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 1.1.1q and 1.1.1r [xx XXX xxxx]
14 Changes between 1.1.1p and 1.1.1q [5 Jul 2022]
16 *) AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
17 implementation would not encrypt the entirety of the data under some
18 circumstances. This could reveal sixteen bytes of data that was
19 preexisting in the memory that wasn't written. In the special case of
20 "in place" encryption, sixteen bytes of the plaintext would be revealed.
22 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
23 they are both unaffected.
25 [Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño]
27 Changes between 1.1.1o and 1.1.1p [21 Jun 2022]
29 *) In addition to the c_rehash shell command injection identified in
30 CVE-2022-1292, further bugs where the c_rehash script does not
31 properly sanitise shell metacharacters to prevent command injection have been
34 When the CVE-2022-1292 was fixed it was not discovered that there
35 are other places in the script where the file names of certificates
36 being hashed were possibly passed to a command executed through the shell.
38 This script is distributed by some operating systems in a manner where
39 it is automatically executed. On such operating systems, an attacker
40 could execute arbitrary commands with the privileges of the script.
42 Use of the c_rehash script is considered obsolete and should be replaced
43 by the OpenSSL rehash command line tool.
45 [Daniel Fiala, Tomáš Mráz]
47 *) When OpenSSL TLS client is connecting without any supported elliptic
48 curves and TLS-1.3 protocol is disabled the connection will no longer fail
49 if a ciphersuite that does not use a key exchange based on elliptic
50 curves can be negotiated.
53 Changes between 1.1.1n and 1.1.1o [3 May 2022]
55 *) Fixed a bug in the c_rehash script which was not properly sanitising shell
56 metacharacters to prevent command injection. This script is distributed
57 by some operating systems in a manner where it is automatically executed.
58 On such operating systems, an attacker could execute arbitrary commands
59 with the privileges of the script.
61 Use of the c_rehash script is considered obsolete and should be replaced
62 by the OpenSSL rehash command line tool.
66 Changes between 1.1.1m and 1.1.1n [15 Mar 2022]
68 *) Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
71 Internally this function is used when parsing certificates that contain
72 elliptic curve public keys in compressed form or explicit elliptic curve
73 parameters with a base point encoded in compressed form.
75 It is possible to trigger the infinite loop by crafting a certificate that
76 has invalid explicit curve parameters.
78 Since certificate parsing happens prior to verification of the certificate
79 signature, any process that parses an externally supplied certificate may
80 thus be subject to a denial of service attack. The infinite loop can also
81 be reached when parsing crafted private keys as they can contain explicit
82 elliptic curve parameters.
84 Thus vulnerable situations include:
86 - TLS clients consuming server certificates
87 - TLS servers consuming client certificates
88 - Hosting providers taking certificates or private keys from customers
89 - Certificate authorities parsing certification requests from subscribers
90 - Anything else which parses ASN.1 elliptic curve parameters
92 Also any other applications that use the BN_mod_sqrt() where the attacker
93 can control the parameter values are vulnerable to this DoS issue.
97 *) Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
98 to the list of ciphersuites providing Perfect Forward Secrecy as
99 required by SECLEVEL >= 3.
101 [Dmitry Belyavskiy, Nicola Tuveri]
103 Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
105 *) Avoid loading of a dynamic engine twice.
109 *) Fixed building on Debian with kfreebsd kernels
113 *) Prioritise DANE TLSA issuer certs over peer certs
117 *) Fixed random API for MacOS prior to 10.12
119 These MacOS versions don't support the CommonCrypto APIs
123 Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
125 *) Fixed an SM2 Decryption Buffer Overflow.
127 In order to decrypt SM2 encrypted data an application is expected to call the
128 API function EVP_PKEY_decrypt(). Typically an application will call this
129 function twice. The first time, on entry, the "out" parameter can be NULL and,
130 on exit, the "outlen" parameter is populated with the buffer size required to
131 hold the decrypted plaintext. The application can then allocate a sufficiently
132 sized buffer and call EVP_PKEY_decrypt() again, but this time passing a non-NULL
133 value for the "out" parameter.
135 A bug in the implementation of the SM2 decryption code means that the
136 calculation of the buffer size required to hold the plaintext returned by the
137 first call to EVP_PKEY_decrypt() can be smaller than the actual size required by
138 the second call. This can lead to a buffer overflow when EVP_PKEY_decrypt() is
139 called by the application a second time with a buffer that is too small.
141 A malicious attacker who is able present SM2 content for decryption to an
142 application could cause attacker chosen data to overflow the buffer by up to a
143 maximum of 62 bytes altering the contents of other data held after the
144 buffer, possibly changing application behaviour or causing the application to
145 crash. The location of the buffer is application dependent but is typically
150 *) Fixed various read buffer overruns processing ASN.1 strings
152 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
153 structure which contains a buffer holding the string data and a field holding
154 the buffer length. This contrasts with normal C strings which are repesented as
155 a buffer for the string data which is terminated with a NUL (0) byte.
157 Although not a strict requirement, ASN.1 strings that are parsed using OpenSSL's
158 own "d2i" functions (and other similar parsing functions) as well as any string
159 whose value has been set with the ASN1_STRING_set() function will additionally
160 NUL terminate the byte array in the ASN1_STRING structure.
162 However, it is possible for applications to directly construct valid ASN1_STRING
163 structures which do not NUL terminate the byte array by directly setting the
164 "data" and "length" fields in the ASN1_STRING array. This can also happen by
165 using the ASN1_STRING_set0() function.
167 Numerous OpenSSL functions that print ASN.1 data have been found to assume that
168 the ASN1_STRING byte array will be NUL terminated, even though this is not
169 guaranteed for strings that have been directly constructed. Where an application
170 requests an ASN.1 structure to be printed, and where that ASN.1 structure
171 contains ASN1_STRINGs that have been directly constructed by the application
172 without NUL terminating the "data" field, then a read buffer overrun can occur.
174 The same thing can also occur during name constraints processing of certificates
175 (for example if a certificate has been directly constructed by the application
176 instead of loading it via the OpenSSL parsing functions, and the certificate
177 contains non NUL terminated ASN1_STRING structures). It can also occur in the
178 X509_get1_email(), X509_REQ_get1_email() and X509_get1_ocsp() functions.
180 If a malicious actor can cause an application to directly construct an
181 ASN1_STRING and then process it through one of the affected OpenSSL functions
182 then this issue could be hit. This might result in a crash (causing a Denial of
183 Service attack). It could also result in the disclosure of private memory
184 contents (such as private keys, or sensitive plaintext).
188 Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
190 *) Fixed a problem with verifying a certificate chain when using the
191 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks
192 of the certificates present in a certificate chain. It is not set by
195 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
196 the chain that have explicitly encoded elliptic curve parameters was added
197 as an additional strict check.
199 An error in the implementation of this check meant that the result of a
200 previous check to confirm that certificates in the chain are valid CA
201 certificates was overwritten. This effectively bypasses the check
202 that non-CA certificates must not be able to issue other certificates.
204 If a "purpose" has been configured then there is a subsequent opportunity
205 for checks that the certificate is a valid CA. All of the named "purpose"
206 values implemented in libcrypto perform this check. Therefore, where
207 a purpose is set the certificate chain will still be rejected even when the
208 strict flag has been used. A purpose is set by default in libssl client and
209 server certificate verification routines, but it can be overridden or
210 removed by an application.
212 In order to be affected, an application must explicitly set the
213 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
214 for the certificate verification or, in the case of TLS client or server
215 applications, override the default purpose.
219 *) Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
220 crafted renegotiation ClientHello message from a client. If a TLSv1.2
221 renegotiation ClientHello omits the signature_algorithms extension (where
222 it was present in the initial ClientHello), but includes a
223 signature_algorithms_cert extension then a NULL pointer dereference will
224 result, leading to a crash and a denial of service attack.
226 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
227 (which is the default configuration). OpenSSL TLS clients are not impacted
230 [Peter Kästle and Samuel Sapalski]
232 Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
234 *) Fixed the X509_issuer_and_serial_hash() function. It attempts to
235 create a unique hash value based on the issuer and serial number data
236 contained within an X509 certificate. However it was failing to correctly
237 handle any errors that may occur while parsing the issuer field (which might
238 occur if the issuer field is maliciously constructed). This may subsequently
239 result in a NULL pointer deref and a crash leading to a potential denial of
244 *) Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
245 padding mode to correctly check for rollback attacks. This is considered a
246 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
250 *) Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
251 functions. Previously they could overflow the output length argument in some
252 cases where the input length is close to the maximum permissable length for
253 an integer on the platform. In such cases the return value from the function
254 call would be 1 (indicating success), but the output length value would be
255 negative. This could cause applications to behave incorrectly or crash.
259 *) Fixed SRP_Calc_client_key so that it runs in constant time. The previous
260 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
261 could be exploited in a side channel attack to recover the password. Since
262 the attack is local host only this is outside of the current OpenSSL
263 threat model and therefore no CVE is assigned.
265 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
269 Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
271 *) Fixed NULL pointer deref in the GENERAL_NAME_cmp function
272 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
273 If an attacker can control both items being compared then this could lead
274 to a possible denial of service attack. OpenSSL itself uses the
275 GENERAL_NAME_cmp function for two purposes:
276 1) Comparing CRL distribution point names between an available CRL and a
277 CRL distribution point embedded in an X509 certificate
278 2) When verifying that a timestamp response token signer matches the
279 timestamp authority name (exposed via the API functions
280 TS_RESP_verify_response and TS_RESP_verify_token)
284 *) Add support for Apple Silicon M1 Macs with the darwin64-arm64-cc target.
287 *) The security callback, which can be customised by application code, supports
288 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
289 in the "other" parameter. In most places this is what is passed. All these
290 places occur server side. However there was one client side call of this
291 security operation and it passed a DH object instead. This is incorrect
292 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
293 of the other locations. Therefore this client side call has been changed to
294 pass an EVP_PKEY instead.
297 *) In 1.1.1h, an expired trusted (root) certificate was not anymore rejected
298 when validating a certificate path. This check is restored in 1.1.1i.
301 Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
303 *) Certificates with explicit curve parameters are now disallowed in
304 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
307 *) The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
308 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
309 conversely, silently ignore DTLS protocol version bounds when configuring
310 TLS-based contexts. The commands can be repeated to set bounds of both
311 types. The same applies with the corresponding "min_protocol" and
312 "max_protocol" command-line switches, in case some application uses both TLS
315 SSL_CTX instances that are created for a fixed protocol version (e.g.
316 TLSv1_server_method()) also silently ignore version bounds. Previously
317 attempts to apply bounds to these protocol versions would result in an
318 error. Now only the "version-flexible" SSL_CTX instances are subject to
319 limits in configuration files in command-line options.
322 *) Handshake now fails if Extended Master Secret extension is dropped
326 *) Accidentally, an expired trusted (root) certificate is not anymore rejected
327 when validating a certificate path.
330 *) The Oracle Developer Studio compiler will start reporting deprecated APIs
332 Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
334 *) Fixed segmentation fault in SSL_check_chain()
335 Server or client applications that call the SSL_check_chain() function
336 during or after a TLS 1.3 handshake may crash due to a NULL pointer
337 dereference as a result of incorrect handling of the
338 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
339 or unrecognised signature algorithm is received from the peer. This could
340 be exploited by a malicious peer in a Denial of Service attack.
344 *) Added AES consttime code for no-asm configurations
345 an optional constant time support for AES was added
346 when building openssl for no-asm.
347 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
348 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
349 At this time this feature is by default disabled.
350 It will be enabled by default in 3.0.
353 Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
355 *) Revert the change of EOF detection while reading in libssl to avoid
356 regressions in applications depending on the current way of reporting
357 the EOF. As the existing method is not fully accurate the change to
358 reporting the EOF via SSL_ERROR_SSL is kept on the current development
359 branch and will be present in the 3.0 release.
362 *) Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
363 when primes for RSA keys are computed.
364 Since we previously always generated primes == 2 (mod 3) for RSA keys,
365 the 2-prime and 3-prime RSA modules were easy to distinguish, since
366 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
367 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
368 This avoids possible fingerprinting of newly generated RSA modules.
371 Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
372 *) Properly detect EOF while reading in libssl. Previously if we hit an EOF
373 while reading in libssl then we would report an error back to the
374 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
375 an error to the stack (which means we instead return SSL_ERROR_SSL) and
376 therefore give a hint as to what went wrong.
379 *) Check that ed25519 and ed448 are allowed by the security level. Previously
380 signature algorithms not using an MD were not being checked that they were
381 allowed by the security level.
384 *) Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
385 was not quite right. The behaviour was not consistent between resumption
386 and normal handshakes, and also not quite consistent with historical
387 behaviour. The behaviour in various scenarios has been clarified and
388 it has been updated to make it match historical behaviour as closely as
392 *) [VMS only] The header files that the VMS compilers include automatically,
393 __DECC_INCLUDE_PROLOGUE.H and __DECC_INCLUDE_EPILOGUE.H, use pragmas that
394 the C++ compiler doesn't understand. This is a shortcoming in the
395 compiler, but can be worked around with __cplusplus guards.
397 C++ applications that use OpenSSL libraries must be compiled using the
398 qualifier '/NAMES=(AS_IS,SHORTENED)' to be able to use all the OpenSSL
399 functions. Otherwise, only functions with symbols of less than 31
400 characters can be used, as the linker will not be able to successfully
401 resolve symbols with longer names.
404 *) Corrected the documentation of the return values from the EVP_DigestSign*
405 set of functions. The documentation mentioned negative values for some
406 errors, but this was never the case, so the mention of negative values
409 Code that followed the documentation and thereby check with something
410 like 'EVP_DigestSignInit(...) <= 0' will continue to work undisturbed.
413 *) Fixed an an overflow bug in the x64_64 Montgomery squaring procedure
414 used in exponentiation with 512-bit moduli. No EC algorithms are
415 affected. Analysis suggests that attacks against 2-prime RSA1024,
416 3-prime RSA1536, and DSA1024 as a result of this defect would be very
417 difficult to perform and are not believed likely. Attacks against DH512
418 are considered just feasible. However, for an attack the target would
419 have to re-use the DH512 private key, which is not recommended anyway.
420 Also applications directly using the low level API BN_mod_exp may be
421 affected if they use BN_FLG_CONSTTIME.
425 *) Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
426 The presence of this system service is determined at run-time.
429 *) Added newline escaping functionality to a filename when using openssl dgst.
430 This output format is to replicate the output format found in the '*sum'
431 checksum programs. This aims to preserve backward compatibility.
432 [Matt Eaton, Richard Levitte, and Paul Dale]
434 *) Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
438 Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
440 *) Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
441 number generator (RNG). This was intended to include protection in the
442 event of a fork() system call in order to ensure that the parent and child
443 processes did not share the same RNG state. However this protection was not
444 being used in the default case.
446 A partial mitigation for this issue is that the output from a high
447 precision timer is mixed into the RNG state so the likelihood of a parent
448 and child process sharing state is significantly reduced.
450 If an application already calls OPENSSL_init_crypto() explicitly using
451 OPENSSL_INIT_ATFORK then this problem does not occur at all.
453 [Matthias St. Pierre]
455 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is
456 used even when parsing explicit parameters, when loading a serialized key
457 or calling `EC_GROUP_new_from_ecpkparameters()`/
458 `EC_GROUP_new_from_ecparameters()`.
459 This prevents bypass of security hardening and performance gains,
460 especially for curves with specialized EC_METHODs.
461 By default, if a key encoded with explicit parameters is loaded and later
462 serialized, the output is still encoded with explicit parameters, even if
463 internally a "named" EC_GROUP is used for computation.
466 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before
467 this change, EC_GROUP_set_generator would accept order and/or cofactor as
468 NULL. After this change, only the cofactor parameter can be NULL. It also
469 does some minimal sanity checks on the passed order.
473 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
474 An attack is simple, if the first CMS_recipientInfo is valid but the
475 second CMS_recipientInfo is chosen ciphertext. If the second
476 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
477 encryption key will be replaced by garbage, and the message cannot be
478 decoded, but if the RSA decryption fails, the correct encryption key is
479 used and the recipient will not notice the attack.
480 As a work around for this potential attack the length of the decrypted
481 key must be equal to the cipher default key length, in case the
482 certifiate is not given and all recipientInfo are tried out.
483 The old behaviour can be re-enabled in the CMS code by setting the
484 CMS_DEBUG_DECRYPT flag.
488 *) Early start up entropy quality from the DEVRANDOM seed source has been
489 improved for older Linux systems. The RAND subsystem will wait for
490 /dev/random to be producing output before seeding from /dev/urandom.
491 The seeded state is stored for future library initialisations using
492 a system global shared memory segment. The shared memory identifier
493 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
494 the desired value. The default identifier is 114.
497 *) Correct the extended master secret constant on EBCDIC systems. Without this
498 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
499 negotiate EMS will fail. Unfortunately this also means that TLS connections
500 between EBCDIC systems with this fix, and EBCDIC systems without this
501 fix will fail if they negotiate EMS.
504 *) Use Windows installation paths in the mingw builds
506 Mingw isn't a POSIX environment per se, which means that Windows
507 paths should be used for installation.
511 *) Changed DH_check to accept parameters with order q and 2q subgroups.
512 With order 2q subgroups the bit 0 of the private key is not secret
513 but DH_generate_key works around that by clearing bit 0 of the
514 private key for those. This avoids leaking bit 0 of the private key.
517 *) Significantly reduce secure memory usage by the randomness pools.
520 *) Revert the DEVRANDOM_WAIT feature for Linux systems
522 The DEVRANDOM_WAIT feature added a select() call to wait for the
523 /dev/random device to become readable before reading from the
526 It turned out that this change had negative side effects on
527 performance which were not acceptable. After some discussion it
528 was decided to revert this feature and leave it up to the OS
529 resp. the platform maintainer to ensure a proper initialization
530 during early boot time.
531 [Matthias St. Pierre]
533 Changes between 1.1.1b and 1.1.1c [28 May 2019]
535 *) Add build tests for C++. These are generated files that only do one
536 thing, to include one public OpenSSL head file each. This tests that
537 the public header files can be usefully included in a C++ application.
539 This test isn't enabled by default. It can be enabled with the option
540 'enable-buildtest-c++'.
543 *) Enable SHA3 pre-hashing for ECDSA and DSA.
546 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
547 This changes the size when using the genpkey app when no size is given. It
548 fixes an omission in earlier changes that changed all RSA, DSA and DH
549 generation apps to use 2048 bits by default.
552 *) Reorganize the manual pages to consistently have RETURN VALUES,
553 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
554 util/fix-doc-nits accordingly.
555 [Paul Yang, Joshua Lock]
557 *) Add the missing accessor EVP_PKEY_get0_engine()
560 *) Have apps like 's_client' and 's_server' output the signature scheme
561 along with other cipher suite parameters when debugging.
564 *) Make OPENSSL_config() error agnostic again.
567 *) Do the error handling in RSA decryption constant time.
570 *) Prevent over long nonces in ChaCha20-Poly1305.
572 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
573 for every encryption operation. RFC 7539 specifies that the nonce value
574 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
575 and front pads the nonce with 0 bytes if it is less than 12
576 bytes. However it also incorrectly allows a nonce to be set of up to 16
577 bytes. In this case only the last 12 bytes are significant and any
578 additional leading bytes are ignored.
580 It is a requirement of using this cipher that nonce values are
581 unique. Messages encrypted using a reused nonce value are susceptible to
582 serious confidentiality and integrity attacks. If an application changes
583 the default nonce length to be longer than 12 bytes and then makes a
584 change to the leading bytes of the nonce expecting the new value to be a
585 new unique nonce then such an application could inadvertently encrypt
586 messages with a reused nonce.
588 Additionally the ignored bytes in a long nonce are not covered by the
589 integrity guarantee of this cipher. Any application that relies on the
590 integrity of these ignored leading bytes of a long nonce may be further
591 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
592 is safe because no such use sets such a long nonce value. However user
593 applications that use this cipher directly and set a non-default nonce
594 length to be longer than 12 bytes may be vulnerable.
596 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
601 *) Add DEVRANDOM_WAIT feature for Linux systems
603 On older Linux systems where the getrandom() system call is not available,
604 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
605 Contrary to getrandom(), the /dev/urandom device will not block during
606 early boot when the kernel CSPRNG has not been seeded yet.
608 To mitigate this known weakness, use select() to wait for /dev/random to
609 become readable before reading from /dev/urandom.
611 *) Ensure that SM2 only uses SM3 as digest algorithm
614 Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
616 *) Added SCA hardening for modular field inversion in EC_GROUP through
617 a new dedicated field_inv() pointer in EC_METHOD.
618 This also addresses a leakage affecting conversions from projective
619 to affine coordinates.
620 [Billy Bob Brumley, Nicola Tuveri]
622 *) Change the info callback signals for the start and end of a post-handshake
623 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
624 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
625 confused by this and assume that a TLSv1.2 renegotiation has started. This
626 can break KeyUpdate handling. Instead we no longer signal the start and end
627 of a post handshake message exchange (although the messages themselves are
628 still signalled). This could break some applications that were expecting
629 the old signals. However without this KeyUpdate is not usable for many
633 *) Fix a bug in the computation of the endpoint-pair shared secret used
634 by DTLS over SCTP. This breaks interoperability with older versions
635 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
636 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
637 interoperability with such broken implementations. However, enabling
638 this switch breaks interoperability with correct implementations.
640 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
641 re-used X509_PUBKEY object if the second PUBKEY is malformed.
644 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
647 *) Remove the 'dist' target and add a tarball building script. The
648 'dist' target has fallen out of use, and it shouldn't be
649 necessary to configure just to create a source distribution.
652 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
654 *) Timing vulnerability in DSA signature generation
656 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
657 timing side channel attack. An attacker could use variations in the signing
658 algorithm to recover the private key.
660 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
664 *) Timing vulnerability in ECDSA signature generation
666 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
667 timing side channel attack. An attacker could use variations in the signing
668 algorithm to recover the private key.
670 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
674 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
675 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
676 are retained for backwards compatibility.
679 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
680 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
681 of two gigabytes and the error handling improved.
683 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
684 categorized as a normal bug, not a security issue, because the DRBG reseeds
685 automatically and is fully functional even without additional randomness
686 provided by the application.
688 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
690 *) Add a new ClientHello callback. Provides a callback interface that gives
691 the application the ability to adjust the nascent SSL object at the
692 earliest stage of ClientHello processing, immediately after extensions have
693 been collected but before they have been processed. In particular, this
694 callback can adjust the supported TLS versions in response to the contents
698 *) Add SM2 base algorithm support.
701 *) s390x assembly pack: add (improved) hardware-support for the following
702 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
703 aes-cfb/cfb8, aes-ecb.
706 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
707 parameter is no longer accepted, as it leads to a corrupt table. NULL
708 pem_str is reserved for alias entries only.
711 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
712 step for prime curves. The new implementation is based on formulae from
713 differential addition-and-doubling in homogeneous projective coordinates
714 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
715 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
716 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
717 to work in projective coordinates.
718 [Billy Bob Brumley, Nicola Tuveri]
720 *) Change generating and checking of primes so that the error rate of not
721 being prime depends on the intended use based on the size of the input.
722 For larger primes this will result in more rounds of Miller-Rabin.
723 The maximal error rate for primes with more than 1080 bits is lowered
725 [Kurt Roeckx, Annie Yousar]
727 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
730 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
731 moving between systems, and to avoid confusion when a Windows build is
732 done with mingw vs with MSVC. For POSIX installs, there's still a
733 symlink or copy named 'tsget' to avoid that confusion as well.
736 *) Revert blinding in ECDSA sign and instead make problematic addition
737 length-invariant. Switch even to fixed-length Montgomery multiplication.
740 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
741 step for binary curves. The new implementation is based on formulae from
742 differential addition-and-doubling in mixed Lopez-Dahab projective
743 coordinates, modified to independently blind the operands.
744 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
746 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
747 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
748 EC_METHODs to implement their own specialized "ladder step", to take
749 advantage of more favorable coordinate systems or more efficient
750 differential addition-and-doubling algorithms.
751 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
753 *) Modified the random device based seed sources to keep the relevant
754 file descriptors open rather than reopening them on each access.
755 This allows such sources to operate in a chroot() jail without
756 the associated device nodes being available. This behaviour can be
757 controlled using RAND_keep_random_devices_open().
760 *) Numerous side-channel attack mitigations have been applied. This may have
761 performance impacts for some algorithms for the benefit of improved
762 security. Specific changes are noted in this change log by their respective
766 *) AIX shared library support overhaul. Switch to AIX "natural" way of
767 handling shared libraries, which means collecting shared objects of
768 different versions and bitnesses in one common archive. This allows to
769 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
770 doesn't affect the way 3rd party applications are linked, only how
771 multi-version installation is managed.
774 *) Make ec_group_do_inverse_ord() more robust and available to other
775 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
776 mitigations are applied to the fallback BN_mod_inverse().
777 When using this function rather than BN_mod_inverse() directly, new
778 EC cryptosystem implementations are then safer-by-default.
781 *) Add coordinate blinding for EC_POINT and implement projective
782 coordinate blinding for generic prime curves as a countermeasure to
783 chosen point SCA attacks.
784 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
786 *) Add blinding to ECDSA and DSA signatures to protect against side channel
787 attacks discovered by Keegan Ryan (NCC Group).
790 *) Enforce checking in the pkeyutl command line app to ensure that the input
791 length does not exceed the maximum supported digest length when performing
792 a sign, verify or verifyrecover operation.
795 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
796 I/O in combination with something like select() or poll() will hang. This
797 can be turned off again using SSL_CTX_clear_mode().
798 Many applications do not properly handle non-application data records, and
799 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
800 around the problems in those applications, but can also break some.
801 It's recommended to read the manpages about SSL_read(), SSL_write(),
802 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
803 SSL_CTX_set_read_ahead() again.
806 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
807 now allow empty (zero character) pass phrases.
810 *) Apply blinding to binary field modular inversion and remove patent
811 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
814 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
815 binary and prime elliptic curves.
818 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
819 constant time fixed point multiplication.
822 *) Revise elliptic curve scalar multiplication with timing attack
823 defenses: ec_wNAF_mul redirects to a constant time implementation
824 when computing fixed point and variable point multiplication (which
825 in OpenSSL are mostly used with secret scalars in keygen, sign,
826 ECDH derive operations).
827 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
830 *) Updated CONTRIBUTING
833 *) Updated DRBG / RAND to request nonce and additional low entropy
834 randomness from the system.
835 [Matthias St. Pierre]
837 *) Updated 'openssl rehash' to use OpenSSL consistent default.
840 *) Moved the load of the ssl_conf module to libcrypto, which helps
841 loading engines that libssl uses before libssl is initialised.
844 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
847 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
848 [Ingo Schwarze, Rich Salz]
850 *) Added output of accepting IP address and port for 'openssl s_server'
853 *) Added a new API for TLSv1.3 ciphersuites:
854 SSL_CTX_set_ciphersuites()
855 SSL_set_ciphersuites()
858 *) Memory allocation failures consistently add an error to the error
862 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
863 in libcrypto when run as setuid/setgid.
866 *) Load any config file by default when libssl is used.
869 *) Added new public header file <openssl/rand_drbg.h> and documentation
870 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
871 [Matthias St. Pierre]
873 *) QNX support removed (cannot find contributors to get their approval
874 for the license change).
877 *) TLSv1.3 replay protection for early data has been implemented. See the
878 SSL_read_early_data() man page for further details.
881 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
882 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
883 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
884 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
885 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
886 configuration has been separated out. See the ciphers man page or the
887 SSL_CTX_set_ciphersuites() man page for more information.
890 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
891 in responder mode now supports the new "-multi" option, which
892 spawns the specified number of child processes to handle OCSP
893 requests. The "-timeout" option now also limits the OCSP
894 responder's patience to wait to receive the full client request
895 on a newly accepted connection. Child processes are respawned
896 as needed, and the CA index file is automatically reloaded
897 when changed. This makes it possible to run the "ocsp" responder
898 as a long-running service, making the OpenSSL CA somewhat more
899 feature-complete. In this mode, most diagnostic messages logged
900 after entering the event loop are logged via syslog(3) rather than
904 *) Added support for X448 and Ed448. Heavily based on original work by
908 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
909 objects loaded. This adds the functions OSSL_STORE_expect() and
910 OSSL_STORE_find() as well as needed tools to construct searches and
911 get the search data out of them.
914 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
915 version of OpenSSL should review their configuration settings to ensure
916 that they are still appropriate for TLSv1.3. For further information see:
917 https://wiki.openssl.org/index.php/TLS1.3
920 *) Grand redesign of the OpenSSL random generator
922 The default RAND method now utilizes an AES-CTR DRBG according to
923 NIST standard SP 800-90Ar1. The new random generator is essentially
924 a port of the default random generator from the OpenSSL FIPS 2.0
925 object module. It is a hybrid deterministic random bit generator
926 using an AES-CTR bit stream and which seeds and reseeds itself
927 automatically using trusted system entropy sources.
929 Some of its new features are:
930 o Support for multiple DRBG instances with seed chaining.
931 o The default RAND method makes use of a DRBG.
932 o There is a public and private DRBG instance.
933 o The DRBG instances are fork-safe.
934 o Keep all global DRBG instances on the secure heap if it is enabled.
935 o The public and private DRBG instance are per thread for lock free
937 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
939 *) Changed Configure so it only says what it does and doesn't dump
940 so much data. Instead, ./configdata.pm should be used as a script
941 to display all sorts of configuration data.
944 *) Added processing of "make variables" to Configure.
947 *) Added SHA512/224 and SHA512/256 algorithm support.
950 *) The last traces of Netware support, first removed in 1.1.0, have
954 *) Get rid of Makefile.shared, and in the process, make the processing
955 of certain files (rc.obj, or the .def/.map/.opt files produced from
956 the ordinal files) more visible and hopefully easier to trace and
957 debug (or make silent).
960 *) Make it possible to have environment variable assignments as
961 arguments to config / Configure.
964 *) Add multi-prime RSA (RFC 8017) support.
967 *) Add SM3 implemented according to GB/T 32905-2016
968 [ Jack Lloyd <jack.lloyd@ribose.com>,
969 Ronald Tse <ronald.tse@ribose.com>,
970 Erick Borsboom <erick.borsboom@ribose.com> ]
972 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
973 as documented in RFC6066.
974 Based on a patch from Tomasz Moń
975 [Filipe Raimundo da Silva]
977 *) Add SM4 implemented according to GB/T 32907-2016.
978 [ Jack Lloyd <jack.lloyd@ribose.com>,
979 Ronald Tse <ronald.tse@ribose.com>,
980 Erick Borsboom <erick.borsboom@ribose.com> ]
982 *) Reimplement -newreq-nodes and ERR_error_string_n; the
983 original author does not agree with the license change.
986 *) Add ARIA AEAD TLS support.
989 *) Some macro definitions to support VS6 have been removed. Visual
990 Studio 6 has not worked since 1.1.0
993 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
994 without clearing the errors.
997 *) Add "atfork" functions. If building on a system that without
998 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
999 requirements. The RAND facility now uses/requires this.
1005 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
1006 not possible to disable entirely. However, it's still possible to
1007 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1010 To disable, configure with 'no-ui-console'. 'no-ui' is still
1011 possible to use as an alias. Check at compile time with the
1012 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1013 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1016 *) Add a STORE module, which implements a uniform and URI based reader of
1017 stores that can contain keys, certificates, CRLs and numerous other
1018 objects. The main API is loosely based on a few stdio functions,
1019 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1020 OSSL_STORE_error and OSSL_STORE_close.
1021 The implementation uses backends called "loaders" to implement arbitrary
1022 URI schemes. There is one built in "loader" for the 'file' scheme.
1025 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
1026 then adjusted to work on FreeBSD 8.4 as well.
1027 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1028 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1031 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
1032 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1033 error code calls like this:
1035 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1037 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1038 that can be encoded in C. For the foreseeable future, this will only
1040 [Richard Levitte and Tim Hudson]
1042 *) Removed BSD cryptodev engine.
1045 *) Add a build target 'build_all_generated', to build all generated files
1046 and only that. This can be used to prepare everything that requires
1047 things like perl for a system that lacks perl and then move everything
1048 to that system and do the rest of the build there.
1051 *) In the UI interface, make it possible to duplicate the user data. This
1052 can be used by engines that need to retain the data for a longer time
1053 than just the call where this user data is passed.
1056 *) Ignore the '-named_curve auto' value for compatibility of applications
1058 [Tomas Mraz <tmraz@fedoraproject.org>]
1060 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1061 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1062 alerts across multiple records (some of which could be empty). In practice
1063 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1064 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1065 support this at all. Supporting it adds significant complexity to the
1066 record layer, and its removal is unlikely to cause interoperability
1070 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1071 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1072 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1076 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1077 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1078 [Richard Levitte, Andy Polyakov]
1080 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1084 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1085 platform rather than 'mingw'.
1088 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
1089 success if they are asked to add an object which already exists
1090 in the store. This change cascades to other functions which load
1091 certificates and CRLs.
1094 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
1095 facilitate stack unwinding even from assembly subroutines.
1098 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1099 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1102 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1103 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1104 which is the minimum version we support.
1107 *) Certificate time validation (X509_cmp_time) enforces stricter
1108 compliance with RFC 5280. Fractional seconds and timezone offsets
1109 are no longer allowed.
1112 *) Add support for ARIA
1115 *) s_client will now send the Server Name Indication (SNI) extension by
1116 default unless the new "-noservername" option is used. The server name is
1117 based on the host provided to the "-connect" option unless overridden by
1118 using "-servername".
1121 *) Add support for SipHash
1124 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1125 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1126 prevent issues where no progress is being made and the peer continually
1127 sends unrecognised record types, using up resources processing them.
1130 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
1131 using the algorithm defined in
1132 https://www.akkadia.org/drepper/SHA-crypt.txt
1135 *) Heartbeat support has been removed; the ABI is changed for now.
1136 [Richard Levitte, Rich Salz]
1138 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1141 *) The RSA "null" method, which was partially supported to avoid patent
1142 issues, has been replaced to always returns NULL.
1146 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
1148 *) Client DoS due to large DH parameter
1150 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1151 malicious server can send a very large prime value to the client. This will
1152 cause the client to spend an unreasonably long period of time generating a
1153 key for this prime resulting in a hang until the client has finished. This
1154 could be exploited in a Denial Of Service attack.
1156 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1160 *) Cache timing vulnerability in RSA Key Generation
1162 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1163 a cache timing side channel attack. An attacker with sufficient access to
1164 mount cache timing attacks during the RSA key generation process could
1165 recover the private key.
1167 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1168 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1172 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1173 parameter is no longer accepted, as it leads to a corrupt table. NULL
1174 pem_str is reserved for alias entries only.
1177 *) Revert blinding in ECDSA sign and instead make problematic addition
1178 length-invariant. Switch even to fixed-length Montgomery multiplication.
1181 *) Change generating and checking of primes so that the error rate of not
1182 being prime depends on the intended use based on the size of the input.
1183 For larger primes this will result in more rounds of Miller-Rabin.
1184 The maximal error rate for primes with more than 1080 bits is lowered
1186 [Kurt Roeckx, Annie Yousar]
1188 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1191 *) Add blinding to ECDSA and DSA signatures to protect against side channel
1192 attacks discovered by Keegan Ryan (NCC Group).
1195 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1196 now allow empty (zero character) pass phrases.
1199 *) Certificate time validation (X509_cmp_time) enforces stricter
1200 compliance with RFC 5280. Fractional seconds and timezone offsets
1201 are no longer allowed.
1204 *) Fixed a text canonicalisation bug in CMS
1206 Where a CMS detached signature is used with text content the text goes
1207 through a canonicalisation process first prior to signing or verifying a
1208 signature. This process strips trailing space at the end of lines, converts
1209 line terminators to CRLF and removes additional trailing line terminators
1210 at the end of a file. A bug in the canonicalisation process meant that
1211 some characters, such as form-feed, were incorrectly treated as whitespace
1212 and removed. This is contrary to the specification (RFC5485). This fix
1213 could mean that detached text data signed with an earlier version of
1214 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
1215 signed with a fixed OpenSSL may fail to verify with an earlier version of
1216 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
1217 and use the "-binary" flag (for the "cms" command line application) or set
1218 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
1221 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
1223 *) Constructed ASN.1 types with a recursive definition could exceed the stack
1225 Constructed ASN.1 types with a recursive definition (such as can be found
1226 in PKCS7) could eventually exceed the stack given malicious input with
1227 excessive recursion. This could result in a Denial Of Service attack. There
1228 are no such structures used within SSL/TLS that come from untrusted sources
1229 so this is considered safe.
1231 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
1236 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
1238 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
1239 effectively reduced to only comparing the least significant bit of each
1240 byte. This allows an attacker to forge messages that would be considered as
1241 authenticated in an amount of tries lower than that guaranteed by the
1242 security claims of the scheme. The module can only be compiled by the
1243 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
1245 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
1250 *) Add a build target 'build_all_generated', to build all generated files
1251 and only that. This can be used to prepare everything that requires
1252 things like perl for a system that lacks perl and then move everything
1253 to that system and do the rest of the build there.
1256 *) Backport SSL_OP_NO_RENGOTIATION
1258 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
1259 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
1260 changes this is no longer possible in 1.1.0. Therefore the new
1261 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
1262 1.1.0 to provide equivalent functionality.
1264 Note that if an application built against 1.1.0h headers (or above) is run
1265 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
1266 accepted but nothing will happen, i.e. renegotiation will not be prevented.
1269 *) Removed the OS390-Unix config target. It relied on a script that doesn't
1273 *) rsaz_1024_mul_avx2 overflow bug on x86_64
1275 There is an overflow bug in the AVX2 Montgomery multiplication procedure
1276 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
1277 Analysis suggests that attacks against RSA and DSA as a result of this
1278 defect would be very difficult to perform and are not believed likely.
1279 Attacks against DH1024 are considered just feasible, because most of the
1280 work necessary to deduce information about a private key may be performed
1281 offline. The amount of resources required for such an attack would be
1282 significant. However, for an attack on TLS to be meaningful, the server
1283 would have to share the DH1024 private key among multiple clients, which is
1284 no longer an option since CVE-2016-0701.
1286 This only affects processors that support the AVX2 but not ADX extensions
1287 like Intel Haswell (4th generation).
1289 This issue was reported to OpenSSL by David Benjamin (Google). The issue
1290 was originally found via the OSS-Fuzz project.
1294 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
1296 *) bn_sqrx8x_internal carry bug on x86_64
1298 There is a carry propagating bug in the x86_64 Montgomery squaring
1299 procedure. No EC algorithms are affected. Analysis suggests that attacks
1300 against RSA and DSA as a result of this defect would be very difficult to
1301 perform and are not believed likely. Attacks against DH are considered just
1302 feasible (although very difficult) because most of the work necessary to
1303 deduce information about a private key may be performed offline. The amount
1304 of resources required for such an attack would be very significant and
1305 likely only accessible to a limited number of attackers. An attacker would
1306 additionally need online access to an unpatched system using the target
1307 private key in a scenario with persistent DH parameters and a private
1308 key that is shared between multiple clients.
1310 This only affects processors that support the BMI1, BMI2 and ADX extensions
1311 like Intel Broadwell (5th generation) and later or AMD Ryzen.
1313 This issue was reported to OpenSSL by the OSS-Fuzz project.
1317 *) Malformed X.509 IPAddressFamily could cause OOB read
1319 If an X.509 certificate has a malformed IPAddressFamily extension,
1320 OpenSSL could do a one-byte buffer overread. The most likely result
1321 would be an erroneous display of the certificate in text format.
1323 This issue was reported to OpenSSL by the OSS-Fuzz project.
1327 Changes between 1.1.0e and 1.1.0f [25 May 2017]
1329 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1330 platform rather than 'mingw'.
1333 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1334 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1335 which is the minimum version we support.
1338 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
1340 *) Encrypt-Then-Mac renegotiation crash
1342 During a renegotiation handshake if the Encrypt-Then-Mac extension is
1343 negotiated where it was not in the original handshake (or vice-versa) then
1344 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
1345 and servers are affected.
1347 This issue was reported to OpenSSL by Joe Orton (Red Hat).
1351 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
1353 *) Truncated packet could crash via OOB read
1355 If one side of an SSL/TLS path is running on a 32-bit host and a specific
1356 cipher is being used, then a truncated packet can cause that host to
1357 perform an out-of-bounds read, usually resulting in a crash.
1359 This issue was reported to OpenSSL by Robert Święcki of Google.
1363 *) Bad (EC)DHE parameters cause a client crash
1365 If a malicious server supplies bad parameters for a DHE or ECDHE key
1366 exchange then this can result in the client attempting to dereference a
1367 NULL pointer leading to a client crash. This could be exploited in a Denial
1370 This issue was reported to OpenSSL by Guido Vranken.
1374 *) BN_mod_exp may produce incorrect results on x86_64
1376 There is a carry propagating bug in the x86_64 Montgomery squaring
1377 procedure. No EC algorithms are affected. Analysis suggests that attacks
1378 against RSA and DSA as a result of this defect would be very difficult to
1379 perform and are not believed likely. Attacks against DH are considered just
1380 feasible (although very difficult) because most of the work necessary to
1381 deduce information about a private key may be performed offline. The amount
1382 of resources required for such an attack would be very significant and
1383 likely only accessible to a limited number of attackers. An attacker would
1384 additionally need online access to an unpatched system using the target
1385 private key in a scenario with persistent DH parameters and a private
1386 key that is shared between multiple clients. For example this can occur by
1387 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1388 similar to CVE-2015-3193 but must be treated as a separate problem.
1390 This issue was reported to OpenSSL by the OSS-Fuzz project.
1394 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1396 *) ChaCha20/Poly1305 heap-buffer-overflow
1398 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1399 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1400 crash. This issue is not considered to be exploitable beyond a DoS.
1402 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1406 *) CMS Null dereference
1408 Applications parsing invalid CMS structures can crash with a NULL pointer
1409 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1410 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1411 structure callback if an attempt is made to free certain invalid encodings.
1412 Only CHOICE structures using a callback which do not handle NULL value are
1415 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1419 *) Montgomery multiplication may produce incorrect results
1421 There is a carry propagating bug in the Broadwell-specific Montgomery
1422 multiplication procedure that handles input lengths divisible by, but
1423 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1424 and DH private keys are impossible. This is because the subroutine in
1425 question is not used in operations with the private key itself and an input
1426 of the attacker's direct choice. Otherwise the bug can manifest itself as
1427 transient authentication and key negotiation failures or reproducible
1428 erroneous outcome of public-key operations with specially crafted input.
1429 Among EC algorithms only Brainpool P-512 curves are affected and one
1430 presumably can attack ECDH key negotiation. Impact was not analyzed in
1431 detail, because pre-requisites for attack are considered unlikely. Namely
1432 multiple clients have to choose the curve in question and the server has to
1433 share the private key among them, neither of which is default behaviour.
1434 Even then only clients that chose the curve will be affected.
1436 This issue was publicly reported as transient failures and was not
1437 initially recognized as a security issue. Thanks to Richard Morgan for
1438 providing reproducible case.
1442 *) Removed automatic addition of RPATH in shared libraries and executables,
1443 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1446 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1448 *) Fix Use After Free for large message sizes
1450 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1451 message larger than approx 16k is received then the underlying buffer to
1452 store the incoming message is reallocated and moved. Unfortunately a
1453 dangling pointer to the old location is left which results in an attempt to
1454 write to the previously freed location. This is likely to result in a
1455 crash, however it could potentially lead to execution of arbitrary code.
1457 This issue only affects OpenSSL 1.1.0a.
1459 This issue was reported to OpenSSL by Robert Święcki.
1463 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1465 *) OCSP Status Request extension unbounded memory growth
1467 A malicious client can send an excessively large OCSP Status Request
1468 extension. If that client continually requests renegotiation, sending a
1469 large OCSP Status Request extension each time, then there will be unbounded
1470 memory growth on the server. This will eventually lead to a Denial Of
1471 Service attack through memory exhaustion. Servers with a default
1472 configuration are vulnerable even if they do not support OCSP. Builds using
1473 the "no-ocsp" build time option are not affected.
1475 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1479 *) SSL_peek() hang on empty record
1481 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1482 sends an empty record. This could be exploited by a malicious peer in a
1483 Denial Of Service attack.
1485 This issue was reported to OpenSSL by Alex Gaynor.
1489 *) Excessive allocation of memory in tls_get_message_header() and
1490 dtls1_preprocess_fragment()
1492 A (D)TLS message includes 3 bytes for its length in the header for the
1493 message. This would allow for messages up to 16Mb in length. Messages of
1494 this length are excessive and OpenSSL includes a check to ensure that a
1495 peer is sending reasonably sized messages in order to avoid too much memory
1496 being consumed to service a connection. A flaw in the logic of version
1497 1.1.0 means that memory for the message is allocated too early, prior to
1498 the excessive message length check. Due to way memory is allocated in
1499 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1500 to service a connection. This could lead to a Denial of Service through
1501 memory exhaustion. However, the excessive message length check still takes
1502 place, and this would cause the connection to immediately fail. Assuming
1503 that the application calls SSL_free() on the failed connection in a timely
1504 manner then the 21Mb of allocated memory will then be immediately freed
1505 again. Therefore the excessive memory allocation will be transitory in
1506 nature. This then means that there is only a security impact if:
1508 1) The application does not call SSL_free() in a timely manner in the event
1509 that the connection fails
1511 2) The application is working in a constrained environment where there is
1512 very little free memory
1514 3) The attacker initiates multiple connection attempts such that there are
1515 multiple connections in a state where memory has been allocated for the
1516 connection; SSL_free() has not yet been called; and there is insufficient
1517 memory to service the multiple requests.
1519 Except in the instance of (1) above any Denial Of Service is likely to be
1520 transitory because as soon as the connection fails the memory is
1521 subsequently freed again in the SSL_free() call. However there is an
1522 increased risk during this period of application crashes due to the lack of
1523 memory - which would then mean a more serious Denial of Service.
1525 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1526 (CVE-2016-6307 and CVE-2016-6308)
1529 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1530 had to be removed. Primary reason is that vendor assembler can't
1531 assemble our modules with -KPIC flag. As result it, assembly
1532 support, was not even available as option. But its lack means
1533 lack of side-channel resistant code, which is incompatible with
1534 security by todays standards. Fortunately gcc is readily available
1535 prepackaged option, which we firmly point at...
1538 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1540 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1541 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1542 (to any value) allows Windows user to access PKCS#12 file generated
1543 with Windows CryptoAPI and protected with non-ASCII password, as well
1544 as files generated under UTF-8 locale on Linux also protected with
1548 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1549 have been disabled by default and removed from DEFAULT, just like RC4.
1550 See the RC4 item below to re-enable both.
1553 *) The method for finding the storage location for the Windows RAND seed file
1554 has changed. First we check %RANDFILE%. If that is not set then we check
1555 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1556 all else fails we fall back to C:\.
1559 *) The EVP_EncryptUpdate() function has had its return type changed from void
1560 to int. A return of 0 indicates and error while a return of 1 indicates
1564 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1565 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1566 off the constant time implementation for RSA, DSA and DH have been made
1567 no-ops and deprecated.
1570 *) Windows RAND implementation was simplified to only get entropy by
1571 calling CryptGenRandom(). Various other RAND-related tickets
1573 [Joseph Wylie Yandle, Rich Salz]
1575 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1576 and OPENSSL_LH_, respectively. The old names are available
1577 with API compatibility. They new names are now completely documented.
1580 *) Unify TYPE_up_ref(obj) methods signature.
1581 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1582 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1583 int (instead of void) like all others TYPE_up_ref() methods.
1584 So now these methods also check the return value of CRYPTO_atomic_add(),
1585 and the validity of object reference counter.
1586 [fdasilvayy@gmail.com]
1588 *) With Windows Visual Studio builds, the .pdb files are installed
1589 alongside the installed libraries and executables. For a static
1590 library installation, ossl_static.pdb is the associate compiler
1591 generated .pdb file to be used when linking programs.
1594 *) Remove openssl.spec. Packaging files belong with the packagers.
1597 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1598 recognise x86_64 architectures automatically. You can still decide
1599 to build for a different bitness with the environment variable
1600 KERNEL_BITS (can be 32 or 64), for example:
1602 KERNEL_BITS=32 ./config
1606 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1607 256 bit AES and HMAC with SHA256.
1610 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1613 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1616 *) To enable users to have their own config files and build file templates,
1617 Configure looks in the directory indicated by the environment variable
1618 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1619 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1620 name and is used as is.
1623 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1624 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1625 X509_CERT_FILE_CTX was removed.
1628 *) "shared" builds are now the default. To create only static libraries use
1629 the "no-shared" Configure option.
1632 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1633 All of these option have not worked for some while and are fundamental
1637 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1638 global cleanup functions are no longer required because they are handled
1639 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1640 Explicitly de-initing can cause problems (e.g. where a library that uses
1641 OpenSSL de-inits, but an application is still using it). The affected
1642 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1643 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1644 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1645 COMP_zlib_cleanup().
1648 *) --strict-warnings no longer enables runtime debugging options
1649 such as REF_DEBUG. Instead, debug options are automatically
1650 enabled with '--debug' builds.
1651 [Andy Polyakov, Emilia Käsper]
1653 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1654 have been moved out of the public header files. New functions for managing
1655 these have been added.
1658 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1659 objects have been moved out of the public header files. New
1660 functions for managing these have been added.
1663 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1664 have been moved out of the public header files. New functions for managing
1665 these have been added.
1668 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1669 moved out of the public header files. New functions for managing these
1673 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1676 *) Removed the mk1mf build scripts.
1679 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1680 it is always safe to #include a header now.
1683 *) Removed the aged BC-32 config and all its supporting scripts
1686 *) Removed support for Ultrix, Netware, and OS/2.
1689 *) Add support for HKDF.
1690 [Alessandro Ghedini]
1692 *) Add support for blake2b and blake2s
1695 *) Added support for "pipelining". Ciphers that have the
1696 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1697 encryptions/decryptions simultaneously. There are currently no built-in
1698 ciphers with this property but the expectation is that engines will be able
1699 to offer it to significantly improve throughput. Support has been extended
1700 into libssl so that multiple records for a single connection can be
1701 processed in one go (for >=TLS 1.1).
1704 *) Added the AFALG engine. This is an async capable engine which is able to
1705 offload work to the Linux kernel. In this initial version it only supports
1706 AES128-CBC. The kernel must be version 4.1.0 or greater.
1709 *) OpenSSL now uses a new threading API. It is no longer necessary to
1710 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1711 are two supported threading models: pthreads and windows threads. It is
1712 also possible to configure OpenSSL at compile time for "no-threads". The
1713 old threading API should no longer be used. The functions have been
1714 replaced with "no-op" compatibility macros.
1715 [Alessandro Ghedini, Matt Caswell]
1717 *) Modify behavior of ALPN to invoke callback after SNI/servername
1718 callback, such that updates to the SSL_CTX affect ALPN.
1721 *) Add SSL_CIPHER queries for authentication and key-exchange.
1724 *) Changes to the DEFAULT cipherlist:
1725 - Prefer (EC)DHE handshakes over plain RSA.
1726 - Prefer AEAD ciphers over legacy ciphers.
1727 - Prefer ECDSA over RSA when both certificates are available.
1728 - Prefer TLSv1.2 ciphers/PRF.
1729 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1733 *) Change the ECC default curve list to be this, in order: x25519,
1734 secp256r1, secp521r1, secp384r1.
1737 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1738 disabled by default. They can be re-enabled using the
1739 enable-weak-ssl-ciphers option to Configure.
1742 *) If the server has ALPN configured, but supports no protocols that the
1743 client advertises, send a fatal "no_application_protocol" alert.
1744 This behaviour is SHALL in RFC 7301, though it isn't universally
1745 implemented by other servers.
1748 *) Add X25519 support.
1749 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1750 for public and private key encoding using the format documented in
1751 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1752 key generation and key derivation.
1754 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1758 *) Deprecate SRP_VBASE_get_by_user.
1759 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1760 In order to fix an unavoidable memory leak (CVE-2016-0798),
1761 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1762 seed, even if the seed is configured.
1764 Users should use SRP_VBASE_get1_by_user instead. Note that in
1765 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1766 also that even though configuring the SRP seed attempts to hide
1767 invalid usernames by continuing the handshake with fake
1768 credentials, this behaviour is not constant time and no strong
1769 guarantees are made that the handshake is indistinguishable from
1770 that of a valid user.
1773 *) Configuration change; it's now possible to build dynamic engines
1774 without having to build shared libraries and vice versa. This
1775 only applies to the engines in engines/, those in crypto/engine/
1776 will always be built into libcrypto (i.e. "static").
1778 Building dynamic engines is enabled by default; to disable, use
1779 the configuration option "disable-dynamic-engine".
1781 The only requirements for building dynamic engines are the
1782 presence of the DSO module and building with position independent
1783 code, so they will also automatically be disabled if configuring
1784 with "disable-dso" or "disable-pic".
1786 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1787 are also taken away from openssl/opensslconf.h, as they are
1791 *) Configuration change; if there is a known flag to compile
1792 position independent code, it will always be applied on the
1793 libcrypto and libssl object files, and never on the application
1794 object files. This means other libraries that use routines from
1795 libcrypto / libssl can be made into shared libraries regardless
1796 of how OpenSSL was configured.
1798 If this isn't desirable, the configuration options "disable-pic"
1799 or "no-pic" can be used to disable the use of PIC. This will
1800 also disable building shared libraries and dynamic engines.
1803 *) Removed JPAKE code. It was experimental and has no wide use.
1806 *) The INSTALL_PREFIX Makefile variable has been renamed to
1807 DESTDIR. That makes for less confusion on what this variable
1808 is for. Also, the configuration option --install_prefix is
1812 *) Heartbeat for TLS has been removed and is disabled by default
1813 for DTLS; configure with enable-heartbeats. Code that uses the
1814 old #define's might need to be updated.
1815 [Emilia Käsper, Rich Salz]
1817 *) Rename REF_CHECK to REF_DEBUG.
1820 *) New "unified" build system
1822 The "unified" build system is aimed to be a common system for all
1823 platforms we support. With it comes new support for VMS.
1825 This system builds supports building in a different directory tree
1826 than the source tree. It produces one Makefile (for unix family
1827 or lookalikes), or one descrip.mms (for VMS).
1829 The source of information to make the Makefile / descrip.mms is
1830 small files called 'build.info', holding the necessary
1831 information for each directory with source to compile, and a
1832 template in Configurations, like unix-Makefile.tmpl or
1835 With this change, the library names were also renamed on Windows
1836 and on VMS. They now have names that are closer to the standard
1837 on Unix, and include the major version number, and in certain
1838 cases, the architecture they are built for. See "Notes on shared
1839 libraries" in INSTALL.
1841 We rely heavily on the perl module Text::Template.
1844 *) Added support for auto-initialisation and de-initialisation of the library.
1845 OpenSSL no longer requires explicit init or deinit routines to be called,
1846 except in certain circumstances. See the OPENSSL_init_crypto() and
1847 OPENSSL_init_ssl() man pages for further information.
1850 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1851 "peer" argument is now expected to be a BIO_ADDR object.
1853 *) Rewrite of BIO networking library. The BIO library lacked consistent
1854 support of IPv6, and adding it required some more extensive
1855 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1856 which hold all types of addresses and chains of address information.
1857 It also introduces a new API, with functions like BIO_socket,
1858 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1859 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1860 have been adapted accordingly.
1863 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1867 *) CRIME protection: disable compression by default, even if OpenSSL is
1868 compiled with zlib enabled. Applications can still enable compression
1869 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1870 using the SSL_CONF library to configure compression.
1873 *) The signature of the session callback configured with
1874 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1875 was explicitly marked as 'const unsigned char*' instead of
1879 *) Always DPURIFY. Remove the use of uninitialized memory in the
1880 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1883 *) Removed many obsolete configuration items, including
1884 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1885 MD2_CHAR, MD2_INT, MD2_LONG
1887 IDEA_SHORT, IDEA_LONG
1888 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1889 [Rich Salz, with advice from Andy Polyakov]
1891 *) Many BN internals have been moved to an internal header file.
1892 [Rich Salz with help from Andy Polyakov]
1894 *) Configuration and writing out the results from it has changed.
1895 Files such as Makefile include/openssl/opensslconf.h and are now
1896 produced through general templates, such as Makefile.in and
1897 crypto/opensslconf.h.in and some help from the perl module
1900 Also, the center of configuration information is no longer
1901 Makefile. Instead, Configure produces a perl module in
1902 configdata.pm which holds most of the config data (in the hash
1903 table %config), the target data that comes from the target
1904 configuration in one of the Configurations/*.conf files (in
1908 *) To clarify their intended purposes, the Configure options
1909 --prefix and --openssldir change their semantics, and become more
1910 straightforward and less interdependent.
1912 --prefix shall be used exclusively to give the location INSTALLTOP
1913 where programs, scripts, libraries, include files and manuals are
1914 going to be installed. The default is now /usr/local.
1916 --openssldir shall be used exclusively to give the default
1917 location OPENSSLDIR where certificates, private keys, CRLs are
1918 managed. This is also where the default openssl.cnf gets
1920 If the directory given with this option is a relative path, the
1921 values of both the --prefix value and the --openssldir value will
1922 be combined to become OPENSSLDIR.
1923 The default for --openssldir is INSTALLTOP/ssl.
1925 Anyone who uses --openssldir to specify where OpenSSL is to be
1926 installed MUST change to use --prefix instead.
1929 *) The GOST engine was out of date and therefore it has been removed. An up
1930 to date GOST engine is now being maintained in an external repository.
1931 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1932 support for GOST ciphersuites (these are only activated if a GOST engine
1936 *) EGD is no longer supported by default; use enable-egd when
1938 [Ben Kaduk and Rich Salz]
1940 *) The distribution now has Makefile.in files, which are used to
1941 create Makefile's when Configure is run. *Configure must be run
1942 before trying to build now.*
1945 *) The return value for SSL_CIPHER_description() for error conditions
1949 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1951 Obtaining and performing DNSSEC validation of TLSA records is
1952 the application's responsibility. The application provides
1953 the TLSA records of its choice to OpenSSL, and these are then
1954 used to authenticate the peer.
1956 The TLSA records need not even come from DNS. They can, for
1957 example, be used to implement local end-entity certificate or
1958 trust-anchor "pinning", where the "pin" data takes the form
1959 of TLSA records, which can augment or replace verification
1960 based on the usual WebPKI public certification authorities.
1963 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1964 continues to support deprecated interfaces in default builds.
1965 However, applications are strongly advised to compile their
1966 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1967 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1968 or the 1.1.0 releases.
1970 In environments in which all applications have been ported to
1971 not use any deprecated interfaces OpenSSL's Configure script
1972 should be used with the --api=1.1.0 option to entirely remove
1973 support for the deprecated features from the library and
1974 unconditionally disable them in the installed headers.
1975 Essentially the same effect can be achieved with the "no-deprecated"
1976 argument to Configure, except that this will always restrict
1977 the build to just the latest API, rather than a fixed API
1980 As applications are ported to future revisions of the API,
1981 they should update their compile-time OPENSSL_API_COMPAT define
1982 accordingly, but in most cases should be able to continue to
1983 compile with later releases.
1985 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1986 0x10000000L and 0x00908000L, respectively. However those
1987 versions did not support the OPENSSL_API_COMPAT feature, and
1988 so applications are not typically tested for explicit support
1989 of just the undeprecated features of either release.
1992 *) Add support for setting the minimum and maximum supported protocol.
1993 It can bet set via the SSL_set_min_proto_version() and
1994 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1995 MaxProtocol. It's recommended to use the new APIs to disable
1996 protocols instead of disabling individual protocols using
1997 SSL_set_options() or SSL_CONF's Protocol. This change also
1998 removes support for disabling TLS 1.2 in the OpenSSL TLS
1999 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2002 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2005 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2006 and integrates ECDSA and ECDH functionality into EC. Implementations can
2007 now redirect key generation and no longer need to convert to or from
2010 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2011 include the ec.h header file instead.
2014 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
2015 ciphers who are no longer supported and drops support the ephemeral RSA key
2016 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2019 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2020 opaque. For HMAC_CTX, the following constructors and destructors
2023 HMAC_CTX *HMAC_CTX_new(void);
2024 void HMAC_CTX_free(HMAC_CTX *ctx);
2026 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2027 destroy such methods has been added. See EVP_MD_meth_new(3) and
2028 EVP_CIPHER_meth_new(3) for documentation.
2031 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2032 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2033 EVP_MD_CTX_reset() should be called instead to reinitialise
2034 an already created structure.
2035 2) For consistency with the majority of our object creators and
2036 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2037 EVP_MD_CTX_(new|free). The old names are retained as macros
2038 for deprecated builds.
2041 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
2042 cryptographic operations to be performed asynchronously as long as an
2043 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2044 further details. Libssl has also had this capability integrated with the
2045 introduction of the new mode SSL_MODE_ASYNC and associated error
2046 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2047 pages. This work was developed in partnership with Intel Corp.
2050 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2051 always enabled now. If you want to disable the support you should
2052 exclude it using the list of supported ciphers. This also means that the
2053 "-no_ecdhe" option has been removed from s_server.
2056 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2057 SSL_{CTX_}set1_curves() which can set a list.
2060 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2061 curve you want to support using SSL_{CTX_}set1_curves().
2064 *) State machine rewrite. The state machine code has been significantly
2065 refactored in order to remove much duplication of code and solve issues
2066 with the old code (see ssl/statem/README for further details). This change
2067 does have some associated API changes. Notably the SSL_state() function
2068 has been removed and replaced by SSL_get_state which now returns an
2069 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
2070 altogether. The previous handshake states defined in ssl.h and ssl3.h have
2074 *) All instances of the string "ssleay" in the public API were replaced
2075 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
2076 Some error codes related to internal RSA_eay API's were renamed.
2079 *) The demo files in crypto/threads were moved to demo/threads.
2082 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
2084 [Matt Caswell, Rich Salz]
2086 *) New ASN.1 embed macro.
2088 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
2089 structure is not allocated: it is part of the parent. That is instead of
2097 This reduces memory fragmentation and make it impossible to accidentally
2098 set a mandatory field to NULL.
2100 This currently only works for some fields specifically a SEQUENCE, CHOICE,
2101 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
2102 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
2106 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
2109 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
2110 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
2111 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
2112 DES and RC4 ciphersuites.
2115 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2116 This changes the decoding behaviour for some invalid messages,
2117 though the change is mostly in the more lenient direction, and
2118 legacy behaviour is preserved as much as possible.
2121 *) Fix no-stdio build.
2122 [ David Woodhouse <David.Woodhouse@intel.com> and also
2123 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
2125 *) New testing framework
2126 The testing framework has been largely rewritten and is now using
2127 perl and the perl modules Test::Harness and an extended variant of
2128 Test::More called OpenSSL::Test to do its work. All test scripts in
2129 test/ have been rewritten into test recipes, and all direct calls to
2130 executables in test/Makefile have become individual recipes using the
2131 simplified testing OpenSSL::Test::Simple.
2133 For documentation on our testing modules, do:
2135 perldoc test/testlib/OpenSSL/Test/Simple.pm
2136 perldoc test/testlib/OpenSSL/Test.pm
2140 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
2141 are used; the latter aborts on memory leaks (usually checked on exit).
2142 Some undocumented "set malloc, etc., hooks" functions were removed
2143 and others were changed. All are now documented.
2146 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2148 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2150 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
2151 from RFC4279, RFC4785, RFC5487, RFC5489.
2153 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
2154 original RSA_PSK patch.
2157 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
2158 era flag was never set throughout the codebase (only read). Also removed
2159 SSL3_FLAGS_POP_BUFFER which was only used if
2160 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
2163 *) Changed the default name options in the "ca", "crl", "req" and "x509"
2164 to be "oneline" instead of "compat".
2167 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
2168 not aware of clients that still exhibit this bug, and the workaround
2169 hasn't been working properly for a while.
2172 *) The return type of BIO_number_read() and BIO_number_written() as well as
2173 the corresponding num_read and num_write members in the BIO structure has
2174 changed from unsigned long to uint64_t. On platforms where an unsigned
2175 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
2179 *) Given the pervasive nature of TLS extensions it is inadvisable to run
2180 OpenSSL without support for them. It also means that maintaining
2181 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
2182 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
2185 *) Removed support for the two export grade static DH ciphersuites
2186 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
2187 were newly added (along with a number of other static DH ciphersuites) to
2188 1.0.2. However the two export ones have *never* worked since they were
2189 introduced. It seems strange in any case to be adding new export
2190 ciphersuites, and given "logjam" it also does not seem correct to fix them.
2193 *) Version negotiation has been rewritten. In particular SSLv23_method(),
2194 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
2195 and turned into macros which simply call the new preferred function names
2196 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
2197 should use the new names instead. Also as part of this change the ssl23.h
2198 header file has been removed.
2201 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
2202 code and the associated standard is no longer considered fit-for-purpose.
2205 *) RT2547 was closed. When generating a private key, try to make the
2206 output file readable only by the owner. This behavior change might
2207 be noticeable when interacting with other software.
2209 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
2213 *) Added HTTP GET support to the ocsp command.
2216 *) Changed default digest for the dgst and enc commands from MD5 to
2220 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
2223 *) Added support for TLS extended master secret from
2224 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
2225 initial patch which was a great help during development.
2228 *) All libssl internal structures have been removed from the public header
2229 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
2230 now redundant). Users should not attempt to access internal structures
2231 directly. Instead they should use the provided API functions.
2234 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
2235 Access to deprecated functions can be re-enabled by running config with
2236 "enable-deprecated". In addition applications wishing to use deprecated
2237 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
2238 will, by default, disable some transitive includes that previously existed
2239 in the header files (e.g. ec.h will no longer, by default, include bn.h)
2242 *) Added support for OCB mode. OpenSSL has been granted a patent license
2243 compatible with the OpenSSL license for use of OCB. Details are available
2244 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
2245 for OCB can be removed by calling config with no-ocb.
2248 *) SSLv2 support has been removed. It still supports receiving a SSLv2
2249 compatible client hello.
2252 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
2253 done while fixing the error code for the key-too-small case.
2254 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
2256 *) CA.sh has been removed; use CA.pl instead.
2259 *) Removed old DES API.
2262 *) Remove various unsupported platforms:
2268 Sinix/ReliantUNIX RM400
2273 16-bit platforms such as WIN16
2276 *) Clean up OPENSSL_NO_xxx #define's
2277 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
2278 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
2279 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
2280 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
2281 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
2282 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
2283 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
2284 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
2285 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
2286 Remove MS_STATIC; it's a relic from platforms <32 bits.
2289 *) Cleaned up dead code
2290 Remove all but one '#ifdef undef' which is to be looked at.
2293 *) Clean up calling of xxx_free routines.
2294 Just like free(), fix most of the xxx_free routines to accept
2295 NULL. Remove the non-null checks from callers. Save much code.
2298 *) Add secure heap for storage of private keys (when possible).
2299 Add BIO_s_secmem(), CBIGNUM, etc.
2300 Contributed by Akamai Technologies under our Corporate CLA.
2303 *) Experimental support for a new, fast, unbiased prime candidate generator,
2304 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
2305 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
2307 *) New output format NSS in the sess_id command line tool. This allows
2308 exporting the session id and the master key in NSS keylog format.
2309 [Martin Kaiser <martin@kaiser.cx>]
2311 *) Harmonize version and its documentation. -f flag is used to display
2313 [mancha <mancha1@zoho.com>]
2315 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2316 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
2317 [mancha <mancha1@zoho.com>]
2319 *) Fix some double frees. These are not thought to be exploitable.
2320 [mancha <mancha1@zoho.com>]
2322 *) A missing bounds check in the handling of the TLS heartbeat extension
2323 can be used to reveal up to 64k of memory to a connected client or
2326 Thanks for Neel Mehta of Google Security for discovering this bug and to
2327 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2328 preparing the fix (CVE-2014-0160)
2329 [Adam Langley, Bodo Moeller]
2331 *) Fix for the attack described in the paper "Recovering OpenSSL
2332 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2333 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2334 http://eprint.iacr.org/2014/140
2336 Thanks to Yuval Yarom and Naomi Benger for discovering this
2337 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2338 [Yuval Yarom and Naomi Benger]
2340 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2341 this fixes a limitation in previous versions of OpenSSL.
2344 *) Experimental encrypt-then-mac support.
2346 Experimental support for encrypt then mac from
2347 draft-gutmann-tls-encrypt-then-mac-02.txt
2349 To enable it set the appropriate extension number (0x42 for the test
2350 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
2352 For non-compliant peers (i.e. just about everything) this should have no
2355 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
2359 *) Add EVP support for key wrapping algorithms, to avoid problems with
2360 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2361 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2362 algorithms and include tests cases.
2365 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2369 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2370 MGF1 digest and OAEP label.
2373 *) Make openssl verify return errors.
2374 [Chris Palmer <palmer@google.com> and Ben Laurie]
2376 *) New function ASN1_TIME_diff to calculate the difference between two
2377 ASN1_TIME structures or one structure and the current time.
2380 *) Update fips_test_suite to support multiple command line options. New
2381 test to induce all self test errors in sequence and check expected
2385 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2386 sign or verify all in one operation.
2389 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2390 test programs and fips_test_suite. Includes functionality to parse
2391 the minimal script output of fipsalgest.pl directly.
2394 *) Add authorisation parameter to FIPS_module_mode_set().
2397 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2400 *) Use separate DRBG fields for internal and external flags. New function
2401 FIPS_drbg_health_check() to perform on demand health checking. Add
2402 generation tests to fips_test_suite with reduced health check interval to
2403 demonstrate periodic health checking. Add "nodh" option to
2404 fips_test_suite to skip very slow DH test.
2407 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2411 *) More extensive health check for DRBG checking many more failure modes.
2412 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2413 combination: call this in fips_test_suite.
2416 *) Add support for canonical generation of DSA parameter 'g'. See
2419 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2420 POST to handle HMAC cases.
2423 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2424 to return numerical and string versions of the FIPS module number.
2427 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2428 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2429 outside the validated module in the FIPS capable OpenSSL.
2432 *) Minor change to DRBG entropy callback semantics. In some cases
2433 there is no multiple of the block length between min_len and
2434 max_len. Allow the callback to return more than max_len bytes
2435 of entropy but discard any extra: it is the callback's responsibility
2436 to ensure that the extra data discarded does not impact the
2437 requested amount of entropy.
2440 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2441 information in FIPS186-3, SP800-57 and SP800-131A.
2444 *) CCM support via EVP. Interface is very similar to GCM case except we
2445 must supply all data in one chunk (i.e. no update, final) and the
2446 message length must be supplied if AAD is used. Add algorithm test
2450 *) Initial version of POST overhaul. Add POST callback to allow the status
2451 of POST to be monitored and/or failures induced. Modify fips_test_suite
2452 to use callback. Always run all selftests even if one fails.
2455 *) XTS support including algorithm test driver in the fips_gcmtest program.
2456 Note: this does increase the maximum key length from 32 to 64 bytes but
2457 there should be no binary compatibility issues as existing applications
2458 will never use XTS mode.
2461 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2462 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2463 performs algorithm blocking for unapproved PRNG types. Also do not
2464 set PRNG type in FIPS_mode_set(): leave this to the application.
2465 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2466 the standard OpenSSL PRNG: set additional data to a date time vector.
2469 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2470 This shouldn't present any incompatibility problems because applications
2471 shouldn't be using these directly and any that are will need to rethink
2472 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2475 *) Extensive self tests and health checking required by SP800-90 DRBG.
2476 Remove strength parameter from FIPS_drbg_instantiate and always
2477 instantiate at maximum supported strength.
2480 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2483 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2486 *) New function DH_compute_key_padded() to compute a DH key and pad with
2487 leading zeroes if needed: this complies with SP800-56A et al.
2490 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2491 anything, incomplete, subject to change and largely untested at present.
2494 *) Modify fipscanisteronly build option to only build the necessary object
2495 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2498 *) Add experimental option FIPSSYMS to give all symbols in
2499 fipscanister.o and FIPS or fips prefix. This will avoid
2500 conflicts with future versions of OpenSSL. Add perl script
2501 util/fipsas.pl to preprocess assembly language source files
2502 and rename any affected symbols.
2505 *) Add selftest checks and algorithm block of non-fips algorithms in
2506 FIPS mode. Remove DES2 from selftests.
2509 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2510 return internal method without any ENGINE dependencies. Add new
2511 tiny fips sign and verify functions.
2514 *) New build option no-ec2m to disable characteristic 2 code.
2517 *) New build option "fipscanisteronly". This only builds fipscanister.o
2518 and (currently) associated fips utilities. Uses the file Makefile.fips
2519 instead of Makefile.org as the prototype.
2522 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2523 Update fips_gcmtest to use IV generator.
2526 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2527 setting output buffer to NULL. The *Final function must be
2528 called although it will not retrieve any additional data. The tag
2529 can be set or retrieved with a ctrl. The IV length is by default 12
2530 bytes (96 bits) but can be set to an alternative value. If the IV
2531 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2535 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2536 underlying do_cipher function handles all cipher semantics itself
2537 including padding and finalisation. This is useful if (for example)
2538 an ENGINE cipher handles block padding itself. The behaviour of
2539 do_cipher is subtly changed if this flag is set: the return value
2540 is the number of characters written to the output buffer (zero is
2541 no longer an error code) or a negative error code. Also if the
2542 input buffer is NULL and length 0 finalisation should be performed.
2545 *) If a candidate issuer certificate is already part of the constructed
2546 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2549 *) Improve forward-security support: add functions
2551 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2552 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2554 for use by SSL/TLS servers; the callback function will be called whenever a
2555 new session is created, and gets to decide whether the session may be
2556 cached to make it resumable (return 0) or not (return 1). (As by the
2557 SSL/TLS protocol specifications, the session_id sent by the server will be
2558 empty to indicate that the session is not resumable; also, the server will
2559 not generate RFC 4507 (RFC 5077) session tickets.)
2561 A simple reasonable callback implementation is to return is_forward_secure.
2562 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2563 by the SSL/TLS server library, indicating whether it can provide forward
2565 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2567 *) New -verify_name option in command line utilities to set verification
2571 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2572 Add CMAC pkey methods.
2575 *) Experimental renegotiation in s_server -www mode. If the client
2576 browses /reneg connection is renegotiated. If /renegcert it is
2577 renegotiated requesting a certificate.
2580 *) Add an "external" session cache for debugging purposes to s_server. This
2581 should help trace issues which normally are only apparent in deployed
2582 multi-process servers.
2585 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2586 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2587 BIO_set_cipher() and some obscure PEM functions were changed so they
2588 can now return an error. The RAND changes required a change to the
2589 RAND_METHOD structure.
2592 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2593 a gcc attribute to warn if the result of a function is ignored. This
2594 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2595 whose return value is often ignored.
2598 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2599 These allow SCTs (signed certificate timestamps) to be requested and
2600 validated when establishing a connection.
2601 [Rob Percival <robpercival@google.com>]
2603 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2605 *) Prevent padding oracle in AES-NI CBC MAC check
2607 A MITM attacker can use a padding oracle attack to decrypt traffic
2608 when the connection uses an AES CBC cipher and the server support
2611 This issue was introduced as part of the fix for Lucky 13 padding
2612 attack (CVE-2013-0169). The padding check was rewritten to be in
2613 constant time by making sure that always the same bytes are read and
2614 compared against either the MAC or padding bytes. But it no longer
2615 checked that there was enough data to have both the MAC and padding
2618 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2622 *) Fix EVP_EncodeUpdate overflow
2624 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2625 Base64 encoding of binary data. If an attacker is able to supply very large
2626 amounts of input data then a length check can overflow resulting in a heap
2629 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2630 the PEM_write_bio* family of functions. These are mainly used within the
2631 OpenSSL command line applications, so any application which processes data
2632 from an untrusted source and outputs it as a PEM file should be considered
2633 vulnerable to this issue. User applications that call these APIs directly
2634 with large amounts of untrusted data may also be vulnerable.
2636 This issue was reported by Guido Vranken.
2640 *) Fix EVP_EncryptUpdate overflow
2642 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2643 is able to supply very large amounts of input data after a previous call to
2644 EVP_EncryptUpdate() with a partial block then a length check can overflow
2645 resulting in a heap corruption. Following an analysis of all OpenSSL
2646 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2647 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2648 the first called function after an EVP_EncryptInit(), and therefore that
2649 specific call must be safe. The second form is where the length passed to
2650 EVP_EncryptUpdate() can be seen from the code to be some small value and
2651 therefore there is no possibility of an overflow. Since all instances are
2652 one of these two forms, it is believed that there can be no overflows in
2653 internal code due to this problem. It should be noted that
2654 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2655 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2656 of these calls have also been analysed too and it is believed there are no
2657 instances in internal usage where an overflow could occur.
2659 This issue was reported by Guido Vranken.
2663 *) Prevent ASN.1 BIO excessive memory allocation
2665 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2666 a short invalid encoding can cause allocation of large amounts of memory
2667 potentially consuming excessive resources or exhausting memory.
2669 Any application parsing untrusted data through d2i BIO functions is
2670 affected. The memory based functions such as d2i_X509() are *not* affected.
2671 Since the memory based functions are used by the TLS library, TLS
2672 applications are not affected.
2674 This issue was reported by Brian Carpenter.
2680 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2681 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2682 in arbitrary stack data being returned in the buffer.
2684 This issue was reported by Guido Vranken.
2688 *) Modify behavior of ALPN to invoke callback after SNI/servername
2689 callback, such that updates to the SSL_CTX affect ALPN.
2692 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2696 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2697 methods are enabled and ssl2 is disabled the methods return NULL.
2700 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2702 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2703 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2704 provide any "EXPORT" or "LOW" strength ciphers.
2707 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2708 is by default disabled at build-time. Builds that are not configured with
2709 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2710 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2711 will need to explicitly call either of:
2713 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2715 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2717 as appropriate. Even if either of those is used, or the application
2718 explicitly uses the version-specific SSLv2_method() or its client and
2719 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2720 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2721 ciphers, and SSLv2 56-bit DES are no longer available.
2725 *) Fix a double-free in DSA code
2727 A double free bug was discovered when OpenSSL parses malformed DSA private
2728 keys and could lead to a DoS attack or memory corruption for applications
2729 that receive DSA private keys from untrusted sources. This scenario is
2732 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2737 *) Disable SRP fake user seed to address a server memory leak.
2739 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2741 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2742 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2743 was changed to ignore the "fake user" SRP seed, even if the seed
2746 Users should use SRP_VBASE_get1_by_user instead. Note that in
2747 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2748 also that even though configuring the SRP seed attempts to hide
2749 invalid usernames by continuing the handshake with fake
2750 credentials, this behaviour is not constant time and no strong
2751 guarantees are made that the handshake is indistinguishable from
2752 that of a valid user.
2756 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2758 In the BN_hex2bn function the number of hex digits is calculated using an
2759 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2760 large values of |i| this can result in |bn_expand| not allocating any
2761 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2762 field as NULL leading to a subsequent NULL ptr deref. For very large values
2763 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2764 In this case memory is allocated to the internal BIGNUM data field, but it
2765 is insufficiently sized leading to heap corruption. A similar issue exists
2766 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2767 is ever called by user applications with very large untrusted hex/dec data.
2768 This is anticipated to be a rare occurrence.
2770 All OpenSSL internal usage of these functions use data that is not expected
2771 to be untrusted, e.g. config file data or application command line
2772 arguments. If user developed applications generate config file data based
2773 on untrusted data then it is possible that this could also lead to security
2774 consequences. This is also anticipated to be rare.
2776 This issue was reported to OpenSSL by Guido Vranken.
2780 *) Fix memory issues in BIO_*printf functions
2782 The internal |fmtstr| function used in processing a "%s" format string in
2783 the BIO_*printf functions could overflow while calculating the length of a
2784 string and cause an OOB read when printing very long strings.
2786 Additionally the internal |doapr_outch| function can attempt to write to an
2787 OOB memory location (at an offset from the NULL pointer) in the event of a
2788 memory allocation failure. In 1.0.2 and below this could be caused where
2789 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2790 could be in processing a very long "%s" format string. Memory leaks can
2793 The first issue may mask the second issue dependent on compiler behaviour.
2794 These problems could enable attacks where large amounts of untrusted data
2795 is passed to the BIO_*printf functions. If applications use these functions
2796 in this way then they could be vulnerable. OpenSSL itself uses these
2797 functions when printing out human-readable dumps of ASN.1 data. Therefore
2798 applications that print this data could be vulnerable if the data is from
2799 untrusted sources. OpenSSL command line applications could also be
2800 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2801 as command line arguments.
2803 Libssl is not considered directly vulnerable. Additionally certificates etc
2804 received via remote connections via libssl are also unlikely to be able to
2805 trigger these issues because of message size limits enforced within libssl.
2807 This issue was reported to OpenSSL Guido Vranken.
2811 *) Side channel attack on modular exponentiation
2813 A side-channel attack was found which makes use of cache-bank conflicts on
2814 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2815 of RSA keys. The ability to exploit this issue is limited as it relies on
2816 an attacker who has control of code in a thread running on the same
2817 hyper-threaded core as the victim thread which is performing decryptions.
2819 This issue was reported to OpenSSL by Yuval Yarom, The University of
2820 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2821 Nadia Heninger, University of Pennsylvania with more information at
2822 http://cachebleed.info.
2826 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2827 if no keysize is specified with default_bits. This fixes an
2828 omission in an earlier change that changed all RSA/DSA key generation
2829 apps to use 2048 bits by default.
2832 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2833 *) DH small subgroups
2835 Historically OpenSSL only ever generated DH parameters based on "safe"
2836 primes. More recently (in version 1.0.2) support was provided for
2837 generating X9.42 style parameter files such as those required for RFC 5114
2838 support. The primes used in such files may not be "safe". Where an
2839 application is using DH configured with parameters based on primes that are
2840 not "safe" then an attacker could use this fact to find a peer's private
2841 DH exponent. This attack requires that the attacker complete multiple
2842 handshakes in which the peer uses the same private DH exponent. For example
2843 this could be used to discover a TLS server's private DH exponent if it's
2844 reusing the private DH exponent or it's using a static DH ciphersuite.
2846 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2847 TLS. It is not on by default. If the option is not set then the server
2848 reuses the same private DH exponent for the life of the server process and
2849 would be vulnerable to this attack. It is believed that many popular
2850 applications do set this option and would therefore not be at risk.
2852 The fix for this issue adds an additional check where a "q" parameter is
2853 available (as is the case in X9.42 based parameters). This detects the
2854 only known attack, and is the only possible defense for static DH
2855 ciphersuites. This could have some performance impact.
2857 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2858 default and cannot be disabled. This could have some performance impact.
2860 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2864 *) SSLv2 doesn't block disabled ciphers
2866 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2867 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2868 been disabled, provided that the SSLv2 protocol was not also disabled via
2871 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2872 and Sebastian Schinzel.
2876 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2878 *) BN_mod_exp may produce incorrect results on x86_64
2880 There is a carry propagating bug in the x86_64 Montgomery squaring
2881 procedure. No EC algorithms are affected. Analysis suggests that attacks
2882 against RSA and DSA as a result of this defect would be very difficult to
2883 perform and are not believed likely. Attacks against DH are considered just
2884 feasible (although very difficult) because most of the work necessary to
2885 deduce information about a private key may be performed offline. The amount
2886 of resources required for such an attack would be very significant and
2887 likely only accessible to a limited number of attackers. An attacker would
2888 additionally need online access to an unpatched system using the target
2889 private key in a scenario with persistent DH parameters and a private
2890 key that is shared between multiple clients. For example this can occur by
2891 default in OpenSSL DHE based SSL/TLS ciphersuites.
2893 This issue was reported to OpenSSL by Hanno Böck.
2897 *) Certificate verify crash with missing PSS parameter
2899 The signature verification routines will crash with a NULL pointer
2900 dereference if presented with an ASN.1 signature using the RSA PSS
2901 algorithm and absent mask generation function parameter. Since these
2902 routines are used to verify certificate signature algorithms this can be
2903 used to crash any certificate verification operation and exploited in a
2904 DoS attack. Any application which performs certificate verification is
2905 vulnerable including OpenSSL clients and servers which enable client
2908 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2912 *) X509_ATTRIBUTE memory leak
2914 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2915 memory. This structure is used by the PKCS#7 and CMS routines so any
2916 application which reads PKCS#7 or CMS data from untrusted sources is
2917 affected. SSL/TLS is not affected.
2919 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2924 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2925 This changes the decoding behaviour for some invalid messages,
2926 though the change is mostly in the more lenient direction, and
2927 legacy behaviour is preserved as much as possible.
2930 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2932 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2934 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2936 *) Alternate chains certificate forgery
2938 During certificate verification, OpenSSL will attempt to find an
2939 alternative certificate chain if the first attempt to build such a chain
2940 fails. An error in the implementation of this logic can mean that an
2941 attacker could cause certain checks on untrusted certificates to be
2942 bypassed, such as the CA flag, enabling them to use a valid leaf
2943 certificate to act as a CA and "issue" an invalid certificate.
2945 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2949 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2951 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2952 incompatibility in the handling of HMAC. The previous ABI has now been
2956 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2958 *) Malformed ECParameters causes infinite loop
2960 When processing an ECParameters structure OpenSSL enters an infinite loop
2961 if the curve specified is over a specially malformed binary polynomial
2964 This can be used to perform denial of service against any
2965 system which processes public keys, certificate requests or
2966 certificates. This includes TLS clients and TLS servers with
2967 client authentication enabled.
2969 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2973 *) Exploitable out-of-bounds read in X509_cmp_time
2975 X509_cmp_time does not properly check the length of the ASN1_TIME
2976 string and can read a few bytes out of bounds. In addition,
2977 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2980 An attacker can use this to craft malformed certificates and CRLs of
2981 various sizes and potentially cause a segmentation fault, resulting in
2982 a DoS on applications that verify certificates or CRLs. TLS clients
2983 that verify CRLs are affected. TLS clients and servers with client
2984 authentication enabled may be affected if they use custom verification
2987 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2988 independently by Hanno Böck.
2992 *) PKCS7 crash with missing EnvelopedContent
2994 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2995 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2996 with missing content and trigger a NULL pointer dereference on parsing.
2998 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2999 structures from untrusted sources are affected. OpenSSL clients and
3000 servers are not affected.
3002 This issue was reported to OpenSSL by Michal Zalewski (Google).
3006 *) CMS verify infinite loop with unknown hash function
3008 When verifying a signedData message the CMS code can enter an infinite loop
3009 if presented with an unknown hash function OID. This can be used to perform
3010 denial of service against any system which verifies signedData messages using
3012 This issue was reported to OpenSSL by Johannes Bauer.
3016 *) Race condition handling NewSessionTicket
3018 If a NewSessionTicket is received by a multi-threaded client when attempting to
3019 reuse a previous ticket then a race condition can occur potentially leading to
3020 a double free of the ticket data.
3024 *) Only support 256-bit or stronger elliptic curves with the
3025 'ecdh_auto' setting (server) or by default (client). Of supported
3026 curves, prefer P-256 (both).
3029 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
3031 *) ClientHello sigalgs DoS fix
3033 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
3034 invalid signature algorithms extension a NULL pointer dereference will
3035 occur. This can be exploited in a DoS attack against the server.
3037 This issue was was reported to OpenSSL by David Ramos of Stanford
3040 [Stephen Henson and Matt Caswell]
3042 *) Multiblock corrupted pointer fix
3044 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
3045 feature only applies on 64 bit x86 architecture platforms that support AES
3046 NI instructions. A defect in the implementation of "multiblock" can cause
3047 OpenSSL's internal write buffer to become incorrectly set to NULL when
3048 using non-blocking IO. Typically, when the user application is using a
3049 socket BIO for writing, this will only result in a failed connection.
3050 However if some other BIO is used then it is likely that a segmentation
3051 fault will be triggered, thus enabling a potential DoS attack.
3053 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
3057 *) Segmentation fault in DTLSv1_listen fix
3059 The DTLSv1_listen function is intended to be stateless and processes the
3060 initial ClientHello from many peers. It is common for user code to loop
3061 over the call to DTLSv1_listen until a valid ClientHello is received with
3062 an associated cookie. A defect in the implementation of DTLSv1_listen means
3063 that state is preserved in the SSL object from one invocation to the next
3064 that can lead to a segmentation fault. Errors processing the initial
3065 ClientHello can trigger this scenario. An example of such an error could be
3066 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
3069 This issue was reported to OpenSSL by Per Allansson.
3073 *) Segmentation fault in ASN1_TYPE_cmp fix
3075 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
3076 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
3077 certificate signature algorithm consistency this can be used to crash any
3078 certificate verification operation and exploited in a DoS attack. Any
3079 application which performs certificate verification is vulnerable including
3080 OpenSSL clients and servers which enable client authentication.
3084 *) Segmentation fault for invalid PSS parameters fix
3086 The signature verification routines will crash with a NULL pointer
3087 dereference if presented with an ASN.1 signature using the RSA PSS
3088 algorithm and invalid parameters. Since these routines are used to verify
3089 certificate signature algorithms this can be used to crash any
3090 certificate verification operation and exploited in a DoS attack. Any
3091 application which performs certificate verification is vulnerable including
3092 OpenSSL clients and servers which enable client authentication.
3094 This issue was was reported to OpenSSL by Brian Carpenter.
3098 *) ASN.1 structure reuse memory corruption fix
3100 Reusing a structure in ASN.1 parsing may allow an attacker to cause
3101 memory corruption via an invalid write. Such reuse is and has been
3102 strongly discouraged and is believed to be rare.
3104 Applications that parse structures containing CHOICE or ANY DEFINED BY
3105 components may be affected. Certificate parsing (d2i_X509 and related
3106 functions) are however not affected. OpenSSL clients and servers are
3111 *) PKCS7 NULL pointer dereferences fix
3113 The PKCS#7 parsing code does not handle missing outer ContentInfo
3114 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
3115 missing content and trigger a NULL pointer dereference on parsing.
3117 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
3118 otherwise parse PKCS#7 structures from untrusted sources are
3119 affected. OpenSSL clients and servers are not affected.
3121 This issue was reported to OpenSSL by Michal Zalewski (Google).
3125 *) DoS via reachable assert in SSLv2 servers fix
3127 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
3128 servers that both support SSLv2 and enable export cipher suites by sending
3129 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
3131 This issue was discovered by Sean Burford (Google) and Emilia Käsper
3132 (OpenSSL development team).
3136 *) Empty CKE with client auth and DHE fix
3138 If client auth is used then a server can seg fault in the event of a DHE
3139 ciphersuite being selected and a zero length ClientKeyExchange message
3140 being sent by the client. This could be exploited in a DoS attack.
3144 *) Handshake with unseeded PRNG fix
3146 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
3147 with an unseeded PRNG. The conditions are:
3148 - The client is on a platform where the PRNG has not been seeded
3149 automatically, and the user has not seeded manually
3150 - A protocol specific client method version has been used (i.e. not
3151 SSL_client_methodv23)
3152 - A ciphersuite is used that does not require additional random data from
3153 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
3155 If the handshake succeeds then the client random that has been used will
3156 have been generated from a PRNG with insufficient entropy and therefore the
3157 output may be predictable.
3159 For example using the following command with an unseeded openssl will
3160 succeed on an unpatched platform:
3162 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
3166 *) Use After Free following d2i_ECPrivatekey error fix
3168 A malformed EC private key file consumed via the d2i_ECPrivateKey function
3169 could cause a use after free condition. This, in turn, could cause a double
3170 free in several private key parsing functions (such as d2i_PrivateKey
3171 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
3172 for applications that receive EC private keys from untrusted
3173 sources. This scenario is considered rare.
3175 This issue was discovered by the BoringSSL project and fixed in their
3180 *) X509_to_X509_REQ NULL pointer deref fix
3182 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
3183 the certificate key is invalid. This function is rarely used in practice.
3185 This issue was discovered by Brian Carpenter.
3189 *) Removed the export ciphers from the DEFAULT ciphers
3192 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
3194 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
3195 ARMv5 through ARMv8, as opposite to "locking" it to single one.
3196 So far those who have to target multiple platforms would compromise
3197 and argue that binary targeting say ARMv5 would still execute on
3198 ARMv8. "Universal" build resolves this compromise by providing
3199 near-optimal performance even on newer platforms.
3202 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
3203 (other platforms pending).
3204 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
3206 *) Add support for the SignedCertificateTimestampList certificate and
3207 OCSP response extensions from RFC6962.
3210 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3211 for corner cases. (Certain input points at infinity could lead to
3212 bogus results, with non-infinity inputs mapped to infinity too.)
3215 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
3216 This covers AES, SHA256/512 and GHASH. "Initial" means that most
3217 common cases are optimized and there still is room for further
3218 improvements. Vector Permutation AES for Altivec is also added.
3221 *) Add support for little-endian ppc64 Linux target.
3222 [Marcelo Cerri (IBM)]
3224 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
3225 SHA1, SHA256 and GHASH. "Initial" means that most common cases
3226 are optimized and there still is room for further improvements.
3227 Both 32- and 64-bit modes are supported.
3228 [Andy Polyakov, Ard Biesheuvel (Linaro)]
3230 *) Improved ARMv7 NEON support.
3233 *) Support for SPARC Architecture 2011 crypto extensions, first
3234 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
3235 SHA256/512, MD5, GHASH and modular exponentiation.
3236 [Andy Polyakov, David Miller]
3238 *) Accelerated modular exponentiation for Intel processors, a.k.a.
3240 [Shay Gueron & Vlad Krasnov (Intel Corp)]
3242 *) Support for new and upcoming Intel processors, including AVX2,
3243 BMI and SHA ISA extensions. This includes additional "stitched"
3244 implementations, AESNI-SHA256 and GCM, and multi-buffer support
3247 This work was sponsored by Intel Corp.
3250 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
3251 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
3252 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
3255 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3256 this fixes a limitation in previous versions of OpenSSL.
3259 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3260 MGF1 digest and OAEP label.
3263 *) Add EVP support for key wrapping algorithms, to avoid problems with
3264 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3265 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3266 algorithms and include tests cases.
3269 *) Add functions to allocate and set the fields of an ECDSA_METHOD
3271 [Douglas E. Engert, Steve Henson]
3273 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
3274 difference in days and seconds between two tm or ASN1_TIME structures.
3277 *) Add -rev test option to s_server to just reverse order of characters
3278 received by client and send back to server. Also prints an abbreviated
3279 summary of the connection parameters.
3282 *) New option -brief for s_client and s_server to print out a brief summary
3283 of connection parameters.
3286 *) Add callbacks for arbitrary TLS extensions.
3287 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
3289 *) New option -crl_download in several openssl utilities to download CRLs
3290 from CRLDP extension in certificates.
3293 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
3296 *) New function X509_CRL_diff to generate a delta CRL from the difference
3297 of two full CRLs. Add support to "crl" utility.
3300 *) New functions to set lookup_crls function and to retrieve
3301 X509_STORE from X509_STORE_CTX.
3304 *) Print out deprecated issuer and subject unique ID fields in
3308 *) Extend OCSP I/O functions so they can be used for simple general purpose
3309 HTTP as well as OCSP. New wrapper function which can be used to download
3310 CRLs using the OCSP API.
3313 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
3316 *) SSL_CONF* functions. These provide a common framework for application
3317 configuration using configuration files or command lines.
3320 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
3321 message callback and prints the results. Needs compile time option
3322 "enable-ssl-trace". New options to s_client and s_server to enable
3326 *) New ctrl and macro to retrieve supported points extensions.
3327 Print out extension in s_server and s_client.
3330 *) New functions to retrieve certificate signature and signature
3334 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
3338 *) New Suite B modes for TLS code. These use and enforce the requirements
3339 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
3340 only use Suite B curves. The Suite B modes can be set by using the
3341 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
3344 *) New chain verification flags for Suite B levels of security. Check
3345 algorithms are acceptable when flags are set in X509_verify_cert.
3348 *) Make tls1_check_chain return a set of flags indicating checks passed
3349 by a certificate chain. Add additional tests to handle client
3350 certificates: checks for matching certificate type and issuer name
3354 *) If an attempt is made to use a signature algorithm not in the peer
3355 preference list abort the handshake. If client has no suitable
3356 signature algorithms in response to a certificate request do not
3357 use the certificate.
3360 *) If server EC tmp key is not in client preference list abort handshake.
3363 *) Add support for certificate stores in CERT structure. This makes it
3364 possible to have different stores per SSL structure or one store in
3365 the parent SSL_CTX. Include distinct stores for certificate chain
3366 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3367 to build and store a certificate chain in CERT structure: returning
3368 an error if the chain cannot be built: this will allow applications
3369 to test if a chain is correctly configured.
3371 Note: if the CERT based stores are not set then the parent SSL_CTX
3372 store is used to retain compatibility with existing behaviour.
3376 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3377 mask based on the current session, check mask when sending client
3378 hello and checking the requested ciphersuite.
3381 *) New ctrls to retrieve and set certificate types in a certificate
3382 request message. Print out received values in s_client. If certificate
3383 types is not set with custom values set sensible values based on
3384 supported signature algorithms.
3387 *) Support for distinct client and server supported signature algorithms.
3390 *) Add certificate callback. If set this is called whenever a certificate
3391 is required by client or server. An application can decide which
3392 certificate chain to present based on arbitrary criteria: for example
3393 supported signature algorithms. Add very simple example to s_server.
3394 This fixes many of the problems and restrictions of the existing client
3395 certificate callback: for example you can now clear an existing
3396 certificate and specify the whole chain.
3399 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3400 the certificate can be used for (if anything). Set valid_flags field
3401 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3402 to have similar checks in it.
3404 Add new "cert_flags" field to CERT structure and include a "strict mode".
3405 This enforces some TLS certificate requirements (such as only permitting
3406 certificate signature algorithms contained in the supported algorithms
3407 extension) which some implementations ignore: this option should be used
3408 with caution as it could cause interoperability issues.
3411 *) Update and tidy signature algorithm extension processing. Work out
3412 shared signature algorithms based on preferences and peer algorithms
3413 and print them out in s_client and s_server. Abort handshake if no
3414 shared signature algorithms.
3417 *) Add new functions to allow customised supported signature algorithms
3418 for SSL and SSL_CTX structures. Add options to s_client and s_server
3422 *) New function SSL_certs_clear() to delete all references to certificates
3423 from an SSL structure. Before this once a certificate had been added
3424 it couldn't be removed.
3427 *) Integrate hostname, email address and IP address checking with certificate
3428 verification. New verify options supporting checking in openssl utility.
3431 *) Fixes and wildcard matching support to hostname and email checking
3432 functions. Add manual page.
3433 [Florian Weimer (Red Hat Product Security Team)]
3435 *) New functions to check a hostname email or IP address against a
3436 certificate. Add options x509 utility to print results of checks against
3440 *) Fix OCSP checking.
3441 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3443 *) Initial experimental support for explicitly trusted non-root CAs.
3444 OpenSSL still tries to build a complete chain to a root but if an
3445 intermediate CA has a trust setting included that is used. The first
3446 setting is used: whether to trust (e.g., -addtrust option to the x509
3450 *) Add -trusted_first option which attempts to find certificates in the
3451 trusted store even if an untrusted chain is also supplied.
3454 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3455 platform support for Linux and Android.
3458 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3461 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3462 When in FIPS mode the approved implementations are used as normal,
3463 when not in FIPS mode the internal unapproved versions are used instead.
3464 This means that the FIPS capable OpenSSL isn't forced to use the
3465 (often lower performance) FIPS implementations outside FIPS mode.
3468 *) Transparently support X9.42 DH parameters when calling
3469 PEM_read_bio_DHparameters. This means existing applications can handle
3470 the new parameter format automatically.
3473 *) Initial experimental support for X9.42 DH parameter format: mainly
3474 to support use of 'q' parameter for RFC5114 parameters.
3477 *) Add DH parameters from RFC5114 including test data to dhtest.
3480 *) Support for automatic EC temporary key parameter selection. If enabled
3481 the most preferred EC parameters are automatically used instead of
3482 hardcoded fixed parameters. Now a server just has to call:
3483 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3484 support ECDH and use the most appropriate parameters.
3487 *) Enhance and tidy EC curve and point format TLS extension code. Use
3488 static structures instead of allocation if default values are used.
3489 New ctrls to set curves we wish to support and to retrieve shared curves.
3490 Print out shared curves in s_server. New options to s_server and s_client
3491 to set list of supported curves.
3494 *) New ctrls to retrieve supported signature algorithms and
3495 supported curve values as an array of NIDs. Extend openssl utility
3496 to print out received values.
3499 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3500 between NIDs and the more common NIST names such as "P-256". Enhance
3501 ecparam utility and ECC method to recognise the NIST names for curves.
3504 *) Enhance SSL/TLS certificate chain handling to support different
3505 chains for each certificate instead of one chain in the parent SSL_CTX.
3508 *) Support for fixed DH ciphersuite client authentication: where both
3509 server and client use DH certificates with common parameters.
3512 *) Support for fixed DH ciphersuites: those requiring DH server
3516 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3518 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3519 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3520 X509_CINF_get_signature were reverted post internal team review.
3522 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3524 *) Build fixes for the Windows and OpenVMS platforms
3525 [Matt Caswell and Richard Levitte]
3527 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3529 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3530 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3531 dereference. This could lead to a Denial Of Service attack. Thanks to
3532 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3536 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3537 dtls1_buffer_record function under certain conditions. In particular this
3538 could occur if an attacker sent repeated DTLS records with the same
3539 sequence number but for the next epoch. The memory leak could be exploited
3540 by an attacker in a Denial of Service attack through memory exhaustion.
3541 Thanks to Chris Mueller for reporting this issue.
3545 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3546 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3547 method would be set to NULL which could later result in a NULL pointer
3548 dereference. Thanks to Frank Schmirler for reporting this issue.
3552 *) Abort handshake if server key exchange message is omitted for ephemeral
3555 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3556 reporting this issue.
3560 *) Remove non-export ephemeral RSA code on client and server. This code
3561 violated the TLS standard by allowing the use of temporary RSA keys in
3562 non-export ciphersuites and could be used by a server to effectively
3563 downgrade the RSA key length used to a value smaller than the server
3564 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3565 INRIA or reporting this issue.
3569 *) Fixed issue where DH client certificates are accepted without verification.
3570 An OpenSSL server will accept a DH certificate for client authentication
3571 without the certificate verify message. This effectively allows a client to
3572 authenticate without the use of a private key. This only affects servers
3573 which trust a client certificate authority which issues certificates
3574 containing DH keys: these are extremely rare and hardly ever encountered.
3575 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3580 *) Ensure that the session ID context of an SSL is updated when its
3581 SSL_CTX is updated via SSL_set_SSL_CTX.
3583 The session ID context is typically set from the parent SSL_CTX,
3584 and can vary with the CTX.
3587 *) Fix various certificate fingerprint issues.
3589 By using non-DER or invalid encodings outside the signed portion of a
3590 certificate the fingerprint can be changed without breaking the signature.
3591 Although no details of the signed portion of the certificate can be changed
3592 this can cause problems with some applications: e.g. those using the
3593 certificate fingerprint for blacklists.
3595 1. Reject signatures with non zero unused bits.
3597 If the BIT STRING containing the signature has non zero unused bits reject
3598 the signature. All current signature algorithms require zero unused bits.
3600 2. Check certificate algorithm consistency.
3602 Check the AlgorithmIdentifier inside TBS matches the one in the
3603 certificate signature. NB: this will result in signature failure
3604 errors for some broken certificates.
3606 Thanks to Konrad Kraszewski from Google for reporting this issue.
3608 3. Check DSA/ECDSA signatures use DER.
3610 Re-encode DSA/ECDSA signatures and compare with the original received
3611 signature. Return an error if there is a mismatch.
3613 This will reject various cases including garbage after signature
3614 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3615 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3616 (negative or with leading zeroes).
3618 Further analysis was conducted and fixes were developed by Stephen Henson
3619 of the OpenSSL core team.
3624 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3625 results on some platforms, including x86_64. This bug occurs at random
3626 with a very low probability, and is not known to be exploitable in any
3627 way, though its exact impact is difficult to determine. Thanks to Pieter
3628 Wuille (Blockstream) who reported this issue and also suggested an initial
3629 fix. Further analysis was conducted by the OpenSSL development team and
3630 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3631 the OpenSSL core team.
3635 *) Do not resume sessions on the server if the negotiated protocol
3636 version does not match the session's version. Resuming with a different
3637 version, while not strictly forbidden by the RFC, is of questionable
3638 sanity and breaks all known clients.
3639 [David Benjamin, Emilia Käsper]
3641 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3642 early CCS messages during renegotiation. (Note that because
3643 renegotiation is encrypted, this early CCS was not exploitable.)
3646 *) Tighten client-side session ticket handling during renegotiation:
3647 ensure that the client only accepts a session ticket if the server sends
3648 the extension anew in the ServerHello. Previously, a TLS client would
3649 reuse the old extension state and thus accept a session ticket if one was
3650 announced in the initial ServerHello.
3652 Similarly, ensure that the client requires a session ticket if one
3653 was advertised in the ServerHello. Previously, a TLS client would
3654 ignore a missing NewSessionTicket message.
3657 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3659 *) SRTP Memory Leak.
3661 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3662 sends a carefully crafted handshake message, to cause OpenSSL to fail
3663 to free up to 64k of memory causing a memory leak. This could be
3664 exploited in a Denial Of Service attack. This issue affects OpenSSL
3665 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3666 whether SRTP is used or configured. Implementations of OpenSSL that
3667 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3669 The fix was developed by the OpenSSL team.
3673 *) Session Ticket Memory Leak.
3675 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3676 integrity of that ticket is first verified. In the event of a session
3677 ticket integrity check failing, OpenSSL will fail to free memory
3678 causing a memory leak. By sending a large number of invalid session
3679 tickets an attacker could exploit this issue in a Denial Of Service
3684 *) Build option no-ssl3 is incomplete.
3686 When OpenSSL is configured with "no-ssl3" as a build option, servers
3687 could accept and complete a SSL 3.0 handshake, and clients could be
3688 configured to send them.
3690 [Akamai and the OpenSSL team]
3692 *) Add support for TLS_FALLBACK_SCSV.
3693 Client applications doing fallback retries should call
3694 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3696 [Adam Langley, Bodo Moeller]
3698 *) Add additional DigestInfo checks.
3700 Re-encode DigestInto in DER and check against the original when
3701 verifying RSA signature: this will reject any improperly encoded
3702 DigestInfo structures.
3704 Note: this is a precautionary measure and no attacks are currently known.
3708 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3710 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3711 SRP code can be overrun an internal buffer. Add sanity check that
3712 g, A, B < N to SRP code.
3714 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3715 Group for discovering this issue.
3719 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3720 TLS 1.0 instead of higher protocol versions when the ClientHello message
3721 is badly fragmented. This allows a man-in-the-middle attacker to force a
3722 downgrade to TLS 1.0 even if both the server and the client support a
3723 higher protocol version, by modifying the client's TLS records.
3725 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3726 researching this issue.
3730 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3731 to a denial of service attack. A malicious server can crash the client
3732 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3733 ciphersuite and sending carefully crafted handshake messages.
3735 Thanks to Felix Gröbert (Google) for discovering and researching this
3740 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3741 to leak memory. This can be exploited through a Denial of Service attack.
3742 Thanks to Adam Langley for discovering and researching this issue.
3746 *) An attacker can force openssl to consume large amounts of memory whilst
3747 processing DTLS handshake messages. This can be exploited through a
3748 Denial of Service attack.
3749 Thanks to Adam Langley for discovering and researching this issue.
3753 *) An attacker can force an error condition which causes openssl to crash
3754 whilst processing DTLS packets due to memory being freed twice. This
3755 can be exploited through a Denial of Service attack.
3756 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3761 *) If a multithreaded client connects to a malicious server using a resumed
3762 session and the server sends an ec point format extension it could write
3763 up to 255 bytes to freed memory.
3765 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3770 *) A malicious server can crash an OpenSSL client with a null pointer
3771 dereference (read) by specifying an SRP ciphersuite even though it was not
3772 properly negotiated with the client. This can be exploited through a
3773 Denial of Service attack.
3775 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3776 discovering and researching this issue.
3780 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3781 X509_name_oneline, X509_name_print_ex et al. to leak some information
3782 from the stack. Applications may be affected if they echo pretty printing
3783 output to the attacker.
3785 Thanks to Ivan Fratric (Google) for discovering this issue.
3787 [Emilia Käsper, and Steve Henson]
3789 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3790 for corner cases. (Certain input points at infinity could lead to
3791 bogus results, with non-infinity inputs mapped to infinity too.)
3794 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3796 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3797 handshake can force the use of weak keying material in OpenSSL
3798 SSL/TLS clients and servers.
3800 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3801 researching this issue. (CVE-2014-0224)
3802 [KIKUCHI Masashi, Steve Henson]
3804 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3805 OpenSSL DTLS client the code can be made to recurse eventually crashing
3808 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3810 [Imre Rad, Steve Henson]
3812 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3813 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3814 client or server. This is potentially exploitable to run arbitrary
3815 code on a vulnerable client or server.
3817 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3818 [Jüri Aedla, Steve Henson]
3820 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3821 are subject to a denial of service attack.
3823 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3824 this issue. (CVE-2014-3470)
3825 [Felix Gröbert, Ivan Fratric, Steve Henson]
3827 *) Harmonize version and its documentation. -f flag is used to display
3829 [mancha <mancha1@zoho.com>]
3831 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3832 in i2d_ECPrivateKey.
3833 [mancha <mancha1@zoho.com>]
3835 *) Fix some double frees. These are not thought to be exploitable.
3836 [mancha <mancha1@zoho.com>]
3838 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3840 *) A missing bounds check in the handling of the TLS heartbeat extension
3841 can be used to reveal up to 64k of memory to a connected client or
3844 Thanks for Neel Mehta of Google Security for discovering this bug and to
3845 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3846 preparing the fix (CVE-2014-0160)
3847 [Adam Langley, Bodo Moeller]
3849 *) Fix for the attack described in the paper "Recovering OpenSSL
3850 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3851 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3852 http://eprint.iacr.org/2014/140
3854 Thanks to Yuval Yarom and Naomi Benger for discovering this
3855 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3856 [Yuval Yarom and Naomi Benger]
3858 *) TLS pad extension: draft-agl-tls-padding-03
3860 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3861 TLS client Hello record length value would otherwise be > 255 and
3862 less that 512 pad with a dummy extension containing zeroes so it
3863 is at least 512 bytes long.
3865 [Adam Langley, Steve Henson]
3867 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3869 *) Fix for TLS record tampering bug. A carefully crafted invalid
3870 handshake could crash OpenSSL with a NULL pointer exception.
3871 Thanks to Anton Johansson for reporting this issues.
3874 *) Keep original DTLS digest and encryption contexts in retransmission
3875 structures so we can use the previous session parameters if they need
3876 to be resent. (CVE-2013-6450)
3879 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3880 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3881 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3882 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3883 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3884 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3885 [Rob Stradling, Adam Langley]
3887 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3889 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3890 supporting platforms or when small records were transferred.
3891 [Andy Polyakov, Steve Henson]
3893 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3895 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3897 This addresses the flaw in CBC record processing discovered by
3898 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3899 at: http://www.isg.rhul.ac.uk/tls/
3901 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3902 Security Group at Royal Holloway, University of London
3903 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3904 Emilia Käsper for the initial patch.
3906 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3908 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3909 ciphersuites which can be exploited in a denial of service attack.
3910 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3911 and detecting this bug and to Wolfgang Ettlinger
3912 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3916 *) Return an error when checking OCSP signatures when key is NULL.
3917 This fixes a DoS attack. (CVE-2013-0166)
3920 *) Make openssl verify return errors.
3921 [Chris Palmer <palmer@google.com> and Ben Laurie]
3923 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3924 the right response is stapled. Also change SSL_get_certificate()
3925 so it returns the certificate actually sent.
3926 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3927 [Rob Stradling <rob.stradling@comodo.com>]
3929 *) Fix possible deadlock when decoding public keys.
3932 *) Don't use TLS 1.0 record version number in initial client hello
3936 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3938 *) Sanity check record length before skipping explicit IV in TLS
3939 1.2, 1.1 and DTLS to fix DoS attack.
3941 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3942 fuzzing as a service testing platform.
3946 *) Initialise tkeylen properly when encrypting CMS messages.
3947 Thanks to Solar Designer of Openwall for reporting this issue.
3950 *) In FIPS mode don't try to use composite ciphers as they are not
3954 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3956 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3957 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3958 mean any application compiled against OpenSSL 1.0.0 headers setting
3959 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3960 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3961 0x10000000L Any application which was previously compiled against
3962 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3963 will need to be recompiled as a result. Letting be results in
3964 inability to disable specifically TLS 1.1 and in client context,
3965 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3968 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3969 disable just protocol X, but all protocols above X *if* there are
3970 protocols *below* X still enabled. In more practical terms it means
3971 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3972 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3973 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3977 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3979 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3980 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3981 in CRYPTO_realloc_clean.
3983 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3984 issue and to Adam Langley <agl@chromium.org> for fixing it.
3986 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3988 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3991 *) Workarounds for some broken servers that "hang" if a client hello
3992 record length exceeds 255 bytes.
3994 1. Do not use record version number > TLS 1.0 in initial client
3995 hello: some (but not all) hanging servers will now work.
3996 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3997 the number of ciphers sent in the client hello. This should be
3998 set to an even number, such as 50, for example by passing:
3999 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
4000 Most broken servers should now work.
4001 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
4002 TLS 1.2 client support entirely.
4005 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
4008 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
4010 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
4011 STRING form instead of a DigestInfo.
4014 *) The format used for MDC2 RSA signatures is inconsistent between EVP
4015 and the RSA_sign/RSA_verify functions. This was made more apparent when
4016 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
4017 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
4018 the correct format in RSA_verify so both forms transparently work.
4021 *) Some servers which support TLS 1.0 can choke if we initially indicate
4022 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
4023 encrypted premaster secret. As a workaround use the maximum permitted
4024 client version in client hello, this should keep such servers happy
4025 and still work with previous versions of OpenSSL.
4028 *) Add support for TLS/DTLS heartbeats.
4029 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4031 *) Add support for SCTP.
4032 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4034 *) Improved PRNG seeding for VOS.
4035 [Paul Green <Paul.Green@stratus.com>]
4037 *) Extensive assembler packs updates, most notably:
4039 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
4040 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
4041 - x86_64: bit-sliced AES implementation;
4042 - ARM: NEON support, contemporary platforms optimizations;
4043 - s390x: z196 support;
4044 - *: GHASH and GF(2^m) multiplication implementations;
4048 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
4049 (removal of unnecessary code)
4050 [Peter Sylvester <peter.sylvester@edelweb.fr>]
4052 *) Add TLS key material exporter from RFC 5705.
4055 *) Add DTLS-SRTP negotiation from RFC 5764.
4058 *) Add Next Protocol Negotiation,
4059 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
4060 disabled with a no-npn flag to config or Configure. Code donated
4062 [Adam Langley <agl@google.com> and Ben Laurie]
4064 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
4065 NIST-P256, NIST-P521, with constant-time single point multiplication on
4066 typical inputs. Compiler support for the nonstandard type __uint128_t is
4067 required to use this (present in gcc 4.4 and later, for 64-bit builds).
4068 Code made available under Apache License version 2.0.
4070 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
4071 line to include this in your build of OpenSSL, and run "make depend" (or
4072 "make update"). This enables the following EC_METHODs:
4074 EC_GFp_nistp224_method()
4075 EC_GFp_nistp256_method()
4076 EC_GFp_nistp521_method()
4078 EC_GROUP_new_by_curve_name() will automatically use these (while
4079 EC_GROUP_new_curve_GFp() currently prefers the more flexible
4081 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
4083 *) Use type ossl_ssize_t instead of ssize_t which isn't available on
4084 all platforms. Move ssize_t definition from e_os.h to the public
4085 header file e_os2.h as it now appears in public header file cms.h
4088 *) New -sigopt option to the ca, req and x509 utilities. Additional
4089 signature parameters can be passed using this option and in
4093 *) Add RSA PSS signing function. This will generate and set the
4094 appropriate AlgorithmIdentifiers for PSS based on those in the
4095 corresponding EVP_MD_CTX structure. No application support yet.
4098 *) Support for companion algorithm specific ASN1 signing routines.
4099 New function ASN1_item_sign_ctx() signs a pre-initialised
4100 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
4101 the appropriate parameters.
4104 *) Add new algorithm specific ASN1 verification initialisation function
4105 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
4106 handling will be the same no matter what EVP_PKEY_METHOD is used.
4107 Add a PSS handler to support verification of PSS signatures: checked
4108 against a number of sample certificates.
4111 *) Add signature printing for PSS. Add PSS OIDs.
4112 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
4114 *) Add algorithm specific signature printing. An individual ASN1 method
4115 can now print out signatures instead of the standard hex dump.
4117 More complex signatures (e.g. PSS) can print out more meaningful
4118 information. Include DSA version that prints out the signature
4122 *) Password based recipient info support for CMS library: implementing
4126 *) Split password based encryption into PBES2 and PBKDF2 functions. This
4127 neatly separates the code into cipher and PBE sections and is required
4128 for some algorithms that split PBES2 into separate pieces (such as
4129 password based CMS).
4132 *) Session-handling fixes:
4133 - Fix handling of connections that are resuming with a session ID,
4134 but also support Session Tickets.
4135 - Fix a bug that suppressed issuing of a new ticket if the client
4136 presented a ticket with an expired session.
4137 - Try to set the ticket lifetime hint to something reasonable.
4138 - Make tickets shorter by excluding irrelevant information.
4139 - On the client side, don't ignore renewed tickets.
4140 [Adam Langley, Bodo Moeller (Google)]
4142 *) Fix PSK session representation.
4145 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
4147 This work was sponsored by Intel.
4150 *) Add GCM support to TLS library. Some custom code is needed to split
4151 the IV between the fixed (from PRF) and explicit (from TLS record)
4152 portions. This adds all GCM ciphersuites supported by RFC5288 and
4153 RFC5289. Generalise some AES* cipherstrings to include GCM and
4154 add a special AESGCM string for GCM only.
4157 *) Expand range of ctrls for AES GCM. Permit setting invocation
4158 field on decrypt and retrieval of invocation field only on encrypt.
4161 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
4162 As required by RFC5289 these ciphersuites cannot be used if for
4163 versions of TLS earlier than 1.2.
4166 *) For FIPS capable OpenSSL interpret a NULL default public key method
4167 as unset and return the appropriate default but do *not* set the default.
4168 This means we can return the appropriate method in applications that
4169 switch between FIPS and non-FIPS modes.
4172 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
4173 ENGINE is used then we cannot handle that in the FIPS module so we
4174 keep original code iff non-FIPS operations are allowed.
4177 *) Add -attime option to openssl utilities.
4178 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
4180 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
4183 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
4184 FIPS EC methods unconditionally for now.
4187 *) New build option no-ec2m to disable characteristic 2 code.
4190 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
4191 all cases can be covered as some introduce binary incompatibilities.
4194 *) Redirect RSA operations to FIPS module including keygen,
4195 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
4198 *) Add similar low level API blocking to ciphers.
4201 *) Low level digest APIs are not approved in FIPS mode: any attempt
4202 to use these will cause a fatal error. Applications that *really* want
4203 to use them can use the private_* version instead.
4206 *) Redirect cipher operations to FIPS module for FIPS builds.
4209 *) Redirect digest operations to FIPS module for FIPS builds.
4212 *) Update build system to add "fips" flag which will link in fipscanister.o
4213 for static and shared library builds embedding a signature if needed.
4216 *) Output TLS supported curves in preference order instead of numerical
4217 order. This is currently hardcoded for the highest order curves first.
4218 This should be configurable so applications can judge speed vs strength.
4221 *) Add TLS v1.2 server support for client authentication.
4224 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
4228 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
4229 FIPS modules versions.
4232 *) Add TLS v1.2 client side support for client authentication. Keep cache
4233 of handshake records longer as we don't know the hash algorithm to use
4234 until after the certificate request message is received.
4237 *) Initial TLS v1.2 client support. Add a default signature algorithms
4238 extension including all the algorithms we support. Parse new signature
4239 format in client key exchange. Relax some ECC signing restrictions for
4240 TLS v1.2 as indicated in RFC5246.
4243 *) Add server support for TLS v1.2 signature algorithms extension. Switch
4244 to new signature format when needed using client digest preference.
4245 All server ciphersuites should now work correctly in TLS v1.2. No client
4246 support yet and no support for client certificates.
4249 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
4250 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
4251 ciphersuites. At present only RSA key exchange ciphersuites work with
4252 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
4253 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
4254 and version checking.
4257 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
4258 with this defined it will not be affected by any changes to ssl internal
4259 structures. Add several utility functions to allow openssl application
4260 to work with OPENSSL_NO_SSL_INTERN defined.
4263 *) A long standing patch to add support for SRP from EdelWeb (Peter
4264 Sylvester and Christophe Renou) was integrated.
4265 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
4266 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
4269 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
4272 *) Permit abbreviated handshakes when renegotiating using the function
4273 SSL_renegotiate_abbreviated().
4274 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4276 *) Add call to ENGINE_register_all_complete() to
4277 ENGINE_load_builtin_engines(), so some implementations get used
4278 automatically instead of needing explicit application support.
4281 *) Add support for TLS key exporter as described in RFC5705.
4282 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
4284 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
4285 a few changes are required:
4287 Add SSL_OP_NO_TLSv1_1 flag.
4288 Add TLSv1_1 methods.
4289 Update version checking logic to handle version 1.1.
4290 Add explicit IV handling (ported from DTLS code).
4291 Add command line options to s_client/s_server.
4294 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
4296 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
4297 in CMS and PKCS7 code. When RSA decryption fails use a random key for
4298 content decryption and always return the same error. Note: this attack
4299 needs on average 2^20 messages so it only affects automated senders. The
4300 old behaviour can be re-enabled in the CMS code by setting the
4301 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
4302 an MMA defence is not necessary.
4303 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
4304 this issue. (CVE-2012-0884)
4307 *) Fix CVE-2011-4619: make sure we really are receiving a
4308 client hello before rejecting multiple SGC restarts. Thanks to
4309 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
4312 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
4314 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
4315 Thanks to Antonio Martin, Enterprise Secure Access Research and
4316 Development, Cisco Systems, Inc. for discovering this bug and
4317 preparing a fix. (CVE-2012-0050)
4320 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
4322 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
4323 of the Vaudenay padding oracle attack on CBC mode encryption
4324 which enables an efficient plaintext recovery attack against
4325 the OpenSSL implementation of DTLS. Their attack exploits timing
4326 differences arising during decryption processing. A research
4327 paper describing this attack can be found at:
4328 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
4329 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4330 Security Group at Royal Holloway, University of London
4331 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
4332 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
4333 for preparing the fix. (CVE-2011-4108)
4334 [Robin Seggelmann, Michael Tuexen]
4336 *) Clear bytes used for block padding of SSL 3.0 records.
4338 [Adam Langley (Google)]
4340 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
4341 Kadianakis <desnacked@gmail.com> for discovering this issue and
4342 Adam Langley for preparing the fix. (CVE-2011-4619)
4343 [Adam Langley (Google)]
4345 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
4346 [Andrey Kulikov <amdeich@gmail.com>]
4348 *) Prevent malformed RFC3779 data triggering an assertion failure.
4349 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
4350 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
4351 [Rob Austein <sra@hactrn.net>]
4353 *) Improved PRNG seeding for VOS.
4354 [Paul Green <Paul.Green@stratus.com>]
4356 *) Fix ssl_ciph.c set-up race.
4357 [Adam Langley (Google)]
4359 *) Fix spurious failures in ecdsatest.c.
4360 [Emilia Käsper (Google)]
4362 *) Fix the BIO_f_buffer() implementation (which was mixing different
4363 interpretations of the '..._len' fields).
4364 [Adam Langley (Google)]
4366 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4367 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4368 threads won't reuse the same blinding coefficients.
4370 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4371 lock to call BN_BLINDING_invert_ex, and avoids one use of
4372 BN_BLINDING_update for each BN_BLINDING structure (previously,
4373 the last update always remained unused).
4374 [Emilia Käsper (Google)]
4376 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4377 [Bob Buckholz (Google)]
4379 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4381 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4382 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4383 [Kaspar Brand <ossl@velox.ch>]
4385 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4386 for multi-threaded use of ECDH. (CVE-2011-3210)
4387 [Adam Langley (Google)]
4389 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4392 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4393 signature public key algorithm by using OID xref utilities instead.
4394 Before this you could only use some ECC ciphersuites with SHA1 only.
4397 *) Add protection against ECDSA timing attacks as mentioned in the paper
4398 by Billy Bob Brumley and Nicola Tuveri, see:
4400 http://eprint.iacr.org/2011/232.pdf
4402 [Billy Bob Brumley and Nicola Tuveri]
4404 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4406 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4407 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4409 *) Fix bug in string printing code: if *any* escaping is enabled we must
4410 escape the escape character (backslash) or the resulting string is
4414 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4416 *) Disable code workaround for ancient and obsolete Netscape browsers
4417 and servers: an attacker can use it in a ciphersuite downgrade attack.
4418 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4421 *) Fixed J-PAKE implementation error, originally discovered by
4422 Sebastien Martini, further info and confirmation from Stefan
4423 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4426 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4428 *) Fix extension code to avoid race conditions which can result in a buffer
4429 overrun vulnerability: resumed sessions must not be modified as they can
4430 be shared by multiple threads. CVE-2010-3864
4433 *) Fix WIN32 build system to correctly link an ENGINE directory into
4437 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4439 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4441 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4443 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4445 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4446 context. The operation can be customised via the ctrl mechanism in
4447 case ENGINEs want to include additional functionality.
4450 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4453 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4454 output hashes compatible with older versions of OpenSSL.
4455 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4457 *) Fix compression algorithm handling: if resuming a session use the
4458 compression algorithm of the resumed session instead of determining
4459 it from client hello again. Don't allow server to change algorithm.
4462 *) Add load_crls() function to apps tidying load_certs() too. Add option
4463 to verify utility to allow additional CRLs to be included.
4466 *) Update OCSP request code to permit adding custom headers to the request:
4467 some responders need this.
4470 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4472 [Julia Lawall <julia@diku.dk>]
4474 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4475 needlessly dereferenced structures, used obsolete functions and
4476 didn't handle all updated verify codes correctly.
4479 *) Disable MD2 in the default configuration.
4482 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4483 indicate the initial BIO being pushed or popped. This makes it possible
4484 to determine whether the BIO is the one explicitly called or as a result
4485 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4486 it handles reference counts correctly and doesn't zero out the I/O bio
4487 when it is not being explicitly popped. WARNING: applications which
4488 included workarounds for the old buggy behaviour will need to be modified
4489 or they could free up already freed BIOs.
4492 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4493 renaming to all platforms (within the 0.9.8 branch, this was
4494 done conditionally on Netware platforms to avoid a name clash).
4495 [Guenter <lists@gknw.net>]
4497 *) Add ECDHE and PSK support to DTLS.
4498 [Michael Tuexen <tuexen@fh-muenster.de>]
4500 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4504 *) Add "missing" function EVP_MD_flags() (without this the only way to
4505 retrieve a digest flags is by accessing the structure directly. Update
4506 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4507 or cipher is registered as in the "from" argument. Print out all
4508 registered digests in the dgst usage message instead of manually
4509 attempting to work them out.
4512 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4513 this allows the use of compression and extensions. Change default cipher
4514 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4515 by default unless an application cipher string requests it.
4518 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4519 key ids to find matching certificates and keys but some PKCS#12 files
4520 don't follow the (somewhat unwritten) rules and this strategy fails.
4521 Now just gather all certificates together and the first private key
4522 then look for the first certificate that matches the key.
4525 *) Support use of registered digest and cipher names for dgst and cipher
4526 commands instead of having to add each one as a special case. So now
4533 openssl dgst -sha256 foo
4535 and this works for ENGINE based algorithms too.
4539 *) Update Gost ENGINE to support parameter files.
4540 [Victor B. Wagner <vitus@cryptocom.ru>]
4542 *) Support GeneralizedTime in ca utility.
4543 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4545 *) Enhance the hash format used for certificate directory links. The new
4546 form uses the canonical encoding (meaning equivalent names will work
4547 even if they aren't identical) and uses SHA1 instead of MD5. This form
4548 is incompatible with the older format and as a result c_rehash should
4549 be used to rebuild symbolic links.
4552 *) Make PKCS#8 the default write format for private keys, replacing the
4553 traditional format. This form is standardised, more secure and doesn't
4554 include an implicit MD5 dependency.
4557 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4558 committed to OpenSSL should pass this lot as a minimum.
4561 *) Add session ticket override functionality for use by EAP-FAST.
4562 [Jouni Malinen <j@w1.fi>]
4564 *) Modify HMAC functions to return a value. Since these can be implemented
4565 in an ENGINE errors can occur.
4568 *) Type-checked OBJ_bsearch_ex.
4571 *) Type-checked OBJ_bsearch. Also some constification necessitated
4572 by type-checking. Still to come: TXT_DB, bsearch(?),
4573 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4577 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4578 seconds to a tm structure directly, instead of going through OS
4579 specific date routines. This avoids any issues with OS routines such
4580 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4581 and X509_time_adj_ex() to cover the extended range. The existing
4582 X509_time_adj() is still usable and will no longer have any date issues.
4585 *) Delta CRL support. New use deltas option which will attempt to locate
4586 and search any appropriate delta CRLs available.
4588 This work was sponsored by Google.
4591 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4592 code and add additional score elements. Validate alternate CRL paths
4593 as part of the CRL checking and indicate a new error "CRL path validation
4594 error" in this case. Applications wanting additional details can use
4595 the verify callback and check the new "parent" field. If this is not
4596 NULL CRL path validation is taking place. Existing applications won't
4597 see this because it requires extended CRL support which is off by
4600 This work was sponsored by Google.
4603 *) Support for freshest CRL extension.
4605 This work was sponsored by Google.
4608 *) Initial indirect CRL support. Currently only supported in the CRLs
4609 passed directly and not via lookup. Process certificate issuer
4610 CRL entry extension and lookup CRL entries by bother issuer name
4611 and serial number. Check and process CRL issuer entry in IDP extension.
4613 This work was sponsored by Google.
4616 *) Add support for distinct certificate and CRL paths. The CRL issuer
4617 certificate is validated separately in this case. Only enabled if
4618 an extended CRL support flag is set: this flag will enable additional
4619 CRL functionality in future.
4621 This work was sponsored by Google.
4624 *) Add support for policy mappings extension.
4626 This work was sponsored by Google.
4629 *) Fixes to pathlength constraint, self issued certificate handling,
4630 policy processing to align with RFC3280 and PKITS tests.
4632 This work was sponsored by Google.
4635 *) Support for name constraints certificate extension. DN, email, DNS
4636 and URI types are currently supported.
4638 This work was sponsored by Google.
4641 *) To cater for systems that provide a pointer-based thread ID rather
4642 than numeric, deprecate the current numeric thread ID mechanism and
4643 replace it with a structure and associated callback type. This
4644 mechanism allows a numeric "hash" to be extracted from a thread ID in
4645 either case, and on platforms where pointers are larger than 'long',
4646 mixing is done to help ensure the numeric 'hash' is usable even if it
4647 can't be guaranteed unique. The default mechanism is to use "&errno"
4648 as a pointer-based thread ID to distinguish between threads.
4650 Applications that want to provide their own thread IDs should now use
4651 CRYPTO_THREADID_set_callback() to register a callback that will call
4652 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4654 Note that ERR_remove_state() is now deprecated, because it is tied
4655 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4656 to free the current thread's error state should be replaced by
4657 ERR_remove_thread_state(NULL).
4659 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4660 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4661 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4662 application was previously providing a numeric thread callback that
4663 was inappropriate for distinguishing threads, then uniqueness might
4664 have been obtained with &errno that happened immediately in the
4665 intermediate development versions of OpenSSL; this is no longer the
4666 case, the numeric thread callback will now override the automatic use
4668 [Geoff Thorpe, with help from Bodo Moeller]
4670 *) Initial support for different CRL issuing certificates. This covers a
4671 simple case where the self issued certificates in the chain exist and
4672 the real CRL issuer is higher in the existing chain.
4674 This work was sponsored by Google.
4677 *) Removed effectively defunct crypto/store from the build.
4680 *) Revamp of STACK to provide stronger type-checking. Still to come:
4681 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4682 ASN1_STRING, CONF_VALUE.
4685 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4686 RAM on SSL connections. This option can save about 34k per idle SSL.
4689 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4690 STACK, TXT_DB, bsearch, qsort.
4693 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4694 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4695 support for data, signedData, compressedData, digestedData and
4696 encryptedData, envelopedData types included. Scripts to check against
4697 RFC4134 examples draft and interop and consistency checks of many
4698 content types and variants.
4701 *) Add options to enc utility to support use of zlib compression BIO.
4704 *) Extend mk1mf to support importing of options and assembly language
4705 files from Configure script, currently only included in VC-WIN32.
4706 The assembly language rules can now optionally generate the source
4707 files from the associated perl scripts.
4710 *) Implement remaining functionality needed to support GOST ciphersuites.
4711 Interop testing has been performed using CryptoPro implementations.
4712 [Victor B. Wagner <vitus@cryptocom.ru>]
4714 *) s390x assembler pack.
4717 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4721 *) Implement Opaque PRF Input TLS extension as specified in
4722 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4723 official specification yet and no extension type assignment by
4724 IANA exists, this extension (for now) will have to be explicitly
4725 enabled when building OpenSSL by providing the extension number
4726 to use. For example, specify an option
4728 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4730 to the "config" or "Configure" script to enable the extension,
4731 assuming extension number 0x9527 (which is a completely arbitrary
4732 and unofficial assignment based on the MD5 hash of the Internet
4733 Draft). Note that by doing so, you potentially lose
4734 interoperability with other TLS implementations since these might
4735 be using the same extension number for other purposes.
4737 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4738 opaque PRF input value to use in the handshake. This will create
4739 an internal copy of the length-'len' string at 'src', and will
4740 return non-zero for success.
4742 To get more control and flexibility, provide a callback function
4745 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4746 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4750 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4753 Callback function 'cb' will be called in handshakes, and is
4754 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4755 Argument 'arg' is for application purposes (the value as given to
4756 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4757 be provided to the callback function). The callback function
4758 has to return non-zero to report success: usually 1 to use opaque
4759 PRF input just if possible, or 2 to enforce use of the opaque PRF
4760 input. In the latter case, the library will abort the handshake
4761 if opaque PRF input is not successfully negotiated.
4763 Arguments 'peerinput' and 'len' given to the callback function
4764 will always be NULL and 0 in the case of a client. A server will
4765 see the client's opaque PRF input through these variables if
4766 available (NULL and 0 otherwise). Note that if the server
4767 provides an opaque PRF input, the length must be the same as the
4768 length of the client's opaque PRF input.
4770 Note that the callback function will only be called when creating
4771 a new session (session resumption can resume whatever was
4772 previously negotiated), and will not be called in SSL 2.0
4773 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4774 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4775 for applications that need to enforce opaque PRF input.
4779 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4782 [Victor B. Wagner <vitus@cryptocom.ru>]
4784 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4785 RFC4507bis. The encrypted ticket format is an encrypted encoded
4786 SSL_SESSION structure, that way new session features are automatically
4789 If a client application caches session in an SSL_SESSION structure
4790 support is transparent because tickets are now stored in the encoded
4793 The SSL_CTX structure automatically generates keys for ticket
4794 protection in servers so again support should be possible
4795 with no application modification.
4797 If a client or server wishes to disable RFC4507 support then the option
4798 SSL_OP_NO_TICKET can be set.
4800 Add a TLS extension debugging callback to allow the contents of any client
4801 or server extensions to be examined.
4803 This work was sponsored by Google.
4806 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4807 OpenSSL should now compile cleanly on gcc 4.2
4808 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4810 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4811 support including streaming MAC support: this is required for GOST
4812 ciphersuite support.
4813 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4815 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4816 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4817 to output in BER and PEM format.
4820 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4821 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4822 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4823 ENGINE support for HMAC keys which are unextractable. New -mac and
4824 -macopt options to dgst utility.
4827 *) New option -sigopt to dgst utility. Update dgst to use
4828 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4829 alternative signing parameters such as X9.31 or PSS in the dgst
4833 *) Change ssl_cipher_apply_rule(), the internal function that does
4834 the work each time a ciphersuite string requests enabling
4835 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4836 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4837 the order of disabled ciphersuites such that those ciphersuites
4838 that most recently went from enabled to disabled not only stay
4839 in order with respect to each other, but also have higher priority
4840 than other disabled ciphersuites the next time ciphersuites are
4843 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4844 the same ciphersuites as with "HIGH" alone, but in a specific
4845 order where the PSK ciphersuites come first (since they are the
4846 most recently disabled ciphersuites when "HIGH" is parsed).
4848 Also, change ssl_create_cipher_list() (using this new
4849 functionality) such that between otherwise identical
4850 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4854 *) Change ssl_create_cipher_list() so that it automatically
4855 arranges the ciphersuites in reasonable order before starting
4856 to process the rule string. Thus, the definition for "DEFAULT"
4857 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4858 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4859 This makes it much easier to arrive at a reasonable default order
4860 in applications for which anonymous ciphers are OK (meaning
4861 that you can't actually use DEFAULT).
4862 [Bodo Moeller; suggested by Victor Duchovni]
4864 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4865 processing) into multiple integers instead of setting
4866 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4867 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4868 (These masks as well as the individual bit definitions are hidden
4869 away into the non-exported interface ssl/ssl_locl.h, so this
4870 change to the definition of the SSL_CIPHER structure shouldn't
4871 affect applications.) This give us more bits for each of these
4872 categories, so there is no longer a need to coagulate AES128 and
4873 AES256 into a single algorithm bit, and to coagulate Camellia128
4874 and Camellia256 into a single algorithm bit, which has led to all
4877 Thus, among other things, the kludge introduced in 0.9.7m and
4878 0.9.8e for masking out AES256 independently of AES128 or masking
4879 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4881 With the change, we also introduce new ciphersuite aliases that
4882 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4886 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4887 Use the leftmost N bytes of the signature input if the input is
4888 larger than the prime q (with N being the size in bytes of q).
4891 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4892 it yet and it is largely untested.
4895 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4898 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4899 some compilers (gcc 4.2 and later) reject their use. Safestack is
4900 reimplemented. Update ASN1 to avoid use of legacy functions.
4903 *) Win32/64 targets are linked with Winsock2.
4906 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4907 to external functions. This can be used to increase CRL handling
4908 efficiency especially when CRLs are very large by (for example) storing
4909 the CRL revoked certificates in a database.
4912 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4913 new CRLs added to a directory can be used. New command line option
4914 -verify_return_error to s_client and s_server. This causes real errors
4915 to be returned by the verify callback instead of carrying on no matter
4916 what. This reflects the way a "real world" verify callback would behave.
4919 *) GOST engine, supporting several GOST algorithms and public key formats.
4920 Kindly donated by Cryptocom.
4923 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4924 partitioned by DP are handled but no indirect CRL or reason partitioning
4925 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4926 selected via a scoring technique which handles IDP and AKID in CRLs.
4929 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4930 will ultimately be used for all verify operations: this will remove the
4931 X509_STORE dependency on certificate verification and allow alternative
4932 lookup methods. X509_STORE based implementations of these two callbacks.
4935 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4936 Modify get_crl() to find a valid (unexpired) CRL if possible.
4939 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4940 this would be called X509_CRL_cmp() but that name is already used by
4941 a function that just compares CRL issuer names. Cache several CRL
4942 extensions in X509_CRL structure and cache CRLDP in X509.
4945 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4946 this maps equivalent X509_NAME structures into a consistent structure.
4947 Name comparison can then be performed rapidly using memcmp().
4950 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4954 *) Allow digests to supply their own micalg string for S/MIME type using
4955 the ctrl EVP_MD_CTRL_MICALG.
4958 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4959 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4960 ctrl. It can then customise the structure before and/or after signing
4964 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4965 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4966 to free up any added signature OIDs.
4969 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4970 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4971 digest and cipher tables. New options added to openssl utility:
4972 list-message-digest-algorithms and list-cipher-algorithms.
4975 *) Change the array representation of binary polynomials: the list
4976 of degrees of non-zero coefficients is now terminated with -1.
4977 Previously it was terminated with 0, which was also part of the
4978 value; thus, the array representation was not applicable to
4979 polynomials where t^0 has coefficient zero. This change makes
4980 the array representation useful in a more general context.
4983 *) Various modifications and fixes to SSL/TLS cipher string
4984 handling. For ECC, the code now distinguishes between fixed ECDH
4985 with RSA certificates on the one hand and with ECDSA certificates
4986 on the other hand, since these are separate ciphersuites. The
4987 unused code for Fortezza ciphersuites has been removed.
4989 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4990 (not "ECDHE"). For consistency with the code for DH
4991 certificates, use of ECDH certificates is now considered ECDH
4992 authentication, not RSA or ECDSA authentication (the latter is
4993 merely the CA's signing algorithm and not actively used in the
4996 The temporary ciphersuite alias "ECCdraft" is no longer
4997 available, and ECC ciphersuites are no longer excluded from "ALL"
4998 and "DEFAULT". The following aliases now exist for RFC 4492
4999 ciphersuites, most of these by analogy with the DH case:
5001 kECDHr - ECDH cert, signed with RSA
5002 kECDHe - ECDH cert, signed with ECDSA
5003 kECDH - ECDH cert (signed with either RSA or ECDSA)
5004 kEECDH - ephemeral ECDH
5005 ECDH - ECDH cert or ephemeral ECDH
5011 AECDH - anonymous ECDH
5012 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
5016 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
5017 Use correct micalg parameters depending on digest(s) in signed message.
5020 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
5021 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
5024 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
5025 an engine to register a method. Add ENGINE lookups for methods and
5026 functional reference processing.
5029 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
5030 EVP_{Sign,Verify}* which allow an application to customise the signature
5034 *) New -resign option to smime utility. This adds one or more signers
5035 to an existing PKCS#7 signedData structure. Also -md option to use an
5036 alternative message digest algorithm for signing.
5039 *) Tidy up PKCS#7 routines and add new functions to make it easier to
5040 create PKCS7 structures containing multiple signers. Update smime
5041 application to support multiple signers.
5044 *) New -macalg option to pkcs12 utility to allow setting of an alternative
5048 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
5049 Reorganize PBE internals to lookup from a static table using NIDs,
5050 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
5051 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
5052 PRF which will be automatically used with PBES2.
5055 *) Replace the algorithm specific calls to generate keys in "req" with the
5059 *) Update PKCS#7 enveloped data routines to use new API. This is now
5060 supported by any public key method supporting the encrypt operation. A
5061 ctrl is added to allow the public key algorithm to examine or modify
5062 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
5066 *) Add a ctrl to asn1 method to allow a public key algorithm to express
5067 a default digest type to use. In most cases this will be SHA1 but some
5068 algorithms (such as GOST) need to specify an alternative digest. The
5069 return value indicates how strong the preference is 1 means optional and
5070 2 is mandatory (that is it is the only supported type). Modify
5071 ASN1_item_sign() to accept a NULL digest argument to indicate it should
5072 use the default md. Update openssl utilities to use the default digest
5073 type for signing if it is not explicitly indicated.
5076 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
5077 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
5078 signing method from the key type. This effectively removes the link
5079 between digests and public key types.
5082 *) Add an OID cross reference table and utility functions. Its purpose is to
5083 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
5084 rsaEncryption. This will allow some of the algorithm specific hackery
5085 needed to use the correct OID to be removed.
5088 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
5089 structures for PKCS7_sign(). They are now set up by the relevant public
5093 *) Add provisional EC pkey method with support for ECDSA and ECDH.
5096 *) Add support for key derivation (agreement) in the API, DH method and
5100 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
5101 public and private key formats. As a side effect these add additional
5102 command line functionality not previously available: DSA signatures can be
5103 generated and verified using pkeyutl and DH key support and generation in
5108 [Oliver Tappe <zooey@hirschkaefer.de>]
5110 *) New make target "install_html_docs" installs HTML renditions of the
5112 [Oliver Tappe <zooey@hirschkaefer.de>]
5114 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
5115 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
5116 support key and parameter generation and add initial key generation
5117 functionality for RSA.
5120 *) Add functions for main EVP_PKEY_method operations. The undocumented
5121 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
5122 EVP_PKEY_{encrypt,decrypt}_old.
5125 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
5126 key API, doesn't do much yet.
5129 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
5130 public key algorithms. New option to openssl utility:
5131 "list-public-key-algorithms" to print out info.
5134 *) Implement the Supported Elliptic Curves Extension for
5135 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5138 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
5139 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
5142 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
5143 utilities such as rsa, dsa, dsaparam etc except they process any key
5147 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
5148 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
5149 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
5153 *) Initial support for pluggable public key ASN1.
5154 De-spaghettify the public key ASN1 handling. Move public and private
5155 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
5156 algorithm specific handling to a single module within the relevant
5157 algorithm directory. Add functions to allow (near) opaque processing
5158 of public and private key structures.
5161 *) Implement the Supported Point Formats Extension for
5162 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5165 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
5166 for the psk identity [hint] and the psk callback functions to the
5167 SSL_SESSION, SSL and SSL_CTX structure.
5170 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
5174 SSL_CTX_use_psk_identity_hint
5175 SSL_get_psk_identity_hint
5176 SSL_get_psk_identity
5177 SSL_use_psk_identity_hint
5179 [Mika Kousa and Pasi Eronen of Nokia Corporation]
5181 *) Add RFC 3161 compliant time stamp request creation, response generation
5182 and response verification functionality.
5183 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
5185 *) Add initial support for TLS extensions, specifically for the server_name
5186 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5187 have new members for a host name. The SSL data structure has an
5188 additional member SSL_CTX *initial_ctx so that new sessions can be
5189 stored in that context to allow for session resumption, even after the
5190 SSL has been switched to a new SSL_CTX in reaction to a client's
5191 server_name extension.
5193 New functions (subject to change):
5195 SSL_get_servername()
5196 SSL_get_servername_type()
5199 New CTRL codes and macros (subject to change):
5201 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5202 - SSL_CTX_set_tlsext_servername_callback()
5203 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5204 - SSL_CTX_set_tlsext_servername_arg()
5205 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5207 openssl s_client has a new '-servername ...' option.
5209 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5210 '-key2 ...', '-servername_fatal' (subject to change). This allows
5211 testing the HostName extension for a specific single host name ('-cert'
5212 and '-key' remain fallbacks for handshakes without HostName
5213 negotiation). If the unrecognized_name alert has to be sent, this by
5214 default is a warning; it becomes fatal with the '-servername_fatal'
5217 [Peter Sylvester, Remy Allais, Christophe Renou]
5219 *) Whirlpool hash implementation is added.
5222 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
5223 bn(64,32). Because of instruction set limitations it doesn't have
5224 any negative impact on performance. This was done mostly in order
5225 to make it possible to share assembler modules, such as bn_mul_mont
5226 implementations, between 32- and 64-bit builds without hassle.
5229 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
5230 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
5234 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
5235 dedicated Montgomery multiplication procedure, is introduced.
5236 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
5237 "64-bit" performance on certain 32-bit targets.
5240 *) New option SSL_OP_NO_COMP to disable use of compression selectively
5241 in SSL structures. New SSL ctrl to set maximum send fragment size.
5242 Save memory by setting the I/O buffer sizes dynamically instead of
5243 using the maximum available value.
5246 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
5247 in addition to the text details.
5250 *) Very, very preliminary EXPERIMENTAL support for printing of general
5251 ASN1 structures. This currently produces rather ugly output and doesn't
5252 handle several customised structures at all.
5255 *) Integrated support for PVK file format and some related formats such
5256 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
5257 these in the 'rsa' and 'dsa' utilities.
5260 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
5263 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
5264 place for the (very old) "NETSCAPE" format certificates which are now
5265 handled using new ASN1 code equivalents.
5268 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
5269 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
5270 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
5273 *) Modify CRL distribution points extension code to print out previously
5274 unsupported fields. Enhance extension setting code to allow setting of
5278 *) Add print and set support for Issuing Distribution Point CRL extension.
5281 *) Change 'Configure' script to enable Camellia by default.
5284 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
5286 *) When rejecting SSL/TLS records due to an incorrect version number, never
5287 update s->server with a new major version number. As of
5288 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
5289 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
5290 the previous behavior could result in a read attempt at NULL when
5291 receiving specific incorrect SSL/TLS records once record payload
5292 protection is active. (CVE-2010-0740)
5293 [Bodo Moeller, Adam Langley <agl@chromium.org>]
5295 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
5296 could be crashed if the relevant tables were not present (e.g. chrooted).
5297 [Tomas Hoger <thoger@redhat.com>]
5299 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
5301 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
5302 [Martin Olsson, Neel Mehta]
5304 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
5305 accommodate for stack sorting, always a write lock!).
5308 *) On some versions of WIN32 Heap32Next is very slow. This can cause
5309 excessive delays in the RAND_poll(): over a minute. As a workaround
5310 include a time check in the inner Heap32Next loop too.
5313 *) The code that handled flushing of data in SSL/TLS originally used the
5314 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
5315 the problem outlined in PR#1949. The fix suggested there however can
5316 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
5317 of Apache). So instead simplify the code to flush unconditionally.
5318 This should be fine since flushing with no data to flush is a no op.
5321 *) Handle TLS versions 2.0 and later properly and correctly use the
5322 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
5323 off ancient servers have a habit of sticking around for a while...
5326 *) Modify compression code so it frees up structures without using the
5327 ex_data callbacks. This works around a problem where some applications
5328 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
5329 restarting) then use compression (e.g. SSL with compression) later.
5330 This results in significant per-connection memory leaks and
5331 has caused some security issues including CVE-2008-1678 and
5335 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
5336 change when encrypting or decrypting.
5339 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
5340 connect and renegotiate with servers which do not support RI.
5341 Until RI is more widely deployed this option is enabled by default.
5344 *) Add "missing" ssl ctrls to clear options and mode.
5347 *) If client attempts to renegotiate and doesn't support RI respond with
5348 a no_renegotiation alert as required by RFC5746. Some renegotiating
5349 TLS clients will continue a connection gracefully when they receive
5350 the alert. Unfortunately OpenSSL mishandled this alert and would hang
5351 waiting for a server hello which it will never receive. Now we treat a
5352 received no_renegotiation alert as a fatal error. This is because
5353 applications requesting a renegotiation might well expect it to succeed
5354 and would have no code in place to handle the server denying it so the
5355 only safe thing to do is to terminate the connection.
5358 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
5359 peer supports secure renegotiation and 0 otherwise. Print out peer
5360 renegotiation support in s_client/s_server.
5363 *) Replace the highly broken and deprecated SPKAC certification method with
5364 the updated NID creation version. This should correctly handle UTF8.
5367 *) Implement RFC5746. Re-enable renegotiation but require the extension
5368 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
5369 turns out to be a bad idea. It has been replaced by
5370 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5371 SSL_CTX_set_options(). This is really not recommended unless you
5372 know what you are doing.
5373 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5375 *) Fixes to stateless session resumption handling. Use initial_ctx when
5376 issuing and attempting to decrypt tickets in case it has changed during
5377 servername handling. Use a non-zero length session ID when attempting
5378 stateless session resumption: this makes it possible to determine if
5379 a resumption has occurred immediately after receiving server hello
5380 (several places in OpenSSL subtly assume this) instead of later in
5384 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5385 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5386 fixes for a few places where the return code is not checked
5388 [Julia Lawall <julia@diku.dk>]
5390 *) Add --strict-warnings option to Configure script to include devteam
5391 warnings in other configurations.
5394 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5395 makes it possible to install openssl libraries in locations which
5396 have names other than "lib", for example "/usr/lib64" which some
5398 [Steve Henson, based on patch from Jeremy Utley]
5400 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5401 X690 8.9.12 and can produce some misleading textual output of OIDs.
5402 [Steve Henson, reported by Dan Kaminsky]
5404 *) Delete MD2 from algorithm tables. This follows the recommendation in
5405 several standards that it is not used in new applications due to
5406 several cryptographic weaknesses. For binary compatibility reasons
5407 the MD2 API is still compiled in by default.
5410 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5414 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5415 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5417 [Guenter <lists@gknw.net>]
5419 *) Fix the server certificate chain building code to use X509_verify_cert(),
5420 it used to have an ad-hoc builder which was unable to cope with anything
5421 other than a simple chain.
5422 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5424 *) Don't check self signed certificate signatures in X509_verify_cert()
5425 by default (a flag can override this): it just wastes time without
5426 adding any security. As a useful side effect self signed root CAs
5427 with non-FIPS digests are now usable in FIPS mode.
5430 *) In dtls1_process_out_of_seq_message() the check if the current message
5431 is already buffered was missing. For every new message was memory
5432 allocated, allowing an attacker to perform an denial of service attack
5433 with sending out of seq handshake messages until there is no memory
5434 left. Additionally every future message was buffered, even if the
5435 sequence number made no sense and would be part of another handshake.
5436 So only messages with sequence numbers less than 10 in advance will be
5437 buffered. (CVE-2009-1378)
5438 [Robin Seggelmann, discovered by Daniel Mentz]
5440 *) Records are buffered if they arrive with a future epoch to be
5441 processed after finishing the corresponding handshake. There is
5442 currently no limitation to this buffer allowing an attacker to perform
5443 a DOS attack with sending records with future epochs until there is no
5444 memory left. This patch adds the pqueue_size() function to determine
5445 the size of a buffer and limits the record buffer to 100 entries.
5447 [Robin Seggelmann, discovered by Daniel Mentz]
5449 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5450 parent structure is freed. (CVE-2009-1379)
5453 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5454 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5457 [Ilya O. <vrghost@gmail.com>]
5459 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5461 *) Disable renegotiation completely - this fixes a severe security
5462 problem (CVE-2009-3555) at the cost of breaking all
5463 renegotiation. Renegotiation can be re-enabled by setting
5464 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5465 run-time. This is really not recommended unless you know what
5469 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5471 *) Don't set val to NULL when freeing up structures, it is freed up by
5472 underlying code. If sizeof(void *) > sizeof(long) this can result in
5473 zeroing past the valid field. (CVE-2009-0789)
5474 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5476 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5477 checked correctly. This would allow some invalid signed attributes to
5478 appear to verify correctly. (CVE-2009-0591)
5479 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5481 *) Reject UniversalString and BMPString types with invalid lengths. This
5482 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5483 a legal length. (CVE-2009-0590)
5486 *) Set S/MIME signing as the default purpose rather than setting it
5487 unconditionally. This allows applications to override it at the store
5491 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5492 to handle some structures.
5495 *) Improve efficiency of mem_gets: don't search whole buffer each time
5497 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5499 *) New -hex option for openssl rand.
5502 *) Print out UTF8String and NumericString when parsing ASN1.
5505 *) Support NumericString type for name components.
5508 *) Allow CC in the environment to override the automatically chosen
5509 compiler. Note that nothing is done to ensure flags work with the
5513 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5515 *) Properly check EVP_VerifyFinal() and similar return values
5517 [Ben Laurie, Bodo Moeller, Google Security Team]
5519 *) Enable TLS extensions by default.
5522 *) Allow the CHIL engine to be loaded, whether the application is
5523 multithreaded or not. (This does not release the developer from the
5524 obligation to set up the dynamic locking callbacks.)
5525 [Sander Temme <sander@temme.net>]
5527 *) Use correct exit code if there is an error in dgst command.
5528 [Steve Henson; problem pointed out by Roland Dirlewanger]
5530 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5531 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5534 *) Add experimental JPAKE support, including demo authentication in
5535 s_client and s_server.
5538 *) Set the comparison function in v3_addr_canonize().
5539 [Rob Austein <sra@hactrn.net>]
5541 *) Add support for XMPP STARTTLS in s_client.
5542 [Philip Paeps <philip@freebsd.org>]
5544 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5545 to ensure that even with this option, only ciphersuites in the
5546 server's preference list will be accepted. (Note that the option
5547 applies only when resuming a session, so the earlier behavior was
5548 just about the algorithm choice for symmetric cryptography.)
5551 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5553 *) Fix NULL pointer dereference if a DTLS server received
5554 ChangeCipherSpec as first record (CVE-2009-1386).
5557 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5558 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5561 *) The fix in 0.9.8c that supposedly got rid of unsafe
5562 double-checked locking was incomplete for RSA blinding,
5563 addressing just one layer of what turns out to have been
5564 doubly unsafe triple-checked locking.
5566 So now fix this for real by retiring the MONT_HELPER macro
5567 in crypto/rsa/rsa_eay.c.
5569 [Bodo Moeller; problem pointed out by Marius Schilder]
5571 *) Various precautionary measures:
5573 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5575 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5576 (NB: This would require knowledge of the secret session ticket key
5577 to exploit, in which case you'd be SOL either way.)
5579 - Change bn_nist.c so that it will properly handle input BIGNUMs
5580 outside the expected range.
5582 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5585 [Neel Mehta, Bodo Moeller]
5587 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5588 the load fails. Useful for distros.
5589 [Ben Laurie and the FreeBSD team]
5591 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5594 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5597 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5599 This work was sponsored by Logica.
5602 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5603 keystores. Support for SSL/TLS client authentication too.
5604 Not compiled unless enable-capieng specified to Configure.
5606 This work was sponsored by Logica.
5609 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5610 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5611 attribute creation routines such as certificate requests and PKCS#12
5615 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5617 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5618 handshake which could lead to a client crash as found using the
5619 Codenomicon TLS test suite (CVE-2008-1672)
5620 [Steve Henson, Mark Cox]
5622 *) Fix double free in TLS server name extensions which could lead to
5623 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5626 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5628 Clear the error queue to ensure that error entries left from
5629 older function calls do not interfere with the correct operation.
5630 [Lutz Jaenicke, Erik de Castro Lopo]
5632 *) Remove root CA certificates of commercial CAs:
5634 The OpenSSL project does not recommend any specific CA and does not
5635 have any policy with respect to including or excluding any CA.
5636 Therefore it does not make any sense to ship an arbitrary selection
5637 of root CA certificates with the OpenSSL software.
5640 *) RSA OAEP patches to fix two separate invalid memory reads.
5641 The first one involves inputs when 'lzero' is greater than
5642 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5643 before the beginning of from). The second one involves inputs where
5644 the 'db' section contains nothing but zeroes (there is a one-byte
5645 invalid read after the end of 'db').
5646 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5648 *) Partial backport from 0.9.9-dev:
5650 Introduce bn_mul_mont (dedicated Montgomery multiplication
5651 procedure) as a candidate for BIGNUM assembler implementation.
5652 While 0.9.9-dev uses assembler for various architectures, only
5653 x86_64 is available by default here in the 0.9.8 branch, and
5654 32-bit x86 is available through a compile-time setting.
5656 To try the 32-bit x86 assembler implementation, use Configure
5657 option "enable-montasm" (which exists only for this backport).
5659 As "enable-montasm" for 32-bit x86 disclaims code stability
5660 anyway, in this constellation we activate additional code
5661 backported from 0.9.9-dev for further performance improvements,
5662 namely BN_from_montgomery_word. (To enable this otherwise,
5663 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5665 [Andy Polyakov (backport partially by Bodo Moeller)]
5667 *) Add TLS session ticket callback. This allows an application to set
5668 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5669 values. This is useful for key rollover for example where several key
5670 sets may exist with different names.
5673 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5674 This was broken until now in 0.9.8 releases, such that the only way
5675 a registered ENGINE could be used (assuming it initialises
5676 successfully on the host) was to explicitly set it as the default
5677 for the relevant algorithms. This is in contradiction with 0.9.7
5678 behaviour and the documentation. With this fix, when an ENGINE is
5679 registered into a given algorithm's table of implementations, the
5680 'uptodate' flag is reset so that auto-discovery will be used next
5681 time a new context for that algorithm attempts to select an
5683 [Ian Lister (tweaked by Geoff Thorpe)]
5685 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5686 implementation in the following ways:
5688 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5691 Lack of BER streaming support means one pass streaming processing is
5692 only supported if data is detached: setting the streaming flag is
5693 ignored for embedded content.
5695 CMS support is disabled by default and must be explicitly enabled
5696 with the enable-cms configuration option.
5699 *) Update the GMP engine glue to do direct copies between BIGNUM and
5700 mpz_t when openssl and GMP use the same limb size. Otherwise the
5701 existing "conversion via a text string export" trick is still used.
5702 [Paul Sheer <paulsheer@gmail.com>]
5704 *) Zlib compression BIO. This is a filter BIO which compressed and
5705 uncompresses any data passed through it.
5708 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5709 RFC3394 compatible AES key wrapping.
5712 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5713 sets string data without copying. X509_ALGOR_set0() and
5714 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5715 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5716 from an X509_ATTRIBUTE structure optionally checking it occurs only
5717 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5721 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5722 to get the expected BN_FLG_CONSTTIME behavior.
5723 [Bodo Moeller (Google)]
5727 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5728 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5729 - added some more tests to do_tests.pl
5730 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5731 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5732 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5733 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5734 - various changes to netware.pl to enable gcc-cross builds on Win32
5736 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5737 - various changes to fix missing prototype warnings
5738 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5739 - added AES, WHIRLPOOL and CPUID assembler code to build files
5740 - added missing AES assembler make rules to mk1mf.pl
5741 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5742 [Guenter Knauf <eflash@gmx.net>]
5744 *) Implement certificate status request TLS extension defined in RFC3546.
5745 A client can set the appropriate parameters and receive the encoded
5746 OCSP response via a callback. A server can query the supplied parameters
5747 and set the encoded OCSP response in the callback. Add simplified examples
5748 to s_client and s_server.
5751 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5753 *) Fix various bugs:
5754 + Binary incompatibility of ssl_ctx_st structure
5755 + DTLS interoperation with non-compliant servers
5756 + Don't call get_session_cb() without proposed session
5757 + Fix ia64 assembler code
5758 [Andy Polyakov, Steve Henson]
5760 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5762 *) DTLS Handshake overhaul. There were longstanding issues with
5763 OpenSSL DTLS implementation, which were making it impossible for
5764 RFC 4347 compliant client to communicate with OpenSSL server.
5765 Unfortunately just fixing these incompatibilities would "cut off"
5766 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5767 server keeps tolerating non RFC compliant syntax. The opposite is
5768 not true, 0.9.8f client can not communicate with earlier server.
5769 This update even addresses CVE-2007-4995.
5772 *) Changes to avoid need for function casts in OpenSSL: some compilers
5773 (gcc 4.2 and later) reject their use.
5774 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5777 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5778 RFC4507bis. The encrypted ticket format is an encrypted encoded
5779 SSL_SESSION structure, that way new session features are automatically
5782 If a client application caches session in an SSL_SESSION structure
5783 support is transparent because tickets are now stored in the encoded
5786 The SSL_CTX structure automatically generates keys for ticket
5787 protection in servers so again support should be possible
5788 with no application modification.
5790 If a client or server wishes to disable RFC4507 support then the option
5791 SSL_OP_NO_TICKET can be set.
5793 Add a TLS extension debugging callback to allow the contents of any client
5794 or server extensions to be examined.
5796 This work was sponsored by Google.
5799 *) Add initial support for TLS extensions, specifically for the server_name
5800 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5801 have new members for a host name. The SSL data structure has an
5802 additional member SSL_CTX *initial_ctx so that new sessions can be
5803 stored in that context to allow for session resumption, even after the
5804 SSL has been switched to a new SSL_CTX in reaction to a client's
5805 server_name extension.
5807 New functions (subject to change):
5809 SSL_get_servername()
5810 SSL_get_servername_type()
5813 New CTRL codes and macros (subject to change):
5815 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5816 - SSL_CTX_set_tlsext_servername_callback()
5817 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5818 - SSL_CTX_set_tlsext_servername_arg()
5819 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5821 openssl s_client has a new '-servername ...' option.
5823 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5824 '-key2 ...', '-servername_fatal' (subject to change). This allows
5825 testing the HostName extension for a specific single host name ('-cert'
5826 and '-key' remain fallbacks for handshakes without HostName
5827 negotiation). If the unrecognized_name alert has to be sent, this by
5828 default is a warning; it becomes fatal with the '-servername_fatal'
5831 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5833 *) Add AES and SSE2 assembly language support to VC++ build.
5836 *) Mitigate attack on final subtraction in Montgomery reduction.
5839 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5840 (which previously caused an internal error).
5843 *) Squeeze another 10% out of IGE mode when in != out.
5846 *) AES IGE mode speedup.
5847 [Dean Gaudet (Google)]
5849 *) Add the Korean symmetric 128-bit cipher SEED (see
5850 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5851 add SEED ciphersuites from RFC 4162:
5853 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5854 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5855 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5856 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5858 To minimize changes between patchlevels in the OpenSSL 0.9.8
5859 series, SEED remains excluded from compilation unless OpenSSL
5860 is configured with 'enable-seed'.
5861 [KISA, Bodo Moeller]
5863 *) Mitigate branch prediction attacks, which can be practical if a
5864 single processor is shared, allowing a spy process to extract
5865 information. For detailed background information, see
5866 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5867 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5868 and Necessary Software Countermeasures"). The core of the change
5869 are new versions BN_div_no_branch() and
5870 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5871 respectively, which are slower, but avoid the security-relevant
5872 conditional branches. These are automatically called by BN_div()
5873 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5874 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5875 remove a conditional branch.
5877 BN_FLG_CONSTTIME is the new name for the previous
5878 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5879 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5880 in the exponent causes BN_mod_exp_mont() to use the alternative
5881 implementation in BN_mod_exp_mont_consttime().) The old name
5882 remains as a deprecated alias.
5884 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5885 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5886 constant-time implementations for more than just exponentiation.
5887 Here too the old name is kept as a deprecated alias.
5889 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5890 the BN_BLINDING structure gets an independent copy of the
5891 modulus. This means that the previous "BIGNUM *m" argument to
5892 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5893 essentially becomes "const BIGNUM *m", although we can't actually
5894 change this in the header file before 0.9.9. It allows
5895 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5896 enable BN_FLG_CONSTTIME.
5898 [Matthew D Wood (Intel Corp)]
5900 *) In the SSL/TLS server implementation, be strict about session ID
5901 context matching (which matters if an application uses a single
5902 external cache for different purposes). Previously,
5903 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5904 set. This did ensure strict client verification, but meant that,
5905 with applications using a single external cache for quite
5906 different requirements, clients could circumvent ciphersuite
5907 restrictions for a given session ID context by starting a session
5908 in a different context.
5911 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5912 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5913 authentication-only ciphersuites.
5916 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5917 not complete and could lead to a possible single byte overflow
5918 (CVE-2007-5135) [Ben Laurie]
5920 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5922 *) Since AES128 and AES256 (and similarly Camellia128 and
5923 Camellia256) share a single mask bit in the logic of
5924 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5925 kludge to work properly if AES128 is available and AES256 isn't
5926 (or if Camellia128 is available and Camellia256 isn't).
5929 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5930 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5931 When a point or a seed is encoded in a BIT STRING, we need to
5932 prevent the removal of trailing zero bits to get the proper DER
5933 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5934 of a NamedBitList, for which trailing 0 bits need to be removed.)
5937 *) Have SSL/TLS server implementation tolerate "mismatched" record
5938 protocol version while receiving ClientHello even if the
5939 ClientHello is fragmented. (The server can't insist on the
5940 particular protocol version it has chosen before the ServerHello
5941 message has informed the client about his choice.)
5944 *) Add RFC 3779 support.
5945 [Rob Austein for ARIN, Ben Laurie]
5947 *) Load error codes if they are not already present instead of using a
5948 static variable. This allows them to be cleanly unloaded and reloaded.
5949 Improve header file function name parsing.
5952 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5953 or CAPABILITY handshake as required by RFCs.
5956 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5958 *) Introduce limits to prevent malicious keys being able to
5959 cause a denial of service. (CVE-2006-2940)
5960 [Steve Henson, Bodo Moeller]
5962 *) Fix ASN.1 parsing of certain invalid structures that can result
5963 in a denial of service. (CVE-2006-2937) [Steve Henson]
5965 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5966 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5968 *) Fix SSL client code which could crash if connecting to a
5969 malicious SSLv2 server. (CVE-2006-4343)
5970 [Tavis Ormandy and Will Drewry, Google Security Team]
5972 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5973 match only those. Before that, "AES256-SHA" would be interpreted
5974 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5975 the same strength classification in 0.9.7h) as we currently only
5976 have a single AES bit in the ciphersuite description bitmap.
5977 That change, however, also applied to ciphersuite strings such as
5978 "RC4-MD5" that intentionally matched multiple ciphersuites --
5979 namely, SSL 2.0 ciphersuites in addition to the more common ones
5980 from SSL 3.0/TLS 1.0.
5982 So we change the selection algorithm again: Naming an explicit
5983 ciphersuite selects this one ciphersuite, and any other similar
5984 ciphersuite (same bitmap) from *other* protocol versions.
5985 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5986 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5988 Since SSL 2.0 does not have any ciphersuites for which the
5989 128/256 bit distinction would be relevant, this works for now.
5990 The proper fix will be to use different bits for AES128 and
5991 AES256, which would have avoided the problems from the beginning;
5992 however, bits are scarce, so we can only do this in a new release
5993 (not just a patchlevel) when we can change the SSL_CIPHER
5994 definition to split the single 'unsigned long mask' bitmap into
5995 multiple values to extend the available space.
5999 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
6001 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6002 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6004 *) Add AES IGE and biIGE modes.
6007 *) Change the Unix randomness entropy gathering to use poll() when
6008 possible instead of select(), since the latter has some
6009 undesirable limitations.
6010 [Darryl Miles via Richard Levitte and Bodo Moeller]
6012 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
6013 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
6014 cannot be implicitly activated as part of, e.g., the "AES" alias.
6015 However, please upgrade to OpenSSL 0.9.9[-dev] for
6016 non-experimental use of the ECC ciphersuites to get TLS extension
6017 support, which is required for curve and point format negotiation
6018 to avoid potential handshake problems.
6021 *) Disable rogue ciphersuites:
6023 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6024 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6025 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6027 The latter two were purportedly from
6028 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6031 Also deactivate the remaining ciphersuites from
6032 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6033 unofficial, and the ID has long expired.
6036 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6037 dual-core machines) and other potential thread-safety issues.
6040 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
6041 versions), which is now available for royalty-free use
6042 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
6043 Also, add Camellia TLS ciphersuites from RFC 4132.
6045 To minimize changes between patchlevels in the OpenSSL 0.9.8
6046 series, Camellia remains excluded from compilation unless OpenSSL
6047 is configured with 'enable-camellia'.
6050 *) Disable the padding bug check when compression is in use. The padding
6051 bug check assumes the first packet is of even length, this is not
6052 necessarily true if compression is enabled and can result in false
6053 positives causing handshake failure. The actual bug test is ancient
6054 code so it is hoped that implementations will either have fixed it by
6055 now or any which still have the bug do not support compression.
6058 Changes between 0.9.8a and 0.9.8b [04 May 2006]
6060 *) When applying a cipher rule check to see if string match is an explicit
6061 cipher suite and only match that one cipher suite if it is.
6064 *) Link in manifests for VC++ if needed.
6065 [Austin Ziegler <halostatue@gmail.com>]
6067 *) Update support for ECC-based TLS ciphersuites according to
6068 draft-ietf-tls-ecc-12.txt with proposed changes (but without
6069 TLS extensions, which are supported starting with the 0.9.9
6070 branch, not in the OpenSSL 0.9.8 branch).
6073 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
6074 opaque EVP_CIPHER_CTX handling.
6077 *) Fixes and enhancements to zlib compression code. We now only use
6078 "zlib1.dll" and use the default __cdecl calling convention on Win32
6079 to conform with the standards mentioned here:
6080 http://www.zlib.net/DLL_FAQ.txt
6081 Static zlib linking now works on Windows and the new --with-zlib-include
6082 --with-zlib-lib options to Configure can be used to supply the location
6083 of the headers and library. Gracefully handle case where zlib library
6087 *) Several fixes and enhancements to the OID generation code. The old code
6088 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
6089 handle numbers larger than ULONG_MAX, truncated printing and had a
6090 non standard OBJ_obj2txt() behaviour.
6093 *) Add support for building of engines under engine/ as shared libraries
6094 under VC++ build system.
6097 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
6098 Hopefully, we will not see any false combination of paths any more.
6101 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
6103 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6104 (part of SSL_OP_ALL). This option used to disable the
6105 countermeasure against man-in-the-middle protocol-version
6106 rollback in the SSL 2.0 server implementation, which is a bad
6107 idea. (CVE-2005-2969)
6109 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6110 for Information Security, National Institute of Advanced Industrial
6111 Science and Technology [AIST], Japan)]
6113 *) Add two function to clear and return the verify parameter flags.
6116 *) Keep cipherlists sorted in the source instead of sorting them at
6117 runtime, thus removing the need for a lock.
6120 *) Avoid some small subgroup attacks in Diffie-Hellman.
6121 [Nick Mathewson and Ben Laurie]
6123 *) Add functions for well-known primes.
6126 *) Extended Windows CE support.
6127 [Satoshi Nakamura and Andy Polyakov]
6129 *) Initialize SSL_METHOD structures at compile time instead of during
6130 runtime, thus removing the need for a lock.
6133 *) Make PKCS7_decrypt() work even if no certificate is supplied by
6134 attempting to decrypt each encrypted key in turn. Add support to
6138 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
6140 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
6143 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
6146 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
6147 key into the same file any more.
6150 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
6153 *) Add -utf8 command line and config file option to 'ca'.
6154 [Stefan <stf@udoma.org]
6156 *) Removed the macro des_crypt(), as it seems to conflict with some
6157 libraries. Use DES_crypt().
6160 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
6161 involves renaming the source and generated shared-libs for
6162 both. The engines will accept the corrected or legacy ids
6163 ('ncipher' and '4758_cca' respectively) when binding. NB,
6164 this only applies when building 'shared'.
6165 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
6167 *) Add attribute functions to EVP_PKEY structure. Modify
6168 PKCS12_create() to recognize a CSP name attribute and
6169 use it. Make -CSP option work again in pkcs12 utility.
6172 *) Add new functionality to the bn blinding code:
6173 - automatic re-creation of the BN_BLINDING parameters after
6174 a fixed number of uses (currently 32)
6175 - add new function for parameter creation
6176 - introduce flags to control the update behaviour of the
6177 BN_BLINDING parameters
6178 - hide BN_BLINDING structure
6179 Add a second BN_BLINDING slot to the RSA structure to improve
6180 performance when a single RSA object is shared among several
6184 *) Add support for DTLS.
6185 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
6187 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
6188 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
6191 *) Remove buggy and incomplete DH cert support from
6192 ssl/ssl_rsa.c and ssl/s3_both.c
6195 *) Use SHA-1 instead of MD5 as the default digest algorithm for
6196 the apps/openssl applications.
6199 *) Compile clean with "-Wall -Wmissing-prototypes
6200 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
6201 DEBUG_SAFESTACK must also be set.
6204 *) Change ./Configure so that certain algorithms can be disabled by default.
6205 The new counterpiece to "no-xxx" is "enable-xxx".
6207 The patented RC5 and MDC2 algorithms will now be disabled unless
6208 "enable-rc5" and "enable-mdc2", respectively, are specified.
6210 (IDEA remains enabled despite being patented. This is because IDEA
6211 is frequently required for interoperability, and there is no license
6212 fee for non-commercial use. As before, "no-idea" can be used to
6213 avoid this algorithm.)
6217 *) Add processing of proxy certificates (see RFC 3820). This work was
6218 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
6219 EGEE (Enabling Grids for E-science in Europe).
6222 *) RC4 performance overhaul on modern architectures/implementations, such
6223 as Intel P4, IA-64 and AMD64.
6226 *) New utility extract-section.pl. This can be used specify an alternative
6227 section number in a pod file instead of having to treat each file as
6228 a separate case in Makefile. This can be done by adding two lines to the
6231 =for comment openssl_section:XXX
6233 The blank line is mandatory.
6237 *) New arguments -certform, -keyform and -pass for s_client and s_server
6238 to allow alternative format key and certificate files and passphrase
6242 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
6243 update associated structures and add various utility functions.
6245 Add new policy related verify parameters, include policy checking in
6246 standard verify code. Enhance 'smime' application with extra parameters
6247 to support policy checking and print out.
6250 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
6251 Nehemiah processors. These extensions support AES encryption in hardware
6252 as well as RNG (though RNG support is currently disabled).
6253 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
6255 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
6258 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
6259 [Andy Polyakov and a number of other people]
6261 *) Improved PowerPC platform support. Most notably BIGNUM assembler
6262 implementation contributed by IBM.
6263 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
6265 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
6266 exponent rather than 'unsigned long'. There is a corresponding change to
6267 the new 'rsa_keygen' element of the RSA_METHOD structure.
6268 [Jelte Jansen, Geoff Thorpe]
6270 *) Functionality for creating the initial serial number file is now
6271 moved from CA.pl to the 'ca' utility with a new option -create_serial.
6273 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
6274 number file to 1, which is bound to cause problems. To avoid
6275 the problems while respecting compatibility between different 0.9.7
6276 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
6277 CA.pl for serial number initialization. With the new release 0.9.8,
6278 we can fix the problem directly in the 'ca' utility.)
6281 *) Reduced header interdependencies by declaring more opaque objects in
6282 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
6283 give fewer recursive includes, which could break lazy source code - so
6284 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
6285 developers should define this symbol when building and using openssl to
6286 ensure they track the recommended behaviour, interfaces, [etc], but
6287 backwards-compatible behaviour prevails when this isn't defined.
6290 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
6293 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
6294 This will generate a random key of the appropriate length based on the
6295 cipher context. The EVP_CIPHER can provide its own random key generation
6296 routine to support keys of a specific form. This is used in the des and
6297 3des routines to generate a key of the correct parity. Update S/MIME
6298 code to use new functions and hence generate correct parity DES keys.
6299 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
6300 valid (weak or incorrect parity).
6303 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
6304 as looking them up. This is useful when the verified structure may contain
6305 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
6306 present unless the new PKCS7_NO_CRL flag is asserted.
6309 *) Extend ASN1 oid configuration module. It now additionally accepts the
6312 shortName = some long name, 1.2.3.4
6315 *) Reimplemented the BN_CTX implementation. There is now no more static
6316 limitation on the number of variables it can handle nor the depth of the
6317 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
6318 information can now expand as required, and rather than having a single
6319 static array of bignums, BN_CTX now uses a linked-list of such arrays
6320 allowing it to expand on demand whilst maintaining the usefulness of
6321 BN_CTX's "bundling".
6324 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
6325 to allow all RSA operations to function using a single BN_CTX.
6328 *) Preliminary support for certificate policy evaluation and checking. This
6329 is initially intended to pass the tests outlined in "Conformance Testing
6330 of Relying Party Client Certificate Path Processing Logic" v1.07.
6333 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
6334 remained unused and not that useful. A variety of other little bignum
6335 tweaks and fixes have also been made continuing on from the audit (see
6339 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
6340 associated ASN1, EVP and SSL functions and old ASN1 macros.
6343 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
6344 and this should never fail. So the return value from the use of
6345 BN_set_word() (which can fail due to needless expansion) is now deprecated;
6346 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
6349 *) BN_CTX_get() should return zero-valued bignums, providing the same
6350 initialised value as BN_new().
6351 [Geoff Thorpe, suggested by Ulf Möller]
6353 *) Support for inhibitAnyPolicy certificate extension.
6356 *) An audit of the BIGNUM code is underway, for which debugging code is
6357 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
6358 is considered valid when processing BIGNUMs, and causes execution to
6359 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
6360 further steps are taken to deliberately pollute unused data in BIGNUM
6361 structures to try and expose faulty code further on. For now, openssl will
6362 (in its default mode of operation) continue to tolerate the inconsistent
6363 forms that it has tolerated in the past, but authors and packagers should
6364 consider trying openssl and their own applications when compiled with
6365 these debugging symbols defined. It will help highlight potential bugs in
6366 their own code, and will improve the test coverage for OpenSSL itself. At
6367 some point, these tighter rules will become openssl's default to improve
6368 maintainability, though the assert()s and other overheads will remain only
6369 in debugging configurations. See bn.h for more details.
6370 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6372 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6373 that can only be obtained through BN_CTX_new() (which implicitly
6374 initialises it). The presence of this function only made it possible
6375 to overwrite an existing structure (and cause memory leaks).
6378 *) Because of the callback-based approach for implementing LHASH as a
6379 template type, lh_insert() adds opaque objects to hash-tables and
6380 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6381 to clean up those corresponding objects before destroying the hash table
6382 (and losing the object pointers). So some over-zealous constifications in
6383 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6384 objects as "const" and the lh_doall[_arg] callback wrappers are not
6385 prototyped to have "const" restrictions on the object pointers they are
6386 given (and so aren't required to cast them away any more).
6389 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6390 (speed) prefers to use its own implementation. The two implementations
6391 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6392 its object type properly exposed (MS_TM) instead of casting to/from "char
6393 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6394 aren't necessarily the greatest nomenclatures - but this is what was used
6395 internally to the implementation so I've used that for now.
6398 *) Ensure that deprecated functions do not get compiled when
6399 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6400 the self-tests were still using deprecated key-generation functions so
6401 these have been updated also.
6404 *) Reorganise PKCS#7 code to separate the digest location functionality
6405 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6406 New function PKCS7_set_digest() to set the digest type for PKCS#7
6407 digestedData type. Add additional code to correctly generate the
6408 digestedData type and add support for this type in PKCS7 initialization
6412 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6413 structure of type "other".
6416 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6417 sure the loop does correctly stop and breaking ("division by zero")
6418 modulus operations are not performed. The (pre-generated) prime
6419 table crypto/bn/bn_prime.h was already correct, but it could not be
6420 re-generated on some platforms because of the "division by zero"
6421 situation in the script.
6422 [Ralf S. Engelschall]
6424 *) Update support for ECC-based TLS ciphersuites according to
6425 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6426 SHA-1 now is only used for "small" curves (where the
6427 representation of a field element takes up to 24 bytes); for
6428 larger curves, the field element resulting from ECDH is directly
6429 used as premaster secret.
6430 [Douglas Stebila (Sun Microsystems Laboratories)]
6432 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6433 curve secp160r1 to the tests.
6434 [Douglas Stebila (Sun Microsystems Laboratories)]
6436 *) Add the possibility to load symbols globally with DSO.
6437 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6439 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6440 control of the error stack.
6443 *) Add support for STORE in ENGINE.
6446 *) Add the STORE type. The intention is to provide a common interface
6447 to certificate and key stores, be they simple file-based stores, or
6448 HSM-type store, or LDAP stores, or...
6449 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6452 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6453 pass a list of arguments to any function as well as provide a way
6454 for a function to pass data back to the caller.
6457 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6458 works like BUF_strdup() but can be used to duplicate a portion of
6459 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6463 *) Add the function sk_find_ex() which works like sk_find(), but will
6464 return an index to an element even if an exact match couldn't be
6465 found. The index is guaranteed to point at the element where the
6466 searched-for key would be inserted to preserve sorting order.
6469 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6470 takes an extra flags argument for optional functionality. Currently,
6471 the following flags are defined:
6473 OBJ_BSEARCH_VALUE_ON_NOMATCH
6474 This one gets OBJ_bsearch_ex() to return a pointer to the first
6475 element where the comparing function returns a negative or zero
6478 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6479 This one gets OBJ_bsearch_ex() to return a pointer to the first
6480 element where the comparing function returns zero. This is useful
6481 if there are more than one element where the comparing function
6485 *) Make it possible to create self-signed certificates with 'openssl ca'
6486 in such a way that the self-signed certificate becomes part of the
6487 CA database and uses the same mechanisms for serial number generation
6488 as all other certificate signing. The new flag '-selfsign' enables
6489 this functionality. Adapt CA.sh and CA.pl.in.
6492 *) Add functionality to check the public key of a certificate request
6493 against a given private. This is useful to check that a certificate
6494 request can be signed by that key (self-signing).
6497 *) Make it possible to have multiple active certificates with the same
6498 subject in the CA index file. This is done only if the keyword
6499 'unique_subject' is set to 'no' in the main CA section (default
6500 if 'CA_default') of the configuration file. The value is saved
6501 with the database itself in a separate index attribute file,
6502 named like the index file with '.attr' appended to the name.
6505 *) Generate multi-valued AVAs using '+' notation in config files for
6509 *) Support for nameConstraints certificate extension.
6512 *) Support for policyConstraints certificate extension.
6515 *) Support for policyMappings certificate extension.
6518 *) Make sure the default DSA_METHOD implementation only uses its
6519 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6520 and change its own handlers to be NULL so as to remove unnecessary
6521 indirection. This lets alternative implementations fallback to the
6522 default implementation more easily.
6525 *) Support for directoryName in GeneralName related extensions
6529 *) Make it possible to link applications using Makefile.shared.
6530 Make that possible even when linking against static libraries!
6533 *) Support for single pass processing for S/MIME signing. This now
6534 means that S/MIME signing can be done from a pipe, in addition
6535 cleartext signing (multipart/signed type) is effectively streaming
6536 and the signed data does not need to be all held in memory.
6538 This is done with a new flag PKCS7_STREAM. When this flag is set
6539 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6540 is done after the data is output (and digests calculated) in
6541 SMIME_write_PKCS7().
6544 *) Add full support for -rpath/-R, both in shared libraries and
6545 applications, at least on the platforms where it's known how
6549 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6550 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6551 will now compute a table of multiples of the generator that
6552 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6553 faster (notably in the case of a single point multiplication,
6554 scalar * generator).
6555 [Nils Larsch, Bodo Moeller]
6557 *) IPv6 support for certificate extensions. The various extensions
6558 which use the IP:a.b.c.d can now take IPv6 addresses using the
6559 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6563 *) Added an ENGINE that implements RSA by performing private key
6564 exponentiations with the GMP library. The conversions to and from
6565 GMP's mpz_t format aren't optimised nor are any montgomery forms
6566 cached, and on x86 it appears OpenSSL's own performance has caught up.
6567 However there are likely to be other architectures where GMP could
6568 provide a boost. This ENGINE is not built in by default, but it can be
6569 specified at Configure time and should be accompanied by the necessary
6570 linker additions, eg;
6571 ./config -DOPENSSL_USE_GMP -lgmp
6574 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6575 testing availability of engines with "-t" - the old behaviour is
6576 produced by increasing the feature's verbosity with "-tt".
6579 *) ECDSA routines: under certain error conditions uninitialized BN objects
6580 could be freed. Solution: make sure initialization is performed early
6581 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6585 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6586 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6587 software implementations. For DSA and DH, parameter generation can
6588 also be overridden by providing the appropriate method callbacks.
6591 *) Change the "progress" mechanism used in key-generation and
6592 primality testing to functions that take a new BN_GENCB pointer in
6593 place of callback/argument pairs. The new API functions have "_ex"
6594 postfixes and the older functions are reimplemented as wrappers for
6595 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6596 declarations of the old functions to help (graceful) attempts to
6597 migrate to the new functions. Also, the new key-generation API
6598 functions operate on a caller-supplied key-structure and return
6599 success/failure rather than returning a key or NULL - this is to
6600 help make "keygen" another member function of RSA_METHOD etc.
6602 Example for using the new callback interface:
6604 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6608 BN_GENCB_set(&my_cb, my_callback, my_arg);
6610 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6611 /* For the meaning of a, b in calls to my_callback(), see the
6612 * documentation of the function that calls the callback.
6613 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6614 * my_callback should return 1 if it wants BN_is_prime_ex()
6615 * to continue, or 0 to stop.
6620 *) Change the ZLIB compression method to be stateful, and make it
6621 available to TLS with the number defined in
6622 draft-ietf-tls-compression-04.txt.
6625 *) Add the ASN.1 structures and functions for CertificatePair, which
6626 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6628 CertificatePair ::= SEQUENCE {
6629 forward [0] Certificate OPTIONAL,
6630 reverse [1] Certificate OPTIONAL,
6631 -- at least one of the pair shall be present -- }
6633 Also implement the PEM functions to read and write certificate
6634 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6636 This needed to be defined, mostly for the sake of the LDAP
6637 attribute crossCertificatePair, but may prove useful elsewhere as
6641 *) Make it possible to inhibit symlinking of shared libraries in
6642 Makefile.shared, for Cygwin's sake.
6645 *) Extend the BIGNUM API by creating a function
6646 void BN_set_negative(BIGNUM *a, int neg);
6647 and a macro that behave like
6648 int BN_is_negative(const BIGNUM *a);
6650 to avoid the need to access 'a->neg' directly in applications.
6653 *) Implement fast modular reduction for pseudo-Mersenne primes
6654 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6655 EC_GROUP_new_curve_GFp() will now automatically use this
6657 [Nils Larsch <nla@trustcenter.de>]
6659 *) Add new lock type (CRYPTO_LOCK_BN).
6662 *) Change the ENGINE framework to automatically load engines
6663 dynamically from specific directories unless they could be
6664 found to already be built in or loaded. Move all the
6665 current engines except for the cryptodev one to a new
6667 The engines in engines/ are built as shared libraries if
6668 the "shared" options was given to ./Configure or ./config.
6669 Otherwise, they are inserted in libcrypto.a.
6670 /usr/local/ssl/engines is the default directory for dynamic
6671 engines, but that can be overridden at configure time through
6672 the usual use of --prefix and/or --openssldir, and at run
6673 time with the environment variable OPENSSL_ENGINES.
6674 [Geoff Thorpe and Richard Levitte]
6676 *) Add Makefile.shared, a helper makefile to build shared
6677 libraries. Adapt Makefile.org.
6680 *) Add version info to Win32 DLLs.
6681 [Peter 'Luna' Runestig" <peter@runestig.com>]
6683 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6684 can be added using this API to created arbitrary PKCS#12
6685 files while avoiding the low level API.
6687 New options to PKCS12_create(), key or cert can be NULL and
6688 will then be omitted from the output file. The encryption
6689 algorithm NIDs can be set to -1 for no encryption, the mac
6690 iteration count can be set to 0 to omit the mac.
6692 Enhance pkcs12 utility by making the -nokeys and -nocerts
6693 options work when creating a PKCS#12 file. New option -nomac
6694 to omit the mac, NONE can be set for an encryption algorithm.
6695 New code is modified to use the enhanced PKCS12_create()
6696 instead of the low level API.
6699 *) Extend ASN1 encoder to support indefinite length constructed
6700 encoding. This can output sequences tags and octet strings in
6701 this form. Modify pk7_asn1.c to support indefinite length
6702 encoding. This is experimental and needs additional code to
6703 be useful, such as an ASN1 bio and some enhanced streaming
6706 Extend template encode functionality so that tagging is passed
6707 down to the template encoder.
6710 *) Let 'openssl req' fail if an argument to '-newkey' is not
6711 recognized instead of using RSA as a default.
6714 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6715 As these are not official, they are not included in "ALL";
6716 the "ECCdraft" ciphersuite group alias can be used to select them.
6717 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6719 *) Add ECDH engine support.
6720 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6722 *) Add ECDH in new directory crypto/ecdh/.
6723 [Douglas Stebila (Sun Microsystems Laboratories)]
6725 *) Let BN_rand_range() abort with an error after 100 iterations
6726 without success (which indicates a broken PRNG).
6729 *) Change BN_mod_sqrt() so that it verifies that the input value
6730 is really the square of the return value. (Previously,
6731 BN_mod_sqrt would show GIGO behaviour.)
6734 *) Add named elliptic curves over binary fields from X9.62, SECG,
6735 and WAP/WTLS; add OIDs that were still missing.
6737 [Sheueling Chang Shantz and Douglas Stebila
6738 (Sun Microsystems Laboratories)]
6740 *) Extend the EC library for elliptic curves over binary fields
6741 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6744 EC_GF2m_simple_method
6748 EC_GROUP_new_curve_GF2m
6749 EC_GROUP_set_curve_GF2m
6750 EC_GROUP_get_curve_GF2m
6751 EC_POINT_set_affine_coordinates_GF2m
6752 EC_POINT_get_affine_coordinates_GF2m
6753 EC_POINT_set_compressed_coordinates_GF2m
6755 Point compression for binary fields is disabled by default for
6756 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6759 As binary polynomials are represented as BIGNUMs, various members
6760 of the EC_GROUP and EC_POINT data structures can be shared
6761 between the implementations for prime fields and binary fields;
6762 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6763 are essentially identical to their ..._GFp counterparts.
6764 (For simplicity, the '..._GFp' prefix has been dropped from
6765 various internal method names.)
6767 An internal 'field_div' method (similar to 'field_mul' and
6768 'field_sqr') has been added; this is used only for binary fields.
6770 [Sheueling Chang Shantz and Douglas Stebila
6771 (Sun Microsystems Laboratories)]
6773 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6774 through methods ('mul', 'precompute_mult').
6776 The generic implementations (now internally called 'ec_wNAF_mul'
6777 and 'ec_wNAF_precomputed_mult') remain the default if these
6778 methods are undefined.
6780 [Sheueling Chang Shantz and Douglas Stebila
6781 (Sun Microsystems Laboratories)]
6783 *) New function EC_GROUP_get_degree, which is defined through
6784 EC_METHOD. For curves over prime fields, this returns the bit
6785 length of the modulus.
6787 [Sheueling Chang Shantz and Douglas Stebila
6788 (Sun Microsystems Laboratories)]
6790 *) New functions EC_GROUP_dup, EC_POINT_dup.
6791 (These simply call ..._new and ..._copy).
6793 [Sheueling Chang Shantz and Douglas Stebila
6794 (Sun Microsystems Laboratories)]
6796 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6797 Polynomials are represented as BIGNUMs (where the sign bit is not
6798 used) in the following functions [macros]:
6801 BN_GF2m_sub [= BN_GF2m_add]
6802 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6803 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6804 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6806 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6807 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6808 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6809 BN_GF2m_cmp [= BN_ucmp]
6811 (Note that only the 'mod' functions are actually for fields GF(2^m).
6812 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6814 For some functions, an the irreducible polynomial defining a
6815 field can be given as an 'unsigned int[]' with strictly
6816 decreasing elements giving the indices of those bits that are set;
6817 i.e., p[] represents the polynomial
6818 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6820 p[0] > p[1] > ... > p[k] = 0.
6821 This applies to the following functions:
6826 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6827 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6829 BN_GF2m_mod_sqrt_arr
6830 BN_GF2m_mod_solve_quad_arr
6834 Conversion can be performed by the following functions:
6839 bntest.c has additional tests for binary polynomial arithmetic.
6841 Two implementations for BN_GF2m_mod_div() are available.
6842 The default algorithm simply uses BN_GF2m_mod_inv() and
6843 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6844 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6845 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6847 [Sheueling Chang Shantz and Douglas Stebila
6848 (Sun Microsystems Laboratories)]
6850 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6851 functionality is disabled at compile-time.
6852 [Douglas Stebila <douglas.stebila@sun.com>]
6854 *) Change default behaviour of 'openssl asn1parse' so that more
6855 information is visible when viewing, e.g., a certificate:
6857 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6858 mode the content of non-printable OCTET STRINGs is output in a
6859 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6860 avoid the appearance of a printable string.
6861 [Nils Larsch <nla@trustcenter.de>]
6863 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6865 EC_GROUP_set_asn1_flag()
6866 EC_GROUP_get_asn1_flag()
6867 EC_GROUP_set_point_conversion_form()
6868 EC_GROUP_get_point_conversion_form()
6869 These control ASN1 encoding details:
6870 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6871 has been set to OPENSSL_EC_NAMED_CURVE.
6872 - Points are encoded in uncompressed form by default; options for
6873 asn1_for are as for point2oct, namely
6874 POINT_CONVERSION_COMPRESSED
6875 POINT_CONVERSION_UNCOMPRESSED
6876 POINT_CONVERSION_HYBRID
6878 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6881 EC_GROUP_get0_seed()
6882 EC_GROUP_get_seed_len()
6883 This is used only for ASN1 purposes (so far).
6884 [Nils Larsch <nla@trustcenter.de>]
6886 *) Add 'field_type' member to EC_METHOD, which holds the NID
6887 of the appropriate field type OID. The new function
6888 EC_METHOD_get_field_type() returns this value.
6889 [Nils Larsch <nla@trustcenter.de>]
6894 EC_POINT_point2hex()
6895 EC_POINT_hex2point()
6896 providing useful interfaces to EC_POINT_point2oct() and
6897 EC_POINT_oct2point().
6898 [Nils Larsch <nla@trustcenter.de>]
6900 *) Change internals of the EC library so that the functions
6901 EC_GROUP_set_generator()
6902 EC_GROUP_get_generator()
6903 EC_GROUP_get_order()
6904 EC_GROUP_get_cofactor()
6905 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6906 to methods, which would lead to unnecessary code duplication when
6907 adding different types of curves.
6908 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6910 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6911 arithmetic, and such that modified wNAFs are generated
6912 (which avoid length expansion in many cases).
6915 *) Add a function EC_GROUP_check_discriminant() (defined via
6916 EC_METHOD) that verifies that the curve discriminant is non-zero.
6918 Add a function EC_GROUP_check() that makes some sanity tests
6919 on a EC_GROUP, its generator and order. This includes
6920 EC_GROUP_check_discriminant().
6921 [Nils Larsch <nla@trustcenter.de>]
6923 *) Add ECDSA in new directory crypto/ecdsa/.
6925 Add applications 'openssl ecparam' and 'openssl ecdsa'
6926 (these are based on 'openssl dsaparam' and 'openssl dsa').
6928 ECDSA support is also included in various other files across the
6929 library. Most notably,
6930 - 'openssl req' now has a '-newkey ecdsa:file' option;
6931 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6932 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6933 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6934 them suitable for ECDSA where domain parameters must be
6935 extracted before the specific public key;
6936 - ECDSA engine support has been added.
6937 [Nils Larsch <nla@trustcenter.de>]
6939 *) Include some named elliptic curves, and add OIDs from X9.62,
6940 SECG, and WAP/WTLS. Each curve can be obtained from the new
6942 EC_GROUP_new_by_curve_name(),
6943 and the list of available named curves can be obtained with
6944 EC_get_builtin_curves().
6945 Also add a 'curve_name' member to EC_GROUP objects, which can be
6947 EC_GROUP_set_curve_name()
6948 EC_GROUP_get_curve_name()
6949 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6951 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6952 was actually never needed) and in BN_mul(). The removal in BN_mul()
6953 required a small change in bn_mul_part_recursive() and the addition
6954 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6955 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6956 bn_sub_words() and bn_add_words() except they take arrays with
6960 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6962 *) Cleanse PEM buffers before freeing them since they may contain
6964 [Benjamin Bennett <ben@psc.edu>]
6966 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6967 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6968 authentication-only ciphersuites.
6971 *) Since AES128 and AES256 share a single mask bit in the logic of
6972 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6973 kludge to work properly if AES128 is available and AES256 isn't.
6976 *) Expand security boundary to match 1.1.1 module.
6979 *) Remove redundant features: hash file source, editing of test vectors
6980 modify fipsld to use external fips_premain.c signature.
6983 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6984 run algorithm test programs.
6987 *) Make algorithm test programs more tolerant of whitespace.
6990 *) Have SSL/TLS server implementation tolerate "mismatched" record
6991 protocol version while receiving ClientHello even if the
6992 ClientHello is fragmented. (The server can't insist on the
6993 particular protocol version it has chosen before the ServerHello
6994 message has informed the client about his choice.)
6997 *) Load error codes if they are not already present instead of using a
6998 static variable. This allows them to be cleanly unloaded and reloaded.
7001 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
7003 *) Introduce limits to prevent malicious keys being able to
7004 cause a denial of service. (CVE-2006-2940)
7005 [Steve Henson, Bodo Moeller]
7007 *) Fix ASN.1 parsing of certain invalid structures that can result
7008 in a denial of service. (CVE-2006-2937) [Steve Henson]
7010 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
7011 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
7013 *) Fix SSL client code which could crash if connecting to a
7014 malicious SSLv2 server. (CVE-2006-4343)
7015 [Tavis Ormandy and Will Drewry, Google Security Team]
7017 *) Change ciphersuite string processing so that an explicit
7018 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
7019 will no longer include "AES128-SHA"), and any other similar
7020 ciphersuite (same bitmap) from *other* protocol versions (so that
7021 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
7022 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
7023 changes from 0.9.8b and 0.9.8d.
7026 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
7028 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
7029 (CVE-2006-4339) [Ben Laurie and Google Security Team]
7031 *) Change the Unix randomness entropy gathering to use poll() when
7032 possible instead of select(), since the latter has some
7033 undesirable limitations.
7034 [Darryl Miles via Richard Levitte and Bodo Moeller]
7036 *) Disable rogue ciphersuites:
7038 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
7039 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
7040 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
7042 The latter two were purportedly from
7043 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
7046 Also deactivate the remaining ciphersuites from
7047 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
7048 unofficial, and the ID has long expired.
7051 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
7052 dual-core machines) and other potential thread-safety issues.
7055 Changes between 0.9.7i and 0.9.7j [04 May 2006]
7057 *) Adapt fipsld and the build system to link against the validated FIPS
7058 module in FIPS mode.
7061 *) Fixes for VC++ 2005 build under Windows.
7064 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
7065 from a Windows bash shell such as MSYS. It is autodetected from the
7066 "config" script when run from a VC++ environment. Modify standard VC++
7067 build to use fipscanister.o from the GNU make build.
7070 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
7072 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
7073 The value now differs depending on if you build for FIPS or not.
7074 BEWARE! A program linked with a shared FIPSed libcrypto can't be
7075 safely run with a non-FIPSed libcrypto, as it may crash because of
7076 the difference induced by this change.
7079 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
7081 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
7082 (part of SSL_OP_ALL). This option used to disable the
7083 countermeasure against man-in-the-middle protocol-version
7084 rollback in the SSL 2.0 server implementation, which is a bad
7085 idea. (CVE-2005-2969)
7087 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
7088 for Information Security, National Institute of Advanced Industrial
7089 Science and Technology [AIST], Japan)]
7091 *) Minimal support for X9.31 signatures and PSS padding modes. This is
7092 mainly for FIPS compliance and not fully integrated at this stage.
7095 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
7096 the exponentiation using a fixed-length exponent. (Otherwise,
7097 the information leaked through timing could expose the secret key
7098 after many signatures; cf. Bleichenbacher's attack on DSA with
7102 *) Make a new fixed-window mod_exp implementation the default for
7103 RSA, DSA, and DH private-key operations so that the sequence of
7104 squares and multiplies and the memory access pattern are
7105 independent of the particular secret key. This will mitigate
7106 cache-timing and potential related attacks.
7108 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
7109 and this is automatically used by BN_mod_exp_mont() if the new flag
7110 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
7111 will use this BN flag for private exponents unless the flag
7112 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
7113 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
7115 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
7117 *) Change the client implementation for SSLv23_method() and
7118 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
7119 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
7120 (Previously, the SSL 2.0 backwards compatible Client Hello
7121 message format would be used even with SSL_OP_NO_SSLv2.)
7124 *) Add support for smime-type MIME parameter in S/MIME messages which some
7128 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
7129 a threadsafe manner. Modify rsa code to use new function and add calls
7130 to dsa and dh code (which had race conditions before).
7133 *) Include the fixed error library code in the C error file definitions
7134 instead of fixing them up at runtime. This keeps the error code
7135 structures constant.
7138 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
7140 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
7143 *) Fixes for newer kerberos headers. NB: the casts are needed because
7144 the 'length' field is signed on one version and unsigned on another
7145 with no (?) obvious way to tell the difference, without these VC++
7146 complains. Also the "definition" of FAR (blank) is no longer included
7147 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
7148 some needed definitions.
7151 *) Undo Cygwin change.
7154 *) Added support for proxy certificates according to RFC 3820.
7155 Because they may be a security thread to unaware applications,
7156 they must be explicitly allowed in run-time. See
7157 docs/HOWTO/proxy_certificates.txt for further information.
7160 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
7162 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
7163 server and client random values. Previously
7164 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
7165 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
7167 This change has negligible security impact because:
7169 1. Server and client random values still have 24 bytes of pseudo random
7172 2. Server and client random values are sent in the clear in the initial
7175 3. The master secret is derived using the premaster secret (48 bytes in
7176 size for static RSA ciphersuites) as well as client server and random
7179 The OpenSSL team would like to thank the UK NISCC for bringing this issue
7182 [Stephen Henson, reported by UK NISCC]
7184 *) Use Windows randomness collection on Cygwin.
7187 *) Fix hang in EGD/PRNGD query when communication socket is closed
7188 prematurely by EGD/PRNGD.
7189 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
7191 *) Prompt for pass phrases when appropriate for PKCS12 input format.
7194 *) Back-port of selected performance improvements from development
7195 branch, as well as improved support for PowerPC platforms.
7198 *) Add lots of checks for memory allocation failure, error codes to indicate
7199 failure and freeing up memory if a failure occurs.
7200 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
7202 *) Add new -passin argument to dgst.
7205 *) Perform some character comparisons of different types in X509_NAME_cmp:
7206 this is needed for some certificates that re-encode DNs into UTF8Strings
7207 (in violation of RFC3280) and can't or won't issue name rollover
7211 *) Make an explicit check during certificate validation to see that
7212 the CA setting in each certificate on the chain is correct. As a
7213 side effect always do the following basic checks on extensions,
7214 not just when there's an associated purpose to the check:
7216 - if there is an unhandled critical extension (unless the user
7217 has chosen to ignore this fault)
7218 - if the path length has been exceeded (if one is set at all)
7219 - that certain extensions fit the associated purpose (if one has
7223 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
7225 *) Avoid a race condition when CRLs are checked in a multi threaded
7226 environment. This would happen due to the reordering of the revoked
7227 entries during signature checking and serial number lookup. Now the
7228 encoding is cached and the serial number sort performed under a lock.
7229 Add new STACK function sk_is_sorted().
7232 *) Add Delta CRL to the extension code.
7235 *) Various fixes to s3_pkt.c so alerts are sent properly.
7236 [David Holmes <d.holmes@f5.com>]
7238 *) Reduce the chances of duplicate issuer name and serial numbers (in
7239 violation of RFC3280) using the OpenSSL certificate creation utilities.
7240 This is done by creating a random 64 bit value for the initial serial
7241 number when a serial number file is created or when a self signed
7242 certificate is created using 'openssl req -x509'. The initial serial
7243 number file is created using 'openssl x509 -next_serial' in CA.pl
7244 rather than being initialized to 1.
7247 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
7249 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7250 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7251 [Joe Orton, Steve Henson]
7253 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
7255 [Joe Orton, Steve Henson]
7257 *) Make it possible to have multiple active certificates with the same
7258 subject in the CA index file. This is done only if the keyword
7259 'unique_subject' is set to 'no' in the main CA section (default
7260 if 'CA_default') of the configuration file. The value is saved
7261 with the database itself in a separate index attribute file,
7262 named like the index file with '.attr' appended to the name.
7265 *) X509 verify fixes. Disable broken certificate workarounds when
7266 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
7267 keyUsage extension present. Don't accept CRLs with unhandled critical
7268 extensions: since verify currently doesn't process CRL extensions this
7269 rejects a CRL with *any* critical extensions. Add new verify error codes
7273 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
7274 A clarification of RFC2560 will require the use of OCTET STRINGs and
7275 some implementations cannot handle the current raw format. Since OpenSSL
7276 copies and compares OCSP nonces as opaque blobs without any attempt at
7277 parsing them this should not create any compatibility issues.
7280 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
7281 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
7282 this HMAC (and other) operations are several times slower than OpenSSL
7286 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
7287 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7289 *) Use the correct content when signing type "other".
7292 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
7294 *) Fix various bugs revealed by running the NISCC test suite:
7296 Stop out of bounds reads in the ASN1 code when presented with
7297 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7299 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
7301 If verify callback ignores invalid public key errors don't try to check
7302 certificate signature with the NULL public key.
7306 *) New -ignore_err option in ocsp application to stop the server
7307 exiting on the first error in a request.
7310 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7311 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7315 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7316 extra data after the compression methods not only for TLS 1.0
7317 but also for SSL 3.0 (as required by the specification).
7318 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7320 *) Change X509_certificate_type() to mark the key as exported/exportable
7321 when it's 512 *bits* long, not 512 bytes.
7324 *) Change AES_cbc_encrypt() so it outputs exact multiple of
7325 blocks during encryption.
7328 *) Various fixes to base64 BIO and non blocking I/O. On write
7329 flushes were not handled properly if the BIO retried. On read
7330 data was not being buffered properly and had various logic bugs.
7331 This also affects blocking I/O when the data being decoded is a
7335 *) Various S/MIME bugfixes and compatibility changes:
7336 output correct application/pkcs7 MIME type if
7337 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
7338 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
7339 of files as .eml work). Correctly handle very long lines in MIME
7343 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
7345 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7346 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7347 a protocol version number mismatch like a decryption error
7348 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7351 *) Turn on RSA blinding by default in the default implementation
7352 to avoid a timing attack. Applications that don't want it can call
7353 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7354 They would be ill-advised to do so in most cases.
7355 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7357 *) Change RSA blinding code so that it works when the PRNG is not
7358 seeded (in this case, the secret RSA exponent is abused as
7359 an unpredictable seed -- if it is not unpredictable, there
7360 is no point in blinding anyway). Make RSA blinding thread-safe
7361 by remembering the creator's thread ID in rsa->blinding and
7362 having all other threads use local one-time blinding factors
7363 (this requires more computation than sharing rsa->blinding, but
7364 avoids excessive locking; and if an RSA object is not shared
7365 between threads, blinding will still be very fast).
7368 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
7369 ENGINE as defaults for all supported algorithms irrespective of
7370 the 'flags' parameter. 'flags' is now honoured, so applications
7371 should make sure they are passing it correctly.
7374 *) Target "mingw" now allows native Windows code to be generated in
7375 the Cygwin environment as well as with the MinGW compiler.
7378 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7380 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7381 via timing by performing a MAC computation even if incorrect
7382 block cipher padding has been found. This is a countermeasure
7383 against active attacks where the attacker has to distinguish
7384 between bad padding and a MAC verification error. (CVE-2003-0078)
7386 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7387 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7388 Martin Vuagnoux (EPFL, Ilion)]
7390 *) Make the no-err option work as intended. The intention with no-err
7391 is not to have the whole error stack handling routines removed from
7392 libcrypto, it's only intended to remove all the function name and
7393 reason texts, thereby removing some of the footprint that may not
7394 be interesting if those errors aren't displayed anyway.
7396 NOTE: it's still possible for any application or module to have its
7397 own set of error texts inserted. The routines are there, just not
7398 used by default when no-err is given.
7401 *) Add support for FreeBSD on IA64.
7402 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7404 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7405 Kerberos function mit_des_cbc_cksum(). Before this change,
7406 the value returned by DES_cbc_cksum() was like the one from
7407 mit_des_cbc_cksum(), except the bytes were swapped.
7408 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7410 *) Allow an application to disable the automatic SSL chain building.
7411 Before this a rather primitive chain build was always performed in
7412 ssl3_output_cert_chain(): an application had no way to send the
7413 correct chain if the automatic operation produced an incorrect result.
7415 Now the chain builder is disabled if either:
7417 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7419 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7421 The reasoning behind this is that an application would not want the
7422 auto chain building to take place if extra chain certificates are
7423 present and it might also want a means of sending no additional
7424 certificates (for example the chain has two certificates and the
7428 *) Add the possibility to build without the ENGINE framework.
7429 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7431 *) Under Win32 gmtime() can return NULL: check return value in
7432 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7435 *) DSA routines: under certain error conditions uninitialized BN objects
7436 could be freed. Solution: make sure initialization is performed early
7437 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7438 Nils Larsch <nla@trustcenter.de> via PR#459)
7441 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7442 checked on reconnect on the client side, therefore session resumption
7443 could still fail with a "ssl session id is different" error. This
7444 behaviour is masked when SSL_OP_ALL is used due to
7445 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7446 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7447 followup to PR #377.
7450 *) IA-32 assembler support enhancements: unified ELF targets, support
7451 for SCO/Caldera platforms, fix for Cygwin shared build.
7454 *) Add support for FreeBSD on sparc64. As a consequence, support for
7455 FreeBSD on non-x86 processors is separate from x86 processors on
7456 the config script, much like the NetBSD support.
7457 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7459 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7461 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7464 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7465 code (06) was taken as the first octet of the session ID and the last
7466 octet was ignored consequently. As a result SSLv2 client side session
7467 caching could not have worked due to the session ID mismatch between
7469 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7473 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7474 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7478 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7479 seems that in spite of existing for more than a year, many application
7480 author have done nothing to provide the necessary callbacks, which
7481 means that this particular engine will not work properly anywhere.
7482 This is a very unfortunate situation which forces us, in the name
7483 of usability, to give the hw_ncipher.c a static lock, which is part
7485 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7486 appear in 0.9.8 or later. We EXPECT application authors to have
7487 dealt properly with this when 0.9.8 is released (unless we actually
7488 make such changes in the libcrypto locking code that changes will
7489 have to be made anyway).
7492 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7493 octets have been read, EOF or an error occurs. Without this change
7494 some truncated ASN1 structures will not produce an error.
7497 *) Disable Heimdal support, since it hasn't been fully implemented.
7498 Still give the possibility to force the use of Heimdal, but with
7499 warnings and a request that patches get sent to openssl-dev.
7502 *) Add the VC-CE target, introduce the WINCE sysname, and add
7503 INSTALL.WCE and appropriate conditionals to make it build.
7504 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7506 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7507 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7508 edit numbers of the version.
7509 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7511 *) Introduce safe string copy and catenation functions
7512 (BUF_strlcpy() and BUF_strlcat()).
7513 [Ben Laurie (CHATS) and Richard Levitte]
7515 *) Avoid using fixed-size buffers for one-line DNs.
7516 [Ben Laurie (CHATS)]
7518 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7519 resizing buffers containing secrets, and use where appropriate.
7520 [Ben Laurie (CHATS)]
7522 *) Avoid using fixed size buffers for configuration file location.
7523 [Ben Laurie (CHATS)]
7525 *) Avoid filename truncation for various CA files.
7526 [Ben Laurie (CHATS)]
7528 *) Use sizeof in preference to magic numbers.
7529 [Ben Laurie (CHATS)]
7531 *) Avoid filename truncation in cert requests.
7532 [Ben Laurie (CHATS)]
7534 *) Add assertions to check for (supposedly impossible) buffer
7536 [Ben Laurie (CHATS)]
7538 *) Don't cache truncated DNS entries in the local cache (this could
7539 potentially lead to a spoofing attack).
7540 [Ben Laurie (CHATS)]
7542 *) Fix various buffers to be large enough for hex/decimal
7543 representations in a platform independent manner.
7544 [Ben Laurie (CHATS)]
7546 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7547 resizing buffers containing secrets, and use where appropriate.
7548 [Ben Laurie (CHATS)]
7550 *) Add BIO_indent() to avoid much slightly worrying code to do
7552 [Ben Laurie (CHATS)]
7554 *) Convert sprintf()/BIO_puts() to BIO_printf().
7555 [Ben Laurie (CHATS)]
7557 *) buffer_gets() could terminate with the buffer only half
7559 [Ben Laurie (CHATS)]
7561 *) Add assertions to prevent user-supplied crypto functions from
7562 overflowing internal buffers by having large block sizes, etc.
7563 [Ben Laurie (CHATS)]
7565 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7567 [Ben Laurie (CHATS)]
7569 *) Eliminate unused copy of key in RC4.
7570 [Ben Laurie (CHATS)]
7572 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7573 [Ben Laurie (CHATS)]
7575 *) Fix off-by-one error in EGD path.
7576 [Ben Laurie (CHATS)]
7578 *) If RANDFILE path is too long, ignore instead of truncating.
7579 [Ben Laurie (CHATS)]
7581 *) Eliminate unused and incorrectly sized X.509 structure
7583 [Ben Laurie (CHATS)]
7585 *) Eliminate unused and dangerous function knumber().
7586 [Ben Laurie (CHATS)]
7588 *) Eliminate unused and dangerous structure, KSSL_ERR.
7589 [Ben Laurie (CHATS)]
7591 *) Protect against overlong session ID context length in an encoded
7592 session object. Since these are local, this does not appear to be
7594 [Ben Laurie (CHATS)]
7596 *) Change from security patch (see 0.9.6e below) that did not affect
7597 the 0.9.6 release series:
7599 Remote buffer overflow in SSL3 protocol - an attacker could
7600 supply an oversized master key in Kerberos-enabled versions.
7602 [Ben Laurie (CHATS)]
7604 *) Change the SSL kerb5 codes to match RFC 2712.
7607 *) Make -nameopt work fully for req and add -reqopt switch.
7608 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7610 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7611 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7613 *) Make sure tests can be performed even if the corresponding algorithms
7614 have been removed entirely. This was also the last step to make
7615 OpenSSL compilable with DJGPP under all reasonable conditions.
7616 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7618 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7619 to allow version independent disabling of normally unselected ciphers,
7620 which may be activated as a side-effect of selecting a single cipher.
7622 (E.g., cipher list string "RSA" enables ciphersuites that are left
7623 out of "ALL" because they do not provide symmetric encryption.
7624 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7625 [Lutz Jaenicke, Bodo Moeller]
7627 *) Add appropriate support for separate platform-dependent build
7628 directories. The recommended way to make a platform-dependent
7629 build directory is the following (tested on Linux), maybe with
7632 # Place yourself outside of the OpenSSL source tree. In
7633 # this example, the environment variable OPENSSL_SOURCE
7634 # is assumed to contain the absolute OpenSSL source directory.
7635 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7636 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7637 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7638 mkdir -p `dirname $F`
7639 ln -s $OPENSSL_SOURCE/$F $F
7642 To be absolutely sure not to disturb the source tree, a "make clean"
7643 is a good thing. If it isn't successful, don't worry about it,
7644 it probably means the source directory is very clean.
7647 *) Make sure any ENGINE control commands make local copies of string
7648 pointers passed to them whenever necessary. Otherwise it is possible
7649 the caller may have overwritten (or deallocated) the original string
7650 data when a later ENGINE operation tries to use the stored values.
7651 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7653 *) Improve diagnostics in file reading and command-line digests.
7654 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7656 *) Add AES modes CFB and OFB to the object database. Correct an
7657 error in AES-CFB decryption.
7660 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7661 allows existing EVP_CIPHER_CTX structures to be reused after
7662 calling EVP_*Final(). This behaviour is used by encryption
7663 BIOs and some applications. This has the side effect that
7664 applications must explicitly clean up cipher contexts with
7665 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7668 *) Check the values of dna and dnb in bn_mul_recursive before calling
7669 bn_mul_comba (a non zero value means the a or b arrays do not contain
7670 n2 elements) and fallback to bn_mul_normal if either is not zero.
7673 *) Fix escaping of non-ASCII characters when using the -subj option
7674 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7677 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7678 form for "surname", serialNumber has no short form.
7679 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7680 therefore remove "mail" short name for "internet 7".
7681 The OID for unique identifiers in X509 certificates is
7682 x500UniqueIdentifier, not uniqueIdentifier.
7683 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7686 *) Add an "init" command to the ENGINE config module and auto initialize
7687 ENGINEs. Without any "init" command the ENGINE will be initialized
7688 after all ctrl commands have been executed on it. If init=1 the
7689 ENGINE is initialized at that point (ctrls before that point are run
7690 on the uninitialized ENGINE and after on the initialized one). If
7691 init=0 then the ENGINE will not be initialized at all.
7694 *) Fix the 'app_verify_callback' interface so that the user-defined
7695 argument is actually passed to the callback: In the
7696 SSL_CTX_set_cert_verify_callback() prototype, the callback
7697 declaration has been changed from
7700 int (*cb)(X509_STORE_CTX *,void *);
7701 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7702 i=s->ctx->app_verify_callback(&ctx)
7703 has been changed into
7704 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7706 To update applications using SSL_CTX_set_cert_verify_callback(),
7707 a dummy argument can be added to their callback functions.
7708 [D. K. Smetters <smetters@parc.xerox.com>]
7710 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7711 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7713 *) Add and OPENSSL_LOAD_CONF define which will cause
7714 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7715 This allows older applications to transparently support certain
7716 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7717 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7718 load the config file and OPENSSL_add_all_algorithms_conf() which will
7719 always load it have also been added.
7722 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7723 Adjust NIDs and EVP layer.
7724 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7726 *) Config modules support in openssl utility.
7728 Most commands now load modules from the config file,
7729 though in a few (such as version) this isn't done
7730 because it couldn't be used for anything.
7732 In the case of ca and req the config file used is
7733 the same as the utility itself: that is the -config
7734 command line option can be used to specify an
7738 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7739 use "openssl_conf" if filename is NULL use default openssl config file.
7742 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7743 config section name. Add a new flag to tolerate a missing config file
7744 and move code to CONF_modules_load_file().
7747 *) Support for crypto accelerator cards from Accelerated Encryption
7748 Processing, www.aep.ie. (Use engine 'aep')
7749 The support was copied from 0.9.6c [engine] and adapted/corrected
7750 to work with the new engine framework.
7751 [AEP Inc. and Richard Levitte]
7753 *) Support for SureWare crypto accelerator cards from Baltimore
7754 Technologies. (Use engine 'sureware')
7755 The support was copied from 0.9.6c [engine] and adapted
7756 to work with the new engine framework.
7759 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7760 make the newer ENGINE framework commands for the CHIL engine work.
7761 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7763 *) Make it possible to produce shared libraries on ReliantUNIX.
7764 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7766 *) Add the configuration target debug-linux-ppro.
7767 Make 'openssl rsa' use the general key loading routines
7768 implemented in apps.c, and make those routines able to
7769 handle the key format FORMAT_NETSCAPE and the variant
7771 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7773 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7774 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7776 *) Add -keyform to rsautl, and document -engine.
7777 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7779 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7780 BIO_R_NO_SUCH_FILE error code rather than the generic
7781 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7784 *) Add new functions
7786 ERR_peek_last_error_line
7787 ERR_peek_last_error_line_data.
7788 These are similar to
7791 ERR_peek_error_line_data,
7792 but report on the latest error recorded rather than the first one
7793 still in the error queue.
7794 [Ben Laurie, Bodo Moeller]
7796 *) default_algorithms option in ENGINE config module. This allows things
7798 default_algorithms = ALL
7799 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7802 *) Preliminary ENGINE config module.
7805 *) New experimental application configuration code.
7808 *) Change the AES code to follow the same name structure as all other
7809 symmetric ciphers, and behave the same way. Move everything to
7810 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7811 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7813 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7814 [Ben Laurie and Theo de Raadt]
7816 *) Add option to output public keys in req command.
7817 [Massimiliano Pala madwolf@openca.org]
7819 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7820 (up to about 10% better than before for P-192 and P-224).
7823 *) New functions/macros
7825 SSL_CTX_set_msg_callback(ctx, cb)
7826 SSL_CTX_set_msg_callback_arg(ctx, arg)
7827 SSL_set_msg_callback(ssl, cb)
7828 SSL_set_msg_callback_arg(ssl, arg)
7830 to request calling a callback function
7832 void cb(int write_p, int version, int content_type,
7833 const void *buf, size_t len, SSL *ssl, void *arg)
7835 whenever a protocol message has been completely received
7836 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7837 protocol version according to which the SSL library interprets
7838 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7839 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7840 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7841 specification (change_cipher_spec(20), alert(21), handshake(22)).
7842 'buf' and 'len' point to the actual message, 'ssl' to the
7843 SSL object, and 'arg' is the application-defined value set by
7844 SSL[_CTX]_set_msg_callback_arg().
7846 'openssl s_client' and 'openssl s_server' have new '-msg' options
7847 to enable a callback that displays all protocol messages.
7850 *) Change the shared library support so shared libraries are built as
7851 soon as the corresponding static library is finished, and thereby get
7852 openssl and the test programs linked against the shared library.
7853 This still only happens when the keyword "shard" has been given to
7854 the configuration scripts.
7856 NOTE: shared library support is still an experimental thing, and
7857 backward binary compatibility is still not guaranteed.
7858 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7860 *) Add support for Subject Information Access extension.
7861 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7863 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7864 additional bytes when new memory had to be allocated, not just
7865 when reusing an existing buffer.
7868 *) New command line and configuration option 'utf8' for the req command.
7869 This allows field values to be specified as UTF8 strings.
7872 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7873 runs for the former and machine-readable output for the latter.
7876 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7877 of the e-mail address in the DN (i.e., it will go into a certificate
7878 extension only). The new configuration file option 'email_in_dn = no'
7879 has the same effect.
7880 [Massimiliano Pala madwolf@openca.org]
7882 *) Change all functions with names starting with des_ to be starting
7883 with DES_ instead. Add wrappers that are compatible with libdes,
7884 but are named _ossl_old_des_*. Finally, add macros that map the
7885 des_* symbols to the corresponding _ossl_old_des_* if libdes
7886 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7887 desired, the des_* symbols will be mapped to DES_*, with one
7890 Since we provide two compatibility mappings, the user needs to
7891 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7892 compatibility is desired. The default (i.e., when that macro
7893 isn't defined) is OpenSSL 0.9.6c compatibility.
7895 There are also macros that enable and disable the support of old
7896 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7897 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7898 are defined, the default will apply: to support the old des routines.
7900 In either case, one must include openssl/des.h to get the correct
7901 definitions. Do not try to just include openssl/des_old.h, that
7904 NOTE: This is a major break of an old API into a new one. Software
7905 authors are encouraged to switch to the DES_ style functions. Some
7906 time in the future, des_old.h and the libdes compatibility functions
7907 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7908 default), and then completely removed.
7911 *) Test for certificates which contain unsupported critical extensions.
7912 If such a certificate is found during a verify operation it is
7913 rejected by default: this behaviour can be overridden by either
7914 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7915 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7916 X509_supported_extension() has also been added which returns 1 if a
7917 particular extension is supported.
7920 *) Modify the behaviour of EVP cipher functions in similar way to digests
7921 to retain compatibility with existing code.
7924 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7925 compatibility with existing code. In particular the 'ctx' parameter does
7926 not have to be to be initialized before the call to EVP_DigestInit() and
7927 it is tidied up after a call to EVP_DigestFinal(). New function
7928 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7929 EVP_MD_CTX_copy() changed to not require the destination to be
7930 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7931 requires the destination to be valid.
7933 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7934 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7937 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7938 so that complete 'Handshake' protocol structures are kept in memory
7939 instead of overwriting 'msg_type' and 'length' with 'body' data.
7942 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7943 [Massimo Santin via Richard Levitte]
7945 *) Major restructuring to the underlying ENGINE code. This includes
7946 reduction of linker bloat, separation of pure "ENGINE" manipulation
7947 (initialisation, etc) from functionality dealing with implementations
7948 of specific crypto interfaces. This change also introduces integrated
7949 support for symmetric ciphers and digest implementations - so ENGINEs
7950 can now accelerate these by providing EVP_CIPHER and EVP_MD
7951 implementations of their own. This is detailed in crypto/engine/README
7952 as it couldn't be adequately described here. However, there are a few
7953 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7954 were changed in the original introduction of ENGINE code have now
7955 reverted back - the hooking from this code to ENGINE is now a good
7956 deal more passive and at run-time, operations deal directly with
7957 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7958 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7959 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7960 they were not being used by the framework as there is no concept of a
7961 BIGNUM_METHOD and they could not be generalised to the new
7962 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7963 ENGINE_cpy() has been removed as it cannot be consistently defined in
7967 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7970 *) Change mkdef.pl to sort symbols that get the same entry number,
7971 and make sure the automatically generated functions ERR_load_*
7972 become part of libeay.num as well.
7975 *) New function SSL_renegotiate_pending(). This returns true once
7976 renegotiation has been requested (either SSL_renegotiate() call
7977 or HelloRequest/ClientHello received from the peer) and becomes
7978 false once a handshake has been completed.
7979 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7980 sends a HelloRequest, but does not ensure that a handshake takes
7981 place. SSL_renegotiate_pending() is useful for checking if the
7982 client has followed the request.)
7985 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7986 By default, clients may request session resumption even during
7987 renegotiation (if session ID contexts permit); with this option,
7988 session resumption is possible only in the first handshake.
7990 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7991 more bits available for options that should not be part of
7992 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7995 *) Add some demos for certificate and certificate request creation.
7998 *) Make maximum certificate chain size accepted from the peer application
7999 settable (SSL*_get/set_max_cert_list()), as proposed by
8000 "Douglas E. Engert" <deengert@anl.gov>.
8003 *) Add support for shared libraries for Unixware-7
8004 (Boyd Lynn Gerber <gerberb@zenez.com>).
8007 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
8008 be done prior to destruction. Use this to unload error strings from
8009 ENGINEs that load their own error strings. NB: This adds two new API
8010 functions to "get" and "set" this destroy handler in an ENGINE.
8013 *) Alter all existing ENGINE implementations (except "openssl" and
8014 "openbsd") to dynamically instantiate their own error strings. This
8015 makes them more flexible to be built both as statically-linked ENGINEs
8016 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
8017 Also, add stub code to each that makes building them as self-contained
8018 shared-libraries easier (see README.ENGINE).
8021 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
8022 implementations into applications that are completely implemented in
8023 self-contained shared-libraries. The "dynamic" ENGINE exposes control
8024 commands that can be used to configure what shared-library to load and
8025 to control aspects of the way it is handled. Also, made an update to
8026 the README.ENGINE file that brings its information up-to-date and
8027 provides some information and instructions on the "dynamic" ENGINE
8028 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
8031 *) Make it possible to unload ranges of ERR strings with a new
8032 "ERR_unload_strings" function.
8035 *) Add a copy() function to EVP_MD.
8038 *) Make EVP_MD routines take a context pointer instead of just the
8039 md_data void pointer.
8042 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
8043 that the digest can only process a single chunk of data
8044 (typically because it is provided by a piece of
8045 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
8046 is only going to provide a single chunk of data, and hence the
8047 framework needn't accumulate the data for oneshot drivers.
8050 *) As with "ERR", make it possible to replace the underlying "ex_data"
8051 functions. This change also alters the storage and management of global
8052 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
8053 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
8054 index counters. The API functions that use this state have been changed
8055 to take a "class_index" rather than pointers to the class's local STACK
8056 and counter, and there is now an API function to dynamically create new
8057 classes. This centralisation allows us to (a) plug a lot of the
8058 thread-safety problems that existed, and (b) makes it possible to clean
8059 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
8060 such data would previously have always leaked in application code and
8061 workarounds were in place to make the memory debugging turn a blind eye
8062 to it. Application code that doesn't use this new function will still
8063 leak as before, but their memory debugging output will announce it now
8064 rather than letting it slide.
8066 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
8067 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
8068 has a return value to indicate success or failure.
8071 *) Make it possible to replace the underlying "ERR" functions such that the
8072 global state (2 LHASH tables and 2 locks) is only used by the "default"
8073 implementation. This change also adds two functions to "get" and "set"
8074 the implementation prior to it being automatically set the first time
8075 any other ERR function takes place. Ie. an application can call "get",
8076 pass the return value to a module it has just loaded, and that module
8077 can call its own "set" function using that value. This means the
8078 module's "ERR" operations will use (and modify) the error state in the
8079 application and not in its own statically linked copy of OpenSSL code.
8082 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
8083 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
8084 the operation, and provides a more encapsulated way for external code
8085 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
8086 to use these functions rather than manually incrementing the counts.
8088 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
8091 *) Add EVP test program.
8094 *) Add symmetric cipher support to ENGINE. Expect the API to change!
8097 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
8098 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
8099 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
8100 These allow a CRL to be built without having to access X509_CRL fields
8101 directly. Modify 'ca' application to use new functions.
8104 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
8105 bug workarounds. Rollback attack detection is a security feature.
8106 The problem will only arise on OpenSSL servers when TLSv1 is not
8107 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
8108 Software authors not wanting to support TLSv1 will have special reasons
8109 for their choice and can explicitly enable this option.
8110 [Bodo Moeller, Lutz Jaenicke]
8112 *) Rationalise EVP so it can be extended: don't include a union of
8113 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
8114 (similar to those existing for EVP_CIPHER_CTX).
8119 EVP_MD_CTX_init(&md); /* new function call */
8120 EVP_DigestInit(&md, EVP_sha1());
8121 EVP_DigestUpdate(&md, in, len);
8122 EVP_DigestFinal(&md, out, NULL);
8123 EVP_MD_CTX_cleanup(&md); /* new function call */
8127 *) Make DES key schedule conform to the usual scheme, as well as
8128 correcting its structure. This means that calls to DES functions
8129 now have to pass a pointer to a des_key_schedule instead of a
8130 plain des_key_schedule (which was actually always a pointer
8133 des_key_schedule ks;
8135 des_set_key_checked(..., &ks);
8136 des_ncbc_encrypt(..., &ks, ...);
8138 (Note that a later change renames 'des_...' into 'DES_...'.)
8141 *) Initial reduction of linker bloat: the use of some functions, such as
8142 PEM causes large amounts of unused functions to be linked in due to
8143 poor organisation. For example pem_all.c contains every PEM function
8144 which has a knock on effect of linking in large amounts of (unused)
8145 ASN1 code. Grouping together similar functions and splitting unrelated
8146 functions prevents this.
8149 *) Cleanup of EVP macros.
8152 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
8153 correct _ecb suffix.
8156 *) Add initial OCSP responder support to ocsp application. The
8157 revocation information is handled using the text based index
8158 use by the ca application. The responder can either handle
8159 requests generated internally, supplied in files (for example
8160 via a CGI script) or using an internal minimal server.
8163 *) Add configuration choices to get zlib compression for TLS.
8166 *) Changes to Kerberos SSL for RFC 2712 compliance:
8167 1. Implemented real KerberosWrapper, instead of just using
8168 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
8169 2. Implemented optional authenticator field of KerberosWrapper.
8171 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
8172 and authenticator structs; see crypto/krb5/.
8174 Generalized Kerberos calls to support multiple Kerberos libraries.
8175 [Vern Staats <staatsvr@asc.hpc.mil>,
8176 Jeffrey Altman <jaltman@columbia.edu>
8177 via Richard Levitte]
8179 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
8180 already does with RSA. testdsa.h now has 'priv_key/pub_key'
8181 values for each of the key sizes rather than having just
8182 parameters (and 'speed' generating keys each time).
8185 *) Speed up EVP routines.
8188 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
8189 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
8190 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
8191 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
8193 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
8194 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
8195 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
8198 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
8200 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
8203 *) Added the OS2-EMX target.
8204 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
8206 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
8207 to support NCONF routines in extension code. New function CONF_set_nconf()
8208 to allow functions which take an NCONF to also handle the old LHASH
8209 structure: this means that the old CONF compatible routines can be
8210 retained (in particular wrt extensions) without having to duplicate the
8211 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
8214 *) Enhance the general user interface with mechanisms for inner control
8215 and with possibilities to have yes/no kind of prompts.
8218 *) Change all calls to low level digest routines in the library and
8219 applications to use EVP. Add missing calls to HMAC_cleanup() and
8220 don't assume HMAC_CTX can be copied using memcpy().
8221 [Verdon Walker <VWalker@novell.com>, Steve Henson]
8223 *) Add the possibility to control engines through control names but with
8224 arbitrary arguments instead of just a string.
8225 Change the key loaders to take a UI_METHOD instead of a callback
8226 function pointer. NOTE: this breaks binary compatibility with earlier
8227 versions of OpenSSL [engine].
8228 Adapt the nCipher code for these new conditions and add a card insertion
8232 *) Enhance the general user interface with mechanisms to better support
8233 dialog box interfaces, application-defined prompts, the possibility
8234 to use defaults (for example default passwords from somewhere else)
8235 and interrupts/cancellations.
8238 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
8239 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
8242 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
8243 tidy up some unnecessarily weird code in 'sk_new()').
8244 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
8246 *) Change the key loading routines for ENGINEs to use the same kind
8247 callback (pem_password_cb) as all other routines that need this
8251 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
8252 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
8253 than this minimum value is recommended.
8256 *) New random seeder for OpenVMS, using the system process statistics
8257 that are easily reachable.
8260 *) Windows apparently can't transparently handle global
8261 variables defined in DLLs. Initialisations such as:
8263 const ASN1_ITEM *it = &ASN1_INTEGER_it;
8265 won't compile. This is used by the any applications that need to
8266 declare their own ASN1 modules. This was fixed by adding the option
8267 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
8268 needed for static libraries under Win32.
8271 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
8272 setting of purpose and trust fields. New X509_STORE trust and
8273 purpose functions and tidy up setting in other SSL functions.
8276 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
8277 structure. These are inherited by X509_STORE_CTX when it is
8278 initialised. This allows various defaults to be set in the
8279 X509_STORE structure (such as flags for CRL checking and custom
8280 purpose or trust settings) for functions which only use X509_STORE_CTX
8281 internally such as S/MIME.
8283 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
8284 trust settings if they are not set in X509_STORE. This allows X509_STORE
8285 purposes and trust (in S/MIME for example) to override any set by default.
8287 Add command line options for CRL checking to smime, s_client and s_server
8291 *) Initial CRL based revocation checking. If the CRL checking flag(s)
8292 are set then the CRL is looked up in the X509_STORE structure and
8293 its validity and signature checked, then if the certificate is found
8294 in the CRL the verify fails with a revoked error.
8296 Various new CRL related callbacks added to X509_STORE_CTX structure.
8298 Command line options added to 'verify' application to support this.
8300 This needs some additional work, such as being able to handle multiple
8301 CRLs with different times, extension based lookup (rather than just
8302 by subject name) and ultimately more complete V2 CRL extension
8306 *) Add a general user interface API (crypto/ui/). This is designed
8307 to replace things like des_read_password and friends (backward
8308 compatibility functions using this new API are provided).
8309 The purpose is to remove prompting functions from the DES code
8310 section as well as provide for prompting through dialog boxes in
8311 a window system and the like.
8314 *) Add "ex_data" support to ENGINE so implementations can add state at a
8315 per-structure level rather than having to store it globally.
8318 *) Make it possible for ENGINE structures to be copied when retrieved by
8319 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
8320 This causes the "original" ENGINE structure to act like a template,
8321 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
8322 operational state can be localised to each ENGINE structure, despite the
8323 fact they all share the same "methods". New ENGINE structures returned in
8324 this case have no functional references and the return value is the single
8325 structural reference. This matches the single structural reference returned
8326 by ENGINE_by_id() normally, when it is incremented on the pre-existing
8330 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
8331 needs to match any other type at all we need to manually clear the
8335 *) Changes to the "openssl engine" utility to include;
8336 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
8337 about an ENGINE's available control commands.
8338 - executing control commands from command line arguments using the
8339 '-pre' and '-post' switches. '-post' is only used if '-t' is
8340 specified and the ENGINE is successfully initialised. The syntax for
8341 the individual commands are colon-separated, for example;
8342 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
8345 *) New dynamic control command support for ENGINEs. ENGINEs can now
8346 declare their own commands (numbers), names (strings), descriptions,
8347 and input types for run-time discovery by calling applications. A
8348 subset of these commands are implicitly classed as "executable"
8349 depending on their input type, and only these can be invoked through
8350 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
8351 can be based on user input, config files, etc). The distinction is
8352 that "executable" commands cannot return anything other than a boolean
8353 result and can only support numeric or string input, whereas some
8354 discoverable commands may only be for direct use through
8355 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
8356 pointers, or other custom uses. The "executable" commands are to
8357 support parameterisations of ENGINE behaviour that can be
8358 unambiguously defined by ENGINEs and used consistently across any
8359 OpenSSL-based application. Commands have been added to all the
8360 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
8361 control over shared-library paths without source code alterations.
8364 *) Changed all ENGINE implementations to dynamically allocate their
8365 ENGINEs rather than declaring them statically. Apart from this being
8366 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
8367 this also allows the implementations to compile without using the
8368 internal engine_int.h header.
8371 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8372 'const' value. Any code that should be able to modify a RAND_METHOD
8373 should already have non-const pointers to it (ie. they should only
8374 modify their own ones).
8377 *) Made a variety of little tweaks to the ENGINE code.
8378 - "atalla" and "ubsec" string definitions were moved from header files
8379 to C code. "nuron" string definitions were placed in variables
8380 rather than hard-coded - allowing parameterisation of these values
8381 later on via ctrl() commands.
8382 - Removed unused "#if 0"'d code.
8383 - Fixed engine list iteration code so it uses ENGINE_free() to release
8384 structural references.
8385 - Constified the RAND_METHOD element of ENGINE structures.
8386 - Constified various get/set functions as appropriate and added
8387 missing functions (including a catch-all ENGINE_cpy that duplicates
8388 all ENGINE values onto a new ENGINE except reference counts/state).
8389 - Removed NULL parameter checks in get/set functions. Setting a method
8390 or function to NULL is a way of cancelling out a previously set
8391 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8392 and doesn't justify the extra error symbols and code.
8393 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8394 flags from engine_int.h to engine.h.
8395 - Changed prototypes for ENGINE handler functions (init(), finish(),
8396 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8399 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8400 to the algorithm using long division. The binary algorithm can be
8401 used only if the modulus is odd. On 32-bit systems, it is faster
8402 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8403 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8404 up to 450 bits. In 64-bit environments, the binary algorithm
8405 appears to be advantageous for much longer moduli; here we use it
8406 for moduli up to 2048 bits.
8409 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8410 could not support the combine flag in choice fields.
8413 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8414 extensions from a certificate request to the certificate.
8417 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8418 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8419 file: this allows the display of the certificate about to be
8420 signed to be customised, to allow certain fields to be included
8421 or excluded and extension details. The old system didn't display
8422 multicharacter strings properly, omitted fields not in the policy
8423 and couldn't display additional details such as extensions.
8426 *) Function EC_POINTs_mul for multiple scalar multiplication
8427 of an arbitrary number of elliptic curve points
8428 \sum scalars[i]*points[i],
8429 optionally including the generator defined for the EC_GROUP:
8430 scalar*generator + \sum scalars[i]*points[i].
8432 EC_POINT_mul is a simple wrapper function for the typical case
8433 that the point list has just one item (besides the optional
8437 *) First EC_METHODs for curves over GF(p):
8439 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8440 operations and provides various method functions that can also
8441 operate with faster implementations of modular arithmetic.
8443 EC_GFp_mont_method() reuses most functions that are part of
8444 EC_GFp_simple_method, but uses Montgomery arithmetic.
8446 [Bodo Moeller; point addition and point doubling
8447 implementation directly derived from source code provided by
8448 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8450 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8451 crypto/ec/ec_lib.c):
8453 Curves are EC_GROUP objects (with an optional group generator)
8454 based on EC_METHODs that are built into the library.
8456 Points are EC_POINT objects based on EC_GROUP objects.
8458 Most of the framework would be able to handle curves over arbitrary
8459 finite fields, but as there are no obvious types for fields other
8460 than GF(p), some functions are limited to that for now.
8463 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8464 that the file contains a complete HTTP response.
8467 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8468 change the def and num file printf format specifier from "%-40sXXX"
8469 to "%-39s XXX". The latter will always guarantee a space after the
8470 field while the former will cause them to run together if the field
8471 is 40 of more characters long.
8474 *) Constify the cipher and digest 'method' functions and structures
8475 and modify related functions to take constant EVP_MD and EVP_CIPHER
8479 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8480 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8483 *) Modify EVP_Digest*() routines so they now return values. Although the
8484 internal software routines can never fail additional hardware versions
8488 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8490 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8491 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8494 ERR_R_NESTED_ASN1_ERROR
8496 ERR_R_MISSING_ASN1_EOS
8497 were 4 .. 9, conflicting with
8498 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8500 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8501 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8503 Add new error code 'ERR_R_INTERNAL_ERROR'.
8506 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8510 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8511 sets the subject name for a new request or supersedes the
8512 subject name in a given request. Formats that can be parsed are
8513 'CN=Some Name, OU=myOU, C=IT'
8515 'CN=Some Name/OU=myOU/C=IT'.
8517 Add options '-batch' and '-verbose' to 'openssl req'.
8518 [Massimiliano Pala <madwolf@hackmasters.net>]
8520 *) Introduce the possibility to access global variables through
8521 functions on platform were that's the best way to handle exporting
8522 global variables in shared libraries. To enable this functionality,
8523 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8524 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8525 is normally done by Configure or something similar).
8527 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8528 in the source file (foo.c) like this:
8530 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8531 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8533 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8534 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8536 OPENSSL_DECLARE_GLOBAL(int,foo);
8537 #define foo OPENSSL_GLOBAL_REF(foo)
8538 OPENSSL_DECLARE_GLOBAL(double,bar);
8539 #define bar OPENSSL_GLOBAL_REF(bar)
8541 The #defines are very important, and therefore so is including the
8542 header file everywhere where the defined globals are used.
8544 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8545 of ASN.1 items, but that structure is a bit different.
8547 The largest change is in util/mkdef.pl which has been enhanced with
8548 better and easier to understand logic to choose which symbols should
8549 go into the Windows .def files as well as a number of fixes and code
8550 cleanup (among others, algorithm keywords are now sorted
8551 lexicographically to avoid constant rewrites).
8554 *) In BN_div() keep a copy of the sign of 'num' before writing the
8555 result to 'rm' because if rm==num the value will be overwritten
8556 and produce the wrong result if 'num' is negative: this caused
8557 problems with BN_mod() and BN_nnmod().
8560 *) Function OCSP_request_verify(). This checks the signature on an
8561 OCSP request and verifies the signer certificate. The signer
8562 certificate is just checked for a generic purpose and OCSP request
8566 *) Add OCSP_check_validity() function to check the validity of OCSP
8567 responses. OCSP responses are prepared in real time and may only
8568 be a few seconds old. Simply checking that the current time lies
8569 between thisUpdate and nextUpdate max reject otherwise valid responses
8570 caused by either OCSP responder or client clock inaccuracy. Instead
8571 we allow thisUpdate and nextUpdate to fall within a certain period of
8572 the current time. The age of the response can also optionally be
8573 checked. Two new options -validity_period and -status_age added to
8577 *) If signature or public key algorithm is unrecognized print out its
8578 OID rather that just UNKNOWN.
8581 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8582 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8583 ID to be generated from the issuer certificate alone which can then be
8584 passed to OCSP_id_issuer_cmp().
8587 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8588 ASN1 modules to export functions returning ASN1_ITEM pointers
8589 instead of the ASN1_ITEM structures themselves. This adds several
8590 new macros which allow the underlying ASN1 function/structure to
8591 be accessed transparently. As a result code should not use ASN1_ITEM
8592 references directly (such as &X509_it) but instead use the relevant
8593 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8594 use of the new ASN1 code on platforms where exporting structures
8595 is problematical (for example in shared libraries) but exporting
8596 functions returning pointers to structures is not.
8599 *) Add support for overriding the generation of SSL/TLS session IDs.
8600 These callbacks can be registered either in an SSL_CTX or per SSL.
8601 The purpose of this is to allow applications to control, if they wish,
8602 the arbitrary values chosen for use as session IDs, particularly as it
8603 can be useful for session caching in multiple-server environments. A
8604 command-line switch for testing this (and any client code that wishes
8605 to use such a feature) has been added to "s_server".
8606 [Geoff Thorpe, Lutz Jaenicke]
8608 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8609 of the form '#if defined(...) || defined(...) || ...' and
8610 '#if !defined(...) && !defined(...) && ...'. This also avoids
8611 the growing number of special cases it was previously handling.
8614 *) Make all configuration macros available for application by making
8615 sure they are available in opensslconf.h, by giving them names starting
8616 with "OPENSSL_" to avoid conflicts with other packages and by making
8617 sure e_os2.h will cover all platform-specific cases together with
8619 Additionally, it is now possible to define configuration/platform-
8620 specific names (called "system identities"). In the C code, these
8621 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8622 macro with the name beginning with "OPENSSL_SYS_", which is determined
8623 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8627 *) New option -set_serial to 'req' and 'x509' this allows the serial
8628 number to use to be specified on the command line. Previously self
8629 signed certificates were hard coded with serial number 0 and the
8630 CA options of 'x509' had to use a serial number in a file which was
8634 *) New options to 'ca' utility to support V2 CRL entry extensions.
8635 Currently CRL reason, invalidity date and hold instruction are
8636 supported. Add new CRL extensions to V3 code and some new objects.
8639 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8640 disable standard block padding (aka PKCS#5 padding) in the EVP
8641 API, which was previously mandatory. This means that the data is
8642 not padded in any way and so the total length much be a multiple
8643 of the block size, otherwise an error occurs.
8646 *) Initial (incomplete) OCSP SSL support.
8649 *) New function OCSP_parse_url(). This splits up a URL into its host,
8650 port and path components: primarily to parse OCSP URLs. New -url
8651 option to ocsp utility.
8654 *) New nonce behavior. The return value of OCSP_check_nonce() now
8655 reflects the various checks performed. Applications can decide
8656 whether to tolerate certain situations such as an absent nonce
8657 in a response when one was present in a request: the ocsp application
8658 just prints out a warning. New function OCSP_add1_basic_nonce()
8659 this is to allow responders to include a nonce in a response even if
8660 the request is nonce-less.
8663 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8664 skipped when using openssl x509 multiple times on a single input file,
8665 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8668 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8669 set string type: to handle setting ASN1_TIME structures. Fix ca
8670 utility to correctly initialize revocation date of CRLs.
8673 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8674 the clients preferred ciphersuites and rather use its own preferences.
8675 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8676 Internet Explorer by ensuring unchanged hash method during stepup.
8677 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8680 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8681 to aes and add a new 'exist' option to print out symbols that don't
8685 *) Additional options to ocsp utility to allow flags to be set and
8686 additional certificates supplied.
8689 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8690 OCSP client a number of certificate to only verify the response
8694 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8695 handle the new API. Currently only ECB, CBC modes supported. Add new
8698 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8699 Encryption Standard (AES) Ciphersuites for Transport Layer
8700 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8701 not enabled by default and were not part of the "ALL" ciphersuite
8702 alias because they were not yet official; they could be
8703 explicitly requested by specifying the "AESdraft" ciphersuite
8704 group alias. In the final release of OpenSSL 0.9.7, the group
8705 alias is called "AES" and is part of "ALL".)
8706 [Ben Laurie, Steve Henson, Bodo Moeller]
8708 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8709 request to response.
8712 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8713 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8714 extract information from a certificate request. OCSP_response_create()
8715 creates a response and optionally adds a basic response structure.
8716 OCSP_basic_add1_status() adds a complete single response to a basic
8717 response and returns the OCSP_SINGLERESP structure just added (to allow
8718 extensions to be included for example). OCSP_basic_add1_cert() adds a
8719 certificate to a basic response and OCSP_basic_sign() signs a basic
8720 response with various flags. New helper functions ASN1_TIME_check()
8721 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8722 (converts ASN1_TIME to GeneralizedTime).
8725 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8726 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8727 structure from a certificate. X509_pubkey_digest() digests the public_key
8728 contents: this is used in various key identifiers.
8731 *) Make sk_sort() tolerate a NULL argument.
8732 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8734 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8735 passed by the function are trusted implicitly. If any of them signed the
8736 response then it is assumed to be valid and is not verified.
8739 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8740 to data. This was previously part of the PKCS7 ASN1 code. This
8741 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8742 [Steve Henson, reported by Kenneth R. Robinette
8743 <support@securenetterm.com>]
8745 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8746 routines: without these tracing memory leaks is very painful.
8747 Fix leaks in PKCS12 and PKCS7 routines.
8750 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8751 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8752 effectively meant GeneralizedTime would never be used. Now it
8753 is initialised to -1 but X509_time_adj() now has to check the value
8754 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8755 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8756 [Steve Henson, reported by Kenneth R. Robinette
8757 <support@securenetterm.com>]
8759 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8760 result in a zero length in the ASN1_INTEGER structure which was
8761 not consistent with the structure when d2i_ASN1_INTEGER() was used
8762 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8763 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8764 where it did not print out a minus for negative ASN1_INTEGER.
8767 *) Add summary printout to ocsp utility. The various functions which
8768 convert status values to strings have been renamed to:
8769 OCSP_response_status_str(), OCSP_cert_status_str() and
8770 OCSP_crl_reason_str() and are no longer static. New options
8771 to verify nonce values and to disable verification. OCSP response
8772 printout format cleaned up.
8775 *) Add additional OCSP certificate checks. These are those specified
8776 in RFC2560. This consists of two separate checks: the CA of the
8777 certificate being checked must either be the OCSP signer certificate
8778 or the issuer of the OCSP signer certificate. In the latter case the
8779 OCSP signer certificate must contain the OCSP signing extended key
8780 usage. This check is performed by attempting to match the OCSP
8781 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8782 in the OCSP_CERTID structures of the response.
8785 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8786 and related routines. This uses the standard OpenSSL certificate
8787 verify routines to perform initial checks (just CA validity) and
8788 to obtain the certificate chain. Then additional checks will be
8789 performed on the chain. Currently the root CA is checked to see
8790 if it is explicitly trusted for OCSP signing. This is used to set
8791 a root CA as a global signing root: that is any certificate that
8792 chains to that CA is an acceptable OCSP signing certificate.
8795 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8796 extensions from a separate configuration file.
8797 As when reading extensions from the main configuration file,
8798 the '-extensions ...' option may be used for specifying the
8800 [Massimiliano Pala <madwolf@comune.modena.it>]
8802 *) New OCSP utility. Allows OCSP requests to be generated or
8803 read. The request can be sent to a responder and the output
8804 parsed, outputted or printed in text form. Not complete yet:
8805 still needs to check the OCSP response validity.
8808 *) New subcommands for 'openssl ca':
8809 'openssl ca -status <serial>' prints the status of the cert with
8810 the given serial number (according to the index file).
8811 'openssl ca -updatedb' updates the expiry status of certificates
8813 [Massimiliano Pala <madwolf@comune.modena.it>]
8815 *) New '-newreq-nodes' command option to CA.pl. This is like
8816 '-newreq', but calls 'openssl req' with the '-nodes' option
8817 so that the resulting key is not encrypted.
8818 [Damien Miller <djm@mindrot.org>]
8820 *) New configuration for the GNU Hurd.
8821 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8823 *) Initial code to implement OCSP basic response verify. This
8824 is currently incomplete. Currently just finds the signer's
8825 certificate and verifies the signature on the response.
8828 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8829 value of OPENSSLDIR. This is available via the new '-d' option
8830 to 'openssl version', and is also included in 'openssl version -a'.
8833 *) Allowing defining memory allocation callbacks that will be given
8834 file name and line number information in additional arguments
8835 (a const char* and an int). The basic functionality remains, as
8836 well as the original possibility to just replace malloc(),
8837 realloc() and free() by functions that do not know about these
8838 additional arguments. To register and find out the current
8839 settings for extended allocation functions, the following
8840 functions are provided:
8842 CRYPTO_set_mem_ex_functions
8843 CRYPTO_set_locked_mem_ex_functions
8844 CRYPTO_get_mem_ex_functions
8845 CRYPTO_get_locked_mem_ex_functions
8847 These work the same way as CRYPTO_set_mem_functions and friends.
8848 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8849 extended allocation function is enabled.
8850 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8851 a conventional allocation function is enabled.
8852 [Richard Levitte, Bodo Moeller]
8854 *) Finish off removing the remaining LHASH function pointer casts.
8855 There should no longer be any prototype-casting required when using
8856 the LHASH abstraction, and any casts that remain are "bugs". See
8857 the callback types and macros at the head of lhash.h for details
8858 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8861 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8862 If /dev/[u]random devices are not available or do not return enough
8863 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8865 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8866 /etc/entropy will be queried once each in this sequence, querying stops
8867 when enough entropy was collected without querying more sockets.
8870 *) Change the Unix RAND_poll() variant to be able to poll several
8871 random devices, as specified by DEVRANDOM, until a sufficient amount
8872 of data has been collected. We spend at most 10 ms on each file
8873 (select timeout) and read in non-blocking mode. DEVRANDOM now
8874 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8875 (previously it was just the string "/dev/urandom"), so on typical
8876 platforms the 10 ms delay will never occur.
8877 Also separate out the Unix variant to its own file, rand_unix.c.
8878 For VMS, there's a currently-empty rand_vms.c.
8881 *) Move OCSP client related routines to ocsp_cl.c. These
8882 provide utility functions which an application needing
8883 to issue a request to an OCSP responder and analyse the
8884 response will typically need: as opposed to those which an
8885 OCSP responder itself would need which will be added later.
8887 OCSP_request_sign() signs an OCSP request with an API similar
8888 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8889 response. OCSP_response_get1_basic() extracts basic response
8890 from response. OCSP_resp_find_status(): finds and extracts status
8891 information from an OCSP_CERTID structure (which will be created
8892 when the request structure is built). These are built from lower
8893 level functions which work on OCSP_SINGLERESP structures but
8894 won't normally be used unless the application wishes to examine
8895 extensions in the OCSP response for example.
8897 Replace nonce routines with a pair of functions.
8898 OCSP_request_add1_nonce() adds a nonce value and optionally
8899 generates a random value. OCSP_check_nonce() checks the
8900 validity of the nonce in an OCSP response.
8903 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8904 This doesn't copy the supplied OCSP_CERTID and avoids the
8905 need to free up the newly created id. Change return type
8906 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8907 This can then be used to add extensions to the request.
8908 Deleted OCSP_request_new(), since most of its functionality
8909 is now in OCSP_REQUEST_new() (and the case insensitive name
8910 clash) apart from the ability to set the request name which
8911 will be added elsewhere.
8914 *) Update OCSP API. Remove obsolete extensions argument from
8915 various functions. Extensions are now handled using the new
8916 OCSP extension code. New simple OCSP HTTP function which
8917 can be used to send requests and parse the response.
8920 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8921 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8922 uses the special reorder version of SET OF to sort the attributes
8923 and reorder them to match the encoded order. This resolves a long
8924 standing problem: a verify on a PKCS7 structure just after signing
8925 it used to fail because the attribute order did not match the
8926 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8927 it uses the received order. This is necessary to tolerate some broken
8928 software that does not order SET OF. This is handled by encoding
8929 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8930 to produce the required SET OF.
8933 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8934 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8935 files to get correct declarations of the ASN.1 item variables.
8938 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8939 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8940 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8941 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8942 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8943 ASN1_ITEM and no wrapper functions.
8946 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8947 replace the old function pointer based I/O routines. Change most of
8948 the *_d2i_bio() and *_d2i_fp() functions to use these.
8951 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8952 lines, recognize more "algorithms" that can be deselected, and make
8953 it complain about algorithm deselection that isn't recognised.
8956 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8957 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8958 to use new functions. Add NO_ASN1_OLD which can be set to remove
8959 some old style ASN1 functions: this can be used to determine if old
8960 code will still work when these eventually go away.
8963 *) New extension functions for OCSP structures, these follow the
8964 same conventions as certificates and CRLs.
8967 *) New function X509V3_add1_i2d(). This automatically encodes and
8968 adds an extension. Its behaviour can be customised with various
8969 flags to append, replace or delete. Various wrappers added for
8970 certificates and CRLs.
8973 *) Fix to avoid calling the underlying ASN1 print routine when
8974 an extension cannot be parsed. Correct a typo in the
8975 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8978 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8979 entries for variables.
8982 *) Add functionality to apps/openssl.c for detecting locking
8983 problems: As the program is single-threaded, all we have
8984 to do is register a locking callback using an array for
8985 storing which locks are currently held by the program.
8988 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8989 SSL_get_ex_data_X509_STORE_idx(), which is used in
8990 ssl_verify_cert_chain() and thus can be called at any time
8991 during TLS/SSL handshakes so that thread-safety is essential.
8992 Unfortunately, the ex_data design is not at all suited
8993 for multi-threaded use, so it probably should be abolished.
8996 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8997 [Broadcom, tweaked and integrated by Geoff Thorpe]
8999 *) Move common extension printing code to new function
9000 X509V3_print_extensions(). Reorganise OCSP print routines and
9001 implement some needed OCSP ASN1 functions. Add OCSP extensions.
9004 *) New function X509_signature_print() to remove duplication in some
9008 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
9009 set (this was treated exactly the same as SET OF previously). This
9010 is used to reorder the STACK representing the structure to match the
9011 encoding. This will be used to get round a problem where a PKCS7
9012 structure which was signed could not be verified because the STACK
9013 order did not reflect the encoded order.
9016 *) Reimplement the OCSP ASN1 module using the new code.
9019 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
9020 for its ASN1 operations. The old style function pointers still exist
9021 for now but they will eventually go away.
9024 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
9025 completely replaces the old ASN1 functionality with a table driven
9026 encoder and decoder which interprets an ASN1_ITEM structure describing
9027 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
9028 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
9029 has also been converted to the new form.
9032 *) Change BN_mod_exp_recp so that negative moduli are tolerated
9033 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
9034 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
9035 for negative moduli.
9038 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
9039 of not touching the result's sign bit.
9042 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
9046 *) Changed the LHASH code to use prototypes for callbacks, and created
9047 macros to declare and implement thin (optionally static) functions
9048 that provide type-safety and avoid function pointer casting for the
9049 type-specific callbacks.
9052 *) Added Kerberos Cipher Suites to be used with TLS, as written in
9054 [Veers Staats <staatsvr@asc.hpc.mil>,
9055 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
9057 *) Reformat the FAQ so the different questions and answers can be divided
9058 in sections depending on the subject.
9061 *) Have the zlib compression code load ZLIB.DLL dynamically under
9065 *) New function BN_mod_sqrt for computing square roots modulo a prime
9066 (using the probabilistic Tonelli-Shanks algorithm unless
9067 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
9068 be handled deterministically).
9069 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9071 *) Make BN_mod_inverse faster by explicitly handling small quotients
9072 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
9073 512 bits], about 30% for larger ones [1024 or 2048 bits].)
9076 *) New function BN_kronecker.
9079 *) Fix BN_gcd so that it works on negative inputs; the result is
9080 positive unless both parameters are zero.
9081 Previously something reasonably close to an infinite loop was
9082 possible because numbers could be growing instead of shrinking
9083 in the implementation of Euclid's algorithm.
9086 *) Fix BN_is_word() and BN_is_one() macros to take into account the
9087 sign of the number in question.
9089 Fix BN_is_word(a,w) to work correctly for w == 0.
9091 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
9092 because its test if the absolute value of 'a' equals 'w'.
9093 Note that BN_abs_is_word does *not* handle w == 0 reliably;
9094 it exists mostly for use in the implementations of BN_is_zero(),
9095 BN_is_one(), and BN_is_word().
9098 *) New function BN_swap.
9101 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
9102 the exponentiation functions are more likely to produce reasonable
9103 results on negative inputs.
9106 *) Change BN_mod_mul so that the result is always non-negative.
9107 Previously, it could be negative if one of the factors was negative;
9108 I don't think anyone really wanted that behaviour.
9111 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
9112 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
9113 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
9114 and add new functions:
9123 BN_mod_lshift1_quick
9127 These functions always generate non-negative results.
9129 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
9130 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
9132 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
9133 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
9134 be reduced modulo m.
9135 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9138 The following entry accidentally appeared in the CHANGES file
9139 distributed with OpenSSL 0.9.7. The modifications described in
9140 it do *not* apply to OpenSSL 0.9.7.
9142 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
9143 was actually never needed) and in BN_mul(). The removal in BN_mul()
9144 required a small change in bn_mul_part_recursive() and the addition
9145 of the functions bn_cmp_part_words(), bn_sub_part_words() and
9146 bn_add_part_words(), which do the same thing as bn_cmp_words(),
9147 bn_sub_words() and bn_add_words() except they take arrays with
9152 *) In 'openssl passwd', verify passwords read from the terminal
9153 unless the '-salt' option is used (which usually means that
9154 verification would just waste user's time since the resulting
9155 hash is going to be compared with some given password hash)
9156 or the new '-noverify' option is used.
9158 This is an incompatible change, but it does not affect
9159 non-interactive use of 'openssl passwd' (passwords on the command
9160 line, '-stdin' option, '-in ...' option) and thus should not
9164 *) Remove all references to RSAref, since there's no more need for it.
9167 *) Make DSO load along a path given through an environment variable
9168 (SHLIB_PATH) with shl_load().
9171 *) Constify the ENGINE code as a result of BIGNUM constification.
9172 Also constify the RSA code and most things related to it. In a
9173 few places, most notable in the depth of the ASN.1 code, ugly
9174 casts back to non-const were required (to be solved at a later
9178 *) Make it so the openssl application has all engines loaded by default.
9181 *) Constify the BIGNUM routines a little more.
9184 *) Add the following functions:
9186 ENGINE_load_cswift()
9188 ENGINE_load_atalla()
9190 ENGINE_load_builtin_engines()
9192 That way, an application can itself choose if external engines that
9193 are built-in in OpenSSL shall ever be used or not. The benefit is
9194 that applications won't have to be linked with libdl or other dso
9195 libraries unless it's really needed.
9197 Changed 'openssl engine' to load all engines on demand.
9198 Changed the engine header files to avoid the duplication of some
9199 declarations (they differed!).
9202 *) 'openssl engine' can now list capabilities.
9205 *) Better error reporting in 'openssl engine'.
9208 *) Never call load_dh_param(NULL) in s_server.
9211 *) Add engine application. It can currently list engines by name and
9212 identity, and test if they are actually available.
9215 *) Improve RPM specification file by forcing symbolic linking and making
9216 sure the installed documentation is also owned by root.root.
9217 [Damien Miller <djm@mindrot.org>]
9219 *) Give the OpenSSL applications more possibilities to make use of
9220 keys (public as well as private) handled by engines.
9223 *) Add OCSP code that comes from CertCo.
9226 *) Add VMS support for the Rijndael code.
9229 *) Added untested support for Nuron crypto accelerator.
9232 *) Add support for external cryptographic devices. This code was
9233 previously distributed separately as the "engine" branch.
9234 [Geoff Thorpe, Richard Levitte]
9236 *) Rework the filename-translation in the DSO code. It is now possible to
9237 have far greater control over how a "name" is turned into a filename
9238 depending on the operating environment and any oddities about the
9239 different shared library filenames on each system.
9242 *) Support threads on FreeBSD-elf in Configure.
9245 *) Fix for SHA1 assembly problem with MASM: it produces
9246 warnings about corrupt line number information when assembling
9247 with debugging information. This is caused by the overlapping
9249 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
9252 NCONF_get_number() has no error checking at all. As a replacement,
9253 NCONF_get_number_e() is defined (_e for "error checking") and is
9254 promoted strongly. The old NCONF_get_number is kept around for
9255 binary backward compatibility.
9256 Make it possible for methods to load from something other than a BIO,
9257 by providing a function pointer that is given a name instead of a BIO.
9258 For example, this could be used to load configuration data from an
9262 *) Fix for non blocking accept BIOs. Added new I/O special reason
9263 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
9264 with non blocking I/O was not possible because no retry code was
9265 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
9269 *) Added the beginnings of Rijndael support.
9272 *) Fix for bug in DirectoryString mask setting. Add support for
9273 X509_NAME_print_ex() in 'req' and X509_print_ex() function
9274 to allow certificate printing to more controllable, additional
9275 'certopt' option to 'x509' to allow new printing options to be
9279 *) Clean old EAY MD5 hack from e_os.h.
9282 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
9284 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
9285 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
9286 [Joe Orton, Steve Henson]
9288 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
9290 *) Fix additional bug revealed by the NISCC test suite:
9292 Stop bug triggering large recursion when presented with
9293 certain ASN.1 tags (CVE-2003-0851)
9296 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
9298 *) Fix various bugs revealed by running the NISCC test suite:
9300 Stop out of bounds reads in the ASN1 code when presented with
9301 invalid tags (CVE-2003-0543 and CVE-2003-0544).
9303 If verify callback ignores invalid public key errors don't try to check
9304 certificate signature with the NULL public key.
9308 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
9309 if the server requested one: as stated in TLS 1.0 and SSL 3.0
9313 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
9314 extra data after the compression methods not only for TLS 1.0
9315 but also for SSL 3.0 (as required by the specification).
9316 [Bodo Moeller; problem pointed out by Matthias Loepfe]
9318 *) Change X509_certificate_type() to mark the key as exported/exportable
9319 when it's 512 *bits* long, not 512 bytes.
9322 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
9324 *) Countermeasure against the Klima-Pokorny-Rosa extension of
9325 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
9326 a protocol version number mismatch like a decryption error
9327 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
9330 *) Turn on RSA blinding by default in the default implementation
9331 to avoid a timing attack. Applications that don't want it can call
9332 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
9333 They would be ill-advised to do so in most cases.
9334 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
9336 *) Change RSA blinding code so that it works when the PRNG is not
9337 seeded (in this case, the secret RSA exponent is abused as
9338 an unpredictable seed -- if it is not unpredictable, there
9339 is no point in blinding anyway). Make RSA blinding thread-safe
9340 by remembering the creator's thread ID in rsa->blinding and
9341 having all other threads use local one-time blinding factors
9342 (this requires more computation than sharing rsa->blinding, but
9343 avoids excessive locking; and if an RSA object is not shared
9344 between threads, blinding will still be very fast).
9347 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
9349 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
9350 via timing by performing a MAC computation even if incorrect
9351 block cipher padding has been found. This is a countermeasure
9352 against active attacks where the attacker has to distinguish
9353 between bad padding and a MAC verification error. (CVE-2003-0078)
9355 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
9356 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
9357 Martin Vuagnoux (EPFL, Ilion)]
9359 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
9361 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
9362 memory from its contents. This is done with a counter that will
9363 place alternating values in each byte. This can be used to solve
9364 two issues: 1) the removal of calls to memset() by highly optimizing
9365 compilers, and 2) cleansing with other values than 0, since those can
9366 be read through on certain media, for example a swap space on disk.
9369 *) Bugfix: client side session caching did not work with external caching,
9370 because the session->cipher setting was not restored when reloading
9371 from the external cache. This problem was masked, when
9372 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9373 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9376 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9377 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9378 [Zeev Lieber <zeev-l@yahoo.com>]
9380 *) Undo an undocumented change introduced in 0.9.6e which caused
9381 repeated calls to OpenSSL_add_all_ciphers() and
9382 OpenSSL_add_all_digests() to be ignored, even after calling
9386 *) Change the default configuration reader to deal with last line not
9387 being properly terminated.
9390 *) Change X509_NAME_cmp() so it applies the special rules on handling
9391 DN values that are of type PrintableString, as well as RDNs of type
9392 emailAddress where the value has the type ia5String.
9393 [stefank@valicert.com via Richard Levitte]
9395 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9396 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9397 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9398 the bitwise-OR of the two for use by the majority of applications
9399 wanting this behaviour, and update the docs. The documented
9400 behaviour and actual behaviour were inconsistent and had been
9401 changing anyway, so this is more a bug-fix than a behavioural
9403 [Geoff Thorpe, diagnosed by Nadav Har'El]
9405 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9406 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9409 *) Fix initialization code race conditions in
9410 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9411 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9412 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9413 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9414 ssl2_get_cipher_by_char(),
9415 ssl3_get_cipher_by_char().
9416 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9418 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9419 the cached sessions are flushed, as the remove_cb() might use ex_data
9420 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9421 (see [openssl.org #212]).
9422 [Geoff Thorpe, Lutz Jaenicke]
9424 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9425 length, instead of the encoding length to d2i_ASN1_OBJECT.
9428 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9430 *) [In 0.9.6g-engine release:]
9431 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9432 [Lynn Gazis <lgazis@rainbow.com>]
9434 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9436 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9437 and get fix the header length calculation.
9438 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9439 Alon Kantor <alonk@checkpoint.com> (and others),
9442 *) Use proper error handling instead of 'assertions' in buffer
9443 overflow checks added in 0.9.6e. This prevents DoS (the
9444 assertions could call abort()).
9445 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9447 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9449 *) Add various sanity checks to asn1_get_length() to reject
9450 the ASN1 length bytes if they exceed sizeof(long), will appear
9451 negative or the content length exceeds the length of the
9453 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9455 *) Fix cipher selection routines: ciphers without encryption had no flags
9456 for the cipher strength set and where therefore not handled correctly
9457 by the selection routines (PR #130).
9460 *) Fix EVP_dsa_sha macro.
9464 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9465 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9466 that was added in OpenSSL 0.9.6d.
9468 As the countermeasure turned out to be incompatible with some
9469 broken SSL implementations, the new option is part of SSL_OP_ALL.
9470 SSL_OP_ALL is usually employed when compatibility with weird SSL
9471 implementations is desired (e.g. '-bugs' option to 's_client' and
9472 's_server'), so the new option is automatically set in many
9476 *) Changes in security patch:
9478 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9479 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9480 Air Force Materiel Command, USAF, under agreement number
9483 *) Add various sanity checks to asn1_get_length() to reject
9484 the ASN1 length bytes if they exceed sizeof(long), will appear
9485 negative or the content length exceeds the length of the
9486 supplied buffer. (CVE-2002-0659)
9487 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9489 *) Assertions for various potential buffer overflows, not known to
9491 [Ben Laurie (CHATS)]
9493 *) Various temporary buffers to hold ASCII versions of integers were
9494 too small for 64 bit platforms. (CVE-2002-0655)
9495 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9497 *) Remote buffer overflow in SSL3 protocol - an attacker could
9498 supply an oversized session ID to a client. (CVE-2002-0656)
9499 [Ben Laurie (CHATS)]
9501 *) Remote buffer overflow in SSL2 protocol - an attacker could
9502 supply an oversized client master key. (CVE-2002-0656)
9503 [Ben Laurie (CHATS)]
9505 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9507 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9508 encoded as NULL) with id-dsa-with-sha1.
9509 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9511 *) Check various X509_...() return values in apps/req.c.
9512 [Nils Larsch <nla@trustcenter.de>]
9514 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9515 an end-of-file condition would erroneously be flagged, when the CRLF
9516 was just at the end of a processed block. The bug was discovered when
9517 processing data through a buffering memory BIO handing the data to a
9518 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9519 <ptsekov@syntrex.com> and Nedelcho Stanev.
9522 *) Implement a countermeasure against a vulnerability recently found
9523 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9524 before application data chunks to avoid the use of known IVs
9525 with data potentially chosen by the attacker.
9528 *) Fix length checks in ssl3_get_client_hello().
9531 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9532 to prevent ssl3_read_internal() from incorrectly assuming that
9533 ssl3_read_bytes() found application data while handshake
9534 processing was enabled when in fact s->s3->in_read_app_data was
9535 merely automatically cleared during the initial handshake.
9536 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9538 *) Fix object definitions for Private and Enterprise: they were not
9539 recognized in their shortname (=lowercase) representation. Extend
9540 obj_dat.pl to issue an error when using undefined keywords instead
9541 of silently ignoring the problem (Svenning Sorensen
9542 <sss@sss.dnsalias.net>).
9545 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9546 generators, i.e. generators other than 2 and 5. (Previously, the
9547 code did not properly initialise the 'add' and 'rem' values to
9548 BN_generate_prime().)
9550 In the new general case, we do not insist that 'generator' is
9551 actually a primitive root: This requirement is rather pointless;
9552 a generator of the order-q subgroup is just as good, if not
9556 *) Map new X509 verification errors to alerts. Discovered and submitted by
9557 Tom Wu <tom@arcot.com>.
9560 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9561 returning non-zero before the data has been completely received
9562 when using non-blocking I/O.
9563 [Bodo Moeller; problem pointed out by John Hughes]
9565 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9566 [Ben Laurie, Lutz Jaenicke]
9568 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9569 Yoram Zahavi <YoramZ@gilian.com>).
9572 *) Add information about CygWin 1.3 and on, and preserve proper
9573 configuration for the versions before that.
9574 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9576 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9577 check whether we deal with a copy of a session and do not delete from
9578 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9579 <izhar@checkpoint.com>.
9582 *) Do not store session data into the internal session cache, if it
9583 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9584 flag is set). Proposed by Aslam <aslam@funk.com>.
9587 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9591 *) [In 0.9.6d-engine release:]
9592 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9593 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9595 *) Add the configuration target linux-s390x.
9596 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9598 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9599 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9600 variable as an indication that a ClientHello message has been
9601 received. As the flag value will be lost between multiple
9602 invocations of ssl3_accept when using non-blocking I/O, the
9603 function may not be aware that a handshake has actually taken
9604 place, thus preventing a new session from being added to the
9607 To avoid this problem, we now set s->new_session to 2 instead of
9608 using a local variable.
9609 [Lutz Jaenicke, Bodo Moeller]
9611 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9612 if the SSL_R_LENGTH_MISMATCH error is detected.
9613 [Geoff Thorpe, Bodo Moeller]
9615 *) New 'shared_ldflag' column in Configure platform table.
9618 *) Fix EVP_CIPHER_mode macro.
9619 ["Dan S. Camper" <dan@bti.net>]
9621 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9622 type, we must throw them away by setting rr->length to 0.
9623 [D P Chang <dpc@qualys.com>]
9625 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9627 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9628 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9629 worked incorrectly for those cases where range = 10..._2 and
9630 3*range is two bits longer than range.)
9633 *) Only add signing time to PKCS7 structures if it is not already
9637 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9638 OBJ_ld_ce should be OBJ_id_ce.
9639 Also some ip-pda OIDs in crypto/objects/objects.txt were
9640 incorrect (cf. RFC 3039).
9641 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9643 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9644 returns early because it has nothing to do.
9645 [Andy Schneider <andy.schneider@bjss.co.uk>]
9647 *) [In 0.9.6c-engine release:]
9648 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9649 [Andy Schneider <andy.schneider@bjss.co.uk>]
9651 *) [In 0.9.6c-engine release:]
9652 Add support for Cryptographic Appliance's keyserver technology.
9653 (Use engine 'keyclient')
9654 [Cryptographic Appliances and Geoff Thorpe]
9656 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9657 is called via tools/c89.sh because arguments have to be
9658 rearranged (all '-L' options must appear before the first object
9660 [Richard Shapiro <rshapiro@abinitio.com>]
9662 *) [In 0.9.6c-engine release:]
9663 Add support for Broadcom crypto accelerator cards, backported
9665 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9667 *) [In 0.9.6c-engine release:]
9668 Add support for SureWare crypto accelerator cards from
9669 Baltimore Technologies. (Use engine 'sureware')
9670 [Baltimore Technologies and Mark Cox]
9672 *) [In 0.9.6c-engine release:]
9673 Add support for crypto accelerator cards from Accelerated
9674 Encryption Processing, www.aep.ie. (Use engine 'aep')
9675 [AEP Inc. and Mark Cox]
9677 *) Add a configuration entry for gcc on UnixWare.
9678 [Gary Benson <gbenson@redhat.com>]
9680 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9681 messages are stored in a single piece (fixed-length part and
9682 variable-length part combined) and fix various bugs found on the way.
9685 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9686 instead. BIO_gethostbyname() does not know what timeouts are
9687 appropriate, so entries would stay in cache even when they have
9689 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9691 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9692 faced with a pathologically small ClientHello fragment that does
9693 not contain client_version: Instead of aborting with an error,
9694 simply choose the highest available protocol version (i.e.,
9695 TLS 1.0 unless it is disabled). In practice, ClientHello
9696 messages are never sent like this, but this change gives us
9697 strictly correct behaviour at least for TLS.
9700 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9701 never resets s->method to s->ctx->method when called from within
9702 one of the SSL handshake functions.
9703 [Bodo Moeller; problem pointed out by Niko Baric]
9705 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9706 (sent using the client's version number) if client_version is
9707 smaller than the protocol version in use. Also change
9708 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9709 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9710 the client will at least see that alert.
9713 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9717 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9718 client receives HelloRequest while in a handshake.
9719 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9721 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9722 should end in 'break', not 'goto end' which circumvents various
9723 cleanups done in state SSL_ST_OK. But session related stuff
9724 must be disabled for SSL_ST_OK in the case that we just sent a
9727 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9728 before just sending a HelloRequest.
9729 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9731 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9732 reveal whether illegal block cipher padding was found or a MAC
9733 verification error occurred. (Neither SSLerr() codes nor alerts
9734 are directly visible to potential attackers, but the information
9735 may leak via logfiles.)
9737 Similar changes are not required for the SSL 2.0 implementation
9738 because the number of padding bytes is sent in clear for SSL 2.0,
9739 and the extra bytes are just ignored. However ssl/s2_pkt.c
9740 failed to verify that the purported number of padding bytes is in
9744 *) Add OpenUNIX-8 support including shared libraries
9745 (Boyd Lynn Gerber <gerberb@zenez.com>).
9748 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9749 'wristwatch attack' using huge encoding parameters (cf.
9750 James H. Manger's CRYPTO 2001 paper). Note that the
9751 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9752 encoding parameters and hence was not vulnerable.
9755 *) BN_sqr() bug fix.
9756 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9758 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9759 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9760 followed by modular reduction.
9761 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9763 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9764 equivalent based on BN_pseudo_rand() instead of BN_rand().
9767 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9768 This function was broken, as the check for a new client hello message
9769 to handle SGC did not allow these large messages.
9770 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9773 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9776 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9777 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9780 *) Rework the configuration and shared library support for Tru64 Unix.
9781 The configuration part makes use of modern compiler features and
9782 still retains old compiler behavior for those that run older versions
9783 of the OS. The shared library support part includes a variant that
9784 uses the RPATH feature, and is available through the special
9785 configuration target "alpha-cc-rpath", which will never be selected
9787 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9789 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9790 with the same message size as in ssl3_get_certificate_request().
9791 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9792 messages might inadvertently be reject as too long.
9793 [Petr Lampa <lampa@fee.vutbr.cz>]
9795 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9798 *) Modified SSL library such that the verify_callback that has been set
9799 specifically for an SSL object with SSL_set_verify() is actually being
9800 used. Before the change, a verify_callback set with this function was
9801 ignored and the verify_callback() set in the SSL_CTX at the time of
9802 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9803 to allow the necessary settings.
9806 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9807 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9808 done automatically (in contradiction to the requirements of the C
9809 standard). This made problems when used from OpenSSH.
9812 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9813 dh->length and always used
9815 BN_rand_range(priv_key, dh->p).
9817 BN_rand_range() is not necessary for Diffie-Hellman, and this
9818 specific range makes Diffie-Hellman unnecessarily inefficient if
9819 dh->length (recommended exponent length) is much smaller than the
9820 length of dh->p. We could use BN_rand_range() if the order of
9821 the subgroup was stored in the DH structure, but we only have
9826 BN_rand(priv_key, l, ...)
9828 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9834 RSA_eay_public_encrypt
9835 RSA_eay_private_decrypt
9836 RSA_eay_private_encrypt (signing)
9837 RSA_eay_public_decrypt (signature verification)
9839 (default implementations for RSA_public_encrypt,
9840 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9841 always reject numbers >= n.
9844 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9845 to synchronize access to 'locking_thread'. This is necessary on
9846 systems where access to 'locking_thread' (an 'unsigned long'
9847 variable) is not atomic.
9850 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9851 *before* setting the 'crypto_lock_rand' flag. The previous code had
9852 a race condition if 0 is a valid thread ID.
9853 [Travis Vitek <vitek@roguewave.com>]
9855 *) Add support for shared libraries under Irix.
9856 [Albert Chin-A-Young <china@thewrittenword.com>]
9858 *) Add configuration option to build on Linux on both big-endian and
9860 [Ralf Baechle <ralf@uni-koblenz.de>]
9862 *) Add the possibility to create shared libraries on HP-UX.
9865 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9867 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9868 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9869 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9870 PRNG state recovery was possible based on the output of
9871 one PRNG request appropriately sized to gain knowledge on
9872 'md' followed by enough consecutive 1-byte PRNG requests
9873 to traverse all of 'state'.
9875 1. When updating 'md_local' (the current thread's copy of 'md')
9876 during PRNG output generation, hash all of the previous
9877 'md_local' value, not just the half used for PRNG output.
9879 2. Make the number of bytes from 'state' included into the hash
9880 independent from the number of PRNG bytes requested.
9882 The first measure alone would be sufficient to avoid
9883 Markku-Juhani's attack. (Actually it had never occurred
9884 to me that the half of 'md_local' used for chaining was the
9885 half from which PRNG output bytes were taken -- I had always
9886 assumed that the secret half would be used.) The second
9887 measure makes sure that additional data from 'state' is never
9888 mixed into 'md_local' in small portions; this heuristically
9889 further strengthens the PRNG.
9892 *) Fix crypto/bn/asm/mips3.s.
9895 *) When only the key is given to "enc", the IV is undefined. Print out
9896 an error message in this case.
9899 *) Handle special case when X509_NAME is empty in X509 printing routines.
9902 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9903 positive and less than q.
9906 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9907 used: it isn't thread safe and the add_lock_callback should handle
9909 [Paul Rose <Paul.Rose@bridge.com>]
9911 *) Verify that incoming data obeys the block size in
9912 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9916 [Ulf Möller, Bodo Möller]
9918 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9919 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9920 when fixing the server behaviour for backwards-compatible 'client
9921 hello' messages. (Note that the attack is impractical against
9922 SSL 3.0 and TLS 1.0 anyway because length and version checking
9923 means that the probability of guessing a valid ciphertext is
9924 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9927 Before 0.9.5, the countermeasure (hide the error by generating a
9928 random 'decryption result') did not work properly because
9929 ERR_clear_error() was missing, meaning that SSL_get_error() would
9930 detect the supposedly ignored error.
9932 Both problems are now fixed.
9935 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9936 (previously it was 1024).
9939 *) Fix for compatibility mode trust settings: ignore trust settings
9940 unless some valid trust or reject settings are present.
9943 *) Fix for blowfish EVP: its a variable length cipher.
9946 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9947 parameters in DSA public key structures and return an error in the
9948 DSA routines if parameters are absent.
9951 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9952 in the current directory if neither $RANDFILE nor $HOME was set.
9953 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9954 caused some confusion to Windows users who haven't defined $HOME.
9955 Thus RAND_file_name() is changed again: e_os.h can define a
9956 DEFAULT_HOME, which will be used if $HOME is not set.
9957 For Windows, we use "C:"; on other platforms, we still require
9958 environment variables.
9960 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9961 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9962 having multiple threads call RAND_poll() concurrently.
9965 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9966 combination of a flag and a thread ID variable.
9967 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9968 flag), *other* threads can enter ssleay_add_bytes without obeying
9969 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9970 that they do not hold after the first thread unsets add_do_not_lock).
9973 *) Change bctest again: '-x' expressions are not available in all
9977 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9979 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9980 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9982 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9983 the default extension for executables, if any. Also, make the perl
9984 scripts that use symlink() to test if it really exists and use "cp"
9985 if it doesn't. All this made OpenSSL compilable and installable in
9989 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9990 If SEQUENCE is length is indefinite just set c->slen to the total
9991 amount of data available.
9992 [Steve Henson, reported by shige@FreeBSD.org]
9993 [This change does not apply to 0.9.7.]
9995 *) Change bctest to avoid here-documents inside command substitution
9996 (workaround for FreeBSD /bin/sh bug).
9997 For compatibility with Ultrix, avoid shell functions (introduced
9998 in the bctest version that searches along $PATH).
10001 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
10002 with des_encrypt() defined on some operating systems, like Solaris
10006 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
10007 On the Importance of Eliminating Errors in Cryptographic
10008 Computations, J. Cryptology 14 (2001) 2, 101-119,
10009 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
10012 *) MIPS assembler BIGNUM division bug fix.
10015 *) Disabled incorrect Alpha assembler code.
10018 *) Fix PKCS#7 decode routines so they correctly update the length
10019 after reading an EOC for the EXPLICIT tag.
10021 [This change does not apply to 0.9.7.]
10023 *) Fix bug in PKCS#12 key generation routines. This was triggered
10024 if a 3DES key was generated with a 0 initial byte. Include
10025 PKCS12_BROKEN_KEYGEN compilation option to retain the old
10026 (but broken) behaviour.
10029 *) Enhance bctest to search for a working bc along $PATH and print
10031 [Tim Rice <tim@multitalents.net> via Richard Levitte]
10033 *) Fix memory leaks in err.c: free err_data string if necessary;
10034 don't write to the wrong index in ERR_set_error_data.
10037 *) Implement ssl23_peek (analogous to ssl23_read), which previously
10041 *) Replace rdtsc with _emit statements for VC++ version 5.
10042 [Jeremy Cooper <jeremy@baymoo.org>]
10044 *) Make it possible to reuse SSLv2 sessions.
10047 *) In copy_email() check for >= 0 as a return value for
10048 X509_NAME_get_index_by_NID() since 0 is a valid index.
10049 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
10051 *) Avoid coredump with unsupported or invalid public keys by checking if
10052 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
10053 PKCS7_verify() fails with non detached data.
10056 *) Don't use getenv in library functions when run as setuid/setgid.
10057 New function OPENSSL_issetugid().
10060 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
10061 due to incorrect handling of multi-threading:
10063 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
10065 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
10067 3. Count how many times MemCheck_off() has been called so that
10068 nested use can be treated correctly. This also avoids
10069 inband-signalling in the previous code (which relied on the
10070 assumption that thread ID 0 is impossible).
10073 *) Add "-rand" option also to s_client and s_server.
10076 *) Fix CPU detection on Irix 6.x.
10077 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
10078 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10080 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
10083 [This change does not apply to 0.9.7.]
10085 *) Use the cached encoding of an X509_NAME structure rather than
10086 copying it. This is apparently the reason for the libsafe "errors"
10087 but the code is actually correct.
10090 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
10091 Bleichenbacher's DSA attack.
10092 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
10093 to be set and top=0 forces the highest bit to be set; top=-1 is new
10094 and leaves the highest bit random.
10095 [Ulf Moeller, Bodo Moeller]
10097 *) In the NCONF_...-based implementations for CONF_... queries
10098 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
10099 a temporary CONF structure with the data component set to NULL
10100 (which gives segmentation faults in lh_retrieve).
10101 Instead, use NULL for the CONF pointer in CONF_get_string and
10102 CONF_get_number (which may use environment variables) and directly
10103 return NULL from CONF_get_section.
10106 *) Fix potential buffer overrun for EBCDIC.
10109 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
10110 keyUsage if basicConstraints absent for a CA.
10113 *) Make SMIME_write_PKCS7() write mail header values with a format that
10114 is more generally accepted (no spaces before the semicolon), since
10115 some programs can't parse those values properly otherwise. Also make
10116 sure BIO's that break lines after each write do not create invalid
10120 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
10121 macros previously used would not encode an empty SEQUENCE OF
10122 and break the signature.
10124 [This change does not apply to 0.9.7.]
10126 *) Zero the premaster secret after deriving the master secret in
10130 *) Add some EVP_add_digest_alias registrations (as found in
10131 OpenSSL_add_all_digests()) to SSL_library_init()
10132 aka OpenSSL_add_ssl_algorithms(). This provides improved
10133 compatibility with peers using X.509 certificates
10134 with unconventional AlgorithmIdentifier OIDs.
10137 *) Fix for Irix with NO_ASM.
10138 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10140 *) ./config script fixes.
10141 [Ulf Moeller, Richard Levitte]
10143 *) Fix 'openssl passwd -1'.
10146 *) Change PKCS12_key_gen_asc() so it can cope with non null
10147 terminated strings whose length is passed in the passlen
10148 parameter, for example from PEM callbacks. This was done
10149 by adding an extra length parameter to asc2uni().
10150 [Steve Henson, reported by <oddissey@samsung.co.kr>]
10152 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
10153 call failed, free the DSA structure.
10156 *) Fix to uni2asc() to cope with zero length Unicode strings.
10157 These are present in some PKCS#12 files.
10160 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
10161 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
10162 when writing a 32767 byte record.
10163 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
10165 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
10166 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
10168 (RSA objects have a reference count access to which is protected
10169 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
10170 so they are meant to be shared between threads.)
10171 [Bodo Moeller, Geoff Thorpe; original patch submitted by
10172 "Reddie, Steven" <Steven.Reddie@ca.com>]
10174 *) Fix a deadlock in CRYPTO_mem_leaks().
10177 *) Use better test patterns in bntest.
10180 *) rand_win.c fix for Borland C.
10183 *) BN_rshift bugfix for n == 0.
10186 *) Add a 'bctest' script that checks for some known 'bc' bugs
10187 so that 'make test' does not abort just because 'bc' is broken.
10190 *) Store verify_result within SSL_SESSION also for client side to
10191 avoid potential security hole. (Re-used sessions on the client side
10192 always resulted in verify_result==X509_V_OK, not using the original
10193 result of the server certificate verification.)
10196 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
10197 SSL3_RT_APPLICATION_DATA, return 0.
10198 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
10202 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
10203 releases, have been re-implemented by renaming the previous
10204 implementations of ssl2_read and ssl3_read to ssl2_read_internal
10205 and ssl3_read_internal, respectively, and adding 'peek' parameters
10206 to them. The new ssl[23]_{read,peek} functions are calls to
10207 ssl[23]_read_internal with the 'peek' flag set appropriately.
10208 A 'peek' parameter has also been added to ssl3_read_bytes, which
10209 does the actual work for ssl3_read_internal.
10212 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
10213 the method-specific "init()" handler. Also clean up ex_data after
10214 calling the method-specific "finish()" handler. Previously, this was
10215 happening the other way round.
10218 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
10219 The previous value, 12, was not always sufficient for BN_mod_exp().
10222 *) Make sure that shared libraries get the internal name engine with
10223 the full version number and not just 0. This should mark the
10224 shared libraries as not backward compatible. Of course, this should
10225 be changed again when we can guarantee backward binary compatibility.
10228 *) Fix typo in get_cert_by_subject() in by_dir.c
10229 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
10231 *) Rework the system to generate shared libraries:
10233 - Make note of the expected extension for the shared libraries and
10234 if there is a need for symbolic links from for example libcrypto.so.0
10235 to libcrypto.so.0.9.7. There is extended info in Configure for
10238 - Make as few rebuilds of the shared libraries as possible.
10240 - Still avoid linking the OpenSSL programs with the shared libraries.
10242 - When installing, install the shared libraries separately from the
10246 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
10248 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
10249 and not in SSL_clear because the latter is also used by the
10250 accept/connect functions; previously, the settings made by
10251 SSL_set_read_ahead would be lost during the handshake.
10252 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
10254 *) Correct util/mkdef.pl to be selective about disabled algorithms.
10255 Previously, it would create entries for disabled algorithms no
10259 *) Added several new manual pages for SSL_* function.
10262 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
10264 *) In ssl23_get_client_hello, generate an error message when faced
10265 with an initial SSL 3.0/TLS record that is too small to contain the
10266 first two bytes of the ClientHello message, i.e. client_version.
10267 (Note that this is a pathologic case that probably has never happened
10268 in real life.) The previous approach was to use the version number
10269 from the record header as a substitute; but our protocol choice
10270 should not depend on that one because it is not authenticated
10271 by the Finished messages.
10274 *) More robust randomness gathering functions for Windows.
10275 [Jeffrey Altman <jaltman@columbia.edu>]
10277 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
10278 not set then we don't setup the error code for issuer check errors
10279 to avoid possibly overwriting other errors which the callback does
10280 handle. If an application does set the flag then we assume it knows
10281 what it is doing and can handle the new informational codes
10285 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
10286 a general "ANY" type, as such it should be able to decode anything
10287 including tagged types. However it didn't check the class so it would
10288 wrongly interpret tagged types in the same way as their universal
10289 counterpart and unknown types were just rejected. Changed so that the
10290 tagged and unknown types are handled in the same way as a SEQUENCE:
10291 that is the encoding is stored intact. There is also a new type
10292 "V_ASN1_OTHER" which is used when the class is not universal, in this
10293 case we have no idea what the actual type is so we just lump them all
10297 *) On VMS, stdout may very well lead to a file that is written to
10298 in a record-oriented fashion. That means that every write() will
10299 write a separate record, which will be read separately by the
10300 programs trying to read from it. This can be very confusing.
10302 The solution is to put a BIO filter in the way that will buffer
10303 text until a linefeed is reached, and then write everything a
10304 line at a time, so every record written will be an actual line,
10305 not chunks of lines and not (usually doesn't happen, but I've
10306 seen it once) several lines in one record. BIO_f_linebuffer() is
10309 Currently, it's a VMS-only method, because that's where it has
10310 been tested well enough.
10313 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
10314 it can return incorrect results.
10315 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
10316 but it was in 0.9.6-beta[12].)
10319 *) Disable the check for content being present when verifying detached
10320 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
10321 include zero length content when signing messages.
10324 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
10325 BIO_ctrl (for BIO pairs).
10328 *) Add DSO method for VMS.
10331 *) Bug fix: Montgomery multiplication could produce results with the
10335 *) Add RPM specification openssl.spec and modify it to build three
10336 packages. The default package contains applications, application
10337 documentation and run-time libraries. The devel package contains
10338 include files, static libraries and function documentation. The
10339 doc package contains the contents of the doc directory. The original
10340 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
10343 *) Add a large number of documentation files for many SSL routines.
10344 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10346 *) Add a configuration entry for Sony News 4.
10347 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
10349 *) Don't set the two most significant bits to one when generating a
10350 random number < q in the DSA library.
10353 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
10354 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
10355 the underlying transport is blocking) if a handshake took place.
10356 (The default behaviour is needed by applications such as s_client
10357 and s_server that use select() to determine when to use SSL_read;
10358 but for applications that know in advance when to expect data, it
10359 just makes things more complicated.)
10362 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
10366 *) Add a few more EBCDIC conditionals that make `req' and `x509'
10367 work better on such systems.
10368 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10370 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10371 Update PKCS12_parse() so it copies the friendlyName and the
10372 keyid to the certificates aux info.
10375 *) Fix bug in PKCS7_verify() which caused an infinite loop
10376 if there was more than one signature.
10377 [Sven Uszpelkat <su@celocom.de>]
10379 *) Major change in util/mkdef.pl to include extra information
10380 about each symbol, as well as presenting variables as well
10381 as functions. This change means that there's n more need
10382 to rebuild the .num files when some algorithms are excluded.
10385 *) Allow the verify time to be set by an application,
10386 rather than always using the current time.
10389 *) Phase 2 verify code reorganisation. The certificate
10390 verify code now looks up an issuer certificate by a
10391 number of criteria: subject name, authority key id
10392 and key usage. It also verifies self signed certificates
10393 by the same criteria. The main comparison function is
10394 X509_check_issued() which performs these checks.
10396 Lot of changes were necessary in order to support this
10397 without completely rewriting the lookup code.
10399 Authority and subject key identifier are now cached.
10401 The LHASH 'certs' is X509_STORE has now been replaced
10402 by a STACK_OF(X509_OBJECT). This is mainly because an
10403 LHASH can't store or retrieve multiple objects with
10404 the same hash value.
10406 As a result various functions (which were all internal
10407 use only) have changed to handle the new X509_STORE
10408 structure. This will break anything that messed round
10409 with X509_STORE internally.
10411 The functions X509_STORE_add_cert() now checks for an
10412 exact match, rather than just subject name.
10414 The X509_STORE API doesn't directly support the retrieval
10415 of multiple certificates matching a given criteria, however
10416 this can be worked round by performing a lookup first
10417 (which will fill the cache with candidate certificates)
10418 and then examining the cache for matches. This is probably
10419 the best we can do without throwing out X509_LOOKUP
10420 entirely (maybe later...).
10422 The X509_VERIFY_CTX structure has been enhanced considerably.
10424 All certificate lookup operations now go via a get_issuer()
10425 callback. Although this currently uses an X509_STORE it
10426 can be replaced by custom lookups. This is a simple way
10427 to bypass the X509_STORE hackery necessary to make this
10428 work and makes it possible to use more efficient techniques
10429 in future. A very simple version which uses a simple
10430 STACK for its trusted certificate store is also provided
10431 using X509_STORE_CTX_trusted_stack().
10433 The verify_cb() and verify() callbacks now have equivalents
10434 in the X509_STORE_CTX structure.
10436 X509_STORE_CTX also has a 'flags' field which can be used
10437 to customise the verify behaviour.
10440 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10441 excludes S/MIME capabilities.
10444 *) When a certificate request is read in keep a copy of the
10445 original encoding of the signed data and use it when outputting
10446 again. Signatures then use the original encoding rather than
10447 a decoded, encoded version which may cause problems if the
10448 request is improperly encoded.
10451 *) For consistency with other BIO_puts implementations, call
10452 buffer_write(b, ...) directly in buffer_puts instead of calling
10455 In BIO_puts, increment b->num_write as in BIO_write.
10456 [Peter.Sylvester@EdelWeb.fr]
10458 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10459 BN_zero, we may not return a BIGNUM with an array consisting of
10460 words set to zero.)
10463 *) Avoid calling abort() from within the library when problems are
10464 detected, except if preprocessor symbols have been defined
10465 (such as REF_CHECK, BN_DEBUG etc.).
10468 *) New openssl application 'rsautl'. This utility can be
10469 used for low level RSA operations. DER public key
10470 BIO/fp routines also added.
10473 *) New Configure entry and patches for compiling on QNX 4.
10474 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10476 *) A demo state-machine implementation was sponsored by
10477 Nuron (http://www.nuron.com/) and is now available in
10478 demos/state_machine.
10481 *) New options added to the 'dgst' utility for signature
10482 generation and verification.
10485 *) Unrecognized PKCS#7 content types are now handled via a
10486 catch all ASN1_TYPE structure. This allows unsupported
10487 types to be stored as a "blob" and an application can
10488 encode and decode it manually.
10491 *) Fix various signed/unsigned issues to make a_strex.c
10492 compile under VC++.
10493 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10495 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10496 length if passed a buffer. ASN1_INTEGER_to_BN failed
10497 if passed a NULL BN and its argument was negative.
10498 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10500 *) Modification to PKCS#7 encoding routines to output definite
10501 length encoding. Since currently the whole structures are in
10502 memory there's not real point in using indefinite length
10503 constructed encoding. However if OpenSSL is compiled with
10504 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10507 *) Added BIO_vprintf() and BIO_vsnprintf().
10510 *) Added more prefixes to parse for in the strings written
10511 through a logging bio, to cover all the levels that are available
10512 through syslog. The prefixes are now:
10514 PANIC, EMERG, EMR => LOG_EMERG
10515 ALERT, ALR => LOG_ALERT
10516 CRIT, CRI => LOG_CRIT
10517 ERROR, ERR => LOG_ERR
10518 WARNING, WARN, WAR => LOG_WARNING
10519 NOTICE, NOTE, NOT => LOG_NOTICE
10520 INFO, INF => LOG_INFO
10521 DEBUG, DBG => LOG_DEBUG
10523 and as before, if none of those prefixes are present at the
10524 beginning of the string, LOG_ERR is chosen.
10526 On Win32, the LOG_* levels are mapped according to this:
10528 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10529 LOG_WARNING => EVENTLOG_WARNING_TYPE
10530 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10534 *) Made it possible to reconfigure with just the configuration
10535 argument "reconf" or "reconfigure". The command line arguments
10536 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10537 and are retrieved from there when reconfiguring.
10540 *) MD4 implemented.
10541 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10543 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10546 *) The obj_dat.pl script was messing up the sorting of object
10547 names. The reason was that it compared the quoted version
10548 of strings as a result "OCSP" > "OCSP Signing" because
10549 " > SPACE. Changed script to store unquoted versions of
10550 names and add quotes on output. It was also omitting some
10551 names from the lookup table if they were given a default
10552 value (that is if SN is missing it is given the same
10553 value as LN and vice versa), these are now added on the
10554 grounds that if an object has a name we should be able to
10555 look it up. Finally added warning output when duplicate
10556 short or long names are found.
10559 *) Changes needed for Tandem NSK.
10560 [Scott Uroff <scott@xypro.com>]
10562 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10563 RSA_padding_check_SSLv23(), special padding was never detected
10564 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10565 version rollback attacks was not effective.
10567 In s23_clnt.c, don't use special rollback-attack detection padding
10568 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10569 client; similarly, in s23_srvr.c, don't do the rollback check if
10570 SSL 2.0 is the only protocol enabled in the server.
10573 *) Make it possible to get hexdumps of unprintable data with 'openssl
10574 asn1parse'. By implication, the functions ASN1_parse_dump() and
10575 BIO_dump_indent() are added.
10578 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10579 these print out strings and name structures based on various
10580 flags including RFC2253 support and proper handling of
10581 multibyte characters. Added options to the 'x509' utility
10582 to allow the various flags to be set.
10585 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10586 Also change the functions X509_cmp_current_time() and
10587 X509_gmtime_adj() work with an ASN1_TIME structure,
10588 this will enable certificates using GeneralizedTime in validity
10589 dates to be checked.
10592 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10593 negative public key encodings) on by default,
10594 NO_NEG_PUBKEY_BUG can be set to disable it.
10597 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10598 content octets. An i2c_ASN1_OBJECT is unnecessary because
10599 the encoding can be trivially obtained from the structure.
10602 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10603 not read locks (CRYPTO_r_[un]lock).
10606 *) A first attempt at creating official support for shared
10607 libraries through configuration. I've kept it so the
10608 default is static libraries only, and the OpenSSL programs
10609 are always statically linked for now, but there are
10610 preparations for dynamic linking in place.
10611 This has been tested on Linux and Tru64.
10614 *) Randomness polling function for Win9x, as described in:
10615 Peter Gutmann, Software Generation of Practically Strong
10619 *) Fix so PRNG is seeded in req if using an already existing
10623 *) New options to smime application. -inform and -outform
10624 allow alternative formats for the S/MIME message including
10625 PEM and DER. The -content option allows the content to be
10626 specified separately. This should allow things like Netscape
10627 form signing output easier to verify.
10630 *) Fix the ASN1 encoding of tags using the 'long form'.
10633 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10634 STRING types. These convert content octets to and from the
10635 underlying type. The actual tag and length octets are
10636 already assumed to have been read in and checked. These
10637 are needed because all other string types have virtually
10638 identical handling apart from the tag. By having versions
10639 of the ASN1 functions that just operate on content octets
10640 IMPLICIT tagging can be handled properly. It also allows
10641 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10642 and ASN1_INTEGER are identical apart from the tag.
10645 *) Change the handling of OID objects as follows:
10647 - New object identifiers are inserted in objects.txt, following
10648 the syntax given in objects.README.
10649 - objects.pl is used to process obj_mac.num and create a new
10651 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10654 This is currently kind of a hack, and the perl code in objects.pl
10655 isn't very elegant, but it works as I intended. The simplest way
10656 to check that it worked correctly is to look in obj_dat.h and
10657 check the array nid_objs and make sure the objects haven't moved
10658 around (this is important!). Additions are OK, as well as
10659 consistent name changes.
10662 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10665 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10666 The given file adds to whatever has already been seeded into the
10667 random pool through the RANDFILE configuration file option or
10668 environment variable, or the default random state file.
10671 *) mkstack.pl now sorts each macro group into lexical order.
10672 Previously the output order depended on the order the files
10673 appeared in the directory, resulting in needless rewriting
10677 *) Patches to make OpenSSL compile under Win32 again. Mostly
10678 work arounds for the VC++ problem that it treats func() as
10679 func(void). Also stripped out the parts of mkdef.pl that
10680 added extra typesafe functions: these no longer exist.
10683 *) Reorganisation of the stack code. The macros are now all
10684 collected in safestack.h . Each macro is defined in terms of
10685 a "stack macro" of the form SKM_<name>(type, a, b). The
10686 DEBUG_SAFESTACK is now handled in terms of function casts,
10687 this has the advantage of retaining type safety without the
10688 use of additional functions. If DEBUG_SAFESTACK is not defined
10689 then the non typesafe macros are used instead. Also modified the
10690 mkstack.pl script to handle the new form. Needs testing to see
10691 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10692 the default if no major problems. Similar behaviour for ASN1_SET_OF
10693 and PKCS12_STACK_OF.
10696 *) When some versions of IIS use the 'NET' form of private key the
10697 key derivation algorithm is different. Normally MD5(password) is
10698 used as a 128 bit RC4 key. In the modified case
10699 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10700 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10701 as the old Netscape_RSA functions except they have an additional
10702 'sgckey' parameter which uses the modified algorithm. Also added
10703 an -sgckey command line option to the rsa utility. Thanks to
10704 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10705 algorithm to openssl-dev.
10708 *) The evp_local.h macros were using 'c.##kname' which resulted in
10709 invalid expansion on some systems (SCO 5.0.5 for example).
10710 Corrected to 'c.kname'.
10711 [Phillip Porch <root@theporch.com>]
10713 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10714 a STACK of email addresses from a certificate or request, these look
10715 in the subject name and the subject alternative name extensions and
10716 omit any duplicate addresses.
10719 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10720 This makes DSA verification about 2 % faster.
10723 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10724 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10725 plus overhead for 1024 bit moduli).
10726 This makes exponentiations about 0.5 % faster for 1024 bit
10727 exponents (as measured by "openssl speed rsa2048").
10730 *) Rename memory handling macros to avoid conflicts with other
10732 Malloc => OPENSSL_malloc
10733 Malloc_locked => OPENSSL_malloc_locked
10734 Realloc => OPENSSL_realloc
10735 Free => OPENSSL_free
10738 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10739 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10742 *) CygWin32 support.
10743 [John Jarvie <jjarvie@newsguy.com>]
10745 *) The type-safe stack code has been rejigged. It is now only compiled
10746 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10747 by default all type-specific stack functions are "#define"d back to
10748 standard stack functions. This results in more streamlined output
10749 but retains the type-safety checking possibilities of the original
10753 *) The STACK code has been cleaned up, and certain type declarations
10754 that didn't make a lot of sense have been brought in line. This has
10755 also involved a cleanup of sorts in safestack.h to more correctly
10756 map type-safe stack functions onto their plain stack counterparts.
10757 This work has also resulted in a variety of "const"ifications of
10758 lots of the code, especially "_cmp" operations which should normally
10759 be prototyped with "const" parameters anyway.
10762 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10763 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10764 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10765 where all of 'md' is used each time the PRNG is used, but 'state'
10766 is used only indexed by a cyclic counter. As entropy may not be
10767 well distributed from the beginning, 'md' is important as a
10768 chaining variable. However, the output function chains only half
10769 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10770 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10771 in all of 'state' being rewritten, with the new values depending
10772 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10775 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10776 the handshake is continued after ssl_verify_cert_chain();
10777 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10778 can lead to 'unexplainable' connection aborts later.
10779 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10781 *) Major EVP API cipher revision.
10782 Add hooks for extra EVP features. This allows various cipher
10783 parameters to be set in the EVP interface. Support added for variable
10784 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10785 setting of RC2 and RC5 parameters.
10787 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10790 Remove lots of duplicated code from the EVP library. For example *every*
10791 cipher init() function handles the 'iv' in the same way according to the
10792 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10793 for CFB and OFB modes they zero ctx->num.
10795 New functionality allows removal of S/MIME code RC2 hack.
10797 Most of the routines have the same form and so can be declared in terms
10800 By shifting this to the top level EVP_CipherInit() it can be removed from
10801 all individual ciphers. If the cipher wants to handle IVs or keys
10802 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10805 Change lots of functions like EVP_EncryptUpdate() to now return a
10806 value: although software versions of the algorithms cannot fail
10807 any installed hardware versions can.
10810 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10811 this option is set, tolerate broken clients that send the negotiated
10812 protocol version number instead of the requested protocol version
10816 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10817 i.e. non-zero for export ciphersuites, zero otherwise.
10818 Previous versions had this flag inverted, inconsistent with
10819 rsa_tmp_cb (..._TMP_RSA_CB).
10820 [Bodo Moeller; problem reported by Amit Chopra]
10822 *) Add missing DSA library text string. Work around for some IIS
10823 key files with invalid SEQUENCE encoding.
10826 *) Add a document (doc/standards.txt) that list all kinds of standards
10827 and so on that are implemented in OpenSSL.
10830 *) Enhance c_rehash script. Old version would mishandle certificates
10831 with the same subject name hash and wouldn't handle CRLs at all.
10832 Added -fingerprint option to crl utility, to support new c_rehash
10836 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10839 *) Fix for SSL server purpose checking. Server checking was
10840 rejecting certificates which had extended key usage present
10841 but no ssl client purpose.
10842 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10844 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10845 is a little unclear about how a blank password is handled.
10846 Since the password in encoded as a BMPString with terminating
10847 double NULL a zero length password would end up as just the
10848 double NULL. However no password at all is different and is
10849 handled differently in the PKCS#12 key generation code. NS
10850 treats a blank password as zero length. MSIE treats it as no
10851 password on export: but it will try both on import. We now do
10852 the same: PKCS12_parse() tries zero length and no password if
10853 the password is set to "" or NULL (NULL is now a valid password:
10854 it wasn't before) as does the pkcs12 application.
10857 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10858 perror when PEM_read_bio_X509_REQ fails, the error message must
10859 be obtained from the error queue.
10862 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10863 it in ERR_remove_state if appropriate, and change ERR_get_state
10864 accordingly to avoid race conditions (this is necessary because
10865 thread_hash is no longer constant once set).
10868 *) Bugfix for linux-elf makefile.one.
10871 *) RSA_get_default_method() will now cause a default
10872 RSA_METHOD to be chosen if one doesn't exist already.
10873 Previously this was only set during a call to RSA_new()
10874 or RSA_new_method(NULL) meaning it was possible for
10875 RSA_get_default_method() to return NULL.
10878 *) Added native name translation to the existing DSO code
10879 that will convert (if the flag to do so is set) filenames
10880 that are sufficiently small and have no path information
10881 into a canonical native form. Eg. "blah" converted to
10882 "libblah.so" or "blah.dll" etc.
10885 *) New function ERR_error_string_n(e, buf, len) which is like
10886 ERR_error_string(e, buf), but writes at most 'len' bytes
10887 including the 0 terminator. For ERR_error_string_n, 'buf'
10889 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10891 *) CONF library reworked to become more general. A new CONF
10892 configuration file reader "class" is implemented as well as a
10893 new functions (NCONF_*, for "New CONF") to handle it. The now
10894 old CONF_* functions are still there, but are reimplemented to
10895 work in terms of the new functions. Also, a set of functions
10896 to handle the internal storage of the configuration data is
10897 provided to make it easier to write new configuration file
10898 reader "classes" (I can definitely see something reading a
10899 configuration file in XML format, for example), called _CONF_*,
10900 or "the configuration storage API"...
10902 The new configuration file reading functions are:
10904 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10905 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10907 NCONF_default, NCONF_WIN32
10909 NCONF_dump_fp, NCONF_dump_bio
10911 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10912 NCONF_new creates a new CONF object. This works in the same way
10913 as other interfaces in OpenSSL, like the BIO interface.
10914 NCONF_dump_* dump the internal storage of the configuration file,
10915 which is useful for debugging. All other functions take the same
10916 arguments as the old CONF_* functions with the exception of the
10917 first that must be a `CONF *' instead of a `LHASH *'.
10919 To make it easier to use the new classes with the old CONF_* functions,
10920 the function CONF_set_default_method is provided.
10923 *) Add '-tls1' option to 'openssl ciphers', which was already
10924 mentioned in the documentation but had not been implemented.
10925 (This option is not yet really useful because even the additional
10926 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10929 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10930 OpenSSL-based applications) load shared libraries and bind to
10931 them in a portable way.
10932 [Geoff Thorpe, with contributions from Richard Levitte]
10934 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10936 *) Make sure _lrotl and _lrotr are only used with MSVC.
10938 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10939 (the default implementation of RAND_status).
10941 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10942 to '-clrext' (= clear extensions), as intended and documented.
10943 [Bodo Moeller; inconsistency pointed out by Michael Attili
10944 <attili@amaxo.com>]
10946 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10947 was larger than the MD block size.
10948 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10950 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10951 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10952 using the passed key: if the passed key was a private key the result
10953 of X509_print(), for example, would be to print out all the private key
10957 *) des_quad_cksum() byte order bug fix.
10958 [Ulf Möller, using the problem description in krb4-0.9.7, where
10959 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10961 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10963 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10965 *) For easily testing in shell scripts whether some command
10966 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10967 returns with exit code 0 iff no command of the given name is available.
10968 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10969 the output goes to stdout and nothing is printed to stderr.
10970 Additional arguments are always ignored.
10972 Since for each cipher there is a command of the same name,
10973 the 'no-cipher' compilation switches can be tested this way.
10975 ('openssl no-XXX' is not able to detect pseudo-commands such
10976 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10979 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10982 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10983 is set; it will be thrown away anyway because each handshake creates
10985 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10986 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10987 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10988 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10991 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10992 'Q' and 'R' lose their special meanings (quit/renegotiate).
10993 This is part of what -quiet does; unlike -quiet, -ign_eof
10994 does not suppress any output.
10997 *) Add compatibility options to the purpose and trust code. The
10998 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10999 accepts a certificate or CA, this was the previous behaviour,
11000 with all the associated security issues.
11002 X509_TRUST_COMPAT is the old trust behaviour: only and
11003 automatically trust self signed roots in certificate store. A
11004 new trust setting X509_TRUST_DEFAULT is used to specify that
11005 a purpose has no associated trust setting and it should instead
11006 use the value in the default purpose.
11009 *) Fix the PKCS#8 DSA private key code so it decodes keys again
11010 and fix a memory leak.
11013 *) In util/mkerr.pl (which implements 'make errors'), preserve
11014 reason strings from the previous version of the .c file, as
11015 the default to have only downcase letters (and digits) in
11016 automatically generated reasons codes is not always appropriate.
11019 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
11020 using strerror. Previously, ERR_reason_error_string() returned
11021 library names as reason strings for SYSerr; but SYSerr is a special
11022 case where small numbers are errno values, not library numbers.
11025 *) Add '-dsaparam' option to 'openssl dhparam' application. This
11026 converts DSA parameters into DH parameters. (When creating parameters,
11027 DSA_generate_parameters is used.)
11030 *) Include 'length' (recommended exponent length) in C code generated
11031 by 'openssl dhparam -C'.
11034 *) The second argument to set_label in perlasm was already being used
11035 so couldn't be used as a "file scope" flag. Moved to third argument
11039 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
11040 instead of RAND_bytes for encryption IVs and salts.
11043 *) Include RAND_status() into RAND_METHOD instead of implementing
11044 it only for md_rand.c Otherwise replacing the PRNG by calling
11045 RAND_set_rand_method would be impossible.
11048 *) Don't let DSA_generate_key() enter an infinite loop if the random
11049 number generation fails.
11052 *) New 'rand' application for creating pseudo-random output.
11055 *) Added configuration support for Linux/IA64
11056 [Rolf Haberrecker <rolf@suse.de>]
11058 *) Assembler module support for Mingw32.
11061 *) Shared library support for HPUX (in shlib/).
11062 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
11064 *) Shared library support for Solaris gcc.
11065 [Lutz Behnke <behnke@trustcenter.de>]
11067 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
11069 *) PKCS7_encrypt() was adding text MIME headers twice because they
11070 were added manually and by SMIME_crlf_copy().
11073 *) In bntest.c don't call BN_rand with zero bits argument.
11074 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
11076 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
11077 case was implemented. This caused BN_div_recp() to fail occasionally.
11080 *) Add an optional second argument to the set_label() in the perl
11081 assembly language builder. If this argument exists and is set
11082 to 1 it signals that the assembler should use a symbol whose
11083 scope is the entire file, not just the current function. This
11084 is needed with MASM which uses the format label:: for this scope.
11085 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
11087 *) Change the ASN1 types so they are typedefs by default. Before
11088 almost all types were #define'd to ASN1_STRING which was causing
11089 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
11093 *) Change names of new functions to the new get1/get0 naming
11094 convention: After 'get1', the caller owns a reference count
11095 and has to call ..._free; 'get0' returns a pointer to some
11096 data structure without incrementing reference counters.
11097 (Some of the existing 'get' functions increment a reference
11098 counter, some don't.)
11099 Similarly, 'set1' and 'add1' functions increase reference
11100 counters or duplicate objects.
11103 *) Allow for the possibility of temp RSA key generation failure:
11104 the code used to assume it always worked and crashed on failure.
11107 *) Fix potential buffer overrun problem in BIO_printf().
11108 [Ulf Möller, using public domain code by Patrick Powell; problem
11109 pointed out by David Sacerdote <das33@cornell.edu>]
11111 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
11112 RAND_egd() and RAND_status(). In the command line application,
11113 the EGD socket can be specified like a seed file using RANDFILE
11117 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
11118 Some CAs (e.g. Verisign) distribute certificates in this form.
11121 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
11122 list to exclude them. This means that no special compilation option
11123 is needed to use anonymous DH: it just needs to be included in the
11127 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
11128 EVP_MD_type. The old functionality is available in a new macro called
11129 EVP_MD_md(). Change code that uses it and update docs.
11132 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
11133 where the 'void *' argument is replaced by a function pointer argument.
11134 Previously 'void *' was abused to point to functions, which works on
11135 many platforms, but is not correct. As these functions are usually
11136 called by macros defined in OpenSSL header files, most source code
11137 should work without changes.
11140 *) <openssl/opensslconf.h> (which is created by Configure) now contains
11141 sections with information on -D... compiler switches used for
11142 compiling the library so that applications can see them. To enable
11143 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
11144 must be defined. E.g.,
11145 #define OPENSSL_ALGORITHM_DEFINES
11146 #include <openssl/opensslconf.h>
11147 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
11148 [Richard Levitte, Ulf and Bodo Möller]
11150 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
11154 *) Change the 'other' type in certificate aux info to a STACK_OF
11155 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
11156 the required ASN1 format: arbitrary types determined by an OID.
11159 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
11160 argument to 'req'. This is not because the function is newer or
11161 better than others it just uses the work 'NEW' in the certificate
11162 request header lines. Some software needs this.
11165 *) Reorganise password command line arguments: now passwords can be
11166 obtained from various sources. Delete the PEM_cb function and make
11167 it the default behaviour: i.e. if the callback is NULL and the
11168 usrdata argument is not NULL interpret it as a null terminated pass
11169 phrase. If usrdata and the callback are NULL then the pass phrase
11170 is prompted for as usual.
11173 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
11174 the support is automatically enabled. The resulting binaries will
11175 autodetect the card and use it if present.
11176 [Ben Laurie and Compaq Inc.]
11178 *) Work around for Netscape hang bug. This sends certificate request
11179 and server done in one record. Since this is perfectly legal in the
11180 SSL/TLS protocol it isn't a "bug" option and is on by default. See
11181 the bugs/SSLv3 entry for more info.
11184 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
11187 *) Add -rand argument to smime and pkcs12 applications and read/write
11191 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
11194 *) Add command line password options to the remaining applications.
11197 *) Bug fix for BN_div_recp() for numerators with an even number of
11201 *) More tests in bntest.c, and changed test_bn output.
11204 *) ./config recognizes MacOS X now.
11207 *) Bug fix for BN_div() when the first words of num and divisor are
11208 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
11211 *) Add support for various broken PKCS#8 formats, and command line
11212 options to produce them.
11215 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
11216 get temporary BIGNUMs from a BN_CTX.
11219 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
11223 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
11224 include a #define from the old name to the new. The original intent
11225 was that statically linked binaries could for example just call
11226 SSLeay_add_all_ciphers() to just add ciphers to the table and not
11227 link with digests. This never worked because SSLeay_add_all_digests()
11228 and SSLeay_add_all_ciphers() were in the same source file so calling
11229 one would link with the other. They are now in separate source files.
11232 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
11235 *) Use a less unusual form of the Miller-Rabin primality test (it used
11236 a binary algorithm for exponentiation integrated into the Miller-Rabin
11237 loop, our standard modexp algorithms are faster).
11240 *) Support for the EBCDIC character set completed.
11241 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
11243 *) Source code cleanups: use const where appropriate, eliminate casts,
11244 use void * instead of char * in lhash.
11247 *) Bugfix: ssl3_send_server_key_exchange was not restartable
11248 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
11249 this the server could overwrite ephemeral keys that the client
11253 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
11254 using 50 iterations of the Rabin-Miller test.
11256 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
11257 iterations of the Rabin-Miller test as required by the appendix
11258 to FIPS PUB 186[-1]) instead of DSA_is_prime.
11259 As BN_is_prime_fasttest includes trial division, DSA parameter
11260 generation becomes much faster.
11262 This implies a change for the callback functions in DSA_is_prime
11263 and DSA_generate_parameters: The callback function is called once
11264 for each positive witness in the Rabin-Miller test, not just
11265 occasionally in the inner loop; and the parameters to the
11266 callback function now provide an iteration count for the outer
11267 loop rather than for the current invocation of the inner loop.
11268 DSA_generate_parameters additionally can call the callback
11269 function with an 'iteration count' of -1, meaning that a
11270 candidate has passed the trial division test (when q is generated
11271 from an application-provided seed, trial division is skipped).
11274 *) New function BN_is_prime_fasttest that optionally does trial
11275 division before starting the Rabin-Miller test and has
11276 an additional BN_CTX * argument (whereas BN_is_prime always
11277 has to allocate at least one BN_CTX).
11278 'callback(1, -1, cb_arg)' is called when a number has passed the
11279 trial division stage.
11282 *) Fix for bug in CRL encoding. The validity dates weren't being handled
11286 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
11289 *) New function BN_pseudo_rand().
11292 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
11293 bignum version of BN_from_montgomery() with the working code from
11294 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
11298 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
11299 made it impossible to use the same SSL_SESSION data structure in
11300 SSL2 clients in multiple threads.
11303 *) The return value of RAND_load_file() no longer counts bytes obtained
11304 by stat(). RAND_load_file(..., -1) is new and uses the complete file
11305 to seed the PRNG (previously an explicit byte count was required).
11306 [Ulf Möller, Bodo Möller]
11308 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
11309 used (char *) instead of (void *) and had casts all over the place.
11312 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
11315 *) Retain source code compatibility for BN_prime_checks macro:
11316 BN_is_prime(..., BN_prime_checks, ...) now uses
11317 BN_prime_checks_for_size to determine the appropriate number of
11318 Rabin-Miller iterations.
11321 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
11322 DH_CHECK_P_NOT_SAFE_PRIME.
11323 (Check if this is true? OpenPGP calls them "strong".)
11326 *) Merge the functionality of "dh" and "gendh" programs into a new program
11327 "dhparam". The old programs are retained for now but will handle DH keys
11328 (instead of parameters) in future.
11331 *) Make the ciphers, s_server and s_client programs check the return values
11332 when a new cipher list is set.
11335 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
11336 ciphers. Before when the 56bit ciphers were enabled the sorting was
11339 The syntax for the cipher sorting has been extended to support sorting by
11340 cipher-strength (using the strength_bits hard coded in the tables).
11341 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
11343 Fix a bug in the cipher-command parser: when supplying a cipher command
11344 string with an "undefined" symbol (neither command nor alphanumeric
11345 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
11346 an error is flagged.
11348 Due to the strength-sorting extension, the code of the
11349 ssl_create_cipher_list() function was completely rearranged. I hope that
11350 the readability was also increased :-)
11351 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
11353 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
11354 for the first serial number and places 2 in the serial number file. This
11355 avoids problems when the root CA is created with serial number zero and
11356 the first user certificate has the same issuer name and serial number
11360 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
11361 the new code. Add documentation for this stuff.
11364 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
11365 X509_*() to X509at_*() on the grounds that they don't handle X509
11366 structures and behave in an analogous way to the X509v3 functions:
11367 they shouldn't be called directly but wrapper functions should be used
11370 So we also now have some wrapper functions that call the X509at functions
11371 when passed certificate requests. (TO DO: similar things can be done with
11372 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11373 things. Some of these need some d2i or i2d and print functionality
11374 because they handle more complex structures.)
11377 *) Add missing #ifndefs that caused missing symbols when building libssl
11378 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11379 NO_RSA in ssl/s2*.c.
11380 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11382 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11383 has a return value which indicates the quality of the random data
11384 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11385 error queue. New function RAND_pseudo_bytes() generates output that is
11386 guaranteed to be unique but not unpredictable. RAND_add is like
11387 RAND_seed, but takes an extra argument for an entropy estimate
11388 (RAND_seed always assumes full entropy).
11391 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11392 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11393 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11394 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11395 false-positive rate of at most 2^-80 for random input.
11398 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11401 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11402 in the 0.9.5 release), this returns the chain
11403 from an X509_CTX structure with a dup of the stack and all
11404 the X509 reference counts upped: so the stack will exist
11405 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11408 Also make SSL_SESSION_print() print out the verify return
11412 *) Add manpage for the pkcs12 command. Also change the default
11413 behaviour so MAC iteration counts are used unless the new
11414 -nomaciter option is used. This improves file security and
11415 only older versions of MSIE (4.0 for example) need it.
11418 *) Honor the no-xxx Configure options when creating .DEF files.
11421 *) Add PKCS#10 attributes to field table: challengePassword,
11422 unstructuredName and unstructuredAddress. These are taken from
11423 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11424 international characters are used.
11426 More changes to X509_ATTRIBUTE code: allow the setting of types
11427 based on strings. Remove the 'loc' parameter when adding
11428 attributes because these will be a SET OF encoding which is sorted
11432 *) Initial changes to the 'req' utility to allow request generation
11433 automation. This will allow an application to just generate a template
11434 file containing all the field values and have req construct the
11437 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11438 used all over the place including certificate requests and PKCS#7
11439 structures. They are currently handled manually where necessary with
11440 some primitive wrappers for PKCS#7. The new functions behave in a
11441 manner analogous to the X509 extension functions: they allow
11442 attributes to be looked up by NID and added.
11444 Later something similar to the X509V3 code would be desirable to
11445 automatically handle the encoding, decoding and printing of the
11446 more complex types. The string types like challengePassword can
11447 be handled by the string table functions.
11449 Also modified the multi byte string table handling. Now there is
11450 a 'global mask' which masks out certain types. The table itself
11451 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11452 is useful when for example there is only one permissible type
11453 (as in countryName) and using the mask might result in no valid
11457 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11458 SSL_get_peer_finished to allow applications to obtain the latest
11459 Finished messages sent to the peer or expected from the peer,
11460 respectively. (SSL_get_peer_finished is usually the Finished message
11461 actually received from the peer, otherwise the protocol will be aborted.)
11463 As the Finished message are message digests of the complete handshake
11464 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11465 be used for external authentication procedures when the authentication
11466 provided by SSL/TLS is not desired or is not enough.
11469 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11470 the host supports BWX extension and if Compaq C is present on the
11471 $PATH. Just exploiting of the BWX extension results in 20-30%
11472 performance kick for some algorithms, e.g. DES and RC4 to mention
11473 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11477 *) Add support for MS "fast SGC". This is arguably a violation of the
11478 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11479 weak crypto and after checking the certificate is SGC a second one
11480 with strong crypto. MS SGC stops the first handshake after receiving
11481 the server certificate message and sends a second client hello. Since
11482 a server will typically do all the time consuming operations before
11483 expecting any further messages from the client (server key exchange
11484 is the most expensive) there is little difference between the two.
11486 To get OpenSSL to support MS SGC we have to permit a second client
11487 hello message after we have sent server done. In addition we have to
11488 reset the MAC if we do get this second client hello.
11491 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11492 if a DER encoded private key is RSA or DSA traditional format. Changed
11493 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11494 format DER encoded private key. Newer code should use PKCS#8 format which
11495 has the key type encoded in the ASN1 structure. Added DER private key
11496 support to pkcs8 application.
11499 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11500 ciphersuites has been selected (as required by the SSL 3/TLS 1
11501 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11502 is set, we interpret this as a request to violate the specification
11503 (the worst that can happen is a handshake failure, and 'correct'
11504 behaviour would result in a handshake failure anyway).
11507 *) In SSL_CTX_add_session, take into account that there might be multiple
11508 SSL_SESSION structures with the same session ID (e.g. when two threads
11509 concurrently obtain them from an external cache).
11510 The internal cache can handle only one SSL_SESSION with a given ID,
11511 so if there's a conflict, we now throw out the old one to achieve
11515 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11516 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11517 some routines that use cipher OIDs: some ciphers do not have OIDs
11518 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11522 *) Simplify the trust setting structure and code. Now we just have
11523 two sequences of OIDs for trusted and rejected settings. These will
11524 typically have values the same as the extended key usage extension
11525 and any application specific purposes.
11527 The trust checking code now has a default behaviour: it will just
11528 check for an object with the same NID as the passed id. Functions can
11529 be provided to override either the default behaviour or the behaviour
11530 for a given id. SSL client, server and email already have functions
11531 in place for compatibility: they check the NID and also return "trusted"
11532 if the certificate is self signed.
11535 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11536 traditional format into an EVP_PKEY structure.
11539 *) Add a password callback function PEM_cb() which either prompts for
11540 a password if usr_data is NULL or otherwise assumes it is a null
11541 terminated password. Allow passwords to be passed on command line
11542 environment or config files in a few more utilities.
11545 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11546 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11547 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11548 Update documentation.
11551 *) Support for ASN1 "NULL" type. This could be handled before by using
11552 ASN1_TYPE but there wasn't any function that would try to read a NULL
11553 and produce an error if it couldn't. For compatibility we also have
11554 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11555 don't allocate anything because they don't need to.
11558 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11560 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11562 *) Rebuild of the memory allocation routines used by OpenSSL code and
11563 possibly others as well. The purpose is to make an interface that
11564 provide hooks so anyone can build a separate set of allocation and
11565 deallocation routines to be used by OpenSSL, for example memory
11566 pool implementations, or something else, which was previously hard
11567 since Malloc(), Realloc() and Free() were defined as macros having
11568 the values malloc, realloc and free, respectively (except for Win32
11569 compilations). The same is provided for memory debugging code.
11570 OpenSSL already comes with functionality to find memory leaks, but
11571 this gives people a chance to debug other memory problems.
11573 With these changes, a new set of functions and macros have appeared:
11575 CRYPTO_set_mem_debug_functions() [F]
11576 CRYPTO_get_mem_debug_functions() [F]
11577 CRYPTO_dbg_set_options() [F]
11578 CRYPTO_dbg_get_options() [F]
11579 CRYPTO_malloc_debug_init() [M]
11581 The memory debug functions are NULL by default, unless the library
11582 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11583 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11584 gives the standard debugging functions that come with OpenSSL) or
11585 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11586 provided by the library user) must be used. When the standard
11587 debugging functions are used, CRYPTO_dbg_set_options can be used to
11588 request additional information:
11589 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11590 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11592 Also, things like CRYPTO_set_mem_functions will always give the
11593 expected result (the new set of functions is used for allocation
11594 and deallocation) at all times, regardless of platform and compiler
11597 To finish it up, some functions that were never use in any other
11598 way than through macros have a new API and new semantic:
11600 CRYPTO_dbg_malloc()
11601 CRYPTO_dbg_realloc()
11604 All macros of value have retained their old syntax.
11605 [Richard Levitte and Bodo Moeller]
11607 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11608 ordering of SMIMECapabilities wasn't in "strength order" and there
11609 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11613 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11614 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11615 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11617 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11618 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11619 functionality to handle multipart/signed properly) and a utility
11620 called 'smime' to call all this stuff. This is based on code I
11621 originally wrote for Celo who have kindly allowed it to be
11622 included in OpenSSL.
11625 *) Add variants des_set_key_checked and des_set_key_unchecked of
11626 des_set_key (aka des_key_sched). Global variable des_check_key
11627 decides which of these is called by des_set_key; this way
11628 des_check_key behaves as it always did, but applications and
11629 the library itself, which was buggy for des_check_key == 1,
11630 have a cleaner way to pick the version they need.
11633 *) New function PKCS12_newpass() which changes the password of a
11637 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11638 dynamic mix. In both cases the ids can be used as an index into the
11639 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11640 functions so they accept a list of the field values and the
11641 application doesn't need to directly manipulate the X509_TRUST
11645 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11649 *) Modify the way the V3 extension code looks up extensions. This now
11650 works in a similar way to the object code: we have some "standard"
11651 extensions in a static table which is searched with OBJ_bsearch()
11652 and the application can add dynamic ones if needed. The file
11653 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11654 updated whenever a new extension is added to the core code and kept
11655 in ext_nid order. There is a simple program 'tabtest.c' which checks
11656 this. New extensions are not added too often so this file can readily
11657 be maintained manually.
11659 There are two big advantages in doing things this way. The extensions
11660 can be looked up immediately and no longer need to be "added" using
11661 X509V3_add_standard_extensions(): this function now does nothing.
11662 [Side note: I get *lots* of email saying the extension code doesn't
11663 work because people forget to call this function]
11664 Also no dynamic allocation is done unless new extensions are added:
11665 so if we don't add custom extensions there is no need to call
11666 X509V3_EXT_cleanup().
11669 *) Modify enc utility's salting as follows: make salting the default. Add a
11670 magic header, so unsalted files fail gracefully instead of just decrypting
11671 to garbage. This is because not salting is a big security hole, so people
11672 should be discouraged from doing it.
11675 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11676 digest to be passed on the command line but it only used this
11677 parameter when signing a certificate. Modified so all relevant
11678 operations are affected by the digest parameter including the
11679 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11680 DSA key was used because it didn't fix the digest.
11683 *) Initial certificate chain verify code. Currently tests the untrusted
11684 certificates for consistency with the verify purpose (which is set
11685 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11687 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11688 this is because it will reject chains with invalid extensions whereas
11689 every previous version of OpenSSL and SSLeay made no checks at all.
11691 Trust code: checks the root CA for the relevant trust settings. Trust
11692 settings have an initial value consistent with the verify purpose: e.g.
11693 if the verify purpose is for SSL client use it expects the CA to be
11694 trusted for SSL client use. However the default value can be changed to
11695 permit custom trust settings: one example of this would be to only trust
11696 certificates from a specific "secure" set of CAs.
11698 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11699 which should be used for version portability: especially since the
11700 verify structure is likely to change more often now.
11702 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11703 to set them. If not set then assume SSL clients will verify SSL servers
11706 Two new options to the verify program: -untrusted allows a set of
11707 untrusted certificates to be passed in and -purpose which sets the
11708 intended purpose of the certificate. If a purpose is set then the
11709 new chain verify code is used to check extension consistency.
11712 *) Support for the authority information access extension.
11715 *) Modify RSA and DSA PEM read routines to transparently handle
11716 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11717 public keys in a format compatible with certificate
11718 SubjectPublicKeyInfo structures. Unfortunately there were already
11719 functions called *_PublicKey_* which used various odd formats so
11720 these are retained for compatibility: however the DSA variants were
11721 never in a public release so they have been deleted. Changed dsa/rsa
11722 utilities to handle the new format: note no releases ever handled public
11723 keys so we should be OK.
11725 The primary motivation for this change is to avoid the same fiasco
11726 that dogs private keys: there are several incompatible private key
11727 formats some of which are standard and some OpenSSL specific and
11728 require various evil hacks to allow partial transparent handling and
11729 even then it doesn't work with DER formats. Given the option anything
11730 other than PKCS#8 should be dumped: but the other formats have to
11731 stay in the name of compatibility.
11733 With public keys and the benefit of hindsight one standard format
11734 is used which works with EVP_PKEY, RSA or DSA structures: though
11735 it clearly returns an error if you try to read the wrong kind of key.
11737 Added a -pubkey option to the 'x509' utility to output the public key.
11738 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11739 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11740 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11741 that do the same as the EVP_PKEY_assign_*() except they up the
11742 reference count of the added key (they don't "swallow" the
11746 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11747 CRLs would fail if the file contained no certificates or no CRLs:
11748 added a new function to read in both types and return the number
11749 read: this means that if none are read it will be an error. The
11750 DER versions of the certificate and CRL reader would always fail
11751 because it isn't possible to mix certificates and CRLs in DER format
11752 without choking one or the other routine. Changed this to just read
11753 a certificate: this is the best we can do. Also modified the code
11754 in apps/verify.c to take notice of return codes: it was previously
11755 attempting to read in certificates from NULL pointers and ignoring
11756 any errors: this is one reason why the cert and CRL reader seemed
11757 to work. It doesn't check return codes from the default certificate
11758 routines: these may well fail if the certificates aren't installed.
11761 *) Code to support otherName option in GeneralName.
11764 *) First update to verify code. Change the verify utility
11765 so it warns if it is passed a self signed certificate:
11766 for consistency with the normal behaviour. X509_verify
11767 has been modified to it will now verify a self signed
11768 certificate if *exactly* the same certificate appears
11769 in the store: it was previously impossible to trust a
11770 single self signed certificate. This means that:
11771 openssl verify ss.pem
11772 now gives a warning about a self signed certificate but
11773 openssl verify -CAfile ss.pem ss.pem
11777 *) For servers, store verify_result in SSL_SESSION data structure
11778 (and add it to external session representation).
11779 This is needed when client certificate verifications fails,
11780 but an application-provided verification callback (set by
11781 SSL_CTX_set_cert_verify_callback) allows accepting the session
11782 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11783 but returns 1): When the session is reused, we have to set
11784 ssl->verify_result to the appropriate error code to avoid
11786 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11788 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11789 case in PKCS7_dataInit() where the signed PKCS7 structure
11790 didn't contain any existing data because it was being created.
11791 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11793 *) Add a salt to the key derivation routines in enc.c. This
11794 forms the first 8 bytes of the encrypted file. Also add a
11795 -S option to allow a salt to be input on the command line.
11798 *) New function X509_cmp(). Oddly enough there wasn't a function
11799 to compare two certificates. We do this by working out the SHA1
11800 hash and comparing that. X509_cmp() will be needed by the trust
11804 *) SSL_get1_session() is like SSL_get_session(), but increments
11805 the reference count in the SSL_SESSION returned.
11806 [Geoff Thorpe <geoff@eu.c2.net>]
11808 *) Fix for 'req': it was adding a null to request attributes.
11809 Also change the X509_LOOKUP and X509_INFO code to handle
11810 certificate auxiliary information.
11813 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11817 *) Add the possibility to add extra information to the memory leak
11818 detecting output, to form tracebacks, showing from where each
11819 allocation was originated: CRYPTO_push_info("constant string") adds
11820 the string plus current file name and line number to a per-thread
11821 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11822 is like calling CYRPTO_pop_info() until the stack is empty.
11823 Also updated memory leak detection code to be multi-thread-safe.
11826 *) Add options -text and -noout to pkcs7 utility and delete the
11827 encryption options which never did anything. Update docs.
11830 *) Add options to some of the utilities to allow the pass phrase
11831 to be included on either the command line (not recommended on
11832 OSes like Unix) or read from the environment. Update the
11833 manpages and fix a few bugs.
11836 *) Add a few manpages for some of the openssl commands.
11839 *) Fix the -revoke option in ca. It was freeing up memory twice,
11840 leaking and not finding already revoked certificates.
11843 *) Extensive changes to support certificate auxiliary information.
11844 This involves the use of X509_CERT_AUX structure and X509_AUX
11845 functions. An X509_AUX function such as PEM_read_X509_AUX()
11846 can still read in a certificate file in the usual way but it
11847 will also read in any additional "auxiliary information". By
11848 doing things this way a fair degree of compatibility can be
11849 retained: existing certificates can have this information added
11850 using the new 'x509' options.
11852 Current auxiliary information includes an "alias" and some trust
11853 settings. The trust settings will ultimately be used in enhanced
11854 certificate chain verification routines: currently a certificate
11855 can only be trusted if it is self signed and then it is trusted
11859 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11860 The problem was that one of the replacement routines had not been working
11861 since SSLeay releases. For now the offending routine has been replaced
11862 with non-optimised assembler. Even so, this now gives around 95%
11863 performance improvement for 1024 bit RSA signs.
11866 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11867 handling. Most clients have the effective key size in bits equal to
11868 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11869 A few however don't do this and instead use the size of the decrypted key
11870 to determine the RC2 key length and the AlgorithmIdentifier to determine
11871 the effective key length. In this case the effective key length can still
11872 be 40 bits but the key length can be 168 bits for example. This is fixed
11873 by manually forcing an RC2 key into the EVP_PKEY structure because the
11874 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11875 the key length and effective key length are equal.
11878 *) Add a bunch of functions that should simplify the creation of
11879 X509_NAME structures. Now you should be able to do:
11880 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11881 and have it automatically work out the correct field type and fill in
11882 the structures. The more adventurous can try:
11883 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11884 and it will (hopefully) work out the correct multibyte encoding.
11887 *) Change the 'req' utility to use the new field handling and multibyte
11888 copy routines. Before the DN field creation was handled in an ad hoc
11889 way in req, ca, and x509 which was rather broken and didn't support
11890 BMPStrings or UTF8Strings. Since some software doesn't implement
11891 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11892 using the dirstring_type option. See the new comment in the default
11893 openssl.cnf for more info.
11896 *) Make crypto/rand/md_rand.c more robust:
11897 - Assure unique random numbers after fork().
11898 - Make sure that concurrent threads access the global counter and
11899 md serializably so that we never lose entropy in them
11900 or use exactly the same state in multiple threads.
11901 Access to the large state is not always serializable because
11902 the additional locking could be a performance killer, and
11903 md should be large enough anyway.
11906 *) New file apps/app_rand.c with commonly needed functionality
11907 for handling the random seed file.
11909 Use the random seed file in some applications that previously did not:
11911 dsaparam -genkey (which also ignored its '-rand' option),
11914 x509 (when signing).
11915 Except on systems with /dev/urandom, it is crucial to have a random
11916 seed file at least for key creation, DSA signing, and for DH exchanges;
11917 for RSA signatures we could do without one.
11919 gendh and gendsa (unlike genrsa) used to read only the first byte
11920 of each file listed in the '-rand' option. The function as previously
11921 found in genrsa is now in app_rand.c and is used by all programs
11922 that support '-rand'.
11925 *) In RAND_write_file, use mode 0600 for creating files;
11926 don't just chmod when it may be too late.
11929 *) Report an error from X509_STORE_load_locations
11930 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11933 *) New function ASN1_mbstring_copy() this copies a string in either
11934 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11935 into an ASN1_STRING type. A mask of permissible types is passed
11936 and it chooses the "minimal" type to use or an error if not type
11940 *) Add function equivalents to the various macros in asn1.h. The old
11941 macros are retained with an M_ prefix. Code inside the library can
11942 use the M_ macros. External code (including the openssl utility)
11943 should *NOT* in order to be "shared library friendly".
11946 *) Add various functions that can check a certificate's extensions
11947 to see if it usable for various purposes such as SSL client,
11948 server or S/MIME and CAs of these types. This is currently
11949 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11950 verification. Also added a -purpose flag to x509 utility to
11951 print out all the purposes.
11954 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11958 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11959 for, obtain and decode and extension and obtain its critical flag.
11960 This allows all the necessary extension code to be handled in a
11961 single function call.
11964 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11965 platforms. See crypto/rc4/rc4_enc.c for further details.
11968 *) New -noout option to asn1parse. This causes no output to be produced
11969 its main use is when combined with -strparse and -out to extract data
11970 from a file (which may not be in ASN.1 format).
11973 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11974 when producing the local key id.
11975 [Richard Levitte <levitte@stacken.kth.se>]
11977 *) New option -dhparam in s_server. This allows a DH parameter file to be
11978 stated explicitly. If it is not stated then it tries the first server
11979 certificate file. The previous behaviour hard coded the filename
11983 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11984 a public key to be input or output. For example:
11985 openssl rsa -in key.pem -pubout -out pubkey.pem
11986 Also added necessary DSA public key functions to handle this.
11989 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11990 in the message. This was handled by allowing
11991 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11992 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11994 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11995 to the end of the strings whereas this didn't. This would cause problems
11996 if strings read with d2i_ASN1_bytes() were later modified.
11997 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11999 *) Fix for base64 decode bug. When a base64 bio reads only one line of
12000 data and it contains EOF it will end up returning an error. This is
12001 caused by input 46 bytes long. The cause is due to the way base64
12002 BIOs find the start of base64 encoded data. They do this by trying a
12003 trial decode on each line until they find one that works. When they
12004 do a flag is set and it starts again knowing it can pass all the
12005 data directly through the decoder. Unfortunately it doesn't reset
12006 the context it uses. This means that if EOF is reached an attempt
12007 is made to pass two EOFs through the context and this causes the
12008 resulting error. This can also cause other problems as well. As is
12009 usual with these problems it takes *ages* to find and the fix is
12010 trivial: move one line.
12011 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
12013 *) Ugly workaround to get s_client and s_server working under Windows. The
12014 old code wouldn't work because it needed to select() on sockets and the
12015 tty (for keypresses and to see if data could be written). Win32 only
12016 supports select() on sockets so we select() with a 1s timeout on the
12017 sockets and then see if any characters are waiting to be read, if none
12018 are present then we retry, we also assume we can always write data to
12019 the tty. This isn't nice because the code then blocks until we've
12020 received a complete line of data and it is effectively polling the
12021 keyboard at 1s intervals: however it's quite a bit better than not
12022 working at all :-) A dedicated Windows application might handle this
12023 with an event loop for example.
12026 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
12027 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
12028 will be called when RSA_sign() and RSA_verify() are used. This is useful
12029 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
12030 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
12031 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
12032 This necessitated the support of an extra signature type NID_md5_sha1
12033 for SSL signatures and modifications to the SSL library to use it instead
12034 of calling RSA_public_decrypt() and RSA_private_encrypt().
12037 *) Add new -verify -CAfile and -CApath options to the crl program, these
12038 will lookup a CRL issuers certificate and verify the signature in a
12039 similar way to the verify program. Tidy up the crl program so it
12040 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
12041 less strict. It will now permit CRL extensions even if it is not
12042 a V2 CRL: this will allow it to tolerate some broken CRLs.
12045 *) Initialize all non-automatic variables each time one of the openssl
12046 sub-programs is started (this is necessary as they may be started
12047 multiple times from the "OpenSSL>" prompt).
12048 [Lennart Bang, Bodo Moeller]
12050 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
12051 removing all other RSA functionality (this is what NO_RSA does). This
12052 is so (for example) those in the US can disable those operations covered
12053 by the RSA patent while allowing storage and parsing of RSA keys and RSA
12057 *) Non-copying interface to BIO pairs.
12058 (still largely untested)
12061 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
12062 ASCII string. This was handled independently in various places before.
12065 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
12066 UTF8 strings a character at a time.
12069 *) Use client_version from client hello to select the protocol
12070 (s23_srvr.c) and for RSA client key exchange verification
12071 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
12074 *) Add various utility functions to handle SPKACs, these were previously
12075 handled by poking round in the structure internals. Added new function
12076 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
12077 print, verify and generate SPKACs. Based on an original idea from
12078 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
12081 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
12084 *) Allow the config file extension section to be overwritten on the
12085 command line. Based on an original idea from Massimiliano Pala
12086 <madwolf@comune.modena.it>. The new option is called -extensions
12087 and can be applied to ca, req and x509. Also -reqexts to override
12088 the request extensions in req and -crlexts to override the crl extensions
12092 *) Add new feature to the SPKAC handling in ca. Now you can include
12093 the same field multiple times by preceding it by "XXXX." for example:
12096 this is the same syntax as used in the req config file.
12099 *) Allow certificate extensions to be added to certificate requests. These
12100 are specified in a 'req_extensions' option of the req section of the
12101 config file. They can be printed out with the -text option to req but
12102 are otherwise ignored at present.
12105 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
12106 data read consists of only the final block it would not decrypted because
12107 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
12108 A misplaced 'break' also meant the decrypted final block might not be
12109 copied until the next read.
12112 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
12113 a few extra parameters to the DH structure: these will be useful if
12114 for example we want the value of 'q' or implement X9.42 DH.
12117 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
12118 provides hooks that allow the default DSA functions or functions on a
12119 "per key" basis to be replaced. This allows hardware acceleration and
12120 hardware key storage to be handled without major modification to the
12121 library. Also added low level modexp hooks and CRYPTO_EX structure and
12122 associated functions.
12125 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
12126 as "read only": it can't be written to and the buffer it points to will
12127 not be freed. Reading from a read only BIO is much more efficient than
12128 a normal memory BIO. This was added because there are several times when
12129 an area of memory needs to be read from a BIO. The previous method was
12130 to create a memory BIO and write the data to it, this results in two
12131 copies of the data and an O(n^2) reading algorithm. There is a new
12132 function BIO_new_mem_buf() which creates a read only memory BIO from
12133 an area of memory. Also modified the PKCS#7 routines to use read only
12137 *) Bugfix: ssl23_get_client_hello did not work properly when called in
12138 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
12139 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
12140 but a retry condition occurred while trying to read the rest.
12143 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
12144 NID_pkcs7_encrypted by default: this was wrong since this should almost
12145 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
12146 the encrypted data type: this is a more sensible place to put it and it
12147 allows the PKCS#12 code to be tidied up that duplicated this
12151 *) Changed obj_dat.pl script so it takes its input and output files on
12152 the command line. This should avoid shell escape redirection problems
12156 *) Initial support for certificate extension requests, these are included
12157 in things like Xenroll certificate requests. Included functions to allow
12158 extensions to be obtained and added.
12161 *) -crlf option to s_client and s_server for sending newlines as
12162 CRLF (as required by many protocols).
12165 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
12167 *) Install libRSAglue.a when OpenSSL is built with RSAref.
12168 [Ralf S. Engelschall]
12170 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
12171 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
12173 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
12177 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
12178 DH parameters/keys (q is lost during that conversion, but the resulting
12179 DH parameters contain its length).
12181 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
12182 much faster than DH_generate_parameters (which creates parameters
12183 where p = 2*q + 1), and also the smaller q makes DH computations
12184 much more efficient (160-bit exponentiation instead of 1024-bit
12185 exponentiation); so this provides a convenient way to support DHE
12186 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
12187 utter importance to use
12188 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12190 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12191 when such DH parameters are used, because otherwise small subgroup
12192 attacks may become possible!
12195 *) Avoid memory leak in i2d_DHparams.
12198 *) Allow the -k option to be used more than once in the enc program:
12199 this allows the same encrypted message to be read by multiple recipients.
12202 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
12203 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
12204 it will always use the numerical form of the OID, even if it has a short
12208 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
12209 method only got called if p,q,dmp1,dmq1,iqmp components were present,
12210 otherwise bn_mod_exp was called. In the case of hardware keys for example
12211 no private key components need be present and it might store extra data
12212 in the RSA structure, which cannot be accessed from bn_mod_exp.
12213 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
12214 private key operations.
12217 *) Added support for SPARC Linux.
12220 *) pem_password_cb function type incompatibly changed from
12221 typedef int pem_password_cb(char *buf, int size, int rwflag);
12223 ....(char *buf, int size, int rwflag, void *userdata);
12224 so that applications can pass data to their callbacks:
12225 The PEM[_ASN1]_{read,write}... functions and macros now take an
12226 additional void * argument, which is just handed through whenever
12227 the password callback is called.
12228 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
12230 New function SSL_CTX_set_default_passwd_cb_userdata.
12232 Compatibility note: As many C implementations push function arguments
12233 onto the stack in reverse order, the new library version is likely to
12234 interoperate with programs that have been compiled with the old
12235 pem_password_cb definition (PEM_whatever takes some data that
12236 happens to be on the stack as its last argument, and the callback
12237 just ignores this garbage); but there is no guarantee whatsoever that
12240 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
12241 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
12242 problems not only on Windows, but also on some Unix platforms.
12243 To avoid problematic command lines, these definitions are now in an
12244 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
12245 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
12248 *) MIPS III/IV assembler module is reimplemented.
12251 *) More DES library cleanups: remove references to srand/rand and
12252 delete an unused file.
12255 *) Add support for the free Netwide assembler (NASM) under Win32,
12256 since not many people have MASM (ml) and it can be hard to obtain.
12257 This is currently experimental but it seems to work OK and pass all
12258 the tests. Check out INSTALL.W32 for info.
12261 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
12262 without temporary keys kept an extra copy of the server key,
12263 and connections with temporary keys did not free everything in case
12267 *) New function RSA_check_key and new openssl rsa option -check
12268 for verifying the consistency of RSA keys.
12269 [Ulf Moeller, Bodo Moeller]
12271 *) Various changes to make Win32 compile work:
12272 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
12273 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
12274 comparison" warnings.
12275 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
12278 *) Add a debugging option to PKCS#5 v2 key generation function: when
12279 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
12280 derived keys are printed to stderr.
12283 *) Copy the flags in ASN1_STRING_dup().
12284 [Roman E. Pavlov <pre@mo.msk.ru>]
12286 *) The x509 application mishandled signing requests containing DSA
12287 keys when the signing key was also DSA and the parameters didn't match.
12289 It was supposed to omit the parameters when they matched the signing key:
12290 the verifying software was then supposed to automatically use the CA's
12291 parameters if they were absent from the end user certificate.
12293 Omitting parameters is no longer recommended. The test was also
12294 the wrong way round! This was probably due to unusual behaviour in
12295 EVP_cmp_parameters() which returns 1 if the parameters match.
12296 This meant that parameters were omitted when they *didn't* match and
12297 the certificate was useless. Certificates signed with 'ca' didn't have
12299 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
12301 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
12302 The interface is as follows:
12303 Applications can use
12304 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
12305 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
12306 "off" is now the default.
12307 The library internally uses
12308 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
12309 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
12310 to disable memory-checking temporarily.
12312 Some inconsistent states that previously were possible (and were
12313 even the default) are now avoided.
12315 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
12316 with each memory chunk allocated; this is occasionally more helpful
12317 than just having a counter.
12319 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
12321 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
12325 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
12326 which largely parallels "options", but is for changing API behaviour,
12327 whereas "options" are about protocol behaviour.
12328 Initial "mode" flags are:
12330 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
12331 a single record has been written.
12332 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
12333 retries use the same buffer location.
12334 (But all of the contents must be
12338 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
12341 *) Fix problems with no-hmac etc.
12342 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
12344 *) New functions RSA_get_default_method(), RSA_set_method() and
12345 RSA_get_method(). These allows replacement of RSA_METHODs without having
12346 to mess around with the internals of an RSA structure.
12349 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
12350 Also really enable memory leak checks in openssl.c and in some
12352 [Chad C. Mulligan, Bodo Moeller]
12354 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
12355 up the length of negative integers. This has now been simplified to just
12356 store the length when it is first determined and use it later, rather
12357 than trying to keep track of where data is copied and updating it to
12359 [Steve Henson, reported by Brien Wheeler
12360 <bwheeler@authentica-security.com>]
12362 *) Add a new function PKCS7_signatureVerify. This allows the verification
12363 of a PKCS#7 signature but with the signing certificate passed to the
12364 function itself. This contrasts with PKCS7_dataVerify which assumes the
12365 certificate is present in the PKCS#7 structure. This isn't always the
12366 case: certificates can be omitted from a PKCS#7 structure and be
12367 distributed by "out of band" means (such as a certificate database).
12370 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12371 function prototypes in pem.h, also change util/mkdef.pl to add the
12372 necessary function names.
12375 *) mk1mf.pl (used by Windows builds) did not properly read the
12376 options set by Configure in the top level Makefile, and Configure
12377 was not even able to write more than one option correctly.
12378 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12381 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12382 file to be loaded from a BIO or FILE pointer. The BIO version will
12383 for example allow memory BIOs to contain config info.
12386 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12387 Whoever hopes to achieve shared-library compatibility across versions
12388 must use this, not the compile-time macro.
12389 (Exercise 0.9.4: Which is the minimum library version required by
12391 Note: All this applies only to multi-threaded programs, others don't
12395 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12396 through a BIO pair triggered the default case, i.e.
12397 SSLerr(...,SSL_R_UNKNOWN_STATE).
12400 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12401 can use the SSL library even if none of the specific BIOs is
12405 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12406 for the encoded length.
12407 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12409 *) Add initial documentation of the X509V3 functions.
12412 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12413 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12414 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12415 secure PKCS#8 private key format with a high iteration count.
12418 *) Fix determination of Perl interpreter: A perl or perl5
12419 _directory_ in $PATH was also accepted as the interpreter.
12420 [Ralf S. Engelschall]
12422 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12423 wrong with it but it was very old and did things like calling
12424 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12425 unusual formatting.
12428 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12429 to use the new extension code.
12432 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12433 with macros. This should make it easier to change their form, add extra
12434 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12438 *) Add to configuration table a new entry that can specify an alternative
12439 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12440 according to Mark Crispin <MRC@Panda.COM>.
12444 *) DES CBC did not update the IV. Weird.
12447 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12448 Changing the behaviour of the former might break existing programs --
12449 where IV updating is needed, des_ncbc_encrypt can be used.
12452 *) When bntest is run from "make test" it drives bc to check its
12453 calculations, as well as internally checking them. If an internal check
12454 fails, it needs to cause bc to give a non-zero result or make test carries
12455 on without noticing the failure. Fixed.
12458 *) DES library cleanups.
12461 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12462 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12463 ciphers. NOTE: although the key derivation function has been verified
12464 against some published test vectors it has not been extensively tested
12465 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12469 *) Instead of "mkdir -p", which is not fully portable, use new
12470 Perl script "util/mkdir-p.pl".
12473 *) Rewrite the way password based encryption (PBE) is handled. It used to
12474 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12475 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12476 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12477 the 'parameter' field of the AlgorithmIdentifier is passed to the
12478 underlying key generation function so it must do its own ASN1 parsing.
12479 This has also changed the EVP_PBE_CipherInit() function which now has a
12480 'parameter' argument instead of literal salt and iteration count values
12481 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12484 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12485 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12486 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12487 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12488 value was just used as a "magic string" and not used directly its
12489 value doesn't matter.
12492 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12496 *) "linux-sparc64" configuration (ultrapenguin).
12497 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12498 "linux-sparc" configuration.
12499 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12501 *) config now generates no-xxx options for missing ciphers.
12504 *) Support the EBCDIC character set (work in progress).
12505 File ebcdic.c not yet included because it has a different license.
12506 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12508 *) Support BS2000/OSD-POSIX.
12509 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12511 *) Make callbacks for key generation use void * instead of char *.
12514 *) Make S/MIME samples compile (not yet tested).
12517 *) Additional typesafe stacks.
12520 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12524 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12526 *) New configuration variant "sco5-gcc".
12528 *) Updated some demos.
12529 [Sean O Riordain, Wade Scholine]
12531 *) Add missing BIO_free at exit of pkcs12 application.
12534 *) Fix memory leak in conf.c.
12537 *) Updates for Win32 to assembler version of MD5.
12540 *) Set #! path to perl in apps/der_chop to where we found it
12541 instead of using a fixed path.
12544 *) SHA library changes for irix64-mips4-cc.
12547 *) Improvements for VMS support.
12551 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12553 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12554 This also avoids the problems with SC4.2 and unpatched SC5.
12555 [Andy Polyakov <appro@fy.chalmers.se>]
12557 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12558 These are required because of the typesafe stack would otherwise break
12559 existing code. If old code used a structure member which used to be STACK
12560 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12561 sk_num or sk_value it would produce an error because the num, data members
12562 are not present in STACK_OF. Now it just produces a warning. sk_set
12563 replaces the old method of assigning a value to sk_value
12564 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12565 that does this will no longer work (and should use sk_set instead) but
12566 this could be regarded as a "questionable" behaviour anyway.
12569 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12570 correctly handle encrypted S/MIME data.
12573 *) Change type of various DES function arguments from des_cblock
12574 (which means, in function argument declarations, pointer to char)
12575 to des_cblock * (meaning pointer to array with 8 char elements),
12576 which allows the compiler to do more typechecking; it was like
12577 that back in SSLeay, but with lots of ugly casts.
12579 Introduce new type const_des_cblock.
12582 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12583 problems: find RecipientInfo structure that matches recipient certificate
12584 and initialise the ASN1 structures properly based on passed cipher.
12587 *) Belatedly make the BN tests actually check the results.
12590 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12591 to and from BNs: it was completely broken. New compilation option
12592 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12593 key elements as negative integers.
12596 *) Reorganize and speed up MD5.
12597 [Andy Polyakov <appro@fy.chalmers.se>]
12600 [Richard Levitte <richard@levitte.org>]
12602 *) New option -out to asn1parse to allow the parsed structure to be
12603 output to a file. This is most useful when combined with the -strparse
12604 option to examine the output of things like OCTET STRINGS.
12607 *) Make SSL library a little more fool-proof by not requiring any longer
12608 that SSL_set_{accept,connect}_state be called before
12609 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12610 in many applications because usually everything *appeared* to work as
12611 intended anyway -- now it really works as intended).
12614 *) Move openssl.cnf out of lib/.
12617 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12618 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12619 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12620 [Ralf S. Engelschall]
12622 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12623 handle PKCS#7 enveloped data properly.
12624 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12626 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12627 copying pointers. The cert_st handling is changed by this in
12628 various ways (and thus what used to be known as ctx->default_cert
12629 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12630 any longer when s->cert does not give us what we need).
12631 ssl_cert_instantiate becomes obsolete by this change.
12632 As soon as we've got the new code right (possibly it already is?),
12633 we have solved a couple of bugs of the earlier code where s->cert
12634 was used as if it could not have been shared with other SSL structures.
12636 Note that using the SSL API in certain dirty ways now will result
12637 in different behaviour than observed with earlier library versions:
12638 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12639 does not influence s as it used to.
12641 In order to clean up things more thoroughly, inside SSL_SESSION
12642 we don't use CERT any longer, but a new structure SESS_CERT
12643 that holds per-session data (if available); currently, this is
12644 the peer's certificate chain and, for clients, the server's certificate
12645 and temporary key. CERT holds only those values that can have
12646 meaningful defaults in an SSL_CTX.
12649 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12650 from the internal representation. Various PKCS#7 fixes: remove some
12651 evil casts and set the enc_dig_alg field properly based on the signing
12655 *) Allow PKCS#12 password to be set from the command line or the
12656 environment. Let 'ca' get its config file name from the environment
12657 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12661 *) Allow certificate policies extension to use an IA5STRING for the
12662 organization field. This is contrary to the PKIX definition but
12663 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12667 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12668 without disallowing inline assembler and the like for non-pedantic builds.
12671 *) Support Borland C++ builder.
12672 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12674 *) Support Mingw32.
12677 *) SHA-1 cleanups and performance enhancements.
12678 [Andy Polyakov <appro@fy.chalmers.se>]
12680 *) Sparc v8plus assembler for the bignum library.
12681 [Andy Polyakov <appro@fy.chalmers.se>]
12683 *) Accept any -xxx and +xxx compiler options in Configure.
12686 *) Update HPUX configuration.
12689 *) Add missing sk_<type>_unshift() function to safestack.h
12690 [Ralf S. Engelschall]
12692 *) New function SSL_CTX_use_certificate_chain_file that sets the
12693 "extra_cert"s in addition to the certificate. (This makes sense
12694 only for "PEM" format files, as chains as a whole are not
12698 *) Support verify_depth from the SSL API.
12699 x509_vfy.c had what can be considered an off-by-one-error:
12700 Its depth (which was not part of the external interface)
12701 was actually counting the number of certificates in a chain;
12702 now it really counts the depth.
12705 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12706 instead of X509err, which often resulted in confusing error
12707 messages since the error codes are not globally unique
12708 (e.g. an alleged error in ssl3_accept when a certificate
12709 didn't match the private key).
12711 *) New function SSL_CTX_set_session_id_context that allows to set a default
12712 value (so that you don't need SSL_set_session_id_context for each
12713 connection using the SSL_CTX).
12716 *) OAEP decoding bug fix.
12719 *) Support INSTALL_PREFIX for package builders, as proposed by
12723 *) New Configure options "threads" and "no-threads". For systems
12724 where the proper compiler options are known (currently Solaris
12725 and Linux), "threads" is the default.
12728 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12731 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12732 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12733 such as /usr/local/bin.
12736 *) "make linux-shared" to build shared libraries.
12737 [Niels Poppe <niels@netbox.org>]
12739 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12742 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12743 extension adding in x509 utility.
12746 *) Remove NOPROTO sections and error code comments.
12749 *) Partial rewrite of the DEF file generator to now parse the ANSI
12753 *) New Configure options --prefix=DIR and --openssldir=DIR.
12756 *) Complete rewrite of the error code script(s). It is all now handled
12757 by one script at the top level which handles error code gathering,
12758 header rewriting and C source file generation. It should be much better
12759 than the old method: it now uses a modified version of Ulf's parser to
12760 read the ANSI prototypes in all header files (thus the old K&R definitions
12761 aren't needed for error creation any more) and do a better job of
12762 translating function codes into names. The old 'ASN1 error code embedded
12763 in a comment' is no longer necessary and it doesn't use .err files which
12764 have now been deleted. Also the error code call doesn't have to appear all
12765 on one line (which resulted in some large lines...).
12768 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12771 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12772 0 (which usually indicates a closed connection), but continue reading.
12775 *) Fix some race conditions.
12778 *) Add support for CRL distribution points extension. Add Certificate
12779 Policies and CRL distribution points documentation.
12782 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12785 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12786 8 of keying material. Merlin has also confirmed interop with this fix
12787 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12788 [Merlin Hughes <merlin@baltimore.ie>]
12790 *) Fix lots of warnings.
12791 [Richard Levitte <levitte@stacken.kth.se>]
12793 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12794 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12795 [Richard Levitte <levitte@stacken.kth.se>]
12797 *) Fix problems with sizeof(long) == 8.
12798 [Andy Polyakov <appro@fy.chalmers.se>]
12800 *) Change functions to ANSI C.
12803 *) Fix typos in error codes.
12804 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12806 *) Remove defunct assembler files from Configure.
12809 *) SPARC v8 assembler BIGNUM implementation.
12810 [Andy Polyakov <appro@fy.chalmers.se>]
12812 *) Support for Certificate Policies extension: both print and set.
12813 Various additions to support the r2i method this uses.
12816 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12817 return a const string when you are expecting an allocated buffer.
12820 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12821 types DirectoryString and DisplayText.
12824 *) Add code to allow r2i extensions to access the configuration database,
12825 add an LHASH database driver and add several ctx helper functions.
12828 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12829 fail when they extended the size of a BIGNUM.
12832 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12833 support typesafe stack.
12836 *) Fix typo in SSL_[gs]et_options().
12837 [Nils Frostberg <nils@medcom.se>]
12839 *) Delete various functions and files that belonged to the (now obsolete)
12840 old X509V3 handling code.
12843 *) New Configure option "rsaref".
12846 *) Don't auto-generate pem.h.
12849 *) Introduce type-safe ASN.1 SETs.
12852 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12853 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12855 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12856 that links with OpenSSL (well at least cause lots of warnings), but fear
12857 not: the conversion is trivial, and it eliminates loads of evil casts. A
12858 few STACKed things have been converted already. Feel free to convert more.
12859 In the fullness of time, I'll do away with the STACK type altogether.
12862 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12863 specified in <certfile> by updating the entry in the index.txt file.
12864 This way one no longer has to edit the index.txt file manually for
12865 revoking a certificate. The -revoke option does the gory details now.
12866 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12868 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12869 `-text' option at all and this way the `-noout -text' combination was
12870 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12871 [Ralf S. Engelschall]
12873 *) Make sure a corresponding plain text error message exists for the
12874 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12875 verify callback function determined that a certificate was revoked.
12876 [Ralf S. Engelschall]
12878 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12879 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12880 all available ciphers including rc5, which was forgotten until now.
12881 In order to let the testing shell script know which algorithms
12882 are available, a new (up to now undocumented) command
12883 "openssl list-cipher-commands" is used.
12886 *) Bugfix: s_client occasionally would sleep in select() when
12887 it should have checked SSL_pending() first.
12890 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12891 the raw DSA values prior to ASN.1 encoding.
12894 *) Tweaks to Configure
12895 [Niels Poppe <niels@netbox.org>]
12897 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12901 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12904 *) New config option to avoid instructions that are illegal on the 80386.
12905 The default code is faster, but requires at least a 486.
12908 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12909 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12910 same as SSL2_VERSION anyway.
12913 *) New "-showcerts" option for s_client.
12916 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12917 application. Various cleanups and fixes.
12920 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12921 modify error routines to work internally. Add error codes and PBE init
12922 to library startup routines.
12925 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12926 packing functions to asn1 and evp. Changed function names and error
12927 codes along the way.
12930 *) PKCS12 integration: and so it begins... First of several patches to
12931 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12932 objects to objects.h
12935 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12936 and display support for Thawte strong extranet extension.
12939 *) Add LinuxPPC support.
12940 [Jeff Dubrule <igor@pobox.org>]
12942 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12943 bn_div_words in alpha.s.
12944 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12946 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12947 OAEP isn't supported when OpenSSL is built with RSAref.
12948 [Ulf Moeller <ulf@fitug.de>]
12950 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12951 so they no longer are missing under -DNOPROTO.
12952 [Soren S. Jorvang <soren@t.dk>]
12955 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12957 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12958 doesn't work when the session is reused. Coming soon!
12961 *) Fix a security hole, that allows sessions to be reused in the wrong
12962 context thus bypassing client cert protection! All software that uses
12963 client certs and session caches in multiple contexts NEEDS PATCHING to
12964 allow session reuse! A fuller solution is in the works.
12965 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12967 *) Some more source tree cleanups (removed obsolete files
12968 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12969 permission on "config" script to be executable) and a fix for the INSTALL
12971 [Ulf Moeller <ulf@fitug.de>]
12973 *) Remove some legacy and erroneous uses of malloc, free instead of
12975 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12977 *) Make rsa_oaep_test return non-zero on error.
12978 [Ulf Moeller <ulf@fitug.de>]
12980 *) Add support for native Solaris shared libraries. Configure
12981 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12982 if someone would make that last step automatic.
12983 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12985 *) ctx_size was not built with the right compiler during "make links". Fixed.
12988 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12989 except NULL ciphers". This means the default cipher list will no longer
12990 enable NULL ciphers. They need to be specifically enabled e.g. with
12991 the string "DEFAULT:eNULL".
12994 *) Fix to RSA private encryption routines: if p < q then it would
12995 occasionally produce an invalid result. This will only happen with
12996 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12999 *) Be less restrictive and allow also `perl util/perlpath.pl
13000 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
13001 because this way one can also use an interpreter named `perl5' (which is
13002 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
13003 installed as `perl').
13004 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13006 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
13007 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13009 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
13010 advapi32.lib to Win32 build and change the pem test comparison
13011 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
13012 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
13013 and crypto/des/ede_cbcm_enc.c.
13016 *) DES quad checksum was broken on big-endian architectures. Fixed.
13019 *) Comment out two functions in bio.h that aren't implemented. Fix up the
13020 Win32 test batch file so it (might) work again. The Win32 test batch file
13021 is horrible: I feel ill....
13024 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
13025 in e_os.h. Audit of header files to check ANSI and non ANSI
13026 sections: 10 functions were absent from non ANSI section and not exported
13027 from Windows DLLs. Fixed up libeay.num for new functions.
13030 *) Make `openssl version' output lines consistent.
13031 [Ralf S. Engelschall]
13033 *) Fix Win32 symbol export lists for BIO functions: Added
13034 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
13035 to ms/libeay{16,32}.def.
13036 [Ralf S. Engelschall]
13038 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
13039 fine under Unix and passes some trivial tests I've now added. But the
13040 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
13041 added to make sure no one expects that this stuff really works in the
13042 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
13043 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
13045 [Ralf S. Engelschall]
13047 *) Fix the generation of two part addresses in perl.
13048 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
13050 *) Add config entry for Linux on MIPS.
13051 [John Tobey <jtobey@channel1.com>]
13053 *) Make links whenever Configure is run, unless we are on Windoze.
13056 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
13057 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
13061 *) Add a useful kludge to allow package maintainers to specify compiler and
13062 other platforms details on the command line without having to patch the
13063 Configure script every time: One now can use ``perl Configure
13064 <id>:<details>'', i.e. platform ids are allowed to have details appended
13065 to them (separated by colons). This is treated as there would be a static
13066 pre-configured entry in Configure's %table under key <id> with value
13067 <details> and ``perl Configure <id>'' is called. So, when you want to
13068 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
13069 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
13070 now, which overrides the FreeBSD-elf entry on-the-fly.
13071 [Ralf S. Engelschall]
13073 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
13076 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
13077 on the `perl Configure ...' command line. This way one can compile
13078 OpenSSL libraries with Position Independent Code (PIC) which is needed
13079 for linking it into DSOs.
13080 [Ralf S. Engelschall]
13082 *) Remarkably, export ciphers were totally broken and no-one had noticed!
13086 *) Cleaned up the LICENSE document: The official contact for any license
13087 questions now is the OpenSSL core team under openssl-core@openssl.org.
13088 And add a paragraph about the dual-license situation to make sure people
13089 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
13090 to the OpenSSL toolkit.
13091 [Ralf S. Engelschall]
13093 *) General source tree makefile cleanups: Made `making xxx in yyy...'
13094 display consistent in the source tree and replaced `/bin/rm' by `rm'.
13095 Additionally cleaned up the `make links' target: Remove unnecessary
13096 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
13097 to speed processing and no longer clutter the display with confusing
13098 stuff. Instead only the actually done links are displayed.
13099 [Ralf S. Engelschall]
13101 *) Permit null encryption ciphersuites, used for authentication only. It used
13102 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
13103 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
13107 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
13108 signed attributes when verifying signatures (this would break them),
13109 the detached data encoding was wrong and public keys obtained using
13110 X509_get_pubkey() weren't freed.
13113 *) Add text documentation for the BUFFER functions. Also added a work around
13114 to a Win95 console bug. This was triggered by the password read stuff: the
13115 last character typed gets carried over to the next fread(). If you were
13116 generating a new cert request using 'req' for example then the last
13117 character of the passphrase would be CR which would then enter the first
13121 *) Added the new `Includes OpenSSL Cryptography Software' button as
13122 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
13123 button and can be used by applications based on OpenSSL to show the
13124 relationship to the OpenSSL project.
13125 [Ralf S. Engelschall]
13127 *) Remove confusing variables in function signatures in files
13128 ssl/ssl_lib.c and ssl/ssl.h.
13129 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13131 *) Don't install bss_file.c under PREFIX/include/
13132 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13134 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
13135 functions that return function pointers and has support for NT specific
13136 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
13137 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
13138 unsigned to signed types: this was killing the Win32 compile.
13141 *) Add new certificate file to stack functions,
13142 SSL_add_dir_cert_subjects_to_stack() and
13143 SSL_add_file_cert_subjects_to_stack(). These largely supplant
13144 SSL_load_client_CA_file(), and can be used to add multiple certs easily
13145 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
13146 This means that Apache-SSL and similar packages don't have to mess around
13147 to add as many CAs as they want to the preferred list.
13150 *) Experiment with doxygen documentation. Currently only partially applied to
13152 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
13153 openssl.doxy as the configuration file.
13156 *) Get rid of remaining C++-style comments which strict C compilers hate.
13157 [Ralf S. Engelschall, pointed out by Carlos Amengual]
13159 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
13160 compiled in by default: it has problems with large keys.
13163 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
13164 DH private keys and/or callback functions which directly correspond to
13165 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
13166 is needed for applications which have to configure certificates on a
13167 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
13169 For the RSA certificate situation is makes no difference, but
13170 for the DSA certificate situation this fixes the "no shared cipher"
13171 problem where the OpenSSL cipher selection procedure failed because the
13172 temporary keys were not overtaken from the context and the API provided
13173 no way to reconfigure them.
13174 The new functions now let applications reconfigure the stuff and they
13175 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
13176 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
13177 non-public-API function ssl_cert_instantiate() is used as a helper
13178 function and also to reduce code redundancy inside ssl_rsa.c.
13179 [Ralf S. Engelschall]
13181 *) Move s_server -dcert and -dkey options out of the undocumented feature
13182 area because they are useful for the DSA situation and should be
13183 recognized by the users.
13184 [Ralf S. Engelschall]
13186 *) Fix the cipher decision scheme for export ciphers: the export bits are
13187 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
13188 SSL_EXP_MASK. So, the original variable has to be used instead of the
13189 already masked variable.
13190 [Richard Levitte <levitte@stacken.kth.se>]
13192 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
13193 [Richard Levitte <levitte@stacken.kth.se>]
13195 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
13196 from `int' to `unsigned int' because it's a length and initialized by
13197 EVP_DigestFinal() which expects an `unsigned int *'.
13198 [Richard Levitte <levitte@stacken.kth.se>]
13200 *) Don't hard-code path to Perl interpreter on shebang line of Configure
13201 script. Instead use the usual Shell->Perl transition trick.
13202 [Ralf S. Engelschall]
13204 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
13205 (in addition to RSA certificates) to match the behaviour of `openssl dsa
13206 -noout -modulus' as it's already the case for `openssl rsa -noout
13207 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
13208 currently the public key is printed (a decision which was already done by
13209 `openssl dsa -modulus' in the past) which serves a similar purpose.
13210 Additionally the NO_RSA no longer completely removes the whole -modulus
13211 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
13213 [Ralf S. Engelschall]
13215 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
13216 BIO. See the source (crypto/evp/bio_ok.c) for more info.
13217 [Arne Ansper <arne@ats.cyber.ee>]
13219 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
13220 to be added. Now both 'req' and 'ca' can use new objects defined in the
13224 *) Add cool BIO that does syslog (or event log on NT).
13225 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
13227 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
13228 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
13229 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
13230 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
13233 *) Add preliminary config info for new extension code.
13236 *) Make RSA_NO_PADDING really use no padding.
13237 [Ulf Moeller <ulf@fitug.de>]
13239 *) Generate errors when private/public key check is done.
13242 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
13243 for some CRL extensions and new objects added.
13246 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
13247 key usage extension and fuller support for authority key id.
13250 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
13251 padding method for RSA, which is recommended for new applications in PKCS
13252 #1 v2.0 (RFC 2437, October 1998).
13253 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
13254 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
13255 against Bleichbacher's attack on RSA.
13256 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
13259 *) Updates to the new SSL compression code
13260 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13262 *) Fix so that the version number in the master secret, when passed
13263 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
13264 (because the server will not accept higher), that the version number
13265 is 0x03,0x01, not 0x03,0x00
13266 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13268 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
13269 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
13270 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
13273 *) Support for RAW extensions where an arbitrary extension can be
13274 created by including its DER encoding. See apps/openssl.cnf for
13278 *) Make sure latest Perl versions don't interpret some generated C array
13279 code as Perl array code in the crypto/err/err_genc.pl script.
13280 [Lars Weber <3weber@informatik.uni-hamburg.de>]
13282 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
13283 not many people have the assembler. Various Win32 compilation fixes and
13284 update to the INSTALL.W32 file with (hopefully) more accurate Win32
13285 build instructions.
13288 *) Modify configure script 'Configure' to automatically create crypto/date.h
13289 file under Win32 and also build pem.h from pem.org. New script
13290 util/mkfiles.pl to create the MINFO file on environments that can't do a
13291 'make files': perl util/mkfiles.pl >MINFO should work.
13294 *) Major rework of DES function declarations, in the pursuit of correctness
13295 and purity. As a result, many evil casts evaporated, and some weirdness,
13296 too. You may find this causes warnings in your code. Zapping your evil
13297 casts will probably fix them. Mostly.
13300 *) Fix for a typo in asn1.h. Bug fix to object creation script
13301 obj_dat.pl. It considered a zero in an object definition to mean
13302 "end of object": none of the objects in objects.h have any zeros
13303 so it wasn't spotted.
13304 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
13306 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
13307 Masking (CBCM). In the absence of test vectors, the best I have been able
13308 to do is check that the decrypt undoes the encrypt, so far. Send me test
13309 vectors if you have them.
13312 *) Correct calculation of key length for export ciphers (too much space was
13313 allocated for null ciphers). This has not been tested!
13316 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
13317 message is now correct (it understands "crypto" and "ssl" on its
13318 command line). There is also now an "update" option. This will update
13319 the util/ssleay.num and util/libeay.num files with any new functions.
13321 perl util/mkdef.pl crypto ssl update
13322 it will update them.
13325 *) Overhauled the Perl interface (perl/*):
13326 - ported BN stuff to OpenSSL's different BN library
13327 - made the perl/ source tree CVS-aware
13328 - renamed the package from SSLeay to OpenSSL (the files still contain
13329 their history because I've copied them in the repository)
13330 - removed obsolete files (the test scripts will be replaced
13331 by better Test::Harness variants in the future)
13332 [Ralf S. Engelschall]
13334 *) First cut for a very conservative source tree cleanup:
13335 1. merge various obsolete readme texts into doc/ssleay.txt
13336 where we collect the old documents and readme texts.
13337 2. remove the first part of files where I'm already sure that we no
13338 longer need them because of three reasons: either they are just temporary
13339 files which were left by Eric or they are preserved original files where
13340 I've verified that the diff is also available in the CVS via "cvs diff
13341 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
13342 the crypto/md/ stuff).
13343 [Ralf S. Engelschall]
13345 *) More extension code. Incomplete support for subject and issuer alt
13346 name, issuer and authority key id. Change the i2v function parameters
13347 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
13348 what that's for :-) Fix to ASN1 macro which messed up
13349 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
13352 *) Preliminary support for ENUMERATED type. This is largely copied from the
13356 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
13357 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13359 *) Make sure `make rehash' target really finds the `openssl' program.
13360 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13362 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
13363 like to hear about it if this slows down other processors.
13366 *) Add CygWin32 platform information to Configure script.
13367 [Alan Batie <batie@aahz.jf.intel.com>]
13369 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
13370 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13372 *) New program nseq to manipulate netscape certificate sequences
13375 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13379 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13380 but the BN code had some problems that would cause failures when
13381 doing certificate verification and some other functions.
13382 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13384 *) Add ASN1 and PEM code to support netscape certificate sequences.
13387 *) Add ASN1 and PEM code to support netscape certificate sequences.
13390 *) Add several PKIX and private extended key usage OIDs.
13393 *) Modify the 'ca' program to handle the new extension code. Modify
13394 openssl.cnf for new extension format, add comments.
13397 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13398 and add a sample to openssl.cnf so req -x509 now adds appropriate
13402 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13403 error code, add initial support to X509_print() and x509 application.
13406 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13407 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13408 stuff is currently isolated and isn't even compiled yet.
13411 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13412 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13413 Removed the versions check from X509 routines when loading extensions:
13414 this allows certain broken certificates that don't set the version
13415 properly to be processed.
13418 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13419 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13420 can still be regenerated with "make depend".
13423 *) Spelling mistake in C version of CAST-128.
13424 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13426 *) Changes to the error generation code. The perl script err-code.pl
13427 now reads in the old error codes and retains the old numbers, only
13428 adding new ones if necessary. It also only changes the .err files if new
13429 codes are added. The makefiles have been modified to only insert errors
13430 when needed (to avoid needlessly modifying header files). This is done
13431 by only inserting errors if the .err file is newer than the auto generated
13432 C file. To rebuild all the error codes from scratch (the old behaviour)
13433 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13434 or delete all the .err files.
13437 *) CAST-128 was incorrectly implemented for short keys. The C version has
13438 been fixed, but is untested. The assembler versions are also fixed, but
13439 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13440 to regenerate it if needed.
13441 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13442 Hagino <itojun@kame.net>]
13444 *) File was opened incorrectly in randfile.c.
13445 [Ulf Möller <ulf@fitug.de>]
13447 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13448 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13449 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13450 al: it's just almost always a UTCTime. Note this patch adds new error
13451 codes so do a "make errors" if there are problems.
13454 *) Correct Linux 1 recognition in config.
13455 [Ulf Möller <ulf@fitug.de>]
13457 *) Remove pointless MD5 hash when using DSA keys in ca.
13458 [Anonymous <nobody@replay.com>]
13460 *) Generate an error if given an empty string as a cert directory. Also
13461 generate an error if handed NULL (previously returned 0 to indicate an
13462 error, but didn't set one).
13463 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13465 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13468 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13469 parameters. This was causing a warning which killed off the Win32 compile.
13472 *) Remove C++ style comments from crypto/bn/bn_local.h.
13473 [Neil Costigan <neil.costigan@celocom.com>]
13475 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13476 based on a text string, looking up short and long names and finally
13477 "dot" format. The "dot" format stuff didn't work. Added new function
13478 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13479 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13480 OID is not part of the table.
13483 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13484 X509_LOOKUP_by_alias().
13487 *) Sort openssl functions by name.
13490 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13491 encryption from sample DSA keys (in case anyone is interested the password
13495 *) Make _all_ *_free functions accept a NULL pointer.
13496 [Frans Heymans <fheymans@isaserver.be>]
13498 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13500 [Anonymous <nobody@replay.com>]
13502 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13503 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13505 *) Don't blow it for numeric -newkey arguments to apps/req.
13506 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13508 *) Temp key "for export" tests were wrong in s3_srvr.c.
13509 [Anonymous <nobody@replay.com>]
13511 *) Add prototype for temp key callback functions
13512 SSL_CTX_set_tmp_{rsa,dh}_callback().
13515 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13516 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13519 *) X509_name_add_entry() freed the wrong thing after an error.
13520 [Arne Ansper <arne@ats.cyber.ee>]
13522 *) rsa_eay.c would attempt to free a NULL context.
13523 [Arne Ansper <arne@ats.cyber.ee>]
13525 *) BIO_s_socket() had a broken should_retry() on Windoze.
13526 [Arne Ansper <arne@ats.cyber.ee>]
13528 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13529 [Arne Ansper <arne@ats.cyber.ee>]
13531 *) Make sure the already existing X509_STORE->depth variable is initialized
13532 in X509_STORE_new(), but document the fact that this variable is still
13533 unused in the certificate verification process.
13534 [Ralf S. Engelschall]
13536 *) Fix the various library and apps files to free up pkeys obtained from
13537 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13540 *) Fix reference counting in X509_PUBKEY_get(). This makes
13541 demos/maurice/example2.c work, amongst others, probably.
13542 [Steve Henson and Ben Laurie]
13544 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13545 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13546 are no longer created. This way we have a single and consistent command
13547 line interface `openssl <command>', similar to `cvs <command>'.
13548 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13550 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13551 BIT STRING wrapper always have zero unused bits.
13554 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13557 *) Make the top-level INSTALL documentation easier to understand.
13560 *) Makefiles updated to exit if an error occurs in a sub-directory
13561 make (including if user presses ^C) [Paul Sutton]
13563 *) Make Montgomery context stuff explicit in RSA data structure.
13566 *) Fix build order of pem and err to allow for generated pem.h.
13569 *) Fix renumbering bug in X509_NAME_delete_entry().
13572 *) Enhanced the err-ins.pl script so it makes the error library number
13573 global and can add a library name. This is needed for external ASN1 and
13574 other error libraries.
13577 *) Fixed sk_insert which never worked properly.
13580 *) Fix ASN1 macros so they can handle indefinite length constructed
13581 EXPLICIT tags. Some non standard certificates use these: they can now
13585 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13586 into a single doc/ssleay.txt bundle. This way the information is still
13587 preserved but no longer messes up this directory. Now it's new room for
13588 the new set of documentation files.
13589 [Ralf S. Engelschall]
13591 *) SETs were incorrectly DER encoded. This was a major pain, because they
13592 shared code with SEQUENCEs, which aren't coded the same. This means that
13593 almost everything to do with SETs or SEQUENCEs has either changed name or
13594 number of arguments.
13595 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13597 *) Fix test data to work with the above.
13600 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13601 was already fixed by Eric for 0.9.1 it seems.
13602 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13604 *) Autodetect FreeBSD3.
13607 *) Fix various bugs in Configure. This affects the following platforms:
13611 unixware-2.0-pentium
13615 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13616 before they are needed.
13619 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13623 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13625 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13626 changed SSLeay to OpenSSL in version strings.
13627 [Ralf S. Engelschall]
13629 *) Some fixups to the top-level documents.
13632 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13633 because the symlink to include/ was missing.
13634 [Ralf S. Engelschall]
13636 *) Incorporated the popular no-RSA/DSA-only patches
13637 which allow to compile a RSA-free SSLeay.
13638 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13640 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13641 when "ssleay" is still not found.
13642 [Ralf S. Engelschall]
13644 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13645 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13647 *) Updated the README file.
13648 [Ralf S. Engelschall]
13650 *) Added various .cvsignore files in the CVS repository subdirs
13651 to make a "cvs update" really silent.
13652 [Ralf S. Engelschall]
13654 *) Recompiled the error-definition header files and added
13655 missing symbols to the Win32 linker tables.
13656 [Ralf S. Engelschall]
13658 *) Cleaned up the top-level documents;
13659 o new files: CHANGES and LICENSE
13660 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13661 o merged COPYRIGHT into LICENSE
13662 o removed obsolete TODO file
13663 o renamed MICROSOFT to INSTALL.W32
13664 [Ralf S. Engelschall]
13666 *) Removed dummy files from the 0.9.1b source tree:
13667 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13668 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13669 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13670 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13671 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13672 [Ralf S. Engelschall]
13674 *) Added various platform portability fixes.
13677 *) The Genesis of the OpenSSL rpject:
13678 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13679 Young and Tim J. Hudson created while they were working for C2Net until
13681 [The OpenSSL Project]
13684 Changes between 0.9.0b and 0.9.1b [not released]
13686 *) Updated a few CA certificates under certs/
13689 *) Changed some BIGNUM api stuff.
13692 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13693 DGUX x86, Linux Alpha, etc.
13696 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13697 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13701 *) Add -strparse option to asn1pars program which parses nested
13703 [Dr Stephen Henson <shenson@bigfoot.com>]
13705 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13708 *) DSA fix for "ca" program.
13711 *) Added "-genkey" option to "dsaparam" program.
13714 *) Added RIPE MD160 (rmd160) message digest.
13717 *) Added -a (all) option to "ssleay version" command.
13720 *) Added PLATFORM define which is the id given to Configure.
13723 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13726 *) Extended the ASN.1 parser routines.
13729 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13732 *) Added a BN_CTX to the BN library.
13735 *) Fixed the weak key values in DES library
13738 *) Changed API in EVP library for cipher aliases.
13741 *) Added support for RC2/64bit cipher.
13744 *) Converted the lhash library to the crypto/mem.c functions.
13747 *) Added more recognized ASN.1 object ids.
13750 *) Added more RSA padding checks for SSL/TLS.
13753 *) Added BIO proxy/filter functionality.
13756 *) Added extra_certs to SSL_CTX which can be used
13757 send extra CA certificates to the client in the CA cert chain sending
13758 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13761 *) Now Fortezza is denied in the authentication phase because
13762 this is key exchange mechanism is not supported by SSLeay at all.
13765 *) Additional PKCS1 checks.
13768 *) Support the string "TLSv1" for all TLS v1 ciphers.
13771 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13772 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13775 *) Fixed a few memory leaks.
13778 *) Fixed various code and comment typos.
13781 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13782 bytes sent in the client random.
13783 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob