
7981 *) Make maximum certificate chain size accepted from the peer application
7982 settable (SSL*_get/set_max_cert_list()), as proposed by
7983 "Douglas E. Engert" <deengert@anl.gov>.
7986 *) Add support for shared libraries for Unixware-7
7987 (Boyd Lynn Gerber <gerberb@zenez.com>).
7990 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7991 be done prior to destruction. Use this to unload error strings from
7992 ENGINEs that load their own error strings. NB: This adds two new API
7993 functions to "get" and "set" this destroy handler in an ENGINE.
7996 *) Alter all existing ENGINE implementations (except "openssl" and
7997 "openbsd") to dynamically instantiate their own error strings. This
7998 makes them more flexible to be built both as statically-linked ENGINEs
7999 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
8000 Also, add stub code to each that makes building them as self-contained
8001 shared-libraries easier (see README.ENGINE).
8004 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
8005 implementations into applications that are completely implemented in
8006 self-contained shared-libraries. The "dynamic" ENGINE exposes control
8007 commands that can be used to configure what shared-library to load and
8008 to control aspects of the way it is handled. Also, made an update to
8009 the README.ENGINE file that brings its information up-to-date and
8010 provides some information and instructions on the "dynamic" ENGINE
8011 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
8014 *) Make it possible to unload ranges of ERR strings with a new
8015 "ERR_unload_strings" function.
8018 *) Add a copy() function to EVP_MD.
8021 *) Make EVP_MD routines take a context pointer instead of just the
8022 md_data void pointer.
8025 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
8026 that the digest can only process a single chunk of data
8027 (typically because it is provided by a piece of
8028 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
8029 is only going to provide a single chunk of data, and hence the
8030 framework needn't accumulate the data for oneshot drivers.
8033 *) As with "ERR", make it possible to replace the underlying "ex_data"
8034 functions. This change also alters the storage and management of global
8035 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
8036 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
8037 index counters. The API functions that use this state have been changed
8038 to take a "class_index" rather than pointers to the class's local STACK
8039 and counter, and there is now an API function to dynamically create new
8040 classes. This centralisation allows us to (a) plug a lot of the
8041 thread-safety problems that existed, and (b) makes it possible to clean
8042 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
8043 such data would previously have always leaked in application code and
8044 workarounds were in place to make the memory debugging turn a blind eye
8045 to it. Application code that doesn't use this new function will still
8046 leak as before, but their memory debugging output will announce it now
8047 rather than letting it slide.
8049 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
8050 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
8051 has a return value to indicate success or failure.
8054 *) Make it possible to replace the underlying "ERR" functions such that the
8055 global state (2 LHASH tables and 2 locks) is only used by the "default"
8056 implementation. This change also adds two functions to "get" and "set"
8057 the implementation prior to it being automatically set the first time
8058 any other ERR function takes place. Ie. an application can call "get",
8059 pass the return value to a module it has just loaded, and that module
8060 can call its own "set" function using that value. This means the
8061 module's "ERR" operations will use (and modify) the error state in the
8062 application and not in its own statically linked copy of OpenSSL code.
8065 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
8066 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
8067 the operation, and provides a more encapsulated way for external code
8068 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
8069 to use these functions rather than manually incrementing the counts.
8071 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
8074 *) Add EVP test program.
8077 *) Add symmetric cipher support to ENGINE. Expect the API to change!
8080 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
8081 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
8082 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
8083 These allow a CRL to be built without having to access X509_CRL fields
8084 directly. Modify 'ca' application to use new functions.
8087 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
8088 bug workarounds. Rollback attack detection is a security feature.
8089 The problem will only arise on OpenSSL servers when TLSv1 is not
8090 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
8091 Software authors not wanting to support TLSv1 will have special reasons
8092 for their choice and can explicitly enable this option.
8093 [Bodo Moeller, Lutz Jaenicke]
8095 *) Rationalise EVP so it can be extended: don't include a union of
8096 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
8097 (similar to those existing for EVP_CIPHER_CTX).
8102 EVP_MD_CTX_init(&md); /* new function call */
8103 EVP_DigestInit(&md, EVP_sha1());
8104 EVP_DigestUpdate(&md, in, len);
8105 EVP_DigestFinal(&md, out, NULL);
8106 EVP_MD_CTX_cleanup(&md); /* new function call */
8110 *) Make DES key schedule conform to the usual scheme, as well as
8111 correcting its structure. This means that calls to DES functions
8112 now have to pass a pointer to a des_key_schedule instead of a
8113 plain des_key_schedule (which was actually always a pointer
8116 des_key_schedule ks;
8118 des_set_key_checked(..., &ks);
8119 des_ncbc_encrypt(..., &ks, ...);
8121 (Note that a later change renames 'des_...' into 'DES_...'.)
8124 *) Initial reduction of linker bloat: the use of some functions, such as
8125 PEM causes large amounts of unused functions to be linked in due to
8126 poor organisation. For example pem_all.c contains every PEM function
8127 which has a knock on effect of linking in large amounts of (unused)
8128 ASN1 code. Grouping together similar functions and splitting unrelated
8129 functions prevents this.
8132 *) Cleanup of EVP macros.
8135 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
8136 correct _ecb suffix.
8139 *) Add initial OCSP responder support to ocsp application. The
8140 revocation information is handled using the text based index
8141 use by the ca application. The responder can either handle
8142 requests generated internally, supplied in files (for example
8143 via a CGI script) or using an internal minimal server.
8146 *) Add configuration choices to get zlib compression for TLS.
8149 *) Changes to Kerberos SSL for RFC 2712 compliance:
8150 1. Implemented real KerberosWrapper, instead of just using
8151 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
8152 2. Implemented optional authenticator field of KerberosWrapper.
8154 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
8155 and authenticator structs; see crypto/krb5/.
8157 Generalized Kerberos calls to support multiple Kerberos libraries.
8158 [Vern Staats <staatsvr@asc.hpc.mil>,
8159 Jeffrey Altman <jaltman@columbia.edu>
8160 via Richard Levitte]
8162 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
8163 already does with RSA. testdsa.h now has 'priv_key/pub_key'
8164 values for each of the key sizes rather than having just
8165 parameters (and 'speed' generating keys each time).
8168 *) Speed up EVP routines.
8171 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
8172 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
8173 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
8174 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
8176 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
8177 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
8178 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
8181 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
8183 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
8186 *) Added the OS2-EMX target.
8187 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
8189 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
8190 to support NCONF routines in extension code. New function CONF_set_nconf()
8191 to allow functions which take an NCONF to also handle the old LHASH
8192 structure: this means that the old CONF compatible routines can be
8193 retained (in particular wrt extensions) without having to duplicate the
8194 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
8197 *) Enhance the general user interface with mechanisms for inner control
8198 and with possibilities to have yes/no kind of prompts.
8201 *) Change all calls to low level digest routines in the library and
8202 applications to use EVP. Add missing calls to HMAC_cleanup() and
8203 don't assume HMAC_CTX can be copied using memcpy().
8204 [Verdon Walker <VWalker@novell.com>, Steve Henson]
8206 *) Add the possibility to control engines through control names but with
8207 arbitrary arguments instead of just a string.
8208 Change the key loaders to take a UI_METHOD instead of a callback
8209 function pointer. NOTE: this breaks binary compatibility with earlier
8210 versions of OpenSSL [engine].
8211 Adapt the nCipher code for these new conditions and add a card insertion
8215 *) Enhance the general user interface with mechanisms to better support
8216 dialog box interfaces, application-defined prompts, the possibility
8217 to use defaults (for example default passwords from somewhere else)
8218 and interrupts/cancellations.
8221 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
8222 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
8225 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
8226 tidy up some unnecessarily weird code in 'sk_new()').
8227 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
8229 *) Change the key loading routines for ENGINEs to use the same kind
8230 callback (pem_password_cb) as all other routines that need this
8234 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
8235 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
8236 than this minimum value is recommended.
8239 *) New random seeder for OpenVMS, using the system process statistics
8240 that are easily reachable.
8243 *) Windows apparently can't transparently handle global
8244 variables defined in DLLs. Initialisations such as:
8246 const ASN1_ITEM *it = &ASN1_INTEGER_it;
8248 won't compile. This is used by the any applications that need to
8249 declare their own ASN1 modules. This was fixed by adding the option
8250 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
8251 needed for static libraries under Win32.
8254 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
8255 setting of purpose and trust fields. New X509_STORE trust and
8256 purpose functions and tidy up setting in other SSL functions.
8259 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
8260 structure. These are inherited by X509_STORE_CTX when it is
8261 initialised. This allows various defaults to be set in the
8262 X509_STORE structure (such as flags for CRL checking and custom
8263 purpose or trust settings) for functions which only use X509_STORE_CTX
8264 internally such as S/MIME.
8266 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
8267 trust settings if they are not set in X509_STORE. This allows X509_STORE
8268 purposes and trust (in S/MIME for example) to override any set by default.
8270 Add command line options for CRL checking to smime, s_client and s_server
8274 *) Initial CRL based revocation checking. If the CRL checking flag(s)
8275 are set then the CRL is looked up in the X509_STORE structure and
8276 its validity and signature checked, then if the certificate is found
8277 in the CRL the verify fails with a revoked error.
8279 Various new CRL related callbacks added to X509_STORE_CTX structure.
8281 Command line options added to 'verify' application to support this.
8283 This needs some additional work, such as being able to handle multiple
8284 CRLs with different times, extension based lookup (rather than just
8285 by subject name) and ultimately more complete V2 CRL extension
8289 *) Add a general user interface API (crypto/ui/). This is designed
8290 to replace things like des_read_password and friends (backward
8291 compatibility functions using this new API are provided).
8292 The purpose is to remove prompting functions from the DES code
8293 section as well as provide for prompting through dialog boxes in
8294 a window system and the like.
8297 *) Add "ex_data" support to ENGINE so implementations can add state at a
8298 per-structure level rather than having to store it globally.
8301 *) Make it possible for ENGINE structures to be copied when retrieved by
8302 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
8303 This causes the "original" ENGINE structure to act like a template,
8304 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
8305 operational state can be localised to each ENGINE structure, despite the
8306 fact they all share the same "methods". New ENGINE structures returned in
8307 this case have no functional references and the return value is the single
8308 structural reference. This matches the single structural reference returned
8309 by ENGINE_by_id() normally, when it is incremented on the pre-existing
8313 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
8314 needs to match any other type at all we need to manually clear the
8318 *) Changes to the "openssl engine" utility to include;
8319 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
8320 about an ENGINE's available control commands.
8321 - executing control commands from command line arguments using the
8322 '-pre' and '-post' switches. '-post' is only used if '-t' is
8323 specified and the ENGINE is successfully initialised. The syntax for
8324 the individual commands are colon-separated, for example;
8325 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
8328 *) New dynamic control command support for ENGINEs. ENGINEs can now
8329 declare their own commands (numbers), names (strings), descriptions,
8330 and input types for run-time discovery by calling applications. A
8331 subset of these commands are implicitly classed as "executable"
8332 depending on their input type, and only these can be invoked through
8333 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
8334 can be based on user input, config files, etc). The distinction is
8335 that "executable" commands cannot return anything other than a boolean
8336 result and can only support numeric or string input, whereas some
8337 discoverable commands may only be for direct use through
8338 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
8339 pointers, or other custom uses. The "executable" commands are to
8340 support parameterisations of ENGINE behaviour that can be
8341 unambiguously defined by ENGINEs and used consistently across any
8342 OpenSSL-based application. Commands have been added to all the
8343 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
8344 control over shared-library paths without source code alterations.
8347 *) Changed all ENGINE implementations to dynamically allocate their
8348 ENGINEs rather than declaring them statically. Apart from this being
8349 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
8350 this also allows the implementations to compile without using the
8351 internal engine_int.h header.
8354 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8355 'const' value. Any code that should be able to modify a RAND_METHOD
8356 should already have non-const pointers to it (ie. they should only
8357 modify their own ones).
8360 *) Made a variety of little tweaks to the ENGINE code.
8361 - "atalla" and "ubsec" string definitions were moved from header files
8362 to C code. "nuron" string definitions were placed in variables
8363 rather than hard-coded - allowing parameterisation of these values
8364 later on via ctrl() commands.
8365 - Removed unused "#if 0"'d code.
8366 - Fixed engine list iteration code so it uses ENGINE_free() to release
8367 structural references.
8368 - Constified the RAND_METHOD element of ENGINE structures.
8369 - Constified various get/set functions as appropriate and added
8370 missing functions (including a catch-all ENGINE_cpy that duplicates
8371 all ENGINE values onto a new ENGINE except reference counts/state).
8372 - Removed NULL parameter checks in get/set functions. Setting a method
8373 or function to NULL is a way of cancelling out a previously set
8374 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8375 and doesn't justify the extra error symbols and code.
8376 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8377 flags from engine_int.h to engine.h.
8378 - Changed prototypes for ENGINE handler functions (init(), finish(),
8379 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8382 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8383 to the algorithm using long division. The binary algorithm can be
8384 used only if the modulus is odd. On 32-bit systems, it is faster
8385 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8386 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8387 up to 450 bits. In 64-bit environments, the binary algorithm
8388 appears to be advantageous for much longer moduli; here we use it
8389 for moduli up to 2048 bits.
8392 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8393 could not support the combine flag in choice fields.
8396 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8397 extensions from a certificate request to the certificate.
8400 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8401 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8402 file: this allows the display of the certificate about to be
8403 signed to be customised, to allow certain fields to be included
8404 or excluded and extension details. The old system didn't display
8405 multicharacter strings properly, omitted fields not in the policy
8406 and couldn't display additional details such as extensions.
8409 *) Function EC_POINTs_mul for multiple scalar multiplication
8410 of an arbitrary number of elliptic curve points
8411 \sum scalars[i]*points[i],
8412 optionally including the generator defined for the EC_GROUP:
8413 scalar*generator + \sum scalars[i]*points[i].
8415 EC_POINT_mul is a simple wrapper function for the typical case
8416 that the point list has just one item (besides the optional
8420 *) First EC_METHODs for curves over GF(p):
8422 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8423 operations and provides various method functions that can also
8424 operate with faster implementations of modular arithmetic.
8426 EC_GFp_mont_method() reuses most functions that are part of
8427 EC_GFp_simple_method, but uses Montgomery arithmetic.
8429 [Bodo Moeller; point addition and point doubling
8430 implementation directly derived from source code provided by
8431 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8433 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8434 crypto/ec/ec_lib.c):
8436 Curves are EC_GROUP objects (with an optional group generator)
8437 based on EC_METHODs that are built into the library.
8439 Points are EC_POINT objects based on EC_GROUP objects.
8441 Most of the framework would be able to handle curves over arbitrary
8442 finite fields, but as there are no obvious types for fields other
8443 than GF(p), some functions are limited to that for now.
8446 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8447 that the file contains a complete HTTP response.
8450 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8451 change the def and num file printf format specifier from "%-40sXXX"
8452 to "%-39s XXX". The latter will always guarantee a space after the
8453 field while the former will cause them to run together if the field
8454 is 40 of more characters long.
8457 *) Constify the cipher and digest 'method' functions and structures
8458 and modify related functions to take constant EVP_MD and EVP_CIPHER
8462 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8463 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8466 *) Modify EVP_Digest*() routines so they now return values. Although the
8467 internal software routines can never fail additional hardware versions
8471 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8473 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8474 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8477 ERR_R_NESTED_ASN1_ERROR
8479 ERR_R_MISSING_ASN1_EOS
8480 were 4 .. 9, conflicting with
8481 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8483 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8484 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8486 Add new error code 'ERR_R_INTERNAL_ERROR'.
8489 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8493 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8494 sets the subject name for a new request or supersedes the
8495 subject name in a given request. Formats that can be parsed are
8496 'CN=Some Name, OU=myOU, C=IT'
8498 'CN=Some Name/OU=myOU/C=IT'.
8500 Add options '-batch' and '-verbose' to 'openssl req'.
8501 [Massimiliano Pala <madwolf@hackmasters.net>]
8503 *) Introduce the possibility to access global variables through
8504 functions on platform were that's the best way to handle exporting
8505 global variables in shared libraries. To enable this functionality,
8506 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8507 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8508 is normally done by Configure or something similar).
8510 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8511 in the source file (foo.c) like this:
8513 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8514 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8516 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8517 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8519 OPENSSL_DECLARE_GLOBAL(int,foo);
8520 #define foo OPENSSL_GLOBAL_REF(foo)
8521 OPENSSL_DECLARE_GLOBAL(double,bar);
8522 #define bar OPENSSL_GLOBAL_REF(bar)
8524 The #defines are very important, and therefore so is including the
8525 header file everywhere where the defined globals are used.
8527 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8528 of ASN.1 items, but that structure is a bit different.
8530 The largest change is in util/mkdef.pl which has been enhanced with
8531 better and easier to understand logic to choose which symbols should
8532 go into the Windows .def files as well as a number of fixes and code
8533 cleanup (among others, algorithm keywords are now sorted
8534 lexicographically to avoid constant rewrites).
8537 *) In BN_div() keep a copy of the sign of 'num' before writing the
8538 result to 'rm' because if rm==num the value will be overwritten
8539 and produce the wrong result if 'num' is negative: this caused
8540 problems with BN_mod() and BN_nnmod().
8543 *) Function OCSP_request_verify(). This checks the signature on an
8544 OCSP request and verifies the signer certificate. The signer
8545 certificate is just checked for a generic purpose and OCSP request
8549 *) Add OCSP_check_validity() function to check the validity of OCSP
8550 responses. OCSP responses are prepared in real time and may only
8551 be a few seconds old. Simply checking that the current time lies
8552 between thisUpdate and nextUpdate max reject otherwise valid responses
8553 caused by either OCSP responder or client clock inaccuracy. Instead
8554 we allow thisUpdate and nextUpdate to fall within a certain period of
8555 the current time. The age of the response can also optionally be
8556 checked. Two new options -validity_period and -status_age added to
8560 *) If signature or public key algorithm is unrecognized print out its
8561 OID rather that just UNKNOWN.
8564 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8565 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8566 ID to be generated from the issuer certificate alone which can then be
8567 passed to OCSP_id_issuer_cmp().
8570 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8571 ASN1 modules to export functions returning ASN1_ITEM pointers
8572 instead of the ASN1_ITEM structures themselves. This adds several
8573 new macros which allow the underlying ASN1 function/structure to
8574 be accessed transparently. As a result code should not use ASN1_ITEM
8575 references directly (such as &X509_it) but instead use the relevant
8576 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8577 use of the new ASN1 code on platforms where exporting structures
8578 is problematical (for example in shared libraries) but exporting
8579 functions returning pointers to structures is not.
8582 *) Add support for overriding the generation of SSL/TLS session IDs.
8583 These callbacks can be registered either in an SSL_CTX or per SSL.
8584 The purpose of this is to allow applications to control, if they wish,
8585 the arbitrary values chosen for use as session IDs, particularly as it
8586 can be useful for session caching in multiple-server environments. A
8587 command-line switch for testing this (and any client code that wishes
8588 to use such a feature) has been added to "s_server".
8589 [Geoff Thorpe, Lutz Jaenicke]
8591 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8592 of the form '#if defined(...) || defined(...) || ...' and
8593 '#if !defined(...) && !defined(...) && ...'. This also avoids
8594 the growing number of special cases it was previously handling.
8597 *) Make all configuration macros available for application by making
8598 sure they are available in opensslconf.h, by giving them names starting
8599 with "OPENSSL_" to avoid conflicts with other packages and by making
8600 sure e_os2.h will cover all platform-specific cases together with
8602 Additionally, it is now possible to define configuration/platform-
8603 specific names (called "system identities"). In the C code, these
8604 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8605 macro with the name beginning with "OPENSSL_SYS_", which is determined
8606 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8610 *) New option -set_serial to 'req' and 'x509' this allows the serial
8611 number to use to be specified on the command line. Previously self
8612 signed certificates were hard coded with serial number 0 and the
8613 CA options of 'x509' had to use a serial number in a file which was
8617 *) New options to 'ca' utility to support V2 CRL entry extensions.
8618 Currently CRL reason, invalidity date and hold instruction are
8619 supported. Add new CRL extensions to V3 code and some new objects.
8622 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8623 disable standard block padding (aka PKCS#5 padding) in the EVP
8624 API, which was previously mandatory. This means that the data is
8625 not padded in any way and so the total length much be a multiple
8626 of the block size, otherwise an error occurs.
8629 *) Initial (incomplete) OCSP SSL support.
8632 *) New function OCSP_parse_url(). This splits up a URL into its host,
8633 port and path components: primarily to parse OCSP URLs. New -url
8634 option to ocsp utility.
8637 *) New nonce behavior. The return value of OCSP_check_nonce() now
8638 reflects the various checks performed. Applications can decide
8639 whether to tolerate certain situations such as an absent nonce
8640 in a response when one was present in a request: the ocsp application
8641 just prints out a warning. New function OCSP_add1_basic_nonce()
8642 this is to allow responders to include a nonce in a response even if
8643 the request is nonce-less.
8646 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8647 skipped when using openssl x509 multiple times on a single input file,
8648 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8651 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8652 set string type: to handle setting ASN1_TIME structures. Fix ca
8653 utility to correctly initialize revocation date of CRLs.
8656 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8657 the clients preferred ciphersuites and rather use its own preferences.
8658 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8659 Internet Explorer by ensuring unchanged hash method during stepup.
8660 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8663 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8664 to aes and add a new 'exist' option to print out symbols that don't
8668 *) Additional options to ocsp utility to allow flags to be set and
8669 additional certificates supplied.
8672 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8673 OCSP client a number of certificate to only verify the response
8677 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8678 handle the new API. Currently only ECB, CBC modes supported. Add new
8681 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8682 Encryption Standard (AES) Ciphersuites for Transport Layer
8683 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8684 not enabled by default and were not part of the "ALL" ciphersuite
8685 alias because they were not yet official; they could be
8686 explicitly requested by specifying the "AESdraft" ciphersuite
8687 group alias. In the final release of OpenSSL 0.9.7, the group
8688 alias is called "AES" and is part of "ALL".)
8689 [Ben Laurie, Steve Henson, Bodo Moeller]
8691 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8692 request to response.
8695 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8696 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8697 extract information from a certificate request. OCSP_response_create()
8698 creates a response and optionally adds a basic response structure.
8699 OCSP_basic_add1_status() adds a complete single response to a basic
8700 response and returns the OCSP_SINGLERESP structure just added (to allow
8701 extensions to be included for example). OCSP_basic_add1_cert() adds a
8702 certificate to a basic response and OCSP_basic_sign() signs a basic
8703 response with various flags. New helper functions ASN1_TIME_check()
8704 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8705 (converts ASN1_TIME to GeneralizedTime).
8708 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8709 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8710 structure from a certificate. X509_pubkey_digest() digests the public_key
8711 contents: this is used in various key identifiers.
8714 *) Make sk_sort() tolerate a NULL argument.
8715 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8717 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8718 passed by the function are trusted implicitly. If any of them signed the
8719 response then it is assumed to be valid and is not verified.
8722 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8723 to data. This was previously part of the PKCS7 ASN1 code. This
8724 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8725 [Steve Henson, reported by Kenneth R. Robinette
8726 <support@securenetterm.com>]
8728 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8729 routines: without these tracing memory leaks is very painful.
8730 Fix leaks in PKCS12 and PKCS7 routines.
8733 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8734 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8735 effectively meant GeneralizedTime would never be used. Now it
8736 is initialised to -1 but X509_time_adj() now has to check the value
8737 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8738 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8739 [Steve Henson, reported by Kenneth R. Robinette
8740 <support@securenetterm.com>]
8742 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8743 result in a zero length in the ASN1_INTEGER structure which was
8744 not consistent with the structure when d2i_ASN1_INTEGER() was used
8745 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8746 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8747 where it did not print out a minus for negative ASN1_INTEGER.
8750 *) Add summary printout to ocsp utility. The various functions which
8751 convert status values to strings have been renamed to:
8752 OCSP_response_status_str(), OCSP_cert_status_str() and
8753 OCSP_crl_reason_str() and are no longer static. New options
8754 to verify nonce values and to disable verification. OCSP response
8755 printout format cleaned up.
8758 *) Add additional OCSP certificate checks. These are those specified
8759 in RFC2560. This consists of two separate checks: the CA of the
8760 certificate being checked must either be the OCSP signer certificate
8761 or the issuer of the OCSP signer certificate. In the latter case the
8762 OCSP signer certificate must contain the OCSP signing extended key
8763 usage. This check is performed by attempting to match the OCSP
8764 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8765 in the OCSP_CERTID structures of the response.
8768 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8769 and related routines. This uses the standard OpenSSL certificate
8770 verify routines to perform initial checks (just CA validity) and
8771 to obtain the certificate chain. Then additional checks will be
8772 performed on the chain. Currently the root CA is checked to see
8773 if it is explicitly trusted for OCSP signing. This is used to set
8774 a root CA as a global signing root: that is any certificate that
8775 chains to that CA is an acceptable OCSP signing certificate.
8778 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8779 extensions from a separate configuration file.
8780 As when reading extensions from the main configuration file,
8781 the '-extensions ...' option may be used for specifying the
8783 [Massimiliano Pala <madwolf@comune.modena.it>]
8785 *) New OCSP utility. Allows OCSP requests to be generated or
8786 read. The request can be sent to a responder and the output
8787 parsed, outputted or printed in text form. Not complete yet:
8788 still needs to check the OCSP response validity.
8791 *) New subcommands for 'openssl ca':
8792 'openssl ca -status <serial>' prints the status of the cert with
8793 the given serial number (according to the index file).
8794 'openssl ca -updatedb' updates the expiry status of certificates
8796 [Massimiliano Pala <madwolf@comune.modena.it>]
8798 *) New '-newreq-nodes' command option to CA.pl. This is like
8799 '-newreq', but calls 'openssl req' with the '-nodes' option
8800 so that the resulting key is not encrypted.
8801 [Damien Miller <djm@mindrot.org>]
8803 *) New configuration for the GNU Hurd.
8804 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8806 *) Initial code to implement OCSP basic response verify. This
8807 is currently incomplete. Currently just finds the signer's
8808 certificate and verifies the signature on the response.
8811 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8812 value of OPENSSLDIR. This is available via the new '-d' option
8813 to 'openssl version', and is also included in 'openssl version -a'.
8816 *) Allowing defining memory allocation callbacks that will be given
8817 file name and line number information in additional arguments
8818 (a const char* and an int). The basic functionality remains, as
8819 well as the original possibility to just replace malloc(),
8820 realloc() and free() by functions that do not know about these
8821 additional arguments. To register and find out the current
8822 settings for extended allocation functions, the following
8823 functions are provided:
8825 CRYPTO_set_mem_ex_functions
8826 CRYPTO_set_locked_mem_ex_functions
8827 CRYPTO_get_mem_ex_functions
8828 CRYPTO_get_locked_mem_ex_functions
8830 These work the same way as CRYPTO_set_mem_functions and friends.
8831 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8832 extended allocation function is enabled.
8833 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8834 a conventional allocation function is enabled.
8835 [Richard Levitte, Bodo Moeller]
8837 *) Finish off removing the remaining LHASH function pointer casts.
8838 There should no longer be any prototype-casting required when using
8839 the LHASH abstraction, and any casts that remain are "bugs". See
8840 the callback types and macros at the head of lhash.h for details
8841 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8844 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8845 If /dev/[u]random devices are not available or do not return enough
8846 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8848 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8849 /etc/entropy will be queried once each in this sequence, querying stops
8850 when enough entropy was collected without querying more sockets.
8853 *) Change the Unix RAND_poll() variant to be able to poll several
8854 random devices, as specified by DEVRANDOM, until a sufficient amount
8855 of data has been collected. We spend at most 10 ms on each file
8856 (select timeout) and read in non-blocking mode. DEVRANDOM now
8857 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8858 (previously it was just the string "/dev/urandom"), so on typical
8859 platforms the 10 ms delay will never occur.
8860 Also separate out the Unix variant to its own file, rand_unix.c.
8861 For VMS, there's a currently-empty rand_vms.c.
8864 *) Move OCSP client related routines to ocsp_cl.c. These
8865 provide utility functions which an application needing
8866 to issue a request to an OCSP responder and analyse the
8867 response will typically need: as opposed to those which an
8868 OCSP responder itself would need which will be added later.
8870 OCSP_request_sign() signs an OCSP request with an API similar
8871 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8872 response. OCSP_response_get1_basic() extracts basic response
8873 from response. OCSP_resp_find_status(): finds and extracts status
8874 information from an OCSP_CERTID structure (which will be created
8875 when the request structure is built). These are built from lower
8876 level functions which work on OCSP_SINGLERESP structures but
8877 won't normally be used unless the application wishes to examine
8878 extensions in the OCSP response for example.
8880 Replace nonce routines with a pair of functions.
8881 OCSP_request_add1_nonce() adds a nonce value and optionally
8882 generates a random value. OCSP_check_nonce() checks the
8883 validity of the nonce in an OCSP response.
8886 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8887 This doesn't copy the supplied OCSP_CERTID and avoids the
8888 need to free up the newly created id. Change return type
8889 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8890 This can then be used to add extensions to the request.
8891 Deleted OCSP_request_new(), since most of its functionality
8892 is now in OCSP_REQUEST_new() (and the case insensitive name
8893 clash) apart from the ability to set the request name which
8894 will be added elsewhere.
8897 *) Update OCSP API. Remove obsolete extensions argument from
8898 various functions. Extensions are now handled using the new
8899 OCSP extension code. New simple OCSP HTTP function which
8900 can be used to send requests and parse the response.
8903 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8904 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8905 uses the special reorder version of SET OF to sort the attributes
8906 and reorder them to match the encoded order. This resolves a long
8907 standing problem: a verify on a PKCS7 structure just after signing
8908 it used to fail because the attribute order did not match the
8909 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8910 it uses the received order. This is necessary to tolerate some broken
8911 software that does not order SET OF. This is handled by encoding
8912 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8913 to produce the required SET OF.
8916 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8917 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8918 files to get correct declarations of the ASN.1 item variables.
8921 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8922 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8923 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8924 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8925 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8926 ASN1_ITEM and no wrapper functions.
8929 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8930 replace the old function pointer based I/O routines. Change most of
8931 the *_d2i_bio() and *_d2i_fp() functions to use these.
8934 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8935 lines, recognize more "algorithms" that can be deselected, and make
8936 it complain about algorithm deselection that isn't recognised.
8939 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8940 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8941 to use new functions. Add NO_ASN1_OLD which can be set to remove
8942 some old style ASN1 functions: this can be used to determine if old
8943 code will still work when these eventually go away.
8946 *) New extension functions for OCSP structures, these follow the
8947 same conventions as certificates and CRLs.
8950 *) New function X509V3_add1_i2d(). This automatically encodes and
8951 adds an extension. Its behaviour can be customised with various
8952 flags to append, replace or delete. Various wrappers added for
8953 certificates and CRLs.
8956 *) Fix to avoid calling the underlying ASN1 print routine when
8957 an extension cannot be parsed. Correct a typo in the
8958 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8961 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8962 entries for variables.
8965 *) Add functionality to apps/openssl.c for detecting locking
8966 problems: As the program is single-threaded, all we have
8967 to do is register a locking callback using an array for
8968 storing which locks are currently held by the program.
8971 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8972 SSL_get_ex_data_X509_STORE_idx(), which is used in
8973 ssl_verify_cert_chain() and thus can be called at any time
8974 during TLS/SSL handshakes so that thread-safety is essential.
8975 Unfortunately, the ex_data design is not at all suited
8976 for multi-threaded use, so it probably should be abolished.
8979 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8980 [Broadcom, tweaked and integrated by Geoff Thorpe]
8982 *) Move common extension printing code to new function
8983 X509V3_print_extensions(). Reorganise OCSP print routines and
8984 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8987 *) New function X509_signature_print() to remove duplication in some
8991 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8992 set (this was treated exactly the same as SET OF previously). This
8993 is used to reorder the STACK representing the structure to match the
8994 encoding. This will be used to get round a problem where a PKCS7
8995 structure which was signed could not be verified because the STACK
8996 order did not reflect the encoded order.
8999 *) Reimplement the OCSP ASN1 module using the new code.
9002 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
9003 for its ASN1 operations. The old style function pointers still exist
9004 for now but they will eventually go away.
9007 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
9008 completely replaces the old ASN1 functionality with a table driven
9009 encoder and decoder which interprets an ASN1_ITEM structure describing
9010 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
9011 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
9012 has also been converted to the new form.
9015 *) Change BN_mod_exp_recp so that negative moduli are tolerated
9016 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
9017 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
9018 for negative moduli.
9021 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
9022 of not touching the result's sign bit.
9025 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
9029 *) Changed the LHASH code to use prototypes for callbacks, and created
9030 macros to declare and implement thin (optionally static) functions
9031 that provide type-safety and avoid function pointer casting for the
9032 type-specific callbacks.
9035 *) Added Kerberos Cipher Suites to be used with TLS, as written in
9037 [Veers Staats <staatsvr@asc.hpc.mil>,
9038 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
9040 *) Reformat the FAQ so the different questions and answers can be divided
9041 in sections depending on the subject.
9044 *) Have the zlib compression code load ZLIB.DLL dynamically under
9048 *) New function BN_mod_sqrt for computing square roots modulo a prime
9049 (using the probabilistic Tonelli-Shanks algorithm unless
9050 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
9051 be handled deterministically).
9052 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9054 *) Make BN_mod_inverse faster by explicitly handling small quotients
9055 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
9056 512 bits], about 30% for larger ones [1024 or 2048 bits].)
9059 *) New function BN_kronecker.
9062 *) Fix BN_gcd so that it works on negative inputs; the result is
9063 positive unless both parameters are zero.
9064 Previously something reasonably close to an infinite loop was
9065 possible because numbers could be growing instead of shrinking
9066 in the implementation of Euclid's algorithm.
9069 *) Fix BN_is_word() and BN_is_one() macros to take into account the
9070 sign of the number in question.
9072 Fix BN_is_word(a,w) to work correctly for w == 0.
9074 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
9075 because its test if the absolute value of 'a' equals 'w'.
9076 Note that BN_abs_is_word does *not* handle w == 0 reliably;
9077 it exists mostly for use in the implementations of BN_is_zero(),
9078 BN_is_one(), and BN_is_word().
9081 *) New function BN_swap.
9084 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
9085 the exponentiation functions are more likely to produce reasonable
9086 results on negative inputs.
9089 *) Change BN_mod_mul so that the result is always non-negative.
9090 Previously, it could be negative if one of the factors was negative;
9091 I don't think anyone really wanted that behaviour.
9094 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
9095 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
9096 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
9097 and add new functions:
9106 BN_mod_lshift1_quick
9110 These functions always generate non-negative results.
9112 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
9113 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
9115 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
9116 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
9117 be reduced modulo m.
9118 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9121 The following entry accidentally appeared in the CHANGES file
9122 distributed with OpenSSL 0.9.7. The modifications described in
9123 it do *not* apply to OpenSSL 0.9.7.
9125 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
9126 was actually never needed) and in BN_mul(). The removal in BN_mul()
9127 required a small change in bn_mul_part_recursive() and the addition
9128 of the functions bn_cmp_part_words(), bn_sub_part_words() and
9129 bn_add_part_words(), which do the same thing as bn_cmp_words(),
9130 bn_sub_words() and bn_add_words() except they take arrays with
9135 *) In 'openssl passwd', verify passwords read from the terminal
9136 unless the '-salt' option is used (which usually means that
9137 verification would just waste user's time since the resulting
9138 hash is going to be compared with some given password hash)
9139 or the new '-noverify' option is used.
9141 This is an incompatible change, but it does not affect
9142 non-interactive use of 'openssl passwd' (passwords on the command
9143 line, '-stdin' option, '-in ...' option) and thus should not
9147 *) Remove all references to RSAref, since there's no more need for it.
9150 *) Make DSO load along a path given through an environment variable
9151 (SHLIB_PATH) with shl_load().
9154 *) Constify the ENGINE code as a result of BIGNUM constification.
9155 Also constify the RSA code and most things related to it. In a
9156 few places, most notable in the depth of the ASN.1 code, ugly
9157 casts back to non-const were required (to be solved at a later
9161 *) Make it so the openssl application has all engines loaded by default.
9164 *) Constify the BIGNUM routines a little more.
9167 *) Add the following functions:
9169 ENGINE_load_cswift()
9171 ENGINE_load_atalla()
9173 ENGINE_load_builtin_engines()
9175 That way, an application can itself choose if external engines that
9176 are built-in in OpenSSL shall ever be used or not. The benefit is
9177 that applications won't have to be linked with libdl or other dso
9178 libraries unless it's really needed.
9180 Changed 'openssl engine' to load all engines on demand.
9181 Changed the engine header files to avoid the duplication of some
9182 declarations (they differed!).
9185 *) 'openssl engine' can now list capabilities.
9188 *) Better error reporting in 'openssl engine'.
9191 *) Never call load_dh_param(NULL) in s_server.
9194 *) Add engine application. It can currently list engines by name and
9195 identity, and test if they are actually available.
9198 *) Improve RPM specification file by forcing symbolic linking and making
9199 sure the installed documentation is also owned by root.root.
9200 [Damien Miller <djm@mindrot.org>]
9202 *) Give the OpenSSL applications more possibilities to make use of
9203 keys (public as well as private) handled by engines.
9206 *) Add OCSP code that comes from CertCo.
9209 *) Add VMS support for the Rijndael code.
9212 *) Added untested support for Nuron crypto accelerator.
9215 *) Add support for external cryptographic devices. This code was
9216 previously distributed separately as the "engine" branch.
9217 [Geoff Thorpe, Richard Levitte]
9219 *) Rework the filename-translation in the DSO code. It is now possible to
9220 have far greater control over how a "name" is turned into a filename
9221 depending on the operating environment and any oddities about the
9222 different shared library filenames on each system.
9225 *) Support threads on FreeBSD-elf in Configure.
9228 *) Fix for SHA1 assembly problem with MASM: it produces
9229 warnings about corrupt line number information when assembling
9230 with debugging information. This is caused by the overlapping
9232 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
9235 NCONF_get_number() has no error checking at all. As a replacement,
9236 NCONF_get_number_e() is defined (_e for "error checking") and is
9237 promoted strongly. The old NCONF_get_number is kept around for
9238 binary backward compatibility.
9239 Make it possible for methods to load from something other than a BIO,
9240 by providing a function pointer that is given a name instead of a BIO.
9241 For example, this could be used to load configuration data from an
9245 *) Fix for non blocking accept BIOs. Added new I/O special reason
9246 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
9247 with non blocking I/O was not possible because no retry code was
9248 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
9252 *) Added the beginnings of Rijndael support.
9255 *) Fix for bug in DirectoryString mask setting. Add support for
9256 X509_NAME_print_ex() in 'req' and X509_print_ex() function
9257 to allow certificate printing to more controllable, additional
9258 'certopt' option to 'x509' to allow new printing options to be
9262 *) Clean old EAY MD5 hack from e_os.h.
9265 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
9267 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
9268 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
9269 [Joe Orton, Steve Henson]
9271 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
9273 *) Fix additional bug revealed by the NISCC test suite:
9275 Stop bug triggering large recursion when presented with
9276 certain ASN.1 tags (CVE-2003-0851)
9279 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
9281 *) Fix various bugs revealed by running the NISCC test suite:
9283 Stop out of bounds reads in the ASN1 code when presented with
9284 invalid tags (CVE-2003-0543 and CVE-2003-0544).
9286 If verify callback ignores invalid public key errors don't try to check
9287 certificate signature with the NULL public key.
9291 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
9292 if the server requested one: as stated in TLS 1.0 and SSL 3.0
9296 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
9297 extra data after the compression methods not only for TLS 1.0
9298 but also for SSL 3.0 (as required by the specification).
9299 [Bodo Moeller; problem pointed out by Matthias Loepfe]
9301 *) Change X509_certificate_type() to mark the key as exported/exportable
9302 when it's 512 *bits* long, not 512 bytes.
9305 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
9307 *) Countermeasure against the Klima-Pokorny-Rosa extension of
9308 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
9309 a protocol version number mismatch like a decryption error
9310 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
9313 *) Turn on RSA blinding by default in the default implementation
9314 to avoid a timing attack. Applications that don't want it can call
9315 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
9316 They would be ill-advised to do so in most cases.
9317 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
9319 *) Change RSA blinding code so that it works when the PRNG is not
9320 seeded (in this case, the secret RSA exponent is abused as
9321 an unpredictable seed -- if it is not unpredictable, there
9322 is no point in blinding anyway). Make RSA blinding thread-safe
9323 by remembering the creator's thread ID in rsa->blinding and
9324 having all other threads use local one-time blinding factors
9325 (this requires more computation than sharing rsa->blinding, but
9326 avoids excessive locking; and if an RSA object is not shared
9327 between threads, blinding will still be very fast).
9330 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
9332 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
9333 via timing by performing a MAC computation even if incorrect
9334 block cipher padding has been found. This is a countermeasure
9335 against active attacks where the attacker has to distinguish
9336 between bad padding and a MAC verification error. (CVE-2003-0078)
9338 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
9339 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
9340 Martin Vuagnoux (EPFL, Ilion)]
9342 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
9344 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
9345 memory from its contents. This is done with a counter that will
9346 place alternating values in each byte. This can be used to solve
9347 two issues: 1) the removal of calls to memset() by highly optimizing
9348 compilers, and 2) cleansing with other values than 0, since those can
9349 be read through on certain media, for example a swap space on disk.
9352 *) Bugfix: client side session caching did not work with external caching,
9353 because the session->cipher setting was not restored when reloading
9354 from the external cache. This problem was masked, when
9355 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9356 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9359 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9360 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9361 [Zeev Lieber <zeev-l@yahoo.com>]
9363 *) Undo an undocumented change introduced in 0.9.6e which caused
9364 repeated calls to OpenSSL_add_all_ciphers() and
9365 OpenSSL_add_all_digests() to be ignored, even after calling
9369 *) Change the default configuration reader to deal with last line not
9370 being properly terminated.
9373 *) Change X509_NAME_cmp() so it applies the special rules on handling
9374 DN values that are of type PrintableString, as well as RDNs of type
9375 emailAddress where the value has the type ia5String.
9376 [stefank@valicert.com via Richard Levitte]
9378 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9379 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9380 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9381 the bitwise-OR of the two for use by the majority of applications
9382 wanting this behaviour, and update the docs. The documented
9383 behaviour and actual behaviour were inconsistent and had been
9384 changing anyway, so this is more a bug-fix than a behavioural
9386 [Geoff Thorpe, diagnosed by Nadav Har'El]
9388 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9389 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9392 *) Fix initialization code race conditions in
9393 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9394 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9395 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9396 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9397 ssl2_get_cipher_by_char(),
9398 ssl3_get_cipher_by_char().
9399 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9401 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9402 the cached sessions are flushed, as the remove_cb() might use ex_data
9403 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9404 (see [openssl.org #212]).
9405 [Geoff Thorpe, Lutz Jaenicke]
9407 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9408 length, instead of the encoding length to d2i_ASN1_OBJECT.
9411 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9413 *) [In 0.9.6g-engine release:]
9414 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9415 [Lynn Gazis <lgazis@rainbow.com>]
9417 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9419 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9420 and get fix the header length calculation.
9421 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9422 Alon Kantor <alonk@checkpoint.com> (and others),
9425 *) Use proper error handling instead of 'assertions' in buffer
9426 overflow checks added in 0.9.6e. This prevents DoS (the
9427 assertions could call abort()).
9428 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9430 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9432 *) Add various sanity checks to asn1_get_length() to reject
9433 the ASN1 length bytes if they exceed sizeof(long), will appear
9434 negative or the content length exceeds the length of the
9436 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9438 *) Fix cipher selection routines: ciphers without encryption had no flags
9439 for the cipher strength set and where therefore not handled correctly
9440 by the selection routines (PR #130).
9443 *) Fix EVP_dsa_sha macro.
9447 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9448 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9449 that was added in OpenSSL 0.9.6d.
9451 As the countermeasure turned out to be incompatible with some
9452 broken SSL implementations, the new option is part of SSL_OP_ALL.
9453 SSL_OP_ALL is usually employed when compatibility with weird SSL
9454 implementations is desired (e.g. '-bugs' option to 's_client' and
9455 's_server'), so the new option is automatically set in many
9459 *) Changes in security patch:
9461 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9462 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9463 Air Force Materiel Command, USAF, under agreement number
9466 *) Add various sanity checks to asn1_get_length() to reject
9467 the ASN1 length bytes if they exceed sizeof(long), will appear
9468 negative or the content length exceeds the length of the
9469 supplied buffer. (CVE-2002-0659)
9470 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9472 *) Assertions for various potential buffer overflows, not known to
9474 [Ben Laurie (CHATS)]
9476 *) Various temporary buffers to hold ASCII versions of integers were
9477 too small for 64 bit platforms. (CVE-2002-0655)
9478 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9480 *) Remote buffer overflow in SSL3 protocol - an attacker could
9481 supply an oversized session ID to a client. (CVE-2002-0656)
9482 [Ben Laurie (CHATS)]
9484 *) Remote buffer overflow in SSL2 protocol - an attacker could
9485 supply an oversized client master key. (CVE-2002-0656)
9486 [Ben Laurie (CHATS)]
9488 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9490 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9491 encoded as NULL) with id-dsa-with-sha1.
9492 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9494 *) Check various X509_...() return values in apps/req.c.
9495 [Nils Larsch <nla@trustcenter.de>]
9497 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9498 an end-of-file condition would erroneously be flagged, when the CRLF
9499 was just at the end of a processed block. The bug was discovered when
9500 processing data through a buffering memory BIO handing the data to a
9501 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9502 <ptsekov@syntrex.com> and Nedelcho Stanev.
9505 *) Implement a countermeasure against a vulnerability recently found
9506 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9507 before application data chunks to avoid the use of known IVs
9508 with data potentially chosen by the attacker.
9511 *) Fix length checks in ssl3_get_client_hello().
9514 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9515 to prevent ssl3_read_internal() from incorrectly assuming that
9516 ssl3_read_bytes() found application data while handshake
9517 processing was enabled when in fact s->s3->in_read_app_data was
9518 merely automatically cleared during the initial handshake.
9519 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9521 *) Fix object definitions for Private and Enterprise: they were not
9522 recognized in their shortname (=lowercase) representation. Extend
9523 obj_dat.pl to issue an error when using undefined keywords instead
9524 of silently ignoring the problem (Svenning Sorensen
9525 <sss@sss.dnsalias.net>).
9528 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9529 generators, i.e. generators other than 2 and 5. (Previously, the
9530 code did not properly initialise the 'add' and 'rem' values to
9531 BN_generate_prime().)
9533 In the new general case, we do not insist that 'generator' is
9534 actually a primitive root: This requirement is rather pointless;
9535 a generator of the order-q subgroup is just as good, if not
9539 *) Map new X509 verification errors to alerts. Discovered and submitted by
9540 Tom Wu <tom@arcot.com>.
9543 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9544 returning non-zero before the data has been completely received
9545 when using non-blocking I/O.
9546 [Bodo Moeller; problem pointed out by John Hughes]
9548 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9549 [Ben Laurie, Lutz Jaenicke]
9551 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9552 Yoram Zahavi <YoramZ@gilian.com>).
9555 *) Add information about CygWin 1.3 and on, and preserve proper
9556 configuration for the versions before that.
9557 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9559 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9560 check whether we deal with a copy of a session and do not delete from
9561 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9562 <izhar@checkpoint.com>.
9565 *) Do not store session data into the internal session cache, if it
9566 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9567 flag is set). Proposed by Aslam <aslam@funk.com>.
9570 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9574 *) [In 0.9.6d-engine release:]
9575 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9576 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9578 *) Add the configuration target linux-s390x.
9579 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9581 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9582 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9583 variable as an indication that a ClientHello message has been
9584 received. As the flag value will be lost between multiple
9585 invocations of ssl3_accept when using non-blocking I/O, the
9586 function may not be aware that a handshake has actually taken
9587 place, thus preventing a new session from being added to the
9590 To avoid this problem, we now set s->new_session to 2 instead of
9591 using a local variable.
9592 [Lutz Jaenicke, Bodo Moeller]
9594 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9595 if the SSL_R_LENGTH_MISMATCH error is detected.
9596 [Geoff Thorpe, Bodo Moeller]
9598 *) New 'shared_ldflag' column in Configure platform table.
9601 *) Fix EVP_CIPHER_mode macro.
9602 ["Dan S. Camper" <dan@bti.net>]
9604 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9605 type, we must throw them away by setting rr->length to 0.
9606 [D P Chang <dpc@qualys.com>]
9608 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9610 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9611 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9612 worked incorrectly for those cases where range = 10..._2 and
9613 3*range is two bits longer than range.)
9616 *) Only add signing time to PKCS7 structures if it is not already
9620 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9621 OBJ_ld_ce should be OBJ_id_ce.
9622 Also some ip-pda OIDs in crypto/objects/objects.txt were
9623 incorrect (cf. RFC 3039).
9624 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9626 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9627 returns early because it has nothing to do.
9628 [Andy Schneider <andy.schneider@bjss.co.uk>]
9630 *) [In 0.9.6c-engine release:]
9631 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9632 [Andy Schneider <andy.schneider@bjss.co.uk>]
9634 *) [In 0.9.6c-engine release:]
9635 Add support for Cryptographic Appliance's keyserver technology.
9636 (Use engine 'keyclient')
9637 [Cryptographic Appliances and Geoff Thorpe]
9639 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9640 is called via tools/c89.sh because arguments have to be
9641 rearranged (all '-L' options must appear before the first object
9643 [Richard Shapiro <rshapiro@abinitio.com>]
9645 *) [In 0.9.6c-engine release:]
9646 Add support for Broadcom crypto accelerator cards, backported
9648 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9650 *) [In 0.9.6c-engine release:]
9651 Add support for SureWare crypto accelerator cards from
9652 Baltimore Technologies. (Use engine 'sureware')
9653 [Baltimore Technologies and Mark Cox]
9655 *) [In 0.9.6c-engine release:]
9656 Add support for crypto accelerator cards from Accelerated
9657 Encryption Processing, www.aep.ie. (Use engine 'aep')
9658 [AEP Inc. and Mark Cox]
9660 *) Add a configuration entry for gcc on UnixWare.
9661 [Gary Benson <gbenson@redhat.com>]
9663 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9664 messages are stored in a single piece (fixed-length part and
9665 variable-length part combined) and fix various bugs found on the way.
9668 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9669 instead. BIO_gethostbyname() does not know what timeouts are
9670 appropriate, so entries would stay in cache even when they have
9672 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9674 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9675 faced with a pathologically small ClientHello fragment that does
9676 not contain client_version: Instead of aborting with an error,
9677 simply choose the highest available protocol version (i.e.,
9678 TLS 1.0 unless it is disabled). In practice, ClientHello
9679 messages are never sent like this, but this change gives us
9680 strictly correct behaviour at least for TLS.
9683 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9684 never resets s->method to s->ctx->method when called from within
9685 one of the SSL handshake functions.
9686 [Bodo Moeller; problem pointed out by Niko Baric]
9688 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9689 (sent using the client's version number) if client_version is
9690 smaller than the protocol version in use. Also change
9691 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9692 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9693 the client will at least see that alert.
9696 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9700 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9701 client receives HelloRequest while in a handshake.
9702 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9704 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9705 should end in 'break', not 'goto end' which circumvents various
9706 cleanups done in state SSL_ST_OK. But session related stuff
9707 must be disabled for SSL_ST_OK in the case that we just sent a
9710 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9711 before just sending a HelloRequest.
9712 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9714 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9715 reveal whether illegal block cipher padding was found or a MAC
9716 verification error occurred. (Neither SSLerr() codes nor alerts
9717 are directly visible to potential attackers, but the information
9718 may leak via logfiles.)
9720 Similar changes are not required for the SSL 2.0 implementation
9721 because the number of padding bytes is sent in clear for SSL 2.0,
9722 and the extra bytes are just ignored. However ssl/s2_pkt.c
9723 failed to verify that the purported number of padding bytes is in
9727 *) Add OpenUNIX-8 support including shared libraries
9728 (Boyd Lynn Gerber <gerberb@zenez.com>).
9731 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9732 'wristwatch attack' using huge encoding parameters (cf.
9733 James H. Manger's CRYPTO 2001 paper). Note that the
9734 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9735 encoding parameters and hence was not vulnerable.
9738 *) BN_sqr() bug fix.
9739 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9741 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9742 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9743 followed by modular reduction.
9744 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9746 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9747 equivalent based on BN_pseudo_rand() instead of BN_rand().
9750 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9751 This function was broken, as the check for a new client hello message
9752 to handle SGC did not allow these large messages.
9753 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9756 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9759 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9760 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9763 *) Rework the configuration and shared library support for Tru64 Unix.
9764 The configuration part makes use of modern compiler features and
9765 still retains old compiler behavior for those that run older versions
9766 of the OS. The shared library support part includes a variant that
9767 uses the RPATH feature, and is available through the special
9768 configuration target "alpha-cc-rpath", which will never be selected
9770 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9772 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9773 with the same message size as in ssl3_get_certificate_request().
9774 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9775 messages might inadvertently be reject as too long.
9776 [Petr Lampa <lampa@fee.vutbr.cz>]
9778 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9781 *) Modified SSL library such that the verify_callback that has been set
9782 specifically for an SSL object with SSL_set_verify() is actually being
9783 used. Before the change, a verify_callback set with this function was
9784 ignored and the verify_callback() set in the SSL_CTX at the time of
9785 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9786 to allow the necessary settings.
9789 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9790 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9791 done automatically (in contradiction to the requirements of the C
9792 standard). This made problems when used from OpenSSH.
9795 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9796 dh->length and always used
9798 BN_rand_range(priv_key, dh->p).
9800 BN_rand_range() is not necessary for Diffie-Hellman, and this
9801 specific range makes Diffie-Hellman unnecessarily inefficient if
9802 dh->length (recommended exponent length) is much smaller than the
9803 length of dh->p. We could use BN_rand_range() if the order of
9804 the subgroup was stored in the DH structure, but we only have
9809 BN_rand(priv_key, l, ...)
9811 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9817 RSA_eay_public_encrypt
9818 RSA_eay_private_decrypt
9819 RSA_eay_private_encrypt (signing)
9820 RSA_eay_public_decrypt (signature verification)
9822 (default implementations for RSA_public_encrypt,
9823 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9824 always reject numbers >= n.
9827 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9828 to synchronize access to 'locking_thread'. This is necessary on
9829 systems where access to 'locking_thread' (an 'unsigned long'
9830 variable) is not atomic.
9833 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9834 *before* setting the 'crypto_lock_rand' flag. The previous code had
9835 a race condition if 0 is a valid thread ID.
9836 [Travis Vitek <vitek@roguewave.com>]
9838 *) Add support for shared libraries under Irix.
9839 [Albert Chin-A-Young <china@thewrittenword.com>]
9841 *) Add configuration option to build on Linux on both big-endian and
9843 [Ralf Baechle <ralf@uni-koblenz.de>]
9845 *) Add the possibility to create shared libraries on HP-UX.
9848 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9850 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9851 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9852 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9853 PRNG state recovery was possible based on the output of
9854 one PRNG request appropriately sized to gain knowledge on
9855 'md' followed by enough consecutive 1-byte PRNG requests
9856 to traverse all of 'state'.
9858 1. When updating 'md_local' (the current thread's copy of 'md')
9859 during PRNG output generation, hash all of the previous
9860 'md_local' value, not just the half used for PRNG output.
9862 2. Make the number of bytes from 'state' included into the hash
9863 independent from the number of PRNG bytes requested.
9865 The first measure alone would be sufficient to avoid
9866 Markku-Juhani's attack. (Actually it had never occurred
9867 to me that the half of 'md_local' used for chaining was the
9868 half from which PRNG output bytes were taken -- I had always
9869 assumed that the secret half would be used.) The second
9870 measure makes sure that additional data from 'state' is never
9871 mixed into 'md_local' in small portions; this heuristically
9872 further strengthens the PRNG.
9875 *) Fix crypto/bn/asm/mips3.s.
9878 *) When only the key is given to "enc", the IV is undefined. Print out
9879 an error message in this case.
9882 *) Handle special case when X509_NAME is empty in X509 printing routines.
9885 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9886 positive and less than q.
9889 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9890 used: it isn't thread safe and the add_lock_callback should handle
9892 [Paul Rose <Paul.Rose@bridge.com>]
9894 *) Verify that incoming data obeys the block size in
9895 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9899 [Ulf Möller, Bodo Möller]
9901 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9902 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9903 when fixing the server behaviour for backwards-compatible 'client
9904 hello' messages. (Note that the attack is impractical against
9905 SSL 3.0 and TLS 1.0 anyway because length and version checking
9906 means that the probability of guessing a valid ciphertext is
9907 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9910 Before 0.9.5, the countermeasure (hide the error by generating a
9911 random 'decryption result') did not work properly because
9912 ERR_clear_error() was missing, meaning that SSL_get_error() would
9913 detect the supposedly ignored error.
9915 Both problems are now fixed.
9918 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9919 (previously it was 1024).
9922 *) Fix for compatibility mode trust settings: ignore trust settings
9923 unless some valid trust or reject settings are present.
9926 *) Fix for blowfish EVP: its a variable length cipher.
9929 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9930 parameters in DSA public key structures and return an error in the
9931 DSA routines if parameters are absent.
9934 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9935 in the current directory if neither $RANDFILE nor $HOME was set.
9936 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9937 caused some confusion to Windows users who haven't defined $HOME.
9938 Thus RAND_file_name() is changed again: e_os.h can define a
9939 DEFAULT_HOME, which will be used if $HOME is not set.
9940 For Windows, we use "C:"; on other platforms, we still require
9941 environment variables.
9943 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9944 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9945 having multiple threads call RAND_poll() concurrently.
9948 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9949 combination of a flag and a thread ID variable.
9950 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9951 flag), *other* threads can enter ssleay_add_bytes without obeying
9952 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9953 that they do not hold after the first thread unsets add_do_not_lock).
9956 *) Change bctest again: '-x' expressions are not available in all
9960 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9962 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9963 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9965 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9966 the default extension for executables, if any. Also, make the perl
9967 scripts that use symlink() to test if it really exists and use "cp"
9968 if it doesn't. All this made OpenSSL compilable and installable in
9972 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9973 If SEQUENCE is length is indefinite just set c->slen to the total
9974 amount of data available.
9975 [Steve Henson, reported by shige@FreeBSD.org]
9976 [This change does not apply to 0.9.7.]
9978 *) Change bctest to avoid here-documents inside command substitution
9979 (workaround for FreeBSD /bin/sh bug).
9980 For compatibility with Ultrix, avoid shell functions (introduced
9981 in the bctest version that searches along $PATH).
9984 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9985 with des_encrypt() defined on some operating systems, like Solaris
9989 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9990 On the Importance of Eliminating Errors in Cryptographic
9991 Computations, J. Cryptology 14 (2001) 2, 101-119,
9992 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9995 *) MIPS assembler BIGNUM division bug fix.
9998 *) Disabled incorrect Alpha assembler code.
10001 *) Fix PKCS#7 decode routines so they correctly update the length
10002 after reading an EOC for the EXPLICIT tag.
10004 [This change does not apply to 0.9.7.]
10006 *) Fix bug in PKCS#12 key generation routines. This was triggered
10007 if a 3DES key was generated with a 0 initial byte. Include
10008 PKCS12_BROKEN_KEYGEN compilation option to retain the old
10009 (but broken) behaviour.
10012 *) Enhance bctest to search for a working bc along $PATH and print
10014 [Tim Rice <tim@multitalents.net> via Richard Levitte]
10016 *) Fix memory leaks in err.c: free err_data string if necessary;
10017 don't write to the wrong index in ERR_set_error_data.
10020 *) Implement ssl23_peek (analogous to ssl23_read), which previously
10024 *) Replace rdtsc with _emit statements for VC++ version 5.
10025 [Jeremy Cooper <jeremy@baymoo.org>]
10027 *) Make it possible to reuse SSLv2 sessions.
10030 *) In copy_email() check for >= 0 as a return value for
10031 X509_NAME_get_index_by_NID() since 0 is a valid index.
10032 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
10034 *) Avoid coredump with unsupported or invalid public keys by checking if
10035 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
10036 PKCS7_verify() fails with non detached data.
10039 *) Don't use getenv in library functions when run as setuid/setgid.
10040 New function OPENSSL_issetugid().
10043 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
10044 due to incorrect handling of multi-threading:
10046 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
10048 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
10050 3. Count how many times MemCheck_off() has been called so that
10051 nested use can be treated correctly. This also avoids
10052 inband-signalling in the previous code (which relied on the
10053 assumption that thread ID 0 is impossible).
10056 *) Add "-rand" option also to s_client and s_server.
10059 *) Fix CPU detection on Irix 6.x.
10060 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
10061 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10063 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
10066 [This change does not apply to 0.9.7.]
10068 *) Use the cached encoding of an X509_NAME structure rather than
10069 copying it. This is apparently the reason for the libsafe "errors"
10070 but the code is actually correct.
10073 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
10074 Bleichenbacher's DSA attack.
10075 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
10076 to be set and top=0 forces the highest bit to be set; top=-1 is new
10077 and leaves the highest bit random.
10078 [Ulf Moeller, Bodo Moeller]
10080 *) In the NCONF_...-based implementations for CONF_... queries
10081 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
10082 a temporary CONF structure with the data component set to NULL
10083 (which gives segmentation faults in lh_retrieve).
10084 Instead, use NULL for the CONF pointer in CONF_get_string and
10085 CONF_get_number (which may use environment variables) and directly
10086 return NULL from CONF_get_section.
10089 *) Fix potential buffer overrun for EBCDIC.
10092 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
10093 keyUsage if basicConstraints absent for a CA.
10096 *) Make SMIME_write_PKCS7() write mail header values with a format that
10097 is more generally accepted (no spaces before the semicolon), since
10098 some programs can't parse those values properly otherwise. Also make
10099 sure BIO's that break lines after each write do not create invalid
10103 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
10104 macros previously used would not encode an empty SEQUENCE OF
10105 and break the signature.
10107 [This change does not apply to 0.9.7.]
10109 *) Zero the premaster secret after deriving the master secret in
10113 *) Add some EVP_add_digest_alias registrations (as found in
10114 OpenSSL_add_all_digests()) to SSL_library_init()
10115 aka OpenSSL_add_ssl_algorithms(). This provides improved
10116 compatibility with peers using X.509 certificates
10117 with unconventional AlgorithmIdentifier OIDs.
10120 *) Fix for Irix with NO_ASM.
10121 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10123 *) ./config script fixes.
10124 [Ulf Moeller, Richard Levitte]
10126 *) Fix 'openssl passwd -1'.
10129 *) Change PKCS12_key_gen_asc() so it can cope with non null
10130 terminated strings whose length is passed in the passlen
10131 parameter, for example from PEM callbacks. This was done
10132 by adding an extra length parameter to asc2uni().
10133 [Steve Henson, reported by <oddissey@samsung.co.kr>]
10135 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
10136 call failed, free the DSA structure.
10139 *) Fix to uni2asc() to cope with zero length Unicode strings.
10140 These are present in some PKCS#12 files.
10143 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
10144 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
10145 when writing a 32767 byte record.
10146 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
10148 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
10149 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
10151 (RSA objects have a reference count access to which is protected
10152 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
10153 so they are meant to be shared between threads.)
10154 [Bodo Moeller, Geoff Thorpe; original patch submitted by
10155 "Reddie, Steven" <Steven.Reddie@ca.com>]
10157 *) Fix a deadlock in CRYPTO_mem_leaks().
10160 *) Use better test patterns in bntest.
10163 *) rand_win.c fix for Borland C.
10166 *) BN_rshift bugfix for n == 0.
10169 *) Add a 'bctest' script that checks for some known 'bc' bugs
10170 so that 'make test' does not abort just because 'bc' is broken.
10173 *) Store verify_result within SSL_SESSION also for client side to
10174 avoid potential security hole. (Re-used sessions on the client side
10175 always resulted in verify_result==X509_V_OK, not using the original
10176 result of the server certificate verification.)
10179 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
10180 SSL3_RT_APPLICATION_DATA, return 0.
10181 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
10185 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
10186 releases, have been re-implemented by renaming the previous
10187 implementations of ssl2_read and ssl3_read to ssl2_read_internal
10188 and ssl3_read_internal, respectively, and adding 'peek' parameters
10189 to them. The new ssl[23]_{read,peek} functions are calls to
10190 ssl[23]_read_internal with the 'peek' flag set appropriately.
10191 A 'peek' parameter has also been added to ssl3_read_bytes, which
10192 does the actual work for ssl3_read_internal.
10195 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
10196 the method-specific "init()" handler. Also clean up ex_data after
10197 calling the method-specific "finish()" handler. Previously, this was
10198 happening the other way round.
10201 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
10202 The previous value, 12, was not always sufficient for BN_mod_exp().
10205 *) Make sure that shared libraries get the internal name engine with
10206 the full version number and not just 0. This should mark the
10207 shared libraries as not backward compatible. Of course, this should
10208 be changed again when we can guarantee backward binary compatibility.
10211 *) Fix typo in get_cert_by_subject() in by_dir.c
10212 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
10214 *) Rework the system to generate shared libraries:
10216 - Make note of the expected extension for the shared libraries and
10217 if there is a need for symbolic links from for example libcrypto.so.0
10218 to libcrypto.so.0.9.7. There is extended info in Configure for
10221 - Make as few rebuilds of the shared libraries as possible.
10223 - Still avoid linking the OpenSSL programs with the shared libraries.
10225 - When installing, install the shared libraries separately from the
10229 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
10231 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
10232 and not in SSL_clear because the latter is also used by the
10233 accept/connect functions; previously, the settings made by
10234 SSL_set_read_ahead would be lost during the handshake.
10235 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
10237 *) Correct util/mkdef.pl to be selective about disabled algorithms.
10238 Previously, it would create entries for disabled algorithms no
10242 *) Added several new manual pages for SSL_* function.
10245 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
10247 *) In ssl23_get_client_hello, generate an error message when faced
10248 with an initial SSL 3.0/TLS record that is too small to contain the
10249 first two bytes of the ClientHello message, i.e. client_version.
10250 (Note that this is a pathologic case that probably has never happened
10251 in real life.) The previous approach was to use the version number
10252 from the record header as a substitute; but our protocol choice
10253 should not depend on that one because it is not authenticated
10254 by the Finished messages.
10257 *) More robust randomness gathering functions for Windows.
10258 [Jeffrey Altman <jaltman@columbia.edu>]
10260 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
10261 not set then we don't setup the error code for issuer check errors
10262 to avoid possibly overwriting other errors which the callback does
10263 handle. If an application does set the flag then we assume it knows
10264 what it is doing and can handle the new informational codes
10268 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
10269 a general "ANY" type, as such it should be able to decode anything
10270 including tagged types. However it didn't check the class so it would
10271 wrongly interpret tagged types in the same way as their universal
10272 counterpart and unknown types were just rejected. Changed so that the
10273 tagged and unknown types are handled in the same way as a SEQUENCE:
10274 that is the encoding is stored intact. There is also a new type
10275 "V_ASN1_OTHER" which is used when the class is not universal, in this
10276 case we have no idea what the actual type is so we just lump them all
10280 *) On VMS, stdout may very well lead to a file that is written to
10281 in a record-oriented fashion. That means that every write() will
10282 write a separate record, which will be read separately by the
10283 programs trying to read from it. This can be very confusing.
10285 The solution is to put a BIO filter in the way that will buffer
10286 text until a linefeed is reached, and then write everything a
10287 line at a time, so every record written will be an actual line,
10288 not chunks of lines and not (usually doesn't happen, but I've
10289 seen it once) several lines in one record. BIO_f_linebuffer() is
10292 Currently, it's a VMS-only method, because that's where it has
10293 been tested well enough.
10296 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
10297 it can return incorrect results.
10298 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
10299 but it was in 0.9.6-beta[12].)
10302 *) Disable the check for content being present when verifying detached
10303 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
10304 include zero length content when signing messages.
10307 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
10308 BIO_ctrl (for BIO pairs).
10311 *) Add DSO method for VMS.
10314 *) Bug fix: Montgomery multiplication could produce results with the
10318 *) Add RPM specification openssl.spec and modify it to build three
10319 packages. The default package contains applications, application
10320 documentation and run-time libraries. The devel package contains
10321 include files, static libraries and function documentation. The
10322 doc package contains the contents of the doc directory. The original
10323 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
10326 *) Add a large number of documentation files for many SSL routines.
10327 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10329 *) Add a configuration entry for Sony News 4.
10330 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
10332 *) Don't set the two most significant bits to one when generating a
10333 random number < q in the DSA library.
10336 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
10337 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
10338 the underlying transport is blocking) if a handshake took place.
10339 (The default behaviour is needed by applications such as s_client
10340 and s_server that use select() to determine when to use SSL_read;
10341 but for applications that know in advance when to expect data, it
10342 just makes things more complicated.)
10345 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
10349 *) Add a few more EBCDIC conditionals that make `req' and `x509'
10350 work better on such systems.
10351 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10353 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10354 Update PKCS12_parse() so it copies the friendlyName and the
10355 keyid to the certificates aux info.
10358 *) Fix bug in PKCS7_verify() which caused an infinite loop
10359 if there was more than one signature.
10360 [Sven Uszpelkat <su@celocom.de>]
10362 *) Major change in util/mkdef.pl to include extra information
10363 about each symbol, as well as presenting variables as well
10364 as functions. This change means that there's n more need
10365 to rebuild the .num files when some algorithms are excluded.
10368 *) Allow the verify time to be set by an application,
10369 rather than always using the current time.
10372 *) Phase 2 verify code reorganisation. The certificate
10373 verify code now looks up an issuer certificate by a
10374 number of criteria: subject name, authority key id
10375 and key usage. It also verifies self signed certificates
10376 by the same criteria. The main comparison function is
10377 X509_check_issued() which performs these checks.
10379 Lot of changes were necessary in order to support this
10380 without completely rewriting the lookup code.
10382 Authority and subject key identifier are now cached.
10384 The LHASH 'certs' is X509_STORE has now been replaced
10385 by a STACK_OF(X509_OBJECT). This is mainly because an
10386 LHASH can't store or retrieve multiple objects with
10387 the same hash value.
10389 As a result various functions (which were all internal
10390 use only) have changed to handle the new X509_STORE
10391 structure. This will break anything that messed round
10392 with X509_STORE internally.
10394 The functions X509_STORE_add_cert() now checks for an
10395 exact match, rather than just subject name.
10397 The X509_STORE API doesn't directly support the retrieval
10398 of multiple certificates matching a given criteria, however
10399 this can be worked round by performing a lookup first
10400 (which will fill the cache with candidate certificates)
10401 and then examining the cache for matches. This is probably
10402 the best we can do without throwing out X509_LOOKUP
10403 entirely (maybe later...).
10405 The X509_VERIFY_CTX structure has been enhanced considerably.
10407 All certificate lookup operations now go via a get_issuer()
10408 callback. Although this currently uses an X509_STORE it
10409 can be replaced by custom lookups. This is a simple way
10410 to bypass the X509_STORE hackery necessary to make this
10411 work and makes it possible to use more efficient techniques
10412 in future. A very simple version which uses a simple
10413 STACK for its trusted certificate store is also provided
10414 using X509_STORE_CTX_trusted_stack().
10416 The verify_cb() and verify() callbacks now have equivalents
10417 in the X509_STORE_CTX structure.
10419 X509_STORE_CTX also has a 'flags' field which can be used
10420 to customise the verify behaviour.
10423 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10424 excludes S/MIME capabilities.
10427 *) When a certificate request is read in keep a copy of the
10428 original encoding of the signed data and use it when outputting
10429 again. Signatures then use the original encoding rather than
10430 a decoded, encoded version which may cause problems if the
10431 request is improperly encoded.
10434 *) For consistency with other BIO_puts implementations, call
10435 buffer_write(b, ...) directly in buffer_puts instead of calling
10438 In BIO_puts, increment b->num_write as in BIO_write.
10439 [Peter.Sylvester@EdelWeb.fr]
10441 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10442 BN_zero, we may not return a BIGNUM with an array consisting of
10443 words set to zero.)
10446 *) Avoid calling abort() from within the library when problems are
10447 detected, except if preprocessor symbols have been defined
10448 (such as REF_CHECK, BN_DEBUG etc.).
10451 *) New openssl application 'rsautl'. This utility can be
10452 used for low level RSA operations. DER public key
10453 BIO/fp routines also added.
10456 *) New Configure entry and patches for compiling on QNX 4.
10457 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10459 *) A demo state-machine implementation was sponsored by
10460 Nuron (http://www.nuron.com/) and is now available in
10461 demos/state_machine.
10464 *) New options added to the 'dgst' utility for signature
10465 generation and verification.
10468 *) Unrecognized PKCS#7 content types are now handled via a
10469 catch all ASN1_TYPE structure. This allows unsupported
10470 types to be stored as a "blob" and an application can
10471 encode and decode it manually.
10474 *) Fix various signed/unsigned issues to make a_strex.c
10475 compile under VC++.
10476 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10478 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10479 length if passed a buffer. ASN1_INTEGER_to_BN failed
10480 if passed a NULL BN and its argument was negative.
10481 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10483 *) Modification to PKCS#7 encoding routines to output definite
10484 length encoding. Since currently the whole structures are in
10485 memory there's not real point in using indefinite length
10486 constructed encoding. However if OpenSSL is compiled with
10487 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10490 *) Added BIO_vprintf() and BIO_vsnprintf().
10493 *) Added more prefixes to parse for in the strings written
10494 through a logging bio, to cover all the levels that are available
10495 through syslog. The prefixes are now:
10497 PANIC, EMERG, EMR => LOG_EMERG
10498 ALERT, ALR => LOG_ALERT
10499 CRIT, CRI => LOG_CRIT
10500 ERROR, ERR => LOG_ERR
10501 WARNING, WARN, WAR => LOG_WARNING
10502 NOTICE, NOTE, NOT => LOG_NOTICE
10503 INFO, INF => LOG_INFO
10504 DEBUG, DBG => LOG_DEBUG
10506 and as before, if none of those prefixes are present at the
10507 beginning of the string, LOG_ERR is chosen.
10509 On Win32, the LOG_* levels are mapped according to this:
10511 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10512 LOG_WARNING => EVENTLOG_WARNING_TYPE
10513 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10517 *) Made it possible to reconfigure with just the configuration
10518 argument "reconf" or "reconfigure". The command line arguments
10519 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10520 and are retrieved from there when reconfiguring.
10523 *) MD4 implemented.
10524 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10526 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10529 *) The obj_dat.pl script was messing up the sorting of object
10530 names. The reason was that it compared the quoted version
10531 of strings as a result "OCSP" > "OCSP Signing" because
10532 " > SPACE. Changed script to store unquoted versions of
10533 names and add quotes on output. It was also omitting some
10534 names from the lookup table if they were given a default
10535 value (that is if SN is missing it is given the same
10536 value as LN and vice versa), these are now added on the
10537 grounds that if an object has a name we should be able to
10538 look it up. Finally added warning output when duplicate
10539 short or long names are found.
10542 *) Changes needed for Tandem NSK.
10543 [Scott Uroff <scott@xypro.com>]
10545 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10546 RSA_padding_check_SSLv23(), special padding was never detected
10547 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10548 version rollback attacks was not effective.
10550 In s23_clnt.c, don't use special rollback-attack detection padding
10551 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10552 client; similarly, in s23_srvr.c, don't do the rollback check if
10553 SSL 2.0 is the only protocol enabled in the server.
10556 *) Make it possible to get hexdumps of unprintable data with 'openssl
10557 asn1parse'. By implication, the functions ASN1_parse_dump() and
10558 BIO_dump_indent() are added.
10561 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10562 these print out strings and name structures based on various
10563 flags including RFC2253 support and proper handling of
10564 multibyte characters. Added options to the 'x509' utility
10565 to allow the various flags to be set.
10568 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10569 Also change the functions X509_cmp_current_time() and
10570 X509_gmtime_adj() work with an ASN1_TIME structure,
10571 this will enable certificates using GeneralizedTime in validity
10572 dates to be checked.
10575 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10576 negative public key encodings) on by default,
10577 NO_NEG_PUBKEY_BUG can be set to disable it.
10580 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10581 content octets. An i2c_ASN1_OBJECT is unnecessary because
10582 the encoding can be trivially obtained from the structure.
10585 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10586 not read locks (CRYPTO_r_[un]lock).
10589 *) A first attempt at creating official support for shared
10590 libraries through configuration. I've kept it so the
10591 default is static libraries only, and the OpenSSL programs
10592 are always statically linked for now, but there are
10593 preparations for dynamic linking in place.
10594 This has been tested on Linux and Tru64.
10597 *) Randomness polling function for Win9x, as described in:
10598 Peter Gutmann, Software Generation of Practically Strong
10602 *) Fix so PRNG is seeded in req if using an already existing
10606 *) New options to smime application. -inform and -outform
10607 allow alternative formats for the S/MIME message including
10608 PEM and DER. The -content option allows the content to be
10609 specified separately. This should allow things like Netscape
10610 form signing output easier to verify.
10613 *) Fix the ASN1 encoding of tags using the 'long form'.
10616 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10617 STRING types. These convert content octets to and from the
10618 underlying type. The actual tag and length octets are
10619 already assumed to have been read in and checked. These
10620 are needed because all other string types have virtually
10621 identical handling apart from the tag. By having versions
10622 of the ASN1 functions that just operate on content octets
10623 IMPLICIT tagging can be handled properly. It also allows
10624 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10625 and ASN1_INTEGER are identical apart from the tag.
10628 *) Change the handling of OID objects as follows:
10630 - New object identifiers are inserted in objects.txt, following
10631 the syntax given in objects.README.
10632 - objects.pl is used to process obj_mac.num and create a new
10634 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10637 This is currently kind of a hack, and the perl code in objects.pl
10638 isn't very elegant, but it works as I intended. The simplest way
10639 to check that it worked correctly is to look in obj_dat.h and
10640 check the array nid_objs and make sure the objects haven't moved
10641 around (this is important!). Additions are OK, as well as
10642 consistent name changes.
10645 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10648 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10649 The given file adds to whatever has already been seeded into the
10650 random pool through the RANDFILE configuration file option or
10651 environment variable, or the default random state file.
10654 *) mkstack.pl now sorts each macro group into lexical order.
10655 Previously the output order depended on the order the files
10656 appeared in the directory, resulting in needless rewriting
10660 *) Patches to make OpenSSL compile under Win32 again. Mostly
10661 work arounds for the VC++ problem that it treats func() as
10662 func(void). Also stripped out the parts of mkdef.pl that
10663 added extra typesafe functions: these no longer exist.
10666 *) Reorganisation of the stack code. The macros are now all
10667 collected in safestack.h . Each macro is defined in terms of
10668 a "stack macro" of the form SKM_<name>(type, a, b). The
10669 DEBUG_SAFESTACK is now handled in terms of function casts,
10670 this has the advantage of retaining type safety without the
10671 use of additional functions. If DEBUG_SAFESTACK is not defined
10672 then the non typesafe macros are used instead. Also modified the
10673 mkstack.pl script to handle the new form. Needs testing to see
10674 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10675 the default if no major problems. Similar behaviour for ASN1_SET_OF
10676 and PKCS12_STACK_OF.
10679 *) When some versions of IIS use the 'NET' form of private key the
10680 key derivation algorithm is different. Normally MD5(password) is
10681 used as a 128 bit RC4 key. In the modified case
10682 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10683 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10684 as the old Netscape_RSA functions except they have an additional
10685 'sgckey' parameter which uses the modified algorithm. Also added
10686 an -sgckey command line option to the rsa utility. Thanks to
10687 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10688 algorithm to openssl-dev.
10691 *) The evp_local.h macros were using 'c.##kname' which resulted in
10692 invalid expansion on some systems (SCO 5.0.5 for example).
10693 Corrected to 'c.kname'.
10694 [Phillip Porch <root@theporch.com>]
10696 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10697 a STACK of email addresses from a certificate or request, these look
10698 in the subject name and the subject alternative name extensions and
10699 omit any duplicate addresses.
10702 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10703 This makes DSA verification about 2 % faster.
10706 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10707 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10708 plus overhead for 1024 bit moduli).
10709 This makes exponentiations about 0.5 % faster for 1024 bit
10710 exponents (as measured by "openssl speed rsa2048").
10713 *) Rename memory handling macros to avoid conflicts with other
10715 Malloc => OPENSSL_malloc
10716 Malloc_locked => OPENSSL_malloc_locked
10717 Realloc => OPENSSL_realloc
10718 Free => OPENSSL_free
10721 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10722 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10725 *) CygWin32 support.
10726 [John Jarvie <jjarvie@newsguy.com>]
10728 *) The type-safe stack code has been rejigged. It is now only compiled
10729 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10730 by default all type-specific stack functions are "#define"d back to
10731 standard stack functions. This results in more streamlined output
10732 but retains the type-safety checking possibilities of the original
10736 *) The STACK code has been cleaned up, and certain type declarations
10737 that didn't make a lot of sense have been brought in line. This has
10738 also involved a cleanup of sorts in safestack.h to more correctly
10739 map type-safe stack functions onto their plain stack counterparts.
10740 This work has also resulted in a variety of "const"ifications of
10741 lots of the code, especially "_cmp" operations which should normally
10742 be prototyped with "const" parameters anyway.
10745 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10746 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10747 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10748 where all of 'md' is used each time the PRNG is used, but 'state'
10749 is used only indexed by a cyclic counter. As entropy may not be
10750 well distributed from the beginning, 'md' is important as a
10751 chaining variable. However, the output function chains only half
10752 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10753 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10754 in all of 'state' being rewritten, with the new values depending
10755 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10758 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10759 the handshake is continued after ssl_verify_cert_chain();
10760 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10761 can lead to 'unexplainable' connection aborts later.
10762 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10764 *) Major EVP API cipher revision.
10765 Add hooks for extra EVP features. This allows various cipher
10766 parameters to be set in the EVP interface. Support added for variable
10767 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10768 setting of RC2 and RC5 parameters.
10770 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10773 Remove lots of duplicated code from the EVP library. For example *every*
10774 cipher init() function handles the 'iv' in the same way according to the
10775 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10776 for CFB and OFB modes they zero ctx->num.
10778 New functionality allows removal of S/MIME code RC2 hack.
10780 Most of the routines have the same form and so can be declared in terms
10783 By shifting this to the top level EVP_CipherInit() it can be removed from
10784 all individual ciphers. If the cipher wants to handle IVs or keys
10785 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10788 Change lots of functions like EVP_EncryptUpdate() to now return a
10789 value: although software versions of the algorithms cannot fail
10790 any installed hardware versions can.
10793 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10794 this option is set, tolerate broken clients that send the negotiated
10795 protocol version number instead of the requested protocol version
10799 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10800 i.e. non-zero for export ciphersuites, zero otherwise.
10801 Previous versions had this flag inverted, inconsistent with
10802 rsa_tmp_cb (..._TMP_RSA_CB).
10803 [Bodo Moeller; problem reported by Amit Chopra]
10805 *) Add missing DSA library text string. Work around for some IIS
10806 key files with invalid SEQUENCE encoding.
10809 *) Add a document (doc/standards.txt) that list all kinds of standards
10810 and so on that are implemented in OpenSSL.
10813 *) Enhance c_rehash script. Old version would mishandle certificates
10814 with the same subject name hash and wouldn't handle CRLs at all.
10815 Added -fingerprint option to crl utility, to support new c_rehash
10819 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10822 *) Fix for SSL server purpose checking. Server checking was
10823 rejecting certificates which had extended key usage present
10824 but no ssl client purpose.
10825 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10827 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10828 is a little unclear about how a blank password is handled.
10829 Since the password in encoded as a BMPString with terminating
10830 double NULL a zero length password would end up as just the
10831 double NULL. However no password at all is different and is
10832 handled differently in the PKCS#12 key generation code. NS
10833 treats a blank password as zero length. MSIE treats it as no
10834 password on export: but it will try both on import. We now do
10835 the same: PKCS12_parse() tries zero length and no password if
10836 the password is set to "" or NULL (NULL is now a valid password:
10837 it wasn't before) as does the pkcs12 application.
10840 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10841 perror when PEM_read_bio_X509_REQ fails, the error message must
10842 be obtained from the error queue.
10845 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10846 it in ERR_remove_state if appropriate, and change ERR_get_state
10847 accordingly to avoid race conditions (this is necessary because
10848 thread_hash is no longer constant once set).
10851 *) Bugfix for linux-elf makefile.one.
10854 *) RSA_get_default_method() will now cause a default
10855 RSA_METHOD to be chosen if one doesn't exist already.
10856 Previously this was only set during a call to RSA_new()
10857 or RSA_new_method(NULL) meaning it was possible for
10858 RSA_get_default_method() to return NULL.
10861 *) Added native name translation to the existing DSO code
10862 that will convert (if the flag to do so is set) filenames
10863 that are sufficiently small and have no path information
10864 into a canonical native form. Eg. "blah" converted to
10865 "libblah.so" or "blah.dll" etc.
10868 *) New function ERR_error_string_n(e, buf, len) which is like
10869 ERR_error_string(e, buf), but writes at most 'len' bytes
10870 including the 0 terminator. For ERR_error_string_n, 'buf'
10872 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10874 *) CONF library reworked to become more general. A new CONF
10875 configuration file reader "class" is implemented as well as a
10876 new functions (NCONF_*, for "New CONF") to handle it. The now
10877 old CONF_* functions are still there, but are reimplemented to
10878 work in terms of the new functions. Also, a set of functions
10879 to handle the internal storage of the configuration data is
10880 provided to make it easier to write new configuration file
10881 reader "classes" (I can definitely see something reading a
10882 configuration file in XML format, for example), called _CONF_*,
10883 or "the configuration storage API"...
10885 The new configuration file reading functions are:
10887 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10888 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10890 NCONF_default, NCONF_WIN32
10892 NCONF_dump_fp, NCONF_dump_bio
10894 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10895 NCONF_new creates a new CONF object. This works in the same way
10896 as other interfaces in OpenSSL, like the BIO interface.
10897 NCONF_dump_* dump the internal storage of the configuration file,
10898 which is useful for debugging. All other functions take the same
10899 arguments as the old CONF_* functions with the exception of the
10900 first that must be a `CONF *' instead of a `LHASH *'.
10902 To make it easier to use the new classes with the old CONF_* functions,
10903 the function CONF_set_default_method is provided.
10906 *) Add '-tls1' option to 'openssl ciphers', which was already
10907 mentioned in the documentation but had not been implemented.
10908 (This option is not yet really useful because even the additional
10909 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10912 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10913 OpenSSL-based applications) load shared libraries and bind to
10914 them in a portable way.
10915 [Geoff Thorpe, with contributions from Richard Levitte]
10917 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10919 *) Make sure _lrotl and _lrotr are only used with MSVC.
10921 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10922 (the default implementation of RAND_status).
10924 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10925 to '-clrext' (= clear extensions), as intended and documented.
10926 [Bodo Moeller; inconsistency pointed out by Michael Attili
10927 <attili@amaxo.com>]
10929 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10930 was larger than the MD block size.
10931 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10933 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10934 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10935 using the passed key: if the passed key was a private key the result
10936 of X509_print(), for example, would be to print out all the private key
10940 *) des_quad_cksum() byte order bug fix.
10941 [Ulf Möller, using the problem description in krb4-0.9.7, where
10942 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10944 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10946 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10948 *) For easily testing in shell scripts whether some command
10949 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10950 returns with exit code 0 iff no command of the given name is available.
10951 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10952 the output goes to stdout and nothing is printed to stderr.
10953 Additional arguments are always ignored.
10955 Since for each cipher there is a command of the same name,
10956 the 'no-cipher' compilation switches can be tested this way.
10958 ('openssl no-XXX' is not able to detect pseudo-commands such
10959 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10962 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10965 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10966 is set; it will be thrown away anyway because each handshake creates
10968 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10969 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10970 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10971 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10974 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10975 'Q' and 'R' lose their special meanings (quit/renegotiate).
10976 This is part of what -quiet does; unlike -quiet, -ign_eof
10977 does not suppress any output.
10980 *) Add compatibility options to the purpose and trust code. The
10981 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10982 accepts a certificate or CA, this was the previous behaviour,
10983 with all the associated security issues.
10985 X509_TRUST_COMPAT is the old trust behaviour: only and
10986 automatically trust self signed roots in certificate store. A
10987 new trust setting X509_TRUST_DEFAULT is used to specify that
10988 a purpose has no associated trust setting and it should instead
10989 use the value in the default purpose.
10992 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10993 and fix a memory leak.
10996 *) In util/mkerr.pl (which implements 'make errors'), preserve
10997 reason strings from the previous version of the .c file, as
10998 the default to have only downcase letters (and digits) in
10999 automatically generated reasons codes is not always appropriate.
11002 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
11003 using strerror. Previously, ERR_reason_error_string() returned
11004 library names as reason strings for SYSerr; but SYSerr is a special
11005 case where small numbers are errno values, not library numbers.
11008 *) Add '-dsaparam' option to 'openssl dhparam' application. This
11009 converts DSA parameters into DH parameters. (When creating parameters,
11010 DSA_generate_parameters is used.)
11013 *) Include 'length' (recommended exponent length) in C code generated
11014 by 'openssl dhparam -C'.
11017 *) The second argument to set_label in perlasm was already being used
11018 so couldn't be used as a "file scope" flag. Moved to third argument
11022 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
11023 instead of RAND_bytes for encryption IVs and salts.
11026 *) Include RAND_status() into RAND_METHOD instead of implementing
11027 it only for md_rand.c Otherwise replacing the PRNG by calling
11028 RAND_set_rand_method would be impossible.
11031 *) Don't let DSA_generate_key() enter an infinite loop if the random
11032 number generation fails.
11035 *) New 'rand' application for creating pseudo-random output.
11038 *) Added configuration support for Linux/IA64
11039 [Rolf Haberrecker <rolf@suse.de>]
11041 *) Assembler module support for Mingw32.
11044 *) Shared library support for HPUX (in shlib/).
11045 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
11047 *) Shared library support for Solaris gcc.
11048 [Lutz Behnke <behnke@trustcenter.de>]
11050 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
11052 *) PKCS7_encrypt() was adding text MIME headers twice because they
11053 were added manually and by SMIME_crlf_copy().
11056 *) In bntest.c don't call BN_rand with zero bits argument.
11057 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
11059 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
11060 case was implemented. This caused BN_div_recp() to fail occasionally.
11063 *) Add an optional second argument to the set_label() in the perl
11064 assembly language builder. If this argument exists and is set
11065 to 1 it signals that the assembler should use a symbol whose
11066 scope is the entire file, not just the current function. This
11067 is needed with MASM which uses the format label:: for this scope.
11068 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
11070 *) Change the ASN1 types so they are typedefs by default. Before
11071 almost all types were #define'd to ASN1_STRING which was causing
11072 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
11076 *) Change names of new functions to the new get1/get0 naming
11077 convention: After 'get1', the caller owns a reference count
11078 and has to call ..._free; 'get0' returns a pointer to some
11079 data structure without incrementing reference counters.
11080 (Some of the existing 'get' functions increment a reference
11081 counter, some don't.)
11082 Similarly, 'set1' and 'add1' functions increase reference
11083 counters or duplicate objects.
11086 *) Allow for the possibility of temp RSA key generation failure:
11087 the code used to assume it always worked and crashed on failure.
11090 *) Fix potential buffer overrun problem in BIO_printf().
11091 [Ulf Möller, using public domain code by Patrick Powell; problem
11092 pointed out by David Sacerdote <das33@cornell.edu>]
11094 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
11095 RAND_egd() and RAND_status(). In the command line application,
11096 the EGD socket can be specified like a seed file using RANDFILE
11100 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
11101 Some CAs (e.g. Verisign) distribute certificates in this form.
11104 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
11105 list to exclude them. This means that no special compilation option
11106 is needed to use anonymous DH: it just needs to be included in the
11110 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
11111 EVP_MD_type. The old functionality is available in a new macro called
11112 EVP_MD_md(). Change code that uses it and update docs.
11115 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
11116 where the 'void *' argument is replaced by a function pointer argument.
11117 Previously 'void *' was abused to point to functions, which works on
11118 many platforms, but is not correct. As these functions are usually
11119 called by macros defined in OpenSSL header files, most source code
11120 should work without changes.
11123 *) <openssl/opensslconf.h> (which is created by Configure) now contains
11124 sections with information on -D... compiler switches used for
11125 compiling the library so that applications can see them. To enable
11126 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
11127 must be defined. E.g.,
11128 #define OPENSSL_ALGORITHM_DEFINES
11129 #include <openssl/opensslconf.h>
11130 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
11131 [Richard Levitte, Ulf and Bodo Möller]
11133 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
11137 *) Change the 'other' type in certificate aux info to a STACK_OF
11138 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
11139 the required ASN1 format: arbitrary types determined by an OID.
11142 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
11143 argument to 'req'. This is not because the function is newer or
11144 better than others it just uses the work 'NEW' in the certificate
11145 request header lines. Some software needs this.
11148 *) Reorganise password command line arguments: now passwords can be
11149 obtained from various sources. Delete the PEM_cb function and make
11150 it the default behaviour: i.e. if the callback is NULL and the
11151 usrdata argument is not NULL interpret it as a null terminated pass
11152 phrase. If usrdata and the callback are NULL then the pass phrase
11153 is prompted for as usual.
11156 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
11157 the support is automatically enabled. The resulting binaries will
11158 autodetect the card and use it if present.
11159 [Ben Laurie and Compaq Inc.]
11161 *) Work around for Netscape hang bug. This sends certificate request
11162 and server done in one record. Since this is perfectly legal in the
11163 SSL/TLS protocol it isn't a "bug" option and is on by default. See
11164 the bugs/SSLv3 entry for more info.
11167 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
11170 *) Add -rand argument to smime and pkcs12 applications and read/write
11174 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
11177 *) Add command line password options to the remaining applications.
11180 *) Bug fix for BN_div_recp() for numerators with an even number of
11184 *) More tests in bntest.c, and changed test_bn output.
11187 *) ./config recognizes MacOS X now.
11190 *) Bug fix for BN_div() when the first words of num and divisor are
11191 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
11194 *) Add support for various broken PKCS#8 formats, and command line
11195 options to produce them.
11198 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
11199 get temporary BIGNUMs from a BN_CTX.
11202 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
11206 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
11207 include a #define from the old name to the new. The original intent
11208 was that statically linked binaries could for example just call
11209 SSLeay_add_all_ciphers() to just add ciphers to the table and not
11210 link with digests. This never worked because SSLeay_add_all_digests()
11211 and SSLeay_add_all_ciphers() were in the same source file so calling
11212 one would link with the other. They are now in separate source files.
11215 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
11218 *) Use a less unusual form of the Miller-Rabin primality test (it used
11219 a binary algorithm for exponentiation integrated into the Miller-Rabin
11220 loop, our standard modexp algorithms are faster).
11223 *) Support for the EBCDIC character set completed.
11224 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
11226 *) Source code cleanups: use const where appropriate, eliminate casts,
11227 use void * instead of char * in lhash.
11230 *) Bugfix: ssl3_send_server_key_exchange was not restartable
11231 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
11232 this the server could overwrite ephemeral keys that the client
11236 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
11237 using 50 iterations of the Rabin-Miller test.
11239 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
11240 iterations of the Rabin-Miller test as required by the appendix
11241 to FIPS PUB 186[-1]) instead of DSA_is_prime.
11242 As BN_is_prime_fasttest includes trial division, DSA parameter
11243 generation becomes much faster.
11245 This implies a change for the callback functions in DSA_is_prime
11246 and DSA_generate_parameters: The callback function is called once
11247 for each positive witness in the Rabin-Miller test, not just
11248 occasionally in the inner loop; and the parameters to the
11249 callback function now provide an iteration count for the outer
11250 loop rather than for the current invocation of the inner loop.
11251 DSA_generate_parameters additionally can call the callback
11252 function with an 'iteration count' of -1, meaning that a
11253 candidate has passed the trial division test (when q is generated
11254 from an application-provided seed, trial division is skipped).
11257 *) New function BN_is_prime_fasttest that optionally does trial
11258 division before starting the Rabin-Miller test and has
11259 an additional BN_CTX * argument (whereas BN_is_prime always
11260 has to allocate at least one BN_CTX).
11261 'callback(1, -1, cb_arg)' is called when a number has passed the
11262 trial division stage.
11265 *) Fix for bug in CRL encoding. The validity dates weren't being handled
11269 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
11272 *) New function BN_pseudo_rand().
11275 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
11276 bignum version of BN_from_montgomery() with the working code from
11277 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
11281 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
11282 made it impossible to use the same SSL_SESSION data structure in
11283 SSL2 clients in multiple threads.
11286 *) The return value of RAND_load_file() no longer counts bytes obtained
11287 by stat(). RAND_load_file(..., -1) is new and uses the complete file
11288 to seed the PRNG (previously an explicit byte count was required).
11289 [Ulf Möller, Bodo Möller]
11291 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
11292 used (char *) instead of (void *) and had casts all over the place.
11295 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
11298 *) Retain source code compatibility for BN_prime_checks macro:
11299 BN_is_prime(..., BN_prime_checks, ...) now uses
11300 BN_prime_checks_for_size to determine the appropriate number of
11301 Rabin-Miller iterations.
11304 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
11305 DH_CHECK_P_NOT_SAFE_PRIME.
11306 (Check if this is true? OpenPGP calls them "strong".)
11309 *) Merge the functionality of "dh" and "gendh" programs into a new program
11310 "dhparam". The old programs are retained for now but will handle DH keys
11311 (instead of parameters) in future.
11314 *) Make the ciphers, s_server and s_client programs check the return values
11315 when a new cipher list is set.
11318 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
11319 ciphers. Before when the 56bit ciphers were enabled the sorting was
11322 The syntax for the cipher sorting has been extended to support sorting by
11323 cipher-strength (using the strength_bits hard coded in the tables).
11324 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
11326 Fix a bug in the cipher-command parser: when supplying a cipher command
11327 string with an "undefined" symbol (neither command nor alphanumeric
11328 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
11329 an error is flagged.
11331 Due to the strength-sorting extension, the code of the
11332 ssl_create_cipher_list() function was completely rearranged. I hope that
11333 the readability was also increased :-)
11334 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
11336 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
11337 for the first serial number and places 2 in the serial number file. This
11338 avoids problems when the root CA is created with serial number zero and
11339 the first user certificate has the same issuer name and serial number
11343 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
11344 the new code. Add documentation for this stuff.
11347 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
11348 X509_*() to X509at_*() on the grounds that they don't handle X509
11349 structures and behave in an analogous way to the X509v3 functions:
11350 they shouldn't be called directly but wrapper functions should be used
11353 So we also now have some wrapper functions that call the X509at functions
11354 when passed certificate requests. (TO DO: similar things can be done with
11355 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11356 things. Some of these need some d2i or i2d and print functionality
11357 because they handle more complex structures.)
11360 *) Add missing #ifndefs that caused missing symbols when building libssl
11361 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11362 NO_RSA in ssl/s2*.c.
11363 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11365 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11366 has a return value which indicates the quality of the random data
11367 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11368 error queue. New function RAND_pseudo_bytes() generates output that is
11369 guaranteed to be unique but not unpredictable. RAND_add is like
11370 RAND_seed, but takes an extra argument for an entropy estimate
11371 (RAND_seed always assumes full entropy).
11374 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11375 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11376 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11377 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11378 false-positive rate of at most 2^-80 for random input.
11381 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11384 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11385 in the 0.9.5 release), this returns the chain
11386 from an X509_CTX structure with a dup of the stack and all
11387 the X509 reference counts upped: so the stack will exist
11388 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11391 Also make SSL_SESSION_print() print out the verify return
11395 *) Add manpage for the pkcs12 command. Also change the default
11396 behaviour so MAC iteration counts are used unless the new
11397 -nomaciter option is used. This improves file security and
11398 only older versions of MSIE (4.0 for example) need it.
11401 *) Honor the no-xxx Configure options when creating .DEF files.
11404 *) Add PKCS#10 attributes to field table: challengePassword,
11405 unstructuredName and unstructuredAddress. These are taken from
11406 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11407 international characters are used.
11409 More changes to X509_ATTRIBUTE code: allow the setting of types
11410 based on strings. Remove the 'loc' parameter when adding
11411 attributes because these will be a SET OF encoding which is sorted
11415 *) Initial changes to the 'req' utility to allow request generation
11416 automation. This will allow an application to just generate a template
11417 file containing all the field values and have req construct the
11420 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11421 used all over the place including certificate requests and PKCS#7
11422 structures. They are currently handled manually where necessary with
11423 some primitive wrappers for PKCS#7. The new functions behave in a
11424 manner analogous to the X509 extension functions: they allow
11425 attributes to be looked up by NID and added.
11427 Later something similar to the X509V3 code would be desirable to
11428 automatically handle the encoding, decoding and printing of the
11429 more complex types. The string types like challengePassword can
11430 be handled by the string table functions.
11432 Also modified the multi byte string table handling. Now there is
11433 a 'global mask' which masks out certain types. The table itself
11434 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11435 is useful when for example there is only one permissible type
11436 (as in countryName) and using the mask might result in no valid
11440 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11441 SSL_get_peer_finished to allow applications to obtain the latest
11442 Finished messages sent to the peer or expected from the peer,
11443 respectively. (SSL_get_peer_finished is usually the Finished message
11444 actually received from the peer, otherwise the protocol will be aborted.)
11446 As the Finished message are message digests of the complete handshake
11447 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11448 be used for external authentication procedures when the authentication
11449 provided by SSL/TLS is not desired or is not enough.
11452 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11453 the host supports BWX extension and if Compaq C is present on the
11454 $PATH. Just exploiting of the BWX extension results in 20-30%
11455 performance kick for some algorithms, e.g. DES and RC4 to mention
11456 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11460 *) Add support for MS "fast SGC". This is arguably a violation of the
11461 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11462 weak crypto and after checking the certificate is SGC a second one
11463 with strong crypto. MS SGC stops the first handshake after receiving
11464 the server certificate message and sends a second client hello. Since
11465 a server will typically do all the time consuming operations before
11466 expecting any further messages from the client (server key exchange
11467 is the most expensive) there is little difference between the two.
11469 To get OpenSSL to support MS SGC we have to permit a second client
11470 hello message after we have sent server done. In addition we have to
11471 reset the MAC if we do get this second client hello.
11474 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11475 if a DER encoded private key is RSA or DSA traditional format. Changed
11476 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11477 format DER encoded private key. Newer code should use PKCS#8 format which
11478 has the key type encoded in the ASN1 structure. Added DER private key
11479 support to pkcs8 application.
11482 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11483 ciphersuites has been selected (as required by the SSL 3/TLS 1
11484 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11485 is set, we interpret this as a request to violate the specification
11486 (the worst that can happen is a handshake failure, and 'correct'
11487 behaviour would result in a handshake failure anyway).
11490 *) In SSL_CTX_add_session, take into account that there might be multiple
11491 SSL_SESSION structures with the same session ID (e.g. when two threads
11492 concurrently obtain them from an external cache).
11493 The internal cache can handle only one SSL_SESSION with a given ID,
11494 so if there's a conflict, we now throw out the old one to achieve
11498 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11499 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11500 some routines that use cipher OIDs: some ciphers do not have OIDs
11501 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11505 *) Simplify the trust setting structure and code. Now we just have
11506 two sequences of OIDs for trusted and rejected settings. These will
11507 typically have values the same as the extended key usage extension
11508 and any application specific purposes.
11510 The trust checking code now has a default behaviour: it will just
11511 check for an object with the same NID as the passed id. Functions can
11512 be provided to override either the default behaviour or the behaviour
11513 for a given id. SSL client, server and email already have functions
11514 in place for compatibility: they check the NID and also return "trusted"
11515 if the certificate is self signed.
11518 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11519 traditional format into an EVP_PKEY structure.
11522 *) Add a password callback function PEM_cb() which either prompts for
11523 a password if usr_data is NULL or otherwise assumes it is a null
11524 terminated password. Allow passwords to be passed on command line
11525 environment or config files in a few more utilities.
11528 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11529 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11530 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11531 Update documentation.
11534 *) Support for ASN1 "NULL" type. This could be handled before by using
11535 ASN1_TYPE but there wasn't any function that would try to read a NULL
11536 and produce an error if it couldn't. For compatibility we also have
11537 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11538 don't allocate anything because they don't need to.
11541 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11543 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11545 *) Rebuild of the memory allocation routines used by OpenSSL code and
11546 possibly others as well. The purpose is to make an interface that
11547 provide hooks so anyone can build a separate set of allocation and
11548 deallocation routines to be used by OpenSSL, for example memory
11549 pool implementations, or something else, which was previously hard
11550 since Malloc(), Realloc() and Free() were defined as macros having
11551 the values malloc, realloc and free, respectively (except for Win32
11552 compilations). The same is provided for memory debugging code.
11553 OpenSSL already comes with functionality to find memory leaks, but
11554 this gives people a chance to debug other memory problems.
11556 With these changes, a new set of functions and macros have appeared:
11558 CRYPTO_set_mem_debug_functions() [F]
11559 CRYPTO_get_mem_debug_functions() [F]
11560 CRYPTO_dbg_set_options() [F]
11561 CRYPTO_dbg_get_options() [F]
11562 CRYPTO_malloc_debug_init() [M]
11564 The memory debug functions are NULL by default, unless the library
11565 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11566 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11567 gives the standard debugging functions that come with OpenSSL) or
11568 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11569 provided by the library user) must be used. When the standard
11570 debugging functions are used, CRYPTO_dbg_set_options can be used to
11571 request additional information:
11572 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11573 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11575 Also, things like CRYPTO_set_mem_functions will always give the
11576 expected result (the new set of functions is used for allocation
11577 and deallocation) at all times, regardless of platform and compiler
11580 To finish it up, some functions that were never use in any other
11581 way than through macros have a new API and new semantic:
11583 CRYPTO_dbg_malloc()
11584 CRYPTO_dbg_realloc()
11587 All macros of value have retained their old syntax.
11588 [Richard Levitte and Bodo Moeller]
11590 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11591 ordering of SMIMECapabilities wasn't in "strength order" and there
11592 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11596 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11597 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11598 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11600 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11601 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11602 functionality to handle multipart/signed properly) and a utility
11603 called 'smime' to call all this stuff. This is based on code I
11604 originally wrote for Celo who have kindly allowed it to be
11605 included in OpenSSL.
11608 *) Add variants des_set_key_checked and des_set_key_unchecked of
11609 des_set_key (aka des_key_sched). Global variable des_check_key
11610 decides which of these is called by des_set_key; this way
11611 des_check_key behaves as it always did, but applications and
11612 the library itself, which was buggy for des_check_key == 1,
11613 have a cleaner way to pick the version they need.
11616 *) New function PKCS12_newpass() which changes the password of a
11620 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11621 dynamic mix. In both cases the ids can be used as an index into the
11622 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11623 functions so they accept a list of the field values and the
11624 application doesn't need to directly manipulate the X509_TRUST
11628 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11632 *) Modify the way the V3 extension code looks up extensions. This now
11633 works in a similar way to the object code: we have some "standard"
11634 extensions in a static table which is searched with OBJ_bsearch()
11635 and the application can add dynamic ones if needed. The file
11636 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11637 updated whenever a new extension is added to the core code and kept
11638 in ext_nid order. There is a simple program 'tabtest.c' which checks
11639 this. New extensions are not added too often so this file can readily
11640 be maintained manually.
11642 There are two big advantages in doing things this way. The extensions
11643 can be looked up immediately and no longer need to be "added" using
11644 X509V3_add_standard_extensions(): this function now does nothing.
11645 [Side note: I get *lots* of email saying the extension code doesn't
11646 work because people forget to call this function]
11647 Also no dynamic allocation is done unless new extensions are added:
11648 so if we don't add custom extensions there is no need to call
11649 X509V3_EXT_cleanup().
11652 *) Modify enc utility's salting as follows: make salting the default. Add a
11653 magic header, so unsalted files fail gracefully instead of just decrypting
11654 to garbage. This is because not salting is a big security hole, so people
11655 should be discouraged from doing it.
11658 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11659 digest to be passed on the command line but it only used this
11660 parameter when signing a certificate. Modified so all relevant
11661 operations are affected by the digest parameter including the
11662 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11663 DSA key was used because it didn't fix the digest.
11666 *) Initial certificate chain verify code. Currently tests the untrusted
11667 certificates for consistency with the verify purpose (which is set
11668 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11670 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11671 this is because it will reject chains with invalid extensions whereas
11672 every previous version of OpenSSL and SSLeay made no checks at all.
11674 Trust code: checks the root CA for the relevant trust settings. Trust
11675 settings have an initial value consistent with the verify purpose: e.g.
11676 if the verify purpose is for SSL client use it expects the CA to be
11677 trusted for SSL client use. However the default value can be changed to
11678 permit custom trust settings: one example of this would be to only trust
11679 certificates from a specific "secure" set of CAs.
11681 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11682 which should be used for version portability: especially since the
11683 verify structure is likely to change more often now.
11685 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11686 to set them. If not set then assume SSL clients will verify SSL servers
11689 Two new options to the verify program: -untrusted allows a set of
11690 untrusted certificates to be passed in and -purpose which sets the
11691 intended purpose of the certificate. If a purpose is set then the
11692 new chain verify code is used to check extension consistency.
11695 *) Support for the authority information access extension.
11698 *) Modify RSA and DSA PEM read routines to transparently handle
11699 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11700 public keys in a format compatible with certificate
11701 SubjectPublicKeyInfo structures. Unfortunately there were already
11702 functions called *_PublicKey_* which used various odd formats so
11703 these are retained for compatibility: however the DSA variants were
11704 never in a public release so they have been deleted. Changed dsa/rsa
11705 utilities to handle the new format: note no releases ever handled public
11706 keys so we should be OK.
11708 The primary motivation for this change is to avoid the same fiasco
11709 that dogs private keys: there are several incompatible private key
11710 formats some of which are standard and some OpenSSL specific and
11711 require various evil hacks to allow partial transparent handling and
11712 even then it doesn't work with DER formats. Given the option anything
11713 other than PKCS#8 should be dumped: but the other formats have to
11714 stay in the name of compatibility.
11716 With public keys and the benefit of hindsight one standard format
11717 is used which works with EVP_PKEY, RSA or DSA structures: though
11718 it clearly returns an error if you try to read the wrong kind of key.
11720 Added a -pubkey option to the 'x509' utility to output the public key.
11721 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11722 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11723 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11724 that do the same as the EVP_PKEY_assign_*() except they up the
11725 reference count of the added key (they don't "swallow" the
11729 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11730 CRLs would fail if the file contained no certificates or no CRLs:
11731 added a new function to read in both types and return the number
11732 read: this means that if none are read it will be an error. The
11733 DER versions of the certificate and CRL reader would always fail
11734 because it isn't possible to mix certificates and CRLs in DER format
11735 without choking one or the other routine. Changed this to just read
11736 a certificate: this is the best we can do. Also modified the code
11737 in apps/verify.c to take notice of return codes: it was previously
11738 attempting to read in certificates from NULL pointers and ignoring
11739 any errors: this is one reason why the cert and CRL reader seemed
11740 to work. It doesn't check return codes from the default certificate
11741 routines: these may well fail if the certificates aren't installed.
11744 *) Code to support otherName option in GeneralName.
11747 *) First update to verify code. Change the verify utility
11748 so it warns if it is passed a self signed certificate:
11749 for consistency with the normal behaviour. X509_verify
11750 has been modified to it will now verify a self signed
11751 certificate if *exactly* the same certificate appears
11752 in the store: it was previously impossible to trust a
11753 single self signed certificate. This means that:
11754 openssl verify ss.pem
11755 now gives a warning about a self signed certificate but
11756 openssl verify -CAfile ss.pem ss.pem
11760 *) For servers, store verify_result in SSL_SESSION data structure
11761 (and add it to external session representation).
11762 This is needed when client certificate verifications fails,
11763 but an application-provided verification callback (set by
11764 SSL_CTX_set_cert_verify_callback) allows accepting the session
11765 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11766 but returns 1): When the session is reused, we have to set
11767 ssl->verify_result to the appropriate error code to avoid
11769 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11771 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11772 case in PKCS7_dataInit() where the signed PKCS7 structure
11773 didn't contain any existing data because it was being created.
11774 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11776 *) Add a salt to the key derivation routines in enc.c. This
11777 forms the first 8 bytes of the encrypted file. Also add a
11778 -S option to allow a salt to be input on the command line.
11781 *) New function X509_cmp(). Oddly enough there wasn't a function
11782 to compare two certificates. We do this by working out the SHA1
11783 hash and comparing that. X509_cmp() will be needed by the trust
11787 *) SSL_get1_session() is like SSL_get_session(), but increments
11788 the reference count in the SSL_SESSION returned.
11789 [Geoff Thorpe <geoff@eu.c2.net>]
11791 *) Fix for 'req': it was adding a null to request attributes.
11792 Also change the X509_LOOKUP and X509_INFO code to handle
11793 certificate auxiliary information.
11796 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11800 *) Add the possibility to add extra information to the memory leak
11801 detecting output, to form tracebacks, showing from where each
11802 allocation was originated: CRYPTO_push_info("constant string") adds
11803 the string plus current file name and line number to a per-thread
11804 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11805 is like calling CYRPTO_pop_info() until the stack is empty.
11806 Also updated memory leak detection code to be multi-thread-safe.
11809 *) Add options -text and -noout to pkcs7 utility and delete the
11810 encryption options which never did anything. Update docs.
11813 *) Add options to some of the utilities to allow the pass phrase
11814 to be included on either the command line (not recommended on
11815 OSes like Unix) or read from the environment. Update the
11816 manpages and fix a few bugs.
11819 *) Add a few manpages for some of the openssl commands.
11822 *) Fix the -revoke option in ca. It was freeing up memory twice,
11823 leaking and not finding already revoked certificates.
11826 *) Extensive changes to support certificate auxiliary information.
11827 This involves the use of X509_CERT_AUX structure and X509_AUX
11828 functions. An X509_AUX function such as PEM_read_X509_AUX()
11829 can still read in a certificate file in the usual way but it
11830 will also read in any additional "auxiliary information". By
11831 doing things this way a fair degree of compatibility can be
11832 retained: existing certificates can have this information added
11833 using the new 'x509' options.
11835 Current auxiliary information includes an "alias" and some trust
11836 settings. The trust settings will ultimately be used in enhanced
11837 certificate chain verification routines: currently a certificate
11838 can only be trusted if it is self signed and then it is trusted
11842 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11843 The problem was that one of the replacement routines had not been working
11844 since SSLeay releases. For now the offending routine has been replaced
11845 with non-optimised assembler. Even so, this now gives around 95%
11846 performance improvement for 1024 bit RSA signs.
11849 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11850 handling. Most clients have the effective key size in bits equal to
11851 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11852 A few however don't do this and instead use the size of the decrypted key
11853 to determine the RC2 key length and the AlgorithmIdentifier to determine
11854 the effective key length. In this case the effective key length can still
11855 be 40 bits but the key length can be 168 bits for example. This is fixed
11856 by manually forcing an RC2 key into the EVP_PKEY structure because the
11857 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11858 the key length and effective key length are equal.
11861 *) Add a bunch of functions that should simplify the creation of
11862 X509_NAME structures. Now you should be able to do:
11863 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11864 and have it automatically work out the correct field type and fill in
11865 the structures. The more adventurous can try:
11866 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11867 and it will (hopefully) work out the correct multibyte encoding.
11870 *) Change the 'req' utility to use the new field handling and multibyte
11871 copy routines. Before the DN field creation was handled in an ad hoc
11872 way in req, ca, and x509 which was rather broken and didn't support
11873 BMPStrings or UTF8Strings. Since some software doesn't implement
11874 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11875 using the dirstring_type option. See the new comment in the default
11876 openssl.cnf for more info.
11879 *) Make crypto/rand/md_rand.c more robust:
11880 - Assure unique random numbers after fork().
11881 - Make sure that concurrent threads access the global counter and
11882 md serializably so that we never lose entropy in them
11883 or use exactly the same state in multiple threads.
11884 Access to the large state is not always serializable because
11885 the additional locking could be a performance killer, and
11886 md should be large enough anyway.
11889 *) New file apps/app_rand.c with commonly needed functionality
11890 for handling the random seed file.
11892 Use the random seed file in some applications that previously did not:
11894 dsaparam -genkey (which also ignored its '-rand' option),
11897 x509 (when signing).
11898 Except on systems with /dev/urandom, it is crucial to have a random
11899 seed file at least for key creation, DSA signing, and for DH exchanges;
11900 for RSA signatures we could do without one.
11902 gendh and gendsa (unlike genrsa) used to read only the first byte
11903 of each file listed in the '-rand' option. The function as previously
11904 found in genrsa is now in app_rand.c and is used by all programs
11905 that support '-rand'.
11908 *) In RAND_write_file, use mode 0600 for creating files;
11909 don't just chmod when it may be too late.
11912 *) Report an error from X509_STORE_load_locations
11913 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11916 *) New function ASN1_mbstring_copy() this copies a string in either
11917 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11918 into an ASN1_STRING type. A mask of permissible types is passed
11919 and it chooses the "minimal" type to use or an error if not type
11923 *) Add function equivalents to the various macros in asn1.h. The old
11924 macros are retained with an M_ prefix. Code inside the library can
11925 use the M_ macros. External code (including the openssl utility)
11926 should *NOT* in order to be "shared library friendly".
11929 *) Add various functions that can check a certificate's extensions
11930 to see if it usable for various purposes such as SSL client,
11931 server or S/MIME and CAs of these types. This is currently
11932 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11933 verification. Also added a -purpose flag to x509 utility to
11934 print out all the purposes.
11937 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11941 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11942 for, obtain and decode and extension and obtain its critical flag.
11943 This allows all the necessary extension code to be handled in a
11944 single function call.
11947 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11948 platforms. See crypto/rc4/rc4_enc.c for further details.
11951 *) New -noout option to asn1parse. This causes no output to be produced
11952 its main use is when combined with -strparse and -out to extract data
11953 from a file (which may not be in ASN.1 format).
11956 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11957 when producing the local key id.
11958 [Richard Levitte <levitte@stacken.kth.se>]
11960 *) New option -dhparam in s_server. This allows a DH parameter file to be
11961 stated explicitly. If it is not stated then it tries the first server
11962 certificate file. The previous behaviour hard coded the filename
11966 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11967 a public key to be input or output. For example:
11968 openssl rsa -in key.pem -pubout -out pubkey.pem
11969 Also added necessary DSA public key functions to handle this.
11972 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11973 in the message. This was handled by allowing
11974 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11975 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11977 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11978 to the end of the strings whereas this didn't. This would cause problems
11979 if strings read with d2i_ASN1_bytes() were later modified.
11980 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11982 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11983 data and it contains EOF it will end up returning an error. This is
11984 caused by input 46 bytes long. The cause is due to the way base64
11985 BIOs find the start of base64 encoded data. They do this by trying a
11986 trial decode on each line until they find one that works. When they
11987 do a flag is set and it starts again knowing it can pass all the
11988 data directly through the decoder. Unfortunately it doesn't reset
11989 the context it uses. This means that if EOF is reached an attempt
11990 is made to pass two EOFs through the context and this causes the
11991 resulting error. This can also cause other problems as well. As is
11992 usual with these problems it takes *ages* to find and the fix is
11993 trivial: move one line.
11994 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11996 *) Ugly workaround to get s_client and s_server working under Windows. The
11997 old code wouldn't work because it needed to select() on sockets and the
11998 tty (for keypresses and to see if data could be written). Win32 only
11999 supports select() on sockets so we select() with a 1s timeout on the
12000 sockets and then see if any characters are waiting to be read, if none
12001 are present then we retry, we also assume we can always write data to
12002 the tty. This isn't nice because the code then blocks until we've
12003 received a complete line of data and it is effectively polling the
12004 keyboard at 1s intervals: however it's quite a bit better than not
12005 working at all :-) A dedicated Windows application might handle this
12006 with an event loop for example.
12009 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
12010 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
12011 will be called when RSA_sign() and RSA_verify() are used. This is useful
12012 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
12013 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
12014 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
12015 This necessitated the support of an extra signature type NID_md5_sha1
12016 for SSL signatures and modifications to the SSL library to use it instead
12017 of calling RSA_public_decrypt() and RSA_private_encrypt().
12020 *) Add new -verify -CAfile and -CApath options to the crl program, these
12021 will lookup a CRL issuers certificate and verify the signature in a
12022 similar way to the verify program. Tidy up the crl program so it
12023 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
12024 less strict. It will now permit CRL extensions even if it is not
12025 a V2 CRL: this will allow it to tolerate some broken CRLs.
12028 *) Initialize all non-automatic variables each time one of the openssl
12029 sub-programs is started (this is necessary as they may be started
12030 multiple times from the "OpenSSL>" prompt).
12031 [Lennart Bang, Bodo Moeller]
12033 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
12034 removing all other RSA functionality (this is what NO_RSA does). This
12035 is so (for example) those in the US can disable those operations covered
12036 by the RSA patent while allowing storage and parsing of RSA keys and RSA
12040 *) Non-copying interface to BIO pairs.
12041 (still largely untested)
12044 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
12045 ASCII string. This was handled independently in various places before.
12048 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
12049 UTF8 strings a character at a time.
12052 *) Use client_version from client hello to select the protocol
12053 (s23_srvr.c) and for RSA client key exchange verification
12054 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
12057 *) Add various utility functions to handle SPKACs, these were previously
12058 handled by poking round in the structure internals. Added new function
12059 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
12060 print, verify and generate SPKACs. Based on an original idea from
12061 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
12064 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
12067 *) Allow the config file extension section to be overwritten on the
12068 command line. Based on an original idea from Massimiliano Pala
12069 <madwolf@comune.modena.it>. The new option is called -extensions
12070 and can be applied to ca, req and x509. Also -reqexts to override
12071 the request extensions in req and -crlexts to override the crl extensions
12075 *) Add new feature to the SPKAC handling in ca. Now you can include
12076 the same field multiple times by preceding it by "XXXX." for example:
12079 this is the same syntax as used in the req config file.
12082 *) Allow certificate extensions to be added to certificate requests. These
12083 are specified in a 'req_extensions' option of the req section of the
12084 config file. They can be printed out with the -text option to req but
12085 are otherwise ignored at present.
12088 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
12089 data read consists of only the final block it would not decrypted because
12090 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
12091 A misplaced 'break' also meant the decrypted final block might not be
12092 copied until the next read.
12095 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
12096 a few extra parameters to the DH structure: these will be useful if
12097 for example we want the value of 'q' or implement X9.42 DH.
12100 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
12101 provides hooks that allow the default DSA functions or functions on a
12102 "per key" basis to be replaced. This allows hardware acceleration and
12103 hardware key storage to be handled without major modification to the
12104 library. Also added low level modexp hooks and CRYPTO_EX structure and
12105 associated functions.
12108 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
12109 as "read only": it can't be written to and the buffer it points to will
12110 not be freed. Reading from a read only BIO is much more efficient than
12111 a normal memory BIO. This was added because there are several times when
12112 an area of memory needs to be read from a BIO. The previous method was
12113 to create a memory BIO and write the data to it, this results in two
12114 copies of the data and an O(n^2) reading algorithm. There is a new
12115 function BIO_new_mem_buf() which creates a read only memory BIO from
12116 an area of memory. Also modified the PKCS#7 routines to use read only
12120 *) Bugfix: ssl23_get_client_hello did not work properly when called in
12121 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
12122 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
12123 but a retry condition occurred while trying to read the rest.
12126 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
12127 NID_pkcs7_encrypted by default: this was wrong since this should almost
12128 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
12129 the encrypted data type: this is a more sensible place to put it and it
12130 allows the PKCS#12 code to be tidied up that duplicated this
12134 *) Changed obj_dat.pl script so it takes its input and output files on
12135 the command line. This should avoid shell escape redirection problems
12139 *) Initial support for certificate extension requests, these are included
12140 in things like Xenroll certificate requests. Included functions to allow
12141 extensions to be obtained and added.
12144 *) -crlf option to s_client and s_server for sending newlines as
12145 CRLF (as required by many protocols).
12148 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
12150 *) Install libRSAglue.a when OpenSSL is built with RSAref.
12151 [Ralf S. Engelschall]
12153 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
12154 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
12156 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
12160 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
12161 DH parameters/keys (q is lost during that conversion, but the resulting
12162 DH parameters contain its length).
12164 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
12165 much faster than DH_generate_parameters (which creates parameters
12166 where p = 2*q + 1), and also the smaller q makes DH computations
12167 much more efficient (160-bit exponentiation instead of 1024-bit
12168 exponentiation); so this provides a convenient way to support DHE
12169 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
12170 utter importance to use
12171 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12173 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12174 when such DH parameters are used, because otherwise small subgroup
12175 attacks may become possible!
12178 *) Avoid memory leak in i2d_DHparams.
12181 *) Allow the -k option to be used more than once in the enc program:
12182 this allows the same encrypted message to be read by multiple recipients.
12185 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
12186 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
12187 it will always use the numerical form of the OID, even if it has a short
12191 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
12192 method only got called if p,q,dmp1,dmq1,iqmp components were present,
12193 otherwise bn_mod_exp was called. In the case of hardware keys for example
12194 no private key components need be present and it might store extra data
12195 in the RSA structure, which cannot be accessed from bn_mod_exp.
12196 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
12197 private key operations.
12200 *) Added support for SPARC Linux.
12203 *) pem_password_cb function type incompatibly changed from
12204 typedef int pem_password_cb(char *buf, int size, int rwflag);
12206 ....(char *buf, int size, int rwflag, void *userdata);
12207 so that applications can pass data to their callbacks:
12208 The PEM[_ASN1]_{read,write}... functions and macros now take an
12209 additional void * argument, which is just handed through whenever
12210 the password callback is called.
12211 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
12213 New function SSL_CTX_set_default_passwd_cb_userdata.
12215 Compatibility note: As many C implementations push function arguments
12216 onto the stack in reverse order, the new library version is likely to
12217 interoperate with programs that have been compiled with the old
12218 pem_password_cb definition (PEM_whatever takes some data that
12219 happens to be on the stack as its last argument, and the callback
12220 just ignores this garbage); but there is no guarantee whatsoever that
12223 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
12224 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
12225 problems not only on Windows, but also on some Unix platforms.
12226 To avoid problematic command lines, these definitions are now in an
12227 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
12228 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
12231 *) MIPS III/IV assembler module is reimplemented.
12234 *) More DES library cleanups: remove references to srand/rand and
12235 delete an unused file.
12238 *) Add support for the free Netwide assembler (NASM) under Win32,
12239 since not many people have MASM (ml) and it can be hard to obtain.
12240 This is currently experimental but it seems to work OK and pass all
12241 the tests. Check out INSTALL.W32 for info.
12244 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
12245 without temporary keys kept an extra copy of the server key,
12246 and connections with temporary keys did not free everything in case
12250 *) New function RSA_check_key and new openssl rsa option -check
12251 for verifying the consistency of RSA keys.
12252 [Ulf Moeller, Bodo Moeller]
12254 *) Various changes to make Win32 compile work:
12255 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
12256 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
12257 comparison" warnings.
12258 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
12261 *) Add a debugging option to PKCS#5 v2 key generation function: when
12262 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
12263 derived keys are printed to stderr.
12266 *) Copy the flags in ASN1_STRING_dup().
12267 [Roman E. Pavlov <pre@mo.msk.ru>]
12269 *) The x509 application mishandled signing requests containing DSA
12270 keys when the signing key was also DSA and the parameters didn't match.
12272 It was supposed to omit the parameters when they matched the signing key:
12273 the verifying software was then supposed to automatically use the CA's
12274 parameters if they were absent from the end user certificate.
12276 Omitting parameters is no longer recommended. The test was also
12277 the wrong way round! This was probably due to unusual behaviour in
12278 EVP_cmp_parameters() which returns 1 if the parameters match.
12279 This meant that parameters were omitted when they *didn't* match and
12280 the certificate was useless. Certificates signed with 'ca' didn't have
12282 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
12284 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
12285 The interface is as follows:
12286 Applications can use
12287 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
12288 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
12289 "off" is now the default.
12290 The library internally uses
12291 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
12292 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
12293 to disable memory-checking temporarily.
12295 Some inconsistent states that previously were possible (and were
12296 even the default) are now avoided.
12298 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
12299 with each memory chunk allocated; this is occasionally more helpful
12300 than just having a counter.
12302 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
12304 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
12308 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
12309 which largely parallels "options", but is for changing API behaviour,
12310 whereas "options" are about protocol behaviour.
12311 Initial "mode" flags are:
12313 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
12314 a single record has been written.
12315 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
12316 retries use the same buffer location.
12317 (But all of the contents must be
12321 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
12324 *) Fix problems with no-hmac etc.
12325 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
12327 *) New functions RSA_get_default_method(), RSA_set_method() and
12328 RSA_get_method(). These allows replacement of RSA_METHODs without having
12329 to mess around with the internals of an RSA structure.
12332 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
12333 Also really enable memory leak checks in openssl.c and in some
12335 [Chad C. Mulligan, Bodo Moeller]
12337 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
12338 up the length of negative integers. This has now been simplified to just
12339 store the length when it is first determined and use it later, rather
12340 than trying to keep track of where data is copied and updating it to
12342 [Steve Henson, reported by Brien Wheeler
12343 <bwheeler@authentica-security.com>]
12345 *) Add a new function PKCS7_signatureVerify. This allows the verification
12346 of a PKCS#7 signature but with the signing certificate passed to the
12347 function itself. This contrasts with PKCS7_dataVerify which assumes the
12348 certificate is present in the PKCS#7 structure. This isn't always the
12349 case: certificates can be omitted from a PKCS#7 structure and be
12350 distributed by "out of band" means (such as a certificate database).
12353 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12354 function prototypes in pem.h, also change util/mkdef.pl to add the
12355 necessary function names.
12358 *) mk1mf.pl (used by Windows builds) did not properly read the
12359 options set by Configure in the top level Makefile, and Configure
12360 was not even able to write more than one option correctly.
12361 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12364 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12365 file to be loaded from a BIO or FILE pointer. The BIO version will
12366 for example allow memory BIOs to contain config info.
12369 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12370 Whoever hopes to achieve shared-library compatibility across versions
12371 must use this, not the compile-time macro.
12372 (Exercise 0.9.4: Which is the minimum library version required by
12374 Note: All this applies only to multi-threaded programs, others don't
12378 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12379 through a BIO pair triggered the default case, i.e.
12380 SSLerr(...,SSL_R_UNKNOWN_STATE).
12383 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12384 can use the SSL library even if none of the specific BIOs is
12388 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12389 for the encoded length.
12390 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12392 *) Add initial documentation of the X509V3 functions.
12395 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12396 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12397 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12398 secure PKCS#8 private key format with a high iteration count.
12401 *) Fix determination of Perl interpreter: A perl or perl5
12402 _directory_ in $PATH was also accepted as the interpreter.
12403 [Ralf S. Engelschall]
12405 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12406 wrong with it but it was very old and did things like calling
12407 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12408 unusual formatting.
12411 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12412 to use the new extension code.
12415 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12416 with macros. This should make it easier to change their form, add extra
12417 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12421 *) Add to configuration table a new entry that can specify an alternative
12422 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12423 according to Mark Crispin <MRC@Panda.COM>.
12427 *) DES CBC did not update the IV. Weird.
12430 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12431 Changing the behaviour of the former might break existing programs --
12432 where IV updating is needed, des_ncbc_encrypt can be used.
12435 *) When bntest is run from "make test" it drives bc to check its
12436 calculations, as well as internally checking them. If an internal check
12437 fails, it needs to cause bc to give a non-zero result or make test carries
12438 on without noticing the failure. Fixed.
12441 *) DES library cleanups.
12444 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12445 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12446 ciphers. NOTE: although the key derivation function has been verified
12447 against some published test vectors it has not been extensively tested
12448 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12452 *) Instead of "mkdir -p", which is not fully portable, use new
12453 Perl script "util/mkdir-p.pl".
12456 *) Rewrite the way password based encryption (PBE) is handled. It used to
12457 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12458 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12459 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12460 the 'parameter' field of the AlgorithmIdentifier is passed to the
12461 underlying key generation function so it must do its own ASN1 parsing.
12462 This has also changed the EVP_PBE_CipherInit() function which now has a
12463 'parameter' argument instead of literal salt and iteration count values
12464 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12467 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12468 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12469 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12470 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12471 value was just used as a "magic string" and not used directly its
12472 value doesn't matter.
12475 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12479 *) "linux-sparc64" configuration (ultrapenguin).
12480 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12481 "linux-sparc" configuration.
12482 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12484 *) config now generates no-xxx options for missing ciphers.
12487 *) Support the EBCDIC character set (work in progress).
12488 File ebcdic.c not yet included because it has a different license.
12489 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12491 *) Support BS2000/OSD-POSIX.
12492 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12494 *) Make callbacks for key generation use void * instead of char *.
12497 *) Make S/MIME samples compile (not yet tested).
12500 *) Additional typesafe stacks.
12503 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12507 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12509 *) New configuration variant "sco5-gcc".
12511 *) Updated some demos.
12512 [Sean O Riordain, Wade Scholine]
12514 *) Add missing BIO_free at exit of pkcs12 application.
12517 *) Fix memory leak in conf.c.
12520 *) Updates for Win32 to assembler version of MD5.
12523 *) Set #! path to perl in apps/der_chop to where we found it
12524 instead of using a fixed path.
12527 *) SHA library changes for irix64-mips4-cc.
12530 *) Improvements for VMS support.
12534 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12536 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12537 This also avoids the problems with SC4.2 and unpatched SC5.
12538 [Andy Polyakov <appro@fy.chalmers.se>]
12540 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12541 These are required because of the typesafe stack would otherwise break
12542 existing code. If old code used a structure member which used to be STACK
12543 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12544 sk_num or sk_value it would produce an error because the num, data members
12545 are not present in STACK_OF. Now it just produces a warning. sk_set
12546 replaces the old method of assigning a value to sk_value
12547 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12548 that does this will no longer work (and should use sk_set instead) but
12549 this could be regarded as a "questionable" behaviour anyway.
12552 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12553 correctly handle encrypted S/MIME data.
12556 *) Change type of various DES function arguments from des_cblock
12557 (which means, in function argument declarations, pointer to char)
12558 to des_cblock * (meaning pointer to array with 8 char elements),
12559 which allows the compiler to do more typechecking; it was like
12560 that back in SSLeay, but with lots of ugly casts.
12562 Introduce new type const_des_cblock.
12565 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12566 problems: find RecipientInfo structure that matches recipient certificate
12567 and initialise the ASN1 structures properly based on passed cipher.
12570 *) Belatedly make the BN tests actually check the results.
12573 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12574 to and from BNs: it was completely broken. New compilation option
12575 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12576 key elements as negative integers.
12579 *) Reorganize and speed up MD5.
12580 [Andy Polyakov <appro@fy.chalmers.se>]
12583 [Richard Levitte <richard@levitte.org>]
12585 *) New option -out to asn1parse to allow the parsed structure to be
12586 output to a file. This is most useful when combined with the -strparse
12587 option to examine the output of things like OCTET STRINGS.
12590 *) Make SSL library a little more fool-proof by not requiring any longer
12591 that SSL_set_{accept,connect}_state be called before
12592 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12593 in many applications because usually everything *appeared* to work as
12594 intended anyway -- now it really works as intended).
12597 *) Move openssl.cnf out of lib/.
12600 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12601 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12602 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12603 [Ralf S. Engelschall]
12605 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12606 handle PKCS#7 enveloped data properly.
12607 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12609 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12610 copying pointers. The cert_st handling is changed by this in
12611 various ways (and thus what used to be known as ctx->default_cert
12612 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12613 any longer when s->cert does not give us what we need).
12614 ssl_cert_instantiate becomes obsolete by this change.
12615 As soon as we've got the new code right (possibly it already is?),
12616 we have solved a couple of bugs of the earlier code where s->cert
12617 was used as if it could not have been shared with other SSL structures.
12619 Note that using the SSL API in certain dirty ways now will result
12620 in different behaviour than observed with earlier library versions:
12621 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12622 does not influence s as it used to.
12624 In order to clean up things more thoroughly, inside SSL_SESSION
12625 we don't use CERT any longer, but a new structure SESS_CERT
12626 that holds per-session data (if available); currently, this is
12627 the peer's certificate chain and, for clients, the server's certificate
12628 and temporary key. CERT holds only those values that can have
12629 meaningful defaults in an SSL_CTX.
12632 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12633 from the internal representation. Various PKCS#7 fixes: remove some
12634 evil casts and set the enc_dig_alg field properly based on the signing
12638 *) Allow PKCS#12 password to be set from the command line or the
12639 environment. Let 'ca' get its config file name from the environment
12640 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12644 *) Allow certificate policies extension to use an IA5STRING for the
12645 organization field. This is contrary to the PKIX definition but
12646 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12650 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12651 without disallowing inline assembler and the like for non-pedantic builds.
12654 *) Support Borland C++ builder.
12655 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12657 *) Support Mingw32.
12660 *) SHA-1 cleanups and performance enhancements.
12661 [Andy Polyakov <appro@fy.chalmers.se>]
12663 *) Sparc v8plus assembler for the bignum library.
12664 [Andy Polyakov <appro@fy.chalmers.se>]
12666 *) Accept any -xxx and +xxx compiler options in Configure.
12669 *) Update HPUX configuration.
12672 *) Add missing sk_<type>_unshift() function to safestack.h
12673 [Ralf S. Engelschall]
12675 *) New function SSL_CTX_use_certificate_chain_file that sets the
12676 "extra_cert"s in addition to the certificate. (This makes sense
12677 only for "PEM" format files, as chains as a whole are not
12681 *) Support verify_depth from the SSL API.
12682 x509_vfy.c had what can be considered an off-by-one-error:
12683 Its depth (which was not part of the external interface)
12684 was actually counting the number of certificates in a chain;
12685 now it really counts the depth.
12688 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12689 instead of X509err, which often resulted in confusing error
12690 messages since the error codes are not globally unique
12691 (e.g. an alleged error in ssl3_accept when a certificate
12692 didn't match the private key).
12694 *) New function SSL_CTX_set_session_id_context that allows to set a default
12695 value (so that you don't need SSL_set_session_id_context for each
12696 connection using the SSL_CTX).
12699 *) OAEP decoding bug fix.
12702 *) Support INSTALL_PREFIX for package builders, as proposed by
12706 *) New Configure options "threads" and "no-threads". For systems
12707 where the proper compiler options are known (currently Solaris
12708 and Linux), "threads" is the default.
12711 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12714 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12715 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12716 such as /usr/local/bin.
12719 *) "make linux-shared" to build shared libraries.
12720 [Niels Poppe <niels@netbox.org>]
12722 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12725 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12726 extension adding in x509 utility.
12729 *) Remove NOPROTO sections and error code comments.
12732 *) Partial rewrite of the DEF file generator to now parse the ANSI
12736 *) New Configure options --prefix=DIR and --openssldir=DIR.
12739 *) Complete rewrite of the error code script(s). It is all now handled
12740 by one script at the top level which handles error code gathering,
12741 header rewriting and C source file generation. It should be much better
12742 than the old method: it now uses a modified version of Ulf's parser to
12743 read the ANSI prototypes in all header files (thus the old K&R definitions
12744 aren't needed for error creation any more) and do a better job of
12745 translating function codes into names. The old 'ASN1 error code embedded
12746 in a comment' is no longer necessary and it doesn't use .err files which
12747 have now been deleted. Also the error code call doesn't have to appear all
12748 on one line (which resulted in some large lines...).
12751 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12754 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12755 0 (which usually indicates a closed connection), but continue reading.
12758 *) Fix some race conditions.
12761 *) Add support for CRL distribution points extension. Add Certificate
12762 Policies and CRL distribution points documentation.
12765 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12768 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12769 8 of keying material. Merlin has also confirmed interop with this fix
12770 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12771 [Merlin Hughes <merlin@baltimore.ie>]
12773 *) Fix lots of warnings.
12774 [Richard Levitte <levitte@stacken.kth.se>]
12776 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12777 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12778 [Richard Levitte <levitte@stacken.kth.se>]
12780 *) Fix problems with sizeof(long) == 8.
12781 [Andy Polyakov <appro@fy.chalmers.se>]
12783 *) Change functions to ANSI C.
12786 *) Fix typos in error codes.
12787 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12789 *) Remove defunct assembler files from Configure.
12792 *) SPARC v8 assembler BIGNUM implementation.
12793 [Andy Polyakov <appro@fy.chalmers.se>]
12795 *) Support for Certificate Policies extension: both print and set.
12796 Various additions to support the r2i method this uses.
12799 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12800 return a const string when you are expecting an allocated buffer.
12803 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12804 types DirectoryString and DisplayText.
12807 *) Add code to allow r2i extensions to access the configuration database,
12808 add an LHASH database driver and add several ctx helper functions.
12811 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12812 fail when they extended the size of a BIGNUM.
12815 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12816 support typesafe stack.
12819 *) Fix typo in SSL_[gs]et_options().
12820 [Nils Frostberg <nils@medcom.se>]
12822 *) Delete various functions and files that belonged to the (now obsolete)
12823 old X509V3 handling code.
12826 *) New Configure option "rsaref".
12829 *) Don't auto-generate pem.h.
12832 *) Introduce type-safe ASN.1 SETs.
12835 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12836 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12838 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12839 that links with OpenSSL (well at least cause lots of warnings), but fear
12840 not: the conversion is trivial, and it eliminates loads of evil casts. A
12841 few STACKed things have been converted already. Feel free to convert more.
12842 In the fullness of time, I'll do away with the STACK type altogether.
12845 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12846 specified in <certfile> by updating the entry in the index.txt file.
12847 This way one no longer has to edit the index.txt file manually for
12848 revoking a certificate. The -revoke option does the gory details now.
12849 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12851 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12852 `-text' option at all and this way the `-noout -text' combination was
12853 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12854 [Ralf S. Engelschall]
12856 *) Make sure a corresponding plain text error message exists for the
12857 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12858 verify callback function determined that a certificate was revoked.
12859 [Ralf S. Engelschall]
12861 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12862 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12863 all available ciphers including rc5, which was forgotten until now.
12864 In order to let the testing shell script know which algorithms
12865 are available, a new (up to now undocumented) command
12866 "openssl list-cipher-commands" is used.
12869 *) Bugfix: s_client occasionally would sleep in select() when
12870 it should have checked SSL_pending() first.
12873 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12874 the raw DSA values prior to ASN.1 encoding.
12877 *) Tweaks to Configure
12878 [Niels Poppe <niels@netbox.org>]
12880 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12884 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12887 *) New config option to avoid instructions that are illegal on the 80386.
12888 The default code is faster, but requires at least a 486.
12891 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12892 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12893 same as SSL2_VERSION anyway.
12896 *) New "-showcerts" option for s_client.
12899 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12900 application. Various cleanups and fixes.
12903 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12904 modify error routines to work internally. Add error codes and PBE init
12905 to library startup routines.
12908 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12909 packing functions to asn1 and evp. Changed function names and error
12910 codes along the way.
12913 *) PKCS12 integration: and so it begins... First of several patches to
12914 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12915 objects to objects.h
12918 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12919 and display support for Thawte strong extranet extension.
12922 *) Add LinuxPPC support.
12923 [Jeff Dubrule <igor@pobox.org>]
12925 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12926 bn_div_words in alpha.s.
12927 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12929 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12930 OAEP isn't supported when OpenSSL is built with RSAref.
12931 [Ulf Moeller <ulf@fitug.de>]
12933 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12934 so they no longer are missing under -DNOPROTO.
12935 [Soren S. Jorvang <soren@t.dk>]
12938 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12940 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12941 doesn't work when the session is reused. Coming soon!
12944 *) Fix a security hole, that allows sessions to be reused in the wrong
12945 context thus bypassing client cert protection! All software that uses
12946 client certs and session caches in multiple contexts NEEDS PATCHING to
12947 allow session reuse! A fuller solution is in the works.
12948 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12950 *) Some more source tree cleanups (removed obsolete files
12951 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12952 permission on "config" script to be executable) and a fix for the INSTALL
12954 [Ulf Moeller <ulf@fitug.de>]
12956 *) Remove some legacy and erroneous uses of malloc, free instead of
12958 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12960 *) Make rsa_oaep_test return non-zero on error.
12961 [Ulf Moeller <ulf@fitug.de>]
12963 *) Add support for native Solaris shared libraries. Configure
12964 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12965 if someone would make that last step automatic.
12966 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12968 *) ctx_size was not built with the right compiler during "make links". Fixed.
12971 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12972 except NULL ciphers". This means the default cipher list will no longer
12973 enable NULL ciphers. They need to be specifically enabled e.g. with
12974 the string "DEFAULT:eNULL".
12977 *) Fix to RSA private encryption routines: if p < q then it would
12978 occasionally produce an invalid result. This will only happen with
12979 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12982 *) Be less restrictive and allow also `perl util/perlpath.pl
12983 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12984 because this way one can also use an interpreter named `perl5' (which is
12985 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12986 installed as `perl').
12987 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12989 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12990 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12992 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12993 advapi32.lib to Win32 build and change the pem test comparison
12994 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12995 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12996 and crypto/des/ede_cbcm_enc.c.
12999 *) DES quad checksum was broken on big-endian architectures. Fixed.
13002 *) Comment out two functions in bio.h that aren't implemented. Fix up the
13003 Win32 test batch file so it (might) work again. The Win32 test batch file
13004 is horrible: I feel ill....
13007 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
13008 in e_os.h. Audit of header files to check ANSI and non ANSI
13009 sections: 10 functions were absent from non ANSI section and not exported
13010 from Windows DLLs. Fixed up libeay.num for new functions.
13013 *) Make `openssl version' output lines consistent.
13014 [Ralf S. Engelschall]
13016 *) Fix Win32 symbol export lists for BIO functions: Added
13017 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
13018 to ms/libeay{16,32}.def.
13019 [Ralf S. Engelschall]
13021 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
13022 fine under Unix and passes some trivial tests I've now added. But the
13023 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
13024 added to make sure no one expects that this stuff really works in the
13025 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
13026 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
13028 [Ralf S. Engelschall]
13030 *) Fix the generation of two part addresses in perl.
13031 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
13033 *) Add config entry for Linux on MIPS.
13034 [John Tobey <jtobey@channel1.com>]
13036 *) Make links whenever Configure is run, unless we are on Windoze.
13039 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
13040 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
13044 *) Add a useful kludge to allow package maintainers to specify compiler and
13045 other platforms details on the command line without having to patch the
13046 Configure script every time: One now can use ``perl Configure
13047 <id>:<details>'', i.e. platform ids are allowed to have details appended
13048 to them (separated by colons). This is treated as there would be a static
13049 pre-configured entry in Configure's %table under key <id> with value
13050 <details> and ``perl Configure <id>'' is called. So, when you want to
13051 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
13052 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
13053 now, which overrides the FreeBSD-elf entry on-the-fly.
13054 [Ralf S. Engelschall]
13056 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
13059 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
13060 on the `perl Configure ...' command line. This way one can compile
13061 OpenSSL libraries with Position Independent Code (PIC) which is needed
13062 for linking it into DSOs.
13063 [Ralf S. Engelschall]
13065 *) Remarkably, export ciphers were totally broken and no-one had noticed!
13069 *) Cleaned up the LICENSE document: The official contact for any license
13070 questions now is the OpenSSL core team under openssl-core@openssl.org.
13071 And add a paragraph about the dual-license situation to make sure people
13072 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
13073 to the OpenSSL toolkit.
13074 [Ralf S. Engelschall]
13076 *) General source tree makefile cleanups: Made `making xxx in yyy...'
13077 display consistent in the source tree and replaced `/bin/rm' by `rm'.
13078 Additionally cleaned up the `make links' target: Remove unnecessary
13079 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
13080 to speed processing and no longer clutter the display with confusing
13081 stuff. Instead only the actually done links are displayed.
13082 [Ralf S. Engelschall]
13084 *) Permit null encryption ciphersuites, used for authentication only. It used
13085 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
13086 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
13090 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
13091 signed attributes when verifying signatures (this would break them),
13092 the detached data encoding was wrong and public keys obtained using
13093 X509_get_pubkey() weren't freed.
13096 *) Add text documentation for the BUFFER functions. Also added a work around
13097 to a Win95 console bug. This was triggered by the password read stuff: the
13098 last character typed gets carried over to the next fread(). If you were
13099 generating a new cert request using 'req' for example then the last
13100 character of the passphrase would be CR which would then enter the first
13104 *) Added the new `Includes OpenSSL Cryptography Software' button as
13105 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
13106 button and can be used by applications based on OpenSSL to show the
13107 relationship to the OpenSSL project.
13108 [Ralf S. Engelschall]
13110 *) Remove confusing variables in function signatures in files
13111 ssl/ssl_lib.c and ssl/ssl.h.
13112 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13114 *) Don't install bss_file.c under PREFIX/include/
13115 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13117 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
13118 functions that return function pointers and has support for NT specific
13119 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
13120 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
13121 unsigned to signed types: this was killing the Win32 compile.
13124 *) Add new certificate file to stack functions,
13125 SSL_add_dir_cert_subjects_to_stack() and
13126 SSL_add_file_cert_subjects_to_stack(). These largely supplant
13127 SSL_load_client_CA_file(), and can be used to add multiple certs easily
13128 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
13129 This means that Apache-SSL and similar packages don't have to mess around
13130 to add as many CAs as they want to the preferred list.
13133 *) Experiment with doxygen documentation. Currently only partially applied to
13135 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
13136 openssl.doxy as the configuration file.
13139 *) Get rid of remaining C++-style comments which strict C compilers hate.
13140 [Ralf S. Engelschall, pointed out by Carlos Amengual]
13142 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
13143 compiled in by default: it has problems with large keys.
13146 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
13147 DH private keys and/or callback functions which directly correspond to
13148 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
13149 is needed for applications which have to configure certificates on a
13150 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
13152 For the RSA certificate situation is makes no difference, but
13153 for the DSA certificate situation this fixes the "no shared cipher"
13154 problem where the OpenSSL cipher selection procedure failed because the
13155 temporary keys were not overtaken from the context and the API provided
13156 no way to reconfigure them.
13157 The new functions now let applications reconfigure the stuff and they
13158 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
13159 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
13160 non-public-API function ssl_cert_instantiate() is used as a helper
13161 function and also to reduce code redundancy inside ssl_rsa.c.
13162 [Ralf S. Engelschall]
13164 *) Move s_server -dcert and -dkey options out of the undocumented feature
13165 area because they are useful for the DSA situation and should be
13166 recognized by the users.
13167 [Ralf S. Engelschall]
13169 *) Fix the cipher decision scheme for export ciphers: the export bits are
13170 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
13171 SSL_EXP_MASK. So, the original variable has to be used instead of the
13172 already masked variable.
13173 [Richard Levitte <levitte@stacken.kth.se>]
13175 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
13176 [Richard Levitte <levitte@stacken.kth.se>]
13178 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
13179 from `int' to `unsigned int' because it's a length and initialized by
13180 EVP_DigestFinal() which expects an `unsigned int *'.
13181 [Richard Levitte <levitte@stacken.kth.se>]
13183 *) Don't hard-code path to Perl interpreter on shebang line of Configure
13184 script. Instead use the usual Shell->Perl transition trick.
13185 [Ralf S. Engelschall]
13187 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
13188 (in addition to RSA certificates) to match the behaviour of `openssl dsa
13189 -noout -modulus' as it's already the case for `openssl rsa -noout
13190 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
13191 currently the public key is printed (a decision which was already done by
13192 `openssl dsa -modulus' in the past) which serves a similar purpose.
13193 Additionally the NO_RSA no longer completely removes the whole -modulus
13194 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
13196 [Ralf S. Engelschall]
13198 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
13199 BIO. See the source (crypto/evp/bio_ok.c) for more info.
13200 [Arne Ansper <arne@ats.cyber.ee>]
13202 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
13203 to be added. Now both 'req' and 'ca' can use new objects defined in the
13207 *) Add cool BIO that does syslog (or event log on NT).
13208 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
13210 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
13211 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
13212 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
13213 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
13216 *) Add preliminary config info for new extension code.
13219 *) Make RSA_NO_PADDING really use no padding.
13220 [Ulf Moeller <ulf@fitug.de>]
13222 *) Generate errors when private/public key check is done.
13225 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
13226 for some CRL extensions and new objects added.
13229 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
13230 key usage extension and fuller support for authority key id.
13233 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
13234 padding method for RSA, which is recommended for new applications in PKCS
13235 #1 v2.0 (RFC 2437, October 1998).
13236 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
13237 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
13238 against Bleichbacher's attack on RSA.
13239 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
13242 *) Updates to the new SSL compression code
13243 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13245 *) Fix so that the version number in the master secret, when passed
13246 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
13247 (because the server will not accept higher), that the version number
13248 is 0x03,0x01, not 0x03,0x00
13249 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13251 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
13252 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
13253 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
13256 *) Support for RAW extensions where an arbitrary extension can be
13257 created by including its DER encoding. See apps/openssl.cnf for
13261 *) Make sure latest Perl versions don't interpret some generated C array
13262 code as Perl array code in the crypto/err/err_genc.pl script.
13263 [Lars Weber <3weber@informatik.uni-hamburg.de>]
13265 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
13266 not many people have the assembler. Various Win32 compilation fixes and
13267 update to the INSTALL.W32 file with (hopefully) more accurate Win32
13268 build instructions.
13271 *) Modify configure script 'Configure' to automatically create crypto/date.h
13272 file under Win32 and also build pem.h from pem.org. New script
13273 util/mkfiles.pl to create the MINFO file on environments that can't do a
13274 'make files': perl util/mkfiles.pl >MINFO should work.
13277 *) Major rework of DES function declarations, in the pursuit of correctness
13278 and purity. As a result, many evil casts evaporated, and some weirdness,
13279 too. You may find this causes warnings in your code. Zapping your evil
13280 casts will probably fix them. Mostly.
13283 *) Fix for a typo in asn1.h. Bug fix to object creation script
13284 obj_dat.pl. It considered a zero in an object definition to mean
13285 "end of object": none of the objects in objects.h have any zeros
13286 so it wasn't spotted.
13287 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
13289 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
13290 Masking (CBCM). In the absence of test vectors, the best I have been able
13291 to do is check that the decrypt undoes the encrypt, so far. Send me test
13292 vectors if you have them.
13295 *) Correct calculation of key length for export ciphers (too much space was
13296 allocated for null ciphers). This has not been tested!
13299 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
13300 message is now correct (it understands "crypto" and "ssl" on its
13301 command line). There is also now an "update" option. This will update
13302 the util/ssleay.num and util/libeay.num files with any new functions.
13304 perl util/mkdef.pl crypto ssl update
13305 it will update them.
13308 *) Overhauled the Perl interface (perl/*):
13309 - ported BN stuff to OpenSSL's different BN library
13310 - made the perl/ source tree CVS-aware
13311 - renamed the package from SSLeay to OpenSSL (the files still contain
13312 their history because I've copied them in the repository)
13313 - removed obsolete files (the test scripts will be replaced
13314 by better Test::Harness variants in the future)
13315 [Ralf S. Engelschall]
13317 *) First cut for a very conservative source tree cleanup:
13318 1. merge various obsolete readme texts into doc/ssleay.txt
13319 where we collect the old documents and readme texts.
13320 2. remove the first part of files where I'm already sure that we no
13321 longer need them because of three reasons: either they are just temporary
13322 files which were left by Eric or they are preserved original files where
13323 I've verified that the diff is also available in the CVS via "cvs diff
13324 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
13325 the crypto/md/ stuff).
13326 [Ralf S. Engelschall]
13328 *) More extension code. Incomplete support for subject and issuer alt
13329 name, issuer and authority key id. Change the i2v function parameters
13330 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
13331 what that's for :-) Fix to ASN1 macro which messed up
13332 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
13335 *) Preliminary support for ENUMERATED type. This is largely copied from the
13339 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
13340 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13342 *) Make sure `make rehash' target really finds the `openssl' program.
13343 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13345 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
13346 like to hear about it if this slows down other processors.
13349 *) Add CygWin32 platform information to Configure script.
13350 [Alan Batie <batie@aahz.jf.intel.com>]
13352 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
13353 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13355 *) New program nseq to manipulate netscape certificate sequences
13358 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13362 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13363 but the BN code had some problems that would cause failures when
13364 doing certificate verification and some other functions.
13365 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13367 *) Add ASN1 and PEM code to support netscape certificate sequences.
13370 *) Add ASN1 and PEM code to support netscape certificate sequences.
13373 *) Add several PKIX and private extended key usage OIDs.
13376 *) Modify the 'ca' program to handle the new extension code. Modify
13377 openssl.cnf for new extension format, add comments.
13380 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13381 and add a sample to openssl.cnf so req -x509 now adds appropriate
13385 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13386 error code, add initial support to X509_print() and x509 application.
13389 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13390 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13391 stuff is currently isolated and isn't even compiled yet.
13394 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13395 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13396 Removed the versions check from X509 routines when loading extensions:
13397 this allows certain broken certificates that don't set the version
13398 properly to be processed.
13401 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13402 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13403 can still be regenerated with "make depend".
13406 *) Spelling mistake in C version of CAST-128.
13407 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13409 *) Changes to the error generation code. The perl script err-code.pl
13410 now reads in the old error codes and retains the old numbers, only
13411 adding new ones if necessary. It also only changes the .err files if new
13412 codes are added. The makefiles have been modified to only insert errors
13413 when needed (to avoid needlessly modifying header files). This is done
13414 by only inserting errors if the .err file is newer than the auto generated
13415 C file. To rebuild all the error codes from scratch (the old behaviour)
13416 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13417 or delete all the .err files.
13420 *) CAST-128 was incorrectly implemented for short keys. The C version has
13421 been fixed, but is untested. The assembler versions are also fixed, but
13422 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13423 to regenerate it if needed.
13424 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13425 Hagino <itojun@kame.net>]
13427 *) File was opened incorrectly in randfile.c.
13428 [Ulf Möller <ulf@fitug.de>]
13430 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13431 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13432 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13433 al: it's just almost always a UTCTime. Note this patch adds new error
13434 codes so do a "make errors" if there are problems.
13437 *) Correct Linux 1 recognition in config.
13438 [Ulf Möller <ulf@fitug.de>]
13440 *) Remove pointless MD5 hash when using DSA keys in ca.
13441 [Anonymous <nobody@replay.com>]
13443 *) Generate an error if given an empty string as a cert directory. Also
13444 generate an error if handed NULL (previously returned 0 to indicate an
13445 error, but didn't set one).
13446 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13448 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13451 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13452 parameters. This was causing a warning which killed off the Win32 compile.
13455 *) Remove C++ style comments from crypto/bn/bn_local.h.
13456 [Neil Costigan <neil.costigan@celocom.com>]
13458 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13459 based on a text string, looking up short and long names and finally
13460 "dot" format. The "dot" format stuff didn't work. Added new function
13461 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13462 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13463 OID is not part of the table.
13466 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13467 X509_LOOKUP_by_alias().
13470 *) Sort openssl functions by name.
13473 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13474 encryption from sample DSA keys (in case anyone is interested the password
13478 *) Make _all_ *_free functions accept a NULL pointer.
13479 [Frans Heymans <fheymans@isaserver.be>]
13481 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13483 [Anonymous <nobody@replay.com>]
13485 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13486 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13488 *) Don't blow it for numeric -newkey arguments to apps/req.
13489 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13491 *) Temp key "for export" tests were wrong in s3_srvr.c.
13492 [Anonymous <nobody@replay.com>]
13494 *) Add prototype for temp key callback functions
13495 SSL_CTX_set_tmp_{rsa,dh}_callback().
13498 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13499 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13502 *) X509_name_add_entry() freed the wrong thing after an error.
13503 [Arne Ansper <arne@ats.cyber.ee>]
13505 *) rsa_eay.c would attempt to free a NULL context.
13506 [Arne Ansper <arne@ats.cyber.ee>]
13508 *) BIO_s_socket() had a broken should_retry() on Windoze.
13509 [Arne Ansper <arne@ats.cyber.ee>]
13511 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13512 [Arne Ansper <arne@ats.cyber.ee>]
13514 *) Make sure the already existing X509_STORE->depth variable is initialized
13515 in X509_STORE_new(), but document the fact that this variable is still
13516 unused in the certificate verification process.
13517 [Ralf S. Engelschall]
13519 *) Fix the various library and apps files to free up pkeys obtained from
13520 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13523 *) Fix reference counting in X509_PUBKEY_get(). This makes
13524 demos/maurice/example2.c work, amongst others, probably.
13525 [Steve Henson and Ben Laurie]
13527 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13528 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13529 are no longer created. This way we have a single and consistent command
13530 line interface `openssl <command>', similar to `cvs <command>'.
13531 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13533 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13534 BIT STRING wrapper always have zero unused bits.
13537 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13540 *) Make the top-level INSTALL documentation easier to understand.
13543 *) Makefiles updated to exit if an error occurs in a sub-directory
13544 make (including if user presses ^C) [Paul Sutton]
13546 *) Make Montgomery context stuff explicit in RSA data structure.
13549 *) Fix build order of pem and err to allow for generated pem.h.
13552 *) Fix renumbering bug in X509_NAME_delete_entry().
13555 *) Enhanced the err-ins.pl script so it makes the error library number
13556 global and can add a library name. This is needed for external ASN1 and
13557 other error libraries.
13560 *) Fixed sk_insert which never worked properly.
13563 *) Fix ASN1 macros so they can handle indefinite length constructed
13564 EXPLICIT tags. Some non standard certificates use these: they can now
13568 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13569 into a single doc/ssleay.txt bundle. This way the information is still
13570 preserved but no longer messes up this directory. Now it's new room for
13571 the new set of documentation files.
13572 [Ralf S. Engelschall]
13574 *) SETs were incorrectly DER encoded. This was a major pain, because they
13575 shared code with SEQUENCEs, which aren't coded the same. This means that
13576 almost everything to do with SETs or SEQUENCEs has either changed name or
13577 number of arguments.
13578 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13580 *) Fix test data to work with the above.
13583 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13584 was already fixed by Eric for 0.9.1 it seems.
13585 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13587 *) Autodetect FreeBSD3.
13590 *) Fix various bugs in Configure. This affects the following platforms:
13594 unixware-2.0-pentium
13598 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13599 before they are needed.
13602 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13606 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13608 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13609 changed SSLeay to OpenSSL in version strings.
13610 [Ralf S. Engelschall]
13612 *) Some fixups to the top-level documents.
13615 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13616 because the symlink to include/ was missing.
13617 [Ralf S. Engelschall]
13619 *) Incorporated the popular no-RSA/DSA-only patches
13620 which allow to compile a RSA-free SSLeay.
13621 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13623 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13624 when "ssleay" is still not found.
13625 [Ralf S. Engelschall]
13627 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13628 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13630 *) Updated the README file.
13631 [Ralf S. Engelschall]
13633 *) Added various .cvsignore files in the CVS repository subdirs
13634 to make a "cvs update" really silent.
13635 [Ralf S. Engelschall]
13637 *) Recompiled the error-definition header files and added
13638 missing symbols to the Win32 linker tables.
13639 [Ralf S. Engelschall]
13641 *) Cleaned up the top-level documents;
13642 o new files: CHANGES and LICENSE
13643 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13644 o merged COPYRIGHT into LICENSE
13645 o removed obsolete TODO file
13646 o renamed MICROSOFT to INSTALL.W32
13647 [Ralf S. Engelschall]
13649 *) Removed dummy files from the 0.9.1b source tree:
13650 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13651 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13652 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13653 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13654 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13655 [Ralf S. Engelschall]
13657 *) Added various platform portability fixes.
13660 *) The Genesis of the OpenSSL rpject:
13661 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13662 Young and Tim J. Hudson created while they were working for C2Net until
13664 [The OpenSSL Project]
13667 Changes between 0.9.0b and 0.9.1b [not released]
13669 *) Updated a few CA certificates under certs/
13672 *) Changed some BIGNUM api stuff.
13675 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13676 DGUX x86, Linux Alpha, etc.
13679 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13680 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13684 *) Add -strparse option to asn1pars program which parses nested
13686 [Dr Stephen Henson <shenson@bigfoot.com>]
13688 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13691 *) DSA fix for "ca" program.
13694 *) Added "-genkey" option to "dsaparam" program.
13697 *) Added RIPE MD160 (rmd160) message digest.
13700 *) Added -a (all) option to "ssleay version" command.
13703 *) Added PLATFORM define which is the id given to Configure.
13706 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13709 *) Extended the ASN.1 parser routines.
13712 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13715 *) Added a BN_CTX to the BN library.
13718 *) Fixed the weak key values in DES library
13721 *) Changed API in EVP library for cipher aliases.
13724 *) Added support for RC2/64bit cipher.
13727 *) Converted the lhash library to the crypto/mem.c functions.
13730 *) Added more recognized ASN.1 object ids.
13733 *) Added more RSA padding checks for SSL/TLS.
13736 *) Added BIO proxy/filter functionality.
13739 *) Added extra_certs to SSL_CTX which can be used
13740 send extra CA certificates to the client in the CA cert chain sending
13741 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13744 *) Now Fortezza is denied in the authentication phase because
13745 this is key exchange mechanism is not supported by SSLeay at all.
13748 *) Additional PKCS1 checks.
13751 *) Support the string "TLSv1" for all TLS v1 ciphers.
13754 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13755 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13758 *) Fixed a few memory leaks.
13761 *) Fixed various code and comment typos.
13764 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13765 bytes sent in the client random.
13766 [Edward Bishop <ebishop@spyglass.com>]