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 1.1.1m and 1.1.1n [15 Mar 2022]
12 *) Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
15 Internally this function is used when parsing certificates that contain
16 elliptic curve public keys in compressed form or explicit elliptic curve
17 parameters with a base point encoded in compressed form.
19 It is possible to trigger the infinite loop by crafting a certificate that
20 has invalid explicit curve parameters.
22 Since certificate parsing happens prior to verification of the certificate
23 signature, any process that parses an externally supplied certificate may
24 thus be subject to a denial of service attack. The infinite loop can also
25 be reached when parsing crafted private keys as they can contain explicit
26 elliptic curve parameters.
28 Thus vulnerable situations include:
30 - TLS clients consuming server certificates
31 - TLS servers consuming client certificates
32 - Hosting providers taking certificates or private keys from customers
33 - Certificate authorities parsing certification requests from subscribers
34 - Anything else which parses ASN.1 elliptic curve parameters
36 Also any other applications that use the BN_mod_sqrt() where the attacker
37 can control the parameter values are vulnerable to this DoS issue.
41 *) Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
42 to the list of ciphersuites providing Perfect Forward Secrecy as
43 required by SECLEVEL >= 3.
45 [Dmitry Belyavskiy, Nicola Tuveri]
47 Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
49 *) Avoid loading of a dynamic engine twice.
53 *) Fixed building on Debian with kfreebsd kernels
57 *) Prioritise DANE TLSA issuer certs over peer certs
61 *) Fixed random API for MacOS prior to 10.12
63 These MacOS versions don't support the CommonCrypto APIs
67 Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
69 *) Fixed an SM2 Decryption Buffer Overflow.
71 In order to decrypt SM2 encrypted data an application is expected to call the
72 API function EVP_PKEY_decrypt(). Typically an application will call this
73 function twice. The first time, on entry, the "out" parameter can be NULL and,
74 on exit, the "outlen" parameter is populated with the buffer size required to
75 hold the decrypted plaintext. The application can then allocate a sufficiently
76 sized buffer and call EVP_PKEY_decrypt() again, but this time passing a non-NULL
77 value for the "out" parameter.
79 A bug in the implementation of the SM2 decryption code means that the
80 calculation of the buffer size required to hold the plaintext returned by the
81 first call to EVP_PKEY_decrypt() can be smaller than the actual size required by
82 the second call. This can lead to a buffer overflow when EVP_PKEY_decrypt() is
83 called by the application a second time with a buffer that is too small.
85 A malicious attacker who is able present SM2 content for decryption to an
86 application could cause attacker chosen data to overflow the buffer by up to a
87 maximum of 62 bytes altering the contents of other data held after the
88 buffer, possibly changing application behaviour or causing the application to
89 crash. The location of the buffer is application dependent but is typically
94 *) Fixed various read buffer overruns processing ASN.1 strings
96 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
97 structure which contains a buffer holding the string data and a field holding
98 the buffer length. This contrasts with normal C strings which are repesented as
99 a buffer for the string data which is terminated with a NUL (0) byte.
101 Although not a strict requirement, ASN.1 strings that are parsed using OpenSSL's
102 own "d2i" functions (and other similar parsing functions) as well as any string
103 whose value has been set with the ASN1_STRING_set() function will additionally
104 NUL terminate the byte array in the ASN1_STRING structure.
106 However, it is possible for applications to directly construct valid ASN1_STRING
107 structures which do not NUL terminate the byte array by directly setting the
108 "data" and "length" fields in the ASN1_STRING array. This can also happen by
109 using the ASN1_STRING_set0() function.
111 Numerous OpenSSL functions that print ASN.1 data have been found to assume that
112 the ASN1_STRING byte array will be NUL terminated, even though this is not
113 guaranteed for strings that have been directly constructed. Where an application
114 requests an ASN.1 structure to be printed, and where that ASN.1 structure
115 contains ASN1_STRINGs that have been directly constructed by the application
116 without NUL terminating the "data" field, then a read buffer overrun can occur.
118 The same thing can also occur during name constraints processing of certificates
119 (for example if a certificate has been directly constructed by the application
120 instead of loading it via the OpenSSL parsing functions, and the certificate
121 contains non NUL terminated ASN1_STRING structures). It can also occur in the
122 X509_get1_email(), X509_REQ_get1_email() and X509_get1_ocsp() functions.
124 If a malicious actor can cause an application to directly construct an
125 ASN1_STRING and then process it through one of the affected OpenSSL functions
126 then this issue could be hit. This might result in a crash (causing a Denial of
127 Service attack). It could also result in the disclosure of private memory
128 contents (such as private keys, or sensitive plaintext).
132 Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
134 *) Fixed a problem with verifying a certificate chain when using the
135 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks
136 of the certificates present in a certificate chain. It is not set by
139 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
140 the chain that have explicitly encoded elliptic curve parameters was added
141 as an additional strict check.
143 An error in the implementation of this check meant that the result of a
144 previous check to confirm that certificates in the chain are valid CA
145 certificates was overwritten. This effectively bypasses the check
146 that non-CA certificates must not be able to issue other certificates.
148 If a "purpose" has been configured then there is a subsequent opportunity
149 for checks that the certificate is a valid CA. All of the named "purpose"
150 values implemented in libcrypto perform this check. Therefore, where
151 a purpose is set the certificate chain will still be rejected even when the
152 strict flag has been used. A purpose is set by default in libssl client and
153 server certificate verification routines, but it can be overridden or
154 removed by an application.
156 In order to be affected, an application must explicitly set the
157 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
158 for the certificate verification or, in the case of TLS client or server
159 applications, override the default purpose.
163 *) Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
164 crafted renegotiation ClientHello message from a client. If a TLSv1.2
165 renegotiation ClientHello omits the signature_algorithms extension (where
166 it was present in the initial ClientHello), but includes a
167 signature_algorithms_cert extension then a NULL pointer dereference will
168 result, leading to a crash and a denial of service attack.
170 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
171 (which is the default configuration). OpenSSL TLS clients are not impacted
174 [Peter Kästle and Samuel Sapalski]
176 Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
178 *) Fixed the X509_issuer_and_serial_hash() function. It attempts to
179 create a unique hash value based on the issuer and serial number data
180 contained within an X509 certificate. However it was failing to correctly
181 handle any errors that may occur while parsing the issuer field (which might
182 occur if the issuer field is maliciously constructed). This may subsequently
183 result in a NULL pointer deref and a crash leading to a potential denial of
188 *) Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
189 padding mode to correctly check for rollback attacks. This is considered a
190 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
194 *) Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
195 functions. Previously they could overflow the output length argument in some
196 cases where the input length is close to the maximum permissable length for
197 an integer on the platform. In such cases the return value from the function
198 call would be 1 (indicating success), but the output length value would be
199 negative. This could cause applications to behave incorrectly or crash.
203 *) Fixed SRP_Calc_client_key so that it runs in constant time. The previous
204 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
205 could be exploited in a side channel attack to recover the password. Since
206 the attack is local host only this is outside of the current OpenSSL
207 threat model and therefore no CVE is assigned.
209 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
213 Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
215 *) Fixed NULL pointer deref in the GENERAL_NAME_cmp function
216 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
217 If an attacker can control both items being compared then this could lead
218 to a possible denial of service attack. OpenSSL itself uses the
219 GENERAL_NAME_cmp function for two purposes:
220 1) Comparing CRL distribution point names between an available CRL and a
221 CRL distribution point embedded in an X509 certificate
222 2) When verifying that a timestamp response token signer matches the
223 timestamp authority name (exposed via the API functions
224 TS_RESP_verify_response and TS_RESP_verify_token)
228 *) Add support for Apple Silicon M1 Macs with the darwin64-arm64-cc target.
231 *) The security callback, which can be customised by application code, supports
232 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
233 in the "other" parameter. In most places this is what is passed. All these
234 places occur server side. However there was one client side call of this
235 security operation and it passed a DH object instead. This is incorrect
236 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
237 of the other locations. Therefore this client side call has been changed to
238 pass an EVP_PKEY instead.
241 *) In 1.1.1h, an expired trusted (root) certificate was not anymore rejected
242 when validating a certificate path. This check is restored in 1.1.1i.
245 Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
247 *) Certificates with explicit curve parameters are now disallowed in
248 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
251 *) The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
252 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
253 conversely, silently ignore DTLS protocol version bounds when configuring
254 TLS-based contexts. The commands can be repeated to set bounds of both
255 types. The same applies with the corresponding "min_protocol" and
256 "max_protocol" command-line switches, in case some application uses both TLS
259 SSL_CTX instances that are created for a fixed protocol version (e.g.
260 TLSv1_server_method()) also silently ignore version bounds. Previously
261 attempts to apply bounds to these protocol versions would result in an
262 error. Now only the "version-flexible" SSL_CTX instances are subject to
263 limits in configuration files in command-line options.
266 *) Handshake now fails if Extended Master Secret extension is dropped
270 *) Accidentally, an expired trusted (root) certificate is not anymore rejected
271 when validating a certificate path.
274 *) The Oracle Developer Studio compiler will start reporting deprecated APIs
276 Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
278 *) Fixed segmentation fault in SSL_check_chain()
279 Server or client applications that call the SSL_check_chain() function
280 during or after a TLS 1.3 handshake may crash due to a NULL pointer
281 dereference as a result of incorrect handling of the
282 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
283 or unrecognised signature algorithm is received from the peer. This could
284 be exploited by a malicious peer in a Denial of Service attack.
288 *) Added AES consttime code for no-asm configurations
289 an optional constant time support for AES was added
290 when building openssl for no-asm.
291 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
292 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
293 At this time this feature is by default disabled.
294 It will be enabled by default in 3.0.
297 Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
299 *) Revert the change of EOF detection while reading in libssl to avoid
300 regressions in applications depending on the current way of reporting
301 the EOF. As the existing method is not fully accurate the change to
302 reporting the EOF via SSL_ERROR_SSL is kept on the current development
303 branch and will be present in the 3.0 release.
306 *) Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
307 when primes for RSA keys are computed.
308 Since we previously always generated primes == 2 (mod 3) for RSA keys,
309 the 2-prime and 3-prime RSA modules were easy to distinguish, since
310 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
311 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
312 This avoids possible fingerprinting of newly generated RSA modules.
315 Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
316 *) Properly detect EOF while reading in libssl. Previously if we hit an EOF
317 while reading in libssl then we would report an error back to the
318 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
319 an error to the stack (which means we instead return SSL_ERROR_SSL) and
320 therefore give a hint as to what went wrong.
323 *) Check that ed25519 and ed448 are allowed by the security level. Previously
324 signature algorithms not using an MD were not being checked that they were
325 allowed by the security level.
328 *) Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
329 was not quite right. The behaviour was not consistent between resumption
330 and normal handshakes, and also not quite consistent with historical
331 behaviour. The behaviour in various scenarios has been clarified and
332 it has been updated to make it match historical behaviour as closely as
336 *) [VMS only] The header files that the VMS compilers include automatically,
337 __DECC_INCLUDE_PROLOGUE.H and __DECC_INCLUDE_EPILOGUE.H, use pragmas that
338 the C++ compiler doesn't understand. This is a shortcoming in the
339 compiler, but can be worked around with __cplusplus guards.
341 C++ applications that use OpenSSL libraries must be compiled using the
342 qualifier '/NAMES=(AS_IS,SHORTENED)' to be able to use all the OpenSSL
343 functions. Otherwise, only functions with symbols of less than 31
344 characters can be used, as the linker will not be able to successfully
345 resolve symbols with longer names.
348 *) Corrected the documentation of the return values from the EVP_DigestSign*
349 set of functions. The documentation mentioned negative values for some
350 errors, but this was never the case, so the mention of negative values
353 Code that followed the documentation and thereby check with something
354 like 'EVP_DigestSignInit(...) <= 0' will continue to work undisturbed.
357 *) Fixed an an overflow bug in the x64_64 Montgomery squaring procedure
358 used in exponentiation with 512-bit moduli. No EC algorithms are
359 affected. Analysis suggests that attacks against 2-prime RSA1024,
360 3-prime RSA1536, and DSA1024 as a result of this defect would be very
361 difficult to perform and are not believed likely. Attacks against DH512
362 are considered just feasible. However, for an attack the target would
363 have to re-use the DH512 private key, which is not recommended anyway.
364 Also applications directly using the low level API BN_mod_exp may be
365 affected if they use BN_FLG_CONSTTIME.
369 *) Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
370 The presence of this system service is determined at run-time.
373 *) Added newline escaping functionality to a filename when using openssl dgst.
374 This output format is to replicate the output format found in the '*sum'
375 checksum programs. This aims to preserve backward compatibility.
376 [Matt Eaton, Richard Levitte, and Paul Dale]
378 *) Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
382 Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
384 *) Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
385 number generator (RNG). This was intended to include protection in the
386 event of a fork() system call in order to ensure that the parent and child
387 processes did not share the same RNG state. However this protection was not
388 being used in the default case.
390 A partial mitigation for this issue is that the output from a high
391 precision timer is mixed into the RNG state so the likelihood of a parent
392 and child process sharing state is significantly reduced.
394 If an application already calls OPENSSL_init_crypto() explicitly using
395 OPENSSL_INIT_ATFORK then this problem does not occur at all.
397 [Matthias St. Pierre]
399 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is
400 used even when parsing explicit parameters, when loading a serialized key
401 or calling `EC_GROUP_new_from_ecpkparameters()`/
402 `EC_GROUP_new_from_ecparameters()`.
403 This prevents bypass of security hardening and performance gains,
404 especially for curves with specialized EC_METHODs.
405 By default, if a key encoded with explicit parameters is loaded and later
406 serialized, the output is still encoded with explicit parameters, even if
407 internally a "named" EC_GROUP is used for computation.
410 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before
411 this change, EC_GROUP_set_generator would accept order and/or cofactor as
412 NULL. After this change, only the cofactor parameter can be NULL. It also
413 does some minimal sanity checks on the passed order.
417 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
418 An attack is simple, if the first CMS_recipientInfo is valid but the
419 second CMS_recipientInfo is chosen ciphertext. If the second
420 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
421 encryption key will be replaced by garbage, and the message cannot be
422 decoded, but if the RSA decryption fails, the correct encryption key is
423 used and the recipient will not notice the attack.
424 As a work around for this potential attack the length of the decrypted
425 key must be equal to the cipher default key length, in case the
426 certifiate is not given and all recipientInfo are tried out.
427 The old behaviour can be re-enabled in the CMS code by setting the
428 CMS_DEBUG_DECRYPT flag.
432 *) Early start up entropy quality from the DEVRANDOM seed source has been
433 improved for older Linux systems. The RAND subsystem will wait for
434 /dev/random to be producing output before seeding from /dev/urandom.
435 The seeded state is stored for future library initialisations using
436 a system global shared memory segment. The shared memory identifier
437 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
438 the desired value. The default identifier is 114.
441 *) Correct the extended master secret constant on EBCDIC systems. Without this
442 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
443 negotiate EMS will fail. Unfortunately this also means that TLS connections
444 between EBCDIC systems with this fix, and EBCDIC systems without this
445 fix will fail if they negotiate EMS.
448 *) Use Windows installation paths in the mingw builds
450 Mingw isn't a POSIX environment per se, which means that Windows
451 paths should be used for installation.
455 *) Changed DH_check to accept parameters with order q and 2q subgroups.
456 With order 2q subgroups the bit 0 of the private key is not secret
457 but DH_generate_key works around that by clearing bit 0 of the
458 private key for those. This avoids leaking bit 0 of the private key.
461 *) Significantly reduce secure memory usage by the randomness pools.
464 *) Revert the DEVRANDOM_WAIT feature for Linux systems
466 The DEVRANDOM_WAIT feature added a select() call to wait for the
467 /dev/random device to become readable before reading from the
470 It turned out that this change had negative side effects on
471 performance which were not acceptable. After some discussion it
472 was decided to revert this feature and leave it up to the OS
473 resp. the platform maintainer to ensure a proper initialization
474 during early boot time.
475 [Matthias St. Pierre]
477 Changes between 1.1.1b and 1.1.1c [28 May 2019]
479 *) Add build tests for C++. These are generated files that only do one
480 thing, to include one public OpenSSL head file each. This tests that
481 the public header files can be usefully included in a C++ application.
483 This test isn't enabled by default. It can be enabled with the option
484 'enable-buildtest-c++'.
487 *) Enable SHA3 pre-hashing for ECDSA and DSA.
490 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
491 This changes the size when using the genpkey app when no size is given. It
492 fixes an omission in earlier changes that changed all RSA, DSA and DH
493 generation apps to use 2048 bits by default.
496 *) Reorganize the manual pages to consistently have RETURN VALUES,
497 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
498 util/fix-doc-nits accordingly.
499 [Paul Yang, Joshua Lock]
501 *) Add the missing accessor EVP_PKEY_get0_engine()
504 *) Have apps like 's_client' and 's_server' output the signature scheme
505 along with other cipher suite parameters when debugging.
508 *) Make OPENSSL_config() error agnostic again.
511 *) Do the error handling in RSA decryption constant time.
514 *) Prevent over long nonces in ChaCha20-Poly1305.
516 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
517 for every encryption operation. RFC 7539 specifies that the nonce value
518 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
519 and front pads the nonce with 0 bytes if it is less than 12
520 bytes. However it also incorrectly allows a nonce to be set of up to 16
521 bytes. In this case only the last 12 bytes are significant and any
522 additional leading bytes are ignored.
524 It is a requirement of using this cipher that nonce values are
525 unique. Messages encrypted using a reused nonce value are susceptible to
526 serious confidentiality and integrity attacks. If an application changes
527 the default nonce length to be longer than 12 bytes and then makes a
528 change to the leading bytes of the nonce expecting the new value to be a
529 new unique nonce then such an application could inadvertently encrypt
530 messages with a reused nonce.
532 Additionally the ignored bytes in a long nonce are not covered by the
533 integrity guarantee of this cipher. Any application that relies on the
534 integrity of these ignored leading bytes of a long nonce may be further
535 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
536 is safe because no such use sets such a long nonce value. However user
537 applications that use this cipher directly and set a non-default nonce
538 length to be longer than 12 bytes may be vulnerable.
540 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
545 *) Add DEVRANDOM_WAIT feature for Linux systems
547 On older Linux systems where the getrandom() system call is not available,
548 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
549 Contrary to getrandom(), the /dev/urandom device will not block during
550 early boot when the kernel CSPRNG has not been seeded yet.
552 To mitigate this known weakness, use select() to wait for /dev/random to
553 become readable before reading from /dev/urandom.
555 *) Ensure that SM2 only uses SM3 as digest algorithm
558 Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
560 *) Added SCA hardening for modular field inversion in EC_GROUP through
561 a new dedicated field_inv() pointer in EC_METHOD.
562 This also addresses a leakage affecting conversions from projective
563 to affine coordinates.
564 [Billy Bob Brumley, Nicola Tuveri]
566 *) Change the info callback signals for the start and end of a post-handshake
567 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
568 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
569 confused by this and assume that a TLSv1.2 renegotiation has started. This
570 can break KeyUpdate handling. Instead we no longer signal the start and end
571 of a post handshake message exchange (although the messages themselves are
572 still signalled). This could break some applications that were expecting
573 the old signals. However without this KeyUpdate is not usable for many
577 *) Fix a bug in the computation of the endpoint-pair shared secret used
578 by DTLS over SCTP. This breaks interoperability with older versions
579 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
580 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
581 interoperability with such broken implementations. However, enabling
582 this switch breaks interoperability with correct implementations.
584 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
585 re-used X509_PUBKEY object if the second PUBKEY is malformed.
588 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
591 *) Remove the 'dist' target and add a tarball building script. The
592 'dist' target has fallen out of use, and it shouldn't be
593 necessary to configure just to create a source distribution.
596 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
598 *) Timing vulnerability in DSA signature generation
600 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
601 timing side channel attack. An attacker could use variations in the signing
602 algorithm to recover the private key.
604 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
608 *) Timing vulnerability in ECDSA signature generation
610 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
611 timing side channel attack. An attacker could use variations in the signing
612 algorithm to recover the private key.
614 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
618 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
619 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
620 are retained for backwards compatibility.
623 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
624 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
625 of two gigabytes and the error handling improved.
627 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
628 categorized as a normal bug, not a security issue, because the DRBG reseeds
629 automatically and is fully functional even without additional randomness
630 provided by the application.
632 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
634 *) Add a new ClientHello callback. Provides a callback interface that gives
635 the application the ability to adjust the nascent SSL object at the
636 earliest stage of ClientHello processing, immediately after extensions have
637 been collected but before they have been processed. In particular, this
638 callback can adjust the supported TLS versions in response to the contents
642 *) Add SM2 base algorithm support.
645 *) s390x assembly pack: add (improved) hardware-support for the following
646 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
647 aes-cfb/cfb8, aes-ecb.
650 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
651 parameter is no longer accepted, as it leads to a corrupt table. NULL
652 pem_str is reserved for alias entries only.
655 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
656 step for prime curves. The new implementation is based on formulae from
657 differential addition-and-doubling in homogeneous projective coordinates
658 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
659 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
660 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
661 to work in projective coordinates.
662 [Billy Bob Brumley, Nicola Tuveri]
664 *) Change generating and checking of primes so that the error rate of not
665 being prime depends on the intended use based on the size of the input.
666 For larger primes this will result in more rounds of Miller-Rabin.
667 The maximal error rate for primes with more than 1080 bits is lowered
669 [Kurt Roeckx, Annie Yousar]
671 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
674 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
675 moving between systems, and to avoid confusion when a Windows build is
676 done with mingw vs with MSVC. For POSIX installs, there's still a
677 symlink or copy named 'tsget' to avoid that confusion as well.
680 *) Revert blinding in ECDSA sign and instead make problematic addition
681 length-invariant. Switch even to fixed-length Montgomery multiplication.
684 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
685 step for binary curves. The new implementation is based on formulae from
686 differential addition-and-doubling in mixed Lopez-Dahab projective
687 coordinates, modified to independently blind the operands.
688 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
690 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
691 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
692 EC_METHODs to implement their own specialized "ladder step", to take
693 advantage of more favorable coordinate systems or more efficient
694 differential addition-and-doubling algorithms.
695 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
697 *) Modified the random device based seed sources to keep the relevant
698 file descriptors open rather than reopening them on each access.
699 This allows such sources to operate in a chroot() jail without
700 the associated device nodes being available. This behaviour can be
701 controlled using RAND_keep_random_devices_open().
704 *) Numerous side-channel attack mitigations have been applied. This may have
705 performance impacts for some algorithms for the benefit of improved
706 security. Specific changes are noted in this change log by their respective
710 *) AIX shared library support overhaul. Switch to AIX "natural" way of
711 handling shared libraries, which means collecting shared objects of
712 different versions and bitnesses in one common archive. This allows to
713 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
714 doesn't affect the way 3rd party applications are linked, only how
715 multi-version installation is managed.
718 *) Make ec_group_do_inverse_ord() more robust and available to other
719 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
720 mitigations are applied to the fallback BN_mod_inverse().
721 When using this function rather than BN_mod_inverse() directly, new
722 EC cryptosystem implementations are then safer-by-default.
725 *) Add coordinate blinding for EC_POINT and implement projective
726 coordinate blinding for generic prime curves as a countermeasure to
727 chosen point SCA attacks.
728 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
730 *) Add blinding to ECDSA and DSA signatures to protect against side channel
731 attacks discovered by Keegan Ryan (NCC Group).
734 *) Enforce checking in the pkeyutl command line app to ensure that the input
735 length does not exceed the maximum supported digest length when performing
736 a sign, verify or verifyrecover operation.
739 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
740 I/O in combination with something like select() or poll() will hang. This
741 can be turned off again using SSL_CTX_clear_mode().
742 Many applications do not properly handle non-application data records, and
743 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
744 around the problems in those applications, but can also break some.
745 It's recommended to read the manpages about SSL_read(), SSL_write(),
746 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
747 SSL_CTX_set_read_ahead() again.
750 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
751 now allow empty (zero character) pass phrases.
754 *) Apply blinding to binary field modular inversion and remove patent
755 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
758 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
759 binary and prime elliptic curves.
762 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
763 constant time fixed point multiplication.
766 *) Revise elliptic curve scalar multiplication with timing attack
767 defenses: ec_wNAF_mul redirects to a constant time implementation
768 when computing fixed point and variable point multiplication (which
769 in OpenSSL are mostly used with secret scalars in keygen, sign,
770 ECDH derive operations).
771 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
774 *) Updated CONTRIBUTING
777 *) Updated DRBG / RAND to request nonce and additional low entropy
778 randomness from the system.
779 [Matthias St. Pierre]
781 *) Updated 'openssl rehash' to use OpenSSL consistent default.
784 *) Moved the load of the ssl_conf module to libcrypto, which helps
785 loading engines that libssl uses before libssl is initialised.
788 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
791 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
792 [Ingo Schwarze, Rich Salz]
794 *) Added output of accepting IP address and port for 'openssl s_server'
797 *) Added a new API for TLSv1.3 ciphersuites:
798 SSL_CTX_set_ciphersuites()
799 SSL_set_ciphersuites()
802 *) Memory allocation failures consistently add an error to the error
806 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
807 in libcrypto when run as setuid/setgid.
810 *) Load any config file by default when libssl is used.
813 *) Added new public header file <openssl/rand_drbg.h> and documentation
814 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
815 [Matthias St. Pierre]
817 *) QNX support removed (cannot find contributors to get their approval
818 for the license change).
821 *) TLSv1.3 replay protection for early data has been implemented. See the
822 SSL_read_early_data() man page for further details.
825 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
826 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
827 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
828 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
829 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
830 configuration has been separated out. See the ciphers man page or the
831 SSL_CTX_set_ciphersuites() man page for more information.
834 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
835 in responder mode now supports the new "-multi" option, which
836 spawns the specified number of child processes to handle OCSP
837 requests. The "-timeout" option now also limits the OCSP
838 responder's patience to wait to receive the full client request
839 on a newly accepted connection. Child processes are respawned
840 as needed, and the CA index file is automatically reloaded
841 when changed. This makes it possible to run the "ocsp" responder
842 as a long-running service, making the OpenSSL CA somewhat more
843 feature-complete. In this mode, most diagnostic messages logged
844 after entering the event loop are logged via syslog(3) rather than
848 *) Added support for X448 and Ed448. Heavily based on original work by
852 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
853 objects loaded. This adds the functions OSSL_STORE_expect() and
854 OSSL_STORE_find() as well as needed tools to construct searches and
855 get the search data out of them.
858 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
859 version of OpenSSL should review their configuration settings to ensure
860 that they are still appropriate for TLSv1.3. For further information see:
861 https://wiki.openssl.org/index.php/TLS1.3
864 *) Grand redesign of the OpenSSL random generator
866 The default RAND method now utilizes an AES-CTR DRBG according to
867 NIST standard SP 800-90Ar1. The new random generator is essentially
868 a port of the default random generator from the OpenSSL FIPS 2.0
869 object module. It is a hybrid deterministic random bit generator
870 using an AES-CTR bit stream and which seeds and reseeds itself
871 automatically using trusted system entropy sources.
873 Some of its new features are:
874 o Support for multiple DRBG instances with seed chaining.
875 o The default RAND method makes use of a DRBG.
876 o There is a public and private DRBG instance.
877 o The DRBG instances are fork-safe.
878 o Keep all global DRBG instances on the secure heap if it is enabled.
879 o The public and private DRBG instance are per thread for lock free
881 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
883 *) Changed Configure so it only says what it does and doesn't dump
884 so much data. Instead, ./configdata.pm should be used as a script
885 to display all sorts of configuration data.
888 *) Added processing of "make variables" to Configure.
891 *) Added SHA512/224 and SHA512/256 algorithm support.
894 *) The last traces of Netware support, first removed in 1.1.0, have
898 *) Get rid of Makefile.shared, and in the process, make the processing
899 of certain files (rc.obj, or the .def/.map/.opt files produced from
900 the ordinal files) more visible and hopefully easier to trace and
901 debug (or make silent).
904 *) Make it possible to have environment variable assignments as
905 arguments to config / Configure.
908 *) Add multi-prime RSA (RFC 8017) support.
911 *) Add SM3 implemented according to GB/T 32905-2016
912 [ Jack Lloyd <jack.lloyd@ribose.com>,
913 Ronald Tse <ronald.tse@ribose.com>,
914 Erick Borsboom <erick.borsboom@ribose.com> ]
916 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
917 as documented in RFC6066.
918 Based on a patch from Tomasz Moń
919 [Filipe Raimundo da Silva]
921 *) Add SM4 implemented according to GB/T 32907-2016.
922 [ Jack Lloyd <jack.lloyd@ribose.com>,
923 Ronald Tse <ronald.tse@ribose.com>,
924 Erick Borsboom <erick.borsboom@ribose.com> ]
926 *) Reimplement -newreq-nodes and ERR_error_string_n; the
927 original author does not agree with the license change.
930 *) Add ARIA AEAD TLS support.
933 *) Some macro definitions to support VS6 have been removed. Visual
934 Studio 6 has not worked since 1.1.0
937 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
938 without clearing the errors.
941 *) Add "atfork" functions. If building on a system that without
942 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
943 requirements. The RAND facility now uses/requires this.
949 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
950 not possible to disable entirely. However, it's still possible to
951 disable the console reading UI method, UI_OpenSSL() (use UI_null()
954 To disable, configure with 'no-ui-console'. 'no-ui' is still
955 possible to use as an alias. Check at compile time with the
956 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
957 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
960 *) Add a STORE module, which implements a uniform and URI based reader of
961 stores that can contain keys, certificates, CRLs and numerous other
962 objects. The main API is loosely based on a few stdio functions,
963 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
964 OSSL_STORE_error and OSSL_STORE_close.
965 The implementation uses backends called "loaders" to implement arbitrary
966 URI schemes. There is one built in "loader" for the 'file' scheme.
969 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
970 then adjusted to work on FreeBSD 8.4 as well.
971 Enable by configuring with 'enable-devcryptoeng'. This is done by default
972 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
975 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
976 util/mkerr.pl, which is adapted to allow those prefixes, leading to
977 error code calls like this:
979 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
981 With this change, we claim the namespaces OSSL and OPENSSL in a manner
982 that can be encoded in C. For the foreseeable future, this will only
984 [Richard Levitte and Tim Hudson]
986 *) Removed BSD cryptodev engine.
989 *) Add a build target 'build_all_generated', to build all generated files
990 and only that. This can be used to prepare everything that requires
991 things like perl for a system that lacks perl and then move everything
992 to that system and do the rest of the build there.
995 *) In the UI interface, make it possible to duplicate the user data. This
996 can be used by engines that need to retain the data for a longer time
997 than just the call where this user data is passed.
1000 *) Ignore the '-named_curve auto' value for compatibility of applications
1002 [Tomas Mraz <tmraz@fedoraproject.org>]
1004 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1005 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1006 alerts across multiple records (some of which could be empty). In practice
1007 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1008 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1009 support this at all. Supporting it adds significant complexity to the
1010 record layer, and its removal is unlikely to cause interoperability
1014 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1015 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1016 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1020 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1021 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1022 [Richard Levitte, Andy Polyakov]
1024 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1028 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1029 platform rather than 'mingw'.
1032 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
1033 success if they are asked to add an object which already exists
1034 in the store. This change cascades to other functions which load
1035 certificates and CRLs.
1038 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
1039 facilitate stack unwinding even from assembly subroutines.
1042 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1043 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1046 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1047 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1048 which is the minimum version we support.
1051 *) Certificate time validation (X509_cmp_time) enforces stricter
1052 compliance with RFC 5280. Fractional seconds and timezone offsets
1053 are no longer allowed.
1056 *) Add support for ARIA
1059 *) s_client will now send the Server Name Indication (SNI) extension by
1060 default unless the new "-noservername" option is used. The server name is
1061 based on the host provided to the "-connect" option unless overridden by
1062 using "-servername".
1065 *) Add support for SipHash
1068 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1069 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1070 prevent issues where no progress is being made and the peer continually
1071 sends unrecognised record types, using up resources processing them.
1074 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
1075 using the algorithm defined in
1076 https://www.akkadia.org/drepper/SHA-crypt.txt
1079 *) Heartbeat support has been removed; the ABI is changed for now.
1080 [Richard Levitte, Rich Salz]
1082 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1085 *) The RSA "null" method, which was partially supported to avoid patent
1086 issues, has been replaced to always returns NULL.
1090 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
1092 *) Client DoS due to large DH parameter
1094 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1095 malicious server can send a very large prime value to the client. This will
1096 cause the client to spend an unreasonably long period of time generating a
1097 key for this prime resulting in a hang until the client has finished. This
1098 could be exploited in a Denial Of Service attack.
1100 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1104 *) Cache timing vulnerability in RSA Key Generation
1106 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1107 a cache timing side channel attack. An attacker with sufficient access to
1108 mount cache timing attacks during the RSA key generation process could
1109 recover the private key.
1111 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1112 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1116 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1117 parameter is no longer accepted, as it leads to a corrupt table. NULL
1118 pem_str is reserved for alias entries only.
1121 *) Revert blinding in ECDSA sign and instead make problematic addition
1122 length-invariant. Switch even to fixed-length Montgomery multiplication.
1125 *) Change generating and checking of primes so that the error rate of not
1126 being prime depends on the intended use based on the size of the input.
1127 For larger primes this will result in more rounds of Miller-Rabin.
1128 The maximal error rate for primes with more than 1080 bits is lowered
1130 [Kurt Roeckx, Annie Yousar]
1132 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1135 *) Add blinding to ECDSA and DSA signatures to protect against side channel
1136 attacks discovered by Keegan Ryan (NCC Group).
1139 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1140 now allow empty (zero character) pass phrases.
1143 *) Certificate time validation (X509_cmp_time) enforces stricter
1144 compliance with RFC 5280. Fractional seconds and timezone offsets
1145 are no longer allowed.
1148 *) Fixed a text canonicalisation bug in CMS
1150 Where a CMS detached signature is used with text content the text goes
1151 through a canonicalisation process first prior to signing or verifying a
1152 signature. This process strips trailing space at the end of lines, converts
1153 line terminators to CRLF and removes additional trailing line terminators
1154 at the end of a file. A bug in the canonicalisation process meant that
1155 some characters, such as form-feed, were incorrectly treated as whitespace
1156 and removed. This is contrary to the specification (RFC5485). This fix
1157 could mean that detached text data signed with an earlier version of
1158 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
1159 signed with a fixed OpenSSL may fail to verify with an earlier version of
1160 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
1161 and use the "-binary" flag (for the "cms" command line application) or set
1162 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
1165 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
1167 *) Constructed ASN.1 types with a recursive definition could exceed the stack
1169 Constructed ASN.1 types with a recursive definition (such as can be found
1170 in PKCS7) could eventually exceed the stack given malicious input with
1171 excessive recursion. This could result in a Denial Of Service attack. There
1172 are no such structures used within SSL/TLS that come from untrusted sources
1173 so this is considered safe.
1175 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
1180 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
1182 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
1183 effectively reduced to only comparing the least significant bit of each
1184 byte. This allows an attacker to forge messages that would be considered as
1185 authenticated in an amount of tries lower than that guaranteed by the
1186 security claims of the scheme. The module can only be compiled by the
1187 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
1189 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
1194 *) Add a build target 'build_all_generated', to build all generated files
1195 and only that. This can be used to prepare everything that requires
1196 things like perl for a system that lacks perl and then move everything
1197 to that system and do the rest of the build there.
1200 *) Backport SSL_OP_NO_RENGOTIATION
1202 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
1203 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
1204 changes this is no longer possible in 1.1.0. Therefore the new
1205 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
1206 1.1.0 to provide equivalent functionality.
1208 Note that if an application built against 1.1.0h headers (or above) is run
1209 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
1210 accepted but nothing will happen, i.e. renegotiation will not be prevented.
1213 *) Removed the OS390-Unix config target. It relied on a script that doesn't
1217 *) rsaz_1024_mul_avx2 overflow bug on x86_64
1219 There is an overflow bug in the AVX2 Montgomery multiplication procedure
1220 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
1221 Analysis suggests that attacks against RSA and DSA as a result of this
1222 defect would be very difficult to perform and are not believed likely.
1223 Attacks against DH1024 are considered just feasible, because most of the
1224 work necessary to deduce information about a private key may be performed
1225 offline. The amount of resources required for such an attack would be
1226 significant. However, for an attack on TLS to be meaningful, the server
1227 would have to share the DH1024 private key among multiple clients, which is
1228 no longer an option since CVE-2016-0701.
1230 This only affects processors that support the AVX2 but not ADX extensions
1231 like Intel Haswell (4th generation).
1233 This issue was reported to OpenSSL by David Benjamin (Google). The issue
1234 was originally found via the OSS-Fuzz project.
1238 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
1240 *) bn_sqrx8x_internal carry bug on x86_64
1242 There is a carry propagating bug in the x86_64 Montgomery squaring
1243 procedure. No EC algorithms are affected. Analysis suggests that attacks
1244 against RSA and DSA as a result of this defect would be very difficult to
1245 perform and are not believed likely. Attacks against DH are considered just
1246 feasible (although very difficult) because most of the work necessary to
1247 deduce information about a private key may be performed offline. The amount
1248 of resources required for such an attack would be very significant and
1249 likely only accessible to a limited number of attackers. An attacker would
1250 additionally need online access to an unpatched system using the target
1251 private key in a scenario with persistent DH parameters and a private
1252 key that is shared between multiple clients.
1254 This only affects processors that support the BMI1, BMI2 and ADX extensions
1255 like Intel Broadwell (5th generation) and later or AMD Ryzen.
1257 This issue was reported to OpenSSL by the OSS-Fuzz project.
1261 *) Malformed X.509 IPAddressFamily could cause OOB read
1263 If an X.509 certificate has a malformed IPAddressFamily extension,
1264 OpenSSL could do a one-byte buffer overread. The most likely result
1265 would be an erroneous display of the certificate in text format.
1267 This issue was reported to OpenSSL by the OSS-Fuzz project.
1271 Changes between 1.1.0e and 1.1.0f [25 May 2017]
1273 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1274 platform rather than 'mingw'.
1277 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1278 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1279 which is the minimum version we support.
1282 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
1284 *) Encrypt-Then-Mac renegotiation crash
1286 During a renegotiation handshake if the Encrypt-Then-Mac extension is
1287 negotiated where it was not in the original handshake (or vice-versa) then
1288 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
1289 and servers are affected.
1291 This issue was reported to OpenSSL by Joe Orton (Red Hat).
1295 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
1297 *) Truncated packet could crash via OOB read
1299 If one side of an SSL/TLS path is running on a 32-bit host and a specific
1300 cipher is being used, then a truncated packet can cause that host to
1301 perform an out-of-bounds read, usually resulting in a crash.
1303 This issue was reported to OpenSSL by Robert Święcki of Google.
1307 *) Bad (EC)DHE parameters cause a client crash
1309 If a malicious server supplies bad parameters for a DHE or ECDHE key
1310 exchange then this can result in the client attempting to dereference a
1311 NULL pointer leading to a client crash. This could be exploited in a Denial
1314 This issue was reported to OpenSSL by Guido Vranken.
1318 *) BN_mod_exp may produce incorrect results on x86_64
1320 There is a carry propagating bug in the x86_64 Montgomery squaring
1321 procedure. No EC algorithms are affected. Analysis suggests that attacks
1322 against RSA and DSA as a result of this defect would be very difficult to
1323 perform and are not believed likely. Attacks against DH are considered just
1324 feasible (although very difficult) because most of the work necessary to
1325 deduce information about a private key may be performed offline. The amount
1326 of resources required for such an attack would be very significant and
1327 likely only accessible to a limited number of attackers. An attacker would
1328 additionally need online access to an unpatched system using the target
1329 private key in a scenario with persistent DH parameters and a private
1330 key that is shared between multiple clients. For example this can occur by
1331 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1332 similar to CVE-2015-3193 but must be treated as a separate problem.
1334 This issue was reported to OpenSSL by the OSS-Fuzz project.
1338 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1340 *) ChaCha20/Poly1305 heap-buffer-overflow
1342 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1343 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1344 crash. This issue is not considered to be exploitable beyond a DoS.
1346 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1350 *) CMS Null dereference
1352 Applications parsing invalid CMS structures can crash with a NULL pointer
1353 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1354 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1355 structure callback if an attempt is made to free certain invalid encodings.
1356 Only CHOICE structures using a callback which do not handle NULL value are
1359 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1363 *) Montgomery multiplication may produce incorrect results
1365 There is a carry propagating bug in the Broadwell-specific Montgomery
1366 multiplication procedure that handles input lengths divisible by, but
1367 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1368 and DH private keys are impossible. This is because the subroutine in
1369 question is not used in operations with the private key itself and an input
1370 of the attacker's direct choice. Otherwise the bug can manifest itself as
1371 transient authentication and key negotiation failures or reproducible
1372 erroneous outcome of public-key operations with specially crafted input.
1373 Among EC algorithms only Brainpool P-512 curves are affected and one
1374 presumably can attack ECDH key negotiation. Impact was not analyzed in
1375 detail, because pre-requisites for attack are considered unlikely. Namely
1376 multiple clients have to choose the curve in question and the server has to
1377 share the private key among them, neither of which is default behaviour.
1378 Even then only clients that chose the curve will be affected.
1380 This issue was publicly reported as transient failures and was not
1381 initially recognized as a security issue. Thanks to Richard Morgan for
1382 providing reproducible case.
1386 *) Removed automatic addition of RPATH in shared libraries and executables,
1387 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1390 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1392 *) Fix Use After Free for large message sizes
1394 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1395 message larger than approx 16k is received then the underlying buffer to
1396 store the incoming message is reallocated and moved. Unfortunately a
1397 dangling pointer to the old location is left which results in an attempt to
1398 write to the previously freed location. This is likely to result in a
1399 crash, however it could potentially lead to execution of arbitrary code.
1401 This issue only affects OpenSSL 1.1.0a.
1403 This issue was reported to OpenSSL by Robert Święcki.
1407 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1409 *) OCSP Status Request extension unbounded memory growth
1411 A malicious client can send an excessively large OCSP Status Request
1412 extension. If that client continually requests renegotiation, sending a
1413 large OCSP Status Request extension each time, then there will be unbounded
1414 memory growth on the server. This will eventually lead to a Denial Of
1415 Service attack through memory exhaustion. Servers with a default
1416 configuration are vulnerable even if they do not support OCSP. Builds using
1417 the "no-ocsp" build time option are not affected.
1419 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1423 *) SSL_peek() hang on empty record
1425 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1426 sends an empty record. This could be exploited by a malicious peer in a
1427 Denial Of Service attack.
1429 This issue was reported to OpenSSL by Alex Gaynor.
1433 *) Excessive allocation of memory in tls_get_message_header() and
1434 dtls1_preprocess_fragment()
1436 A (D)TLS message includes 3 bytes for its length in the header for the
1437 message. This would allow for messages up to 16Mb in length. Messages of
1438 this length are excessive and OpenSSL includes a check to ensure that a
1439 peer is sending reasonably sized messages in order to avoid too much memory
1440 being consumed to service a connection. A flaw in the logic of version
1441 1.1.0 means that memory for the message is allocated too early, prior to
1442 the excessive message length check. Due to way memory is allocated in
1443 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1444 to service a connection. This could lead to a Denial of Service through
1445 memory exhaustion. However, the excessive message length check still takes
1446 place, and this would cause the connection to immediately fail. Assuming
1447 that the application calls SSL_free() on the failed connection in a timely
1448 manner then the 21Mb of allocated memory will then be immediately freed
1449 again. Therefore the excessive memory allocation will be transitory in
1450 nature. This then means that there is only a security impact if:
1452 1) The application does not call SSL_free() in a timely manner in the event
1453 that the connection fails
1455 2) The application is working in a constrained environment where there is
1456 very little free memory
1458 3) The attacker initiates multiple connection attempts such that there are
1459 multiple connections in a state where memory has been allocated for the
1460 connection; SSL_free() has not yet been called; and there is insufficient
1461 memory to service the multiple requests.
1463 Except in the instance of (1) above any Denial Of Service is likely to be
1464 transitory because as soon as the connection fails the memory is
1465 subsequently freed again in the SSL_free() call. However there is an
1466 increased risk during this period of application crashes due to the lack of
1467 memory - which would then mean a more serious Denial of Service.
1469 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1470 (CVE-2016-6307 and CVE-2016-6308)
1473 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1474 had to be removed. Primary reason is that vendor assembler can't
1475 assemble our modules with -KPIC flag. As result it, assembly
1476 support, was not even available as option. But its lack means
1477 lack of side-channel resistant code, which is incompatible with
1478 security by todays standards. Fortunately gcc is readily available
1479 prepackaged option, which we firmly point at...
1482 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1484 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1485 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1486 (to any value) allows Windows user to access PKCS#12 file generated
1487 with Windows CryptoAPI and protected with non-ASCII password, as well
1488 as files generated under UTF-8 locale on Linux also protected with
1492 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1493 have been disabled by default and removed from DEFAULT, just like RC4.
1494 See the RC4 item below to re-enable both.
1497 *) The method for finding the storage location for the Windows RAND seed file
1498 has changed. First we check %RANDFILE%. If that is not set then we check
1499 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1500 all else fails we fall back to C:\.
1503 *) The EVP_EncryptUpdate() function has had its return type changed from void
1504 to int. A return of 0 indicates and error while a return of 1 indicates
1508 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1509 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1510 off the constant time implementation for RSA, DSA and DH have been made
1511 no-ops and deprecated.
1514 *) Windows RAND implementation was simplified to only get entropy by
1515 calling CryptGenRandom(). Various other RAND-related tickets
1517 [Joseph Wylie Yandle, Rich Salz]
1519 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1520 and OPENSSL_LH_, respectively. The old names are available
1521 with API compatibility. They new names are now completely documented.
1524 *) Unify TYPE_up_ref(obj) methods signature.
1525 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1526 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1527 int (instead of void) like all others TYPE_up_ref() methods.
1528 So now these methods also check the return value of CRYPTO_atomic_add(),
1529 and the validity of object reference counter.
1530 [fdasilvayy@gmail.com]
1532 *) With Windows Visual Studio builds, the .pdb files are installed
1533 alongside the installed libraries and executables. For a static
1534 library installation, ossl_static.pdb is the associate compiler
1535 generated .pdb file to be used when linking programs.
1538 *) Remove openssl.spec. Packaging files belong with the packagers.
1541 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1542 recognise x86_64 architectures automatically. You can still decide
1543 to build for a different bitness with the environment variable
1544 KERNEL_BITS (can be 32 or 64), for example:
1546 KERNEL_BITS=32 ./config
1550 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1551 256 bit AES and HMAC with SHA256.
1554 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1557 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1560 *) To enable users to have their own config files and build file templates,
1561 Configure looks in the directory indicated by the environment variable
1562 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1563 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1564 name and is used as is.
1567 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1568 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1569 X509_CERT_FILE_CTX was removed.
1572 *) "shared" builds are now the default. To create only static libraries use
1573 the "no-shared" Configure option.
1576 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1577 All of these option have not worked for some while and are fundamental
1581 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1582 global cleanup functions are no longer required because they are handled
1583 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1584 Explicitly de-initing can cause problems (e.g. where a library that uses
1585 OpenSSL de-inits, but an application is still using it). The affected
1586 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1587 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1588 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1589 COMP_zlib_cleanup().
1592 *) --strict-warnings no longer enables runtime debugging options
1593 such as REF_DEBUG. Instead, debug options are automatically
1594 enabled with '--debug' builds.
1595 [Andy Polyakov, Emilia Käsper]
1597 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1598 have been moved out of the public header files. New functions for managing
1599 these have been added.
1602 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1603 objects have been moved out of the public header files. New
1604 functions for managing these have been added.
1607 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1608 have been moved out of the public header files. New functions for managing
1609 these have been added.
1612 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1613 moved out of the public header files. New functions for managing these
1617 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1620 *) Removed the mk1mf build scripts.
1623 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1624 it is always safe to #include a header now.
1627 *) Removed the aged BC-32 config and all its supporting scripts
1630 *) Removed support for Ultrix, Netware, and OS/2.
1633 *) Add support for HKDF.
1634 [Alessandro Ghedini]
1636 *) Add support for blake2b and blake2s
1639 *) Added support for "pipelining". Ciphers that have the
1640 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1641 encryptions/decryptions simultaneously. There are currently no built-in
1642 ciphers with this property but the expectation is that engines will be able
1643 to offer it to significantly improve throughput. Support has been extended
1644 into libssl so that multiple records for a single connection can be
1645 processed in one go (for >=TLS 1.1).
1648 *) Added the AFALG engine. This is an async capable engine which is able to
1649 offload work to the Linux kernel. In this initial version it only supports
1650 AES128-CBC. The kernel must be version 4.1.0 or greater.
1653 *) OpenSSL now uses a new threading API. It is no longer necessary to
1654 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1655 are two supported threading models: pthreads and windows threads. It is
1656 also possible to configure OpenSSL at compile time for "no-threads". The
1657 old threading API should no longer be used. The functions have been
1658 replaced with "no-op" compatibility macros.
1659 [Alessandro Ghedini, Matt Caswell]
1661 *) Modify behavior of ALPN to invoke callback after SNI/servername
1662 callback, such that updates to the SSL_CTX affect ALPN.
1665 *) Add SSL_CIPHER queries for authentication and key-exchange.
1668 *) Changes to the DEFAULT cipherlist:
1669 - Prefer (EC)DHE handshakes over plain RSA.
1670 - Prefer AEAD ciphers over legacy ciphers.
1671 - Prefer ECDSA over RSA when both certificates are available.
1672 - Prefer TLSv1.2 ciphers/PRF.
1673 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1677 *) Change the ECC default curve list to be this, in order: x25519,
1678 secp256r1, secp521r1, secp384r1.
1681 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1682 disabled by default. They can be re-enabled using the
1683 enable-weak-ssl-ciphers option to Configure.
1686 *) If the server has ALPN configured, but supports no protocols that the
1687 client advertises, send a fatal "no_application_protocol" alert.
1688 This behaviour is SHALL in RFC 7301, though it isn't universally
1689 implemented by other servers.
1692 *) Add X25519 support.
1693 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1694 for public and private key encoding using the format documented in
1695 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1696 key generation and key derivation.
1698 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1702 *) Deprecate SRP_VBASE_get_by_user.
1703 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1704 In order to fix an unavoidable memory leak (CVE-2016-0798),
1705 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1706 seed, even if the seed is configured.
1708 Users should use SRP_VBASE_get1_by_user instead. Note that in
1709 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1710 also that even though configuring the SRP seed attempts to hide
1711 invalid usernames by continuing the handshake with fake
1712 credentials, this behaviour is not constant time and no strong
1713 guarantees are made that the handshake is indistinguishable from
1714 that of a valid user.
1717 *) Configuration change; it's now possible to build dynamic engines
1718 without having to build shared libraries and vice versa. This
1719 only applies to the engines in engines/, those in crypto/engine/
1720 will always be built into libcrypto (i.e. "static").
1722 Building dynamic engines is enabled by default; to disable, use
1723 the configuration option "disable-dynamic-engine".
1725 The only requirements for building dynamic engines are the
1726 presence of the DSO module and building with position independent
1727 code, so they will also automatically be disabled if configuring
1728 with "disable-dso" or "disable-pic".
1730 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1731 are also taken away from openssl/opensslconf.h, as they are
1735 *) Configuration change; if there is a known flag to compile
1736 position independent code, it will always be applied on the
1737 libcrypto and libssl object files, and never on the application
1738 object files. This means other libraries that use routines from
1739 libcrypto / libssl can be made into shared libraries regardless
1740 of how OpenSSL was configured.
1742 If this isn't desirable, the configuration options "disable-pic"
1743 or "no-pic" can be used to disable the use of PIC. This will
1744 also disable building shared libraries and dynamic engines.
1747 *) Removed JPAKE code. It was experimental and has no wide use.
1750 *) The INSTALL_PREFIX Makefile variable has been renamed to
1751 DESTDIR. That makes for less confusion on what this variable
1752 is for. Also, the configuration option --install_prefix is
1756 *) Heartbeat for TLS has been removed and is disabled by default
1757 for DTLS; configure with enable-heartbeats. Code that uses the
1758 old #define's might need to be updated.
1759 [Emilia Käsper, Rich Salz]
1761 *) Rename REF_CHECK to REF_DEBUG.
1764 *) New "unified" build system
1766 The "unified" build system is aimed to be a common system for all
1767 platforms we support. With it comes new support for VMS.
1769 This system builds supports building in a different directory tree
1770 than the source tree. It produces one Makefile (for unix family
1771 or lookalikes), or one descrip.mms (for VMS).
1773 The source of information to make the Makefile / descrip.mms is
1774 small files called 'build.info', holding the necessary
1775 information for each directory with source to compile, and a
1776 template in Configurations, like unix-Makefile.tmpl or
1779 With this change, the library names were also renamed on Windows
1780 and on VMS. They now have names that are closer to the standard
1781 on Unix, and include the major version number, and in certain
1782 cases, the architecture they are built for. See "Notes on shared
1783 libraries" in INSTALL.
1785 We rely heavily on the perl module Text::Template.
1788 *) Added support for auto-initialisation and de-initialisation of the library.
1789 OpenSSL no longer requires explicit init or deinit routines to be called,
1790 except in certain circumstances. See the OPENSSL_init_crypto() and
1791 OPENSSL_init_ssl() man pages for further information.
1794 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1795 "peer" argument is now expected to be a BIO_ADDR object.
1797 *) Rewrite of BIO networking library. The BIO library lacked consistent
1798 support of IPv6, and adding it required some more extensive
1799 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1800 which hold all types of addresses and chains of address information.
1801 It also introduces a new API, with functions like BIO_socket,
1802 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1803 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1804 have been adapted accordingly.
1807 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1811 *) CRIME protection: disable compression by default, even if OpenSSL is
1812 compiled with zlib enabled. Applications can still enable compression
1813 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1814 using the SSL_CONF library to configure compression.
1817 *) The signature of the session callback configured with
1818 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1819 was explicitly marked as 'const unsigned char*' instead of
1823 *) Always DPURIFY. Remove the use of uninitialized memory in the
1824 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1827 *) Removed many obsolete configuration items, including
1828 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1829 MD2_CHAR, MD2_INT, MD2_LONG
1831 IDEA_SHORT, IDEA_LONG
1832 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1833 [Rich Salz, with advice from Andy Polyakov]
1835 *) Many BN internals have been moved to an internal header file.
1836 [Rich Salz with help from Andy Polyakov]
1838 *) Configuration and writing out the results from it has changed.
1839 Files such as Makefile include/openssl/opensslconf.h and are now
1840 produced through general templates, such as Makefile.in and
1841 crypto/opensslconf.h.in and some help from the perl module
1844 Also, the center of configuration information is no longer
1845 Makefile. Instead, Configure produces a perl module in
1846 configdata.pm which holds most of the config data (in the hash
1847 table %config), the target data that comes from the target
1848 configuration in one of the Configurations/*.conf files (in
1852 *) To clarify their intended purposes, the Configure options
1853 --prefix and --openssldir change their semantics, and become more
1854 straightforward and less interdependent.
1856 --prefix shall be used exclusively to give the location INSTALLTOP
1857 where programs, scripts, libraries, include files and manuals are
1858 going to be installed. The default is now /usr/local.
1860 --openssldir shall be used exclusively to give the default
1861 location OPENSSLDIR where certificates, private keys, CRLs are
1862 managed. This is also where the default openssl.cnf gets
1864 If the directory given with this option is a relative path, the
1865 values of both the --prefix value and the --openssldir value will
1866 be combined to become OPENSSLDIR.
1867 The default for --openssldir is INSTALLTOP/ssl.
1869 Anyone who uses --openssldir to specify where OpenSSL is to be
1870 installed MUST change to use --prefix instead.
1873 *) The GOST engine was out of date and therefore it has been removed. An up
1874 to date GOST engine is now being maintained in an external repository.
1875 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1876 support for GOST ciphersuites (these are only activated if a GOST engine
1880 *) EGD is no longer supported by default; use enable-egd when
1882 [Ben Kaduk and Rich Salz]
1884 *) The distribution now has Makefile.in files, which are used to
1885 create Makefile's when Configure is run. *Configure must be run
1886 before trying to build now.*
1889 *) The return value for SSL_CIPHER_description() for error conditions
1893 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1895 Obtaining and performing DNSSEC validation of TLSA records is
1896 the application's responsibility. The application provides
1897 the TLSA records of its choice to OpenSSL, and these are then
1898 used to authenticate the peer.
1900 The TLSA records need not even come from DNS. They can, for
1901 example, be used to implement local end-entity certificate or
1902 trust-anchor "pinning", where the "pin" data takes the form
1903 of TLSA records, which can augment or replace verification
1904 based on the usual WebPKI public certification authorities.
1907 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1908 continues to support deprecated interfaces in default builds.
1909 However, applications are strongly advised to compile their
1910 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1911 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1912 or the 1.1.0 releases.
1914 In environments in which all applications have been ported to
1915 not use any deprecated interfaces OpenSSL's Configure script
1916 should be used with the --api=1.1.0 option to entirely remove
1917 support for the deprecated features from the library and
1918 unconditionally disable them in the installed headers.
1919 Essentially the same effect can be achieved with the "no-deprecated"
1920 argument to Configure, except that this will always restrict
1921 the build to just the latest API, rather than a fixed API
1924 As applications are ported to future revisions of the API,
1925 they should update their compile-time OPENSSL_API_COMPAT define
1926 accordingly, but in most cases should be able to continue to
1927 compile with later releases.
1929 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1930 0x10000000L and 0x00908000L, respectively. However those
1931 versions did not support the OPENSSL_API_COMPAT feature, and
1932 so applications are not typically tested for explicit support
1933 of just the undeprecated features of either release.
1936 *) Add support for setting the minimum and maximum supported protocol.
1937 It can bet set via the SSL_set_min_proto_version() and
1938 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1939 MaxProtocol. It's recommended to use the new APIs to disable
1940 protocols instead of disabling individual protocols using
1941 SSL_set_options() or SSL_CONF's Protocol. This change also
1942 removes support for disabling TLS 1.2 in the OpenSSL TLS
1943 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1946 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1949 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1950 and integrates ECDSA and ECDH functionality into EC. Implementations can
1951 now redirect key generation and no longer need to convert to or from
1954 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1955 include the ec.h header file instead.
1958 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1959 ciphers who are no longer supported and drops support the ephemeral RSA key
1960 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1963 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1964 opaque. For HMAC_CTX, the following constructors and destructors
1967 HMAC_CTX *HMAC_CTX_new(void);
1968 void HMAC_CTX_free(HMAC_CTX *ctx);
1970 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1971 destroy such methods has been added. See EVP_MD_meth_new(3) and
1972 EVP_CIPHER_meth_new(3) for documentation.
1975 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1976 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1977 EVP_MD_CTX_reset() should be called instead to reinitialise
1978 an already created structure.
1979 2) For consistency with the majority of our object creators and
1980 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1981 EVP_MD_CTX_(new|free). The old names are retained as macros
1982 for deprecated builds.
1985 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1986 cryptographic operations to be performed asynchronously as long as an
1987 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1988 further details. Libssl has also had this capability integrated with the
1989 introduction of the new mode SSL_MODE_ASYNC and associated error
1990 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1991 pages. This work was developed in partnership with Intel Corp.
1994 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1995 always enabled now. If you want to disable the support you should
1996 exclude it using the list of supported ciphers. This also means that the
1997 "-no_ecdhe" option has been removed from s_server.
2000 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2001 SSL_{CTX_}set1_curves() which can set a list.
2004 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2005 curve you want to support using SSL_{CTX_}set1_curves().
2008 *) State machine rewrite. The state machine code has been significantly
2009 refactored in order to remove much duplication of code and solve issues
2010 with the old code (see ssl/statem/README for further details). This change
2011 does have some associated API changes. Notably the SSL_state() function
2012 has been removed and replaced by SSL_get_state which now returns an
2013 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
2014 altogether. The previous handshake states defined in ssl.h and ssl3.h have
2018 *) All instances of the string "ssleay" in the public API were replaced
2019 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
2020 Some error codes related to internal RSA_eay API's were renamed.
2023 *) The demo files in crypto/threads were moved to demo/threads.
2026 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
2028 [Matt Caswell, Rich Salz]
2030 *) New ASN.1 embed macro.
2032 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
2033 structure is not allocated: it is part of the parent. That is instead of
2041 This reduces memory fragmentation and make it impossible to accidentally
2042 set a mandatory field to NULL.
2044 This currently only works for some fields specifically a SEQUENCE, CHOICE,
2045 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
2046 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
2050 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
2053 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
2054 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
2055 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
2056 DES and RC4 ciphersuites.
2059 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2060 This changes the decoding behaviour for some invalid messages,
2061 though the change is mostly in the more lenient direction, and
2062 legacy behaviour is preserved as much as possible.
2065 *) Fix no-stdio build.
2066 [ David Woodhouse <David.Woodhouse@intel.com> and also
2067 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
2069 *) New testing framework
2070 The testing framework has been largely rewritten and is now using
2071 perl and the perl modules Test::Harness and an extended variant of
2072 Test::More called OpenSSL::Test to do its work. All test scripts in
2073 test/ have been rewritten into test recipes, and all direct calls to
2074 executables in test/Makefile have become individual recipes using the
2075 simplified testing OpenSSL::Test::Simple.
2077 For documentation on our testing modules, do:
2079 perldoc test/testlib/OpenSSL/Test/Simple.pm
2080 perldoc test/testlib/OpenSSL/Test.pm
2084 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
2085 are used; the latter aborts on memory leaks (usually checked on exit).
2086 Some undocumented "set malloc, etc., hooks" functions were removed
2087 and others were changed. All are now documented.
2090 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2092 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2094 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
2095 from RFC4279, RFC4785, RFC5487, RFC5489.
2097 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
2098 original RSA_PSK patch.
2101 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
2102 era flag was never set throughout the codebase (only read). Also removed
2103 SSL3_FLAGS_POP_BUFFER which was only used if
2104 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
2107 *) Changed the default name options in the "ca", "crl", "req" and "x509"
2108 to be "oneline" instead of "compat".
2111 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
2112 not aware of clients that still exhibit this bug, and the workaround
2113 hasn't been working properly for a while.
2116 *) The return type of BIO_number_read() and BIO_number_written() as well as
2117 the corresponding num_read and num_write members in the BIO structure has
2118 changed from unsigned long to uint64_t. On platforms where an unsigned
2119 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
2123 *) Given the pervasive nature of TLS extensions it is inadvisable to run
2124 OpenSSL without support for them. It also means that maintaining
2125 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
2126 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
2129 *) Removed support for the two export grade static DH ciphersuites
2130 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
2131 were newly added (along with a number of other static DH ciphersuites) to
2132 1.0.2. However the two export ones have *never* worked since they were
2133 introduced. It seems strange in any case to be adding new export
2134 ciphersuites, and given "logjam" it also does not seem correct to fix them.
2137 *) Version negotiation has been rewritten. In particular SSLv23_method(),
2138 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
2139 and turned into macros which simply call the new preferred function names
2140 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
2141 should use the new names instead. Also as part of this change the ssl23.h
2142 header file has been removed.
2145 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
2146 code and the associated standard is no longer considered fit-for-purpose.
2149 *) RT2547 was closed. When generating a private key, try to make the
2150 output file readable only by the owner. This behavior change might
2151 be noticeable when interacting with other software.
2153 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
2157 *) Added HTTP GET support to the ocsp command.
2160 *) Changed default digest for the dgst and enc commands from MD5 to
2164 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
2167 *) Added support for TLS extended master secret from
2168 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
2169 initial patch which was a great help during development.
2172 *) All libssl internal structures have been removed from the public header
2173 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
2174 now redundant). Users should not attempt to access internal structures
2175 directly. Instead they should use the provided API functions.
2178 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
2179 Access to deprecated functions can be re-enabled by running config with
2180 "enable-deprecated". In addition applications wishing to use deprecated
2181 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
2182 will, by default, disable some transitive includes that previously existed
2183 in the header files (e.g. ec.h will no longer, by default, include bn.h)
2186 *) Added support for OCB mode. OpenSSL has been granted a patent license
2187 compatible with the OpenSSL license for use of OCB. Details are available
2188 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
2189 for OCB can be removed by calling config with no-ocb.
2192 *) SSLv2 support has been removed. It still supports receiving a SSLv2
2193 compatible client hello.
2196 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
2197 done while fixing the error code for the key-too-small case.
2198 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
2200 *) CA.sh has been removed; use CA.pl instead.
2203 *) Removed old DES API.
2206 *) Remove various unsupported platforms:
2212 Sinix/ReliantUNIX RM400
2217 16-bit platforms such as WIN16
2220 *) Clean up OPENSSL_NO_xxx #define's
2221 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
2222 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
2223 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
2224 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
2225 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
2226 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
2227 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
2228 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
2229 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
2230 Remove MS_STATIC; it's a relic from platforms <32 bits.
2233 *) Cleaned up dead code
2234 Remove all but one '#ifdef undef' which is to be looked at.
2237 *) Clean up calling of xxx_free routines.
2238 Just like free(), fix most of the xxx_free routines to accept
2239 NULL. Remove the non-null checks from callers. Save much code.
2242 *) Add secure heap for storage of private keys (when possible).
2243 Add BIO_s_secmem(), CBIGNUM, etc.
2244 Contributed by Akamai Technologies under our Corporate CLA.
2247 *) Experimental support for a new, fast, unbiased prime candidate generator,
2248 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
2249 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
2251 *) New output format NSS in the sess_id command line tool. This allows
2252 exporting the session id and the master key in NSS keylog format.
2253 [Martin Kaiser <martin@kaiser.cx>]
2255 *) Harmonize version and its documentation. -f flag is used to display
2257 [mancha <mancha1@zoho.com>]
2259 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2260 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
2261 [mancha <mancha1@zoho.com>]
2263 *) Fix some double frees. These are not thought to be exploitable.
2264 [mancha <mancha1@zoho.com>]
2266 *) A missing bounds check in the handling of the TLS heartbeat extension
2267 can be used to reveal up to 64k of memory to a connected client or
2270 Thanks for Neel Mehta of Google Security for discovering this bug and to
2271 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2272 preparing the fix (CVE-2014-0160)
2273 [Adam Langley, Bodo Moeller]
2275 *) Fix for the attack described in the paper "Recovering OpenSSL
2276 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2277 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2278 http://eprint.iacr.org/2014/140
2280 Thanks to Yuval Yarom and Naomi Benger for discovering this
2281 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2282 [Yuval Yarom and Naomi Benger]
2284 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2285 this fixes a limitation in previous versions of OpenSSL.
2288 *) Experimental encrypt-then-mac support.
2290 Experimental support for encrypt then mac from
2291 draft-gutmann-tls-encrypt-then-mac-02.txt
2293 To enable it set the appropriate extension number (0x42 for the test
2294 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
2296 For non-compliant peers (i.e. just about everything) this should have no
2299 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
2303 *) Add EVP support for key wrapping algorithms, to avoid problems with
2304 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2305 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2306 algorithms and include tests cases.
2309 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2313 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2314 MGF1 digest and OAEP label.
2317 *) Make openssl verify return errors.
2318 [Chris Palmer <palmer@google.com> and Ben Laurie]
2320 *) New function ASN1_TIME_diff to calculate the difference between two
2321 ASN1_TIME structures or one structure and the current time.
2324 *) Update fips_test_suite to support multiple command line options. New
2325 test to induce all self test errors in sequence and check expected
2329 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2330 sign or verify all in one operation.
2333 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2334 test programs and fips_test_suite. Includes functionality to parse
2335 the minimal script output of fipsalgest.pl directly.
2338 *) Add authorisation parameter to FIPS_module_mode_set().
2341 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2344 *) Use separate DRBG fields for internal and external flags. New function
2345 FIPS_drbg_health_check() to perform on demand health checking. Add
2346 generation tests to fips_test_suite with reduced health check interval to
2347 demonstrate periodic health checking. Add "nodh" option to
2348 fips_test_suite to skip very slow DH test.
2351 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2355 *) More extensive health check for DRBG checking many more failure modes.
2356 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2357 combination: call this in fips_test_suite.
2360 *) Add support for canonical generation of DSA parameter 'g'. See
2363 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2364 POST to handle HMAC cases.
2367 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2368 to return numerical and string versions of the FIPS module number.
2371 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2372 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2373 outside the validated module in the FIPS capable OpenSSL.
2376 *) Minor change to DRBG entropy callback semantics. In some cases
2377 there is no multiple of the block length between min_len and
2378 max_len. Allow the callback to return more than max_len bytes
2379 of entropy but discard any extra: it is the callback's responsibility
2380 to ensure that the extra data discarded does not impact the
2381 requested amount of entropy.
2384 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2385 information in FIPS186-3, SP800-57 and SP800-131A.
2388 *) CCM support via EVP. Interface is very similar to GCM case except we
2389 must supply all data in one chunk (i.e. no update, final) and the
2390 message length must be supplied if AAD is used. Add algorithm test
2394 *) Initial version of POST overhaul. Add POST callback to allow the status
2395 of POST to be monitored and/or failures induced. Modify fips_test_suite
2396 to use callback. Always run all selftests even if one fails.
2399 *) XTS support including algorithm test driver in the fips_gcmtest program.
2400 Note: this does increase the maximum key length from 32 to 64 bytes but
2401 there should be no binary compatibility issues as existing applications
2402 will never use XTS mode.
2405 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2406 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2407 performs algorithm blocking for unapproved PRNG types. Also do not
2408 set PRNG type in FIPS_mode_set(): leave this to the application.
2409 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2410 the standard OpenSSL PRNG: set additional data to a date time vector.
2413 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2414 This shouldn't present any incompatibility problems because applications
2415 shouldn't be using these directly and any that are will need to rethink
2416 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2419 *) Extensive self tests and health checking required by SP800-90 DRBG.
2420 Remove strength parameter from FIPS_drbg_instantiate and always
2421 instantiate at maximum supported strength.
2424 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2427 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2430 *) New function DH_compute_key_padded() to compute a DH key and pad with
2431 leading zeroes if needed: this complies with SP800-56A et al.
2434 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2435 anything, incomplete, subject to change and largely untested at present.
2438 *) Modify fipscanisteronly build option to only build the necessary object
2439 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2442 *) Add experimental option FIPSSYMS to give all symbols in
2443 fipscanister.o and FIPS or fips prefix. This will avoid
2444 conflicts with future versions of OpenSSL. Add perl script
2445 util/fipsas.pl to preprocess assembly language source files
2446 and rename any affected symbols.
2449 *) Add selftest checks and algorithm block of non-fips algorithms in
2450 FIPS mode. Remove DES2 from selftests.
2453 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2454 return internal method without any ENGINE dependencies. Add new
2455 tiny fips sign and verify functions.
2458 *) New build option no-ec2m to disable characteristic 2 code.
2461 *) New build option "fipscanisteronly". This only builds fipscanister.o
2462 and (currently) associated fips utilities. Uses the file Makefile.fips
2463 instead of Makefile.org as the prototype.
2466 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2467 Update fips_gcmtest to use IV generator.
2470 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2471 setting output buffer to NULL. The *Final function must be
2472 called although it will not retrieve any additional data. The tag
2473 can be set or retrieved with a ctrl. The IV length is by default 12
2474 bytes (96 bits) but can be set to an alternative value. If the IV
2475 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2479 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2480 underlying do_cipher function handles all cipher semantics itself
2481 including padding and finalisation. This is useful if (for example)
2482 an ENGINE cipher handles block padding itself. The behaviour of
2483 do_cipher is subtly changed if this flag is set: the return value
2484 is the number of characters written to the output buffer (zero is
2485 no longer an error code) or a negative error code. Also if the
2486 input buffer is NULL and length 0 finalisation should be performed.
2489 *) If a candidate issuer certificate is already part of the constructed
2490 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2493 *) Improve forward-security support: add functions
2495 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2496 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2498 for use by SSL/TLS servers; the callback function will be called whenever a
2499 new session is created, and gets to decide whether the session may be
2500 cached to make it resumable (return 0) or not (return 1). (As by the
2501 SSL/TLS protocol specifications, the session_id sent by the server will be
2502 empty to indicate that the session is not resumable; also, the server will
2503 not generate RFC 4507 (RFC 5077) session tickets.)
2505 A simple reasonable callback implementation is to return is_forward_secure.
2506 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2507 by the SSL/TLS server library, indicating whether it can provide forward
2509 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2511 *) New -verify_name option in command line utilities to set verification
2515 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2516 Add CMAC pkey methods.
2519 *) Experimental renegotiation in s_server -www mode. If the client
2520 browses /reneg connection is renegotiated. If /renegcert it is
2521 renegotiated requesting a certificate.
2524 *) Add an "external" session cache for debugging purposes to s_server. This
2525 should help trace issues which normally are only apparent in deployed
2526 multi-process servers.
2529 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2530 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2531 BIO_set_cipher() and some obscure PEM functions were changed so they
2532 can now return an error. The RAND changes required a change to the
2533 RAND_METHOD structure.
2536 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2537 a gcc attribute to warn if the result of a function is ignored. This
2538 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2539 whose return value is often ignored.
2542 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2543 These allow SCTs (signed certificate timestamps) to be requested and
2544 validated when establishing a connection.
2545 [Rob Percival <robpercival@google.com>]
2547 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2549 *) Prevent padding oracle in AES-NI CBC MAC check
2551 A MITM attacker can use a padding oracle attack to decrypt traffic
2552 when the connection uses an AES CBC cipher and the server support
2555 This issue was introduced as part of the fix for Lucky 13 padding
2556 attack (CVE-2013-0169). The padding check was rewritten to be in
2557 constant time by making sure that always the same bytes are read and
2558 compared against either the MAC or padding bytes. But it no longer
2559 checked that there was enough data to have both the MAC and padding
2562 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2566 *) Fix EVP_EncodeUpdate overflow
2568 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2569 Base64 encoding of binary data. If an attacker is able to supply very large
2570 amounts of input data then a length check can overflow resulting in a heap
2573 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2574 the PEM_write_bio* family of functions. These are mainly used within the
2575 OpenSSL command line applications, so any application which processes data
2576 from an untrusted source and outputs it as a PEM file should be considered
2577 vulnerable to this issue. User applications that call these APIs directly
2578 with large amounts of untrusted data may also be vulnerable.
2580 This issue was reported by Guido Vranken.
2584 *) Fix EVP_EncryptUpdate overflow
2586 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2587 is able to supply very large amounts of input data after a previous call to
2588 EVP_EncryptUpdate() with a partial block then a length check can overflow
2589 resulting in a heap corruption. Following an analysis of all OpenSSL
2590 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2591 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2592 the first called function after an EVP_EncryptInit(), and therefore that
2593 specific call must be safe. The second form is where the length passed to
2594 EVP_EncryptUpdate() can be seen from the code to be some small value and
2595 therefore there is no possibility of an overflow. Since all instances are
2596 one of these two forms, it is believed that there can be no overflows in
2597 internal code due to this problem. It should be noted that
2598 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2599 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2600 of these calls have also been analysed too and it is believed there are no
2601 instances in internal usage where an overflow could occur.
2603 This issue was reported by Guido Vranken.
2607 *) Prevent ASN.1 BIO excessive memory allocation
2609 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2610 a short invalid encoding can cause allocation of large amounts of memory
2611 potentially consuming excessive resources or exhausting memory.
2613 Any application parsing untrusted data through d2i BIO functions is
2614 affected. The memory based functions such as d2i_X509() are *not* affected.
2615 Since the memory based functions are used by the TLS library, TLS
2616 applications are not affected.
2618 This issue was reported by Brian Carpenter.
2624 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2625 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2626 in arbitrary stack data being returned in the buffer.
2628 This issue was reported by Guido Vranken.
2632 *) Modify behavior of ALPN to invoke callback after SNI/servername
2633 callback, such that updates to the SSL_CTX affect ALPN.
2636 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2640 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2641 methods are enabled and ssl2 is disabled the methods return NULL.
2644 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2646 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2647 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2648 provide any "EXPORT" or "LOW" strength ciphers.
2651 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2652 is by default disabled at build-time. Builds that are not configured with
2653 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2654 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2655 will need to explicitly call either of:
2657 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2659 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2661 as appropriate. Even if either of those is used, or the application
2662 explicitly uses the version-specific SSLv2_method() or its client and
2663 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2664 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2665 ciphers, and SSLv2 56-bit DES are no longer available.
2669 *) Fix a double-free in DSA code
2671 A double free bug was discovered when OpenSSL parses malformed DSA private
2672 keys and could lead to a DoS attack or memory corruption for applications
2673 that receive DSA private keys from untrusted sources. This scenario is
2676 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2681 *) Disable SRP fake user seed to address a server memory leak.
2683 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2685 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2686 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2687 was changed to ignore the "fake user" SRP seed, even if the seed
2690 Users should use SRP_VBASE_get1_by_user instead. Note that in
2691 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2692 also that even though configuring the SRP seed attempts to hide
2693 invalid usernames by continuing the handshake with fake
2694 credentials, this behaviour is not constant time and no strong
2695 guarantees are made that the handshake is indistinguishable from
2696 that of a valid user.
2700 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2702 In the BN_hex2bn function the number of hex digits is calculated using an
2703 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2704 large values of |i| this can result in |bn_expand| not allocating any
2705 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2706 field as NULL leading to a subsequent NULL ptr deref. For very large values
2707 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2708 In this case memory is allocated to the internal BIGNUM data field, but it
2709 is insufficiently sized leading to heap corruption. A similar issue exists
2710 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2711 is ever called by user applications with very large untrusted hex/dec data.
2712 This is anticipated to be a rare occurrence.
2714 All OpenSSL internal usage of these functions use data that is not expected
2715 to be untrusted, e.g. config file data or application command line
2716 arguments. If user developed applications generate config file data based
2717 on untrusted data then it is possible that this could also lead to security
2718 consequences. This is also anticipated to be rare.
2720 This issue was reported to OpenSSL by Guido Vranken.
2724 *) Fix memory issues in BIO_*printf functions
2726 The internal |fmtstr| function used in processing a "%s" format string in
2727 the BIO_*printf functions could overflow while calculating the length of a
2728 string and cause an OOB read when printing very long strings.
2730 Additionally the internal |doapr_outch| function can attempt to write to an
2731 OOB memory location (at an offset from the NULL pointer) in the event of a
2732 memory allocation failure. In 1.0.2 and below this could be caused where
2733 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2734 could be in processing a very long "%s" format string. Memory leaks can
2737 The first issue may mask the second issue dependent on compiler behaviour.
2738 These problems could enable attacks where large amounts of untrusted data
2739 is passed to the BIO_*printf functions. If applications use these functions
2740 in this way then they could be vulnerable. OpenSSL itself uses these
2741 functions when printing out human-readable dumps of ASN.1 data. Therefore
2742 applications that print this data could be vulnerable if the data is from
2743 untrusted sources. OpenSSL command line applications could also be
2744 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2745 as command line arguments.
2747 Libssl is not considered directly vulnerable. Additionally certificates etc
2748 received via remote connections via libssl are also unlikely to be able to
2749 trigger these issues because of message size limits enforced within libssl.
2751 This issue was reported to OpenSSL Guido Vranken.
2755 *) Side channel attack on modular exponentiation
2757 A side-channel attack was found which makes use of cache-bank conflicts on
2758 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2759 of RSA keys. The ability to exploit this issue is limited as it relies on
2760 an attacker who has control of code in a thread running on the same
2761 hyper-threaded core as the victim thread which is performing decryptions.
2763 This issue was reported to OpenSSL by Yuval Yarom, The University of
2764 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2765 Nadia Heninger, University of Pennsylvania with more information at
2766 http://cachebleed.info.
2770 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2771 if no keysize is specified with default_bits. This fixes an
2772 omission in an earlier change that changed all RSA/DSA key generation
2773 apps to use 2048 bits by default.
2776 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2777 *) DH small subgroups
2779 Historically OpenSSL only ever generated DH parameters based on "safe"
2780 primes. More recently (in version 1.0.2) support was provided for
2781 generating X9.42 style parameter files such as those required for RFC 5114
2782 support. The primes used in such files may not be "safe". Where an
2783 application is using DH configured with parameters based on primes that are
2784 not "safe" then an attacker could use this fact to find a peer's private
2785 DH exponent. This attack requires that the attacker complete multiple
2786 handshakes in which the peer uses the same private DH exponent. For example
2787 this could be used to discover a TLS server's private DH exponent if it's
2788 reusing the private DH exponent or it's using a static DH ciphersuite.
2790 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2791 TLS. It is not on by default. If the option is not set then the server
2792 reuses the same private DH exponent for the life of the server process and
2793 would be vulnerable to this attack. It is believed that many popular
2794 applications do set this option and would therefore not be at risk.
2796 The fix for this issue adds an additional check where a "q" parameter is
2797 available (as is the case in X9.42 based parameters). This detects the
2798 only known attack, and is the only possible defense for static DH
2799 ciphersuites. This could have some performance impact.
2801 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2802 default and cannot be disabled. This could have some performance impact.
2804 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2808 *) SSLv2 doesn't block disabled ciphers
2810 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2811 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2812 been disabled, provided that the SSLv2 protocol was not also disabled via
2815 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2816 and Sebastian Schinzel.
2820 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2822 *) BN_mod_exp may produce incorrect results on x86_64
2824 There is a carry propagating bug in the x86_64 Montgomery squaring
2825 procedure. No EC algorithms are affected. Analysis suggests that attacks
2826 against RSA and DSA as a result of this defect would be very difficult to
2827 perform and are not believed likely. Attacks against DH are considered just
2828 feasible (although very difficult) because most of the work necessary to
2829 deduce information about a private key may be performed offline. The amount
2830 of resources required for such an attack would be very significant and
2831 likely only accessible to a limited number of attackers. An attacker would
2832 additionally need online access to an unpatched system using the target
2833 private key in a scenario with persistent DH parameters and a private
2834 key that is shared between multiple clients. For example this can occur by
2835 default in OpenSSL DHE based SSL/TLS ciphersuites.
2837 This issue was reported to OpenSSL by Hanno Böck.
2841 *) Certificate verify crash with missing PSS parameter
2843 The signature verification routines will crash with a NULL pointer
2844 dereference if presented with an ASN.1 signature using the RSA PSS
2845 algorithm and absent mask generation function parameter. Since these
2846 routines are used to verify certificate signature algorithms this can be
2847 used to crash any certificate verification operation and exploited in a
2848 DoS attack. Any application which performs certificate verification is
2849 vulnerable including OpenSSL clients and servers which enable client
2852 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2856 *) X509_ATTRIBUTE memory leak
2858 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2859 memory. This structure is used by the PKCS#7 and CMS routines so any
2860 application which reads PKCS#7 or CMS data from untrusted sources is
2861 affected. SSL/TLS is not affected.
2863 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2868 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2869 This changes the decoding behaviour for some invalid messages,
2870 though the change is mostly in the more lenient direction, and
2871 legacy behaviour is preserved as much as possible.
2874 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2876 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2878 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2880 *) Alternate chains certificate forgery
2882 During certificate verification, OpenSSL will attempt to find an
2883 alternative certificate chain if the first attempt to build such a chain
2884 fails. An error in the implementation of this logic can mean that an
2885 attacker could cause certain checks on untrusted certificates to be
2886 bypassed, such as the CA flag, enabling them to use a valid leaf
2887 certificate to act as a CA and "issue" an invalid certificate.
2889 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2893 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2895 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2896 incompatibility in the handling of HMAC. The previous ABI has now been
2900 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2902 *) Malformed ECParameters causes infinite loop
2904 When processing an ECParameters structure OpenSSL enters an infinite loop
2905 if the curve specified is over a specially malformed binary polynomial
2908 This can be used to perform denial of service against any
2909 system which processes public keys, certificate requests or
2910 certificates. This includes TLS clients and TLS servers with
2911 client authentication enabled.
2913 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2917 *) Exploitable out-of-bounds read in X509_cmp_time
2919 X509_cmp_time does not properly check the length of the ASN1_TIME
2920 string and can read a few bytes out of bounds. In addition,
2921 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2924 An attacker can use this to craft malformed certificates and CRLs of
2925 various sizes and potentially cause a segmentation fault, resulting in
2926 a DoS on applications that verify certificates or CRLs. TLS clients
2927 that verify CRLs are affected. TLS clients and servers with client
2928 authentication enabled may be affected if they use custom verification
2931 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2932 independently by Hanno Böck.
2936 *) PKCS7 crash with missing EnvelopedContent
2938 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2939 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2940 with missing content and trigger a NULL pointer dereference on parsing.
2942 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2943 structures from untrusted sources are affected. OpenSSL clients and
2944 servers are not affected.
2946 This issue was reported to OpenSSL by Michal Zalewski (Google).
2950 *) CMS verify infinite loop with unknown hash function
2952 When verifying a signedData message the CMS code can enter an infinite loop
2953 if presented with an unknown hash function OID. This can be used to perform
2954 denial of service against any system which verifies signedData messages using
2956 This issue was reported to OpenSSL by Johannes Bauer.
2960 *) Race condition handling NewSessionTicket
2962 If a NewSessionTicket is received by a multi-threaded client when attempting to
2963 reuse a previous ticket then a race condition can occur potentially leading to
2964 a double free of the ticket data.
2968 *) Only support 256-bit or stronger elliptic curves with the
2969 'ecdh_auto' setting (server) or by default (client). Of supported
2970 curves, prefer P-256 (both).
2973 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2975 *) ClientHello sigalgs DoS fix
2977 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2978 invalid signature algorithms extension a NULL pointer dereference will
2979 occur. This can be exploited in a DoS attack against the server.
2981 This issue was was reported to OpenSSL by David Ramos of Stanford
2984 [Stephen Henson and Matt Caswell]
2986 *) Multiblock corrupted pointer fix
2988 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2989 feature only applies on 64 bit x86 architecture platforms that support AES
2990 NI instructions. A defect in the implementation of "multiblock" can cause
2991 OpenSSL's internal write buffer to become incorrectly set to NULL when
2992 using non-blocking IO. Typically, when the user application is using a
2993 socket BIO for writing, this will only result in a failed connection.
2994 However if some other BIO is used then it is likely that a segmentation
2995 fault will be triggered, thus enabling a potential DoS attack.
2997 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
3001 *) Segmentation fault in DTLSv1_listen fix
3003 The DTLSv1_listen function is intended to be stateless and processes the
3004 initial ClientHello from many peers. It is common for user code to loop
3005 over the call to DTLSv1_listen until a valid ClientHello is received with
3006 an associated cookie. A defect in the implementation of DTLSv1_listen means
3007 that state is preserved in the SSL object from one invocation to the next
3008 that can lead to a segmentation fault. Errors processing the initial
3009 ClientHello can trigger this scenario. An example of such an error could be
3010 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
3013 This issue was reported to OpenSSL by Per Allansson.
3017 *) Segmentation fault in ASN1_TYPE_cmp fix
3019 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
3020 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
3021 certificate signature algorithm consistency this can be used to crash any
3022 certificate verification operation and exploited in a DoS attack. Any
3023 application which performs certificate verification is vulnerable including
3024 OpenSSL clients and servers which enable client authentication.
3028 *) Segmentation fault for invalid PSS parameters fix
3030 The signature verification routines will crash with a NULL pointer
3031 dereference if presented with an ASN.1 signature using the RSA PSS
3032 algorithm and invalid parameters. Since these routines are used to verify
3033 certificate signature algorithms this can be used to crash any
3034 certificate verification operation and exploited in a DoS attack. Any
3035 application which performs certificate verification is vulnerable including
3036 OpenSSL clients and servers which enable client authentication.
3038 This issue was was reported to OpenSSL by Brian Carpenter.
3042 *) ASN.1 structure reuse memory corruption fix
3044 Reusing a structure in ASN.1 parsing may allow an attacker to cause
3045 memory corruption via an invalid write. Such reuse is and has been
3046 strongly discouraged and is believed to be rare.
3048 Applications that parse structures containing CHOICE or ANY DEFINED BY
3049 components may be affected. Certificate parsing (d2i_X509 and related
3050 functions) are however not affected. OpenSSL clients and servers are
3055 *) PKCS7 NULL pointer dereferences fix
3057 The PKCS#7 parsing code does not handle missing outer ContentInfo
3058 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
3059 missing content and trigger a NULL pointer dereference on parsing.
3061 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
3062 otherwise parse PKCS#7 structures from untrusted sources are
3063 affected. OpenSSL clients and servers are not affected.
3065 This issue was reported to OpenSSL by Michal Zalewski (Google).
3069 *) DoS via reachable assert in SSLv2 servers fix
3071 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
3072 servers that both support SSLv2 and enable export cipher suites by sending
3073 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
3075 This issue was discovered by Sean Burford (Google) and Emilia Käsper
3076 (OpenSSL development team).
3080 *) Empty CKE with client auth and DHE fix
3082 If client auth is used then a server can seg fault in the event of a DHE
3083 ciphersuite being selected and a zero length ClientKeyExchange message
3084 being sent by the client. This could be exploited in a DoS attack.
3088 *) Handshake with unseeded PRNG fix
3090 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
3091 with an unseeded PRNG. The conditions are:
3092 - The client is on a platform where the PRNG has not been seeded
3093 automatically, and the user has not seeded manually
3094 - A protocol specific client method version has been used (i.e. not
3095 SSL_client_methodv23)
3096 - A ciphersuite is used that does not require additional random data from
3097 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
3099 If the handshake succeeds then the client random that has been used will
3100 have been generated from a PRNG with insufficient entropy and therefore the
3101 output may be predictable.
3103 For example using the following command with an unseeded openssl will
3104 succeed on an unpatched platform:
3106 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
3110 *) Use After Free following d2i_ECPrivatekey error fix
3112 A malformed EC private key file consumed via the d2i_ECPrivateKey function
3113 could cause a use after free condition. This, in turn, could cause a double
3114 free in several private key parsing functions (such as d2i_PrivateKey
3115 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
3116 for applications that receive EC private keys from untrusted
3117 sources. This scenario is considered rare.
3119 This issue was discovered by the BoringSSL project and fixed in their
3124 *) X509_to_X509_REQ NULL pointer deref fix
3126 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
3127 the certificate key is invalid. This function is rarely used in practice.
3129 This issue was discovered by Brian Carpenter.
3133 *) Removed the export ciphers from the DEFAULT ciphers
3136 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
3138 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
3139 ARMv5 through ARMv8, as opposite to "locking" it to single one.
3140 So far those who have to target multiple platforms would compromise
3141 and argue that binary targeting say ARMv5 would still execute on
3142 ARMv8. "Universal" build resolves this compromise by providing
3143 near-optimal performance even on newer platforms.
3146 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
3147 (other platforms pending).
3148 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
3150 *) Add support for the SignedCertificateTimestampList certificate and
3151 OCSP response extensions from RFC6962.
3154 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3155 for corner cases. (Certain input points at infinity could lead to
3156 bogus results, with non-infinity inputs mapped to infinity too.)
3159 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
3160 This covers AES, SHA256/512 and GHASH. "Initial" means that most
3161 common cases are optimized and there still is room for further
3162 improvements. Vector Permutation AES for Altivec is also added.
3165 *) Add support for little-endian ppc64 Linux target.
3166 [Marcelo Cerri (IBM)]
3168 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
3169 SHA1, SHA256 and GHASH. "Initial" means that most common cases
3170 are optimized and there still is room for further improvements.
3171 Both 32- and 64-bit modes are supported.
3172 [Andy Polyakov, Ard Biesheuvel (Linaro)]
3174 *) Improved ARMv7 NEON support.
3177 *) Support for SPARC Architecture 2011 crypto extensions, first
3178 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
3179 SHA256/512, MD5, GHASH and modular exponentiation.
3180 [Andy Polyakov, David Miller]
3182 *) Accelerated modular exponentiation for Intel processors, a.k.a.
3184 [Shay Gueron & Vlad Krasnov (Intel Corp)]
3186 *) Support for new and upcoming Intel processors, including AVX2,
3187 BMI and SHA ISA extensions. This includes additional "stitched"
3188 implementations, AESNI-SHA256 and GCM, and multi-buffer support
3191 This work was sponsored by Intel Corp.
3194 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
3195 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
3196 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
3199 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3200 this fixes a limitation in previous versions of OpenSSL.
3203 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3204 MGF1 digest and OAEP label.
3207 *) Add EVP support for key wrapping algorithms, to avoid problems with
3208 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3209 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3210 algorithms and include tests cases.
3213 *) Add functions to allocate and set the fields of an ECDSA_METHOD
3215 [Douglas E. Engert, Steve Henson]
3217 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
3218 difference in days and seconds between two tm or ASN1_TIME structures.
3221 *) Add -rev test option to s_server to just reverse order of characters
3222 received by client and send back to server. Also prints an abbreviated
3223 summary of the connection parameters.
3226 *) New option -brief for s_client and s_server to print out a brief summary
3227 of connection parameters.
3230 *) Add callbacks for arbitrary TLS extensions.
3231 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
3233 *) New option -crl_download in several openssl utilities to download CRLs
3234 from CRLDP extension in certificates.
3237 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
3240 *) New function X509_CRL_diff to generate a delta CRL from the difference
3241 of two full CRLs. Add support to "crl" utility.
3244 *) New functions to set lookup_crls function and to retrieve
3245 X509_STORE from X509_STORE_CTX.
3248 *) Print out deprecated issuer and subject unique ID fields in
3252 *) Extend OCSP I/O functions so they can be used for simple general purpose
3253 HTTP as well as OCSP. New wrapper function which can be used to download
3254 CRLs using the OCSP API.
3257 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
3260 *) SSL_CONF* functions. These provide a common framework for application
3261 configuration using configuration files or command lines.
3264 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
3265 message callback and prints the results. Needs compile time option
3266 "enable-ssl-trace". New options to s_client and s_server to enable
3270 *) New ctrl and macro to retrieve supported points extensions.
3271 Print out extension in s_server and s_client.
3274 *) New functions to retrieve certificate signature and signature
3278 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
3282 *) New Suite B modes for TLS code. These use and enforce the requirements
3283 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
3284 only use Suite B curves. The Suite B modes can be set by using the
3285 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
3288 *) New chain verification flags for Suite B levels of security. Check
3289 algorithms are acceptable when flags are set in X509_verify_cert.
3292 *) Make tls1_check_chain return a set of flags indicating checks passed
3293 by a certificate chain. Add additional tests to handle client
3294 certificates: checks for matching certificate type and issuer name
3298 *) If an attempt is made to use a signature algorithm not in the peer
3299 preference list abort the handshake. If client has no suitable
3300 signature algorithms in response to a certificate request do not
3301 use the certificate.
3304 *) If server EC tmp key is not in client preference list abort handshake.
3307 *) Add support for certificate stores in CERT structure. This makes it
3308 possible to have different stores per SSL structure or one store in
3309 the parent SSL_CTX. Include distinct stores for certificate chain
3310 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3311 to build and store a certificate chain in CERT structure: returning
3312 an error if the chain cannot be built: this will allow applications
3313 to test if a chain is correctly configured.
3315 Note: if the CERT based stores are not set then the parent SSL_CTX
3316 store is used to retain compatibility with existing behaviour.
3320 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3321 mask based on the current session, check mask when sending client
3322 hello and checking the requested ciphersuite.
3325 *) New ctrls to retrieve and set certificate types in a certificate
3326 request message. Print out received values in s_client. If certificate
3327 types is not set with custom values set sensible values based on
3328 supported signature algorithms.
3331 *) Support for distinct client and server supported signature algorithms.
3334 *) Add certificate callback. If set this is called whenever a certificate
3335 is required by client or server. An application can decide which
3336 certificate chain to present based on arbitrary criteria: for example
3337 supported signature algorithms. Add very simple example to s_server.
3338 This fixes many of the problems and restrictions of the existing client
3339 certificate callback: for example you can now clear an existing
3340 certificate and specify the whole chain.
3343 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3344 the certificate can be used for (if anything). Set valid_flags field
3345 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3346 to have similar checks in it.
3348 Add new "cert_flags" field to CERT structure and include a "strict mode".
3349 This enforces some TLS certificate requirements (such as only permitting
3350 certificate signature algorithms contained in the supported algorithms
3351 extension) which some implementations ignore: this option should be used
3352 with caution as it could cause interoperability issues.
3355 *) Update and tidy signature algorithm extension processing. Work out
3356 shared signature algorithms based on preferences and peer algorithms
3357 and print them out in s_client and s_server. Abort handshake if no
3358 shared signature algorithms.
3361 *) Add new functions to allow customised supported signature algorithms
3362 for SSL and SSL_CTX structures. Add options to s_client and s_server
3366 *) New function SSL_certs_clear() to delete all references to certificates
3367 from an SSL structure. Before this once a certificate had been added
3368 it couldn't be removed.
3371 *) Integrate hostname, email address and IP address checking with certificate
3372 verification. New verify options supporting checking in openssl utility.
3375 *) Fixes and wildcard matching support to hostname and email checking
3376 functions. Add manual page.
3377 [Florian Weimer (Red Hat Product Security Team)]
3379 *) New functions to check a hostname email or IP address against a
3380 certificate. Add options x509 utility to print results of checks against
3384 *) Fix OCSP checking.
3385 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3387 *) Initial experimental support for explicitly trusted non-root CAs.
3388 OpenSSL still tries to build a complete chain to a root but if an
3389 intermediate CA has a trust setting included that is used. The first
3390 setting is used: whether to trust (e.g., -addtrust option to the x509
3394 *) Add -trusted_first option which attempts to find certificates in the
3395 trusted store even if an untrusted chain is also supplied.
3398 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3399 platform support for Linux and Android.
3402 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3405 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3406 When in FIPS mode the approved implementations are used as normal,
3407 when not in FIPS mode the internal unapproved versions are used instead.
3408 This means that the FIPS capable OpenSSL isn't forced to use the
3409 (often lower performance) FIPS implementations outside FIPS mode.
3412 *) Transparently support X9.42 DH parameters when calling
3413 PEM_read_bio_DHparameters. This means existing applications can handle
3414 the new parameter format automatically.
3417 *) Initial experimental support for X9.42 DH parameter format: mainly
3418 to support use of 'q' parameter for RFC5114 parameters.
3421 *) Add DH parameters from RFC5114 including test data to dhtest.
3424 *) Support for automatic EC temporary key parameter selection. If enabled
3425 the most preferred EC parameters are automatically used instead of
3426 hardcoded fixed parameters. Now a server just has to call:
3427 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3428 support ECDH and use the most appropriate parameters.
3431 *) Enhance and tidy EC curve and point format TLS extension code. Use
3432 static structures instead of allocation if default values are used.
3433 New ctrls to set curves we wish to support and to retrieve shared curves.
3434 Print out shared curves in s_server. New options to s_server and s_client
3435 to set list of supported curves.
3438 *) New ctrls to retrieve supported signature algorithms and
3439 supported curve values as an array of NIDs. Extend openssl utility
3440 to print out received values.
3443 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3444 between NIDs and the more common NIST names such as "P-256". Enhance
3445 ecparam utility and ECC method to recognise the NIST names for curves.
3448 *) Enhance SSL/TLS certificate chain handling to support different
3449 chains for each certificate instead of one chain in the parent SSL_CTX.
3452 *) Support for fixed DH ciphersuite client authentication: where both
3453 server and client use DH certificates with common parameters.
3456 *) Support for fixed DH ciphersuites: those requiring DH server
3460 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3462 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3463 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3464 X509_CINF_get_signature were reverted post internal team review.
3466 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3468 *) Build fixes for the Windows and OpenVMS platforms
3469 [Matt Caswell and Richard Levitte]
3471 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3473 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3474 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3475 dereference. This could lead to a Denial Of Service attack. Thanks to
3476 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3480 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3481 dtls1_buffer_record function under certain conditions. In particular this
3482 could occur if an attacker sent repeated DTLS records with the same
3483 sequence number but for the next epoch. The memory leak could be exploited
3484 by an attacker in a Denial of Service attack through memory exhaustion.
3485 Thanks to Chris Mueller for reporting this issue.
3489 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3490 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3491 method would be set to NULL which could later result in a NULL pointer
3492 dereference. Thanks to Frank Schmirler for reporting this issue.
3496 *) Abort handshake if server key exchange message is omitted for ephemeral
3499 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3500 reporting this issue.
3504 *) Remove non-export ephemeral RSA code on client and server. This code
3505 violated the TLS standard by allowing the use of temporary RSA keys in
3506 non-export ciphersuites and could be used by a server to effectively
3507 downgrade the RSA key length used to a value smaller than the server
3508 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3509 INRIA or reporting this issue.
3513 *) Fixed issue where DH client certificates are accepted without verification.
3514 An OpenSSL server will accept a DH certificate for client authentication
3515 without the certificate verify message. This effectively allows a client to
3516 authenticate without the use of a private key. This only affects servers
3517 which trust a client certificate authority which issues certificates
3518 containing DH keys: these are extremely rare and hardly ever encountered.
3519 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3524 *) Ensure that the session ID context of an SSL is updated when its
3525 SSL_CTX is updated via SSL_set_SSL_CTX.
3527 The session ID context is typically set from the parent SSL_CTX,
3528 and can vary with the CTX.
3531 *) Fix various certificate fingerprint issues.
3533 By using non-DER or invalid encodings outside the signed portion of a
3534 certificate the fingerprint can be changed without breaking the signature.
3535 Although no details of the signed portion of the certificate can be changed
3536 this can cause problems with some applications: e.g. those using the
3537 certificate fingerprint for blacklists.
3539 1. Reject signatures with non zero unused bits.
3541 If the BIT STRING containing the signature has non zero unused bits reject
3542 the signature. All current signature algorithms require zero unused bits.
3544 2. Check certificate algorithm consistency.
3546 Check the AlgorithmIdentifier inside TBS matches the one in the
3547 certificate signature. NB: this will result in signature failure
3548 errors for some broken certificates.
3550 Thanks to Konrad Kraszewski from Google for reporting this issue.
3552 3. Check DSA/ECDSA signatures use DER.
3554 Re-encode DSA/ECDSA signatures and compare with the original received
3555 signature. Return an error if there is a mismatch.
3557 This will reject various cases including garbage after signature
3558 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3559 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3560 (negative or with leading zeroes).
3562 Further analysis was conducted and fixes were developed by Stephen Henson
3563 of the OpenSSL core team.
3568 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3569 results on some platforms, including x86_64. This bug occurs at random
3570 with a very low probability, and is not known to be exploitable in any
3571 way, though its exact impact is difficult to determine. Thanks to Pieter
3572 Wuille (Blockstream) who reported this issue and also suggested an initial
3573 fix. Further analysis was conducted by the OpenSSL development team and
3574 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3575 the OpenSSL core team.
3579 *) Do not resume sessions on the server if the negotiated protocol
3580 version does not match the session's version. Resuming with a different
3581 version, while not strictly forbidden by the RFC, is of questionable
3582 sanity and breaks all known clients.
3583 [David Benjamin, Emilia Käsper]
3585 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3586 early CCS messages during renegotiation. (Note that because
3587 renegotiation is encrypted, this early CCS was not exploitable.)
3590 *) Tighten client-side session ticket handling during renegotiation:
3591 ensure that the client only accepts a session ticket if the server sends
3592 the extension anew in the ServerHello. Previously, a TLS client would
3593 reuse the old extension state and thus accept a session ticket if one was
3594 announced in the initial ServerHello.
3596 Similarly, ensure that the client requires a session ticket if one
3597 was advertised in the ServerHello. Previously, a TLS client would
3598 ignore a missing NewSessionTicket message.
3601 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3603 *) SRTP Memory Leak.
3605 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3606 sends a carefully crafted handshake message, to cause OpenSSL to fail
3607 to free up to 64k of memory causing a memory leak. This could be
3608 exploited in a Denial Of Service attack. This issue affects OpenSSL
3609 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3610 whether SRTP is used or configured. Implementations of OpenSSL that
3611 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3613 The fix was developed by the OpenSSL team.
3617 *) Session Ticket Memory Leak.
3619 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3620 integrity of that ticket is first verified. In the event of a session
3621 ticket integrity check failing, OpenSSL will fail to free memory
3622 causing a memory leak. By sending a large number of invalid session
3623 tickets an attacker could exploit this issue in a Denial Of Service
3628 *) Build option no-ssl3 is incomplete.
3630 When OpenSSL is configured with "no-ssl3" as a build option, servers
3631 could accept and complete a SSL 3.0 handshake, and clients could be
3632 configured to send them.
3634 [Akamai and the OpenSSL team]
3636 *) Add support for TLS_FALLBACK_SCSV.
3637 Client applications doing fallback retries should call
3638 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3640 [Adam Langley, Bodo Moeller]
3642 *) Add additional DigestInfo checks.
3644 Re-encode DigestInto in DER and check against the original when
3645 verifying RSA signature: this will reject any improperly encoded
3646 DigestInfo structures.
3648 Note: this is a precautionary measure and no attacks are currently known.
3652 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3654 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3655 SRP code can be overrun an internal buffer. Add sanity check that
3656 g, A, B < N to SRP code.
3658 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3659 Group for discovering this issue.
3663 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3664 TLS 1.0 instead of higher protocol versions when the ClientHello message
3665 is badly fragmented. This allows a man-in-the-middle attacker to force a
3666 downgrade to TLS 1.0 even if both the server and the client support a
3667 higher protocol version, by modifying the client's TLS records.
3669 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3670 researching this issue.
3674 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3675 to a denial of service attack. A malicious server can crash the client
3676 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3677 ciphersuite and sending carefully crafted handshake messages.
3679 Thanks to Felix Gröbert (Google) for discovering and researching this
3684 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3685 to leak memory. This can be exploited through a Denial of Service attack.
3686 Thanks to Adam Langley for discovering and researching this issue.
3690 *) An attacker can force openssl to consume large amounts of memory whilst
3691 processing DTLS handshake messages. This can be exploited through a
3692 Denial of Service attack.
3693 Thanks to Adam Langley for discovering and researching this issue.
3697 *) An attacker can force an error condition which causes openssl to crash
3698 whilst processing DTLS packets due to memory being freed twice. This
3699 can be exploited through a Denial of Service attack.
3700 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3705 *) If a multithreaded client connects to a malicious server using a resumed
3706 session and the server sends an ec point format extension it could write
3707 up to 255 bytes to freed memory.
3709 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3714 *) A malicious server can crash an OpenSSL client with a null pointer
3715 dereference (read) by specifying an SRP ciphersuite even though it was not
3716 properly negotiated with the client. This can be exploited through a
3717 Denial of Service attack.
3719 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3720 discovering and researching this issue.
3724 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3725 X509_name_oneline, X509_name_print_ex et al. to leak some information
3726 from the stack. Applications may be affected if they echo pretty printing
3727 output to the attacker.
3729 Thanks to Ivan Fratric (Google) for discovering this issue.
3731 [Emilia Käsper, and Steve Henson]
3733 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3734 for corner cases. (Certain input points at infinity could lead to
3735 bogus results, with non-infinity inputs mapped to infinity too.)
3738 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3740 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3741 handshake can force the use of weak keying material in OpenSSL
3742 SSL/TLS clients and servers.
3744 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3745 researching this issue. (CVE-2014-0224)
3746 [KIKUCHI Masashi, Steve Henson]
3748 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3749 OpenSSL DTLS client the code can be made to recurse eventually crashing
3752 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3754 [Imre Rad, Steve Henson]
3756 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3757 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3758 client or server. This is potentially exploitable to run arbitrary
3759 code on a vulnerable client or server.
3761 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3762 [Jüri Aedla, Steve Henson]
3764 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3765 are subject to a denial of service attack.
3767 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3768 this issue. (CVE-2014-3470)
3769 [Felix Gröbert, Ivan Fratric, Steve Henson]
3771 *) Harmonize version and its documentation. -f flag is used to display
3773 [mancha <mancha1@zoho.com>]
3775 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3776 in i2d_ECPrivateKey.
3777 [mancha <mancha1@zoho.com>]
3779 *) Fix some double frees. These are not thought to be exploitable.
3780 [mancha <mancha1@zoho.com>]
3782 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3784 *) A missing bounds check in the handling of the TLS heartbeat extension
3785 can be used to reveal up to 64k of memory to a connected client or
3788 Thanks for Neel Mehta of Google Security for discovering this bug and to
3789 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3790 preparing the fix (CVE-2014-0160)
3791 [Adam Langley, Bodo Moeller]
3793 *) Fix for the attack described in the paper "Recovering OpenSSL
3794 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3795 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3796 http://eprint.iacr.org/2014/140
3798 Thanks to Yuval Yarom and Naomi Benger for discovering this
3799 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3800 [Yuval Yarom and Naomi Benger]
3802 *) TLS pad extension: draft-agl-tls-padding-03
3804 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3805 TLS client Hello record length value would otherwise be > 255 and
3806 less that 512 pad with a dummy extension containing zeroes so it
3807 is at least 512 bytes long.
3809 [Adam Langley, Steve Henson]
3811 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3813 *) Fix for TLS record tampering bug. A carefully crafted invalid
3814 handshake could crash OpenSSL with a NULL pointer exception.
3815 Thanks to Anton Johansson for reporting this issues.
3818 *) Keep original DTLS digest and encryption contexts in retransmission
3819 structures so we can use the previous session parameters if they need
3820 to be resent. (CVE-2013-6450)
3823 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3824 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3825 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3826 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3827 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3828 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3829 [Rob Stradling, Adam Langley]
3831 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3833 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3834 supporting platforms or when small records were transferred.
3835 [Andy Polyakov, Steve Henson]
3837 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3839 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3841 This addresses the flaw in CBC record processing discovered by
3842 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3843 at: http://www.isg.rhul.ac.uk/tls/
3845 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3846 Security Group at Royal Holloway, University of London
3847 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3848 Emilia Käsper for the initial patch.
3850 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3852 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3853 ciphersuites which can be exploited in a denial of service attack.
3854 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3855 and detecting this bug and to Wolfgang Ettlinger
3856 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3860 *) Return an error when checking OCSP signatures when key is NULL.
3861 This fixes a DoS attack. (CVE-2013-0166)
3864 *) Make openssl verify return errors.
3865 [Chris Palmer <palmer@google.com> and Ben Laurie]
3867 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3868 the right response is stapled. Also change SSL_get_certificate()
3869 so it returns the certificate actually sent.
3870 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3871 [Rob Stradling <rob.stradling@comodo.com>]
3873 *) Fix possible deadlock when decoding public keys.
3876 *) Don't use TLS 1.0 record version number in initial client hello
3880 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3882 *) Sanity check record length before skipping explicit IV in TLS
3883 1.2, 1.1 and DTLS to fix DoS attack.
3885 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3886 fuzzing as a service testing platform.
3890 *) Initialise tkeylen properly when encrypting CMS messages.
3891 Thanks to Solar Designer of Openwall for reporting this issue.
3894 *) In FIPS mode don't try to use composite ciphers as they are not
3898 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3900 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3901 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3902 mean any application compiled against OpenSSL 1.0.0 headers setting
3903 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3904 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3905 0x10000000L Any application which was previously compiled against
3906 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3907 will need to be recompiled as a result. Letting be results in
3908 inability to disable specifically TLS 1.1 and in client context,
3909 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3912 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3913 disable just protocol X, but all protocols above X *if* there are
3914 protocols *below* X still enabled. In more practical terms it means
3915 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3916 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3917 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3921 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3923 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3924 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3925 in CRYPTO_realloc_clean.
3927 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3928 issue and to Adam Langley <agl@chromium.org> for fixing it.
3930 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3932 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3935 *) Workarounds for some broken servers that "hang" if a client hello
3936 record length exceeds 255 bytes.
3938 1. Do not use record version number > TLS 1.0 in initial client
3939 hello: some (but not all) hanging servers will now work.
3940 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3941 the number of ciphers sent in the client hello. This should be
3942 set to an even number, such as 50, for example by passing:
3943 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3944 Most broken servers should now work.
3945 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3946 TLS 1.2 client support entirely.
3949 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3952 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3954 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3955 STRING form instead of a DigestInfo.
3958 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3959 and the RSA_sign/RSA_verify functions. This was made more apparent when
3960 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3961 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3962 the correct format in RSA_verify so both forms transparently work.
3965 *) Some servers which support TLS 1.0 can choke if we initially indicate
3966 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3967 encrypted premaster secret. As a workaround use the maximum permitted
3968 client version in client hello, this should keep such servers happy
3969 and still work with previous versions of OpenSSL.
3972 *) Add support for TLS/DTLS heartbeats.
3973 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3975 *) Add support for SCTP.
3976 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3978 *) Improved PRNG seeding for VOS.
3979 [Paul Green <Paul.Green@stratus.com>]
3981 *) Extensive assembler packs updates, most notably:
3983 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3984 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3985 - x86_64: bit-sliced AES implementation;
3986 - ARM: NEON support, contemporary platforms optimizations;
3987 - s390x: z196 support;
3988 - *: GHASH and GF(2^m) multiplication implementations;
3992 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3993 (removal of unnecessary code)
3994 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3996 *) Add TLS key material exporter from RFC 5705.
3999 *) Add DTLS-SRTP negotiation from RFC 5764.
4002 *) Add Next Protocol Negotiation,
4003 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
4004 disabled with a no-npn flag to config or Configure. Code donated
4006 [Adam Langley <agl@google.com> and Ben Laurie]
4008 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
4009 NIST-P256, NIST-P521, with constant-time single point multiplication on
4010 typical inputs. Compiler support for the nonstandard type __uint128_t is
4011 required to use this (present in gcc 4.4 and later, for 64-bit builds).
4012 Code made available under Apache License version 2.0.
4014 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
4015 line to include this in your build of OpenSSL, and run "make depend" (or
4016 "make update"). This enables the following EC_METHODs:
4018 EC_GFp_nistp224_method()
4019 EC_GFp_nistp256_method()
4020 EC_GFp_nistp521_method()
4022 EC_GROUP_new_by_curve_name() will automatically use these (while
4023 EC_GROUP_new_curve_GFp() currently prefers the more flexible
4025 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
4027 *) Use type ossl_ssize_t instead of ssize_t which isn't available on
4028 all platforms. Move ssize_t definition from e_os.h to the public
4029 header file e_os2.h as it now appears in public header file cms.h
4032 *) New -sigopt option to the ca, req and x509 utilities. Additional
4033 signature parameters can be passed using this option and in
4037 *) Add RSA PSS signing function. This will generate and set the
4038 appropriate AlgorithmIdentifiers for PSS based on those in the
4039 corresponding EVP_MD_CTX structure. No application support yet.
4042 *) Support for companion algorithm specific ASN1 signing routines.
4043 New function ASN1_item_sign_ctx() signs a pre-initialised
4044 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
4045 the appropriate parameters.
4048 *) Add new algorithm specific ASN1 verification initialisation function
4049 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
4050 handling will be the same no matter what EVP_PKEY_METHOD is used.
4051 Add a PSS handler to support verification of PSS signatures: checked
4052 against a number of sample certificates.
4055 *) Add signature printing for PSS. Add PSS OIDs.
4056 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
4058 *) Add algorithm specific signature printing. An individual ASN1 method
4059 can now print out signatures instead of the standard hex dump.
4061 More complex signatures (e.g. PSS) can print out more meaningful
4062 information. Include DSA version that prints out the signature
4066 *) Password based recipient info support for CMS library: implementing
4070 *) Split password based encryption into PBES2 and PBKDF2 functions. This
4071 neatly separates the code into cipher and PBE sections and is required
4072 for some algorithms that split PBES2 into separate pieces (such as
4073 password based CMS).
4076 *) Session-handling fixes:
4077 - Fix handling of connections that are resuming with a session ID,
4078 but also support Session Tickets.
4079 - Fix a bug that suppressed issuing of a new ticket if the client
4080 presented a ticket with an expired session.
4081 - Try to set the ticket lifetime hint to something reasonable.
4082 - Make tickets shorter by excluding irrelevant information.
4083 - On the client side, don't ignore renewed tickets.
4084 [Adam Langley, Bodo Moeller (Google)]
4086 *) Fix PSK session representation.
4089 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
4091 This work was sponsored by Intel.
4094 *) Add GCM support to TLS library. Some custom code is needed to split
4095 the IV between the fixed (from PRF) and explicit (from TLS record)
4096 portions. This adds all GCM ciphersuites supported by RFC5288 and
4097 RFC5289. Generalise some AES* cipherstrings to include GCM and
4098 add a special AESGCM string for GCM only.
4101 *) Expand range of ctrls for AES GCM. Permit setting invocation
4102 field on decrypt and retrieval of invocation field only on encrypt.
4105 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
4106 As required by RFC5289 these ciphersuites cannot be used if for
4107 versions of TLS earlier than 1.2.
4110 *) For FIPS capable OpenSSL interpret a NULL default public key method
4111 as unset and return the appropriate default but do *not* set the default.
4112 This means we can return the appropriate method in applications that
4113 switch between FIPS and non-FIPS modes.
4116 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
4117 ENGINE is used then we cannot handle that in the FIPS module so we
4118 keep original code iff non-FIPS operations are allowed.
4121 *) Add -attime option to openssl utilities.
4122 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
4124 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
4127 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
4128 FIPS EC methods unconditionally for now.
4131 *) New build option no-ec2m to disable characteristic 2 code.
4134 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
4135 all cases can be covered as some introduce binary incompatibilities.
4138 *) Redirect RSA operations to FIPS module including keygen,
4139 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
4142 *) Add similar low level API blocking to ciphers.
4145 *) Low level digest APIs are not approved in FIPS mode: any attempt
4146 to use these will cause a fatal error. Applications that *really* want
4147 to use them can use the private_* version instead.
4150 *) Redirect cipher operations to FIPS module for FIPS builds.
4153 *) Redirect digest operations to FIPS module for FIPS builds.
4156 *) Update build system to add "fips" flag which will link in fipscanister.o
4157 for static and shared library builds embedding a signature if needed.
4160 *) Output TLS supported curves in preference order instead of numerical
4161 order. This is currently hardcoded for the highest order curves first.
4162 This should be configurable so applications can judge speed vs strength.
4165 *) Add TLS v1.2 server support for client authentication.
4168 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
4172 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
4173 FIPS modules versions.
4176 *) Add TLS v1.2 client side support for client authentication. Keep cache
4177 of handshake records longer as we don't know the hash algorithm to use
4178 until after the certificate request message is received.
4181 *) Initial TLS v1.2 client support. Add a default signature algorithms
4182 extension including all the algorithms we support. Parse new signature
4183 format in client key exchange. Relax some ECC signing restrictions for
4184 TLS v1.2 as indicated in RFC5246.
4187 *) Add server support for TLS v1.2 signature algorithms extension. Switch
4188 to new signature format when needed using client digest preference.
4189 All server ciphersuites should now work correctly in TLS v1.2. No client
4190 support yet and no support for client certificates.
4193 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
4194 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
4195 ciphersuites. At present only RSA key exchange ciphersuites work with
4196 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
4197 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
4198 and version checking.
4201 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
4202 with this defined it will not be affected by any changes to ssl internal
4203 structures. Add several utility functions to allow openssl application
4204 to work with OPENSSL_NO_SSL_INTERN defined.
4207 *) A long standing patch to add support for SRP from EdelWeb (Peter
4208 Sylvester and Christophe Renou) was integrated.
4209 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
4210 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
4213 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
4216 *) Permit abbreviated handshakes when renegotiating using the function
4217 SSL_renegotiate_abbreviated().
4218 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4220 *) Add call to ENGINE_register_all_complete() to
4221 ENGINE_load_builtin_engines(), so some implementations get used
4222 automatically instead of needing explicit application support.
4225 *) Add support for TLS key exporter as described in RFC5705.
4226 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
4228 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
4229 a few changes are required:
4231 Add SSL_OP_NO_TLSv1_1 flag.
4232 Add TLSv1_1 methods.
4233 Update version checking logic to handle version 1.1.
4234 Add explicit IV handling (ported from DTLS code).
4235 Add command line options to s_client/s_server.
4238 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
4240 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
4241 in CMS and PKCS7 code. When RSA decryption fails use a random key for
4242 content decryption and always return the same error. Note: this attack
4243 needs on average 2^20 messages so it only affects automated senders. The
4244 old behaviour can be re-enabled in the CMS code by setting the
4245 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
4246 an MMA defence is not necessary.
4247 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
4248 this issue. (CVE-2012-0884)
4251 *) Fix CVE-2011-4619: make sure we really are receiving a
4252 client hello before rejecting multiple SGC restarts. Thanks to
4253 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
4256 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
4258 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
4259 Thanks to Antonio Martin, Enterprise Secure Access Research and
4260 Development, Cisco Systems, Inc. for discovering this bug and
4261 preparing a fix. (CVE-2012-0050)
4264 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
4266 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
4267 of the Vaudenay padding oracle attack on CBC mode encryption
4268 which enables an efficient plaintext recovery attack against
4269 the OpenSSL implementation of DTLS. Their attack exploits timing
4270 differences arising during decryption processing. A research
4271 paper describing this attack can be found at:
4272 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
4273 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4274 Security Group at Royal Holloway, University of London
4275 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
4276 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
4277 for preparing the fix. (CVE-2011-4108)
4278 [Robin Seggelmann, Michael Tuexen]
4280 *) Clear bytes used for block padding of SSL 3.0 records.
4282 [Adam Langley (Google)]
4284 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
4285 Kadianakis <desnacked@gmail.com> for discovering this issue and
4286 Adam Langley for preparing the fix. (CVE-2011-4619)
4287 [Adam Langley (Google)]
4289 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
4290 [Andrey Kulikov <amdeich@gmail.com>]
4292 *) Prevent malformed RFC3779 data triggering an assertion failure.
4293 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
4294 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
4295 [Rob Austein <sra@hactrn.net>]
4297 *) Improved PRNG seeding for VOS.
4298 [Paul Green <Paul.Green@stratus.com>]
4300 *) Fix ssl_ciph.c set-up race.
4301 [Adam Langley (Google)]
4303 *) Fix spurious failures in ecdsatest.c.
4304 [Emilia Käsper (Google)]
4306 *) Fix the BIO_f_buffer() implementation (which was mixing different
4307 interpretations of the '..._len' fields).
4308 [Adam Langley (Google)]
4310 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4311 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4312 threads won't reuse the same blinding coefficients.
4314 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4315 lock to call BN_BLINDING_invert_ex, and avoids one use of
4316 BN_BLINDING_update for each BN_BLINDING structure (previously,
4317 the last update always remained unused).
4318 [Emilia Käsper (Google)]
4320 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4321 [Bob Buckholz (Google)]
4323 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4325 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4326 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4327 [Kaspar Brand <ossl@velox.ch>]
4329 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4330 for multi-threaded use of ECDH. (CVE-2011-3210)
4331 [Adam Langley (Google)]
4333 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4336 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4337 signature public key algorithm by using OID xref utilities instead.
4338 Before this you could only use some ECC ciphersuites with SHA1 only.
4341 *) Add protection against ECDSA timing attacks as mentioned in the paper
4342 by Billy Bob Brumley and Nicola Tuveri, see:
4344 http://eprint.iacr.org/2011/232.pdf
4346 [Billy Bob Brumley and Nicola Tuveri]
4348 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4350 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4351 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4353 *) Fix bug in string printing code: if *any* escaping is enabled we must
4354 escape the escape character (backslash) or the resulting string is
4358 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4360 *) Disable code workaround for ancient and obsolete Netscape browsers
4361 and servers: an attacker can use it in a ciphersuite downgrade attack.
4362 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4365 *) Fixed J-PAKE implementation error, originally discovered by
4366 Sebastien Martini, further info and confirmation from Stefan
4367 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4370 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4372 *) Fix extension code to avoid race conditions which can result in a buffer
4373 overrun vulnerability: resumed sessions must not be modified as they can
4374 be shared by multiple threads. CVE-2010-3864
4377 *) Fix WIN32 build system to correctly link an ENGINE directory into
4381 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4383 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4385 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4387 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4389 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4390 context. The operation can be customised via the ctrl mechanism in
4391 case ENGINEs want to include additional functionality.
4394 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4397 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4398 output hashes compatible with older versions of OpenSSL.
4399 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4401 *) Fix compression algorithm handling: if resuming a session use the
4402 compression algorithm of the resumed session instead of determining
4403 it from client hello again. Don't allow server to change algorithm.
4406 *) Add load_crls() function to apps tidying load_certs() too. Add option
4407 to verify utility to allow additional CRLs to be included.
4410 *) Update OCSP request code to permit adding custom headers to the request:
4411 some responders need this.
4414 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4416 [Julia Lawall <julia@diku.dk>]
4418 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4419 needlessly dereferenced structures, used obsolete functions and
4420 didn't handle all updated verify codes correctly.
4423 *) Disable MD2 in the default configuration.
4426 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4427 indicate the initial BIO being pushed or popped. This makes it possible
4428 to determine whether the BIO is the one explicitly called or as a result
4429 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4430 it handles reference counts correctly and doesn't zero out the I/O bio
4431 when it is not being explicitly popped. WARNING: applications which
4432 included workarounds for the old buggy behaviour will need to be modified
4433 or they could free up already freed BIOs.
4436 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4437 renaming to all platforms (within the 0.9.8 branch, this was
4438 done conditionally on Netware platforms to avoid a name clash).
4439 [Guenter <lists@gknw.net>]
4441 *) Add ECDHE and PSK support to DTLS.
4442 [Michael Tuexen <tuexen@fh-muenster.de>]
4444 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4448 *) Add "missing" function EVP_MD_flags() (without this the only way to
4449 retrieve a digest flags is by accessing the structure directly. Update
4450 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4451 or cipher is registered as in the "from" argument. Print out all
4452 registered digests in the dgst usage message instead of manually
4453 attempting to work them out.
4456 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4457 this allows the use of compression and extensions. Change default cipher
4458 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4459 by default unless an application cipher string requests it.
4462 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4463 key ids to find matching certificates and keys but some PKCS#12 files
4464 don't follow the (somewhat unwritten) rules and this strategy fails.
4465 Now just gather all certificates together and the first private key
4466 then look for the first certificate that matches the key.
4469 *) Support use of registered digest and cipher names for dgst and cipher
4470 commands instead of having to add each one as a special case. So now
4477 openssl dgst -sha256 foo
4479 and this works for ENGINE based algorithms too.
4483 *) Update Gost ENGINE to support parameter files.
4484 [Victor B. Wagner <vitus@cryptocom.ru>]
4486 *) Support GeneralizedTime in ca utility.
4487 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4489 *) Enhance the hash format used for certificate directory links. The new
4490 form uses the canonical encoding (meaning equivalent names will work
4491 even if they aren't identical) and uses SHA1 instead of MD5. This form
4492 is incompatible with the older format and as a result c_rehash should
4493 be used to rebuild symbolic links.
4496 *) Make PKCS#8 the default write format for private keys, replacing the
4497 traditional format. This form is standardised, more secure and doesn't
4498 include an implicit MD5 dependency.
4501 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4502 committed to OpenSSL should pass this lot as a minimum.
4505 *) Add session ticket override functionality for use by EAP-FAST.
4506 [Jouni Malinen <j@w1.fi>]
4508 *) Modify HMAC functions to return a value. Since these can be implemented
4509 in an ENGINE errors can occur.
4512 *) Type-checked OBJ_bsearch_ex.
4515 *) Type-checked OBJ_bsearch. Also some constification necessitated
4516 by type-checking. Still to come: TXT_DB, bsearch(?),
4517 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4521 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4522 seconds to a tm structure directly, instead of going through OS
4523 specific date routines. This avoids any issues with OS routines such
4524 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4525 and X509_time_adj_ex() to cover the extended range. The existing
4526 X509_time_adj() is still usable and will no longer have any date issues.
4529 *) Delta CRL support. New use deltas option which will attempt to locate
4530 and search any appropriate delta CRLs available.
4532 This work was sponsored by Google.
4535 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4536 code and add additional score elements. Validate alternate CRL paths
4537 as part of the CRL checking and indicate a new error "CRL path validation
4538 error" in this case. Applications wanting additional details can use
4539 the verify callback and check the new "parent" field. If this is not
4540 NULL CRL path validation is taking place. Existing applications won't
4541 see this because it requires extended CRL support which is off by
4544 This work was sponsored by Google.
4547 *) Support for freshest CRL extension.
4549 This work was sponsored by Google.
4552 *) Initial indirect CRL support. Currently only supported in the CRLs
4553 passed directly and not via lookup. Process certificate issuer
4554 CRL entry extension and lookup CRL entries by bother issuer name
4555 and serial number. Check and process CRL issuer entry in IDP extension.
4557 This work was sponsored by Google.
4560 *) Add support for distinct certificate and CRL paths. The CRL issuer
4561 certificate is validated separately in this case. Only enabled if
4562 an extended CRL support flag is set: this flag will enable additional
4563 CRL functionality in future.
4565 This work was sponsored by Google.
4568 *) Add support for policy mappings extension.
4570 This work was sponsored by Google.
4573 *) Fixes to pathlength constraint, self issued certificate handling,
4574 policy processing to align with RFC3280 and PKITS tests.
4576 This work was sponsored by Google.
4579 *) Support for name constraints certificate extension. DN, email, DNS
4580 and URI types are currently supported.
4582 This work was sponsored by Google.
4585 *) To cater for systems that provide a pointer-based thread ID rather
4586 than numeric, deprecate the current numeric thread ID mechanism and
4587 replace it with a structure and associated callback type. This
4588 mechanism allows a numeric "hash" to be extracted from a thread ID in
4589 either case, and on platforms where pointers are larger than 'long',
4590 mixing is done to help ensure the numeric 'hash' is usable even if it
4591 can't be guaranteed unique. The default mechanism is to use "&errno"
4592 as a pointer-based thread ID to distinguish between threads.
4594 Applications that want to provide their own thread IDs should now use
4595 CRYPTO_THREADID_set_callback() to register a callback that will call
4596 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4598 Note that ERR_remove_state() is now deprecated, because it is tied
4599 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4600 to free the current thread's error state should be replaced by
4601 ERR_remove_thread_state(NULL).
4603 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4604 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4605 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4606 application was previously providing a numeric thread callback that
4607 was inappropriate for distinguishing threads, then uniqueness might
4608 have been obtained with &errno that happened immediately in the
4609 intermediate development versions of OpenSSL; this is no longer the
4610 case, the numeric thread callback will now override the automatic use
4612 [Geoff Thorpe, with help from Bodo Moeller]
4614 *) Initial support for different CRL issuing certificates. This covers a
4615 simple case where the self issued certificates in the chain exist and
4616 the real CRL issuer is higher in the existing chain.
4618 This work was sponsored by Google.
4621 *) Removed effectively defunct crypto/store from the build.
4624 *) Revamp of STACK to provide stronger type-checking. Still to come:
4625 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4626 ASN1_STRING, CONF_VALUE.
4629 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4630 RAM on SSL connections. This option can save about 34k per idle SSL.
4633 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4634 STACK, TXT_DB, bsearch, qsort.
4637 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4638 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4639 support for data, signedData, compressedData, digestedData and
4640 encryptedData, envelopedData types included. Scripts to check against
4641 RFC4134 examples draft and interop and consistency checks of many
4642 content types and variants.
4645 *) Add options to enc utility to support use of zlib compression BIO.
4648 *) Extend mk1mf to support importing of options and assembly language
4649 files from Configure script, currently only included in VC-WIN32.
4650 The assembly language rules can now optionally generate the source
4651 files from the associated perl scripts.
4654 *) Implement remaining functionality needed to support GOST ciphersuites.
4655 Interop testing has been performed using CryptoPro implementations.
4656 [Victor B. Wagner <vitus@cryptocom.ru>]
4658 *) s390x assembler pack.
4661 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4665 *) Implement Opaque PRF Input TLS extension as specified in
4666 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4667 official specification yet and no extension type assignment by
4668 IANA exists, this extension (for now) will have to be explicitly
4669 enabled when building OpenSSL by providing the extension number
4670 to use. For example, specify an option
4672 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4674 to the "config" or "Configure" script to enable the extension,
4675 assuming extension number 0x9527 (which is a completely arbitrary
4676 and unofficial assignment based on the MD5 hash of the Internet
4677 Draft). Note that by doing so, you potentially lose
4678 interoperability with other TLS implementations since these might
4679 be using the same extension number for other purposes.
4681 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4682 opaque PRF input value to use in the handshake. This will create
4683 an internal copy of the length-'len' string at 'src', and will
4684 return non-zero for success.
4686 To get more control and flexibility, provide a callback function
4689 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4690 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4694 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4697 Callback function 'cb' will be called in handshakes, and is
4698 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4699 Argument 'arg' is for application purposes (the value as given to
4700 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4701 be provided to the callback function). The callback function
4702 has to return non-zero to report success: usually 1 to use opaque
4703 PRF input just if possible, or 2 to enforce use of the opaque PRF
4704 input. In the latter case, the library will abort the handshake
4705 if opaque PRF input is not successfully negotiated.
4707 Arguments 'peerinput' and 'len' given to the callback function
4708 will always be NULL and 0 in the case of a client. A server will
4709 see the client's opaque PRF input through these variables if
4710 available (NULL and 0 otherwise). Note that if the server
4711 provides an opaque PRF input, the length must be the same as the
4712 length of the client's opaque PRF input.
4714 Note that the callback function will only be called when creating
4715 a new session (session resumption can resume whatever was
4716 previously negotiated), and will not be called in SSL 2.0
4717 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4718 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4719 for applications that need to enforce opaque PRF input.
4723 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4726 [Victor B. Wagner <vitus@cryptocom.ru>]
4728 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4729 RFC4507bis. The encrypted ticket format is an encrypted encoded
4730 SSL_SESSION structure, that way new session features are automatically
4733 If a client application caches session in an SSL_SESSION structure
4734 support is transparent because tickets are now stored in the encoded
4737 The SSL_CTX structure automatically generates keys for ticket
4738 protection in servers so again support should be possible
4739 with no application modification.
4741 If a client or server wishes to disable RFC4507 support then the option
4742 SSL_OP_NO_TICKET can be set.
4744 Add a TLS extension debugging callback to allow the contents of any client
4745 or server extensions to be examined.
4747 This work was sponsored by Google.
4750 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4751 OpenSSL should now compile cleanly on gcc 4.2
4752 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4754 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4755 support including streaming MAC support: this is required for GOST
4756 ciphersuite support.
4757 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4759 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4760 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4761 to output in BER and PEM format.
4764 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4765 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4766 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4767 ENGINE support for HMAC keys which are unextractable. New -mac and
4768 -macopt options to dgst utility.
4771 *) New option -sigopt to dgst utility. Update dgst to use
4772 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4773 alternative signing parameters such as X9.31 or PSS in the dgst
4777 *) Change ssl_cipher_apply_rule(), the internal function that does
4778 the work each time a ciphersuite string requests enabling
4779 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4780 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4781 the order of disabled ciphersuites such that those ciphersuites
4782 that most recently went from enabled to disabled not only stay
4783 in order with respect to each other, but also have higher priority
4784 than other disabled ciphersuites the next time ciphersuites are
4787 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4788 the same ciphersuites as with "HIGH" alone, but in a specific
4789 order where the PSK ciphersuites come first (since they are the
4790 most recently disabled ciphersuites when "HIGH" is parsed).
4792 Also, change ssl_create_cipher_list() (using this new
4793 functionality) such that between otherwise identical
4794 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4798 *) Change ssl_create_cipher_list() so that it automatically
4799 arranges the ciphersuites in reasonable order before starting
4800 to process the rule string. Thus, the definition for "DEFAULT"
4801 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4802 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4803 This makes it much easier to arrive at a reasonable default order
4804 in applications for which anonymous ciphers are OK (meaning
4805 that you can't actually use DEFAULT).
4806 [Bodo Moeller; suggested by Victor Duchovni]
4808 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4809 processing) into multiple integers instead of setting
4810 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4811 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4812 (These masks as well as the individual bit definitions are hidden
4813 away into the non-exported interface ssl/ssl_locl.h, so this
4814 change to the definition of the SSL_CIPHER structure shouldn't
4815 affect applications.) This give us more bits for each of these
4816 categories, so there is no longer a need to coagulate AES128 and
4817 AES256 into a single algorithm bit, and to coagulate Camellia128
4818 and Camellia256 into a single algorithm bit, which has led to all
4821 Thus, among other things, the kludge introduced in 0.9.7m and
4822 0.9.8e for masking out AES256 independently of AES128 or masking
4823 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4825 With the change, we also introduce new ciphersuite aliases that
4826 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4830 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4831 Use the leftmost N bytes of the signature input if the input is
4832 larger than the prime q (with N being the size in bytes of q).
4835 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4836 it yet and it is largely untested.
4839 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4842 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4843 some compilers (gcc 4.2 and later) reject their use. Safestack is
4844 reimplemented. Update ASN1 to avoid use of legacy functions.
4847 *) Win32/64 targets are linked with Winsock2.
4850 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4851 to external functions. This can be used to increase CRL handling
4852 efficiency especially when CRLs are very large by (for example) storing
4853 the CRL revoked certificates in a database.
4856 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4857 new CRLs added to a directory can be used. New command line option
4858 -verify_return_error to s_client and s_server. This causes real errors
4859 to be returned by the verify callback instead of carrying on no matter
4860 what. This reflects the way a "real world" verify callback would behave.
4863 *) GOST engine, supporting several GOST algorithms and public key formats.
4864 Kindly donated by Cryptocom.
4867 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4868 partitioned by DP are handled but no indirect CRL or reason partitioning
4869 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4870 selected via a scoring technique which handles IDP and AKID in CRLs.
4873 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4874 will ultimately be used for all verify operations: this will remove the
4875 X509_STORE dependency on certificate verification and allow alternative
4876 lookup methods. X509_STORE based implementations of these two callbacks.
4879 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4880 Modify get_crl() to find a valid (unexpired) CRL if possible.
4883 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4884 this would be called X509_CRL_cmp() but that name is already used by
4885 a function that just compares CRL issuer names. Cache several CRL
4886 extensions in X509_CRL structure and cache CRLDP in X509.
4889 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4890 this maps equivalent X509_NAME structures into a consistent structure.
4891 Name comparison can then be performed rapidly using memcmp().
4894 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4898 *) Allow digests to supply their own micalg string for S/MIME type using
4899 the ctrl EVP_MD_CTRL_MICALG.
4902 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4903 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4904 ctrl. It can then customise the structure before and/or after signing
4908 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4909 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4910 to free up any added signature OIDs.
4913 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4914 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4915 digest and cipher tables. New options added to openssl utility:
4916 list-message-digest-algorithms and list-cipher-algorithms.
4919 *) Change the array representation of binary polynomials: the list
4920 of degrees of non-zero coefficients is now terminated with -1.
4921 Previously it was terminated with 0, which was also part of the
4922 value; thus, the array representation was not applicable to
4923 polynomials where t^0 has coefficient zero. This change makes
4924 the array representation useful in a more general context.
4927 *) Various modifications and fixes to SSL/TLS cipher string
4928 handling. For ECC, the code now distinguishes between fixed ECDH
4929 with RSA certificates on the one hand and with ECDSA certificates
4930 on the other hand, since these are separate ciphersuites. The
4931 unused code for Fortezza ciphersuites has been removed.
4933 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4934 (not "ECDHE"). For consistency with the code for DH
4935 certificates, use of ECDH certificates is now considered ECDH
4936 authentication, not RSA or ECDSA authentication (the latter is
4937 merely the CA's signing algorithm and not actively used in the
4940 The temporary ciphersuite alias "ECCdraft" is no longer
4941 available, and ECC ciphersuites are no longer excluded from "ALL"
4942 and "DEFAULT". The following aliases now exist for RFC 4492
4943 ciphersuites, most of these by analogy with the DH case:
4945 kECDHr - ECDH cert, signed with RSA
4946 kECDHe - ECDH cert, signed with ECDSA
4947 kECDH - ECDH cert (signed with either RSA or ECDSA)
4948 kEECDH - ephemeral ECDH
4949 ECDH - ECDH cert or ephemeral ECDH
4955 AECDH - anonymous ECDH
4956 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4960 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4961 Use correct micalg parameters depending on digest(s) in signed message.
4964 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4965 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4968 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4969 an engine to register a method. Add ENGINE lookups for methods and
4970 functional reference processing.
4973 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4974 EVP_{Sign,Verify}* which allow an application to customise the signature
4978 *) New -resign option to smime utility. This adds one or more signers
4979 to an existing PKCS#7 signedData structure. Also -md option to use an
4980 alternative message digest algorithm for signing.
4983 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4984 create PKCS7 structures containing multiple signers. Update smime
4985 application to support multiple signers.
4988 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4992 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4993 Reorganize PBE internals to lookup from a static table using NIDs,
4994 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4995 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4996 PRF which will be automatically used with PBES2.
4999 *) Replace the algorithm specific calls to generate keys in "req" with the
5003 *) Update PKCS#7 enveloped data routines to use new API. This is now
5004 supported by any public key method supporting the encrypt operation. A
5005 ctrl is added to allow the public key algorithm to examine or modify
5006 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
5010 *) Add a ctrl to asn1 method to allow a public key algorithm to express
5011 a default digest type to use. In most cases this will be SHA1 but some
5012 algorithms (such as GOST) need to specify an alternative digest. The
5013 return value indicates how strong the preference is 1 means optional and
5014 2 is mandatory (that is it is the only supported type). Modify
5015 ASN1_item_sign() to accept a NULL digest argument to indicate it should
5016 use the default md. Update openssl utilities to use the default digest
5017 type for signing if it is not explicitly indicated.
5020 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
5021 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
5022 signing method from the key type. This effectively removes the link
5023 between digests and public key types.
5026 *) Add an OID cross reference table and utility functions. Its purpose is to
5027 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
5028 rsaEncryption. This will allow some of the algorithm specific hackery
5029 needed to use the correct OID to be removed.
5032 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
5033 structures for PKCS7_sign(). They are now set up by the relevant public
5037 *) Add provisional EC pkey method with support for ECDSA and ECDH.
5040 *) Add support for key derivation (agreement) in the API, DH method and
5044 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
5045 public and private key formats. As a side effect these add additional
5046 command line functionality not previously available: DSA signatures can be
5047 generated and verified using pkeyutl and DH key support and generation in
5052 [Oliver Tappe <zooey@hirschkaefer.de>]
5054 *) New make target "install_html_docs" installs HTML renditions of the
5056 [Oliver Tappe <zooey@hirschkaefer.de>]
5058 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
5059 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
5060 support key and parameter generation and add initial key generation
5061 functionality for RSA.
5064 *) Add functions for main EVP_PKEY_method operations. The undocumented
5065 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
5066 EVP_PKEY_{encrypt,decrypt}_old.
5069 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
5070 key API, doesn't do much yet.
5073 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
5074 public key algorithms. New option to openssl utility:
5075 "list-public-key-algorithms" to print out info.
5078 *) Implement the Supported Elliptic Curves Extension for
5079 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5082 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
5083 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
5086 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
5087 utilities such as rsa, dsa, dsaparam etc except they process any key
5091 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
5092 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
5093 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
5097 *) Initial support for pluggable public key ASN1.
5098 De-spaghettify the public key ASN1 handling. Move public and private
5099 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
5100 algorithm specific handling to a single module within the relevant
5101 algorithm directory. Add functions to allow (near) opaque processing
5102 of public and private key structures.
5105 *) Implement the Supported Point Formats Extension for
5106 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5109 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
5110 for the psk identity [hint] and the psk callback functions to the
5111 SSL_SESSION, SSL and SSL_CTX structure.
5114 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
5118 SSL_CTX_use_psk_identity_hint
5119 SSL_get_psk_identity_hint
5120 SSL_get_psk_identity
5121 SSL_use_psk_identity_hint
5123 [Mika Kousa and Pasi Eronen of Nokia Corporation]
5125 *) Add RFC 3161 compliant time stamp request creation, response generation
5126 and response verification functionality.
5127 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
5129 *) Add initial support for TLS extensions, specifically for the server_name
5130 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5131 have new members for a host name. The SSL data structure has an
5132 additional member SSL_CTX *initial_ctx so that new sessions can be
5133 stored in that context to allow for session resumption, even after the
5134 SSL has been switched to a new SSL_CTX in reaction to a client's
5135 server_name extension.
5137 New functions (subject to change):
5139 SSL_get_servername()
5140 SSL_get_servername_type()
5143 New CTRL codes and macros (subject to change):
5145 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5146 - SSL_CTX_set_tlsext_servername_callback()
5147 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5148 - SSL_CTX_set_tlsext_servername_arg()
5149 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5151 openssl s_client has a new '-servername ...' option.
5153 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5154 '-key2 ...', '-servername_fatal' (subject to change). This allows
5155 testing the HostName extension for a specific single host name ('-cert'
5156 and '-key' remain fallbacks for handshakes without HostName
5157 negotiation). If the unrecognized_name alert has to be sent, this by
5158 default is a warning; it becomes fatal with the '-servername_fatal'
5161 [Peter Sylvester, Remy Allais, Christophe Renou]
5163 *) Whirlpool hash implementation is added.
5166 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
5167 bn(64,32). Because of instruction set limitations it doesn't have
5168 any negative impact on performance. This was done mostly in order
5169 to make it possible to share assembler modules, such as bn_mul_mont
5170 implementations, between 32- and 64-bit builds without hassle.
5173 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
5174 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
5178 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
5179 dedicated Montgomery multiplication procedure, is introduced.
5180 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
5181 "64-bit" performance on certain 32-bit targets.
5184 *) New option SSL_OP_NO_COMP to disable use of compression selectively
5185 in SSL structures. New SSL ctrl to set maximum send fragment size.
5186 Save memory by setting the I/O buffer sizes dynamically instead of
5187 using the maximum available value.
5190 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
5191 in addition to the text details.
5194 *) Very, very preliminary EXPERIMENTAL support for printing of general
5195 ASN1 structures. This currently produces rather ugly output and doesn't
5196 handle several customised structures at all.
5199 *) Integrated support for PVK file format and some related formats such
5200 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
5201 these in the 'rsa' and 'dsa' utilities.
5204 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
5207 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
5208 place for the (very old) "NETSCAPE" format certificates which are now
5209 handled using new ASN1 code equivalents.
5212 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
5213 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
5214 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
5217 *) Modify CRL distribution points extension code to print out previously
5218 unsupported fields. Enhance extension setting code to allow setting of
5222 *) Add print and set support for Issuing Distribution Point CRL extension.
5225 *) Change 'Configure' script to enable Camellia by default.
5228 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
5230 *) When rejecting SSL/TLS records due to an incorrect version number, never
5231 update s->server with a new major version number. As of
5232 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
5233 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
5234 the previous behavior could result in a read attempt at NULL when
5235 receiving specific incorrect SSL/TLS records once record payload
5236 protection is active. (CVE-2010-0740)
5237 [Bodo Moeller, Adam Langley <agl@chromium.org>]
5239 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
5240 could be crashed if the relevant tables were not present (e.g. chrooted).
5241 [Tomas Hoger <thoger@redhat.com>]
5243 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
5245 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
5246 [Martin Olsson, Neel Mehta]
5248 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
5249 accommodate for stack sorting, always a write lock!).
5252 *) On some versions of WIN32 Heap32Next is very slow. This can cause
5253 excessive delays in the RAND_poll(): over a minute. As a workaround
5254 include a time check in the inner Heap32Next loop too.
5257 *) The code that handled flushing of data in SSL/TLS originally used the
5258 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
5259 the problem outlined in PR#1949. The fix suggested there however can
5260 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
5261 of Apache). So instead simplify the code to flush unconditionally.
5262 This should be fine since flushing with no data to flush is a no op.
5265 *) Handle TLS versions 2.0 and later properly and correctly use the
5266 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
5267 off ancient servers have a habit of sticking around for a while...
5270 *) Modify compression code so it frees up structures without using the
5271 ex_data callbacks. This works around a problem where some applications
5272 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
5273 restarting) then use compression (e.g. SSL with compression) later.
5274 This results in significant per-connection memory leaks and
5275 has caused some security issues including CVE-2008-1678 and
5279 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
5280 change when encrypting or decrypting.
5283 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
5284 connect and renegotiate with servers which do not support RI.
5285 Until RI is more widely deployed this option is enabled by default.
5288 *) Add "missing" ssl ctrls to clear options and mode.
5291 *) If client attempts to renegotiate and doesn't support RI respond with
5292 a no_renegotiation alert as required by RFC5746. Some renegotiating
5293 TLS clients will continue a connection gracefully when they receive
5294 the alert. Unfortunately OpenSSL mishandled this alert and would hang
5295 waiting for a server hello which it will never receive. Now we treat a
5296 received no_renegotiation alert as a fatal error. This is because
5297 applications requesting a renegotiation might well expect it to succeed
5298 and would have no code in place to handle the server denying it so the
5299 only safe thing to do is to terminate the connection.
5302 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
5303 peer supports secure renegotiation and 0 otherwise. Print out peer
5304 renegotiation support in s_client/s_server.
5307 *) Replace the highly broken and deprecated SPKAC certification method with
5308 the updated NID creation version. This should correctly handle UTF8.
5311 *) Implement RFC5746. Re-enable renegotiation but require the extension
5312 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
5313 turns out to be a bad idea. It has been replaced by
5314 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5315 SSL_CTX_set_options(). This is really not recommended unless you
5316 know what you are doing.
5317 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5319 *) Fixes to stateless session resumption handling. Use initial_ctx when
5320 issuing and attempting to decrypt tickets in case it has changed during
5321 servername handling. Use a non-zero length session ID when attempting
5322 stateless session resumption: this makes it possible to determine if
5323 a resumption has occurred immediately after receiving server hello
5324 (several places in OpenSSL subtly assume this) instead of later in
5328 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5329 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5330 fixes for a few places where the return code is not checked
5332 [Julia Lawall <julia@diku.dk>]
5334 *) Add --strict-warnings option to Configure script to include devteam
5335 warnings in other configurations.
5338 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5339 makes it possible to install openssl libraries in locations which
5340 have names other than "lib", for example "/usr/lib64" which some
5342 [Steve Henson, based on patch from Jeremy Utley]
5344 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5345 X690 8.9.12 and can produce some misleading textual output of OIDs.
5346 [Steve Henson, reported by Dan Kaminsky]
5348 *) Delete MD2 from algorithm tables. This follows the recommendation in
5349 several standards that it is not used in new applications due to
5350 several cryptographic weaknesses. For binary compatibility reasons
5351 the MD2 API is still compiled in by default.
5354 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5358 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5359 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5361 [Guenter <lists@gknw.net>]
5363 *) Fix the server certificate chain building code to use X509_verify_cert(),
5364 it used to have an ad-hoc builder which was unable to cope with anything
5365 other than a simple chain.
5366 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5368 *) Don't check self signed certificate signatures in X509_verify_cert()
5369 by default (a flag can override this): it just wastes time without
5370 adding any security. As a useful side effect self signed root CAs
5371 with non-FIPS digests are now usable in FIPS mode.
5374 *) In dtls1_process_out_of_seq_message() the check if the current message
5375 is already buffered was missing. For every new message was memory
5376 allocated, allowing an attacker to perform an denial of service attack
5377 with sending out of seq handshake messages until there is no memory
5378 left. Additionally every future message was buffered, even if the
5379 sequence number made no sense and would be part of another handshake.
5380 So only messages with sequence numbers less than 10 in advance will be
5381 buffered. (CVE-2009-1378)
5382 [Robin Seggelmann, discovered by Daniel Mentz]
5384 *) Records are buffered if they arrive with a future epoch to be
5385 processed after finishing the corresponding handshake. There is
5386 currently no limitation to this buffer allowing an attacker to perform
5387 a DOS attack with sending records with future epochs until there is no
5388 memory left. This patch adds the pqueue_size() function to determine
5389 the size of a buffer and limits the record buffer to 100 entries.
5391 [Robin Seggelmann, discovered by Daniel Mentz]
5393 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5394 parent structure is freed. (CVE-2009-1379)
5397 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5398 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5401 [Ilya O. <vrghost@gmail.com>]
5403 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5405 *) Disable renegotiation completely - this fixes a severe security
5406 problem (CVE-2009-3555) at the cost of breaking all
5407 renegotiation. Renegotiation can be re-enabled by setting
5408 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5409 run-time. This is really not recommended unless you know what
5413 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5415 *) Don't set val to NULL when freeing up structures, it is freed up by
5416 underlying code. If sizeof(void *) > sizeof(long) this can result in
5417 zeroing past the valid field. (CVE-2009-0789)
5418 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5420 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5421 checked correctly. This would allow some invalid signed attributes to
5422 appear to verify correctly. (CVE-2009-0591)
5423 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5425 *) Reject UniversalString and BMPString types with invalid lengths. This
5426 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5427 a legal length. (CVE-2009-0590)
5430 *) Set S/MIME signing as the default purpose rather than setting it
5431 unconditionally. This allows applications to override it at the store
5435 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5436 to handle some structures.
5439 *) Improve efficiency of mem_gets: don't search whole buffer each time
5441 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5443 *) New -hex option for openssl rand.
5446 *) Print out UTF8String and NumericString when parsing ASN1.
5449 *) Support NumericString type for name components.
5452 *) Allow CC in the environment to override the automatically chosen
5453 compiler. Note that nothing is done to ensure flags work with the
5457 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5459 *) Properly check EVP_VerifyFinal() and similar return values
5461 [Ben Laurie, Bodo Moeller, Google Security Team]
5463 *) Enable TLS extensions by default.
5466 *) Allow the CHIL engine to be loaded, whether the application is
5467 multithreaded or not. (This does not release the developer from the
5468 obligation to set up the dynamic locking callbacks.)
5469 [Sander Temme <sander@temme.net>]
5471 *) Use correct exit code if there is an error in dgst command.
5472 [Steve Henson; problem pointed out by Roland Dirlewanger]
5474 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5475 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5478 *) Add experimental JPAKE support, including demo authentication in
5479 s_client and s_server.
5482 *) Set the comparison function in v3_addr_canonize().
5483 [Rob Austein <sra@hactrn.net>]
5485 *) Add support for XMPP STARTTLS in s_client.
5486 [Philip Paeps <philip@freebsd.org>]
5488 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5489 to ensure that even with this option, only ciphersuites in the
5490 server's preference list will be accepted. (Note that the option
5491 applies only when resuming a session, so the earlier behavior was
5492 just about the algorithm choice for symmetric cryptography.)
5495 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5497 *) Fix NULL pointer dereference if a DTLS server received
5498 ChangeCipherSpec as first record (CVE-2009-1386).
5501 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5502 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5505 *) The fix in 0.9.8c that supposedly got rid of unsafe
5506 double-checked locking was incomplete for RSA blinding,
5507 addressing just one layer of what turns out to have been
5508 doubly unsafe triple-checked locking.
5510 So now fix this for real by retiring the MONT_HELPER macro
5511 in crypto/rsa/rsa_eay.c.
5513 [Bodo Moeller; problem pointed out by Marius Schilder]
5515 *) Various precautionary measures:
5517 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5519 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5520 (NB: This would require knowledge of the secret session ticket key
5521 to exploit, in which case you'd be SOL either way.)
5523 - Change bn_nist.c so that it will properly handle input BIGNUMs
5524 outside the expected range.
5526 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5529 [Neel Mehta, Bodo Moeller]
5531 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5532 the load fails. Useful for distros.
5533 [Ben Laurie and the FreeBSD team]
5535 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5538 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5541 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5543 This work was sponsored by Logica.
5546 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5547 keystores. Support for SSL/TLS client authentication too.
5548 Not compiled unless enable-capieng specified to Configure.
5550 This work was sponsored by Logica.
5553 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5554 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5555 attribute creation routines such as certificate requests and PKCS#12
5559 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5561 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5562 handshake which could lead to a client crash as found using the
5563 Codenomicon TLS test suite (CVE-2008-1672)
5564 [Steve Henson, Mark Cox]
5566 *) Fix double free in TLS server name extensions which could lead to
5567 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5570 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5572 Clear the error queue to ensure that error entries left from
5573 older function calls do not interfere with the correct operation.
5574 [Lutz Jaenicke, Erik de Castro Lopo]
5576 *) Remove root CA certificates of commercial CAs:
5578 The OpenSSL project does not recommend any specific CA and does not
5579 have any policy with respect to including or excluding any CA.
5580 Therefore it does not make any sense to ship an arbitrary selection
5581 of root CA certificates with the OpenSSL software.
5584 *) RSA OAEP patches to fix two separate invalid memory reads.
5585 The first one involves inputs when 'lzero' is greater than
5586 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5587 before the beginning of from). The second one involves inputs where
5588 the 'db' section contains nothing but zeroes (there is a one-byte
5589 invalid read after the end of 'db').
5590 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5592 *) Partial backport from 0.9.9-dev:
5594 Introduce bn_mul_mont (dedicated Montgomery multiplication
5595 procedure) as a candidate for BIGNUM assembler implementation.
5596 While 0.9.9-dev uses assembler for various architectures, only
5597 x86_64 is available by default here in the 0.9.8 branch, and
5598 32-bit x86 is available through a compile-time setting.
5600 To try the 32-bit x86 assembler implementation, use Configure
5601 option "enable-montasm" (which exists only for this backport).
5603 As "enable-montasm" for 32-bit x86 disclaims code stability
5604 anyway, in this constellation we activate additional code
5605 backported from 0.9.9-dev for further performance improvements,
5606 namely BN_from_montgomery_word. (To enable this otherwise,
5607 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5609 [Andy Polyakov (backport partially by Bodo Moeller)]
5611 *) Add TLS session ticket callback. This allows an application to set
5612 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5613 values. This is useful for key rollover for example where several key
5614 sets may exist with different names.
5617 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5618 This was broken until now in 0.9.8 releases, such that the only way
5619 a registered ENGINE could be used (assuming it initialises
5620 successfully on the host) was to explicitly set it as the default
5621 for the relevant algorithms. This is in contradiction with 0.9.7
5622 behaviour and the documentation. With this fix, when an ENGINE is
5623 registered into a given algorithm's table of implementations, the
5624 'uptodate' flag is reset so that auto-discovery will be used next
5625 time a new context for that algorithm attempts to select an
5627 [Ian Lister (tweaked by Geoff Thorpe)]
5629 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5630 implementation in the following ways:
5632 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5635 Lack of BER streaming support means one pass streaming processing is
5636 only supported if data is detached: setting the streaming flag is
5637 ignored for embedded content.
5639 CMS support is disabled by default and must be explicitly enabled
5640 with the enable-cms configuration option.
5643 *) Update the GMP engine glue to do direct copies between BIGNUM and
5644 mpz_t when openssl and GMP use the same limb size. Otherwise the
5645 existing "conversion via a text string export" trick is still used.
5646 [Paul Sheer <paulsheer@gmail.com>]
5648 *) Zlib compression BIO. This is a filter BIO which compressed and
5649 uncompresses any data passed through it.
5652 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5653 RFC3394 compatible AES key wrapping.
5656 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5657 sets string data without copying. X509_ALGOR_set0() and
5658 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5659 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5660 from an X509_ATTRIBUTE structure optionally checking it occurs only
5661 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5665 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5666 to get the expected BN_FLG_CONSTTIME behavior.
5667 [Bodo Moeller (Google)]
5671 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5672 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5673 - added some more tests to do_tests.pl
5674 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5675 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5676 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5677 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5678 - various changes to netware.pl to enable gcc-cross builds on Win32
5680 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5681 - various changes to fix missing prototype warnings
5682 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5683 - added AES, WHIRLPOOL and CPUID assembler code to build files
5684 - added missing AES assembler make rules to mk1mf.pl
5685 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5686 [Guenter Knauf <eflash@gmx.net>]
5688 *) Implement certificate status request TLS extension defined in RFC3546.
5689 A client can set the appropriate parameters and receive the encoded
5690 OCSP response via a callback. A server can query the supplied parameters
5691 and set the encoded OCSP response in the callback. Add simplified examples
5692 to s_client and s_server.
5695 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5697 *) Fix various bugs:
5698 + Binary incompatibility of ssl_ctx_st structure
5699 + DTLS interoperation with non-compliant servers
5700 + Don't call get_session_cb() without proposed session
5701 + Fix ia64 assembler code
5702 [Andy Polyakov, Steve Henson]
5704 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5706 *) DTLS Handshake overhaul. There were longstanding issues with
5707 OpenSSL DTLS implementation, which were making it impossible for
5708 RFC 4347 compliant client to communicate with OpenSSL server.
5709 Unfortunately just fixing these incompatibilities would "cut off"
5710 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5711 server keeps tolerating non RFC compliant syntax. The opposite is
5712 not true, 0.9.8f client can not communicate with earlier server.
5713 This update even addresses CVE-2007-4995.
5716 *) Changes to avoid need for function casts in OpenSSL: some compilers
5717 (gcc 4.2 and later) reject their use.
5718 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5721 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5722 RFC4507bis. The encrypted ticket format is an encrypted encoded
5723 SSL_SESSION structure, that way new session features are automatically
5726 If a client application caches session in an SSL_SESSION structure
5727 support is transparent because tickets are now stored in the encoded
5730 The SSL_CTX structure automatically generates keys for ticket
5731 protection in servers so again support should be possible
5732 with no application modification.
5734 If a client or server wishes to disable RFC4507 support then the option
5735 SSL_OP_NO_TICKET can be set.
5737 Add a TLS extension debugging callback to allow the contents of any client
5738 or server extensions to be examined.
5740 This work was sponsored by Google.
5743 *) Add initial support for TLS extensions, specifically for the server_name
5744 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5745 have new members for a host name. The SSL data structure has an
5746 additional member SSL_CTX *initial_ctx so that new sessions can be
5747 stored in that context to allow for session resumption, even after the
5748 SSL has been switched to a new SSL_CTX in reaction to a client's
5749 server_name extension.
5751 New functions (subject to change):
5753 SSL_get_servername()
5754 SSL_get_servername_type()
5757 New CTRL codes and macros (subject to change):
5759 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5760 - SSL_CTX_set_tlsext_servername_callback()
5761 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5762 - SSL_CTX_set_tlsext_servername_arg()
5763 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5765 openssl s_client has a new '-servername ...' option.
5767 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5768 '-key2 ...', '-servername_fatal' (subject to change). This allows
5769 testing the HostName extension for a specific single host name ('-cert'
5770 and '-key' remain fallbacks for handshakes without HostName
5771 negotiation). If the unrecognized_name alert has to be sent, this by
5772 default is a warning; it becomes fatal with the '-servername_fatal'
5775 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5777 *) Add AES and SSE2 assembly language support to VC++ build.
5780 *) Mitigate attack on final subtraction in Montgomery reduction.
5783 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5784 (which previously caused an internal error).
5787 *) Squeeze another 10% out of IGE mode when in != out.
5790 *) AES IGE mode speedup.
5791 [Dean Gaudet (Google)]
5793 *) Add the Korean symmetric 128-bit cipher SEED (see
5794 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5795 add SEED ciphersuites from RFC 4162:
5797 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5798 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5799 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5800 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5802 To minimize changes between patchlevels in the OpenSSL 0.9.8
5803 series, SEED remains excluded from compilation unless OpenSSL
5804 is configured with 'enable-seed'.
5805 [KISA, Bodo Moeller]
5807 *) Mitigate branch prediction attacks, which can be practical if a
5808 single processor is shared, allowing a spy process to extract
5809 information. For detailed background information, see
5810 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5811 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5812 and Necessary Software Countermeasures"). The core of the change
5813 are new versions BN_div_no_branch() and
5814 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5815 respectively, which are slower, but avoid the security-relevant
5816 conditional branches. These are automatically called by BN_div()
5817 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5818 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5819 remove a conditional branch.
5821 BN_FLG_CONSTTIME is the new name for the previous
5822 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5823 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5824 in the exponent causes BN_mod_exp_mont() to use the alternative
5825 implementation in BN_mod_exp_mont_consttime().) The old name
5826 remains as a deprecated alias.
5828 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5829 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5830 constant-time implementations for more than just exponentiation.
5831 Here too the old name is kept as a deprecated alias.
5833 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5834 the BN_BLINDING structure gets an independent copy of the
5835 modulus. This means that the previous "BIGNUM *m" argument to
5836 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5837 essentially becomes "const BIGNUM *m", although we can't actually
5838 change this in the header file before 0.9.9. It allows
5839 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5840 enable BN_FLG_CONSTTIME.
5842 [Matthew D Wood (Intel Corp)]
5844 *) In the SSL/TLS server implementation, be strict about session ID
5845 context matching (which matters if an application uses a single
5846 external cache for different purposes). Previously,
5847 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5848 set. This did ensure strict client verification, but meant that,
5849 with applications using a single external cache for quite
5850 different requirements, clients could circumvent ciphersuite
5851 restrictions for a given session ID context by starting a session
5852 in a different context.
5855 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5856 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5857 authentication-only ciphersuites.
5860 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5861 not complete and could lead to a possible single byte overflow
5862 (CVE-2007-5135) [Ben Laurie]
5864 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5866 *) Since AES128 and AES256 (and similarly Camellia128 and
5867 Camellia256) share a single mask bit in the logic of
5868 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5869 kludge to work properly if AES128 is available and AES256 isn't
5870 (or if Camellia128 is available and Camellia256 isn't).
5873 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5874 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5875 When a point or a seed is encoded in a BIT STRING, we need to
5876 prevent the removal of trailing zero bits to get the proper DER
5877 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5878 of a NamedBitList, for which trailing 0 bits need to be removed.)
5881 *) Have SSL/TLS server implementation tolerate "mismatched" record
5882 protocol version while receiving ClientHello even if the
5883 ClientHello is fragmented. (The server can't insist on the
5884 particular protocol version it has chosen before the ServerHello
5885 message has informed the client about his choice.)
5888 *) Add RFC 3779 support.
5889 [Rob Austein for ARIN, Ben Laurie]
5891 *) Load error codes if they are not already present instead of using a
5892 static variable. This allows them to be cleanly unloaded and reloaded.
5893 Improve header file function name parsing.
5896 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5897 or CAPABILITY handshake as required by RFCs.
5900 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5902 *) Introduce limits to prevent malicious keys being able to
5903 cause a denial of service. (CVE-2006-2940)
5904 [Steve Henson, Bodo Moeller]
5906 *) Fix ASN.1 parsing of certain invalid structures that can result
5907 in a denial of service. (CVE-2006-2937) [Steve Henson]
5909 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5910 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5912 *) Fix SSL client code which could crash if connecting to a
5913 malicious SSLv2 server. (CVE-2006-4343)
5914 [Tavis Ormandy and Will Drewry, Google Security Team]
5916 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5917 match only those. Before that, "AES256-SHA" would be interpreted
5918 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5919 the same strength classification in 0.9.7h) as we currently only
5920 have a single AES bit in the ciphersuite description bitmap.
5921 That change, however, also applied to ciphersuite strings such as
5922 "RC4-MD5" that intentionally matched multiple ciphersuites --
5923 namely, SSL 2.0 ciphersuites in addition to the more common ones
5924 from SSL 3.0/TLS 1.0.
5926 So we change the selection algorithm again: Naming an explicit
5927 ciphersuite selects this one ciphersuite, and any other similar
5928 ciphersuite (same bitmap) from *other* protocol versions.
5929 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5930 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5932 Since SSL 2.0 does not have any ciphersuites for which the
5933 128/256 bit distinction would be relevant, this works for now.
5934 The proper fix will be to use different bits for AES128 and
5935 AES256, which would have avoided the problems from the beginning;
5936 however, bits are scarce, so we can only do this in a new release
5937 (not just a patchlevel) when we can change the SSL_CIPHER
5938 definition to split the single 'unsigned long mask' bitmap into
5939 multiple values to extend the available space.
5943 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5945 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5946 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5948 *) Add AES IGE and biIGE modes.
5951 *) Change the Unix randomness entropy gathering to use poll() when
5952 possible instead of select(), since the latter has some
5953 undesirable limitations.
5954 [Darryl Miles via Richard Levitte and Bodo Moeller]
5956 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5957 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5958 cannot be implicitly activated as part of, e.g., the "AES" alias.
5959 However, please upgrade to OpenSSL 0.9.9[-dev] for
5960 non-experimental use of the ECC ciphersuites to get TLS extension
5961 support, which is required for curve and point format negotiation
5962 to avoid potential handshake problems.
5965 *) Disable rogue ciphersuites:
5967 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5968 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5969 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5971 The latter two were purportedly from
5972 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5975 Also deactivate the remaining ciphersuites from
5976 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5977 unofficial, and the ID has long expired.
5980 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5981 dual-core machines) and other potential thread-safety issues.
5984 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5985 versions), which is now available for royalty-free use
5986 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5987 Also, add Camellia TLS ciphersuites from RFC 4132.
5989 To minimize changes between patchlevels in the OpenSSL 0.9.8
5990 series, Camellia remains excluded from compilation unless OpenSSL
5991 is configured with 'enable-camellia'.
5994 *) Disable the padding bug check when compression is in use. The padding
5995 bug check assumes the first packet is of even length, this is not
5996 necessarily true if compression is enabled and can result in false
5997 positives causing handshake failure. The actual bug test is ancient
5998 code so it is hoped that implementations will either have fixed it by
5999 now or any which still have the bug do not support compression.
6002 Changes between 0.9.8a and 0.9.8b [04 May 2006]
6004 *) When applying a cipher rule check to see if string match is an explicit
6005 cipher suite and only match that one cipher suite if it is.
6008 *) Link in manifests for VC++ if needed.
6009 [Austin Ziegler <halostatue@gmail.com>]
6011 *) Update support for ECC-based TLS ciphersuites according to
6012 draft-ietf-tls-ecc-12.txt with proposed changes (but without
6013 TLS extensions, which are supported starting with the 0.9.9
6014 branch, not in the OpenSSL 0.9.8 branch).
6017 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
6018 opaque EVP_CIPHER_CTX handling.
6021 *) Fixes and enhancements to zlib compression code. We now only use
6022 "zlib1.dll" and use the default __cdecl calling convention on Win32
6023 to conform with the standards mentioned here:
6024 http://www.zlib.net/DLL_FAQ.txt
6025 Static zlib linking now works on Windows and the new --with-zlib-include
6026 --with-zlib-lib options to Configure can be used to supply the location
6027 of the headers and library. Gracefully handle case where zlib library
6031 *) Several fixes and enhancements to the OID generation code. The old code
6032 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
6033 handle numbers larger than ULONG_MAX, truncated printing and had a
6034 non standard OBJ_obj2txt() behaviour.
6037 *) Add support for building of engines under engine/ as shared libraries
6038 under VC++ build system.
6041 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
6042 Hopefully, we will not see any false combination of paths any more.
6045 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
6047 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6048 (part of SSL_OP_ALL). This option used to disable the
6049 countermeasure against man-in-the-middle protocol-version
6050 rollback in the SSL 2.0 server implementation, which is a bad
6051 idea. (CVE-2005-2969)
6053 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6054 for Information Security, National Institute of Advanced Industrial
6055 Science and Technology [AIST], Japan)]
6057 *) Add two function to clear and return the verify parameter flags.
6060 *) Keep cipherlists sorted in the source instead of sorting them at
6061 runtime, thus removing the need for a lock.
6064 *) Avoid some small subgroup attacks in Diffie-Hellman.
6065 [Nick Mathewson and Ben Laurie]
6067 *) Add functions for well-known primes.
6070 *) Extended Windows CE support.
6071 [Satoshi Nakamura and Andy Polyakov]
6073 *) Initialize SSL_METHOD structures at compile time instead of during
6074 runtime, thus removing the need for a lock.
6077 *) Make PKCS7_decrypt() work even if no certificate is supplied by
6078 attempting to decrypt each encrypted key in turn. Add support to
6082 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
6084 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
6087 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
6090 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
6091 key into the same file any more.
6094 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
6097 *) Add -utf8 command line and config file option to 'ca'.
6098 [Stefan <stf@udoma.org]
6100 *) Removed the macro des_crypt(), as it seems to conflict with some
6101 libraries. Use DES_crypt().
6104 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
6105 involves renaming the source and generated shared-libs for
6106 both. The engines will accept the corrected or legacy ids
6107 ('ncipher' and '4758_cca' respectively) when binding. NB,
6108 this only applies when building 'shared'.
6109 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
6111 *) Add attribute functions to EVP_PKEY structure. Modify
6112 PKCS12_create() to recognize a CSP name attribute and
6113 use it. Make -CSP option work again in pkcs12 utility.
6116 *) Add new functionality to the bn blinding code:
6117 - automatic re-creation of the BN_BLINDING parameters after
6118 a fixed number of uses (currently 32)
6119 - add new function for parameter creation
6120 - introduce flags to control the update behaviour of the
6121 BN_BLINDING parameters
6122 - hide BN_BLINDING structure
6123 Add a second BN_BLINDING slot to the RSA structure to improve
6124 performance when a single RSA object is shared among several
6128 *) Add support for DTLS.
6129 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
6131 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
6132 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
6135 *) Remove buggy and incomplete DH cert support from
6136 ssl/ssl_rsa.c and ssl/s3_both.c
6139 *) Use SHA-1 instead of MD5 as the default digest algorithm for
6140 the apps/openssl applications.
6143 *) Compile clean with "-Wall -Wmissing-prototypes
6144 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
6145 DEBUG_SAFESTACK must also be set.
6148 *) Change ./Configure so that certain algorithms can be disabled by default.
6149 The new counterpiece to "no-xxx" is "enable-xxx".
6151 The patented RC5 and MDC2 algorithms will now be disabled unless
6152 "enable-rc5" and "enable-mdc2", respectively, are specified.
6154 (IDEA remains enabled despite being patented. This is because IDEA
6155 is frequently required for interoperability, and there is no license
6156 fee for non-commercial use. As before, "no-idea" can be used to
6157 avoid this algorithm.)
6161 *) Add processing of proxy certificates (see RFC 3820). This work was
6162 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
6163 EGEE (Enabling Grids for E-science in Europe).
6166 *) RC4 performance overhaul on modern architectures/implementations, such
6167 as Intel P4, IA-64 and AMD64.
6170 *) New utility extract-section.pl. This can be used specify an alternative
6171 section number in a pod file instead of having to treat each file as
6172 a separate case in Makefile. This can be done by adding two lines to the
6175 =for comment openssl_section:XXX
6177 The blank line is mandatory.
6181 *) New arguments -certform, -keyform and -pass for s_client and s_server
6182 to allow alternative format key and certificate files and passphrase
6186 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
6187 update associated structures and add various utility functions.
6189 Add new policy related verify parameters, include policy checking in
6190 standard verify code. Enhance 'smime' application with extra parameters
6191 to support policy checking and print out.
6194 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
6195 Nehemiah processors. These extensions support AES encryption in hardware
6196 as well as RNG (though RNG support is currently disabled).
6197 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
6199 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
6202 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
6203 [Andy Polyakov and a number of other people]
6205 *) Improved PowerPC platform support. Most notably BIGNUM assembler
6206 implementation contributed by IBM.
6207 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
6209 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
6210 exponent rather than 'unsigned long'. There is a corresponding change to
6211 the new 'rsa_keygen' element of the RSA_METHOD structure.
6212 [Jelte Jansen, Geoff Thorpe]
6214 *) Functionality for creating the initial serial number file is now
6215 moved from CA.pl to the 'ca' utility with a new option -create_serial.
6217 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
6218 number file to 1, which is bound to cause problems. To avoid
6219 the problems while respecting compatibility between different 0.9.7
6220 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
6221 CA.pl for serial number initialization. With the new release 0.9.8,
6222 we can fix the problem directly in the 'ca' utility.)
6225 *) Reduced header interdependencies by declaring more opaque objects in
6226 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
6227 give fewer recursive includes, which could break lazy source code - so
6228 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
6229 developers should define this symbol when building and using openssl to
6230 ensure they track the recommended behaviour, interfaces, [etc], but
6231 backwards-compatible behaviour prevails when this isn't defined.
6234 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
6237 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
6238 This will generate a random key of the appropriate length based on the
6239 cipher context. The EVP_CIPHER can provide its own random key generation
6240 routine to support keys of a specific form. This is used in the des and
6241 3des routines to generate a key of the correct parity. Update S/MIME
6242 code to use new functions and hence generate correct parity DES keys.
6243 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
6244 valid (weak or incorrect parity).
6247 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
6248 as looking them up. This is useful when the verified structure may contain
6249 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
6250 present unless the new PKCS7_NO_CRL flag is asserted.
6253 *) Extend ASN1 oid configuration module. It now additionally accepts the
6256 shortName = some long name, 1.2.3.4
6259 *) Reimplemented the BN_CTX implementation. There is now no more static
6260 limitation on the number of variables it can handle nor the depth of the
6261 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
6262 information can now expand as required, and rather than having a single
6263 static array of bignums, BN_CTX now uses a linked-list of such arrays
6264 allowing it to expand on demand whilst maintaining the usefulness of
6265 BN_CTX's "bundling".
6268 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
6269 to allow all RSA operations to function using a single BN_CTX.
6272 *) Preliminary support for certificate policy evaluation and checking. This
6273 is initially intended to pass the tests outlined in "Conformance Testing
6274 of Relying Party Client Certificate Path Processing Logic" v1.07.
6277 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
6278 remained unused and not that useful. A variety of other little bignum
6279 tweaks and fixes have also been made continuing on from the audit (see
6283 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
6284 associated ASN1, EVP and SSL functions and old ASN1 macros.
6287 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
6288 and this should never fail. So the return value from the use of
6289 BN_set_word() (which can fail due to needless expansion) is now deprecated;
6290 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
6293 *) BN_CTX_get() should return zero-valued bignums, providing the same
6294 initialised value as BN_new().
6295 [Geoff Thorpe, suggested by Ulf Möller]
6297 *) Support for inhibitAnyPolicy certificate extension.
6300 *) An audit of the BIGNUM code is underway, for which debugging code is
6301 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
6302 is considered valid when processing BIGNUMs, and causes execution to
6303 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
6304 further steps are taken to deliberately pollute unused data in BIGNUM
6305 structures to try and expose faulty code further on. For now, openssl will
6306 (in its default mode of operation) continue to tolerate the inconsistent
6307 forms that it has tolerated in the past, but authors and packagers should
6308 consider trying openssl and their own applications when compiled with
6309 these debugging symbols defined. It will help highlight potential bugs in
6310 their own code, and will improve the test coverage for OpenSSL itself. At
6311 some point, these tighter rules will become openssl's default to improve
6312 maintainability, though the assert()s and other overheads will remain only
6313 in debugging configurations. See bn.h for more details.
6314 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6316 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6317 that can only be obtained through BN_CTX_new() (which implicitly
6318 initialises it). The presence of this function only made it possible
6319 to overwrite an existing structure (and cause memory leaks).
6322 *) Because of the callback-based approach for implementing LHASH as a
6323 template type, lh_insert() adds opaque objects to hash-tables and
6324 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6325 to clean up those corresponding objects before destroying the hash table
6326 (and losing the object pointers). So some over-zealous constifications in
6327 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6328 objects as "const" and the lh_doall[_arg] callback wrappers are not
6329 prototyped to have "const" restrictions on the object pointers they are
6330 given (and so aren't required to cast them away any more).
6333 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6334 (speed) prefers to use its own implementation. The two implementations
6335 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6336 its object type properly exposed (MS_TM) instead of casting to/from "char
6337 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6338 aren't necessarily the greatest nomenclatures - but this is what was used
6339 internally to the implementation so I've used that for now.
6342 *) Ensure that deprecated functions do not get compiled when
6343 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6344 the self-tests were still using deprecated key-generation functions so
6345 these have been updated also.
6348 *) Reorganise PKCS#7 code to separate the digest location functionality
6349 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6350 New function PKCS7_set_digest() to set the digest type for PKCS#7
6351 digestedData type. Add additional code to correctly generate the
6352 digestedData type and add support for this type in PKCS7 initialization
6356 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6357 structure of type "other".
6360 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6361 sure the loop does correctly stop and breaking ("division by zero")
6362 modulus operations are not performed. The (pre-generated) prime
6363 table crypto/bn/bn_prime.h was already correct, but it could not be
6364 re-generated on some platforms because of the "division by zero"
6365 situation in the script.
6366 [Ralf S. Engelschall]
6368 *) Update support for ECC-based TLS ciphersuites according to
6369 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6370 SHA-1 now is only used for "small" curves (where the
6371 representation of a field element takes up to 24 bytes); for
6372 larger curves, the field element resulting from ECDH is directly
6373 used as premaster secret.
6374 [Douglas Stebila (Sun Microsystems Laboratories)]
6376 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6377 curve secp160r1 to the tests.
6378 [Douglas Stebila (Sun Microsystems Laboratories)]
6380 *) Add the possibility to load symbols globally with DSO.
6381 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6383 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6384 control of the error stack.
6387 *) Add support for STORE in ENGINE.
6390 *) Add the STORE type. The intention is to provide a common interface
6391 to certificate and key stores, be they simple file-based stores, or
6392 HSM-type store, or LDAP stores, or...
6393 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6396 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6397 pass a list of arguments to any function as well as provide a way
6398 for a function to pass data back to the caller.
6401 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6402 works like BUF_strdup() but can be used to duplicate a portion of
6403 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6407 *) Add the function sk_find_ex() which works like sk_find(), but will
6408 return an index to an element even if an exact match couldn't be
6409 found. The index is guaranteed to point at the element where the
6410 searched-for key would be inserted to preserve sorting order.
6413 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6414 takes an extra flags argument for optional functionality. Currently,
6415 the following flags are defined:
6417 OBJ_BSEARCH_VALUE_ON_NOMATCH
6418 This one gets OBJ_bsearch_ex() to return a pointer to the first
6419 element where the comparing function returns a negative or zero
6422 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6423 This one gets OBJ_bsearch_ex() to return a pointer to the first
6424 element where the comparing function returns zero. This is useful
6425 if there are more than one element where the comparing function
6429 *) Make it possible to create self-signed certificates with 'openssl ca'
6430 in such a way that the self-signed certificate becomes part of the
6431 CA database and uses the same mechanisms for serial number generation
6432 as all other certificate signing. The new flag '-selfsign' enables
6433 this functionality. Adapt CA.sh and CA.pl.in.
6436 *) Add functionality to check the public key of a certificate request
6437 against a given private. This is useful to check that a certificate
6438 request can be signed by that key (self-signing).
6441 *) Make it possible to have multiple active certificates with the same
6442 subject in the CA index file. This is done only if the keyword
6443 'unique_subject' is set to 'no' in the main CA section (default
6444 if 'CA_default') of the configuration file. The value is saved
6445 with the database itself in a separate index attribute file,
6446 named like the index file with '.attr' appended to the name.
6449 *) Generate multi-valued AVAs using '+' notation in config files for
6453 *) Support for nameConstraints certificate extension.
6456 *) Support for policyConstraints certificate extension.
6459 *) Support for policyMappings certificate extension.
6462 *) Make sure the default DSA_METHOD implementation only uses its
6463 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6464 and change its own handlers to be NULL so as to remove unnecessary
6465 indirection. This lets alternative implementations fallback to the
6466 default implementation more easily.
6469 *) Support for directoryName in GeneralName related extensions
6473 *) Make it possible to link applications using Makefile.shared.
6474 Make that possible even when linking against static libraries!
6477 *) Support for single pass processing for S/MIME signing. This now
6478 means that S/MIME signing can be done from a pipe, in addition
6479 cleartext signing (multipart/signed type) is effectively streaming
6480 and the signed data does not need to be all held in memory.
6482 This is done with a new flag PKCS7_STREAM. When this flag is set
6483 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6484 is done after the data is output (and digests calculated) in
6485 SMIME_write_PKCS7().
6488 *) Add full support for -rpath/-R, both in shared libraries and
6489 applications, at least on the platforms where it's known how
6493 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6494 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6495 will now compute a table of multiples of the generator that
6496 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6497 faster (notably in the case of a single point multiplication,
6498 scalar * generator).
6499 [Nils Larsch, Bodo Moeller]
6501 *) IPv6 support for certificate extensions. The various extensions
6502 which use the IP:a.b.c.d can now take IPv6 addresses using the
6503 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6507 *) Added an ENGINE that implements RSA by performing private key
6508 exponentiations with the GMP library. The conversions to and from
6509 GMP's mpz_t format aren't optimised nor are any montgomery forms
6510 cached, and on x86 it appears OpenSSL's own performance has caught up.
6511 However there are likely to be other architectures where GMP could
6512 provide a boost. This ENGINE is not built in by default, but it can be
6513 specified at Configure time and should be accompanied by the necessary
6514 linker additions, eg;
6515 ./config -DOPENSSL_USE_GMP -lgmp
6518 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6519 testing availability of engines with "-t" - the old behaviour is
6520 produced by increasing the feature's verbosity with "-tt".
6523 *) ECDSA routines: under certain error conditions uninitialized BN objects
6524 could be freed. Solution: make sure initialization is performed early
6525 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6529 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6530 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6531 software implementations. For DSA and DH, parameter generation can
6532 also be overridden by providing the appropriate method callbacks.
6535 *) Change the "progress" mechanism used in key-generation and
6536 primality testing to functions that take a new BN_GENCB pointer in
6537 place of callback/argument pairs. The new API functions have "_ex"
6538 postfixes and the older functions are reimplemented as wrappers for
6539 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6540 declarations of the old functions to help (graceful) attempts to
6541 migrate to the new functions. Also, the new key-generation API
6542 functions operate on a caller-supplied key-structure and return
6543 success/failure rather than returning a key or NULL - this is to
6544 help make "keygen" another member function of RSA_METHOD etc.
6546 Example for using the new callback interface:
6548 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6552 BN_GENCB_set(&my_cb, my_callback, my_arg);
6554 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6555 /* For the meaning of a, b in calls to my_callback(), see the
6556 * documentation of the function that calls the callback.
6557 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6558 * my_callback should return 1 if it wants BN_is_prime_ex()
6559 * to continue, or 0 to stop.
6564 *) Change the ZLIB compression method to be stateful, and make it
6565 available to TLS with the number defined in
6566 draft-ietf-tls-compression-04.txt.
6569 *) Add the ASN.1 structures and functions for CertificatePair, which
6570 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6572 CertificatePair ::= SEQUENCE {
6573 forward [0] Certificate OPTIONAL,
6574 reverse [1] Certificate OPTIONAL,
6575 -- at least one of the pair shall be present -- }
6577 Also implement the PEM functions to read and write certificate
6578 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6580 This needed to be defined, mostly for the sake of the LDAP
6581 attribute crossCertificatePair, but may prove useful elsewhere as
6585 *) Make it possible to inhibit symlinking of shared libraries in
6586 Makefile.shared, for Cygwin's sake.
6589 *) Extend the BIGNUM API by creating a function
6590 void BN_set_negative(BIGNUM *a, int neg);
6591 and a macro that behave like
6592 int BN_is_negative(const BIGNUM *a);
6594 to avoid the need to access 'a->neg' directly in applications.
6597 *) Implement fast modular reduction for pseudo-Mersenne primes
6598 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6599 EC_GROUP_new_curve_GFp() will now automatically use this
6601 [Nils Larsch <nla@trustcenter.de>]
6603 *) Add new lock type (CRYPTO_LOCK_BN).
6606 *) Change the ENGINE framework to automatically load engines
6607 dynamically from specific directories unless they could be
6608 found to already be built in or loaded. Move all the
6609 current engines except for the cryptodev one to a new
6611 The engines in engines/ are built as shared libraries if
6612 the "shared" options was given to ./Configure or ./config.
6613 Otherwise, they are inserted in libcrypto.a.
6614 /usr/local/ssl/engines is the default directory for dynamic
6615 engines, but that can be overridden at configure time through
6616 the usual use of --prefix and/or --openssldir, and at run
6617 time with the environment variable OPENSSL_ENGINES.
6618 [Geoff Thorpe and Richard Levitte]
6620 *) Add Makefile.shared, a helper makefile to build shared
6621 libraries. Adapt Makefile.org.
6624 *) Add version info to Win32 DLLs.
6625 [Peter 'Luna' Runestig" <peter@runestig.com>]
6627 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6628 can be added using this API to created arbitrary PKCS#12
6629 files while avoiding the low level API.
6631 New options to PKCS12_create(), key or cert can be NULL and
6632 will then be omitted from the output file. The encryption
6633 algorithm NIDs can be set to -1 for no encryption, the mac
6634 iteration count can be set to 0 to omit the mac.
6636 Enhance pkcs12 utility by making the -nokeys and -nocerts
6637 options work when creating a PKCS#12 file. New option -nomac
6638 to omit the mac, NONE can be set for an encryption algorithm.
6639 New code is modified to use the enhanced PKCS12_create()
6640 instead of the low level API.
6643 *) Extend ASN1 encoder to support indefinite length constructed
6644 encoding. This can output sequences tags and octet strings in
6645 this form. Modify pk7_asn1.c to support indefinite length
6646 encoding. This is experimental and needs additional code to
6647 be useful, such as an ASN1 bio and some enhanced streaming
6650 Extend template encode functionality so that tagging is passed
6651 down to the template encoder.
6654 *) Let 'openssl req' fail if an argument to '-newkey' is not
6655 recognized instead of using RSA as a default.
6658 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6659 As these are not official, they are not included in "ALL";
6660 the "ECCdraft" ciphersuite group alias can be used to select them.
6661 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6663 *) Add ECDH engine support.
6664 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6666 *) Add ECDH in new directory crypto/ecdh/.
6667 [Douglas Stebila (Sun Microsystems Laboratories)]
6669 *) Let BN_rand_range() abort with an error after 100 iterations
6670 without success (which indicates a broken PRNG).
6673 *) Change BN_mod_sqrt() so that it verifies that the input value
6674 is really the square of the return value. (Previously,
6675 BN_mod_sqrt would show GIGO behaviour.)
6678 *) Add named elliptic curves over binary fields from X9.62, SECG,
6679 and WAP/WTLS; add OIDs that were still missing.
6681 [Sheueling Chang Shantz and Douglas Stebila
6682 (Sun Microsystems Laboratories)]
6684 *) Extend the EC library for elliptic curves over binary fields
6685 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6688 EC_GF2m_simple_method
6692 EC_GROUP_new_curve_GF2m
6693 EC_GROUP_set_curve_GF2m
6694 EC_GROUP_get_curve_GF2m
6695 EC_POINT_set_affine_coordinates_GF2m
6696 EC_POINT_get_affine_coordinates_GF2m
6697 EC_POINT_set_compressed_coordinates_GF2m
6699 Point compression for binary fields is disabled by default for
6700 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6703 As binary polynomials are represented as BIGNUMs, various members
6704 of the EC_GROUP and EC_POINT data structures can be shared
6705 between the implementations for prime fields and binary fields;
6706 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6707 are essentially identical to their ..._GFp counterparts.
6708 (For simplicity, the '..._GFp' prefix has been dropped from
6709 various internal method names.)
6711 An internal 'field_div' method (similar to 'field_mul' and
6712 'field_sqr') has been added; this is used only for binary fields.
6714 [Sheueling Chang Shantz and Douglas Stebila
6715 (Sun Microsystems Laboratories)]
6717 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6718 through methods ('mul', 'precompute_mult').
6720 The generic implementations (now internally called 'ec_wNAF_mul'
6721 and 'ec_wNAF_precomputed_mult') remain the default if these
6722 methods are undefined.
6724 [Sheueling Chang Shantz and Douglas Stebila
6725 (Sun Microsystems Laboratories)]
6727 *) New function EC_GROUP_get_degree, which is defined through
6728 EC_METHOD. For curves over prime fields, this returns the bit
6729 length of the modulus.
6731 [Sheueling Chang Shantz and Douglas Stebila
6732 (Sun Microsystems Laboratories)]
6734 *) New functions EC_GROUP_dup, EC_POINT_dup.
6735 (These simply call ..._new and ..._copy).
6737 [Sheueling Chang Shantz and Douglas Stebila
6738 (Sun Microsystems Laboratories)]
6740 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6741 Polynomials are represented as BIGNUMs (where the sign bit is not
6742 used) in the following functions [macros]:
6745 BN_GF2m_sub [= BN_GF2m_add]
6746 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6747 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6748 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6750 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6751 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6752 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6753 BN_GF2m_cmp [= BN_ucmp]
6755 (Note that only the 'mod' functions are actually for fields GF(2^m).
6756 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6758 For some functions, an the irreducible polynomial defining a
6759 field can be given as an 'unsigned int[]' with strictly
6760 decreasing elements giving the indices of those bits that are set;
6761 i.e., p[] represents the polynomial
6762 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6764 p[0] > p[1] > ... > p[k] = 0.
6765 This applies to the following functions:
6770 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6771 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6773 BN_GF2m_mod_sqrt_arr
6774 BN_GF2m_mod_solve_quad_arr
6778 Conversion can be performed by the following functions:
6783 bntest.c has additional tests for binary polynomial arithmetic.
6785 Two implementations for BN_GF2m_mod_div() are available.
6786 The default algorithm simply uses BN_GF2m_mod_inv() and
6787 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6788 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6789 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6791 [Sheueling Chang Shantz and Douglas Stebila
6792 (Sun Microsystems Laboratories)]
6794 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6795 functionality is disabled at compile-time.
6796 [Douglas Stebila <douglas.stebila@sun.com>]
6798 *) Change default behaviour of 'openssl asn1parse' so that more
6799 information is visible when viewing, e.g., a certificate:
6801 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6802 mode the content of non-printable OCTET STRINGs is output in a
6803 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6804 avoid the appearance of a printable string.
6805 [Nils Larsch <nla@trustcenter.de>]
6807 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6809 EC_GROUP_set_asn1_flag()
6810 EC_GROUP_get_asn1_flag()
6811 EC_GROUP_set_point_conversion_form()
6812 EC_GROUP_get_point_conversion_form()
6813 These control ASN1 encoding details:
6814 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6815 has been set to OPENSSL_EC_NAMED_CURVE.
6816 - Points are encoded in uncompressed form by default; options for
6817 asn1_for are as for point2oct, namely
6818 POINT_CONVERSION_COMPRESSED
6819 POINT_CONVERSION_UNCOMPRESSED
6820 POINT_CONVERSION_HYBRID
6822 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6825 EC_GROUP_get0_seed()
6826 EC_GROUP_get_seed_len()
6827 This is used only for ASN1 purposes (so far).
6828 [Nils Larsch <nla@trustcenter.de>]
6830 *) Add 'field_type' member to EC_METHOD, which holds the NID
6831 of the appropriate field type OID. The new function
6832 EC_METHOD_get_field_type() returns this value.
6833 [Nils Larsch <nla@trustcenter.de>]
6838 EC_POINT_point2hex()
6839 EC_POINT_hex2point()
6840 providing useful interfaces to EC_POINT_point2oct() and
6841 EC_POINT_oct2point().
6842 [Nils Larsch <nla@trustcenter.de>]
6844 *) Change internals of the EC library so that the functions
6845 EC_GROUP_set_generator()
6846 EC_GROUP_get_generator()
6847 EC_GROUP_get_order()
6848 EC_GROUP_get_cofactor()
6849 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6850 to methods, which would lead to unnecessary code duplication when
6851 adding different types of curves.
6852 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6854 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6855 arithmetic, and such that modified wNAFs are generated
6856 (which avoid length expansion in many cases).
6859 *) Add a function EC_GROUP_check_discriminant() (defined via
6860 EC_METHOD) that verifies that the curve discriminant is non-zero.
6862 Add a function EC_GROUP_check() that makes some sanity tests
6863 on a EC_GROUP, its generator and order. This includes
6864 EC_GROUP_check_discriminant().
6865 [Nils Larsch <nla@trustcenter.de>]
6867 *) Add ECDSA in new directory crypto/ecdsa/.
6869 Add applications 'openssl ecparam' and 'openssl ecdsa'
6870 (these are based on 'openssl dsaparam' and 'openssl dsa').
6872 ECDSA support is also included in various other files across the
6873 library. Most notably,
6874 - 'openssl req' now has a '-newkey ecdsa:file' option;
6875 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6876 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6877 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6878 them suitable for ECDSA where domain parameters must be
6879 extracted before the specific public key;
6880 - ECDSA engine support has been added.
6881 [Nils Larsch <nla@trustcenter.de>]
6883 *) Include some named elliptic curves, and add OIDs from X9.62,
6884 SECG, and WAP/WTLS. Each curve can be obtained from the new
6886 EC_GROUP_new_by_curve_name(),
6887 and the list of available named curves can be obtained with
6888 EC_get_builtin_curves().
6889 Also add a 'curve_name' member to EC_GROUP objects, which can be
6891 EC_GROUP_set_curve_name()
6892 EC_GROUP_get_curve_name()
6893 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6895 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6896 was actually never needed) and in BN_mul(). The removal in BN_mul()
6897 required a small change in bn_mul_part_recursive() and the addition
6898 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6899 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6900 bn_sub_words() and bn_add_words() except they take arrays with
6904 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6906 *) Cleanse PEM buffers before freeing them since they may contain
6908 [Benjamin Bennett <ben@psc.edu>]
6910 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6911 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6912 authentication-only ciphersuites.
6915 *) Since AES128 and AES256 share a single mask bit in the logic of
6916 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6917 kludge to work properly if AES128 is available and AES256 isn't.
6920 *) Expand security boundary to match 1.1.1 module.
6923 *) Remove redundant features: hash file source, editing of test vectors
6924 modify fipsld to use external fips_premain.c signature.
6927 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6928 run algorithm test programs.
6931 *) Make algorithm test programs more tolerant of whitespace.
6934 *) Have SSL/TLS server implementation tolerate "mismatched" record
6935 protocol version while receiving ClientHello even if the
6936 ClientHello is fragmented. (The server can't insist on the
6937 particular protocol version it has chosen before the ServerHello
6938 message has informed the client about his choice.)
6941 *) Load error codes if they are not already present instead of using a
6942 static variable. This allows them to be cleanly unloaded and reloaded.
6945 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6947 *) Introduce limits to prevent malicious keys being able to
6948 cause a denial of service. (CVE-2006-2940)
6949 [Steve Henson, Bodo Moeller]
6951 *) Fix ASN.1 parsing of certain invalid structures that can result
6952 in a denial of service. (CVE-2006-2937) [Steve Henson]
6954 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6955 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6957 *) Fix SSL client code which could crash if connecting to a
6958 malicious SSLv2 server. (CVE-2006-4343)
6959 [Tavis Ormandy and Will Drewry, Google Security Team]
6961 *) Change ciphersuite string processing so that an explicit
6962 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6963 will no longer include "AES128-SHA"), and any other similar
6964 ciphersuite (same bitmap) from *other* protocol versions (so that
6965 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6966 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6967 changes from 0.9.8b and 0.9.8d.
6970 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6972 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6973 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6975 *) Change the Unix randomness entropy gathering to use poll() when
6976 possible instead of select(), since the latter has some
6977 undesirable limitations.
6978 [Darryl Miles via Richard Levitte and Bodo Moeller]
6980 *) Disable rogue ciphersuites:
6982 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6983 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6984 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6986 The latter two were purportedly from
6987 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6990 Also deactivate the remaining ciphersuites from
6991 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6992 unofficial, and the ID has long expired.
6995 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6996 dual-core machines) and other potential thread-safety issues.
6999 Changes between 0.9.7i and 0.9.7j [04 May 2006]
7001 *) Adapt fipsld and the build system to link against the validated FIPS
7002 module in FIPS mode.
7005 *) Fixes for VC++ 2005 build under Windows.
7008 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
7009 from a Windows bash shell such as MSYS. It is autodetected from the
7010 "config" script when run from a VC++ environment. Modify standard VC++
7011 build to use fipscanister.o from the GNU make build.
7014 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
7016 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
7017 The value now differs depending on if you build for FIPS or not.
7018 BEWARE! A program linked with a shared FIPSed libcrypto can't be
7019 safely run with a non-FIPSed libcrypto, as it may crash because of
7020 the difference induced by this change.
7023 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
7025 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
7026 (part of SSL_OP_ALL). This option used to disable the
7027 countermeasure against man-in-the-middle protocol-version
7028 rollback in the SSL 2.0 server implementation, which is a bad
7029 idea. (CVE-2005-2969)
7031 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
7032 for Information Security, National Institute of Advanced Industrial
7033 Science and Technology [AIST], Japan)]
7035 *) Minimal support for X9.31 signatures and PSS padding modes. This is
7036 mainly for FIPS compliance and not fully integrated at this stage.
7039 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
7040 the exponentiation using a fixed-length exponent. (Otherwise,
7041 the information leaked through timing could expose the secret key
7042 after many signatures; cf. Bleichenbacher's attack on DSA with
7046 *) Make a new fixed-window mod_exp implementation the default for
7047 RSA, DSA, and DH private-key operations so that the sequence of
7048 squares and multiplies and the memory access pattern are
7049 independent of the particular secret key. This will mitigate
7050 cache-timing and potential related attacks.
7052 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
7053 and this is automatically used by BN_mod_exp_mont() if the new flag
7054 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
7055 will use this BN flag for private exponents unless the flag
7056 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
7057 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
7059 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
7061 *) Change the client implementation for SSLv23_method() and
7062 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
7063 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
7064 (Previously, the SSL 2.0 backwards compatible Client Hello
7065 message format would be used even with SSL_OP_NO_SSLv2.)
7068 *) Add support for smime-type MIME parameter in S/MIME messages which some
7072 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
7073 a threadsafe manner. Modify rsa code to use new function and add calls
7074 to dsa and dh code (which had race conditions before).
7077 *) Include the fixed error library code in the C error file definitions
7078 instead of fixing them up at runtime. This keeps the error code
7079 structures constant.
7082 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
7084 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
7087 *) Fixes for newer kerberos headers. NB: the casts are needed because
7088 the 'length' field is signed on one version and unsigned on another
7089 with no (?) obvious way to tell the difference, without these VC++
7090 complains. Also the "definition" of FAR (blank) is no longer included
7091 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
7092 some needed definitions.
7095 *) Undo Cygwin change.
7098 *) Added support for proxy certificates according to RFC 3820.
7099 Because they may be a security thread to unaware applications,
7100 they must be explicitly allowed in run-time. See
7101 docs/HOWTO/proxy_certificates.txt for further information.
7104 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
7106 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
7107 server and client random values. Previously
7108 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
7109 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
7111 This change has negligible security impact because:
7113 1. Server and client random values still have 24 bytes of pseudo random
7116 2. Server and client random values are sent in the clear in the initial
7119 3. The master secret is derived using the premaster secret (48 bytes in
7120 size for static RSA ciphersuites) as well as client server and random
7123 The OpenSSL team would like to thank the UK NISCC for bringing this issue
7126 [Stephen Henson, reported by UK NISCC]
7128 *) Use Windows randomness collection on Cygwin.
7131 *) Fix hang in EGD/PRNGD query when communication socket is closed
7132 prematurely by EGD/PRNGD.
7133 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
7135 *) Prompt for pass phrases when appropriate for PKCS12 input format.
7138 *) Back-port of selected performance improvements from development
7139 branch, as well as improved support for PowerPC platforms.
7142 *) Add lots of checks for memory allocation failure, error codes to indicate
7143 failure and freeing up memory if a failure occurs.
7144 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
7146 *) Add new -passin argument to dgst.
7149 *) Perform some character comparisons of different types in X509_NAME_cmp:
7150 this is needed for some certificates that re-encode DNs into UTF8Strings
7151 (in violation of RFC3280) and can't or won't issue name rollover
7155 *) Make an explicit check during certificate validation to see that
7156 the CA setting in each certificate on the chain is correct. As a
7157 side effect always do the following basic checks on extensions,
7158 not just when there's an associated purpose to the check:
7160 - if there is an unhandled critical extension (unless the user
7161 has chosen to ignore this fault)
7162 - if the path length has been exceeded (if one is set at all)
7163 - that certain extensions fit the associated purpose (if one has
7167 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
7169 *) Avoid a race condition when CRLs are checked in a multi threaded
7170 environment. This would happen due to the reordering of the revoked
7171 entries during signature checking and serial number lookup. Now the
7172 encoding is cached and the serial number sort performed under a lock.
7173 Add new STACK function sk_is_sorted().
7176 *) Add Delta CRL to the extension code.
7179 *) Various fixes to s3_pkt.c so alerts are sent properly.
7180 [David Holmes <d.holmes@f5.com>]
7182 *) Reduce the chances of duplicate issuer name and serial numbers (in
7183 violation of RFC3280) using the OpenSSL certificate creation utilities.
7184 This is done by creating a random 64 bit value for the initial serial
7185 number when a serial number file is created or when a self signed
7186 certificate is created using 'openssl req -x509'. The initial serial
7187 number file is created using 'openssl x509 -next_serial' in CA.pl
7188 rather than being initialized to 1.
7191 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
7193 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7194 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7195 [Joe Orton, Steve Henson]
7197 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
7199 [Joe Orton, Steve Henson]
7201 *) Make it possible to have multiple active certificates with the same
7202 subject in the CA index file. This is done only if the keyword
7203 'unique_subject' is set to 'no' in the main CA section (default
7204 if 'CA_default') of the configuration file. The value is saved
7205 with the database itself in a separate index attribute file,
7206 named like the index file with '.attr' appended to the name.
7209 *) X509 verify fixes. Disable broken certificate workarounds when
7210 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
7211 keyUsage extension present. Don't accept CRLs with unhandled critical
7212 extensions: since verify currently doesn't process CRL extensions this
7213 rejects a CRL with *any* critical extensions. Add new verify error codes
7217 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
7218 A clarification of RFC2560 will require the use of OCTET STRINGs and
7219 some implementations cannot handle the current raw format. Since OpenSSL
7220 copies and compares OCSP nonces as opaque blobs without any attempt at
7221 parsing them this should not create any compatibility issues.
7224 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
7225 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
7226 this HMAC (and other) operations are several times slower than OpenSSL
7230 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
7231 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7233 *) Use the correct content when signing type "other".
7236 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
7238 *) Fix various bugs revealed by running the NISCC test suite:
7240 Stop out of bounds reads in the ASN1 code when presented with
7241 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7243 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
7245 If verify callback ignores invalid public key errors don't try to check
7246 certificate signature with the NULL public key.
7250 *) New -ignore_err option in ocsp application to stop the server
7251 exiting on the first error in a request.
7254 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7255 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7259 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7260 extra data after the compression methods not only for TLS 1.0
7261 but also for SSL 3.0 (as required by the specification).
7262 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7264 *) Change X509_certificate_type() to mark the key as exported/exportable
7265 when it's 512 *bits* long, not 512 bytes.
7268 *) Change AES_cbc_encrypt() so it outputs exact multiple of
7269 blocks during encryption.
7272 *) Various fixes to base64 BIO and non blocking I/O. On write
7273 flushes were not handled properly if the BIO retried. On read
7274 data was not being buffered properly and had various logic bugs.
7275 This also affects blocking I/O when the data being decoded is a
7279 *) Various S/MIME bugfixes and compatibility changes:
7280 output correct application/pkcs7 MIME type if
7281 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
7282 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
7283 of files as .eml work). Correctly handle very long lines in MIME
7287 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
7289 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7290 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7291 a protocol version number mismatch like a decryption error
7292 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7295 *) Turn on RSA blinding by default in the default implementation
7296 to avoid a timing attack. Applications that don't want it can call
7297 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7298 They would be ill-advised to do so in most cases.
7299 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7301 *) Change RSA blinding code so that it works when the PRNG is not
7302 seeded (in this case, the secret RSA exponent is abused as
7303 an unpredictable seed -- if it is not unpredictable, there
7304 is no point in blinding anyway). Make RSA blinding thread-safe
7305 by remembering the creator's thread ID in rsa->blinding and
7306 having all other threads use local one-time blinding factors
7307 (this requires more computation than sharing rsa->blinding, but
7308 avoids excessive locking; and if an RSA object is not shared
7309 between threads, blinding will still be very fast).
7312 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
7313 ENGINE as defaults for all supported algorithms irrespective of
7314 the 'flags' parameter. 'flags' is now honoured, so applications
7315 should make sure they are passing it correctly.
7318 *) Target "mingw" now allows native Windows code to be generated in
7319 the Cygwin environment as well as with the MinGW compiler.
7322 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7324 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7325 via timing by performing a MAC computation even if incorrect
7326 block cipher padding has been found. This is a countermeasure
7327 against active attacks where the attacker has to distinguish
7328 between bad padding and a MAC verification error. (CVE-2003-0078)
7330 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7331 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7332 Martin Vuagnoux (EPFL, Ilion)]
7334 *) Make the no-err option work as intended. The intention with no-err
7335 is not to have the whole error stack handling routines removed from
7336 libcrypto, it's only intended to remove all the function name and
7337 reason texts, thereby removing some of the footprint that may not
7338 be interesting if those errors aren't displayed anyway.
7340 NOTE: it's still possible for any application or module to have its
7341 own set of error texts inserted. The routines are there, just not
7342 used by default when no-err is given.
7345 *) Add support for FreeBSD on IA64.
7346 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7348 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7349 Kerberos function mit_des_cbc_cksum(). Before this change,
7350 the value returned by DES_cbc_cksum() was like the one from
7351 mit_des_cbc_cksum(), except the bytes were swapped.
7352 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7354 *) Allow an application to disable the automatic SSL chain building.
7355 Before this a rather primitive chain build was always performed in
7356 ssl3_output_cert_chain(): an application had no way to send the
7357 correct chain if the automatic operation produced an incorrect result.
7359 Now the chain builder is disabled if either:
7361 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7363 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7365 The reasoning behind this is that an application would not want the
7366 auto chain building to take place if extra chain certificates are
7367 present and it might also want a means of sending no additional
7368 certificates (for example the chain has two certificates and the
7372 *) Add the possibility to build without the ENGINE framework.
7373 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7375 *) Under Win32 gmtime() can return NULL: check return value in
7376 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7379 *) DSA routines: under certain error conditions uninitialized BN objects
7380 could be freed. Solution: make sure initialization is performed early
7381 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7382 Nils Larsch <nla@trustcenter.de> via PR#459)
7385 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7386 checked on reconnect on the client side, therefore session resumption
7387 could still fail with a "ssl session id is different" error. This
7388 behaviour is masked when SSL_OP_ALL is used due to
7389 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7390 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7391 followup to PR #377.
7394 *) IA-32 assembler support enhancements: unified ELF targets, support
7395 for SCO/Caldera platforms, fix for Cygwin shared build.
7398 *) Add support for FreeBSD on sparc64. As a consequence, support for
7399 FreeBSD on non-x86 processors is separate from x86 processors on
7400 the config script, much like the NetBSD support.
7401 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7403 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7405 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7408 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7409 code (06) was taken as the first octet of the session ID and the last
7410 octet was ignored consequently. As a result SSLv2 client side session
7411 caching could not have worked due to the session ID mismatch between
7413 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7417 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7418 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7422 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7423 seems that in spite of existing for more than a year, many application
7424 author have done nothing to provide the necessary callbacks, which
7425 means that this particular engine will not work properly anywhere.
7426 This is a very unfortunate situation which forces us, in the name
7427 of usability, to give the hw_ncipher.c a static lock, which is part
7429 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7430 appear in 0.9.8 or later. We EXPECT application authors to have
7431 dealt properly with this when 0.9.8 is released (unless we actually
7432 make such changes in the libcrypto locking code that changes will
7433 have to be made anyway).
7436 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7437 octets have been read, EOF or an error occurs. Without this change
7438 some truncated ASN1 structures will not produce an error.
7441 *) Disable Heimdal support, since it hasn't been fully implemented.
7442 Still give the possibility to force the use of Heimdal, but with
7443 warnings and a request that patches get sent to openssl-dev.
7446 *) Add the VC-CE target, introduce the WINCE sysname, and add
7447 INSTALL.WCE and appropriate conditionals to make it build.
7448 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7450 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7451 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7452 edit numbers of the version.
7453 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7455 *) Introduce safe string copy and catenation functions
7456 (BUF_strlcpy() and BUF_strlcat()).
7457 [Ben Laurie (CHATS) and Richard Levitte]
7459 *) Avoid using fixed-size buffers for one-line DNs.
7460 [Ben Laurie (CHATS)]
7462 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7463 resizing buffers containing secrets, and use where appropriate.
7464 [Ben Laurie (CHATS)]
7466 *) Avoid using fixed size buffers for configuration file location.
7467 [Ben Laurie (CHATS)]
7469 *) Avoid filename truncation for various CA files.
7470 [Ben Laurie (CHATS)]
7472 *) Use sizeof in preference to magic numbers.
7473 [Ben Laurie (CHATS)]
7475 *) Avoid filename truncation in cert requests.
7476 [Ben Laurie (CHATS)]
7478 *) Add assertions to check for (supposedly impossible) buffer
7480 [Ben Laurie (CHATS)]
7482 *) Don't cache truncated DNS entries in the local cache (this could
7483 potentially lead to a spoofing attack).
7484 [Ben Laurie (CHATS)]
7486 *) Fix various buffers to be large enough for hex/decimal
7487 representations in a platform independent manner.
7488 [Ben Laurie (CHATS)]
7490 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7491 resizing buffers containing secrets, and use where appropriate.
7492 [Ben Laurie (CHATS)]
7494 *) Add BIO_indent() to avoid much slightly worrying code to do
7496 [Ben Laurie (CHATS)]
7498 *) Convert sprintf()/BIO_puts() to BIO_printf().
7499 [Ben Laurie (CHATS)]
7501 *) buffer_gets() could terminate with the buffer only half
7503 [Ben Laurie (CHATS)]
7505 *) Add assertions to prevent user-supplied crypto functions from
7506 overflowing internal buffers by having large block sizes, etc.
7507 [Ben Laurie (CHATS)]
7509 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7511 [Ben Laurie (CHATS)]
7513 *) Eliminate unused copy of key in RC4.
7514 [Ben Laurie (CHATS)]
7516 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7517 [Ben Laurie (CHATS)]
7519 *) Fix off-by-one error in EGD path.
7520 [Ben Laurie (CHATS)]
7522 *) If RANDFILE path is too long, ignore instead of truncating.
7523 [Ben Laurie (CHATS)]
7525 *) Eliminate unused and incorrectly sized X.509 structure
7527 [Ben Laurie (CHATS)]
7529 *) Eliminate unused and dangerous function knumber().
7530 [Ben Laurie (CHATS)]
7532 *) Eliminate unused and dangerous structure, KSSL_ERR.
7533 [Ben Laurie (CHATS)]
7535 *) Protect against overlong session ID context length in an encoded
7536 session object. Since these are local, this does not appear to be
7538 [Ben Laurie (CHATS)]
7540 *) Change from security patch (see 0.9.6e below) that did not affect
7541 the 0.9.6 release series:
7543 Remote buffer overflow in SSL3 protocol - an attacker could
7544 supply an oversized master key in Kerberos-enabled versions.
7546 [Ben Laurie (CHATS)]
7548 *) Change the SSL kerb5 codes to match RFC 2712.
7551 *) Make -nameopt work fully for req and add -reqopt switch.
7552 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7554 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7555 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7557 *) Make sure tests can be performed even if the corresponding algorithms
7558 have been removed entirely. This was also the last step to make
7559 OpenSSL compilable with DJGPP under all reasonable conditions.
7560 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7562 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7563 to allow version independent disabling of normally unselected ciphers,
7564 which may be activated as a side-effect of selecting a single cipher.
7566 (E.g., cipher list string "RSA" enables ciphersuites that are left
7567 out of "ALL" because they do not provide symmetric encryption.
7568 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7569 [Lutz Jaenicke, Bodo Moeller]
7571 *) Add appropriate support for separate platform-dependent build
7572 directories. The recommended way to make a platform-dependent
7573 build directory is the following (tested on Linux), maybe with
7576 # Place yourself outside of the OpenSSL source tree. In
7577 # this example, the environment variable OPENSSL_SOURCE
7578 # is assumed to contain the absolute OpenSSL source directory.
7579 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7580 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7581 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7582 mkdir -p `dirname $F`
7583 ln -s $OPENSSL_SOURCE/$F $F
7586 To be absolutely sure not to disturb the source tree, a "make clean"
7587 is a good thing. If it isn't successful, don't worry about it,
7588 it probably means the source directory is very clean.
7591 *) Make sure any ENGINE control commands make local copies of string
7592 pointers passed to them whenever necessary. Otherwise it is possible
7593 the caller may have overwritten (or deallocated) the original string
7594 data when a later ENGINE operation tries to use the stored values.
7595 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7597 *) Improve diagnostics in file reading and command-line digests.
7598 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7600 *) Add AES modes CFB and OFB to the object database. Correct an
7601 error in AES-CFB decryption.
7604 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7605 allows existing EVP_CIPHER_CTX structures to be reused after
7606 calling EVP_*Final(). This behaviour is used by encryption
7607 BIOs and some applications. This has the side effect that
7608 applications must explicitly clean up cipher contexts with
7609 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7612 *) Check the values of dna and dnb in bn_mul_recursive before calling
7613 bn_mul_comba (a non zero value means the a or b arrays do not contain
7614 n2 elements) and fallback to bn_mul_normal if either is not zero.
7617 *) Fix escaping of non-ASCII characters when using the -subj option
7618 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7621 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7622 form for "surname", serialNumber has no short form.
7623 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7624 therefore remove "mail" short name for "internet 7".
7625 The OID for unique identifiers in X509 certificates is
7626 x500UniqueIdentifier, not uniqueIdentifier.
7627 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7630 *) Add an "init" command to the ENGINE config module and auto initialize
7631 ENGINEs. Without any "init" command the ENGINE will be initialized
7632 after all ctrl commands have been executed on it. If init=1 the
7633 ENGINE is initialized at that point (ctrls before that point are run
7634 on the uninitialized ENGINE and after on the initialized one). If
7635 init=0 then the ENGINE will not be initialized at all.
7638 *) Fix the 'app_verify_callback' interface so that the user-defined
7639 argument is actually passed to the callback: In the
7640 SSL_CTX_set_cert_verify_callback() prototype, the callback
7641 declaration has been changed from
7644 int (*cb)(X509_STORE_CTX *,void *);
7645 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7646 i=s->ctx->app_verify_callback(&ctx)
7647 has been changed into
7648 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7650 To update applications using SSL_CTX_set_cert_verify_callback(),
7651 a dummy argument can be added to their callback functions.
7652 [D. K. Smetters <smetters@parc.xerox.com>]
7654 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7655 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7657 *) Add and OPENSSL_LOAD_CONF define which will cause
7658 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7659 This allows older applications to transparently support certain
7660 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7661 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7662 load the config file and OPENSSL_add_all_algorithms_conf() which will
7663 always load it have also been added.
7666 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7667 Adjust NIDs and EVP layer.
7668 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7670 *) Config modules support in openssl utility.
7672 Most commands now load modules from the config file,
7673 though in a few (such as version) this isn't done
7674 because it couldn't be used for anything.
7676 In the case of ca and req the config file used is
7677 the same as the utility itself: that is the -config
7678 command line option can be used to specify an
7682 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7683 use "openssl_conf" if filename is NULL use default openssl config file.
7686 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7687 config section name. Add a new flag to tolerate a missing config file
7688 and move code to CONF_modules_load_file().
7691 *) Support for crypto accelerator cards from Accelerated Encryption
7692 Processing, www.aep.ie. (Use engine 'aep')
7693 The support was copied from 0.9.6c [engine] and adapted/corrected
7694 to work with the new engine framework.
7695 [AEP Inc. and Richard Levitte]
7697 *) Support for SureWare crypto accelerator cards from Baltimore
7698 Technologies. (Use engine 'sureware')
7699 The support was copied from 0.9.6c [engine] and adapted
7700 to work with the new engine framework.
7703 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7704 make the newer ENGINE framework commands for the CHIL engine work.
7705 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7707 *) Make it possible to produce shared libraries on ReliantUNIX.
7708 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7710 *) Add the configuration target debug-linux-ppro.
7711 Make 'openssl rsa' use the general key loading routines
7712 implemented in apps.c, and make those routines able to
7713 handle the key format FORMAT_NETSCAPE and the variant
7715 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7717 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7718 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7720 *) Add -keyform to rsautl, and document -engine.
7721 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7723 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7724 BIO_R_NO_SUCH_FILE error code rather than the generic
7725 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7728 *) Add new functions
7730 ERR_peek_last_error_line
7731 ERR_peek_last_error_line_data.
7732 These are similar to
7735 ERR_peek_error_line_data,
7736 but report on the latest error recorded rather than the first one
7737 still in the error queue.
7738 [Ben Laurie, Bodo Moeller]
7740 *) default_algorithms option in ENGINE config module. This allows things
7742 default_algorithms = ALL
7743 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7746 *) Preliminary ENGINE config module.
7749 *) New experimental application configuration code.
7752 *) Change the AES code to follow the same name structure as all other
7753 symmetric ciphers, and behave the same way. Move everything to
7754 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7755 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7757 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7758 [Ben Laurie and Theo de Raadt]
7760 *) Add option to output public keys in req command.
7761 [Massimiliano Pala madwolf@openca.org]
7763 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7764 (up to about 10% better than before for P-192 and P-224).
7767 *) New functions/macros
7769 SSL_CTX_set_msg_callback(ctx, cb)
7770 SSL_CTX_set_msg_callback_arg(ctx, arg)
7771 SSL_set_msg_callback(ssl, cb)
7772 SSL_set_msg_callback_arg(ssl, arg)
7774 to request calling a callback function
7776 void cb(int write_p, int version, int content_type,
7777 const void *buf, size_t len, SSL *ssl, void *arg)
7779 whenever a protocol message has been completely received
7780 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7781 protocol version according to which the SSL library interprets
7782 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7783 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7784 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7785 specification (change_cipher_spec(20), alert(21), handshake(22)).
7786 'buf' and 'len' point to the actual message, 'ssl' to the
7787 SSL object, and 'arg' is the application-defined value set by
7788 SSL[_CTX]_set_msg_callback_arg().
7790 'openssl s_client' and 'openssl s_server' have new '-msg' options
7791 to enable a callback that displays all protocol messages.
7794 *) Change the shared library support so shared libraries are built as
7795 soon as the corresponding static library is finished, and thereby get
7796 openssl and the test programs linked against the shared library.
7797 This still only happens when the keyword "shard" has been given to
7798 the configuration scripts.
7800 NOTE: shared library support is still an experimental thing, and
7801 backward binary compatibility is still not guaranteed.
7802 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7804 *) Add support for Subject Information Access extension.
7805 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7807 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7808 additional bytes when new memory had to be allocated, not just
7809 when reusing an existing buffer.
7812 *) New command line and configuration option 'utf8' for the req command.
7813 This allows field values to be specified as UTF8 strings.
7816 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7817 runs for the former and machine-readable output for the latter.
7820 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7821 of the e-mail address in the DN (i.e., it will go into a certificate
7822 extension only). The new configuration file option 'email_in_dn = no'
7823 has the same effect.
7824 [Massimiliano Pala madwolf@openca.org]
7826 *) Change all functions with names starting with des_ to be starting
7827 with DES_ instead. Add wrappers that are compatible with libdes,
7828 but are named _ossl_old_des_*. Finally, add macros that map the
7829 des_* symbols to the corresponding _ossl_old_des_* if libdes
7830 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7831 desired, the des_* symbols will be mapped to DES_*, with one
7834 Since we provide two compatibility mappings, the user needs to
7835 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7836 compatibility is desired. The default (i.e., when that macro
7837 isn't defined) is OpenSSL 0.9.6c compatibility.
7839 There are also macros that enable and disable the support of old
7840 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7841 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7842 are defined, the default will apply: to support the old des routines.
7844 In either case, one must include openssl/des.h to get the correct
7845 definitions. Do not try to just include openssl/des_old.h, that
7848 NOTE: This is a major break of an old API into a new one. Software
7849 authors are encouraged to switch to the DES_ style functions. Some
7850 time in the future, des_old.h and the libdes compatibility functions
7851 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7852 default), and then completely removed.
7855 *) Test for certificates which contain unsupported critical extensions.
7856 If such a certificate is found during a verify operation it is
7857 rejected by default: this behaviour can be overridden by either
7858 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7859 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7860 X509_supported_extension() has also been added which returns 1 if a
7861 particular extension is supported.
7864 *) Modify the behaviour of EVP cipher functions in similar way to digests
7865 to retain compatibility with existing code.
7868 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7869 compatibility with existing code. In particular the 'ctx' parameter does
7870 not have to be to be initialized before the call to EVP_DigestInit() and
7871 it is tidied up after a call to EVP_DigestFinal(). New function
7872 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7873 EVP_MD_CTX_copy() changed to not require the destination to be
7874 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7875 requires the destination to be valid.
7877 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7878 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7881 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7882 so that complete 'Handshake' protocol structures are kept in memory
7883 instead of overwriting 'msg_type' and 'length' with 'body' data.
7886 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7887 [Massimo Santin via Richard Levitte]
7889 *) Major restructuring to the underlying ENGINE code. This includes
7890 reduction of linker bloat, separation of pure "ENGINE" manipulation
7891 (initialisation, etc) from functionality dealing with implementations
7892 of specific crypto interfaces. This change also introduces integrated
7893 support for symmetric ciphers and digest implementations - so ENGINEs
7894 can now accelerate these by providing EVP_CIPHER and EVP_MD
7895 implementations of their own. This is detailed in crypto/engine/README
7896 as it couldn't be adequately described here. However, there are a few
7897 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7898 were changed in the original introduction of ENGINE code have now
7899 reverted back - the hooking from this code to ENGINE is now a good
7900 deal more passive and at run-time, operations deal directly with
7901 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7902 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7903 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7904 they were not being used by the framework as there is no concept of a
7905 BIGNUM_METHOD and they could not be generalised to the new
7906 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7907 ENGINE_cpy() has been removed as it cannot be consistently defined in
7911 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7914 *) Change mkdef.pl to sort symbols that get the same entry number,
7915 and make sure the automatically generated functions ERR_load_*
7916 become part of libeay.num as well.
7919 *) New function SSL_renegotiate_pending(). This returns true once
7920 renegotiation has been requested (either SSL_renegotiate() call
7921 or HelloRequest/ClientHello received from the peer) and becomes
7922 false once a handshake has been completed.
7923 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7924 sends a HelloRequest, but does not ensure that a handshake takes
7925 place. SSL_renegotiate_pending() is useful for checking if the
7926 client has followed the request.)
7929 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7930 By default, clients may request session resumption even during
7931 renegotiation (if session ID contexts permit); with this option,
7932 session resumption is possible only in the first handshake.
7934 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7935 more bits available for options that should not be part of
7936 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7939 *) Add some demos for certificate and certificate request creation.
7942 *) Make maximum certificate chain size accepted from the peer application
7943 settable (SSL*_get/set_max_cert_list()), as proposed by
7944 "Douglas E. Engert" <deengert@anl.gov>.
7947 *) Add support for shared libraries for Unixware-7
7948 (Boyd Lynn Gerber <gerberb@zenez.com>).
7951 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7952 be done prior to destruction. Use this to unload error strings from
7953 ENGINEs that load their own error strings. NB: This adds two new API
7954 functions to "get" and "set" this destroy handler in an ENGINE.
7957 *) Alter all existing ENGINE implementations (except "openssl" and
7958 "openbsd") to dynamically instantiate their own error strings. This
7959 makes them more flexible to be built both as statically-linked ENGINEs
7960 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7961 Also, add stub code to each that makes building them as self-contained
7962 shared-libraries easier (see README.ENGINE).
7965 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7966 implementations into applications that are completely implemented in
7967 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7968 commands that can be used to configure what shared-library to load and
7969 to control aspects of the way it is handled. Also, made an update to
7970 the README.ENGINE file that brings its information up-to-date and
7971 provides some information and instructions on the "dynamic" ENGINE
7972 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7975 *) Make it possible to unload ranges of ERR strings with a new
7976 "ERR_unload_strings" function.
7979 *) Add a copy() function to EVP_MD.
7982 *) Make EVP_MD routines take a context pointer instead of just the
7983 md_data void pointer.
7986 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7987 that the digest can only process a single chunk of data
7988 (typically because it is provided by a piece of
7989 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7990 is only going to provide a single chunk of data, and hence the
7991 framework needn't accumulate the data for oneshot drivers.
7994 *) As with "ERR", make it possible to replace the underlying "ex_data"
7995 functions. This change also alters the storage and management of global
7996 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7997 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7998 index counters. The API functions that use this state have been changed
7999 to take a "class_index" rather than pointers to the class's local STACK
8000 and counter, and there is now an API function to dynamically create new
8001 classes. This centralisation allows us to (a) plug a lot of the
8002 thread-safety problems that existed, and (b) makes it possible to clean
8003 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
8004 such data would previously have always leaked in application code and
8005 workarounds were in place to make the memory debugging turn a blind eye
8006 to it. Application code that doesn't use this new function will still
8007 leak as before, but their memory debugging output will announce it now
8008 rather than letting it slide.
8010 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
8011 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
8012 has a return value to indicate success or failure.
8015 *) Make it possible to replace the underlying "ERR" functions such that the
8016 global state (2 LHASH tables and 2 locks) is only used by the "default"
8017 implementation. This change also adds two functions to "get" and "set"
8018 the implementation prior to it being automatically set the first time
8019 any other ERR function takes place. Ie. an application can call "get",
8020 pass the return value to a module it has just loaded, and that module
8021 can call its own "set" function using that value. This means the
8022 module's "ERR" operations will use (and modify) the error state in the
8023 application and not in its own statically linked copy of OpenSSL code.
8026 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
8027 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
8028 the operation, and provides a more encapsulated way for external code
8029 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
8030 to use these functions rather than manually incrementing the counts.
8032 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
8035 *) Add EVP test program.
8038 *) Add symmetric cipher support to ENGINE. Expect the API to change!
8041 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
8042 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
8043 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
8044 These allow a CRL to be built without having to access X509_CRL fields
8045 directly. Modify 'ca' application to use new functions.
8048 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
8049 bug workarounds. Rollback attack detection is a security feature.
8050 The problem will only arise on OpenSSL servers when TLSv1 is not
8051 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
8052 Software authors not wanting to support TLSv1 will have special reasons
8053 for their choice and can explicitly enable this option.
8054 [Bodo Moeller, Lutz Jaenicke]
8056 *) Rationalise EVP so it can be extended: don't include a union of
8057 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
8058 (similar to those existing for EVP_CIPHER_CTX).
8063 EVP_MD_CTX_init(&md); /* new function call */
8064 EVP_DigestInit(&md, EVP_sha1());
8065 EVP_DigestUpdate(&md, in, len);
8066 EVP_DigestFinal(&md, out, NULL);
8067 EVP_MD_CTX_cleanup(&md); /* new function call */
8071 *) Make DES key schedule conform to the usual scheme, as well as
8072 correcting its structure. This means that calls to DES functions
8073 now have to pass a pointer to a des_key_schedule instead of a
8074 plain des_key_schedule (which was actually always a pointer
8077 des_key_schedule ks;
8079 des_set_key_checked(..., &ks);
8080 des_ncbc_encrypt(..., &ks, ...);
8082 (Note that a later change renames 'des_...' into 'DES_...'.)
8085 *) Initial reduction of linker bloat: the use of some functions, such as
8086 PEM causes large amounts of unused functions to be linked in due to
8087 poor organisation. For example pem_all.c contains every PEM function
8088 which has a knock on effect of linking in large amounts of (unused)
8089 ASN1 code. Grouping together similar functions and splitting unrelated
8090 functions prevents this.
8093 *) Cleanup of EVP macros.
8096 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
8097 correct _ecb suffix.
8100 *) Add initial OCSP responder support to ocsp application. The
8101 revocation information is handled using the text based index
8102 use by the ca application. The responder can either handle
8103 requests generated internally, supplied in files (for example
8104 via a CGI script) or using an internal minimal server.
8107 *) Add configuration choices to get zlib compression for TLS.
8110 *) Changes to Kerberos SSL for RFC 2712 compliance:
8111 1. Implemented real KerberosWrapper, instead of just using
8112 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
8113 2. Implemented optional authenticator field of KerberosWrapper.
8115 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
8116 and authenticator structs; see crypto/krb5/.
8118 Generalized Kerberos calls to support multiple Kerberos libraries.
8119 [Vern Staats <staatsvr@asc.hpc.mil>,
8120 Jeffrey Altman <jaltman@columbia.edu>
8121 via Richard Levitte]
8123 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
8124 already does with RSA. testdsa.h now has 'priv_key/pub_key'
8125 values for each of the key sizes rather than having just
8126 parameters (and 'speed' generating keys each time).
8129 *) Speed up EVP routines.
8132 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
8133 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
8134 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
8135 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
8137 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
8138 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
8139 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
8142 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
8144 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
8147 *) Added the OS2-EMX target.
8148 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
8150 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
8151 to support NCONF routines in extension code. New function CONF_set_nconf()
8152 to allow functions which take an NCONF to also handle the old LHASH
8153 structure: this means that the old CONF compatible routines can be
8154 retained (in particular wrt extensions) without having to duplicate the
8155 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
8158 *) Enhance the general user interface with mechanisms for inner control
8159 and with possibilities to have yes/no kind of prompts.
8162 *) Change all calls to low level digest routines in the library and
8163 applications to use EVP. Add missing calls to HMAC_cleanup() and
8164 don't assume HMAC_CTX can be copied using memcpy().
8165 [Verdon Walker <VWalker@novell.com>, Steve Henson]
8167 *) Add the possibility to control engines through control names but with
8168 arbitrary arguments instead of just a string.
8169 Change the key loaders to take a UI_METHOD instead of a callback
8170 function pointer. NOTE: this breaks binary compatibility with earlier
8171 versions of OpenSSL [engine].
8172 Adapt the nCipher code for these new conditions and add a card insertion
8176 *) Enhance the general user interface with mechanisms to better support
8177 dialog box interfaces, application-defined prompts, the possibility
8178 to use defaults (for example default passwords from somewhere else)
8179 and interrupts/cancellations.
8182 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
8183 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
8186 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
8187 tidy up some unnecessarily weird code in 'sk_new()').
8188 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
8190 *) Change the key loading routines for ENGINEs to use the same kind
8191 callback (pem_password_cb) as all other routines that need this
8195 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
8196 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
8197 than this minimum value is recommended.
8200 *) New random seeder for OpenVMS, using the system process statistics
8201 that are easily reachable.
8204 *) Windows apparently can't transparently handle global
8205 variables defined in DLLs. Initialisations such as:
8207 const ASN1_ITEM *it = &ASN1_INTEGER_it;
8209 won't compile. This is used by the any applications that need to
8210 declare their own ASN1 modules. This was fixed by adding the option
8211 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
8212 needed for static libraries under Win32.
8215 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
8216 setting of purpose and trust fields. New X509_STORE trust and
8217 purpose functions and tidy up setting in other SSL functions.
8220 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
8221 structure. These are inherited by X509_STORE_CTX when it is
8222 initialised. This allows various defaults to be set in the
8223 X509_STORE structure (such as flags for CRL checking and custom
8224 purpose or trust settings) for functions which only use X509_STORE_CTX
8225 internally such as S/MIME.
8227 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
8228 trust settings if they are not set in X509_STORE. This allows X509_STORE
8229 purposes and trust (in S/MIME for example) to override any set by default.
8231 Add command line options for CRL checking to smime, s_client and s_server
8235 *) Initial CRL based revocation checking. If the CRL checking flag(s)
8236 are set then the CRL is looked up in the X509_STORE structure and
8237 its validity and signature checked, then if the certificate is found
8238 in the CRL the verify fails with a revoked error.
8240 Various new CRL related callbacks added to X509_STORE_CTX structure.
8242 Command line options added to 'verify' application to support this.
8244 This needs some additional work, such as being able to handle multiple
8245 CRLs with different times, extension based lookup (rather than just
8246 by subject name) and ultimately more complete V2 CRL extension
8250 *) Add a general user interface API (crypto/ui/). This is designed
8251 to replace things like des_read_password and friends (backward
8252 compatibility functions using this new API are provided).
8253 The purpose is to remove prompting functions from the DES code
8254 section as well as provide for prompting through dialog boxes in
8255 a window system and the like.
8258 *) Add "ex_data" support to ENGINE so implementations can add state at a
8259 per-structure level rather than having to store it globally.
8262 *) Make it possible for ENGINE structures to be copied when retrieved by
8263 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
8264 This causes the "original" ENGINE structure to act like a template,
8265 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
8266 operational state can be localised to each ENGINE structure, despite the
8267 fact they all share the same "methods". New ENGINE structures returned in
8268 this case have no functional references and the return value is the single
8269 structural reference. This matches the single structural reference returned
8270 by ENGINE_by_id() normally, when it is incremented on the pre-existing
8274 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
8275 needs to match any other type at all we need to manually clear the
8279 *) Changes to the "openssl engine" utility to include;
8280 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
8281 about an ENGINE's available control commands.
8282 - executing control commands from command line arguments using the
8283 '-pre' and '-post' switches. '-post' is only used if '-t' is
8284 specified and the ENGINE is successfully initialised. The syntax for
8285 the individual commands are colon-separated, for example;
8286 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
8289 *) New dynamic control command support for ENGINEs. ENGINEs can now
8290 declare their own commands (numbers), names (strings), descriptions,
8291 and input types for run-time discovery by calling applications. A
8292 subset of these commands are implicitly classed as "executable"
8293 depending on their input type, and only these can be invoked through
8294 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
8295 can be based on user input, config files, etc). The distinction is
8296 that "executable" commands cannot return anything other than a boolean
8297 result and can only support numeric or string input, whereas some
8298 discoverable commands may only be for direct use through
8299 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
8300 pointers, or other custom uses. The "executable" commands are to
8301 support parameterisations of ENGINE behaviour that can be
8302 unambiguously defined by ENGINEs and used consistently across any
8303 OpenSSL-based application. Commands have been added to all the
8304 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
8305 control over shared-library paths without source code alterations.
8308 *) Changed all ENGINE implementations to dynamically allocate their
8309 ENGINEs rather than declaring them statically. Apart from this being
8310 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
8311 this also allows the implementations to compile without using the
8312 internal engine_int.h header.
8315 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8316 'const' value. Any code that should be able to modify a RAND_METHOD
8317 should already have non-const pointers to it (ie. they should only
8318 modify their own ones).
8321 *) Made a variety of little tweaks to the ENGINE code.
8322 - "atalla" and "ubsec" string definitions were moved from header files
8323 to C code. "nuron" string definitions were placed in variables
8324 rather than hard-coded - allowing parameterisation of these values
8325 later on via ctrl() commands.
8326 - Removed unused "#if 0"'d code.
8327 - Fixed engine list iteration code so it uses ENGINE_free() to release
8328 structural references.
8329 - Constified the RAND_METHOD element of ENGINE structures.
8330 - Constified various get/set functions as appropriate and added
8331 missing functions (including a catch-all ENGINE_cpy that duplicates
8332 all ENGINE values onto a new ENGINE except reference counts/state).
8333 - Removed NULL parameter checks in get/set functions. Setting a method
8334 or function to NULL is a way of cancelling out a previously set
8335 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8336 and doesn't justify the extra error symbols and code.
8337 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8338 flags from engine_int.h to engine.h.
8339 - Changed prototypes for ENGINE handler functions (init(), finish(),
8340 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8343 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8344 to the algorithm using long division. The binary algorithm can be
8345 used only if the modulus is odd. On 32-bit systems, it is faster
8346 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8347 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8348 up to 450 bits. In 64-bit environments, the binary algorithm
8349 appears to be advantageous for much longer moduli; here we use it
8350 for moduli up to 2048 bits.
8353 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8354 could not support the combine flag in choice fields.
8357 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8358 extensions from a certificate request to the certificate.
8361 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8362 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8363 file: this allows the display of the certificate about to be
8364 signed to be customised, to allow certain fields to be included
8365 or excluded and extension details. The old system didn't display
8366 multicharacter strings properly, omitted fields not in the policy
8367 and couldn't display additional details such as extensions.
8370 *) Function EC_POINTs_mul for multiple scalar multiplication
8371 of an arbitrary number of elliptic curve points
8372 \sum scalars[i]*points[i],
8373 optionally including the generator defined for the EC_GROUP:
8374 scalar*generator + \sum scalars[i]*points[i].
8376 EC_POINT_mul is a simple wrapper function for the typical case
8377 that the point list has just one item (besides the optional
8381 *) First EC_METHODs for curves over GF(p):
8383 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8384 operations and provides various method functions that can also
8385 operate with faster implementations of modular arithmetic.
8387 EC_GFp_mont_method() reuses most functions that are part of
8388 EC_GFp_simple_method, but uses Montgomery arithmetic.
8390 [Bodo Moeller; point addition and point doubling
8391 implementation directly derived from source code provided by
8392 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8394 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8395 crypto/ec/ec_lib.c):
8397 Curves are EC_GROUP objects (with an optional group generator)
8398 based on EC_METHODs that are built into the library.
8400 Points are EC_POINT objects based on EC_GROUP objects.
8402 Most of the framework would be able to handle curves over arbitrary
8403 finite fields, but as there are no obvious types for fields other
8404 than GF(p), some functions are limited to that for now.
8407 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8408 that the file contains a complete HTTP response.
8411 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8412 change the def and num file printf format specifier from "%-40sXXX"
8413 to "%-39s XXX". The latter will always guarantee a space after the
8414 field while the former will cause them to run together if the field
8415 is 40 of more characters long.
8418 *) Constify the cipher and digest 'method' functions and structures
8419 and modify related functions to take constant EVP_MD and EVP_CIPHER
8423 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8424 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8427 *) Modify EVP_Digest*() routines so they now return values. Although the
8428 internal software routines can never fail additional hardware versions
8432 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8434 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8435 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8438 ERR_R_NESTED_ASN1_ERROR
8440 ERR_R_MISSING_ASN1_EOS
8441 were 4 .. 9, conflicting with
8442 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8444 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8445 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8447 Add new error code 'ERR_R_INTERNAL_ERROR'.
8450 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8454 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8455 sets the subject name for a new request or supersedes the
8456 subject name in a given request. Formats that can be parsed are
8457 'CN=Some Name, OU=myOU, C=IT'
8459 'CN=Some Name/OU=myOU/C=IT'.
8461 Add options '-batch' and '-verbose' to 'openssl req'.
8462 [Massimiliano Pala <madwolf@hackmasters.net>]
8464 *) Introduce the possibility to access global variables through
8465 functions on platform were that's the best way to handle exporting
8466 global variables in shared libraries. To enable this functionality,
8467 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8468 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8469 is normally done by Configure or something similar).
8471 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8472 in the source file (foo.c) like this:
8474 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8475 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8477 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8478 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8480 OPENSSL_DECLARE_GLOBAL(int,foo);
8481 #define foo OPENSSL_GLOBAL_REF(foo)
8482 OPENSSL_DECLARE_GLOBAL(double,bar);
8483 #define bar OPENSSL_GLOBAL_REF(bar)
8485 The #defines are very important, and therefore so is including the
8486 header file everywhere where the defined globals are used.
8488 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8489 of ASN.1 items, but that structure is a bit different.
8491 The largest change is in util/mkdef.pl which has been enhanced with
8492 better and easier to understand logic to choose which symbols should
8493 go into the Windows .def files as well as a number of fixes and code
8494 cleanup (among others, algorithm keywords are now sorted
8495 lexicographically to avoid constant rewrites).
8498 *) In BN_div() keep a copy of the sign of 'num' before writing the
8499 result to 'rm' because if rm==num the value will be overwritten
8500 and produce the wrong result if 'num' is negative: this caused
8501 problems with BN_mod() and BN_nnmod().
8504 *) Function OCSP_request_verify(). This checks the signature on an
8505 OCSP request and verifies the signer certificate. The signer
8506 certificate is just checked for a generic purpose and OCSP request
8510 *) Add OCSP_check_validity() function to check the validity of OCSP
8511 responses. OCSP responses are prepared in real time and may only
8512 be a few seconds old. Simply checking that the current time lies
8513 between thisUpdate and nextUpdate max reject otherwise valid responses
8514 caused by either OCSP responder or client clock inaccuracy. Instead
8515 we allow thisUpdate and nextUpdate to fall within a certain period of
8516 the current time. The age of the response can also optionally be
8517 checked. Two new options -validity_period and -status_age added to
8521 *) If signature or public key algorithm is unrecognized print out its
8522 OID rather that just UNKNOWN.
8525 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8526 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8527 ID to be generated from the issuer certificate alone which can then be
8528 passed to OCSP_id_issuer_cmp().
8531 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8532 ASN1 modules to export functions returning ASN1_ITEM pointers
8533 instead of the ASN1_ITEM structures themselves. This adds several
8534 new macros which allow the underlying ASN1 function/structure to
8535 be accessed transparently. As a result code should not use ASN1_ITEM
8536 references directly (such as &X509_it) but instead use the relevant
8537 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8538 use of the new ASN1 code on platforms where exporting structures
8539 is problematical (for example in shared libraries) but exporting
8540 functions returning pointers to structures is not.
8543 *) Add support for overriding the generation of SSL/TLS session IDs.
8544 These callbacks can be registered either in an SSL_CTX or per SSL.
8545 The purpose of this is to allow applications to control, if they wish,
8546 the arbitrary values chosen for use as session IDs, particularly as it
8547 can be useful for session caching in multiple-server environments. A
8548 command-line switch for testing this (and any client code that wishes
8549 to use such a feature) has been added to "s_server".
8550 [Geoff Thorpe, Lutz Jaenicke]
8552 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8553 of the form '#if defined(...) || defined(...) || ...' and
8554 '#if !defined(...) && !defined(...) && ...'. This also avoids
8555 the growing number of special cases it was previously handling.
8558 *) Make all configuration macros available for application by making
8559 sure they are available in opensslconf.h, by giving them names starting
8560 with "OPENSSL_" to avoid conflicts with other packages and by making
8561 sure e_os2.h will cover all platform-specific cases together with
8563 Additionally, it is now possible to define configuration/platform-
8564 specific names (called "system identities"). In the C code, these
8565 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8566 macro with the name beginning with "OPENSSL_SYS_", which is determined
8567 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8571 *) New option -set_serial to 'req' and 'x509' this allows the serial
8572 number to use to be specified on the command line. Previously self
8573 signed certificates were hard coded with serial number 0 and the
8574 CA options of 'x509' had to use a serial number in a file which was
8578 *) New options to 'ca' utility to support V2 CRL entry extensions.
8579 Currently CRL reason, invalidity date and hold instruction are
8580 supported. Add new CRL extensions to V3 code and some new objects.
8583 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8584 disable standard block padding (aka PKCS#5 padding) in the EVP
8585 API, which was previously mandatory. This means that the data is
8586 not padded in any way and so the total length much be a multiple
8587 of the block size, otherwise an error occurs.
8590 *) Initial (incomplete) OCSP SSL support.
8593 *) New function OCSP_parse_url(). This splits up a URL into its host,
8594 port and path components: primarily to parse OCSP URLs. New -url
8595 option to ocsp utility.
8598 *) New nonce behavior. The return value of OCSP_check_nonce() now
8599 reflects the various checks performed. Applications can decide
8600 whether to tolerate certain situations such as an absent nonce
8601 in a response when one was present in a request: the ocsp application
8602 just prints out a warning. New function OCSP_add1_basic_nonce()
8603 this is to allow responders to include a nonce in a response even if
8604 the request is nonce-less.
8607 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8608 skipped when using openssl x509 multiple times on a single input file,
8609 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8612 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8613 set string type: to handle setting ASN1_TIME structures. Fix ca
8614 utility to correctly initialize revocation date of CRLs.
8617 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8618 the clients preferred ciphersuites and rather use its own preferences.
8619 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8620 Internet Explorer by ensuring unchanged hash method during stepup.
8621 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8624 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8625 to aes and add a new 'exist' option to print out symbols that don't
8629 *) Additional options to ocsp utility to allow flags to be set and
8630 additional certificates supplied.
8633 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8634 OCSP client a number of certificate to only verify the response
8638 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8639 handle the new API. Currently only ECB, CBC modes supported. Add new
8642 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8643 Encryption Standard (AES) Ciphersuites for Transport Layer
8644 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8645 not enabled by default and were not part of the "ALL" ciphersuite
8646 alias because they were not yet official; they could be
8647 explicitly requested by specifying the "AESdraft" ciphersuite
8648 group alias. In the final release of OpenSSL 0.9.7, the group
8649 alias is called "AES" and is part of "ALL".)
8650 [Ben Laurie, Steve Henson, Bodo Moeller]
8652 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8653 request to response.
8656 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8657 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8658 extract information from a certificate request. OCSP_response_create()
8659 creates a response and optionally adds a basic response structure.
8660 OCSP_basic_add1_status() adds a complete single response to a basic
8661 response and returns the OCSP_SINGLERESP structure just added (to allow
8662 extensions to be included for example). OCSP_basic_add1_cert() adds a
8663 certificate to a basic response and OCSP_basic_sign() signs a basic
8664 response with various flags. New helper functions ASN1_TIME_check()
8665 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8666 (converts ASN1_TIME to GeneralizedTime).
8669 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8670 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8671 structure from a certificate. X509_pubkey_digest() digests the public_key
8672 contents: this is used in various key identifiers.
8675 *) Make sk_sort() tolerate a NULL argument.
8676 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8678 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8679 passed by the function are trusted implicitly. If any of them signed the
8680 response then it is assumed to be valid and is not verified.
8683 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8684 to data. This was previously part of the PKCS7 ASN1 code. This
8685 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8686 [Steve Henson, reported by Kenneth R. Robinette
8687 <support@securenetterm.com>]
8689 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8690 routines: without these tracing memory leaks is very painful.
8691 Fix leaks in PKCS12 and PKCS7 routines.
8694 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8695 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8696 effectively meant GeneralizedTime would never be used. Now it
8697 is initialised to -1 but X509_time_adj() now has to check the value
8698 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8699 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8700 [Steve Henson, reported by Kenneth R. Robinette
8701 <support@securenetterm.com>]
8703 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8704 result in a zero length in the ASN1_INTEGER structure which was
8705 not consistent with the structure when d2i_ASN1_INTEGER() was used
8706 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8707 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8708 where it did not print out a minus for negative ASN1_INTEGER.
8711 *) Add summary printout to ocsp utility. The various functions which
8712 convert status values to strings have been renamed to:
8713 OCSP_response_status_str(), OCSP_cert_status_str() and
8714 OCSP_crl_reason_str() and are no longer static. New options
8715 to verify nonce values and to disable verification. OCSP response
8716 printout format cleaned up.
8719 *) Add additional OCSP certificate checks. These are those specified
8720 in RFC2560. This consists of two separate checks: the CA of the
8721 certificate being checked must either be the OCSP signer certificate
8722 or the issuer of the OCSP signer certificate. In the latter case the
8723 OCSP signer certificate must contain the OCSP signing extended key
8724 usage. This check is performed by attempting to match the OCSP
8725 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8726 in the OCSP_CERTID structures of the response.
8729 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8730 and related routines. This uses the standard OpenSSL certificate
8731 verify routines to perform initial checks (just CA validity) and
8732 to obtain the certificate chain. Then additional checks will be
8733 performed on the chain. Currently the root CA is checked to see
8734 if it is explicitly trusted for OCSP signing. This is used to set
8735 a root CA as a global signing root: that is any certificate that
8736 chains to that CA is an acceptable OCSP signing certificate.
8739 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8740 extensions from a separate configuration file.
8741 As when reading extensions from the main configuration file,
8742 the '-extensions ...' option may be used for specifying the
8744 [Massimiliano Pala <madwolf@comune.modena.it>]
8746 *) New OCSP utility. Allows OCSP requests to be generated or
8747 read. The request can be sent to a responder and the output
8748 parsed, outputted or printed in text form. Not complete yet:
8749 still needs to check the OCSP response validity.
8752 *) New subcommands for 'openssl ca':
8753 'openssl ca -status <serial>' prints the status of the cert with
8754 the given serial number (according to the index file).
8755 'openssl ca -updatedb' updates the expiry status of certificates
8757 [Massimiliano Pala <madwolf@comune.modena.it>]
8759 *) New '-newreq-nodes' command option to CA.pl. This is like
8760 '-newreq', but calls 'openssl req' with the '-nodes' option
8761 so that the resulting key is not encrypted.
8762 [Damien Miller <djm@mindrot.org>]
8764 *) New configuration for the GNU Hurd.
8765 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8767 *) Initial code to implement OCSP basic response verify. This
8768 is currently incomplete. Currently just finds the signer's
8769 certificate and verifies the signature on the response.
8772 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8773 value of OPENSSLDIR. This is available via the new '-d' option
8774 to 'openssl version', and is also included in 'openssl version -a'.
8777 *) Allowing defining memory allocation callbacks that will be given
8778 file name and line number information in additional arguments
8779 (a const char* and an int). The basic functionality remains, as
8780 well as the original possibility to just replace malloc(),
8781 realloc() and free() by functions that do not know about these
8782 additional arguments. To register and find out the current
8783 settings for extended allocation functions, the following
8784 functions are provided:
8786 CRYPTO_set_mem_ex_functions
8787 CRYPTO_set_locked_mem_ex_functions
8788 CRYPTO_get_mem_ex_functions
8789 CRYPTO_get_locked_mem_ex_functions
8791 These work the same way as CRYPTO_set_mem_functions and friends.
8792 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8793 extended allocation function is enabled.
8794 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8795 a conventional allocation function is enabled.
8796 [Richard Levitte, Bodo Moeller]
8798 *) Finish off removing the remaining LHASH function pointer casts.
8799 There should no longer be any prototype-casting required when using
8800 the LHASH abstraction, and any casts that remain are "bugs". See
8801 the callback types and macros at the head of lhash.h for details
8802 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8805 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8806 If /dev/[u]random devices are not available or do not return enough
8807 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8809 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8810 /etc/entropy will be queried once each in this sequence, querying stops
8811 when enough entropy was collected without querying more sockets.
8814 *) Change the Unix RAND_poll() variant to be able to poll several
8815 random devices, as specified by DEVRANDOM, until a sufficient amount
8816 of data has been collected. We spend at most 10 ms on each file
8817 (select timeout) and read in non-blocking mode. DEVRANDOM now
8818 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8819 (previously it was just the string "/dev/urandom"), so on typical
8820 platforms the 10 ms delay will never occur.
8821 Also separate out the Unix variant to its own file, rand_unix.c.
8822 For VMS, there's a currently-empty rand_vms.c.
8825 *) Move OCSP client related routines to ocsp_cl.c. These
8826 provide utility functions which an application needing
8827 to issue a request to an OCSP responder and analyse the
8828 response will typically need: as opposed to those which an
8829 OCSP responder itself would need which will be added later.
8831 OCSP_request_sign() signs an OCSP request with an API similar
8832 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8833 response. OCSP_response_get1_basic() extracts basic response
8834 from response. OCSP_resp_find_status(): finds and extracts status
8835 information from an OCSP_CERTID structure (which will be created
8836 when the request structure is built). These are built from lower
8837 level functions which work on OCSP_SINGLERESP structures but
8838 won't normally be used unless the application wishes to examine
8839 extensions in the OCSP response for example.
8841 Replace nonce routines with a pair of functions.
8842 OCSP_request_add1_nonce() adds a nonce value and optionally
8843 generates a random value. OCSP_check_nonce() checks the
8844 validity of the nonce in an OCSP response.
8847 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8848 This doesn't copy the supplied OCSP_CERTID and avoids the
8849 need to free up the newly created id. Change return type
8850 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8851 This can then be used to add extensions to the request.
8852 Deleted OCSP_request_new(), since most of its functionality
8853 is now in OCSP_REQUEST_new() (and the case insensitive name
8854 clash) apart from the ability to set the request name which
8855 will be added elsewhere.
8858 *) Update OCSP API. Remove obsolete extensions argument from
8859 various functions. Extensions are now handled using the new
8860 OCSP extension code. New simple OCSP HTTP function which
8861 can be used to send requests and parse the response.
8864 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8865 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8866 uses the special reorder version of SET OF to sort the attributes
8867 and reorder them to match the encoded order. This resolves a long
8868 standing problem: a verify on a PKCS7 structure just after signing
8869 it used to fail because the attribute order did not match the
8870 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8871 it uses the received order. This is necessary to tolerate some broken
8872 software that does not order SET OF. This is handled by encoding
8873 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8874 to produce the required SET OF.
8877 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8878 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8879 files to get correct declarations of the ASN.1 item variables.
8882 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8883 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8884 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8885 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8886 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8887 ASN1_ITEM and no wrapper functions.
8890 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8891 replace the old function pointer based I/O routines. Change most of
8892 the *_d2i_bio() and *_d2i_fp() functions to use these.
8895 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8896 lines, recognize more "algorithms" that can be deselected, and make
8897 it complain about algorithm deselection that isn't recognised.
8900 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8901 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8902 to use new functions. Add NO_ASN1_OLD which can be set to remove
8903 some old style ASN1 functions: this can be used to determine if old
8904 code will still work when these eventually go away.
8907 *) New extension functions for OCSP structures, these follow the
8908 same conventions as certificates and CRLs.
8911 *) New function X509V3_add1_i2d(). This automatically encodes and
8912 adds an extension. Its behaviour can be customised with various
8913 flags to append, replace or delete. Various wrappers added for
8914 certificates and CRLs.
8917 *) Fix to avoid calling the underlying ASN1 print routine when
8918 an extension cannot be parsed. Correct a typo in the
8919 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8922 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8923 entries for variables.
8926 *) Add functionality to apps/openssl.c for detecting locking
8927 problems: As the program is single-threaded, all we have
8928 to do is register a locking callback using an array for
8929 storing which locks are currently held by the program.
8932 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8933 SSL_get_ex_data_X509_STORE_idx(), which is used in
8934 ssl_verify_cert_chain() and thus can be called at any time
8935 during TLS/SSL handshakes so that thread-safety is essential.
8936 Unfortunately, the ex_data design is not at all suited
8937 for multi-threaded use, so it probably should be abolished.
8940 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8941 [Broadcom, tweaked and integrated by Geoff Thorpe]
8943 *) Move common extension printing code to new function
8944 X509V3_print_extensions(). Reorganise OCSP print routines and
8945 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8948 *) New function X509_signature_print() to remove duplication in some
8952 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8953 set (this was treated exactly the same as SET OF previously). This
8954 is used to reorder the STACK representing the structure to match the
8955 encoding. This will be used to get round a problem where a PKCS7
8956 structure which was signed could not be verified because the STACK
8957 order did not reflect the encoded order.
8960 *) Reimplement the OCSP ASN1 module using the new code.
8963 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8964 for its ASN1 operations. The old style function pointers still exist
8965 for now but they will eventually go away.
8968 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8969 completely replaces the old ASN1 functionality with a table driven
8970 encoder and decoder which interprets an ASN1_ITEM structure describing
8971 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8972 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8973 has also been converted to the new form.
8976 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8977 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8978 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8979 for negative moduli.
8982 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8983 of not touching the result's sign bit.
8986 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8990 *) Changed the LHASH code to use prototypes for callbacks, and created
8991 macros to declare and implement thin (optionally static) functions
8992 that provide type-safety and avoid function pointer casting for the
8993 type-specific callbacks.
8996 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8998 [Veers Staats <staatsvr@asc.hpc.mil>,
8999 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
9001 *) Reformat the FAQ so the different questions and answers can be divided
9002 in sections depending on the subject.
9005 *) Have the zlib compression code load ZLIB.DLL dynamically under
9009 *) New function BN_mod_sqrt for computing square roots modulo a prime
9010 (using the probabilistic Tonelli-Shanks algorithm unless
9011 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
9012 be handled deterministically).
9013 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9015 *) Make BN_mod_inverse faster by explicitly handling small quotients
9016 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
9017 512 bits], about 30% for larger ones [1024 or 2048 bits].)
9020 *) New function BN_kronecker.
9023 *) Fix BN_gcd so that it works on negative inputs; the result is
9024 positive unless both parameters are zero.
9025 Previously something reasonably close to an infinite loop was
9026 possible because numbers could be growing instead of shrinking
9027 in the implementation of Euclid's algorithm.
9030 *) Fix BN_is_word() and BN_is_one() macros to take into account the
9031 sign of the number in question.
9033 Fix BN_is_word(a,w) to work correctly for w == 0.
9035 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
9036 because its test if the absolute value of 'a' equals 'w'.
9037 Note that BN_abs_is_word does *not* handle w == 0 reliably;
9038 it exists mostly for use in the implementations of BN_is_zero(),
9039 BN_is_one(), and BN_is_word().
9042 *) New function BN_swap.
9045 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
9046 the exponentiation functions are more likely to produce reasonable
9047 results on negative inputs.
9050 *) Change BN_mod_mul so that the result is always non-negative.
9051 Previously, it could be negative if one of the factors was negative;
9052 I don't think anyone really wanted that behaviour.
9055 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
9056 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
9057 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
9058 and add new functions:
9067 BN_mod_lshift1_quick
9071 These functions always generate non-negative results.
9073 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
9074 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
9076 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
9077 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
9078 be reduced modulo m.
9079 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9082 The following entry accidentally appeared in the CHANGES file
9083 distributed with OpenSSL 0.9.7. The modifications described in
9084 it do *not* apply to OpenSSL 0.9.7.
9086 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
9087 was actually never needed) and in BN_mul(). The removal in BN_mul()
9088 required a small change in bn_mul_part_recursive() and the addition
9089 of the functions bn_cmp_part_words(), bn_sub_part_words() and
9090 bn_add_part_words(), which do the same thing as bn_cmp_words(),
9091 bn_sub_words() and bn_add_words() except they take arrays with
9096 *) In 'openssl passwd', verify passwords read from the terminal
9097 unless the '-salt' option is used (which usually means that
9098 verification would just waste user's time since the resulting
9099 hash is going to be compared with some given password hash)
9100 or the new '-noverify' option is used.
9102 This is an incompatible change, but it does not affect
9103 non-interactive use of 'openssl passwd' (passwords on the command
9104 line, '-stdin' option, '-in ...' option) and thus should not
9108 *) Remove all references to RSAref, since there's no more need for it.
9111 *) Make DSO load along a path given through an environment variable
9112 (SHLIB_PATH) with shl_load().
9115 *) Constify the ENGINE code as a result of BIGNUM constification.
9116 Also constify the RSA code and most things related to it. In a
9117 few places, most notable in the depth of the ASN.1 code, ugly
9118 casts back to non-const were required (to be solved at a later
9122 *) Make it so the openssl application has all engines loaded by default.
9125 *) Constify the BIGNUM routines a little more.
9128 *) Add the following functions:
9130 ENGINE_load_cswift()
9132 ENGINE_load_atalla()
9134 ENGINE_load_builtin_engines()
9136 That way, an application can itself choose if external engines that
9137 are built-in in OpenSSL shall ever be used or not. The benefit is
9138 that applications won't have to be linked with libdl or other dso
9139 libraries unless it's really needed.
9141 Changed 'openssl engine' to load all engines on demand.
9142 Changed the engine header files to avoid the duplication of some
9143 declarations (they differed!).
9146 *) 'openssl engine' can now list capabilities.
9149 *) Better error reporting in 'openssl engine'.
9152 *) Never call load_dh_param(NULL) in s_server.
9155 *) Add engine application. It can currently list engines by name and
9156 identity, and test if they are actually available.
9159 *) Improve RPM specification file by forcing symbolic linking and making
9160 sure the installed documentation is also owned by root.root.
9161 [Damien Miller <djm@mindrot.org>]
9163 *) Give the OpenSSL applications more possibilities to make use of
9164 keys (public as well as private) handled by engines.
9167 *) Add OCSP code that comes from CertCo.
9170 *) Add VMS support for the Rijndael code.
9173 *) Added untested support for Nuron crypto accelerator.
9176 *) Add support for external cryptographic devices. This code was
9177 previously distributed separately as the "engine" branch.
9178 [Geoff Thorpe, Richard Levitte]
9180 *) Rework the filename-translation in the DSO code. It is now possible to
9181 have far greater control over how a "name" is turned into a filename
9182 depending on the operating environment and any oddities about the
9183 different shared library filenames on each system.
9186 *) Support threads on FreeBSD-elf in Configure.
9189 *) Fix for SHA1 assembly problem with MASM: it produces
9190 warnings about corrupt line number information when assembling
9191 with debugging information. This is caused by the overlapping
9193 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
9196 NCONF_get_number() has no error checking at all. As a replacement,
9197 NCONF_get_number_e() is defined (_e for "error checking") and is
9198 promoted strongly. The old NCONF_get_number is kept around for
9199 binary backward compatibility.
9200 Make it possible for methods to load from something other than a BIO,
9201 by providing a function pointer that is given a name instead of a BIO.
9202 For example, this could be used to load configuration data from an
9206 *) Fix for non blocking accept BIOs. Added new I/O special reason
9207 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
9208 with non blocking I/O was not possible because no retry code was
9209 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
9213 *) Added the beginnings of Rijndael support.
9216 *) Fix for bug in DirectoryString mask setting. Add support for
9217 X509_NAME_print_ex() in 'req' and X509_print_ex() function
9218 to allow certificate printing to more controllable, additional
9219 'certopt' option to 'x509' to allow new printing options to be
9223 *) Clean old EAY MD5 hack from e_os.h.
9226 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
9228 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
9229 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
9230 [Joe Orton, Steve Henson]
9232 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
9234 *) Fix additional bug revealed by the NISCC test suite:
9236 Stop bug triggering large recursion when presented with
9237 certain ASN.1 tags (CVE-2003-0851)
9240 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
9242 *) Fix various bugs revealed by running the NISCC test suite:
9244 Stop out of bounds reads in the ASN1 code when presented with
9245 invalid tags (CVE-2003-0543 and CVE-2003-0544).
9247 If verify callback ignores invalid public key errors don't try to check
9248 certificate signature with the NULL public key.
9252 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
9253 if the server requested one: as stated in TLS 1.0 and SSL 3.0
9257 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
9258 extra data after the compression methods not only for TLS 1.0
9259 but also for SSL 3.0 (as required by the specification).
9260 [Bodo Moeller; problem pointed out by Matthias Loepfe]
9262 *) Change X509_certificate_type() to mark the key as exported/exportable
9263 when it's 512 *bits* long, not 512 bytes.
9266 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
9268 *) Countermeasure against the Klima-Pokorny-Rosa extension of
9269 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
9270 a protocol version number mismatch like a decryption error
9271 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
9274 *) Turn on RSA blinding by default in the default implementation
9275 to avoid a timing attack. Applications that don't want it can call
9276 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
9277 They would be ill-advised to do so in most cases.
9278 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
9280 *) Change RSA blinding code so that it works when the PRNG is not
9281 seeded (in this case, the secret RSA exponent is abused as
9282 an unpredictable seed -- if it is not unpredictable, there
9283 is no point in blinding anyway). Make RSA blinding thread-safe
9284 by remembering the creator's thread ID in rsa->blinding and
9285 having all other threads use local one-time blinding factors
9286 (this requires more computation than sharing rsa->blinding, but
9287 avoids excessive locking; and if an RSA object is not shared
9288 between threads, blinding will still be very fast).
9291 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
9293 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
9294 via timing by performing a MAC computation even if incorrect
9295 block cipher padding has been found. This is a countermeasure
9296 against active attacks where the attacker has to distinguish
9297 between bad padding and a MAC verification error. (CVE-2003-0078)
9299 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
9300 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
9301 Martin Vuagnoux (EPFL, Ilion)]
9303 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
9305 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
9306 memory from its contents. This is done with a counter that will
9307 place alternating values in each byte. This can be used to solve
9308 two issues: 1) the removal of calls to memset() by highly optimizing
9309 compilers, and 2) cleansing with other values than 0, since those can
9310 be read through on certain media, for example a swap space on disk.
9313 *) Bugfix: client side session caching did not work with external caching,
9314 because the session->cipher setting was not restored when reloading
9315 from the external cache. This problem was masked, when
9316 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9317 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9320 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9321 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9322 [Zeev Lieber <zeev-l@yahoo.com>]
9324 *) Undo an undocumented change introduced in 0.9.6e which caused
9325 repeated calls to OpenSSL_add_all_ciphers() and
9326 OpenSSL_add_all_digests() to be ignored, even after calling
9330 *) Change the default configuration reader to deal with last line not
9331 being properly terminated.
9334 *) Change X509_NAME_cmp() so it applies the special rules on handling
9335 DN values that are of type PrintableString, as well as RDNs of type
9336 emailAddress where the value has the type ia5String.
9337 [stefank@valicert.com via Richard Levitte]
9339 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9340 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9341 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9342 the bitwise-OR of the two for use by the majority of applications
9343 wanting this behaviour, and update the docs. The documented
9344 behaviour and actual behaviour were inconsistent and had been
9345 changing anyway, so this is more a bug-fix than a behavioural
9347 [Geoff Thorpe, diagnosed by Nadav Har'El]
9349 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9350 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9353 *) Fix initialization code race conditions in
9354 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9355 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9356 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9357 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9358 ssl2_get_cipher_by_char(),
9359 ssl3_get_cipher_by_char().
9360 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9362 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9363 the cached sessions are flushed, as the remove_cb() might use ex_data
9364 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9365 (see [openssl.org #212]).
9366 [Geoff Thorpe, Lutz Jaenicke]
9368 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9369 length, instead of the encoding length to d2i_ASN1_OBJECT.
9372 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9374 *) [In 0.9.6g-engine release:]
9375 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9376 [Lynn Gazis <lgazis@rainbow.com>]
9378 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9380 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9381 and get fix the header length calculation.
9382 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9383 Alon Kantor <alonk@checkpoint.com> (and others),
9386 *) Use proper error handling instead of 'assertions' in buffer
9387 overflow checks added in 0.9.6e. This prevents DoS (the
9388 assertions could call abort()).
9389 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9391 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9393 *) Add various sanity checks to asn1_get_length() to reject
9394 the ASN1 length bytes if they exceed sizeof(long), will appear
9395 negative or the content length exceeds the length of the
9397 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9399 *) Fix cipher selection routines: ciphers without encryption had no flags
9400 for the cipher strength set and where therefore not handled correctly
9401 by the selection routines (PR #130).
9404 *) Fix EVP_dsa_sha macro.
9408 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9409 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9410 that was added in OpenSSL 0.9.6d.
9412 As the countermeasure turned out to be incompatible with some
9413 broken SSL implementations, the new option is part of SSL_OP_ALL.
9414 SSL_OP_ALL is usually employed when compatibility with weird SSL
9415 implementations is desired (e.g. '-bugs' option to 's_client' and
9416 's_server'), so the new option is automatically set in many
9420 *) Changes in security patch:
9422 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9423 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9424 Air Force Materiel Command, USAF, under agreement number
9427 *) Add various sanity checks to asn1_get_length() to reject
9428 the ASN1 length bytes if they exceed sizeof(long), will appear
9429 negative or the content length exceeds the length of the
9430 supplied buffer. (CVE-2002-0659)
9431 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9433 *) Assertions for various potential buffer overflows, not known to
9435 [Ben Laurie (CHATS)]
9437 *) Various temporary buffers to hold ASCII versions of integers were
9438 too small for 64 bit platforms. (CVE-2002-0655)
9439 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9441 *) Remote buffer overflow in SSL3 protocol - an attacker could
9442 supply an oversized session ID to a client. (CVE-2002-0656)
9443 [Ben Laurie (CHATS)]
9445 *) Remote buffer overflow in SSL2 protocol - an attacker could
9446 supply an oversized client master key. (CVE-2002-0656)
9447 [Ben Laurie (CHATS)]
9449 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9451 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9452 encoded as NULL) with id-dsa-with-sha1.
9453 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9455 *) Check various X509_...() return values in apps/req.c.
9456 [Nils Larsch <nla@trustcenter.de>]
9458 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9459 an end-of-file condition would erroneously be flagged, when the CRLF
9460 was just at the end of a processed block. The bug was discovered when
9461 processing data through a buffering memory BIO handing the data to a
9462 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9463 <ptsekov@syntrex.com> and Nedelcho Stanev.
9466 *) Implement a countermeasure against a vulnerability recently found
9467 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9468 before application data chunks to avoid the use of known IVs
9469 with data potentially chosen by the attacker.
9472 *) Fix length checks in ssl3_get_client_hello().
9475 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9476 to prevent ssl3_read_internal() from incorrectly assuming that
9477 ssl3_read_bytes() found application data while handshake
9478 processing was enabled when in fact s->s3->in_read_app_data was
9479 merely automatically cleared during the initial handshake.
9480 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9482 *) Fix object definitions for Private and Enterprise: they were not
9483 recognized in their shortname (=lowercase) representation. Extend
9484 obj_dat.pl to issue an error when using undefined keywords instead
9485 of silently ignoring the problem (Svenning Sorensen
9486 <sss@sss.dnsalias.net>).
9489 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9490 generators, i.e. generators other than 2 and 5. (Previously, the
9491 code did not properly initialise the 'add' and 'rem' values to
9492 BN_generate_prime().)
9494 In the new general case, we do not insist that 'generator' is
9495 actually a primitive root: This requirement is rather pointless;
9496 a generator of the order-q subgroup is just as good, if not
9500 *) Map new X509 verification errors to alerts. Discovered and submitted by
9501 Tom Wu <tom@arcot.com>.
9504 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9505 returning non-zero before the data has been completely received
9506 when using non-blocking I/O.
9507 [Bodo Moeller; problem pointed out by John Hughes]
9509 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9510 [Ben Laurie, Lutz Jaenicke]
9512 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9513 Yoram Zahavi <YoramZ@gilian.com>).
9516 *) Add information about CygWin 1.3 and on, and preserve proper
9517 configuration for the versions before that.
9518 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9520 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9521 check whether we deal with a copy of a session and do not delete from
9522 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9523 <izhar@checkpoint.com>.
9526 *) Do not store session data into the internal session cache, if it
9527 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9528 flag is set). Proposed by Aslam <aslam@funk.com>.
9531 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9535 *) [In 0.9.6d-engine release:]
9536 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9537 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9539 *) Add the configuration target linux-s390x.
9540 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9542 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9543 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9544 variable as an indication that a ClientHello message has been
9545 received. As the flag value will be lost between multiple
9546 invocations of ssl3_accept when using non-blocking I/O, the
9547 function may not be aware that a handshake has actually taken
9548 place, thus preventing a new session from being added to the
9551 To avoid this problem, we now set s->new_session to 2 instead of
9552 using a local variable.
9553 [Lutz Jaenicke, Bodo Moeller]
9555 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9556 if the SSL_R_LENGTH_MISMATCH error is detected.
9557 [Geoff Thorpe, Bodo Moeller]
9559 *) New 'shared_ldflag' column in Configure platform table.
9562 *) Fix EVP_CIPHER_mode macro.
9563 ["Dan S. Camper" <dan@bti.net>]
9565 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9566 type, we must throw them away by setting rr->length to 0.
9567 [D P Chang <dpc@qualys.com>]
9569 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9571 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9572 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9573 worked incorrectly for those cases where range = 10..._2 and
9574 3*range is two bits longer than range.)
9577 *) Only add signing time to PKCS7 structures if it is not already
9581 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9582 OBJ_ld_ce should be OBJ_id_ce.
9583 Also some ip-pda OIDs in crypto/objects/objects.txt were
9584 incorrect (cf. RFC 3039).
9585 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9587 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9588 returns early because it has nothing to do.
9589 [Andy Schneider <andy.schneider@bjss.co.uk>]
9591 *) [In 0.9.6c-engine release:]
9592 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9593 [Andy Schneider <andy.schneider@bjss.co.uk>]
9595 *) [In 0.9.6c-engine release:]
9596 Add support for Cryptographic Appliance's keyserver technology.
9597 (Use engine 'keyclient')
9598 [Cryptographic Appliances and Geoff Thorpe]
9600 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9601 is called via tools/c89.sh because arguments have to be
9602 rearranged (all '-L' options must appear before the first object
9604 [Richard Shapiro <rshapiro@abinitio.com>]
9606 *) [In 0.9.6c-engine release:]
9607 Add support for Broadcom crypto accelerator cards, backported
9609 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9611 *) [In 0.9.6c-engine release:]
9612 Add support for SureWare crypto accelerator cards from
9613 Baltimore Technologies. (Use engine 'sureware')
9614 [Baltimore Technologies and Mark Cox]
9616 *) [In 0.9.6c-engine release:]
9617 Add support for crypto accelerator cards from Accelerated
9618 Encryption Processing, www.aep.ie. (Use engine 'aep')
9619 [AEP Inc. and Mark Cox]
9621 *) Add a configuration entry for gcc on UnixWare.
9622 [Gary Benson <gbenson@redhat.com>]
9624 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9625 messages are stored in a single piece (fixed-length part and
9626 variable-length part combined) and fix various bugs found on the way.
9629 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9630 instead. BIO_gethostbyname() does not know what timeouts are
9631 appropriate, so entries would stay in cache even when they have
9633 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9635 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9636 faced with a pathologically small ClientHello fragment that does
9637 not contain client_version: Instead of aborting with an error,
9638 simply choose the highest available protocol version (i.e.,
9639 TLS 1.0 unless it is disabled). In practice, ClientHello
9640 messages are never sent like this, but this change gives us
9641 strictly correct behaviour at least for TLS.
9644 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9645 never resets s->method to s->ctx->method when called from within
9646 one of the SSL handshake functions.
9647 [Bodo Moeller; problem pointed out by Niko Baric]
9649 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9650 (sent using the client's version number) if client_version is
9651 smaller than the protocol version in use. Also change
9652 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9653 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9654 the client will at least see that alert.
9657 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9661 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9662 client receives HelloRequest while in a handshake.
9663 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9665 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9666 should end in 'break', not 'goto end' which circumvents various
9667 cleanups done in state SSL_ST_OK. But session related stuff
9668 must be disabled for SSL_ST_OK in the case that we just sent a
9671 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9672 before just sending a HelloRequest.
9673 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9675 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9676 reveal whether illegal block cipher padding was found or a MAC
9677 verification error occurred. (Neither SSLerr() codes nor alerts
9678 are directly visible to potential attackers, but the information
9679 may leak via logfiles.)
9681 Similar changes are not required for the SSL 2.0 implementation
9682 because the number of padding bytes is sent in clear for SSL 2.0,
9683 and the extra bytes are just ignored. However ssl/s2_pkt.c
9684 failed to verify that the purported number of padding bytes is in
9688 *) Add OpenUNIX-8 support including shared libraries
9689 (Boyd Lynn Gerber <gerberb@zenez.com>).
9692 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9693 'wristwatch attack' using huge encoding parameters (cf.
9694 James H. Manger's CRYPTO 2001 paper). Note that the
9695 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9696 encoding parameters and hence was not vulnerable.
9699 *) BN_sqr() bug fix.
9700 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9702 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9703 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9704 followed by modular reduction.
9705 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9707 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9708 equivalent based on BN_pseudo_rand() instead of BN_rand().
9711 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9712 This function was broken, as the check for a new client hello message
9713 to handle SGC did not allow these large messages.
9714 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9717 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9720 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9721 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9724 *) Rework the configuration and shared library support for Tru64 Unix.
9725 The configuration part makes use of modern compiler features and
9726 still retains old compiler behavior for those that run older versions
9727 of the OS. The shared library support part includes a variant that
9728 uses the RPATH feature, and is available through the special
9729 configuration target "alpha-cc-rpath", which will never be selected
9731 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9733 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9734 with the same message size as in ssl3_get_certificate_request().
9735 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9736 messages might inadvertently be reject as too long.
9737 [Petr Lampa <lampa@fee.vutbr.cz>]
9739 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9742 *) Modified SSL library such that the verify_callback that has been set
9743 specifically for an SSL object with SSL_set_verify() is actually being
9744 used. Before the change, a verify_callback set with this function was
9745 ignored and the verify_callback() set in the SSL_CTX at the time of
9746 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9747 to allow the necessary settings.
9750 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9751 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9752 done automatically (in contradiction to the requirements of the C
9753 standard). This made problems when used from OpenSSH.
9756 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9757 dh->length and always used
9759 BN_rand_range(priv_key, dh->p).
9761 BN_rand_range() is not necessary for Diffie-Hellman, and this
9762 specific range makes Diffie-Hellman unnecessarily inefficient if
9763 dh->length (recommended exponent length) is much smaller than the
9764 length of dh->p. We could use BN_rand_range() if the order of
9765 the subgroup was stored in the DH structure, but we only have
9770 BN_rand(priv_key, l, ...)
9772 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9778 RSA_eay_public_encrypt
9779 RSA_eay_private_decrypt
9780 RSA_eay_private_encrypt (signing)
9781 RSA_eay_public_decrypt (signature verification)
9783 (default implementations for RSA_public_encrypt,
9784 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9785 always reject numbers >= n.
9788 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9789 to synchronize access to 'locking_thread'. This is necessary on
9790 systems where access to 'locking_thread' (an 'unsigned long'
9791 variable) is not atomic.
9794 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9795 *before* setting the 'crypto_lock_rand' flag. The previous code had
9796 a race condition if 0 is a valid thread ID.
9797 [Travis Vitek <vitek@roguewave.com>]
9799 *) Add support for shared libraries under Irix.
9800 [Albert Chin-A-Young <china@thewrittenword.com>]
9802 *) Add configuration option to build on Linux on both big-endian and
9804 [Ralf Baechle <ralf@uni-koblenz.de>]
9806 *) Add the possibility to create shared libraries on HP-UX.
9809 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9811 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9812 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9813 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9814 PRNG state recovery was possible based on the output of
9815 one PRNG request appropriately sized to gain knowledge on
9816 'md' followed by enough consecutive 1-byte PRNG requests
9817 to traverse all of 'state'.
9819 1. When updating 'md_local' (the current thread's copy of 'md')
9820 during PRNG output generation, hash all of the previous
9821 'md_local' value, not just the half used for PRNG output.
9823 2. Make the number of bytes from 'state' included into the hash
9824 independent from the number of PRNG bytes requested.
9826 The first measure alone would be sufficient to avoid
9827 Markku-Juhani's attack. (Actually it had never occurred
9828 to me that the half of 'md_local' used for chaining was the
9829 half from which PRNG output bytes were taken -- I had always
9830 assumed that the secret half would be used.) The second
9831 measure makes sure that additional data from 'state' is never
9832 mixed into 'md_local' in small portions; this heuristically
9833 further strengthens the PRNG.
9836 *) Fix crypto/bn/asm/mips3.s.
9839 *) When only the key is given to "enc", the IV is undefined. Print out
9840 an error message in this case.
9843 *) Handle special case when X509_NAME is empty in X509 printing routines.
9846 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9847 positive and less than q.
9850 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9851 used: it isn't thread safe and the add_lock_callback should handle
9853 [Paul Rose <Paul.Rose@bridge.com>]
9855 *) Verify that incoming data obeys the block size in
9856 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9860 [Ulf Möller, Bodo Möller]
9862 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9863 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9864 when fixing the server behaviour for backwards-compatible 'client
9865 hello' messages. (Note that the attack is impractical against
9866 SSL 3.0 and TLS 1.0 anyway because length and version checking
9867 means that the probability of guessing a valid ciphertext is
9868 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9871 Before 0.9.5, the countermeasure (hide the error by generating a
9872 random 'decryption result') did not work properly because
9873 ERR_clear_error() was missing, meaning that SSL_get_error() would
9874 detect the supposedly ignored error.
9876 Both problems are now fixed.
9879 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9880 (previously it was 1024).
9883 *) Fix for compatibility mode trust settings: ignore trust settings
9884 unless some valid trust or reject settings are present.
9887 *) Fix for blowfish EVP: its a variable length cipher.
9890 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9891 parameters in DSA public key structures and return an error in the
9892 DSA routines if parameters are absent.
9895 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9896 in the current directory if neither $RANDFILE nor $HOME was set.
9897 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9898 caused some confusion to Windows users who haven't defined $HOME.
9899 Thus RAND_file_name() is changed again: e_os.h can define a
9900 DEFAULT_HOME, which will be used if $HOME is not set.
9901 For Windows, we use "C:"; on other platforms, we still require
9902 environment variables.
9904 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9905 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9906 having multiple threads call RAND_poll() concurrently.
9909 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9910 combination of a flag and a thread ID variable.
9911 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9912 flag), *other* threads can enter ssleay_add_bytes without obeying
9913 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9914 that they do not hold after the first thread unsets add_do_not_lock).
9917 *) Change bctest again: '-x' expressions are not available in all
9921 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9923 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9924 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9926 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9927 the default extension for executables, if any. Also, make the perl
9928 scripts that use symlink() to test if it really exists and use "cp"
9929 if it doesn't. All this made OpenSSL compilable and installable in
9933 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9934 If SEQUENCE is length is indefinite just set c->slen to the total
9935 amount of data available.
9936 [Steve Henson, reported by shige@FreeBSD.org]
9937 [This change does not apply to 0.9.7.]
9939 *) Change bctest to avoid here-documents inside command substitution
9940 (workaround for FreeBSD /bin/sh bug).
9941 For compatibility with Ultrix, avoid shell functions (introduced
9942 in the bctest version that searches along $PATH).
9945 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9946 with des_encrypt() defined on some operating systems, like Solaris
9950 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9951 On the Importance of Eliminating Errors in Cryptographic
9952 Computations, J. Cryptology 14 (2001) 2, 101-119,
9953 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9956 *) MIPS assembler BIGNUM division bug fix.
9959 *) Disabled incorrect Alpha assembler code.
9962 *) Fix PKCS#7 decode routines so they correctly update the length
9963 after reading an EOC for the EXPLICIT tag.
9965 [This change does not apply to 0.9.7.]
9967 *) Fix bug in PKCS#12 key generation routines. This was triggered
9968 if a 3DES key was generated with a 0 initial byte. Include
9969 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9970 (but broken) behaviour.
9973 *) Enhance bctest to search for a working bc along $PATH and print
9975 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9977 *) Fix memory leaks in err.c: free err_data string if necessary;
9978 don't write to the wrong index in ERR_set_error_data.
9981 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9985 *) Replace rdtsc with _emit statements for VC++ version 5.
9986 [Jeremy Cooper <jeremy@baymoo.org>]
9988 *) Make it possible to reuse SSLv2 sessions.
9991 *) In copy_email() check for >= 0 as a return value for
9992 X509_NAME_get_index_by_NID() since 0 is a valid index.
9993 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9995 *) Avoid coredump with unsupported or invalid public keys by checking if
9996 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9997 PKCS7_verify() fails with non detached data.
10000 *) Don't use getenv in library functions when run as setuid/setgid.
10001 New function OPENSSL_issetugid().
10004 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
10005 due to incorrect handling of multi-threading:
10007 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
10009 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
10011 3. Count how many times MemCheck_off() has been called so that
10012 nested use can be treated correctly. This also avoids
10013 inband-signalling in the previous code (which relied on the
10014 assumption that thread ID 0 is impossible).
10017 *) Add "-rand" option also to s_client and s_server.
10020 *) Fix CPU detection on Irix 6.x.
10021 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
10022 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10024 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
10027 [This change does not apply to 0.9.7.]
10029 *) Use the cached encoding of an X509_NAME structure rather than
10030 copying it. This is apparently the reason for the libsafe "errors"
10031 but the code is actually correct.
10034 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
10035 Bleichenbacher's DSA attack.
10036 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
10037 to be set and top=0 forces the highest bit to be set; top=-1 is new
10038 and leaves the highest bit random.
10039 [Ulf Moeller, Bodo Moeller]
10041 *) In the NCONF_...-based implementations for CONF_... queries
10042 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
10043 a temporary CONF structure with the data component set to NULL
10044 (which gives segmentation faults in lh_retrieve).
10045 Instead, use NULL for the CONF pointer in CONF_get_string and
10046 CONF_get_number (which may use environment variables) and directly
10047 return NULL from CONF_get_section.
10050 *) Fix potential buffer overrun for EBCDIC.
10053 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
10054 keyUsage if basicConstraints absent for a CA.
10057 *) Make SMIME_write_PKCS7() write mail header values with a format that
10058 is more generally accepted (no spaces before the semicolon), since
10059 some programs can't parse those values properly otherwise. Also make
10060 sure BIO's that break lines after each write do not create invalid
10064 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
10065 macros previously used would not encode an empty SEQUENCE OF
10066 and break the signature.
10068 [This change does not apply to 0.9.7.]
10070 *) Zero the premaster secret after deriving the master secret in
10074 *) Add some EVP_add_digest_alias registrations (as found in
10075 OpenSSL_add_all_digests()) to SSL_library_init()
10076 aka OpenSSL_add_ssl_algorithms(). This provides improved
10077 compatibility with peers using X.509 certificates
10078 with unconventional AlgorithmIdentifier OIDs.
10081 *) Fix for Irix with NO_ASM.
10082 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10084 *) ./config script fixes.
10085 [Ulf Moeller, Richard Levitte]
10087 *) Fix 'openssl passwd -1'.
10090 *) Change PKCS12_key_gen_asc() so it can cope with non null
10091 terminated strings whose length is passed in the passlen
10092 parameter, for example from PEM callbacks. This was done
10093 by adding an extra length parameter to asc2uni().
10094 [Steve Henson, reported by <oddissey@samsung.co.kr>]
10096 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
10097 call failed, free the DSA structure.
10100 *) Fix to uni2asc() to cope with zero length Unicode strings.
10101 These are present in some PKCS#12 files.
10104 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
10105 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
10106 when writing a 32767 byte record.
10107 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
10109 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
10110 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
10112 (RSA objects have a reference count access to which is protected
10113 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
10114 so they are meant to be shared between threads.)
10115 [Bodo Moeller, Geoff Thorpe; original patch submitted by
10116 "Reddie, Steven" <Steven.Reddie@ca.com>]
10118 *) Fix a deadlock in CRYPTO_mem_leaks().
10121 *) Use better test patterns in bntest.
10124 *) rand_win.c fix for Borland C.
10127 *) BN_rshift bugfix for n == 0.
10130 *) Add a 'bctest' script that checks for some known 'bc' bugs
10131 so that 'make test' does not abort just because 'bc' is broken.
10134 *) Store verify_result within SSL_SESSION also for client side to
10135 avoid potential security hole. (Re-used sessions on the client side
10136 always resulted in verify_result==X509_V_OK, not using the original
10137 result of the server certificate verification.)
10140 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
10141 SSL3_RT_APPLICATION_DATA, return 0.
10142 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
10146 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
10147 releases, have been re-implemented by renaming the previous
10148 implementations of ssl2_read and ssl3_read to ssl2_read_internal
10149 and ssl3_read_internal, respectively, and adding 'peek' parameters
10150 to them. The new ssl[23]_{read,peek} functions are calls to
10151 ssl[23]_read_internal with the 'peek' flag set appropriately.
10152 A 'peek' parameter has also been added to ssl3_read_bytes, which
10153 does the actual work for ssl3_read_internal.
10156 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
10157 the method-specific "init()" handler. Also clean up ex_data after
10158 calling the method-specific "finish()" handler. Previously, this was
10159 happening the other way round.
10162 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
10163 The previous value, 12, was not always sufficient for BN_mod_exp().
10166 *) Make sure that shared libraries get the internal name engine with
10167 the full version number and not just 0. This should mark the
10168 shared libraries as not backward compatible. Of course, this should
10169 be changed again when we can guarantee backward binary compatibility.
10172 *) Fix typo in get_cert_by_subject() in by_dir.c
10173 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
10175 *) Rework the system to generate shared libraries:
10177 - Make note of the expected extension for the shared libraries and
10178 if there is a need for symbolic links from for example libcrypto.so.0
10179 to libcrypto.so.0.9.7. There is extended info in Configure for
10182 - Make as few rebuilds of the shared libraries as possible.
10184 - Still avoid linking the OpenSSL programs with the shared libraries.
10186 - When installing, install the shared libraries separately from the
10190 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
10192 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
10193 and not in SSL_clear because the latter is also used by the
10194 accept/connect functions; previously, the settings made by
10195 SSL_set_read_ahead would be lost during the handshake.
10196 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
10198 *) Correct util/mkdef.pl to be selective about disabled algorithms.
10199 Previously, it would create entries for disabled algorithms no
10203 *) Added several new manual pages for SSL_* function.
10206 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
10208 *) In ssl23_get_client_hello, generate an error message when faced
10209 with an initial SSL 3.0/TLS record that is too small to contain the
10210 first two bytes of the ClientHello message, i.e. client_version.
10211 (Note that this is a pathologic case that probably has never happened
10212 in real life.) The previous approach was to use the version number
10213 from the record header as a substitute; but our protocol choice
10214 should not depend on that one because it is not authenticated
10215 by the Finished messages.
10218 *) More robust randomness gathering functions for Windows.
10219 [Jeffrey Altman <jaltman@columbia.edu>]
10221 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
10222 not set then we don't setup the error code for issuer check errors
10223 to avoid possibly overwriting other errors which the callback does
10224 handle. If an application does set the flag then we assume it knows
10225 what it is doing and can handle the new informational codes
10229 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
10230 a general "ANY" type, as such it should be able to decode anything
10231 including tagged types. However it didn't check the class so it would
10232 wrongly interpret tagged types in the same way as their universal
10233 counterpart and unknown types were just rejected. Changed so that the
10234 tagged and unknown types are handled in the same way as a SEQUENCE:
10235 that is the encoding is stored intact. There is also a new type
10236 "V_ASN1_OTHER" which is used when the class is not universal, in this
10237 case we have no idea what the actual type is so we just lump them all
10241 *) On VMS, stdout may very well lead to a file that is written to
10242 in a record-oriented fashion. That means that every write() will
10243 write a separate record, which will be read separately by the
10244 programs trying to read from it. This can be very confusing.
10246 The solution is to put a BIO filter in the way that will buffer
10247 text until a linefeed is reached, and then write everything a
10248 line at a time, so every record written will be an actual line,
10249 not chunks of lines and not (usually doesn't happen, but I've
10250 seen it once) several lines in one record. BIO_f_linebuffer() is
10253 Currently, it's a VMS-only method, because that's where it has
10254 been tested well enough.
10257 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
10258 it can return incorrect results.
10259 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
10260 but it was in 0.9.6-beta[12].)
10263 *) Disable the check for content being present when verifying detached
10264 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
10265 include zero length content when signing messages.
10268 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
10269 BIO_ctrl (for BIO pairs).
10272 *) Add DSO method for VMS.
10275 *) Bug fix: Montgomery multiplication could produce results with the
10279 *) Add RPM specification openssl.spec and modify it to build three
10280 packages. The default package contains applications, application
10281 documentation and run-time libraries. The devel package contains
10282 include files, static libraries and function documentation. The
10283 doc package contains the contents of the doc directory. The original
10284 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
10287 *) Add a large number of documentation files for many SSL routines.
10288 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10290 *) Add a configuration entry for Sony News 4.
10291 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
10293 *) Don't set the two most significant bits to one when generating a
10294 random number < q in the DSA library.
10297 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
10298 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
10299 the underlying transport is blocking) if a handshake took place.
10300 (The default behaviour is needed by applications such as s_client
10301 and s_server that use select() to determine when to use SSL_read;
10302 but for applications that know in advance when to expect data, it
10303 just makes things more complicated.)
10306 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
10310 *) Add a few more EBCDIC conditionals that make `req' and `x509'
10311 work better on such systems.
10312 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10314 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10315 Update PKCS12_parse() so it copies the friendlyName and the
10316 keyid to the certificates aux info.
10319 *) Fix bug in PKCS7_verify() which caused an infinite loop
10320 if there was more than one signature.
10321 [Sven Uszpelkat <su@celocom.de>]
10323 *) Major change in util/mkdef.pl to include extra information
10324 about each symbol, as well as presenting variables as well
10325 as functions. This change means that there's n more need
10326 to rebuild the .num files when some algorithms are excluded.
10329 *) Allow the verify time to be set by an application,
10330 rather than always using the current time.
10333 *) Phase 2 verify code reorganisation. The certificate
10334 verify code now looks up an issuer certificate by a
10335 number of criteria: subject name, authority key id
10336 and key usage. It also verifies self signed certificates
10337 by the same criteria. The main comparison function is
10338 X509_check_issued() which performs these checks.
10340 Lot of changes were necessary in order to support this
10341 without completely rewriting the lookup code.
10343 Authority and subject key identifier are now cached.
10345 The LHASH 'certs' is X509_STORE has now been replaced
10346 by a STACK_OF(X509_OBJECT). This is mainly because an
10347 LHASH can't store or retrieve multiple objects with
10348 the same hash value.
10350 As a result various functions (which were all internal
10351 use only) have changed to handle the new X509_STORE
10352 structure. This will break anything that messed round
10353 with X509_STORE internally.
10355 The functions X509_STORE_add_cert() now checks for an
10356 exact match, rather than just subject name.
10358 The X509_STORE API doesn't directly support the retrieval
10359 of multiple certificates matching a given criteria, however
10360 this can be worked round by performing a lookup first
10361 (which will fill the cache with candidate certificates)
10362 and then examining the cache for matches. This is probably
10363 the best we can do without throwing out X509_LOOKUP
10364 entirely (maybe later...).
10366 The X509_VERIFY_CTX structure has been enhanced considerably.
10368 All certificate lookup operations now go via a get_issuer()
10369 callback. Although this currently uses an X509_STORE it
10370 can be replaced by custom lookups. This is a simple way
10371 to bypass the X509_STORE hackery necessary to make this
10372 work and makes it possible to use more efficient techniques
10373 in future. A very simple version which uses a simple
10374 STACK for its trusted certificate store is also provided
10375 using X509_STORE_CTX_trusted_stack().
10377 The verify_cb() and verify() callbacks now have equivalents
10378 in the X509_STORE_CTX structure.
10380 X509_STORE_CTX also has a 'flags' field which can be used
10381 to customise the verify behaviour.
10384 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10385 excludes S/MIME capabilities.
10388 *) When a certificate request is read in keep a copy of the
10389 original encoding of the signed data and use it when outputting
10390 again. Signatures then use the original encoding rather than
10391 a decoded, encoded version which may cause problems if the
10392 request is improperly encoded.
10395 *) For consistency with other BIO_puts implementations, call
10396 buffer_write(b, ...) directly in buffer_puts instead of calling
10399 In BIO_puts, increment b->num_write as in BIO_write.
10400 [Peter.Sylvester@EdelWeb.fr]
10402 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10403 BN_zero, we may not return a BIGNUM with an array consisting of
10404 words set to zero.)
10407 *) Avoid calling abort() from within the library when problems are
10408 detected, except if preprocessor symbols have been defined
10409 (such as REF_CHECK, BN_DEBUG etc.).
10412 *) New openssl application 'rsautl'. This utility can be
10413 used for low level RSA operations. DER public key
10414 BIO/fp routines also added.
10417 *) New Configure entry and patches for compiling on QNX 4.
10418 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10420 *) A demo state-machine implementation was sponsored by
10421 Nuron (http://www.nuron.com/) and is now available in
10422 demos/state_machine.
10425 *) New options added to the 'dgst' utility for signature
10426 generation and verification.
10429 *) Unrecognized PKCS#7 content types are now handled via a
10430 catch all ASN1_TYPE structure. This allows unsupported
10431 types to be stored as a "blob" and an application can
10432 encode and decode it manually.
10435 *) Fix various signed/unsigned issues to make a_strex.c
10436 compile under VC++.
10437 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10439 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10440 length if passed a buffer. ASN1_INTEGER_to_BN failed
10441 if passed a NULL BN and its argument was negative.
10442 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10444 *) Modification to PKCS#7 encoding routines to output definite
10445 length encoding. Since currently the whole structures are in
10446 memory there's not real point in using indefinite length
10447 constructed encoding. However if OpenSSL is compiled with
10448 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10451 *) Added BIO_vprintf() and BIO_vsnprintf().
10454 *) Added more prefixes to parse for in the strings written
10455 through a logging bio, to cover all the levels that are available
10456 through syslog. The prefixes are now:
10458 PANIC, EMERG, EMR => LOG_EMERG
10459 ALERT, ALR => LOG_ALERT
10460 CRIT, CRI => LOG_CRIT
10461 ERROR, ERR => LOG_ERR
10462 WARNING, WARN, WAR => LOG_WARNING
10463 NOTICE, NOTE, NOT => LOG_NOTICE
10464 INFO, INF => LOG_INFO
10465 DEBUG, DBG => LOG_DEBUG
10467 and as before, if none of those prefixes are present at the
10468 beginning of the string, LOG_ERR is chosen.
10470 On Win32, the LOG_* levels are mapped according to this:
10472 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10473 LOG_WARNING => EVENTLOG_WARNING_TYPE
10474 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10478 *) Made it possible to reconfigure with just the configuration
10479 argument "reconf" or "reconfigure". The command line arguments
10480 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10481 and are retrieved from there when reconfiguring.
10484 *) MD4 implemented.
10485 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10487 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10490 *) The obj_dat.pl script was messing up the sorting of object
10491 names. The reason was that it compared the quoted version
10492 of strings as a result "OCSP" > "OCSP Signing" because
10493 " > SPACE. Changed script to store unquoted versions of
10494 names and add quotes on output. It was also omitting some
10495 names from the lookup table if they were given a default
10496 value (that is if SN is missing it is given the same
10497 value as LN and vice versa), these are now added on the
10498 grounds that if an object has a name we should be able to
10499 look it up. Finally added warning output when duplicate
10500 short or long names are found.
10503 *) Changes needed for Tandem NSK.
10504 [Scott Uroff <scott@xypro.com>]
10506 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10507 RSA_padding_check_SSLv23(), special padding was never detected
10508 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10509 version rollback attacks was not effective.
10511 In s23_clnt.c, don't use special rollback-attack detection padding
10512 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10513 client; similarly, in s23_srvr.c, don't do the rollback check if
10514 SSL 2.0 is the only protocol enabled in the server.
10517 *) Make it possible to get hexdumps of unprintable data with 'openssl
10518 asn1parse'. By implication, the functions ASN1_parse_dump() and
10519 BIO_dump_indent() are added.
10522 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10523 these print out strings and name structures based on various
10524 flags including RFC2253 support and proper handling of
10525 multibyte characters. Added options to the 'x509' utility
10526 to allow the various flags to be set.
10529 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10530 Also change the functions X509_cmp_current_time() and
10531 X509_gmtime_adj() work with an ASN1_TIME structure,
10532 this will enable certificates using GeneralizedTime in validity
10533 dates to be checked.
10536 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10537 negative public key encodings) on by default,
10538 NO_NEG_PUBKEY_BUG can be set to disable it.
10541 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10542 content octets. An i2c_ASN1_OBJECT is unnecessary because
10543 the encoding can be trivially obtained from the structure.
10546 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10547 not read locks (CRYPTO_r_[un]lock).
10550 *) A first attempt at creating official support for shared
10551 libraries through configuration. I've kept it so the
10552 default is static libraries only, and the OpenSSL programs
10553 are always statically linked for now, but there are
10554 preparations for dynamic linking in place.
10555 This has been tested on Linux and Tru64.
10558 *) Randomness polling function for Win9x, as described in:
10559 Peter Gutmann, Software Generation of Practically Strong
10563 *) Fix so PRNG is seeded in req if using an already existing
10567 *) New options to smime application. -inform and -outform
10568 allow alternative formats for the S/MIME message including
10569 PEM and DER. The -content option allows the content to be
10570 specified separately. This should allow things like Netscape
10571 form signing output easier to verify.
10574 *) Fix the ASN1 encoding of tags using the 'long form'.
10577 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10578 STRING types. These convert content octets to and from the
10579 underlying type. The actual tag and length octets are
10580 already assumed to have been read in and checked. These
10581 are needed because all other string types have virtually
10582 identical handling apart from the tag. By having versions
10583 of the ASN1 functions that just operate on content octets
10584 IMPLICIT tagging can be handled properly. It also allows
10585 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10586 and ASN1_INTEGER are identical apart from the tag.
10589 *) Change the handling of OID objects as follows:
10591 - New object identifiers are inserted in objects.txt, following
10592 the syntax given in objects.README.
10593 - objects.pl is used to process obj_mac.num and create a new
10595 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10598 This is currently kind of a hack, and the perl code in objects.pl
10599 isn't very elegant, but it works as I intended. The simplest way
10600 to check that it worked correctly is to look in obj_dat.h and
10601 check the array nid_objs and make sure the objects haven't moved
10602 around (this is important!). Additions are OK, as well as
10603 consistent name changes.
10606 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10609 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10610 The given file adds to whatever has already been seeded into the
10611 random pool through the RANDFILE configuration file option or
10612 environment variable, or the default random state file.
10615 *) mkstack.pl now sorts each macro group into lexical order.
10616 Previously the output order depended on the order the files
10617 appeared in the directory, resulting in needless rewriting
10621 *) Patches to make OpenSSL compile under Win32 again. Mostly
10622 work arounds for the VC++ problem that it treats func() as
10623 func(void). Also stripped out the parts of mkdef.pl that
10624 added extra typesafe functions: these no longer exist.
10627 *) Reorganisation of the stack code. The macros are now all
10628 collected in safestack.h . Each macro is defined in terms of
10629 a "stack macro" of the form SKM_<name>(type, a, b). The
10630 DEBUG_SAFESTACK is now handled in terms of function casts,
10631 this has the advantage of retaining type safety without the
10632 use of additional functions. If DEBUG_SAFESTACK is not defined
10633 then the non typesafe macros are used instead. Also modified the
10634 mkstack.pl script to handle the new form. Needs testing to see
10635 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10636 the default if no major problems. Similar behaviour for ASN1_SET_OF
10637 and PKCS12_STACK_OF.
10640 *) When some versions of IIS use the 'NET' form of private key the
10641 key derivation algorithm is different. Normally MD5(password) is
10642 used as a 128 bit RC4 key. In the modified case
10643 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10644 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10645 as the old Netscape_RSA functions except they have an additional
10646 'sgckey' parameter which uses the modified algorithm. Also added
10647 an -sgckey command line option to the rsa utility. Thanks to
10648 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10649 algorithm to openssl-dev.
10652 *) The evp_local.h macros were using 'c.##kname' which resulted in
10653 invalid expansion on some systems (SCO 5.0.5 for example).
10654 Corrected to 'c.kname'.
10655 [Phillip Porch <root@theporch.com>]
10657 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10658 a STACK of email addresses from a certificate or request, these look
10659 in the subject name and the subject alternative name extensions and
10660 omit any duplicate addresses.
10663 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10664 This makes DSA verification about 2 % faster.
10667 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10668 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10669 plus overhead for 1024 bit moduli).
10670 This makes exponentiations about 0.5 % faster for 1024 bit
10671 exponents (as measured by "openssl speed rsa2048").
10674 *) Rename memory handling macros to avoid conflicts with other
10676 Malloc => OPENSSL_malloc
10677 Malloc_locked => OPENSSL_malloc_locked
10678 Realloc => OPENSSL_realloc
10679 Free => OPENSSL_free
10682 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10683 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10686 *) CygWin32 support.
10687 [John Jarvie <jjarvie@newsguy.com>]
10689 *) The type-safe stack code has been rejigged. It is now only compiled
10690 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10691 by default all type-specific stack functions are "#define"d back to
10692 standard stack functions. This results in more streamlined output
10693 but retains the type-safety checking possibilities of the original
10697 *) The STACK code has been cleaned up, and certain type declarations
10698 that didn't make a lot of sense have been brought in line. This has
10699 also involved a cleanup of sorts in safestack.h to more correctly
10700 map type-safe stack functions onto their plain stack counterparts.
10701 This work has also resulted in a variety of "const"ifications of
10702 lots of the code, especially "_cmp" operations which should normally
10703 be prototyped with "const" parameters anyway.
10706 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10707 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10708 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10709 where all of 'md' is used each time the PRNG is used, but 'state'
10710 is used only indexed by a cyclic counter. As entropy may not be
10711 well distributed from the beginning, 'md' is important as a
10712 chaining variable. However, the output function chains only half
10713 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10714 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10715 in all of 'state' being rewritten, with the new values depending
10716 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10719 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10720 the handshake is continued after ssl_verify_cert_chain();
10721 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10722 can lead to 'unexplainable' connection aborts later.
10723 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10725 *) Major EVP API cipher revision.
10726 Add hooks for extra EVP features. This allows various cipher
10727 parameters to be set in the EVP interface. Support added for variable
10728 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10729 setting of RC2 and RC5 parameters.
10731 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10734 Remove lots of duplicated code from the EVP library. For example *every*
10735 cipher init() function handles the 'iv' in the same way according to the
10736 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10737 for CFB and OFB modes they zero ctx->num.
10739 New functionality allows removal of S/MIME code RC2 hack.
10741 Most of the routines have the same form and so can be declared in terms
10744 By shifting this to the top level EVP_CipherInit() it can be removed from
10745 all individual ciphers. If the cipher wants to handle IVs or keys
10746 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10749 Change lots of functions like EVP_EncryptUpdate() to now return a
10750 value: although software versions of the algorithms cannot fail
10751 any installed hardware versions can.
10754 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10755 this option is set, tolerate broken clients that send the negotiated
10756 protocol version number instead of the requested protocol version
10760 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10761 i.e. non-zero for export ciphersuites, zero otherwise.
10762 Previous versions had this flag inverted, inconsistent with
10763 rsa_tmp_cb (..._TMP_RSA_CB).
10764 [Bodo Moeller; problem reported by Amit Chopra]
10766 *) Add missing DSA library text string. Work around for some IIS
10767 key files with invalid SEQUENCE encoding.
10770 *) Add a document (doc/standards.txt) that list all kinds of standards
10771 and so on that are implemented in OpenSSL.
10774 *) Enhance c_rehash script. Old version would mishandle certificates
10775 with the same subject name hash and wouldn't handle CRLs at all.
10776 Added -fingerprint option to crl utility, to support new c_rehash
10780 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10783 *) Fix for SSL server purpose checking. Server checking was
10784 rejecting certificates which had extended key usage present
10785 but no ssl client purpose.
10786 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10788 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10789 is a little unclear about how a blank password is handled.
10790 Since the password in encoded as a BMPString with terminating
10791 double NULL a zero length password would end up as just the
10792 double NULL. However no password at all is different and is
10793 handled differently in the PKCS#12 key generation code. NS
10794 treats a blank password as zero length. MSIE treats it as no
10795 password on export: but it will try both on import. We now do
10796 the same: PKCS12_parse() tries zero length and no password if
10797 the password is set to "" or NULL (NULL is now a valid password:
10798 it wasn't before) as does the pkcs12 application.
10801 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10802 perror when PEM_read_bio_X509_REQ fails, the error message must
10803 be obtained from the error queue.
10806 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10807 it in ERR_remove_state if appropriate, and change ERR_get_state
10808 accordingly to avoid race conditions (this is necessary because
10809 thread_hash is no longer constant once set).
10812 *) Bugfix for linux-elf makefile.one.
10815 *) RSA_get_default_method() will now cause a default
10816 RSA_METHOD to be chosen if one doesn't exist already.
10817 Previously this was only set during a call to RSA_new()
10818 or RSA_new_method(NULL) meaning it was possible for
10819 RSA_get_default_method() to return NULL.
10822 *) Added native name translation to the existing DSO code
10823 that will convert (if the flag to do so is set) filenames
10824 that are sufficiently small and have no path information
10825 into a canonical native form. Eg. "blah" converted to
10826 "libblah.so" or "blah.dll" etc.
10829 *) New function ERR_error_string_n(e, buf, len) which is like
10830 ERR_error_string(e, buf), but writes at most 'len' bytes
10831 including the 0 terminator. For ERR_error_string_n, 'buf'
10833 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10835 *) CONF library reworked to become more general. A new CONF
10836 configuration file reader "class" is implemented as well as a
10837 new functions (NCONF_*, for "New CONF") to handle it. The now
10838 old CONF_* functions are still there, but are reimplemented to
10839 work in terms of the new functions. Also, a set of functions
10840 to handle the internal storage of the configuration data is
10841 provided to make it easier to write new configuration file
10842 reader "classes" (I can definitely see something reading a
10843 configuration file in XML format, for example), called _CONF_*,
10844 or "the configuration storage API"...
10846 The new configuration file reading functions are:
10848 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10849 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10851 NCONF_default, NCONF_WIN32
10853 NCONF_dump_fp, NCONF_dump_bio
10855 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10856 NCONF_new creates a new CONF object. This works in the same way
10857 as other interfaces in OpenSSL, like the BIO interface.
10858 NCONF_dump_* dump the internal storage of the configuration file,
10859 which is useful for debugging. All other functions take the same
10860 arguments as the old CONF_* functions with the exception of the
10861 first that must be a `CONF *' instead of a `LHASH *'.
10863 To make it easier to use the new classes with the old CONF_* functions,
10864 the function CONF_set_default_method is provided.
10867 *) Add '-tls1' option to 'openssl ciphers', which was already
10868 mentioned in the documentation but had not been implemented.
10869 (This option is not yet really useful because even the additional
10870 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10873 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10874 OpenSSL-based applications) load shared libraries and bind to
10875 them in a portable way.
10876 [Geoff Thorpe, with contributions from Richard Levitte]
10878 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10880 *) Make sure _lrotl and _lrotr are only used with MSVC.
10882 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10883 (the default implementation of RAND_status).
10885 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10886 to '-clrext' (= clear extensions), as intended and documented.
10887 [Bodo Moeller; inconsistency pointed out by Michael Attili
10888 <attili@amaxo.com>]
10890 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10891 was larger than the MD block size.
10892 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10894 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10895 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10896 using the passed key: if the passed key was a private key the result
10897 of X509_print(), for example, would be to print out all the private key
10901 *) des_quad_cksum() byte order bug fix.
10902 [Ulf Möller, using the problem description in krb4-0.9.7, where
10903 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10905 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10907 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10909 *) For easily testing in shell scripts whether some command
10910 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10911 returns with exit code 0 iff no command of the given name is available.
10912 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10913 the output goes to stdout and nothing is printed to stderr.
10914 Additional arguments are always ignored.
10916 Since for each cipher there is a command of the same name,
10917 the 'no-cipher' compilation switches can be tested this way.
10919 ('openssl no-XXX' is not able to detect pseudo-commands such
10920 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10923 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10926 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10927 is set; it will be thrown away anyway because each handshake creates
10929 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10930 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10931 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10932 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10935 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10936 'Q' and 'R' lose their special meanings (quit/renegotiate).
10937 This is part of what -quiet does; unlike -quiet, -ign_eof
10938 does not suppress any output.
10941 *) Add compatibility options to the purpose and trust code. The
10942 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10943 accepts a certificate or CA, this was the previous behaviour,
10944 with all the associated security issues.
10946 X509_TRUST_COMPAT is the old trust behaviour: only and
10947 automatically trust self signed roots in certificate store. A
10948 new trust setting X509_TRUST_DEFAULT is used to specify that
10949 a purpose has no associated trust setting and it should instead
10950 use the value in the default purpose.
10953 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10954 and fix a memory leak.
10957 *) In util/mkerr.pl (which implements 'make errors'), preserve
10958 reason strings from the previous version of the .c file, as
10959 the default to have only downcase letters (and digits) in
10960 automatically generated reasons codes is not always appropriate.
10963 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10964 using strerror. Previously, ERR_reason_error_string() returned
10965 library names as reason strings for SYSerr; but SYSerr is a special
10966 case where small numbers are errno values, not library numbers.
10969 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10970 converts DSA parameters into DH parameters. (When creating parameters,
10971 DSA_generate_parameters is used.)
10974 *) Include 'length' (recommended exponent length) in C code generated
10975 by 'openssl dhparam -C'.
10978 *) The second argument to set_label in perlasm was already being used
10979 so couldn't be used as a "file scope" flag. Moved to third argument
10983 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10984 instead of RAND_bytes for encryption IVs and salts.
10987 *) Include RAND_status() into RAND_METHOD instead of implementing
10988 it only for md_rand.c Otherwise replacing the PRNG by calling
10989 RAND_set_rand_method would be impossible.
10992 *) Don't let DSA_generate_key() enter an infinite loop if the random
10993 number generation fails.
10996 *) New 'rand' application for creating pseudo-random output.
10999 *) Added configuration support for Linux/IA64
11000 [Rolf Haberrecker <rolf@suse.de>]
11002 *) Assembler module support for Mingw32.
11005 *) Shared library support for HPUX (in shlib/).
11006 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
11008 *) Shared library support for Solaris gcc.
11009 [Lutz Behnke <behnke@trustcenter.de>]
11011 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
11013 *) PKCS7_encrypt() was adding text MIME headers twice because they
11014 were added manually and by SMIME_crlf_copy().
11017 *) In bntest.c don't call BN_rand with zero bits argument.
11018 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
11020 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
11021 case was implemented. This caused BN_div_recp() to fail occasionally.
11024 *) Add an optional second argument to the set_label() in the perl
11025 assembly language builder. If this argument exists and is set
11026 to 1 it signals that the assembler should use a symbol whose
11027 scope is the entire file, not just the current function. This
11028 is needed with MASM which uses the format label:: for this scope.
11029 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
11031 *) Change the ASN1 types so they are typedefs by default. Before
11032 almost all types were #define'd to ASN1_STRING which was causing
11033 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
11037 *) Change names of new functions to the new get1/get0 naming
11038 convention: After 'get1', the caller owns a reference count
11039 and has to call ..._free; 'get0' returns a pointer to some
11040 data structure without incrementing reference counters.
11041 (Some of the existing 'get' functions increment a reference
11042 counter, some don't.)
11043 Similarly, 'set1' and 'add1' functions increase reference
11044 counters or duplicate objects.
11047 *) Allow for the possibility of temp RSA key generation failure:
11048 the code used to assume it always worked and crashed on failure.
11051 *) Fix potential buffer overrun problem in BIO_printf().
11052 [Ulf Möller, using public domain code by Patrick Powell; problem
11053 pointed out by David Sacerdote <das33@cornell.edu>]
11055 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
11056 RAND_egd() and RAND_status(). In the command line application,
11057 the EGD socket can be specified like a seed file using RANDFILE
11061 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
11062 Some CAs (e.g. Verisign) distribute certificates in this form.
11065 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
11066 list to exclude them. This means that no special compilation option
11067 is needed to use anonymous DH: it just needs to be included in the
11071 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
11072 EVP_MD_type. The old functionality is available in a new macro called
11073 EVP_MD_md(). Change code that uses it and update docs.
11076 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
11077 where the 'void *' argument is replaced by a function pointer argument.
11078 Previously 'void *' was abused to point to functions, which works on
11079 many platforms, but is not correct. As these functions are usually
11080 called by macros defined in OpenSSL header files, most source code
11081 should work without changes.
11084 *) <openssl/opensslconf.h> (which is created by Configure) now contains
11085 sections with information on -D... compiler switches used for
11086 compiling the library so that applications can see them. To enable
11087 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
11088 must be defined. E.g.,
11089 #define OPENSSL_ALGORITHM_DEFINES
11090 #include <openssl/opensslconf.h>
11091 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
11092 [Richard Levitte, Ulf and Bodo Möller]
11094 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
11098 *) Change the 'other' type in certificate aux info to a STACK_OF
11099 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
11100 the required ASN1 format: arbitrary types determined by an OID.
11103 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
11104 argument to 'req'. This is not because the function is newer or
11105 better than others it just uses the work 'NEW' in the certificate
11106 request header lines. Some software needs this.
11109 *) Reorganise password command line arguments: now passwords can be
11110 obtained from various sources. Delete the PEM_cb function and make
11111 it the default behaviour: i.e. if the callback is NULL and the
11112 usrdata argument is not NULL interpret it as a null terminated pass
11113 phrase. If usrdata and the callback are NULL then the pass phrase
11114 is prompted for as usual.
11117 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
11118 the support is automatically enabled. The resulting binaries will
11119 autodetect the card and use it if present.
11120 [Ben Laurie and Compaq Inc.]
11122 *) Work around for Netscape hang bug. This sends certificate request
11123 and server done in one record. Since this is perfectly legal in the
11124 SSL/TLS protocol it isn't a "bug" option and is on by default. See
11125 the bugs/SSLv3 entry for more info.
11128 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
11131 *) Add -rand argument to smime and pkcs12 applications and read/write
11135 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
11138 *) Add command line password options to the remaining applications.
11141 *) Bug fix for BN_div_recp() for numerators with an even number of
11145 *) More tests in bntest.c, and changed test_bn output.
11148 *) ./config recognizes MacOS X now.
11151 *) Bug fix for BN_div() when the first words of num and divisor are
11152 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
11155 *) Add support for various broken PKCS#8 formats, and command line
11156 options to produce them.
11159 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
11160 get temporary BIGNUMs from a BN_CTX.
11163 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
11167 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
11168 include a #define from the old name to the new. The original intent
11169 was that statically linked binaries could for example just call
11170 SSLeay_add_all_ciphers() to just add ciphers to the table and not
11171 link with digests. This never worked because SSLeay_add_all_digests()
11172 and SSLeay_add_all_ciphers() were in the same source file so calling
11173 one would link with the other. They are now in separate source files.
11176 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
11179 *) Use a less unusual form of the Miller-Rabin primality test (it used
11180 a binary algorithm for exponentiation integrated into the Miller-Rabin
11181 loop, our standard modexp algorithms are faster).
11184 *) Support for the EBCDIC character set completed.
11185 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
11187 *) Source code cleanups: use const where appropriate, eliminate casts,
11188 use void * instead of char * in lhash.
11191 *) Bugfix: ssl3_send_server_key_exchange was not restartable
11192 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
11193 this the server could overwrite ephemeral keys that the client
11197 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
11198 using 50 iterations of the Rabin-Miller test.
11200 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
11201 iterations of the Rabin-Miller test as required by the appendix
11202 to FIPS PUB 186[-1]) instead of DSA_is_prime.
11203 As BN_is_prime_fasttest includes trial division, DSA parameter
11204 generation becomes much faster.
11206 This implies a change for the callback functions in DSA_is_prime
11207 and DSA_generate_parameters: The callback function is called once
11208 for each positive witness in the Rabin-Miller test, not just
11209 occasionally in the inner loop; and the parameters to the
11210 callback function now provide an iteration count for the outer
11211 loop rather than for the current invocation of the inner loop.
11212 DSA_generate_parameters additionally can call the callback
11213 function with an 'iteration count' of -1, meaning that a
11214 candidate has passed the trial division test (when q is generated
11215 from an application-provided seed, trial division is skipped).
11218 *) New function BN_is_prime_fasttest that optionally does trial
11219 division before starting the Rabin-Miller test and has
11220 an additional BN_CTX * argument (whereas BN_is_prime always
11221 has to allocate at least one BN_CTX).
11222 'callback(1, -1, cb_arg)' is called when a number has passed the
11223 trial division stage.
11226 *) Fix for bug in CRL encoding. The validity dates weren't being handled
11230 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
11233 *) New function BN_pseudo_rand().
11236 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
11237 bignum version of BN_from_montgomery() with the working code from
11238 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
11242 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
11243 made it impossible to use the same SSL_SESSION data structure in
11244 SSL2 clients in multiple threads.
11247 *) The return value of RAND_load_file() no longer counts bytes obtained
11248 by stat(). RAND_load_file(..., -1) is new and uses the complete file
11249 to seed the PRNG (previously an explicit byte count was required).
11250 [Ulf Möller, Bodo Möller]
11252 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
11253 used (char *) instead of (void *) and had casts all over the place.
11256 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
11259 *) Retain source code compatibility for BN_prime_checks macro:
11260 BN_is_prime(..., BN_prime_checks, ...) now uses
11261 BN_prime_checks_for_size to determine the appropriate number of
11262 Rabin-Miller iterations.
11265 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
11266 DH_CHECK_P_NOT_SAFE_PRIME.
11267 (Check if this is true? OpenPGP calls them "strong".)
11270 *) Merge the functionality of "dh" and "gendh" programs into a new program
11271 "dhparam". The old programs are retained for now but will handle DH keys
11272 (instead of parameters) in future.
11275 *) Make the ciphers, s_server and s_client programs check the return values
11276 when a new cipher list is set.
11279 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
11280 ciphers. Before when the 56bit ciphers were enabled the sorting was
11283 The syntax for the cipher sorting has been extended to support sorting by
11284 cipher-strength (using the strength_bits hard coded in the tables).
11285 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
11287 Fix a bug in the cipher-command parser: when supplying a cipher command
11288 string with an "undefined" symbol (neither command nor alphanumeric
11289 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
11290 an error is flagged.
11292 Due to the strength-sorting extension, the code of the
11293 ssl_create_cipher_list() function was completely rearranged. I hope that
11294 the readability was also increased :-)
11295 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
11297 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
11298 for the first serial number and places 2 in the serial number file. This
11299 avoids problems when the root CA is created with serial number zero and
11300 the first user certificate has the same issuer name and serial number
11304 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
11305 the new code. Add documentation for this stuff.
11308 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
11309 X509_*() to X509at_*() on the grounds that they don't handle X509
11310 structures and behave in an analogous way to the X509v3 functions:
11311 they shouldn't be called directly but wrapper functions should be used
11314 So we also now have some wrapper functions that call the X509at functions
11315 when passed certificate requests. (TO DO: similar things can be done with
11316 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11317 things. Some of these need some d2i or i2d and print functionality
11318 because they handle more complex structures.)
11321 *) Add missing #ifndefs that caused missing symbols when building libssl
11322 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11323 NO_RSA in ssl/s2*.c.
11324 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11326 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11327 has a return value which indicates the quality of the random data
11328 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11329 error queue. New function RAND_pseudo_bytes() generates output that is
11330 guaranteed to be unique but not unpredictable. RAND_add is like
11331 RAND_seed, but takes an extra argument for an entropy estimate
11332 (RAND_seed always assumes full entropy).
11335 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11336 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11337 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11338 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11339 false-positive rate of at most 2^-80 for random input.
11342 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11345 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11346 in the 0.9.5 release), this returns the chain
11347 from an X509_CTX structure with a dup of the stack and all
11348 the X509 reference counts upped: so the stack will exist
11349 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11352 Also make SSL_SESSION_print() print out the verify return
11356 *) Add manpage for the pkcs12 command. Also change the default
11357 behaviour so MAC iteration counts are used unless the new
11358 -nomaciter option is used. This improves file security and
11359 only older versions of MSIE (4.0 for example) need it.
11362 *) Honor the no-xxx Configure options when creating .DEF files.
11365 *) Add PKCS#10 attributes to field table: challengePassword,
11366 unstructuredName and unstructuredAddress. These are taken from
11367 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11368 international characters are used.
11370 More changes to X509_ATTRIBUTE code: allow the setting of types
11371 based on strings. Remove the 'loc' parameter when adding
11372 attributes because these will be a SET OF encoding which is sorted
11376 *) Initial changes to the 'req' utility to allow request generation
11377 automation. This will allow an application to just generate a template
11378 file containing all the field values and have req construct the
11381 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11382 used all over the place including certificate requests and PKCS#7
11383 structures. They are currently handled manually where necessary with
11384 some primitive wrappers for PKCS#7. The new functions behave in a
11385 manner analogous to the X509 extension functions: they allow
11386 attributes to be looked up by NID and added.
11388 Later something similar to the X509V3 code would be desirable to
11389 automatically handle the encoding, decoding and printing of the
11390 more complex types. The string types like challengePassword can
11391 be handled by the string table functions.
11393 Also modified the multi byte string table handling. Now there is
11394 a 'global mask' which masks out certain types. The table itself
11395 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11396 is useful when for example there is only one permissible type
11397 (as in countryName) and using the mask might result in no valid
11401 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11402 SSL_get_peer_finished to allow applications to obtain the latest
11403 Finished messages sent to the peer or expected from the peer,
11404 respectively. (SSL_get_peer_finished is usually the Finished message
11405 actually received from the peer, otherwise the protocol will be aborted.)
11407 As the Finished message are message digests of the complete handshake
11408 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11409 be used for external authentication procedures when the authentication
11410 provided by SSL/TLS is not desired or is not enough.
11413 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11414 the host supports BWX extension and if Compaq C is present on the
11415 $PATH. Just exploiting of the BWX extension results in 20-30%
11416 performance kick for some algorithms, e.g. DES and RC4 to mention
11417 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11421 *) Add support for MS "fast SGC". This is arguably a violation of the
11422 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11423 weak crypto and after checking the certificate is SGC a second one
11424 with strong crypto. MS SGC stops the first handshake after receiving
11425 the server certificate message and sends a second client hello. Since
11426 a server will typically do all the time consuming operations before
11427 expecting any further messages from the client (server key exchange
11428 is the most expensive) there is little difference between the two.
11430 To get OpenSSL to support MS SGC we have to permit a second client
11431 hello message after we have sent server done. In addition we have to
11432 reset the MAC if we do get this second client hello.
11435 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11436 if a DER encoded private key is RSA or DSA traditional format. Changed
11437 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11438 format DER encoded private key. Newer code should use PKCS#8 format which
11439 has the key type encoded in the ASN1 structure. Added DER private key
11440 support to pkcs8 application.
11443 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11444 ciphersuites has been selected (as required by the SSL 3/TLS 1
11445 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11446 is set, we interpret this as a request to violate the specification
11447 (the worst that can happen is a handshake failure, and 'correct'
11448 behaviour would result in a handshake failure anyway).
11451 *) In SSL_CTX_add_session, take into account that there might be multiple
11452 SSL_SESSION structures with the same session ID (e.g. when two threads
11453 concurrently obtain them from an external cache).
11454 The internal cache can handle only one SSL_SESSION with a given ID,
11455 so if there's a conflict, we now throw out the old one to achieve
11459 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11460 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11461 some routines that use cipher OIDs: some ciphers do not have OIDs
11462 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11466 *) Simplify the trust setting structure and code. Now we just have
11467 two sequences of OIDs for trusted and rejected settings. These will
11468 typically have values the same as the extended key usage extension
11469 and any application specific purposes.
11471 The trust checking code now has a default behaviour: it will just
11472 check for an object with the same NID as the passed id. Functions can
11473 be provided to override either the default behaviour or the behaviour
11474 for a given id. SSL client, server and email already have functions
11475 in place for compatibility: they check the NID and also return "trusted"
11476 if the certificate is self signed.
11479 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11480 traditional format into an EVP_PKEY structure.
11483 *) Add a password callback function PEM_cb() which either prompts for
11484 a password if usr_data is NULL or otherwise assumes it is a null
11485 terminated password. Allow passwords to be passed on command line
11486 environment or config files in a few more utilities.
11489 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11490 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11491 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11492 Update documentation.
11495 *) Support for ASN1 "NULL" type. This could be handled before by using
11496 ASN1_TYPE but there wasn't any function that would try to read a NULL
11497 and produce an error if it couldn't. For compatibility we also have
11498 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11499 don't allocate anything because they don't need to.
11502 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11504 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11506 *) Rebuild of the memory allocation routines used by OpenSSL code and
11507 possibly others as well. The purpose is to make an interface that
11508 provide hooks so anyone can build a separate set of allocation and
11509 deallocation routines to be used by OpenSSL, for example memory
11510 pool implementations, or something else, which was previously hard
11511 since Malloc(), Realloc() and Free() were defined as macros having
11512 the values malloc, realloc and free, respectively (except for Win32
11513 compilations). The same is provided for memory debugging code.
11514 OpenSSL already comes with functionality to find memory leaks, but
11515 this gives people a chance to debug other memory problems.
11517 With these changes, a new set of functions and macros have appeared:
11519 CRYPTO_set_mem_debug_functions() [F]
11520 CRYPTO_get_mem_debug_functions() [F]
11521 CRYPTO_dbg_set_options() [F]
11522 CRYPTO_dbg_get_options() [F]
11523 CRYPTO_malloc_debug_init() [M]
11525 The memory debug functions are NULL by default, unless the library
11526 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11527 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11528 gives the standard debugging functions that come with OpenSSL) or
11529 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11530 provided by the library user) must be used. When the standard
11531 debugging functions are used, CRYPTO_dbg_set_options can be used to
11532 request additional information:
11533 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11534 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11536 Also, things like CRYPTO_set_mem_functions will always give the
11537 expected result (the new set of functions is used for allocation
11538 and deallocation) at all times, regardless of platform and compiler
11541 To finish it up, some functions that were never use in any other
11542 way than through macros have a new API and new semantic:
11544 CRYPTO_dbg_malloc()
11545 CRYPTO_dbg_realloc()
11548 All macros of value have retained their old syntax.
11549 [Richard Levitte and Bodo Moeller]
11551 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11552 ordering of SMIMECapabilities wasn't in "strength order" and there
11553 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11557 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11558 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11559 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11561 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11562 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11563 functionality to handle multipart/signed properly) and a utility
11564 called 'smime' to call all this stuff. This is based on code I
11565 originally wrote for Celo who have kindly allowed it to be
11566 included in OpenSSL.
11569 *) Add variants des_set_key_checked and des_set_key_unchecked of
11570 des_set_key (aka des_key_sched). Global variable des_check_key
11571 decides which of these is called by des_set_key; this way
11572 des_check_key behaves as it always did, but applications and
11573 the library itself, which was buggy for des_check_key == 1,
11574 have a cleaner way to pick the version they need.
11577 *) New function PKCS12_newpass() which changes the password of a
11581 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11582 dynamic mix. In both cases the ids can be used as an index into the
11583 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11584 functions so they accept a list of the field values and the
11585 application doesn't need to directly manipulate the X509_TRUST
11589 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11593 *) Modify the way the V3 extension code looks up extensions. This now
11594 works in a similar way to the object code: we have some "standard"
11595 extensions in a static table which is searched with OBJ_bsearch()
11596 and the application can add dynamic ones if needed. The file
11597 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11598 updated whenever a new extension is added to the core code and kept
11599 in ext_nid order. There is a simple program 'tabtest.c' which checks
11600 this. New extensions are not added too often so this file can readily
11601 be maintained manually.
11603 There are two big advantages in doing things this way. The extensions
11604 can be looked up immediately and no longer need to be "added" using
11605 X509V3_add_standard_extensions(): this function now does nothing.
11606 [Side note: I get *lots* of email saying the extension code doesn't
11607 work because people forget to call this function]
11608 Also no dynamic allocation is done unless new extensions are added:
11609 so if we don't add custom extensions there is no need to call
11610 X509V3_EXT_cleanup().
11613 *) Modify enc utility's salting as follows: make salting the default. Add a
11614 magic header, so unsalted files fail gracefully instead of just decrypting
11615 to garbage. This is because not salting is a big security hole, so people
11616 should be discouraged from doing it.
11619 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11620 digest to be passed on the command line but it only used this
11621 parameter when signing a certificate. Modified so all relevant
11622 operations are affected by the digest parameter including the
11623 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11624 DSA key was used because it didn't fix the digest.
11627 *) Initial certificate chain verify code. Currently tests the untrusted
11628 certificates for consistency with the verify purpose (which is set
11629 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11631 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11632 this is because it will reject chains with invalid extensions whereas
11633 every previous version of OpenSSL and SSLeay made no checks at all.
11635 Trust code: checks the root CA for the relevant trust settings. Trust
11636 settings have an initial value consistent with the verify purpose: e.g.
11637 if the verify purpose is for SSL client use it expects the CA to be
11638 trusted for SSL client use. However the default value can be changed to
11639 permit custom trust settings: one example of this would be to only trust
11640 certificates from a specific "secure" set of CAs.
11642 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11643 which should be used for version portability: especially since the
11644 verify structure is likely to change more often now.
11646 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11647 to set them. If not set then assume SSL clients will verify SSL servers
11650 Two new options to the verify program: -untrusted allows a set of
11651 untrusted certificates to be passed in and -purpose which sets the
11652 intended purpose of the certificate. If a purpose is set then the
11653 new chain verify code is used to check extension consistency.
11656 *) Support for the authority information access extension.
11659 *) Modify RSA and DSA PEM read routines to transparently handle
11660 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11661 public keys in a format compatible with certificate
11662 SubjectPublicKeyInfo structures. Unfortunately there were already
11663 functions called *_PublicKey_* which used various odd formats so
11664 these are retained for compatibility: however the DSA variants were
11665 never in a public release so they have been deleted. Changed dsa/rsa
11666 utilities to handle the new format: note no releases ever handled public
11667 keys so we should be OK.
11669 The primary motivation for this change is to avoid the same fiasco
11670 that dogs private keys: there are several incompatible private key
11671 formats some of which are standard and some OpenSSL specific and
11672 require various evil hacks to allow partial transparent handling and
11673 even then it doesn't work with DER formats. Given the option anything
11674 other than PKCS#8 should be dumped: but the other formats have to
11675 stay in the name of compatibility.
11677 With public keys and the benefit of hindsight one standard format
11678 is used which works with EVP_PKEY, RSA or DSA structures: though
11679 it clearly returns an error if you try to read the wrong kind of key.
11681 Added a -pubkey option to the 'x509' utility to output the public key.
11682 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11683 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11684 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11685 that do the same as the EVP_PKEY_assign_*() except they up the
11686 reference count of the added key (they don't "swallow" the
11690 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11691 CRLs would fail if the file contained no certificates or no CRLs:
11692 added a new function to read in both types and return the number
11693 read: this means that if none are read it will be an error. The
11694 DER versions of the certificate and CRL reader would always fail
11695 because it isn't possible to mix certificates and CRLs in DER format
11696 without choking one or the other routine. Changed this to just read
11697 a certificate: this is the best we can do. Also modified the code
11698 in apps/verify.c to take notice of return codes: it was previously
11699 attempting to read in certificates from NULL pointers and ignoring
11700 any errors: this is one reason why the cert and CRL reader seemed
11701 to work. It doesn't check return codes from the default certificate
11702 routines: these may well fail if the certificates aren't installed.
11705 *) Code to support otherName option in GeneralName.
11708 *) First update to verify code. Change the verify utility
11709 so it warns if it is passed a self signed certificate:
11710 for consistency with the normal behaviour. X509_verify
11711 has been modified to it will now verify a self signed
11712 certificate if *exactly* the same certificate appears
11713 in the store: it was previously impossible to trust a
11714 single self signed certificate. This means that:
11715 openssl verify ss.pem
11716 now gives a warning about a self signed certificate but
11717 openssl verify -CAfile ss.pem ss.pem
11721 *) For servers, store verify_result in SSL_SESSION data structure
11722 (and add it to external session representation).
11723 This is needed when client certificate verifications fails,
11724 but an application-provided verification callback (set by
11725 SSL_CTX_set_cert_verify_callback) allows accepting the session
11726 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11727 but returns 1): When the session is reused, we have to set
11728 ssl->verify_result to the appropriate error code to avoid
11730 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11732 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11733 case in PKCS7_dataInit() where the signed PKCS7 structure
11734 didn't contain any existing data because it was being created.
11735 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11737 *) Add a salt to the key derivation routines in enc.c. This
11738 forms the first 8 bytes of the encrypted file. Also add a
11739 -S option to allow a salt to be input on the command line.
11742 *) New function X509_cmp(). Oddly enough there wasn't a function
11743 to compare two certificates. We do this by working out the SHA1
11744 hash and comparing that. X509_cmp() will be needed by the trust
11748 *) SSL_get1_session() is like SSL_get_session(), but increments
11749 the reference count in the SSL_SESSION returned.
11750 [Geoff Thorpe <geoff@eu.c2.net>]
11752 *) Fix for 'req': it was adding a null to request attributes.
11753 Also change the X509_LOOKUP and X509_INFO code to handle
11754 certificate auxiliary information.
11757 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11761 *) Add the possibility to add extra information to the memory leak
11762 detecting output, to form tracebacks, showing from where each
11763 allocation was originated: CRYPTO_push_info("constant string") adds
11764 the string plus current file name and line number to a per-thread
11765 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11766 is like calling CYRPTO_pop_info() until the stack is empty.
11767 Also updated memory leak detection code to be multi-thread-safe.
11770 *) Add options -text and -noout to pkcs7 utility and delete the
11771 encryption options which never did anything. Update docs.
11774 *) Add options to some of the utilities to allow the pass phrase
11775 to be included on either the command line (not recommended on
11776 OSes like Unix) or read from the environment. Update the
11777 manpages and fix a few bugs.
11780 *) Add a few manpages for some of the openssl commands.
11783 *) Fix the -revoke option in ca. It was freeing up memory twice,
11784 leaking and not finding already revoked certificates.
11787 *) Extensive changes to support certificate auxiliary information.
11788 This involves the use of X509_CERT_AUX structure and X509_AUX
11789 functions. An X509_AUX function such as PEM_read_X509_AUX()
11790 can still read in a certificate file in the usual way but it
11791 will also read in any additional "auxiliary information". By
11792 doing things this way a fair degree of compatibility can be
11793 retained: existing certificates can have this information added
11794 using the new 'x509' options.
11796 Current auxiliary information includes an "alias" and some trust
11797 settings. The trust settings will ultimately be used in enhanced
11798 certificate chain verification routines: currently a certificate
11799 can only be trusted if it is self signed and then it is trusted
11803 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11804 The problem was that one of the replacement routines had not been working
11805 since SSLeay releases. For now the offending routine has been replaced
11806 with non-optimised assembler. Even so, this now gives around 95%
11807 performance improvement for 1024 bit RSA signs.
11810 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11811 handling. Most clients have the effective key size in bits equal to
11812 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11813 A few however don't do this and instead use the size of the decrypted key
11814 to determine the RC2 key length and the AlgorithmIdentifier to determine
11815 the effective key length. In this case the effective key length can still
11816 be 40 bits but the key length can be 168 bits for example. This is fixed
11817 by manually forcing an RC2 key into the EVP_PKEY structure because the
11818 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11819 the key length and effective key length are equal.
11822 *) Add a bunch of functions that should simplify the creation of
11823 X509_NAME structures. Now you should be able to do:
11824 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11825 and have it automatically work out the correct field type and fill in
11826 the structures. The more adventurous can try:
11827 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11828 and it will (hopefully) work out the correct multibyte encoding.
11831 *) Change the 'req' utility to use the new field handling and multibyte
11832 copy routines. Before the DN field creation was handled in an ad hoc
11833 way in req, ca, and x509 which was rather broken and didn't support
11834 BMPStrings or UTF8Strings. Since some software doesn't implement
11835 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11836 using the dirstring_type option. See the new comment in the default
11837 openssl.cnf for more info.
11840 *) Make crypto/rand/md_rand.c more robust:
11841 - Assure unique random numbers after fork().
11842 - Make sure that concurrent threads access the global counter and
11843 md serializably so that we never lose entropy in them
11844 or use exactly the same state in multiple threads.
11845 Access to the large state is not always serializable because
11846 the additional locking could be a performance killer, and
11847 md should be large enough anyway.
11850 *) New file apps/app_rand.c with commonly needed functionality
11851 for handling the random seed file.
11853 Use the random seed file in some applications that previously did not:
11855 dsaparam -genkey (which also ignored its '-rand' option),
11858 x509 (when signing).
11859 Except on systems with /dev/urandom, it is crucial to have a random
11860 seed file at least for key creation, DSA signing, and for DH exchanges;
11861 for RSA signatures we could do without one.
11863 gendh and gendsa (unlike genrsa) used to read only the first byte
11864 of each file listed in the '-rand' option. The function as previously
11865 found in genrsa is now in app_rand.c and is used by all programs
11866 that support '-rand'.
11869 *) In RAND_write_file, use mode 0600 for creating files;
11870 don't just chmod when it may be too late.
11873 *) Report an error from X509_STORE_load_locations
11874 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11877 *) New function ASN1_mbstring_copy() this copies a string in either
11878 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11879 into an ASN1_STRING type. A mask of permissible types is passed
11880 and it chooses the "minimal" type to use or an error if not type
11884 *) Add function equivalents to the various macros in asn1.h. The old
11885 macros are retained with an M_ prefix. Code inside the library can
11886 use the M_ macros. External code (including the openssl utility)
11887 should *NOT* in order to be "shared library friendly".
11890 *) Add various functions that can check a certificate's extensions
11891 to see if it usable for various purposes such as SSL client,
11892 server or S/MIME and CAs of these types. This is currently
11893 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11894 verification. Also added a -purpose flag to x509 utility to
11895 print out all the purposes.
11898 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11902 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11903 for, obtain and decode and extension and obtain its critical flag.
11904 This allows all the necessary extension code to be handled in a
11905 single function call.
11908 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11909 platforms. See crypto/rc4/rc4_enc.c for further details.
11912 *) New -noout option to asn1parse. This causes no output to be produced
11913 its main use is when combined with -strparse and -out to extract data
11914 from a file (which may not be in ASN.1 format).
11917 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11918 when producing the local key id.
11919 [Richard Levitte <levitte@stacken.kth.se>]
11921 *) New option -dhparam in s_server. This allows a DH parameter file to be
11922 stated explicitly. If it is not stated then it tries the first server
11923 certificate file. The previous behaviour hard coded the filename
11927 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11928 a public key to be input or output. For example:
11929 openssl rsa -in key.pem -pubout -out pubkey.pem
11930 Also added necessary DSA public key functions to handle this.
11933 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11934 in the message. This was handled by allowing
11935 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11936 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11938 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11939 to the end of the strings whereas this didn't. This would cause problems
11940 if strings read with d2i_ASN1_bytes() were later modified.
11941 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11943 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11944 data and it contains EOF it will end up returning an error. This is
11945 caused by input 46 bytes long. The cause is due to the way base64
11946 BIOs find the start of base64 encoded data. They do this by trying a
11947 trial decode on each line until they find one that works. When they
11948 do a flag is set and it starts again knowing it can pass all the
11949 data directly through the decoder. Unfortunately it doesn't reset
11950 the context it uses. This means that if EOF is reached an attempt
11951 is made to pass two EOFs through the context and this causes the
11952 resulting error. This can also cause other problems as well. As is
11953 usual with these problems it takes *ages* to find and the fix is
11954 trivial: move one line.
11955 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11957 *) Ugly workaround to get s_client and s_server working under Windows. The
11958 old code wouldn't work because it needed to select() on sockets and the
11959 tty (for keypresses and to see if data could be written). Win32 only
11960 supports select() on sockets so we select() with a 1s timeout on the
11961 sockets and then see if any characters are waiting to be read, if none
11962 are present then we retry, we also assume we can always write data to
11963 the tty. This isn't nice because the code then blocks until we've
11964 received a complete line of data and it is effectively polling the
11965 keyboard at 1s intervals: however it's quite a bit better than not
11966 working at all :-) A dedicated Windows application might handle this
11967 with an event loop for example.
11970 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11971 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11972 will be called when RSA_sign() and RSA_verify() are used. This is useful
11973 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11974 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11975 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11976 This necessitated the support of an extra signature type NID_md5_sha1
11977 for SSL signatures and modifications to the SSL library to use it instead
11978 of calling RSA_public_decrypt() and RSA_private_encrypt().
11981 *) Add new -verify -CAfile and -CApath options to the crl program, these
11982 will lookup a CRL issuers certificate and verify the signature in a
11983 similar way to the verify program. Tidy up the crl program so it
11984 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11985 less strict. It will now permit CRL extensions even if it is not
11986 a V2 CRL: this will allow it to tolerate some broken CRLs.
11989 *) Initialize all non-automatic variables each time one of the openssl
11990 sub-programs is started (this is necessary as they may be started
11991 multiple times from the "OpenSSL>" prompt).
11992 [Lennart Bang, Bodo Moeller]
11994 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11995 removing all other RSA functionality (this is what NO_RSA does). This
11996 is so (for example) those in the US can disable those operations covered
11997 by the RSA patent while allowing storage and parsing of RSA keys and RSA
12001 *) Non-copying interface to BIO pairs.
12002 (still largely untested)
12005 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
12006 ASCII string. This was handled independently in various places before.
12009 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
12010 UTF8 strings a character at a time.
12013 *) Use client_version from client hello to select the protocol
12014 (s23_srvr.c) and for RSA client key exchange verification
12015 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
12018 *) Add various utility functions to handle SPKACs, these were previously
12019 handled by poking round in the structure internals. Added new function
12020 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
12021 print, verify and generate SPKACs. Based on an original idea from
12022 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
12025 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
12028 *) Allow the config file extension section to be overwritten on the
12029 command line. Based on an original idea from Massimiliano Pala
12030 <madwolf@comune.modena.it>. The new option is called -extensions
12031 and can be applied to ca, req and x509. Also -reqexts to override
12032 the request extensions in req and -crlexts to override the crl extensions
12036 *) Add new feature to the SPKAC handling in ca. Now you can include
12037 the same field multiple times by preceding it by "XXXX." for example:
12040 this is the same syntax as used in the req config file.
12043 *) Allow certificate extensions to be added to certificate requests. These
12044 are specified in a 'req_extensions' option of the req section of the
12045 config file. They can be printed out with the -text option to req but
12046 are otherwise ignored at present.
12049 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
12050 data read consists of only the final block it would not decrypted because
12051 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
12052 A misplaced 'break' also meant the decrypted final block might not be
12053 copied until the next read.
12056 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
12057 a few extra parameters to the DH structure: these will be useful if
12058 for example we want the value of 'q' or implement X9.42 DH.
12061 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
12062 provides hooks that allow the default DSA functions or functions on a
12063 "per key" basis to be replaced. This allows hardware acceleration and
12064 hardware key storage to be handled without major modification to the
12065 library. Also added low level modexp hooks and CRYPTO_EX structure and
12066 associated functions.
12069 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
12070 as "read only": it can't be written to and the buffer it points to will
12071 not be freed. Reading from a read only BIO is much more efficient than
12072 a normal memory BIO. This was added because there are several times when
12073 an area of memory needs to be read from a BIO. The previous method was
12074 to create a memory BIO and write the data to it, this results in two
12075 copies of the data and an O(n^2) reading algorithm. There is a new
12076 function BIO_new_mem_buf() which creates a read only memory BIO from
12077 an area of memory. Also modified the PKCS#7 routines to use read only
12081 *) Bugfix: ssl23_get_client_hello did not work properly when called in
12082 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
12083 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
12084 but a retry condition occurred while trying to read the rest.
12087 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
12088 NID_pkcs7_encrypted by default: this was wrong since this should almost
12089 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
12090 the encrypted data type: this is a more sensible place to put it and it
12091 allows the PKCS#12 code to be tidied up that duplicated this
12095 *) Changed obj_dat.pl script so it takes its input and output files on
12096 the command line. This should avoid shell escape redirection problems
12100 *) Initial support for certificate extension requests, these are included
12101 in things like Xenroll certificate requests. Included functions to allow
12102 extensions to be obtained and added.
12105 *) -crlf option to s_client and s_server for sending newlines as
12106 CRLF (as required by many protocols).
12109 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
12111 *) Install libRSAglue.a when OpenSSL is built with RSAref.
12112 [Ralf S. Engelschall]
12114 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
12115 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
12117 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
12121 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
12122 DH parameters/keys (q is lost during that conversion, but the resulting
12123 DH parameters contain its length).
12125 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
12126 much faster than DH_generate_parameters (which creates parameters
12127 where p = 2*q + 1), and also the smaller q makes DH computations
12128 much more efficient (160-bit exponentiation instead of 1024-bit
12129 exponentiation); so this provides a convenient way to support DHE
12130 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
12131 utter importance to use
12132 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12134 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12135 when such DH parameters are used, because otherwise small subgroup
12136 attacks may become possible!
12139 *) Avoid memory leak in i2d_DHparams.
12142 *) Allow the -k option to be used more than once in the enc program:
12143 this allows the same encrypted message to be read by multiple recipients.
12146 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
12147 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
12148 it will always use the numerical form of the OID, even if it has a short
12152 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
12153 method only got called if p,q,dmp1,dmq1,iqmp components were present,
12154 otherwise bn_mod_exp was called. In the case of hardware keys for example
12155 no private key components need be present and it might store extra data
12156 in the RSA structure, which cannot be accessed from bn_mod_exp.
12157 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
12158 private key operations.
12161 *) Added support for SPARC Linux.
12164 *) pem_password_cb function type incompatibly changed from
12165 typedef int pem_password_cb(char *buf, int size, int rwflag);
12167 ....(char *buf, int size, int rwflag, void *userdata);
12168 so that applications can pass data to their callbacks:
12169 The PEM[_ASN1]_{read,write}... functions and macros now take an
12170 additional void * argument, which is just handed through whenever
12171 the password callback is called.
12172 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
12174 New function SSL_CTX_set_default_passwd_cb_userdata.
12176 Compatibility note: As many C implementations push function arguments
12177 onto the stack in reverse order, the new library version is likely to
12178 interoperate with programs that have been compiled with the old
12179 pem_password_cb definition (PEM_whatever takes some data that
12180 happens to be on the stack as its last argument, and the callback
12181 just ignores this garbage); but there is no guarantee whatsoever that
12184 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
12185 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
12186 problems not only on Windows, but also on some Unix platforms.
12187 To avoid problematic command lines, these definitions are now in an
12188 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
12189 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
12192 *) MIPS III/IV assembler module is reimplemented.
12195 *) More DES library cleanups: remove references to srand/rand and
12196 delete an unused file.
12199 *) Add support for the free Netwide assembler (NASM) under Win32,
12200 since not many people have MASM (ml) and it can be hard to obtain.
12201 This is currently experimental but it seems to work OK and pass all
12202 the tests. Check out INSTALL.W32 for info.
12205 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
12206 without temporary keys kept an extra copy of the server key,
12207 and connections with temporary keys did not free everything in case
12211 *) New function RSA_check_key and new openssl rsa option -check
12212 for verifying the consistency of RSA keys.
12213 [Ulf Moeller, Bodo Moeller]
12215 *) Various changes to make Win32 compile work:
12216 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
12217 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
12218 comparison" warnings.
12219 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
12222 *) Add a debugging option to PKCS#5 v2 key generation function: when
12223 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
12224 derived keys are printed to stderr.
12227 *) Copy the flags in ASN1_STRING_dup().
12228 [Roman E. Pavlov <pre@mo.msk.ru>]
12230 *) The x509 application mishandled signing requests containing DSA
12231 keys when the signing key was also DSA and the parameters didn't match.
12233 It was supposed to omit the parameters when they matched the signing key:
12234 the verifying software was then supposed to automatically use the CA's
12235 parameters if they were absent from the end user certificate.
12237 Omitting parameters is no longer recommended. The test was also
12238 the wrong way round! This was probably due to unusual behaviour in
12239 EVP_cmp_parameters() which returns 1 if the parameters match.
12240 This meant that parameters were omitted when they *didn't* match and
12241 the certificate was useless. Certificates signed with 'ca' didn't have
12243 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
12245 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
12246 The interface is as follows:
12247 Applications can use
12248 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
12249 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
12250 "off" is now the default.
12251 The library internally uses
12252 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
12253 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
12254 to disable memory-checking temporarily.
12256 Some inconsistent states that previously were possible (and were
12257 even the default) are now avoided.
12259 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
12260 with each memory chunk allocated; this is occasionally more helpful
12261 than just having a counter.
12263 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
12265 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
12269 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
12270 which largely parallels "options", but is for changing API behaviour,
12271 whereas "options" are about protocol behaviour.
12272 Initial "mode" flags are:
12274 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
12275 a single record has been written.
12276 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
12277 retries use the same buffer location.
12278 (But all of the contents must be
12282 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
12285 *) Fix problems with no-hmac etc.
12286 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
12288 *) New functions RSA_get_default_method(), RSA_set_method() and
12289 RSA_get_method(). These allows replacement of RSA_METHODs without having
12290 to mess around with the internals of an RSA structure.
12293 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
12294 Also really enable memory leak checks in openssl.c and in some
12296 [Chad C. Mulligan, Bodo Moeller]
12298 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
12299 up the length of negative integers. This has now been simplified to just
12300 store the length when it is first determined and use it later, rather
12301 than trying to keep track of where data is copied and updating it to
12303 [Steve Henson, reported by Brien Wheeler
12304 <bwheeler@authentica-security.com>]
12306 *) Add a new function PKCS7_signatureVerify. This allows the verification
12307 of a PKCS#7 signature but with the signing certificate passed to the
12308 function itself. This contrasts with PKCS7_dataVerify which assumes the
12309 certificate is present in the PKCS#7 structure. This isn't always the
12310 case: certificates can be omitted from a PKCS#7 structure and be
12311 distributed by "out of band" means (such as a certificate database).
12314 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12315 function prototypes in pem.h, also change util/mkdef.pl to add the
12316 necessary function names.
12319 *) mk1mf.pl (used by Windows builds) did not properly read the
12320 options set by Configure in the top level Makefile, and Configure
12321 was not even able to write more than one option correctly.
12322 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12325 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12326 file to be loaded from a BIO or FILE pointer. The BIO version will
12327 for example allow memory BIOs to contain config info.
12330 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12331 Whoever hopes to achieve shared-library compatibility across versions
12332 must use this, not the compile-time macro.
12333 (Exercise 0.9.4: Which is the minimum library version required by
12335 Note: All this applies only to multi-threaded programs, others don't
12339 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12340 through a BIO pair triggered the default case, i.e.
12341 SSLerr(...,SSL_R_UNKNOWN_STATE).
12344 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12345 can use the SSL library even if none of the specific BIOs is
12349 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12350 for the encoded length.
12351 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12353 *) Add initial documentation of the X509V3 functions.
12356 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12357 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12358 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12359 secure PKCS#8 private key format with a high iteration count.
12362 *) Fix determination of Perl interpreter: A perl or perl5
12363 _directory_ in $PATH was also accepted as the interpreter.
12364 [Ralf S. Engelschall]
12366 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12367 wrong with it but it was very old and did things like calling
12368 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12369 unusual formatting.
12372 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12373 to use the new extension code.
12376 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12377 with macros. This should make it easier to change their form, add extra
12378 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12382 *) Add to configuration table a new entry that can specify an alternative
12383 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12384 according to Mark Crispin <MRC@Panda.COM>.
12388 *) DES CBC did not update the IV. Weird.
12391 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12392 Changing the behaviour of the former might break existing programs --
12393 where IV updating is needed, des_ncbc_encrypt can be used.
12396 *) When bntest is run from "make test" it drives bc to check its
12397 calculations, as well as internally checking them. If an internal check
12398 fails, it needs to cause bc to give a non-zero result or make test carries
12399 on without noticing the failure. Fixed.
12402 *) DES library cleanups.
12405 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12406 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12407 ciphers. NOTE: although the key derivation function has been verified
12408 against some published test vectors it has not been extensively tested
12409 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12413 *) Instead of "mkdir -p", which is not fully portable, use new
12414 Perl script "util/mkdir-p.pl".
12417 *) Rewrite the way password based encryption (PBE) is handled. It used to
12418 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12419 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12420 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12421 the 'parameter' field of the AlgorithmIdentifier is passed to the
12422 underlying key generation function so it must do its own ASN1 parsing.
12423 This has also changed the EVP_PBE_CipherInit() function which now has a
12424 'parameter' argument instead of literal salt and iteration count values
12425 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12428 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12429 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12430 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12431 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12432 value was just used as a "magic string" and not used directly its
12433 value doesn't matter.
12436 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12440 *) "linux-sparc64" configuration (ultrapenguin).
12441 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12442 "linux-sparc" configuration.
12443 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12445 *) config now generates no-xxx options for missing ciphers.
12448 *) Support the EBCDIC character set (work in progress).
12449 File ebcdic.c not yet included because it has a different license.
12450 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12452 *) Support BS2000/OSD-POSIX.
12453 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12455 *) Make callbacks for key generation use void * instead of char *.
12458 *) Make S/MIME samples compile (not yet tested).
12461 *) Additional typesafe stacks.
12464 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12468 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12470 *) New configuration variant "sco5-gcc".
12472 *) Updated some demos.
12473 [Sean O Riordain, Wade Scholine]
12475 *) Add missing BIO_free at exit of pkcs12 application.
12478 *) Fix memory leak in conf.c.
12481 *) Updates for Win32 to assembler version of MD5.
12484 *) Set #! path to perl in apps/der_chop to where we found it
12485 instead of using a fixed path.
12488 *) SHA library changes for irix64-mips4-cc.
12491 *) Improvements for VMS support.
12495 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12497 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12498 This also avoids the problems with SC4.2 and unpatched SC5.
12499 [Andy Polyakov <appro@fy.chalmers.se>]
12501 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12502 These are required because of the typesafe stack would otherwise break
12503 existing code. If old code used a structure member which used to be STACK
12504 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12505 sk_num or sk_value it would produce an error because the num, data members
12506 are not present in STACK_OF. Now it just produces a warning. sk_set
12507 replaces the old method of assigning a value to sk_value
12508 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12509 that does this will no longer work (and should use sk_set instead) but
12510 this could be regarded as a "questionable" behaviour anyway.
12513 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12514 correctly handle encrypted S/MIME data.
12517 *) Change type of various DES function arguments from des_cblock
12518 (which means, in function argument declarations, pointer to char)
12519 to des_cblock * (meaning pointer to array with 8 char elements),
12520 which allows the compiler to do more typechecking; it was like
12521 that back in SSLeay, but with lots of ugly casts.
12523 Introduce new type const_des_cblock.
12526 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12527 problems: find RecipientInfo structure that matches recipient certificate
12528 and initialise the ASN1 structures properly based on passed cipher.
12531 *) Belatedly make the BN tests actually check the results.
12534 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12535 to and from BNs: it was completely broken. New compilation option
12536 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12537 key elements as negative integers.
12540 *) Reorganize and speed up MD5.
12541 [Andy Polyakov <appro@fy.chalmers.se>]
12544 [Richard Levitte <richard@levitte.org>]
12546 *) New option -out to asn1parse to allow the parsed structure to be
12547 output to a file. This is most useful when combined with the -strparse
12548 option to examine the output of things like OCTET STRINGS.
12551 *) Make SSL library a little more fool-proof by not requiring any longer
12552 that SSL_set_{accept,connect}_state be called before
12553 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12554 in many applications because usually everything *appeared* to work as
12555 intended anyway -- now it really works as intended).
12558 *) Move openssl.cnf out of lib/.
12561 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12562 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12563 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12564 [Ralf S. Engelschall]
12566 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12567 handle PKCS#7 enveloped data properly.
12568 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12570 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12571 copying pointers. The cert_st handling is changed by this in
12572 various ways (and thus what used to be known as ctx->default_cert
12573 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12574 any longer when s->cert does not give us what we need).
12575 ssl_cert_instantiate becomes obsolete by this change.
12576 As soon as we've got the new code right (possibly it already is?),
12577 we have solved a couple of bugs of the earlier code where s->cert
12578 was used as if it could not have been shared with other SSL structures.
12580 Note that using the SSL API in certain dirty ways now will result
12581 in different behaviour than observed with earlier library versions:
12582 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12583 does not influence s as it used to.
12585 In order to clean up things more thoroughly, inside SSL_SESSION
12586 we don't use CERT any longer, but a new structure SESS_CERT
12587 that holds per-session data (if available); currently, this is
12588 the peer's certificate chain and, for clients, the server's certificate
12589 and temporary key. CERT holds only those values that can have
12590 meaningful defaults in an SSL_CTX.
12593 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12594 from the internal representation. Various PKCS#7 fixes: remove some
12595 evil casts and set the enc_dig_alg field properly based on the signing
12599 *) Allow PKCS#12 password to be set from the command line or the
12600 environment. Let 'ca' get its config file name from the environment
12601 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12605 *) Allow certificate policies extension to use an IA5STRING for the
12606 organization field. This is contrary to the PKIX definition but
12607 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12611 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12612 without disallowing inline assembler and the like for non-pedantic builds.
12615 *) Support Borland C++ builder.
12616 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12618 *) Support Mingw32.
12621 *) SHA-1 cleanups and performance enhancements.
12622 [Andy Polyakov <appro@fy.chalmers.se>]
12624 *) Sparc v8plus assembler for the bignum library.
12625 [Andy Polyakov <appro@fy.chalmers.se>]
12627 *) Accept any -xxx and +xxx compiler options in Configure.
12630 *) Update HPUX configuration.
12633 *) Add missing sk_<type>_unshift() function to safestack.h
12634 [Ralf S. Engelschall]
12636 *) New function SSL_CTX_use_certificate_chain_file that sets the
12637 "extra_cert"s in addition to the certificate. (This makes sense
12638 only for "PEM" format files, as chains as a whole are not
12642 *) Support verify_depth from the SSL API.
12643 x509_vfy.c had what can be considered an off-by-one-error:
12644 Its depth (which was not part of the external interface)
12645 was actually counting the number of certificates in a chain;
12646 now it really counts the depth.
12649 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12650 instead of X509err, which often resulted in confusing error
12651 messages since the error codes are not globally unique
12652 (e.g. an alleged error in ssl3_accept when a certificate
12653 didn't match the private key).
12655 *) New function SSL_CTX_set_session_id_context that allows to set a default
12656 value (so that you don't need SSL_set_session_id_context for each
12657 connection using the SSL_CTX).
12660 *) OAEP decoding bug fix.
12663 *) Support INSTALL_PREFIX for package builders, as proposed by
12667 *) New Configure options "threads" and "no-threads". For systems
12668 where the proper compiler options are known (currently Solaris
12669 and Linux), "threads" is the default.
12672 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12675 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12676 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12677 such as /usr/local/bin.
12680 *) "make linux-shared" to build shared libraries.
12681 [Niels Poppe <niels@netbox.org>]
12683 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12686 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12687 extension adding in x509 utility.
12690 *) Remove NOPROTO sections and error code comments.
12693 *) Partial rewrite of the DEF file generator to now parse the ANSI
12697 *) New Configure options --prefix=DIR and --openssldir=DIR.
12700 *) Complete rewrite of the error code script(s). It is all now handled
12701 by one script at the top level which handles error code gathering,
12702 header rewriting and C source file generation. It should be much better
12703 than the old method: it now uses a modified version of Ulf's parser to
12704 read the ANSI prototypes in all header files (thus the old K&R definitions
12705 aren't needed for error creation any more) and do a better job of
12706 translating function codes into names. The old 'ASN1 error code embedded
12707 in a comment' is no longer necessary and it doesn't use .err files which
12708 have now been deleted. Also the error code call doesn't have to appear all
12709 on one line (which resulted in some large lines...).
12712 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12715 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12716 0 (which usually indicates a closed connection), but continue reading.
12719 *) Fix some race conditions.
12722 *) Add support for CRL distribution points extension. Add Certificate
12723 Policies and CRL distribution points documentation.
12726 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12729 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12730 8 of keying material. Merlin has also confirmed interop with this fix
12731 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12732 [Merlin Hughes <merlin@baltimore.ie>]
12734 *) Fix lots of warnings.
12735 [Richard Levitte <levitte@stacken.kth.se>]
12737 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12738 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12739 [Richard Levitte <levitte@stacken.kth.se>]
12741 *) Fix problems with sizeof(long) == 8.
12742 [Andy Polyakov <appro@fy.chalmers.se>]
12744 *) Change functions to ANSI C.
12747 *) Fix typos in error codes.
12748 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12750 *) Remove defunct assembler files from Configure.
12753 *) SPARC v8 assembler BIGNUM implementation.
12754 [Andy Polyakov <appro@fy.chalmers.se>]
12756 *) Support for Certificate Policies extension: both print and set.
12757 Various additions to support the r2i method this uses.
12760 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12761 return a const string when you are expecting an allocated buffer.
12764 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12765 types DirectoryString and DisplayText.
12768 *) Add code to allow r2i extensions to access the configuration database,
12769 add an LHASH database driver and add several ctx helper functions.
12772 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12773 fail when they extended the size of a BIGNUM.
12776 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12777 support typesafe stack.
12780 *) Fix typo in SSL_[gs]et_options().
12781 [Nils Frostberg <nils@medcom.se>]
12783 *) Delete various functions and files that belonged to the (now obsolete)
12784 old X509V3 handling code.
12787 *) New Configure option "rsaref".
12790 *) Don't auto-generate pem.h.
12793 *) Introduce type-safe ASN.1 SETs.
12796 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12797 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12799 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12800 that links with OpenSSL (well at least cause lots of warnings), but fear
12801 not: the conversion is trivial, and it eliminates loads of evil casts. A
12802 few STACKed things have been converted already. Feel free to convert more.
12803 In the fullness of time, I'll do away with the STACK type altogether.
12806 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12807 specified in <certfile> by updating the entry in the index.txt file.
12808 This way one no longer has to edit the index.txt file manually for
12809 revoking a certificate. The -revoke option does the gory details now.
12810 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12812 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12813 `-text' option at all and this way the `-noout -text' combination was
12814 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12815 [Ralf S. Engelschall]
12817 *) Make sure a corresponding plain text error message exists for the
12818 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12819 verify callback function determined that a certificate was revoked.
12820 [Ralf S. Engelschall]
12822 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12823 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12824 all available ciphers including rc5, which was forgotten until now.
12825 In order to let the testing shell script know which algorithms
12826 are available, a new (up to now undocumented) command
12827 "openssl list-cipher-commands" is used.
12830 *) Bugfix: s_client occasionally would sleep in select() when
12831 it should have checked SSL_pending() first.
12834 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12835 the raw DSA values prior to ASN.1 encoding.
12838 *) Tweaks to Configure
12839 [Niels Poppe <niels@netbox.org>]
12841 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12845 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12848 *) New config option to avoid instructions that are illegal on the 80386.
12849 The default code is faster, but requires at least a 486.
12852 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12853 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12854 same as SSL2_VERSION anyway.
12857 *) New "-showcerts" option for s_client.
12860 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12861 application. Various cleanups and fixes.
12864 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12865 modify error routines to work internally. Add error codes and PBE init
12866 to library startup routines.
12869 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12870 packing functions to asn1 and evp. Changed function names and error
12871 codes along the way.
12874 *) PKCS12 integration: and so it begins... First of several patches to
12875 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12876 objects to objects.h
12879 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12880 and display support for Thawte strong extranet extension.
12883 *) Add LinuxPPC support.
12884 [Jeff Dubrule <igor@pobox.org>]
12886 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12887 bn_div_words in alpha.s.
12888 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12890 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12891 OAEP isn't supported when OpenSSL is built with RSAref.
12892 [Ulf Moeller <ulf@fitug.de>]
12894 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12895 so they no longer are missing under -DNOPROTO.
12896 [Soren S. Jorvang <soren@t.dk>]
12899 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12901 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12902 doesn't work when the session is reused. Coming soon!
12905 *) Fix a security hole, that allows sessions to be reused in the wrong
12906 context thus bypassing client cert protection! All software that uses
12907 client certs and session caches in multiple contexts NEEDS PATCHING to
12908 allow session reuse! A fuller solution is in the works.
12909 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12911 *) Some more source tree cleanups (removed obsolete files
12912 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12913 permission on "config" script to be executable) and a fix for the INSTALL
12915 [Ulf Moeller <ulf@fitug.de>]
12917 *) Remove some legacy and erroneous uses of malloc, free instead of
12919 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12921 *) Make rsa_oaep_test return non-zero on error.
12922 [Ulf Moeller <ulf@fitug.de>]
12924 *) Add support for native Solaris shared libraries. Configure
12925 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12926 if someone would make that last step automatic.
12927 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12929 *) ctx_size was not built with the right compiler during "make links". Fixed.
12932 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12933 except NULL ciphers". This means the default cipher list will no longer
12934 enable NULL ciphers. They need to be specifically enabled e.g. with
12935 the string "DEFAULT:eNULL".
12938 *) Fix to RSA private encryption routines: if p < q then it would
12939 occasionally produce an invalid result. This will only happen with
12940 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12943 *) Be less restrictive and allow also `perl util/perlpath.pl
12944 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12945 because this way one can also use an interpreter named `perl5' (which is
12946 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12947 installed as `perl').
12948 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12950 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12951 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12953 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12954 advapi32.lib to Win32 build and change the pem test comparison
12955 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12956 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12957 and crypto/des/ede_cbcm_enc.c.
12960 *) DES quad checksum was broken on big-endian architectures. Fixed.
12963 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12964 Win32 test batch file so it (might) work again. The Win32 test batch file
12965 is horrible: I feel ill....
12968 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12969 in e_os.h. Audit of header files to check ANSI and non ANSI
12970 sections: 10 functions were absent from non ANSI section and not exported
12971 from Windows DLLs. Fixed up libeay.num for new functions.
12974 *) Make `openssl version' output lines consistent.
12975 [Ralf S. Engelschall]
12977 *) Fix Win32 symbol export lists for BIO functions: Added
12978 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12979 to ms/libeay{16,32}.def.
12980 [Ralf S. Engelschall]
12982 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12983 fine under Unix and passes some trivial tests I've now added. But the
12984 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12985 added to make sure no one expects that this stuff really works in the
12986 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12987 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12989 [Ralf S. Engelschall]
12991 *) Fix the generation of two part addresses in perl.
12992 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12994 *) Add config entry for Linux on MIPS.
12995 [John Tobey <jtobey@channel1.com>]
12997 *) Make links whenever Configure is run, unless we are on Windoze.
13000 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
13001 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
13005 *) Add a useful kludge to allow package maintainers to specify compiler and
13006 other platforms details on the command line without having to patch the
13007 Configure script every time: One now can use ``perl Configure
13008 <id>:<details>'', i.e. platform ids are allowed to have details appended
13009 to them (separated by colons). This is treated as there would be a static
13010 pre-configured entry in Configure's %table under key <id> with value
13011 <details> and ``perl Configure <id>'' is called. So, when you want to
13012 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
13013 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
13014 now, which overrides the FreeBSD-elf entry on-the-fly.
13015 [Ralf S. Engelschall]
13017 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
13020 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
13021 on the `perl Configure ...' command line. This way one can compile
13022 OpenSSL libraries with Position Independent Code (PIC) which is needed
13023 for linking it into DSOs.
13024 [Ralf S. Engelschall]
13026 *) Remarkably, export ciphers were totally broken and no-one had noticed!
13030 *) Cleaned up the LICENSE document: The official contact for any license
13031 questions now is the OpenSSL core team under openssl-core@openssl.org.
13032 And add a paragraph about the dual-license situation to make sure people
13033 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
13034 to the OpenSSL toolkit.
13035 [Ralf S. Engelschall]
13037 *) General source tree makefile cleanups: Made `making xxx in yyy...'
13038 display consistent in the source tree and replaced `/bin/rm' by `rm'.
13039 Additionally cleaned up the `make links' target: Remove unnecessary
13040 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
13041 to speed processing and no longer clutter the display with confusing
13042 stuff. Instead only the actually done links are displayed.
13043 [Ralf S. Engelschall]
13045 *) Permit null encryption ciphersuites, used for authentication only. It used
13046 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
13047 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
13051 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
13052 signed attributes when verifying signatures (this would break them),
13053 the detached data encoding was wrong and public keys obtained using
13054 X509_get_pubkey() weren't freed.
13057 *) Add text documentation for the BUFFER functions. Also added a work around
13058 to a Win95 console bug. This was triggered by the password read stuff: the
13059 last character typed gets carried over to the next fread(). If you were
13060 generating a new cert request using 'req' for example then the last
13061 character of the passphrase would be CR which would then enter the first
13065 *) Added the new `Includes OpenSSL Cryptography Software' button as
13066 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
13067 button and can be used by applications based on OpenSSL to show the
13068 relationship to the OpenSSL project.
13069 [Ralf S. Engelschall]
13071 *) Remove confusing variables in function signatures in files
13072 ssl/ssl_lib.c and ssl/ssl.h.
13073 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13075 *) Don't install bss_file.c under PREFIX/include/
13076 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13078 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
13079 functions that return function pointers and has support for NT specific
13080 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
13081 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
13082 unsigned to signed types: this was killing the Win32 compile.
13085 *) Add new certificate file to stack functions,
13086 SSL_add_dir_cert_subjects_to_stack() and
13087 SSL_add_file_cert_subjects_to_stack(). These largely supplant
13088 SSL_load_client_CA_file(), and can be used to add multiple certs easily
13089 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
13090 This means that Apache-SSL and similar packages don't have to mess around
13091 to add as many CAs as they want to the preferred list.
13094 *) Experiment with doxygen documentation. Currently only partially applied to
13096 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
13097 openssl.doxy as the configuration file.
13100 *) Get rid of remaining C++-style comments which strict C compilers hate.
13101 [Ralf S. Engelschall, pointed out by Carlos Amengual]
13103 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
13104 compiled in by default: it has problems with large keys.
13107 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
13108 DH private keys and/or callback functions which directly correspond to
13109 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
13110 is needed for applications which have to configure certificates on a
13111 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
13113 For the RSA certificate situation is makes no difference, but
13114 for the DSA certificate situation this fixes the "no shared cipher"
13115 problem where the OpenSSL cipher selection procedure failed because the
13116 temporary keys were not overtaken from the context and the API provided
13117 no way to reconfigure them.
13118 The new functions now let applications reconfigure the stuff and they
13119 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
13120 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
13121 non-public-API function ssl_cert_instantiate() is used as a helper
13122 function and also to reduce code redundancy inside ssl_rsa.c.
13123 [Ralf S. Engelschall]
13125 *) Move s_server -dcert and -dkey options out of the undocumented feature
13126 area because they are useful for the DSA situation and should be
13127 recognized by the users.
13128 [Ralf S. Engelschall]
13130 *) Fix the cipher decision scheme for export ciphers: the export bits are
13131 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
13132 SSL_EXP_MASK. So, the original variable has to be used instead of the
13133 already masked variable.
13134 [Richard Levitte <levitte@stacken.kth.se>]
13136 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
13137 [Richard Levitte <levitte@stacken.kth.se>]
13139 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
13140 from `int' to `unsigned int' because it's a length and initialized by
13141 EVP_DigestFinal() which expects an `unsigned int *'.
13142 [Richard Levitte <levitte@stacken.kth.se>]
13144 *) Don't hard-code path to Perl interpreter on shebang line of Configure
13145 script. Instead use the usual Shell->Perl transition trick.
13146 [Ralf S. Engelschall]
13148 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
13149 (in addition to RSA certificates) to match the behaviour of `openssl dsa
13150 -noout -modulus' as it's already the case for `openssl rsa -noout
13151 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
13152 currently the public key is printed (a decision which was already done by
13153 `openssl dsa -modulus' in the past) which serves a similar purpose.
13154 Additionally the NO_RSA no longer completely removes the whole -modulus
13155 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
13157 [Ralf S. Engelschall]
13159 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
13160 BIO. See the source (crypto/evp/bio_ok.c) for more info.
13161 [Arne Ansper <arne@ats.cyber.ee>]
13163 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
13164 to be added. Now both 'req' and 'ca' can use new objects defined in the
13168 *) Add cool BIO that does syslog (or event log on NT).
13169 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
13171 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
13172 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
13173 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
13174 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
13177 *) Add preliminary config info for new extension code.
13180 *) Make RSA_NO_PADDING really use no padding.
13181 [Ulf Moeller <ulf@fitug.de>]
13183 *) Generate errors when private/public key check is done.
13186 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
13187 for some CRL extensions and new objects added.
13190 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
13191 key usage extension and fuller support for authority key id.
13194 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
13195 padding method for RSA, which is recommended for new applications in PKCS
13196 #1 v2.0 (RFC 2437, October 1998).
13197 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
13198 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
13199 against Bleichbacher's attack on RSA.
13200 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
13203 *) Updates to the new SSL compression code
13204 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13206 *) Fix so that the version number in the master secret, when passed
13207 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
13208 (because the server will not accept higher), that the version number
13209 is 0x03,0x01, not 0x03,0x00
13210 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13212 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
13213 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
13214 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
13217 *) Support for RAW extensions where an arbitrary extension can be
13218 created by including its DER encoding. See apps/openssl.cnf for
13222 *) Make sure latest Perl versions don't interpret some generated C array
13223 code as Perl array code in the crypto/err/err_genc.pl script.
13224 [Lars Weber <3weber@informatik.uni-hamburg.de>]
13226 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
13227 not many people have the assembler. Various Win32 compilation fixes and
13228 update to the INSTALL.W32 file with (hopefully) more accurate Win32
13229 build instructions.
13232 *) Modify configure script 'Configure' to automatically create crypto/date.h
13233 file under Win32 and also build pem.h from pem.org. New script
13234 util/mkfiles.pl to create the MINFO file on environments that can't do a
13235 'make files': perl util/mkfiles.pl >MINFO should work.
13238 *) Major rework of DES function declarations, in the pursuit of correctness
13239 and purity. As a result, many evil casts evaporated, and some weirdness,
13240 too. You may find this causes warnings in your code. Zapping your evil
13241 casts will probably fix them. Mostly.
13244 *) Fix for a typo in asn1.h. Bug fix to object creation script
13245 obj_dat.pl. It considered a zero in an object definition to mean
13246 "end of object": none of the objects in objects.h have any zeros
13247 so it wasn't spotted.
13248 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
13250 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
13251 Masking (CBCM). In the absence of test vectors, the best I have been able
13252 to do is check that the decrypt undoes the encrypt, so far. Send me test
13253 vectors if you have them.
13256 *) Correct calculation of key length for export ciphers (too much space was
13257 allocated for null ciphers). This has not been tested!
13260 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
13261 message is now correct (it understands "crypto" and "ssl" on its
13262 command line). There is also now an "update" option. This will update
13263 the util/ssleay.num and util/libeay.num files with any new functions.
13265 perl util/mkdef.pl crypto ssl update
13266 it will update them.
13269 *) Overhauled the Perl interface (perl/*):
13270 - ported BN stuff to OpenSSL's different BN library
13271 - made the perl/ source tree CVS-aware
13272 - renamed the package from SSLeay to OpenSSL (the files still contain
13273 their history because I've copied them in the repository)
13274 - removed obsolete files (the test scripts will be replaced
13275 by better Test::Harness variants in the future)
13276 [Ralf S. Engelschall]
13278 *) First cut for a very conservative source tree cleanup:
13279 1. merge various obsolete readme texts into doc/ssleay.txt
13280 where we collect the old documents and readme texts.
13281 2. remove the first part of files where I'm already sure that we no
13282 longer need them because of three reasons: either they are just temporary
13283 files which were left by Eric or they are preserved original files where
13284 I've verified that the diff is also available in the CVS via "cvs diff
13285 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
13286 the crypto/md/ stuff).
13287 [Ralf S. Engelschall]
13289 *) More extension code. Incomplete support for subject and issuer alt
13290 name, issuer and authority key id. Change the i2v function parameters
13291 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
13292 what that's for :-) Fix to ASN1 macro which messed up
13293 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
13296 *) Preliminary support for ENUMERATED type. This is largely copied from the
13300 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
13301 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13303 *) Make sure `make rehash' target really finds the `openssl' program.
13304 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13306 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
13307 like to hear about it if this slows down other processors.
13310 *) Add CygWin32 platform information to Configure script.
13311 [Alan Batie <batie@aahz.jf.intel.com>]
13313 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
13314 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13316 *) New program nseq to manipulate netscape certificate sequences
13319 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13323 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13324 but the BN code had some problems that would cause failures when
13325 doing certificate verification and some other functions.
13326 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13328 *) Add ASN1 and PEM code to support netscape certificate sequences.
13331 *) Add ASN1 and PEM code to support netscape certificate sequences.
13334 *) Add several PKIX and private extended key usage OIDs.
13337 *) Modify the 'ca' program to handle the new extension code. Modify
13338 openssl.cnf for new extension format, add comments.
13341 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13342 and add a sample to openssl.cnf so req -x509 now adds appropriate
13346 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13347 error code, add initial support to X509_print() and x509 application.
13350 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13351 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13352 stuff is currently isolated and isn't even compiled yet.
13355 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13356 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13357 Removed the versions check from X509 routines when loading extensions:
13358 this allows certain broken certificates that don't set the version
13359 properly to be processed.
13362 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13363 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13364 can still be regenerated with "make depend".
13367 *) Spelling mistake in C version of CAST-128.
13368 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13370 *) Changes to the error generation code. The perl script err-code.pl
13371 now reads in the old error codes and retains the old numbers, only
13372 adding new ones if necessary. It also only changes the .err files if new
13373 codes are added. The makefiles have been modified to only insert errors
13374 when needed (to avoid needlessly modifying header files). This is done
13375 by only inserting errors if the .err file is newer than the auto generated
13376 C file. To rebuild all the error codes from scratch (the old behaviour)
13377 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13378 or delete all the .err files.
13381 *) CAST-128 was incorrectly implemented for short keys. The C version has
13382 been fixed, but is untested. The assembler versions are also fixed, but
13383 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13384 to regenerate it if needed.
13385 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13386 Hagino <itojun@kame.net>]
13388 *) File was opened incorrectly in randfile.c.
13389 [Ulf Möller <ulf@fitug.de>]
13391 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13392 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13393 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13394 al: it's just almost always a UTCTime. Note this patch adds new error
13395 codes so do a "make errors" if there are problems.
13398 *) Correct Linux 1 recognition in config.
13399 [Ulf Möller <ulf@fitug.de>]
13401 *) Remove pointless MD5 hash when using DSA keys in ca.
13402 [Anonymous <nobody@replay.com>]
13404 *) Generate an error if given an empty string as a cert directory. Also
13405 generate an error if handed NULL (previously returned 0 to indicate an
13406 error, but didn't set one).
13407 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13409 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13412 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13413 parameters. This was causing a warning which killed off the Win32 compile.
13416 *) Remove C++ style comments from crypto/bn/bn_local.h.
13417 [Neil Costigan <neil.costigan@celocom.com>]
13419 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13420 based on a text string, looking up short and long names and finally
13421 "dot" format. The "dot" format stuff didn't work. Added new function
13422 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13423 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13424 OID is not part of the table.
13427 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13428 X509_LOOKUP_by_alias().
13431 *) Sort openssl functions by name.
13434 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13435 encryption from sample DSA keys (in case anyone is interested the password
13439 *) Make _all_ *_free functions accept a NULL pointer.
13440 [Frans Heymans <fheymans@isaserver.be>]
13442 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13444 [Anonymous <nobody@replay.com>]
13446 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13447 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13449 *) Don't blow it for numeric -newkey arguments to apps/req.
13450 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13452 *) Temp key "for export" tests were wrong in s3_srvr.c.
13453 [Anonymous <nobody@replay.com>]
13455 *) Add prototype for temp key callback functions
13456 SSL_CTX_set_tmp_{rsa,dh}_callback().
13459 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13460 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13463 *) X509_name_add_entry() freed the wrong thing after an error.
13464 [Arne Ansper <arne@ats.cyber.ee>]
13466 *) rsa_eay.c would attempt to free a NULL context.
13467 [Arne Ansper <arne@ats.cyber.ee>]
13469 *) BIO_s_socket() had a broken should_retry() on Windoze.
13470 [Arne Ansper <arne@ats.cyber.ee>]
13472 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13473 [Arne Ansper <arne@ats.cyber.ee>]
13475 *) Make sure the already existing X509_STORE->depth variable is initialized
13476 in X509_STORE_new(), but document the fact that this variable is still
13477 unused in the certificate verification process.
13478 [Ralf S. Engelschall]
13480 *) Fix the various library and apps files to free up pkeys obtained from
13481 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13484 *) Fix reference counting in X509_PUBKEY_get(). This makes
13485 demos/maurice/example2.c work, amongst others, probably.
13486 [Steve Henson and Ben Laurie]
13488 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13489 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13490 are no longer created. This way we have a single and consistent command
13491 line interface `openssl <command>', similar to `cvs <command>'.
13492 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13494 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13495 BIT STRING wrapper always have zero unused bits.
13498 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13501 *) Make the top-level INSTALL documentation easier to understand.
13504 *) Makefiles updated to exit if an error occurs in a sub-directory
13505 make (including if user presses ^C) [Paul Sutton]
13507 *) Make Montgomery context stuff explicit in RSA data structure.
13510 *) Fix build order of pem and err to allow for generated pem.h.
13513 *) Fix renumbering bug in X509_NAME_delete_entry().
13516 *) Enhanced the err-ins.pl script so it makes the error library number
13517 global and can add a library name. This is needed for external ASN1 and
13518 other error libraries.
13521 *) Fixed sk_insert which never worked properly.
13524 *) Fix ASN1 macros so they can handle indefinite length constructed
13525 EXPLICIT tags. Some non standard certificates use these: they can now
13529 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13530 into a single doc/ssleay.txt bundle. This way the information is still
13531 preserved but no longer messes up this directory. Now it's new room for
13532 the new set of documentation files.
13533 [Ralf S. Engelschall]
13535 *) SETs were incorrectly DER encoded. This was a major pain, because they
13536 shared code with SEQUENCEs, which aren't coded the same. This means that
13537 almost everything to do with SETs or SEQUENCEs has either changed name or
13538 number of arguments.
13539 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13541 *) Fix test data to work with the above.
13544 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13545 was already fixed by Eric for 0.9.1 it seems.
13546 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13548 *) Autodetect FreeBSD3.
13551 *) Fix various bugs in Configure. This affects the following platforms:
13555 unixware-2.0-pentium
13559 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13560 before they are needed.
13563 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13567 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13569 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13570 changed SSLeay to OpenSSL in version strings.
13571 [Ralf S. Engelschall]
13573 *) Some fixups to the top-level documents.
13576 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13577 because the symlink to include/ was missing.
13578 [Ralf S. Engelschall]
13580 *) Incorporated the popular no-RSA/DSA-only patches
13581 which allow to compile a RSA-free SSLeay.
13582 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13584 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13585 when "ssleay" is still not found.
13586 [Ralf S. Engelschall]
13588 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13589 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13591 *) Updated the README file.
13592 [Ralf S. Engelschall]
13594 *) Added various .cvsignore files in the CVS repository subdirs
13595 to make a "cvs update" really silent.
13596 [Ralf S. Engelschall]
13598 *) Recompiled the error-definition header files and added
13599 missing symbols to the Win32 linker tables.
13600 [Ralf S. Engelschall]
13602 *) Cleaned up the top-level documents;
13603 o new files: CHANGES and LICENSE
13604 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13605 o merged COPYRIGHT into LICENSE
13606 o removed obsolete TODO file
13607 o renamed MICROSOFT to INSTALL.W32
13608 [Ralf S. Engelschall]
13610 *) Removed dummy files from the 0.9.1b source tree:
13611 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13612 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13613 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13614 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13615 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13616 [Ralf S. Engelschall]
13618 *) Added various platform portability fixes.
13621 *) The Genesis of the OpenSSL rpject:
13622 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13623 Young and Tim J. Hudson created while they were working for C2Net until
13625 [The OpenSSL Project]
13628 Changes between 0.9.0b and 0.9.1b [not released]
13630 *) Updated a few CA certificates under certs/
13633 *) Changed some BIGNUM api stuff.
13636 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13637 DGUX x86, Linux Alpha, etc.
13640 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13641 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13645 *) Add -strparse option to asn1pars program which parses nested
13647 [Dr Stephen Henson <shenson@bigfoot.com>]
13649 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13652 *) DSA fix for "ca" program.
13655 *) Added "-genkey" option to "dsaparam" program.
13658 *) Added RIPE MD160 (rmd160) message digest.
13661 *) Added -a (all) option to "ssleay version" command.
13664 *) Added PLATFORM define which is the id given to Configure.
13667 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13670 *) Extended the ASN.1 parser routines.
13673 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13676 *) Added a BN_CTX to the BN library.
13679 *) Fixed the weak key values in DES library
13682 *) Changed API in EVP library for cipher aliases.
13685 *) Added support for RC2/64bit cipher.
13688 *) Converted the lhash library to the crypto/mem.c functions.
13691 *) Added more recognized ASN.1 object ids.
13694 *) Added more RSA padding checks for SSL/TLS.
13697 *) Added BIO proxy/filter functionality.
13700 *) Added extra_certs to SSL_CTX which can be used
13701 send extra CA certificates to the client in the CA cert chain sending
13702 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13705 *) Now Fortezza is denied in the authentication phase because
13706 this is key exchange mechanism is not supported by SSLeay at all.
13709 *) Additional PKCS1 checks.
13712 *) Support the string "TLSv1" for all TLS v1 ciphers.
13715 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13716 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13719 *) Fixed a few memory leaks.
13722 *) Fixed various code and comment typos.
13725 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13726 bytes sent in the client random.
13727 [Edward Bishop <ebishop@spyglass.com>]