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 1.1.1t and 1.1.1u [xx XXX xxxx]
12 *) Fixed an issue where invalid certificate policies in leaf certificates are
13 silently ignored by OpenSSL and other certificate policy checks are skipped
14 for that certificate. A malicious CA could use this to deliberately assert
15 invalid certificate policies in order to circumvent policy checking on the
16 certificate altogether. (CVE-2023-0465)
19 *) Limited the number of nodes created in a policy tree to mitigate
20 against CVE-2023-0464. The default limit is set to 1000 nodes, which
21 should be sufficient for most installations. If required, the limit
22 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
23 time define to a desired maximum number of nodes or zero to allow
24 unlimited growth. (CVE-2023-0464)
27 Changes between 1.1.1s and 1.1.1t [7 Feb 2023]
29 *) Fixed X.400 address type confusion in X.509 GeneralName.
31 There is a type confusion vulnerability relating to X.400 address processing
32 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
33 but subsequently interpreted by GENERAL_NAME_cmp as an ASN1_TYPE. This
34 vulnerability may allow an attacker who can provide a certificate chain and
35 CRL (neither of which need have a valid signature) to pass arbitrary
36 pointers to a memcmp call, creating a possible read primitive, subject to
37 some constraints. Refer to the advisory for more information. Thanks to
38 David Benjamin for discovering this issue. (CVE-2023-0286)
40 This issue has been fixed by changing the public header file definition of
41 GENERAL_NAME so that x400Address reflects the implementation. It was not
42 possible for any existing application to successfully use the existing
43 definition; however, if any application references the x400Address field
44 (e.g. in dead code), note that the type of this field has changed. There is
48 *) Fixed Use-after-free following BIO_new_NDEF.
50 The public API function BIO_new_NDEF is a helper function used for
51 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
52 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
53 be called directly by end user applications.
55 The function receives a BIO from the caller, prepends a new BIO_f_asn1
56 filter BIO onto the front of it to form a BIO chain, and then returns
57 the new head of the BIO chain to the caller. Under certain conditions,
58 for example if a CMS recipient public key is invalid, the new filter BIO
59 is freed and the function returns a NULL result indicating a failure.
60 However, in this case, the BIO chain is not properly cleaned up and the
61 BIO passed by the caller still retains internal pointers to the previously
62 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
63 then a use-after-free will occur. This will most likely result in a crash.
65 [Viktor Dukhovni, Matt Caswell]
67 *) Fixed Double free after calling PEM_read_bio_ex.
69 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
70 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
71 data. If the function succeeds then the "name_out", "header" and "data"
72 arguments are populated with pointers to buffers containing the relevant
73 decoded data. The caller is responsible for freeing those buffers. It is
74 possible to construct a PEM file that results in 0 bytes of payload data.
75 In this case PEM_read_bio_ex() will return a failure code but will populate
76 the header argument with a pointer to a buffer that has already been freed.
77 If the caller also frees this buffer then a double free will occur. This
78 will most likely lead to a crash.
80 The functions PEM_read_bio() and PEM_read() are simple wrappers around
81 PEM_read_bio_ex() and therefore these functions are also directly affected.
83 These functions are also called indirectly by a number of other OpenSSL
84 functions including PEM_X509_INFO_read_bio_ex() and
85 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
86 internal uses of these functions are not vulnerable because the caller does
87 not free the header argument if PEM_read_bio_ex() returns a failure code.
89 [Kurt Roeckx, Matt Caswell]
91 *) Fixed Timing Oracle in RSA Decryption.
93 A timing based side channel exists in the OpenSSL RSA Decryption
94 implementation which could be sufficient to recover a plaintext across
95 a network in a Bleichenbacher style attack. To achieve a successful
96 decryption an attacker would have to be able to send a very large number
97 of trial messages for decryption. The vulnerability affects all RSA padding
98 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
100 [Dmitry Belyavsky, Hubert Kario]
102 Changes between 1.1.1r and 1.1.1s [1 Nov 2022]
104 *) Fixed a regression introduced in 1.1.1r version not refreshing the
105 certificate data to be signed before signing the certificate.
108 Changes between 1.1.1q and 1.1.1r [11 Oct 2022]
110 *) Fixed the linux-mips64 Configure target which was missing the
111 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
115 *) Fixed a strict aliasing problem in bn_nist. Clang-14 optimisation was
116 causing incorrect results in some cases as a result.
119 *) Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
120 report correct results in some cases
123 *) Fixed a regression introduced in 1.1.1o for re-signing certificates with
127 *) Added the loongarch64 target
130 *) Fixed a DRBG seed propagation thread safety issue
133 *) Fixed a memory leak in tls13_generate_secret
136 *) Fixed reported performance degradation on aarch64. Restored the
137 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
138 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
139 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
140 The new algorithm is still used for 32 bit targets.
143 *) Added a missing header for memcmp that caused compilation failure on some
147 Changes between 1.1.1p and 1.1.1q [5 Jul 2022]
149 *) AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
150 implementation would not encrypt the entirety of the data under some
151 circumstances. This could reveal sixteen bytes of data that was
152 preexisting in the memory that wasn't written. In the special case of
153 "in place" encryption, sixteen bytes of the plaintext would be revealed.
155 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
156 they are both unaffected.
158 [Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño]
160 Changes between 1.1.1o and 1.1.1p [21 Jun 2022]
162 *) In addition to the c_rehash shell command injection identified in
163 CVE-2022-1292, further bugs where the c_rehash script does not
164 properly sanitise shell metacharacters to prevent command injection have been
167 When the CVE-2022-1292 was fixed it was not discovered that there
168 are other places in the script where the file names of certificates
169 being hashed were possibly passed to a command executed through the shell.
171 This script is distributed by some operating systems in a manner where
172 it is automatically executed. On such operating systems, an attacker
173 could execute arbitrary commands with the privileges of the script.
175 Use of the c_rehash script is considered obsolete and should be replaced
176 by the OpenSSL rehash command line tool.
178 [Daniel Fiala, Tomáš Mráz]
180 *) When OpenSSL TLS client is connecting without any supported elliptic
181 curves and TLS-1.3 protocol is disabled the connection will no longer fail
182 if a ciphersuite that does not use a key exchange based on elliptic
183 curves can be negotiated.
186 Changes between 1.1.1n and 1.1.1o [3 May 2022]
188 *) Fixed a bug in the c_rehash script which was not properly sanitising shell
189 metacharacters to prevent command injection. This script is distributed
190 by some operating systems in a manner where it is automatically executed.
191 On such operating systems, an attacker could execute arbitrary commands
192 with the privileges of the script.
194 Use of the c_rehash script is considered obsolete and should be replaced
195 by the OpenSSL rehash command line tool.
199 Changes between 1.1.1m and 1.1.1n [15 Mar 2022]
201 *) Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
202 for non-prime moduli.
204 Internally this function is used when parsing certificates that contain
205 elliptic curve public keys in compressed form or explicit elliptic curve
206 parameters with a base point encoded in compressed form.
208 It is possible to trigger the infinite loop by crafting a certificate that
209 has invalid explicit curve parameters.
211 Since certificate parsing happens prior to verification of the certificate
212 signature, any process that parses an externally supplied certificate may
213 thus be subject to a denial of service attack. The infinite loop can also
214 be reached when parsing crafted private keys as they can contain explicit
215 elliptic curve parameters.
217 Thus vulnerable situations include:
219 - TLS clients consuming server certificates
220 - TLS servers consuming client certificates
221 - Hosting providers taking certificates or private keys from customers
222 - Certificate authorities parsing certification requests from subscribers
223 - Anything else which parses ASN.1 elliptic curve parameters
225 Also any other applications that use the BN_mod_sqrt() where the attacker
226 can control the parameter values are vulnerable to this DoS issue.
230 *) Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
231 to the list of ciphersuites providing Perfect Forward Secrecy as
232 required by SECLEVEL >= 3.
234 [Dmitry Belyavskiy, Nicola Tuveri]
236 Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
238 *) Avoid loading of a dynamic engine twice.
242 *) Fixed building on Debian with kfreebsd kernels
246 *) Prioritise DANE TLSA issuer certs over peer certs
250 *) Fixed random API for MacOS prior to 10.12
252 These MacOS versions don't support the CommonCrypto APIs
256 Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
258 *) Fixed an SM2 Decryption Buffer Overflow.
260 In order to decrypt SM2 encrypted data an application is expected to call the
261 API function EVP_PKEY_decrypt(). Typically an application will call this
262 function twice. The first time, on entry, the "out" parameter can be NULL and,
263 on exit, the "outlen" parameter is populated with the buffer size required to
264 hold the decrypted plaintext. The application can then allocate a sufficiently
265 sized buffer and call EVP_PKEY_decrypt() again, but this time passing a non-NULL
266 value for the "out" parameter.
268 A bug in the implementation of the SM2 decryption code means that the
269 calculation of the buffer size required to hold the plaintext returned by the
270 first call to EVP_PKEY_decrypt() can be smaller than the actual size required by
271 the second call. This can lead to a buffer overflow when EVP_PKEY_decrypt() is
272 called by the application a second time with a buffer that is too small.
274 A malicious attacker who is able present SM2 content for decryption to an
275 application could cause attacker chosen data to overflow the buffer by up to a
276 maximum of 62 bytes altering the contents of other data held after the
277 buffer, possibly changing application behaviour or causing the application to
278 crash. The location of the buffer is application dependent but is typically
283 *) Fixed various read buffer overruns processing ASN.1 strings
285 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
286 structure which contains a buffer holding the string data and a field holding
287 the buffer length. This contrasts with normal C strings which are repesented as
288 a buffer for the string data which is terminated with a NUL (0) byte.
290 Although not a strict requirement, ASN.1 strings that are parsed using OpenSSL's
291 own "d2i" functions (and other similar parsing functions) as well as any string
292 whose value has been set with the ASN1_STRING_set() function will additionally
293 NUL terminate the byte array in the ASN1_STRING structure.
295 However, it is possible for applications to directly construct valid ASN1_STRING
296 structures which do not NUL terminate the byte array by directly setting the
297 "data" and "length" fields in the ASN1_STRING array. This can also happen by
298 using the ASN1_STRING_set0() function.
300 Numerous OpenSSL functions that print ASN.1 data have been found to assume that
301 the ASN1_STRING byte array will be NUL terminated, even though this is not
302 guaranteed for strings that have been directly constructed. Where an application
303 requests an ASN.1 structure to be printed, and where that ASN.1 structure
304 contains ASN1_STRINGs that have been directly constructed by the application
305 without NUL terminating the "data" field, then a read buffer overrun can occur.
307 The same thing can also occur during name constraints processing of certificates
308 (for example if a certificate has been directly constructed by the application
309 instead of loading it via the OpenSSL parsing functions, and the certificate
310 contains non NUL terminated ASN1_STRING structures). It can also occur in the
311 X509_get1_email(), X509_REQ_get1_email() and X509_get1_ocsp() functions.
313 If a malicious actor can cause an application to directly construct an
314 ASN1_STRING and then process it through one of the affected OpenSSL functions
315 then this issue could be hit. This might result in a crash (causing a Denial of
316 Service attack). It could also result in the disclosure of private memory
317 contents (such as private keys, or sensitive plaintext).
321 Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
323 *) Fixed a problem with verifying a certificate chain when using the
324 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks
325 of the certificates present in a certificate chain. It is not set by
328 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
329 the chain that have explicitly encoded elliptic curve parameters was added
330 as an additional strict check.
332 An error in the implementation of this check meant that the result of a
333 previous check to confirm that certificates in the chain are valid CA
334 certificates was overwritten. This effectively bypasses the check
335 that non-CA certificates must not be able to issue other certificates.
337 If a "purpose" has been configured then there is a subsequent opportunity
338 for checks that the certificate is a valid CA. All of the named "purpose"
339 values implemented in libcrypto perform this check. Therefore, where
340 a purpose is set the certificate chain will still be rejected even when the
341 strict flag has been used. A purpose is set by default in libssl client and
342 server certificate verification routines, but it can be overridden or
343 removed by an application.
345 In order to be affected, an application must explicitly set the
346 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
347 for the certificate verification or, in the case of TLS client or server
348 applications, override the default purpose.
352 *) Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
353 crafted renegotiation ClientHello message from a client. If a TLSv1.2
354 renegotiation ClientHello omits the signature_algorithms extension (where
355 it was present in the initial ClientHello), but includes a
356 signature_algorithms_cert extension then a NULL pointer dereference will
357 result, leading to a crash and a denial of service attack.
359 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
360 (which is the default configuration). OpenSSL TLS clients are not impacted
363 [Peter Kästle and Samuel Sapalski]
365 Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
367 *) Fixed the X509_issuer_and_serial_hash() function. It attempts to
368 create a unique hash value based on the issuer and serial number data
369 contained within an X509 certificate. However it was failing to correctly
370 handle any errors that may occur while parsing the issuer field (which might
371 occur if the issuer field is maliciously constructed). This may subsequently
372 result in a NULL pointer deref and a crash leading to a potential denial of
377 *) Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
378 padding mode to correctly check for rollback attacks. This is considered a
379 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
383 *) Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
384 functions. Previously they could overflow the output length argument in some
385 cases where the input length is close to the maximum permissable length for
386 an integer on the platform. In such cases the return value from the function
387 call would be 1 (indicating success), but the output length value would be
388 negative. This could cause applications to behave incorrectly or crash.
392 *) Fixed SRP_Calc_client_key so that it runs in constant time. The previous
393 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
394 could be exploited in a side channel attack to recover the password. Since
395 the attack is local host only this is outside of the current OpenSSL
396 threat model and therefore no CVE is assigned.
398 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
402 Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
404 *) Fixed NULL pointer deref in the GENERAL_NAME_cmp function
405 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
406 If an attacker can control both items being compared then this could lead
407 to a possible denial of service attack. OpenSSL itself uses the
408 GENERAL_NAME_cmp function for two purposes:
409 1) Comparing CRL distribution point names between an available CRL and a
410 CRL distribution point embedded in an X509 certificate
411 2) When verifying that a timestamp response token signer matches the
412 timestamp authority name (exposed via the API functions
413 TS_RESP_verify_response and TS_RESP_verify_token)
417 *) Add support for Apple Silicon M1 Macs with the darwin64-arm64-cc target.
420 *) The security callback, which can be customised by application code, supports
421 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
422 in the "other" parameter. In most places this is what is passed. All these
423 places occur server side. However there was one client side call of this
424 security operation and it passed a DH object instead. This is incorrect
425 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
426 of the other locations. Therefore this client side call has been changed to
427 pass an EVP_PKEY instead.
430 *) In 1.1.1h, an expired trusted (root) certificate was not anymore rejected
431 when validating a certificate path. This check is restored in 1.1.1i.
434 Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
436 *) Certificates with explicit curve parameters are now disallowed in
437 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
440 *) The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
441 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
442 conversely, silently ignore DTLS protocol version bounds when configuring
443 TLS-based contexts. The commands can be repeated to set bounds of both
444 types. The same applies with the corresponding "min_protocol" and
445 "max_protocol" command-line switches, in case some application uses both TLS
448 SSL_CTX instances that are created for a fixed protocol version (e.g.
449 TLSv1_server_method()) also silently ignore version bounds. Previously
450 attempts to apply bounds to these protocol versions would result in an
451 error. Now only the "version-flexible" SSL_CTX instances are subject to
452 limits in configuration files in command-line options.
455 *) Handshake now fails if Extended Master Secret extension is dropped
459 *) Accidentally, an expired trusted (root) certificate is not anymore rejected
460 when validating a certificate path.
463 *) The Oracle Developer Studio compiler will start reporting deprecated APIs
465 Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
467 *) Fixed segmentation fault in SSL_check_chain()
468 Server or client applications that call the SSL_check_chain() function
469 during or after a TLS 1.3 handshake may crash due to a NULL pointer
470 dereference as a result of incorrect handling of the
471 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
472 or unrecognised signature algorithm is received from the peer. This could
473 be exploited by a malicious peer in a Denial of Service attack.
477 *) Added AES consttime code for no-asm configurations
478 an optional constant time support for AES was added
479 when building openssl for no-asm.
480 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
481 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
482 At this time this feature is by default disabled.
483 It will be enabled by default in 3.0.
486 Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
488 *) Revert the change of EOF detection while reading in libssl to avoid
489 regressions in applications depending on the current way of reporting
490 the EOF. As the existing method is not fully accurate the change to
491 reporting the EOF via SSL_ERROR_SSL is kept on the current development
492 branch and will be present in the 3.0 release.
495 *) Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
496 when primes for RSA keys are computed.
497 Since we previously always generated primes == 2 (mod 3) for RSA keys,
498 the 2-prime and 3-prime RSA modules were easy to distinguish, since
499 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
500 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
501 This avoids possible fingerprinting of newly generated RSA modules.
504 Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
505 *) Properly detect EOF while reading in libssl. Previously if we hit an EOF
506 while reading in libssl then we would report an error back to the
507 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
508 an error to the stack (which means we instead return SSL_ERROR_SSL) and
509 therefore give a hint as to what went wrong.
512 *) Check that ed25519 and ed448 are allowed by the security level. Previously
513 signature algorithms not using an MD were not being checked that they were
514 allowed by the security level.
517 *) Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
518 was not quite right. The behaviour was not consistent between resumption
519 and normal handshakes, and also not quite consistent with historical
520 behaviour. The behaviour in various scenarios has been clarified and
521 it has been updated to make it match historical behaviour as closely as
525 *) [VMS only] The header files that the VMS compilers include automatically,
526 __DECC_INCLUDE_PROLOGUE.H and __DECC_INCLUDE_EPILOGUE.H, use pragmas that
527 the C++ compiler doesn't understand. This is a shortcoming in the
528 compiler, but can be worked around with __cplusplus guards.
530 C++ applications that use OpenSSL libraries must be compiled using the
531 qualifier '/NAMES=(AS_IS,SHORTENED)' to be able to use all the OpenSSL
532 functions. Otherwise, only functions with symbols of less than 31
533 characters can be used, as the linker will not be able to successfully
534 resolve symbols with longer names.
537 *) Corrected the documentation of the return values from the EVP_DigestSign*
538 set of functions. The documentation mentioned negative values for some
539 errors, but this was never the case, so the mention of negative values
542 Code that followed the documentation and thereby check with something
543 like 'EVP_DigestSignInit(...) <= 0' will continue to work undisturbed.
546 *) Fixed an an overflow bug in the x64_64 Montgomery squaring procedure
547 used in exponentiation with 512-bit moduli. No EC algorithms are
548 affected. Analysis suggests that attacks against 2-prime RSA1024,
549 3-prime RSA1536, and DSA1024 as a result of this defect would be very
550 difficult to perform and are not believed likely. Attacks against DH512
551 are considered just feasible. However, for an attack the target would
552 have to re-use the DH512 private key, which is not recommended anyway.
553 Also applications directly using the low level API BN_mod_exp may be
554 affected if they use BN_FLG_CONSTTIME.
558 *) Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
559 The presence of this system service is determined at run-time.
562 *) Added newline escaping functionality to a filename when using openssl dgst.
563 This output format is to replicate the output format found in the '*sum'
564 checksum programs. This aims to preserve backward compatibility.
565 [Matt Eaton, Richard Levitte, and Paul Dale]
567 *) Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
571 Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
573 *) Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
574 number generator (RNG). This was intended to include protection in the
575 event of a fork() system call in order to ensure that the parent and child
576 processes did not share the same RNG state. However this protection was not
577 being used in the default case.
579 A partial mitigation for this issue is that the output from a high
580 precision timer is mixed into the RNG state so the likelihood of a parent
581 and child process sharing state is significantly reduced.
583 If an application already calls OPENSSL_init_crypto() explicitly using
584 OPENSSL_INIT_ATFORK then this problem does not occur at all.
586 [Matthias St. Pierre]
588 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is
589 used even when parsing explicit parameters, when loading a serialized key
590 or calling `EC_GROUP_new_from_ecpkparameters()`/
591 `EC_GROUP_new_from_ecparameters()`.
592 This prevents bypass of security hardening and performance gains,
593 especially for curves with specialized EC_METHODs.
594 By default, if a key encoded with explicit parameters is loaded and later
595 serialized, the output is still encoded with explicit parameters, even if
596 internally a "named" EC_GROUP is used for computation.
599 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before
600 this change, EC_GROUP_set_generator would accept order and/or cofactor as
601 NULL. After this change, only the cofactor parameter can be NULL. It also
602 does some minimal sanity checks on the passed order.
606 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
607 An attack is simple, if the first CMS_recipientInfo is valid but the
608 second CMS_recipientInfo is chosen ciphertext. If the second
609 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
610 encryption key will be replaced by garbage, and the message cannot be
611 decoded, but if the RSA decryption fails, the correct encryption key is
612 used and the recipient will not notice the attack.
613 As a work around for this potential attack the length of the decrypted
614 key must be equal to the cipher default key length, in case the
615 certifiate is not given and all recipientInfo are tried out.
616 The old behaviour can be re-enabled in the CMS code by setting the
617 CMS_DEBUG_DECRYPT flag.
621 *) Early start up entropy quality from the DEVRANDOM seed source has been
622 improved for older Linux systems. The RAND subsystem will wait for
623 /dev/random to be producing output before seeding from /dev/urandom.
624 The seeded state is stored for future library initialisations using
625 a system global shared memory segment. The shared memory identifier
626 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
627 the desired value. The default identifier is 114.
630 *) Correct the extended master secret constant on EBCDIC systems. Without this
631 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
632 negotiate EMS will fail. Unfortunately this also means that TLS connections
633 between EBCDIC systems with this fix, and EBCDIC systems without this
634 fix will fail if they negotiate EMS.
637 *) Use Windows installation paths in the mingw builds
639 Mingw isn't a POSIX environment per se, which means that Windows
640 paths should be used for installation.
644 *) Changed DH_check to accept parameters with order q and 2q subgroups.
645 With order 2q subgroups the bit 0 of the private key is not secret
646 but DH_generate_key works around that by clearing bit 0 of the
647 private key for those. This avoids leaking bit 0 of the private key.
650 *) Significantly reduce secure memory usage by the randomness pools.
653 *) Revert the DEVRANDOM_WAIT feature for Linux systems
655 The DEVRANDOM_WAIT feature added a select() call to wait for the
656 /dev/random device to become readable before reading from the
659 It turned out that this change had negative side effects on
660 performance which were not acceptable. After some discussion it
661 was decided to revert this feature and leave it up to the OS
662 resp. the platform maintainer to ensure a proper initialization
663 during early boot time.
664 [Matthias St. Pierre]
666 Changes between 1.1.1b and 1.1.1c [28 May 2019]
668 *) Add build tests for C++. These are generated files that only do one
669 thing, to include one public OpenSSL head file each. This tests that
670 the public header files can be usefully included in a C++ application.
672 This test isn't enabled by default. It can be enabled with the option
673 'enable-buildtest-c++'.
676 *) Enable SHA3 pre-hashing for ECDSA and DSA.
679 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
680 This changes the size when using the genpkey app when no size is given. It
681 fixes an omission in earlier changes that changed all RSA, DSA and DH
682 generation apps to use 2048 bits by default.
685 *) Reorganize the manual pages to consistently have RETURN VALUES,
686 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
687 util/fix-doc-nits accordingly.
688 [Paul Yang, Joshua Lock]
690 *) Add the missing accessor EVP_PKEY_get0_engine()
693 *) Have apps like 's_client' and 's_server' output the signature scheme
694 along with other cipher suite parameters when debugging.
697 *) Make OPENSSL_config() error agnostic again.
700 *) Do the error handling in RSA decryption constant time.
703 *) Prevent over long nonces in ChaCha20-Poly1305.
705 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
706 for every encryption operation. RFC 7539 specifies that the nonce value
707 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
708 and front pads the nonce with 0 bytes if it is less than 12
709 bytes. However it also incorrectly allows a nonce to be set of up to 16
710 bytes. In this case only the last 12 bytes are significant and any
711 additional leading bytes are ignored.
713 It is a requirement of using this cipher that nonce values are
714 unique. Messages encrypted using a reused nonce value are susceptible to
715 serious confidentiality and integrity attacks. If an application changes
716 the default nonce length to be longer than 12 bytes and then makes a
717 change to the leading bytes of the nonce expecting the new value to be a
718 new unique nonce then such an application could inadvertently encrypt
719 messages with a reused nonce.
721 Additionally the ignored bytes in a long nonce are not covered by the
722 integrity guarantee of this cipher. Any application that relies on the
723 integrity of these ignored leading bytes of a long nonce may be further
724 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
725 is safe because no such use sets such a long nonce value. However user
726 applications that use this cipher directly and set a non-default nonce
727 length to be longer than 12 bytes may be vulnerable.
729 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
734 *) Add DEVRANDOM_WAIT feature for Linux systems
736 On older Linux systems where the getrandom() system call is not available,
737 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
738 Contrary to getrandom(), the /dev/urandom device will not block during
739 early boot when the kernel CSPRNG has not been seeded yet.
741 To mitigate this known weakness, use select() to wait for /dev/random to
742 become readable before reading from /dev/urandom.
744 *) Ensure that SM2 only uses SM3 as digest algorithm
747 Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
749 *) Added SCA hardening for modular field inversion in EC_GROUP through
750 a new dedicated field_inv() pointer in EC_METHOD.
751 This also addresses a leakage affecting conversions from projective
752 to affine coordinates.
753 [Billy Bob Brumley, Nicola Tuveri]
755 *) Change the info callback signals for the start and end of a post-handshake
756 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
757 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
758 confused by this and assume that a TLSv1.2 renegotiation has started. This
759 can break KeyUpdate handling. Instead we no longer signal the start and end
760 of a post handshake message exchange (although the messages themselves are
761 still signalled). This could break some applications that were expecting
762 the old signals. However without this KeyUpdate is not usable for many
766 *) Fix a bug in the computation of the endpoint-pair shared secret used
767 by DTLS over SCTP. This breaks interoperability with older versions
768 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
769 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
770 interoperability with such broken implementations. However, enabling
771 this switch breaks interoperability with correct implementations.
773 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
774 re-used X509_PUBKEY object if the second PUBKEY is malformed.
777 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
780 *) Remove the 'dist' target and add a tarball building script. The
781 'dist' target has fallen out of use, and it shouldn't be
782 necessary to configure just to create a source distribution.
785 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
787 *) Timing vulnerability in DSA signature generation
789 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
790 timing side channel attack. An attacker could use variations in the signing
791 algorithm to recover the private key.
793 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
797 *) Timing vulnerability in ECDSA signature generation
799 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
800 timing side channel attack. An attacker could use variations in the signing
801 algorithm to recover the private key.
803 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
807 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
808 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
809 are retained for backwards compatibility.
812 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
813 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
814 of two gigabytes and the error handling improved.
816 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
817 categorized as a normal bug, not a security issue, because the DRBG reseeds
818 automatically and is fully functional even without additional randomness
819 provided by the application.
821 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
823 *) Add a new ClientHello callback. Provides a callback interface that gives
824 the application the ability to adjust the nascent SSL object at the
825 earliest stage of ClientHello processing, immediately after extensions have
826 been collected but before they have been processed. In particular, this
827 callback can adjust the supported TLS versions in response to the contents
831 *) Add SM2 base algorithm support.
834 *) s390x assembly pack: add (improved) hardware-support for the following
835 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
836 aes-cfb/cfb8, aes-ecb.
839 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
840 parameter is no longer accepted, as it leads to a corrupt table. NULL
841 pem_str is reserved for alias entries only.
844 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
845 step for prime curves. The new implementation is based on formulae from
846 differential addition-and-doubling in homogeneous projective coordinates
847 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
848 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
849 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
850 to work in projective coordinates.
851 [Billy Bob Brumley, Nicola Tuveri]
853 *) Change generating and checking of primes so that the error rate of not
854 being prime depends on the intended use based on the size of the input.
855 For larger primes this will result in more rounds of Miller-Rabin.
856 The maximal error rate for primes with more than 1080 bits is lowered
858 [Kurt Roeckx, Annie Yousar]
860 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
863 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
864 moving between systems, and to avoid confusion when a Windows build is
865 done with mingw vs with MSVC. For POSIX installs, there's still a
866 symlink or copy named 'tsget' to avoid that confusion as well.
869 *) Revert blinding in ECDSA sign and instead make problematic addition
870 length-invariant. Switch even to fixed-length Montgomery multiplication.
873 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
874 step for binary curves. The new implementation is based on formulae from
875 differential addition-and-doubling in mixed Lopez-Dahab projective
876 coordinates, modified to independently blind the operands.
877 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
879 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
880 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
881 EC_METHODs to implement their own specialized "ladder step", to take
882 advantage of more favorable coordinate systems or more efficient
883 differential addition-and-doubling algorithms.
884 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
886 *) Modified the random device based seed sources to keep the relevant
887 file descriptors open rather than reopening them on each access.
888 This allows such sources to operate in a chroot() jail without
889 the associated device nodes being available. This behaviour can be
890 controlled using RAND_keep_random_devices_open().
893 *) Numerous side-channel attack mitigations have been applied. This may have
894 performance impacts for some algorithms for the benefit of improved
895 security. Specific changes are noted in this change log by their respective
899 *) AIX shared library support overhaul. Switch to AIX "natural" way of
900 handling shared libraries, which means collecting shared objects of
901 different versions and bitnesses in one common archive. This allows to
902 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
903 doesn't affect the way 3rd party applications are linked, only how
904 multi-version installation is managed.
907 *) Make ec_group_do_inverse_ord() more robust and available to other
908 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
909 mitigations are applied to the fallback BN_mod_inverse().
910 When using this function rather than BN_mod_inverse() directly, new
911 EC cryptosystem implementations are then safer-by-default.
914 *) Add coordinate blinding for EC_POINT and implement projective
915 coordinate blinding for generic prime curves as a countermeasure to
916 chosen point SCA attacks.
917 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
919 *) Add blinding to ECDSA and DSA signatures to protect against side channel
920 attacks discovered by Keegan Ryan (NCC Group).
923 *) Enforce checking in the pkeyutl command line app to ensure that the input
924 length does not exceed the maximum supported digest length when performing
925 a sign, verify or verifyrecover operation.
928 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
929 I/O in combination with something like select() or poll() will hang. This
930 can be turned off again using SSL_CTX_clear_mode().
931 Many applications do not properly handle non-application data records, and
932 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
933 around the problems in those applications, but can also break some.
934 It's recommended to read the manpages about SSL_read(), SSL_write(),
935 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
936 SSL_CTX_set_read_ahead() again.
939 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
940 now allow empty (zero character) pass phrases.
943 *) Apply blinding to binary field modular inversion and remove patent
944 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
947 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
948 binary and prime elliptic curves.
951 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
952 constant time fixed point multiplication.
955 *) Revise elliptic curve scalar multiplication with timing attack
956 defenses: ec_wNAF_mul redirects to a constant time implementation
957 when computing fixed point and variable point multiplication (which
958 in OpenSSL are mostly used with secret scalars in keygen, sign,
959 ECDH derive operations).
960 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
963 *) Updated CONTRIBUTING
966 *) Updated DRBG / RAND to request nonce and additional low entropy
967 randomness from the system.
968 [Matthias St. Pierre]
970 *) Updated 'openssl rehash' to use OpenSSL consistent default.
973 *) Moved the load of the ssl_conf module to libcrypto, which helps
974 loading engines that libssl uses before libssl is initialised.
977 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
980 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
981 [Ingo Schwarze, Rich Salz]
983 *) Added output of accepting IP address and port for 'openssl s_server'
986 *) Added a new API for TLSv1.3 ciphersuites:
987 SSL_CTX_set_ciphersuites()
988 SSL_set_ciphersuites()
991 *) Memory allocation failures consistently add an error to the error
995 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
996 in libcrypto when run as setuid/setgid.
999 *) Load any config file by default when libssl is used.
1002 *) Added new public header file <openssl/rand_drbg.h> and documentation
1003 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1004 [Matthias St. Pierre]
1006 *) QNX support removed (cannot find contributors to get their approval
1007 for the license change).
1010 *) TLSv1.3 replay protection for early data has been implemented. See the
1011 SSL_read_early_data() man page for further details.
1014 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1015 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1016 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1017 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1018 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1019 configuration has been separated out. See the ciphers man page or the
1020 SSL_CTX_set_ciphersuites() man page for more information.
1023 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1024 in responder mode now supports the new "-multi" option, which
1025 spawns the specified number of child processes to handle OCSP
1026 requests. The "-timeout" option now also limits the OCSP
1027 responder's patience to wait to receive the full client request
1028 on a newly accepted connection. Child processes are respawned
1029 as needed, and the CA index file is automatically reloaded
1030 when changed. This makes it possible to run the "ocsp" responder
1031 as a long-running service, making the OpenSSL CA somewhat more
1032 feature-complete. In this mode, most diagnostic messages logged
1033 after entering the event loop are logged via syslog(3) rather than
1037 *) Added support for X448 and Ed448. Heavily based on original work by
1041 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
1042 objects loaded. This adds the functions OSSL_STORE_expect() and
1043 OSSL_STORE_find() as well as needed tools to construct searches and
1044 get the search data out of them.
1047 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
1048 version of OpenSSL should review their configuration settings to ensure
1049 that they are still appropriate for TLSv1.3. For further information see:
1050 https://wiki.openssl.org/index.php/TLS1.3
1053 *) Grand redesign of the OpenSSL random generator
1055 The default RAND method now utilizes an AES-CTR DRBG according to
1056 NIST standard SP 800-90Ar1. The new random generator is essentially
1057 a port of the default random generator from the OpenSSL FIPS 2.0
1058 object module. It is a hybrid deterministic random bit generator
1059 using an AES-CTR bit stream and which seeds and reseeds itself
1060 automatically using trusted system entropy sources.
1062 Some of its new features are:
1063 o Support for multiple DRBG instances with seed chaining.
1064 o The default RAND method makes use of a DRBG.
1065 o There is a public and private DRBG instance.
1066 o The DRBG instances are fork-safe.
1067 o Keep all global DRBG instances on the secure heap if it is enabled.
1068 o The public and private DRBG instance are per thread for lock free
1070 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
1072 *) Changed Configure so it only says what it does and doesn't dump
1073 so much data. Instead, ./configdata.pm should be used as a script
1074 to display all sorts of configuration data.
1077 *) Added processing of "make variables" to Configure.
1080 *) Added SHA512/224 and SHA512/256 algorithm support.
1083 *) The last traces of Netware support, first removed in 1.1.0, have
1087 *) Get rid of Makefile.shared, and in the process, make the processing
1088 of certain files (rc.obj, or the .def/.map/.opt files produced from
1089 the ordinal files) more visible and hopefully easier to trace and
1090 debug (or make silent).
1093 *) Make it possible to have environment variable assignments as
1094 arguments to config / Configure.
1097 *) Add multi-prime RSA (RFC 8017) support.
1100 *) Add SM3 implemented according to GB/T 32905-2016
1101 [ Jack Lloyd <jack.lloyd@ribose.com>,
1102 Ronald Tse <ronald.tse@ribose.com>,
1103 Erick Borsboom <erick.borsboom@ribose.com> ]
1105 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
1106 as documented in RFC6066.
1107 Based on a patch from Tomasz Moń
1108 [Filipe Raimundo da Silva]
1110 *) Add SM4 implemented according to GB/T 32907-2016.
1111 [ Jack Lloyd <jack.lloyd@ribose.com>,
1112 Ronald Tse <ronald.tse@ribose.com>,
1113 Erick Borsboom <erick.borsboom@ribose.com> ]
1115 *) Reimplement -newreq-nodes and ERR_error_string_n; the
1116 original author does not agree with the license change.
1119 *) Add ARIA AEAD TLS support.
1122 *) Some macro definitions to support VS6 have been removed. Visual
1123 Studio 6 has not worked since 1.1.0
1126 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
1127 without clearing the errors.
1130 *) Add "atfork" functions. If building on a system that without
1131 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1132 requirements. The RAND facility now uses/requires this.
1138 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
1139 not possible to disable entirely. However, it's still possible to
1140 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1143 To disable, configure with 'no-ui-console'. 'no-ui' is still
1144 possible to use as an alias. Check at compile time with the
1145 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1146 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1149 *) Add a STORE module, which implements a uniform and URI based reader of
1150 stores that can contain keys, certificates, CRLs and numerous other
1151 objects. The main API is loosely based on a few stdio functions,
1152 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1153 OSSL_STORE_error and OSSL_STORE_close.
1154 The implementation uses backends called "loaders" to implement arbitrary
1155 URI schemes. There is one built in "loader" for the 'file' scheme.
1158 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
1159 then adjusted to work on FreeBSD 8.4 as well.
1160 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1161 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1164 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
1165 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1166 error code calls like this:
1168 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1170 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1171 that can be encoded in C. For the foreseeable future, this will only
1173 [Richard Levitte and Tim Hudson]
1175 *) Removed BSD cryptodev engine.
1178 *) Add a build target 'build_all_generated', to build all generated files
1179 and only that. This can be used to prepare everything that requires
1180 things like perl for a system that lacks perl and then move everything
1181 to that system and do the rest of the build there.
1184 *) In the UI interface, make it possible to duplicate the user data. This
1185 can be used by engines that need to retain the data for a longer time
1186 than just the call where this user data is passed.
1189 *) Ignore the '-named_curve auto' value for compatibility of applications
1191 [Tomas Mraz <tmraz@fedoraproject.org>]
1193 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1194 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1195 alerts across multiple records (some of which could be empty). In practice
1196 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1197 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1198 support this at all. Supporting it adds significant complexity to the
1199 record layer, and its removal is unlikely to cause interoperability
1203 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1204 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1205 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1209 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1210 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1211 [Richard Levitte, Andy Polyakov]
1213 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1217 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1218 platform rather than 'mingw'.
1221 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
1222 success if they are asked to add an object which already exists
1223 in the store. This change cascades to other functions which load
1224 certificates and CRLs.
1227 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
1228 facilitate stack unwinding even from assembly subroutines.
1231 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1232 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1235 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1236 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1237 which is the minimum version we support.
1240 *) Certificate time validation (X509_cmp_time) enforces stricter
1241 compliance with RFC 5280. Fractional seconds and timezone offsets
1242 are no longer allowed.
1245 *) Add support for ARIA
1248 *) s_client will now send the Server Name Indication (SNI) extension by
1249 default unless the new "-noservername" option is used. The server name is
1250 based on the host provided to the "-connect" option unless overridden by
1251 using "-servername".
1254 *) Add support for SipHash
1257 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1258 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1259 prevent issues where no progress is being made and the peer continually
1260 sends unrecognised record types, using up resources processing them.
1263 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
1264 using the algorithm defined in
1265 https://www.akkadia.org/drepper/SHA-crypt.txt
1268 *) Heartbeat support has been removed; the ABI is changed for now.
1269 [Richard Levitte, Rich Salz]
1271 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1274 *) The RSA "null" method, which was partially supported to avoid patent
1275 issues, has been replaced to always returns NULL.
1279 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
1281 *) Client DoS due to large DH parameter
1283 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1284 malicious server can send a very large prime value to the client. This will
1285 cause the client to spend an unreasonably long period of time generating a
1286 key for this prime resulting in a hang until the client has finished. This
1287 could be exploited in a Denial Of Service attack.
1289 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1293 *) Cache timing vulnerability in RSA Key Generation
1295 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1296 a cache timing side channel attack. An attacker with sufficient access to
1297 mount cache timing attacks during the RSA key generation process could
1298 recover the private key.
1300 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1301 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1305 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1306 parameter is no longer accepted, as it leads to a corrupt table. NULL
1307 pem_str is reserved for alias entries only.
1310 *) Revert blinding in ECDSA sign and instead make problematic addition
1311 length-invariant. Switch even to fixed-length Montgomery multiplication.
1314 *) Change generating and checking of primes so that the error rate of not
1315 being prime depends on the intended use based on the size of the input.
1316 For larger primes this will result in more rounds of Miller-Rabin.
1317 The maximal error rate for primes with more than 1080 bits is lowered
1319 [Kurt Roeckx, Annie Yousar]
1321 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1324 *) Add blinding to ECDSA and DSA signatures to protect against side channel
1325 attacks discovered by Keegan Ryan (NCC Group).
1328 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1329 now allow empty (zero character) pass phrases.
1332 *) Certificate time validation (X509_cmp_time) enforces stricter
1333 compliance with RFC 5280. Fractional seconds and timezone offsets
1334 are no longer allowed.
1337 *) Fixed a text canonicalisation bug in CMS
1339 Where a CMS detached signature is used with text content the text goes
1340 through a canonicalisation process first prior to signing or verifying a
1341 signature. This process strips trailing space at the end of lines, converts
1342 line terminators to CRLF and removes additional trailing line terminators
1343 at the end of a file. A bug in the canonicalisation process meant that
1344 some characters, such as form-feed, were incorrectly treated as whitespace
1345 and removed. This is contrary to the specification (RFC5485). This fix
1346 could mean that detached text data signed with an earlier version of
1347 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
1348 signed with a fixed OpenSSL may fail to verify with an earlier version of
1349 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
1350 and use the "-binary" flag (for the "cms" command line application) or set
1351 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
1354 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
1356 *) Constructed ASN.1 types with a recursive definition could exceed the stack
1358 Constructed ASN.1 types with a recursive definition (such as can be found
1359 in PKCS7) could eventually exceed the stack given malicious input with
1360 excessive recursion. This could result in a Denial Of Service attack. There
1361 are no such structures used within SSL/TLS that come from untrusted sources
1362 so this is considered safe.
1364 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
1369 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
1371 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
1372 effectively reduced to only comparing the least significant bit of each
1373 byte. This allows an attacker to forge messages that would be considered as
1374 authenticated in an amount of tries lower than that guaranteed by the
1375 security claims of the scheme. The module can only be compiled by the
1376 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
1378 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
1383 *) Add a build target 'build_all_generated', to build all generated files
1384 and only that. This can be used to prepare everything that requires
1385 things like perl for a system that lacks perl and then move everything
1386 to that system and do the rest of the build there.
1389 *) Backport SSL_OP_NO_RENGOTIATION
1391 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
1392 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
1393 changes this is no longer possible in 1.1.0. Therefore the new
1394 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
1395 1.1.0 to provide equivalent functionality.
1397 Note that if an application built against 1.1.0h headers (or above) is run
1398 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
1399 accepted but nothing will happen, i.e. renegotiation will not be prevented.
1402 *) Removed the OS390-Unix config target. It relied on a script that doesn't
1406 *) rsaz_1024_mul_avx2 overflow bug on x86_64
1408 There is an overflow bug in the AVX2 Montgomery multiplication procedure
1409 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
1410 Analysis suggests that attacks against RSA and DSA as a result of this
1411 defect would be very difficult to perform and are not believed likely.
1412 Attacks against DH1024 are considered just feasible, because most of the
1413 work necessary to deduce information about a private key may be performed
1414 offline. The amount of resources required for such an attack would be
1415 significant. However, for an attack on TLS to be meaningful, the server
1416 would have to share the DH1024 private key among multiple clients, which is
1417 no longer an option since CVE-2016-0701.
1419 This only affects processors that support the AVX2 but not ADX extensions
1420 like Intel Haswell (4th generation).
1422 This issue was reported to OpenSSL by David Benjamin (Google). The issue
1423 was originally found via the OSS-Fuzz project.
1427 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
1429 *) bn_sqrx8x_internal carry bug on x86_64
1431 There is a carry propagating bug in the x86_64 Montgomery squaring
1432 procedure. No EC algorithms are affected. Analysis suggests that attacks
1433 against RSA and DSA as a result of this defect would be very difficult to
1434 perform and are not believed likely. Attacks against DH are considered just
1435 feasible (although very difficult) because most of the work necessary to
1436 deduce information about a private key may be performed offline. The amount
1437 of resources required for such an attack would be very significant and
1438 likely only accessible to a limited number of attackers. An attacker would
1439 additionally need online access to an unpatched system using the target
1440 private key in a scenario with persistent DH parameters and a private
1441 key that is shared between multiple clients.
1443 This only affects processors that support the BMI1, BMI2 and ADX extensions
1444 like Intel Broadwell (5th generation) and later or AMD Ryzen.
1446 This issue was reported to OpenSSL by the OSS-Fuzz project.
1450 *) Malformed X.509 IPAddressFamily could cause OOB read
1452 If an X.509 certificate has a malformed IPAddressFamily extension,
1453 OpenSSL could do a one-byte buffer overread. The most likely result
1454 would be an erroneous display of the certificate in text format.
1456 This issue was reported to OpenSSL by the OSS-Fuzz project.
1460 Changes between 1.1.0e and 1.1.0f [25 May 2017]
1462 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1463 platform rather than 'mingw'.
1466 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1467 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1468 which is the minimum version we support.
1471 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
1473 *) Encrypt-Then-Mac renegotiation crash
1475 During a renegotiation handshake if the Encrypt-Then-Mac extension is
1476 negotiated where it was not in the original handshake (or vice-versa) then
1477 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
1478 and servers are affected.
1480 This issue was reported to OpenSSL by Joe Orton (Red Hat).
1484 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
1486 *) Truncated packet could crash via OOB read
1488 If one side of an SSL/TLS path is running on a 32-bit host and a specific
1489 cipher is being used, then a truncated packet can cause that host to
1490 perform an out-of-bounds read, usually resulting in a crash.
1492 This issue was reported to OpenSSL by Robert Święcki of Google.
1496 *) Bad (EC)DHE parameters cause a client crash
1498 If a malicious server supplies bad parameters for a DHE or ECDHE key
1499 exchange then this can result in the client attempting to dereference a
1500 NULL pointer leading to a client crash. This could be exploited in a Denial
1503 This issue was reported to OpenSSL by Guido Vranken.
1507 *) BN_mod_exp may produce incorrect results on x86_64
1509 There is a carry propagating bug in the x86_64 Montgomery squaring
1510 procedure. No EC algorithms are affected. Analysis suggests that attacks
1511 against RSA and DSA as a result of this defect would be very difficult to
1512 perform and are not believed likely. Attacks against DH are considered just
1513 feasible (although very difficult) because most of the work necessary to
1514 deduce information about a private key may be performed offline. The amount
1515 of resources required for such an attack would be very significant and
1516 likely only accessible to a limited number of attackers. An attacker would
1517 additionally need online access to an unpatched system using the target
1518 private key in a scenario with persistent DH parameters and a private
1519 key that is shared between multiple clients. For example this can occur by
1520 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1521 similar to CVE-2015-3193 but must be treated as a separate problem.
1523 This issue was reported to OpenSSL by the OSS-Fuzz project.
1527 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1529 *) ChaCha20/Poly1305 heap-buffer-overflow
1531 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1532 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1533 crash. This issue is not considered to be exploitable beyond a DoS.
1535 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1539 *) CMS Null dereference
1541 Applications parsing invalid CMS structures can crash with a NULL pointer
1542 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1543 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1544 structure callback if an attempt is made to free certain invalid encodings.
1545 Only CHOICE structures using a callback which do not handle NULL value are
1548 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1552 *) Montgomery multiplication may produce incorrect results
1554 There is a carry propagating bug in the Broadwell-specific Montgomery
1555 multiplication procedure that handles input lengths divisible by, but
1556 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1557 and DH private keys are impossible. This is because the subroutine in
1558 question is not used in operations with the private key itself and an input
1559 of the attacker's direct choice. Otherwise the bug can manifest itself as
1560 transient authentication and key negotiation failures or reproducible
1561 erroneous outcome of public-key operations with specially crafted input.
1562 Among EC algorithms only Brainpool P-512 curves are affected and one
1563 presumably can attack ECDH key negotiation. Impact was not analyzed in
1564 detail, because pre-requisites for attack are considered unlikely. Namely
1565 multiple clients have to choose the curve in question and the server has to
1566 share the private key among them, neither of which is default behaviour.
1567 Even then only clients that chose the curve will be affected.
1569 This issue was publicly reported as transient failures and was not
1570 initially recognized as a security issue. Thanks to Richard Morgan for
1571 providing reproducible case.
1575 *) Removed automatic addition of RPATH in shared libraries and executables,
1576 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1579 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1581 *) Fix Use After Free for large message sizes
1583 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1584 message larger than approx 16k is received then the underlying buffer to
1585 store the incoming message is reallocated and moved. Unfortunately a
1586 dangling pointer to the old location is left which results in an attempt to
1587 write to the previously freed location. This is likely to result in a
1588 crash, however it could potentially lead to execution of arbitrary code.
1590 This issue only affects OpenSSL 1.1.0a.
1592 This issue was reported to OpenSSL by Robert Święcki.
1596 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1598 *) OCSP Status Request extension unbounded memory growth
1600 A malicious client can send an excessively large OCSP Status Request
1601 extension. If that client continually requests renegotiation, sending a
1602 large OCSP Status Request extension each time, then there will be unbounded
1603 memory growth on the server. This will eventually lead to a Denial Of
1604 Service attack through memory exhaustion. Servers with a default
1605 configuration are vulnerable even if they do not support OCSP. Builds using
1606 the "no-ocsp" build time option are not affected.
1608 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1612 *) SSL_peek() hang on empty record
1614 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1615 sends an empty record. This could be exploited by a malicious peer in a
1616 Denial Of Service attack.
1618 This issue was reported to OpenSSL by Alex Gaynor.
1622 *) Excessive allocation of memory in tls_get_message_header() and
1623 dtls1_preprocess_fragment()
1625 A (D)TLS message includes 3 bytes for its length in the header for the
1626 message. This would allow for messages up to 16Mb in length. Messages of
1627 this length are excessive and OpenSSL includes a check to ensure that a
1628 peer is sending reasonably sized messages in order to avoid too much memory
1629 being consumed to service a connection. A flaw in the logic of version
1630 1.1.0 means that memory for the message is allocated too early, prior to
1631 the excessive message length check. Due to way memory is allocated in
1632 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1633 to service a connection. This could lead to a Denial of Service through
1634 memory exhaustion. However, the excessive message length check still takes
1635 place, and this would cause the connection to immediately fail. Assuming
1636 that the application calls SSL_free() on the failed connection in a timely
1637 manner then the 21Mb of allocated memory will then be immediately freed
1638 again. Therefore the excessive memory allocation will be transitory in
1639 nature. This then means that there is only a security impact if:
1641 1) The application does not call SSL_free() in a timely manner in the event
1642 that the connection fails
1644 2) The application is working in a constrained environment where there is
1645 very little free memory
1647 3) The attacker initiates multiple connection attempts such that there are
1648 multiple connections in a state where memory has been allocated for the
1649 connection; SSL_free() has not yet been called; and there is insufficient
1650 memory to service the multiple requests.
1652 Except in the instance of (1) above any Denial Of Service is likely to be
1653 transitory because as soon as the connection fails the memory is
1654 subsequently freed again in the SSL_free() call. However there is an
1655 increased risk during this period of application crashes due to the lack of
1656 memory - which would then mean a more serious Denial of Service.
1658 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1659 (CVE-2016-6307 and CVE-2016-6308)
1662 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1663 had to be removed. Primary reason is that vendor assembler can't
1664 assemble our modules with -KPIC flag. As result it, assembly
1665 support, was not even available as option. But its lack means
1666 lack of side-channel resistant code, which is incompatible with
1667 security by todays standards. Fortunately gcc is readily available
1668 prepackaged option, which we firmly point at...
1671 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1673 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1674 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1675 (to any value) allows Windows user to access PKCS#12 file generated
1676 with Windows CryptoAPI and protected with non-ASCII password, as well
1677 as files generated under UTF-8 locale on Linux also protected with
1681 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1682 have been disabled by default and removed from DEFAULT, just like RC4.
1683 See the RC4 item below to re-enable both.
1686 *) The method for finding the storage location for the Windows RAND seed file
1687 has changed. First we check %RANDFILE%. If that is not set then we check
1688 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1689 all else fails we fall back to C:\.
1692 *) The EVP_EncryptUpdate() function has had its return type changed from void
1693 to int. A return of 0 indicates and error while a return of 1 indicates
1697 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1698 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1699 off the constant time implementation for RSA, DSA and DH have been made
1700 no-ops and deprecated.
1703 *) Windows RAND implementation was simplified to only get entropy by
1704 calling CryptGenRandom(). Various other RAND-related tickets
1706 [Joseph Wylie Yandle, Rich Salz]
1708 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1709 and OPENSSL_LH_, respectively. The old names are available
1710 with API compatibility. They new names are now completely documented.
1713 *) Unify TYPE_up_ref(obj) methods signature.
1714 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1715 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1716 int (instead of void) like all others TYPE_up_ref() methods.
1717 So now these methods also check the return value of CRYPTO_atomic_add(),
1718 and the validity of object reference counter.
1719 [fdasilvayy@gmail.com]
1721 *) With Windows Visual Studio builds, the .pdb files are installed
1722 alongside the installed libraries and executables. For a static
1723 library installation, ossl_static.pdb is the associate compiler
1724 generated .pdb file to be used when linking programs.
1727 *) Remove openssl.spec. Packaging files belong with the packagers.
1730 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1731 recognise x86_64 architectures automatically. You can still decide
1732 to build for a different bitness with the environment variable
1733 KERNEL_BITS (can be 32 or 64), for example:
1735 KERNEL_BITS=32 ./config
1739 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1740 256 bit AES and HMAC with SHA256.
1743 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1746 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1749 *) To enable users to have their own config files and build file templates,
1750 Configure looks in the directory indicated by the environment variable
1751 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1752 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1753 name and is used as is.
1756 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1757 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1758 X509_CERT_FILE_CTX was removed.
1761 *) "shared" builds are now the default. To create only static libraries use
1762 the "no-shared" Configure option.
1765 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1766 All of these option have not worked for some while and are fundamental
1770 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1771 global cleanup functions are no longer required because they are handled
1772 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1773 Explicitly de-initing can cause problems (e.g. where a library that uses
1774 OpenSSL de-inits, but an application is still using it). The affected
1775 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1776 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1777 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1778 COMP_zlib_cleanup().
1781 *) --strict-warnings no longer enables runtime debugging options
1782 such as REF_DEBUG. Instead, debug options are automatically
1783 enabled with '--debug' builds.
1784 [Andy Polyakov, Emilia Käsper]
1786 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1787 have been moved out of the public header files. New functions for managing
1788 these have been added.
1791 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1792 objects have been moved out of the public header files. New
1793 functions for managing these have been added.
1796 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1797 have been moved out of the public header files. New functions for managing
1798 these have been added.
1801 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1802 moved out of the public header files. New functions for managing these
1806 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1809 *) Removed the mk1mf build scripts.
1812 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1813 it is always safe to #include a header now.
1816 *) Removed the aged BC-32 config and all its supporting scripts
1819 *) Removed support for Ultrix, Netware, and OS/2.
1822 *) Add support for HKDF.
1823 [Alessandro Ghedini]
1825 *) Add support for blake2b and blake2s
1828 *) Added support for "pipelining". Ciphers that have the
1829 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1830 encryptions/decryptions simultaneously. There are currently no built-in
1831 ciphers with this property but the expectation is that engines will be able
1832 to offer it to significantly improve throughput. Support has been extended
1833 into libssl so that multiple records for a single connection can be
1834 processed in one go (for >=TLS 1.1).
1837 *) Added the AFALG engine. This is an async capable engine which is able to
1838 offload work to the Linux kernel. In this initial version it only supports
1839 AES128-CBC. The kernel must be version 4.1.0 or greater.
1842 *) OpenSSL now uses a new threading API. It is no longer necessary to
1843 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1844 are two supported threading models: pthreads and windows threads. It is
1845 also possible to configure OpenSSL at compile time for "no-threads". The
1846 old threading API should no longer be used. The functions have been
1847 replaced with "no-op" compatibility macros.
1848 [Alessandro Ghedini, Matt Caswell]
1850 *) Modify behavior of ALPN to invoke callback after SNI/servername
1851 callback, such that updates to the SSL_CTX affect ALPN.
1854 *) Add SSL_CIPHER queries for authentication and key-exchange.
1857 *) Changes to the DEFAULT cipherlist:
1858 - Prefer (EC)DHE handshakes over plain RSA.
1859 - Prefer AEAD ciphers over legacy ciphers.
1860 - Prefer ECDSA over RSA when both certificates are available.
1861 - Prefer TLSv1.2 ciphers/PRF.
1862 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1866 *) Change the ECC default curve list to be this, in order: x25519,
1867 secp256r1, secp521r1, secp384r1.
1870 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1871 disabled by default. They can be re-enabled using the
1872 enable-weak-ssl-ciphers option to Configure.
1875 *) If the server has ALPN configured, but supports no protocols that the
1876 client advertises, send a fatal "no_application_protocol" alert.
1877 This behaviour is SHALL in RFC 7301, though it isn't universally
1878 implemented by other servers.
1881 *) Add X25519 support.
1882 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1883 for public and private key encoding using the format documented in
1884 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1885 key generation and key derivation.
1887 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1891 *) Deprecate SRP_VBASE_get_by_user.
1892 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1893 In order to fix an unavoidable memory leak (CVE-2016-0798),
1894 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1895 seed, even if the seed is configured.
1897 Users should use SRP_VBASE_get1_by_user instead. Note that in
1898 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1899 also that even though configuring the SRP seed attempts to hide
1900 invalid usernames by continuing the handshake with fake
1901 credentials, this behaviour is not constant time and no strong
1902 guarantees are made that the handshake is indistinguishable from
1903 that of a valid user.
1906 *) Configuration change; it's now possible to build dynamic engines
1907 without having to build shared libraries and vice versa. This
1908 only applies to the engines in engines/, those in crypto/engine/
1909 will always be built into libcrypto (i.e. "static").
1911 Building dynamic engines is enabled by default; to disable, use
1912 the configuration option "disable-dynamic-engine".
1914 The only requirements for building dynamic engines are the
1915 presence of the DSO module and building with position independent
1916 code, so they will also automatically be disabled if configuring
1917 with "disable-dso" or "disable-pic".
1919 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1920 are also taken away from openssl/opensslconf.h, as they are
1924 *) Configuration change; if there is a known flag to compile
1925 position independent code, it will always be applied on the
1926 libcrypto and libssl object files, and never on the application
1927 object files. This means other libraries that use routines from
1928 libcrypto / libssl can be made into shared libraries regardless
1929 of how OpenSSL was configured.
1931 If this isn't desirable, the configuration options "disable-pic"
1932 or "no-pic" can be used to disable the use of PIC. This will
1933 also disable building shared libraries and dynamic engines.
1936 *) Removed JPAKE code. It was experimental and has no wide use.
1939 *) The INSTALL_PREFIX Makefile variable has been renamed to
1940 DESTDIR. That makes for less confusion on what this variable
1941 is for. Also, the configuration option --install_prefix is
1945 *) Heartbeat for TLS has been removed and is disabled by default
1946 for DTLS; configure with enable-heartbeats. Code that uses the
1947 old #define's might need to be updated.
1948 [Emilia Käsper, Rich Salz]
1950 *) Rename REF_CHECK to REF_DEBUG.
1953 *) New "unified" build system
1955 The "unified" build system is aimed to be a common system for all
1956 platforms we support. With it comes new support for VMS.
1958 This system builds supports building in a different directory tree
1959 than the source tree. It produces one Makefile (for unix family
1960 or lookalikes), or one descrip.mms (for VMS).
1962 The source of information to make the Makefile / descrip.mms is
1963 small files called 'build.info', holding the necessary
1964 information for each directory with source to compile, and a
1965 template in Configurations, like unix-Makefile.tmpl or
1968 With this change, the library names were also renamed on Windows
1969 and on VMS. They now have names that are closer to the standard
1970 on Unix, and include the major version number, and in certain
1971 cases, the architecture they are built for. See "Notes on shared
1972 libraries" in INSTALL.
1974 We rely heavily on the perl module Text::Template.
1977 *) Added support for auto-initialisation and de-initialisation of the library.
1978 OpenSSL no longer requires explicit init or deinit routines to be called,
1979 except in certain circumstances. See the OPENSSL_init_crypto() and
1980 OPENSSL_init_ssl() man pages for further information.
1983 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1984 "peer" argument is now expected to be a BIO_ADDR object.
1986 *) Rewrite of BIO networking library. The BIO library lacked consistent
1987 support of IPv6, and adding it required some more extensive
1988 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1989 which hold all types of addresses and chains of address information.
1990 It also introduces a new API, with functions like BIO_socket,
1991 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1992 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1993 have been adapted accordingly.
1996 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2000 *) CRIME protection: disable compression by default, even if OpenSSL is
2001 compiled with zlib enabled. Applications can still enable compression
2002 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2003 using the SSL_CONF library to configure compression.
2006 *) The signature of the session callback configured with
2007 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2008 was explicitly marked as 'const unsigned char*' instead of
2012 *) Always DPURIFY. Remove the use of uninitialized memory in the
2013 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2016 *) Removed many obsolete configuration items, including
2017 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2018 MD2_CHAR, MD2_INT, MD2_LONG
2020 IDEA_SHORT, IDEA_LONG
2021 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2022 [Rich Salz, with advice from Andy Polyakov]
2024 *) Many BN internals have been moved to an internal header file.
2025 [Rich Salz with help from Andy Polyakov]
2027 *) Configuration and writing out the results from it has changed.
2028 Files such as Makefile include/openssl/opensslconf.h and are now
2029 produced through general templates, such as Makefile.in and
2030 crypto/opensslconf.h.in and some help from the perl module
2033 Also, the center of configuration information is no longer
2034 Makefile. Instead, Configure produces a perl module in
2035 configdata.pm which holds most of the config data (in the hash
2036 table %config), the target data that comes from the target
2037 configuration in one of the Configurations/*.conf files (in
2041 *) To clarify their intended purposes, the Configure options
2042 --prefix and --openssldir change their semantics, and become more
2043 straightforward and less interdependent.
2045 --prefix shall be used exclusively to give the location INSTALLTOP
2046 where programs, scripts, libraries, include files and manuals are
2047 going to be installed. The default is now /usr/local.
2049 --openssldir shall be used exclusively to give the default
2050 location OPENSSLDIR where certificates, private keys, CRLs are
2051 managed. This is also where the default openssl.cnf gets
2053 If the directory given with this option is a relative path, the
2054 values of both the --prefix value and the --openssldir value will
2055 be combined to become OPENSSLDIR.
2056 The default for --openssldir is INSTALLTOP/ssl.
2058 Anyone who uses --openssldir to specify where OpenSSL is to be
2059 installed MUST change to use --prefix instead.
2062 *) The GOST engine was out of date and therefore it has been removed. An up
2063 to date GOST engine is now being maintained in an external repository.
2064 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
2065 support for GOST ciphersuites (these are only activated if a GOST engine
2069 *) EGD is no longer supported by default; use enable-egd when
2071 [Ben Kaduk and Rich Salz]
2073 *) The distribution now has Makefile.in files, which are used to
2074 create Makefile's when Configure is run. *Configure must be run
2075 before trying to build now.*
2078 *) The return value for SSL_CIPHER_description() for error conditions
2082 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2084 Obtaining and performing DNSSEC validation of TLSA records is
2085 the application's responsibility. The application provides
2086 the TLSA records of its choice to OpenSSL, and these are then
2087 used to authenticate the peer.
2089 The TLSA records need not even come from DNS. They can, for
2090 example, be used to implement local end-entity certificate or
2091 trust-anchor "pinning", where the "pin" data takes the form
2092 of TLSA records, which can augment or replace verification
2093 based on the usual WebPKI public certification authorities.
2096 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2097 continues to support deprecated interfaces in default builds.
2098 However, applications are strongly advised to compile their
2099 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2100 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2101 or the 1.1.0 releases.
2103 In environments in which all applications have been ported to
2104 not use any deprecated interfaces OpenSSL's Configure script
2105 should be used with the --api=1.1.0 option to entirely remove
2106 support for the deprecated features from the library and
2107 unconditionally disable them in the installed headers.
2108 Essentially the same effect can be achieved with the "no-deprecated"
2109 argument to Configure, except that this will always restrict
2110 the build to just the latest API, rather than a fixed API
2113 As applications are ported to future revisions of the API,
2114 they should update their compile-time OPENSSL_API_COMPAT define
2115 accordingly, but in most cases should be able to continue to
2116 compile with later releases.
2118 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2119 0x10000000L and 0x00908000L, respectively. However those
2120 versions did not support the OPENSSL_API_COMPAT feature, and
2121 so applications are not typically tested for explicit support
2122 of just the undeprecated features of either release.
2125 *) Add support for setting the minimum and maximum supported protocol.
2126 It can bet set via the SSL_set_min_proto_version() and
2127 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2128 MaxProtocol. It's recommended to use the new APIs to disable
2129 protocols instead of disabling individual protocols using
2130 SSL_set_options() or SSL_CONF's Protocol. This change also
2131 removes support for disabling TLS 1.2 in the OpenSSL TLS
2132 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2135 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2138 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2139 and integrates ECDSA and ECDH functionality into EC. Implementations can
2140 now redirect key generation and no longer need to convert to or from
2143 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2144 include the ec.h header file instead.
2147 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
2148 ciphers who are no longer supported and drops support the ephemeral RSA key
2149 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2152 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2153 opaque. For HMAC_CTX, the following constructors and destructors
2156 HMAC_CTX *HMAC_CTX_new(void);
2157 void HMAC_CTX_free(HMAC_CTX *ctx);
2159 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2160 destroy such methods has been added. See EVP_MD_meth_new(3) and
2161 EVP_CIPHER_meth_new(3) for documentation.
2164 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2165 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2166 EVP_MD_CTX_reset() should be called instead to reinitialise
2167 an already created structure.
2168 2) For consistency with the majority of our object creators and
2169 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2170 EVP_MD_CTX_(new|free). The old names are retained as macros
2171 for deprecated builds.
2174 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
2175 cryptographic operations to be performed asynchronously as long as an
2176 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2177 further details. Libssl has also had this capability integrated with the
2178 introduction of the new mode SSL_MODE_ASYNC and associated error
2179 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2180 pages. This work was developed in partnership with Intel Corp.
2183 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2184 always enabled now. If you want to disable the support you should
2185 exclude it using the list of supported ciphers. This also means that the
2186 "-no_ecdhe" option has been removed from s_server.
2189 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2190 SSL_{CTX_}set1_curves() which can set a list.
2193 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2194 curve you want to support using SSL_{CTX_}set1_curves().
2197 *) State machine rewrite. The state machine code has been significantly
2198 refactored in order to remove much duplication of code and solve issues
2199 with the old code (see ssl/statem/README for further details). This change
2200 does have some associated API changes. Notably the SSL_state() function
2201 has been removed and replaced by SSL_get_state which now returns an
2202 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
2203 altogether. The previous handshake states defined in ssl.h and ssl3.h have
2207 *) All instances of the string "ssleay" in the public API were replaced
2208 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
2209 Some error codes related to internal RSA_eay API's were renamed.
2212 *) The demo files in crypto/threads were moved to demo/threads.
2215 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
2217 [Matt Caswell, Rich Salz]
2219 *) New ASN.1 embed macro.
2221 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
2222 structure is not allocated: it is part of the parent. That is instead of
2230 This reduces memory fragmentation and make it impossible to accidentally
2231 set a mandatory field to NULL.
2233 This currently only works for some fields specifically a SEQUENCE, CHOICE,
2234 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
2235 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
2239 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
2242 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
2243 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
2244 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
2245 DES and RC4 ciphersuites.
2248 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2249 This changes the decoding behaviour for some invalid messages,
2250 though the change is mostly in the more lenient direction, and
2251 legacy behaviour is preserved as much as possible.
2254 *) Fix no-stdio build.
2255 [ David Woodhouse <David.Woodhouse@intel.com> and also
2256 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
2258 *) New testing framework
2259 The testing framework has been largely rewritten and is now using
2260 perl and the perl modules Test::Harness and an extended variant of
2261 Test::More called OpenSSL::Test to do its work. All test scripts in
2262 test/ have been rewritten into test recipes, and all direct calls to
2263 executables in test/Makefile have become individual recipes using the
2264 simplified testing OpenSSL::Test::Simple.
2266 For documentation on our testing modules, do:
2268 perldoc test/testlib/OpenSSL/Test/Simple.pm
2269 perldoc test/testlib/OpenSSL/Test.pm
2273 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
2274 are used; the latter aborts on memory leaks (usually checked on exit).
2275 Some undocumented "set malloc, etc., hooks" functions were removed
2276 and others were changed. All are now documented.
2279 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2281 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2283 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
2284 from RFC4279, RFC4785, RFC5487, RFC5489.
2286 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
2287 original RSA_PSK patch.
2290 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
2291 era flag was never set throughout the codebase (only read). Also removed
2292 SSL3_FLAGS_POP_BUFFER which was only used if
2293 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
2296 *) Changed the default name options in the "ca", "crl", "req" and "x509"
2297 to be "oneline" instead of "compat".
2300 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
2301 not aware of clients that still exhibit this bug, and the workaround
2302 hasn't been working properly for a while.
2305 *) The return type of BIO_number_read() and BIO_number_written() as well as
2306 the corresponding num_read and num_write members in the BIO structure has
2307 changed from unsigned long to uint64_t. On platforms where an unsigned
2308 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
2312 *) Given the pervasive nature of TLS extensions it is inadvisable to run
2313 OpenSSL without support for them. It also means that maintaining
2314 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
2315 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
2318 *) Removed support for the two export grade static DH ciphersuites
2319 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
2320 were newly added (along with a number of other static DH ciphersuites) to
2321 1.0.2. However the two export ones have *never* worked since they were
2322 introduced. It seems strange in any case to be adding new export
2323 ciphersuites, and given "logjam" it also does not seem correct to fix them.
2326 *) Version negotiation has been rewritten. In particular SSLv23_method(),
2327 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
2328 and turned into macros which simply call the new preferred function names
2329 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
2330 should use the new names instead. Also as part of this change the ssl23.h
2331 header file has been removed.
2334 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
2335 code and the associated standard is no longer considered fit-for-purpose.
2338 *) RT2547 was closed. When generating a private key, try to make the
2339 output file readable only by the owner. This behavior change might
2340 be noticeable when interacting with other software.
2342 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
2346 *) Added HTTP GET support to the ocsp command.
2349 *) Changed default digest for the dgst and enc commands from MD5 to
2353 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
2356 *) Added support for TLS extended master secret from
2357 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
2358 initial patch which was a great help during development.
2361 *) All libssl internal structures have been removed from the public header
2362 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
2363 now redundant). Users should not attempt to access internal structures
2364 directly. Instead they should use the provided API functions.
2367 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
2368 Access to deprecated functions can be re-enabled by running config with
2369 "enable-deprecated". In addition applications wishing to use deprecated
2370 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
2371 will, by default, disable some transitive includes that previously existed
2372 in the header files (e.g. ec.h will no longer, by default, include bn.h)
2375 *) Added support for OCB mode. OpenSSL has been granted a patent license
2376 compatible with the OpenSSL license for use of OCB. Details are available
2377 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
2378 for OCB can be removed by calling config with no-ocb.
2381 *) SSLv2 support has been removed. It still supports receiving a SSLv2
2382 compatible client hello.
2385 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
2386 done while fixing the error code for the key-too-small case.
2387 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
2389 *) CA.sh has been removed; use CA.pl instead.
2392 *) Removed old DES API.
2395 *) Remove various unsupported platforms:
2401 Sinix/ReliantUNIX RM400
2406 16-bit platforms such as WIN16
2409 *) Clean up OPENSSL_NO_xxx #define's
2410 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
2411 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
2412 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
2413 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
2414 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
2415 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
2416 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
2417 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
2418 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
2419 Remove MS_STATIC; it's a relic from platforms <32 bits.
2422 *) Cleaned up dead code
2423 Remove all but one '#ifdef undef' which is to be looked at.
2426 *) Clean up calling of xxx_free routines.
2427 Just like free(), fix most of the xxx_free routines to accept
2428 NULL. Remove the non-null checks from callers. Save much code.
2431 *) Add secure heap for storage of private keys (when possible).
2432 Add BIO_s_secmem(), CBIGNUM, etc.
2433 Contributed by Akamai Technologies under our Corporate CLA.
2436 *) Experimental support for a new, fast, unbiased prime candidate generator,
2437 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
2438 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
2440 *) New output format NSS in the sess_id command line tool. This allows
2441 exporting the session id and the master key in NSS keylog format.
2442 [Martin Kaiser <martin@kaiser.cx>]
2444 *) Harmonize version and its documentation. -f flag is used to display
2446 [mancha <mancha1@zoho.com>]
2448 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2449 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
2450 [mancha <mancha1@zoho.com>]
2452 *) Fix some double frees. These are not thought to be exploitable.
2453 [mancha <mancha1@zoho.com>]
2455 *) A missing bounds check in the handling of the TLS heartbeat extension
2456 can be used to reveal up to 64k of memory to a connected client or
2459 Thanks for Neel Mehta of Google Security for discovering this bug and to
2460 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2461 preparing the fix (CVE-2014-0160)
2462 [Adam Langley, Bodo Moeller]
2464 *) Fix for the attack described in the paper "Recovering OpenSSL
2465 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2466 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2467 http://eprint.iacr.org/2014/140
2469 Thanks to Yuval Yarom and Naomi Benger for discovering this
2470 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2471 [Yuval Yarom and Naomi Benger]
2473 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2474 this fixes a limitation in previous versions of OpenSSL.
2477 *) Experimental encrypt-then-mac support.
2479 Experimental support for encrypt then mac from
2480 draft-gutmann-tls-encrypt-then-mac-02.txt
2482 To enable it set the appropriate extension number (0x42 for the test
2483 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
2485 For non-compliant peers (i.e. just about everything) this should have no
2488 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
2492 *) Add EVP support for key wrapping algorithms, to avoid problems with
2493 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2494 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2495 algorithms and include tests cases.
2498 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2502 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2503 MGF1 digest and OAEP label.
2506 *) Make openssl verify return errors.
2507 [Chris Palmer <palmer@google.com> and Ben Laurie]
2509 *) New function ASN1_TIME_diff to calculate the difference between two
2510 ASN1_TIME structures or one structure and the current time.
2513 *) Update fips_test_suite to support multiple command line options. New
2514 test to induce all self test errors in sequence and check expected
2518 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2519 sign or verify all in one operation.
2522 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2523 test programs and fips_test_suite. Includes functionality to parse
2524 the minimal script output of fipsalgest.pl directly.
2527 *) Add authorisation parameter to FIPS_module_mode_set().
2530 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2533 *) Use separate DRBG fields for internal and external flags. New function
2534 FIPS_drbg_health_check() to perform on demand health checking. Add
2535 generation tests to fips_test_suite with reduced health check interval to
2536 demonstrate periodic health checking. Add "nodh" option to
2537 fips_test_suite to skip very slow DH test.
2540 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2544 *) More extensive health check for DRBG checking many more failure modes.
2545 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2546 combination: call this in fips_test_suite.
2549 *) Add support for canonical generation of DSA parameter 'g'. See
2552 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2553 POST to handle HMAC cases.
2556 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2557 to return numerical and string versions of the FIPS module number.
2560 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2561 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2562 outside the validated module in the FIPS capable OpenSSL.
2565 *) Minor change to DRBG entropy callback semantics. In some cases
2566 there is no multiple of the block length between min_len and
2567 max_len. Allow the callback to return more than max_len bytes
2568 of entropy but discard any extra: it is the callback's responsibility
2569 to ensure that the extra data discarded does not impact the
2570 requested amount of entropy.
2573 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2574 information in FIPS186-3, SP800-57 and SP800-131A.
2577 *) CCM support via EVP. Interface is very similar to GCM case except we
2578 must supply all data in one chunk (i.e. no update, final) and the
2579 message length must be supplied if AAD is used. Add algorithm test
2583 *) Initial version of POST overhaul. Add POST callback to allow the status
2584 of POST to be monitored and/or failures induced. Modify fips_test_suite
2585 to use callback. Always run all selftests even if one fails.
2588 *) XTS support including algorithm test driver in the fips_gcmtest program.
2589 Note: this does increase the maximum key length from 32 to 64 bytes but
2590 there should be no binary compatibility issues as existing applications
2591 will never use XTS mode.
2594 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2595 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2596 performs algorithm blocking for unapproved PRNG types. Also do not
2597 set PRNG type in FIPS_mode_set(): leave this to the application.
2598 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2599 the standard OpenSSL PRNG: set additional data to a date time vector.
2602 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2603 This shouldn't present any incompatibility problems because applications
2604 shouldn't be using these directly and any that are will need to rethink
2605 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2608 *) Extensive self tests and health checking required by SP800-90 DRBG.
2609 Remove strength parameter from FIPS_drbg_instantiate and always
2610 instantiate at maximum supported strength.
2613 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2616 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2619 *) New function DH_compute_key_padded() to compute a DH key and pad with
2620 leading zeroes if needed: this complies with SP800-56A et al.
2623 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2624 anything, incomplete, subject to change and largely untested at present.
2627 *) Modify fipscanisteronly build option to only build the necessary object
2628 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2631 *) Add experimental option FIPSSYMS to give all symbols in
2632 fipscanister.o and FIPS or fips prefix. This will avoid
2633 conflicts with future versions of OpenSSL. Add perl script
2634 util/fipsas.pl to preprocess assembly language source files
2635 and rename any affected symbols.
2638 *) Add selftest checks and algorithm block of non-fips algorithms in
2639 FIPS mode. Remove DES2 from selftests.
2642 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2643 return internal method without any ENGINE dependencies. Add new
2644 tiny fips sign and verify functions.
2647 *) New build option no-ec2m to disable characteristic 2 code.
2650 *) New build option "fipscanisteronly". This only builds fipscanister.o
2651 and (currently) associated fips utilities. Uses the file Makefile.fips
2652 instead of Makefile.org as the prototype.
2655 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2656 Update fips_gcmtest to use IV generator.
2659 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2660 setting output buffer to NULL. The *Final function must be
2661 called although it will not retrieve any additional data. The tag
2662 can be set or retrieved with a ctrl. The IV length is by default 12
2663 bytes (96 bits) but can be set to an alternative value. If the IV
2664 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2668 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2669 underlying do_cipher function handles all cipher semantics itself
2670 including padding and finalisation. This is useful if (for example)
2671 an ENGINE cipher handles block padding itself. The behaviour of
2672 do_cipher is subtly changed if this flag is set: the return value
2673 is the number of characters written to the output buffer (zero is
2674 no longer an error code) or a negative error code. Also if the
2675 input buffer is NULL and length 0 finalisation should be performed.
2678 *) If a candidate issuer certificate is already part of the constructed
2679 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2682 *) Improve forward-security support: add functions
2684 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2685 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2687 for use by SSL/TLS servers; the callback function will be called whenever a
2688 new session is created, and gets to decide whether the session may be
2689 cached to make it resumable (return 0) or not (return 1). (As by the
2690 SSL/TLS protocol specifications, the session_id sent by the server will be
2691 empty to indicate that the session is not resumable; also, the server will
2692 not generate RFC 4507 (RFC 5077) session tickets.)
2694 A simple reasonable callback implementation is to return is_forward_secure.
2695 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2696 by the SSL/TLS server library, indicating whether it can provide forward
2698 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2700 *) New -verify_name option in command line utilities to set verification
2704 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2705 Add CMAC pkey methods.
2708 *) Experimental renegotiation in s_server -www mode. If the client
2709 browses /reneg connection is renegotiated. If /renegcert it is
2710 renegotiated requesting a certificate.
2713 *) Add an "external" session cache for debugging purposes to s_server. This
2714 should help trace issues which normally are only apparent in deployed
2715 multi-process servers.
2718 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2719 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2720 BIO_set_cipher() and some obscure PEM functions were changed so they
2721 can now return an error. The RAND changes required a change to the
2722 RAND_METHOD structure.
2725 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2726 a gcc attribute to warn if the result of a function is ignored. This
2727 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2728 whose return value is often ignored.
2731 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2732 These allow SCTs (signed certificate timestamps) to be requested and
2733 validated when establishing a connection.
2734 [Rob Percival <robpercival@google.com>]
2736 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2738 *) Prevent padding oracle in AES-NI CBC MAC check
2740 A MITM attacker can use a padding oracle attack to decrypt traffic
2741 when the connection uses an AES CBC cipher and the server support
2744 This issue was introduced as part of the fix for Lucky 13 padding
2745 attack (CVE-2013-0169). The padding check was rewritten to be in
2746 constant time by making sure that always the same bytes are read and
2747 compared against either the MAC or padding bytes. But it no longer
2748 checked that there was enough data to have both the MAC and padding
2751 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2755 *) Fix EVP_EncodeUpdate overflow
2757 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2758 Base64 encoding of binary data. If an attacker is able to supply very large
2759 amounts of input data then a length check can overflow resulting in a heap
2762 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2763 the PEM_write_bio* family of functions. These are mainly used within the
2764 OpenSSL command line applications, so any application which processes data
2765 from an untrusted source and outputs it as a PEM file should be considered
2766 vulnerable to this issue. User applications that call these APIs directly
2767 with large amounts of untrusted data may also be vulnerable.
2769 This issue was reported by Guido Vranken.
2773 *) Fix EVP_EncryptUpdate overflow
2775 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2776 is able to supply very large amounts of input data after a previous call to
2777 EVP_EncryptUpdate() with a partial block then a length check can overflow
2778 resulting in a heap corruption. Following an analysis of all OpenSSL
2779 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2780 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2781 the first called function after an EVP_EncryptInit(), and therefore that
2782 specific call must be safe. The second form is where the length passed to
2783 EVP_EncryptUpdate() can be seen from the code to be some small value and
2784 therefore there is no possibility of an overflow. Since all instances are
2785 one of these two forms, it is believed that there can be no overflows in
2786 internal code due to this problem. It should be noted that
2787 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2788 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2789 of these calls have also been analysed too and it is believed there are no
2790 instances in internal usage where an overflow could occur.
2792 This issue was reported by Guido Vranken.
2796 *) Prevent ASN.1 BIO excessive memory allocation
2798 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2799 a short invalid encoding can cause allocation of large amounts of memory
2800 potentially consuming excessive resources or exhausting memory.
2802 Any application parsing untrusted data through d2i BIO functions is
2803 affected. The memory based functions such as d2i_X509() are *not* affected.
2804 Since the memory based functions are used by the TLS library, TLS
2805 applications are not affected.
2807 This issue was reported by Brian Carpenter.
2813 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2814 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2815 in arbitrary stack data being returned in the buffer.
2817 This issue was reported by Guido Vranken.
2821 *) Modify behavior of ALPN to invoke callback after SNI/servername
2822 callback, such that updates to the SSL_CTX affect ALPN.
2825 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2829 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2830 methods are enabled and ssl2 is disabled the methods return NULL.
2833 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2835 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2836 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2837 provide any "EXPORT" or "LOW" strength ciphers.
2840 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2841 is by default disabled at build-time. Builds that are not configured with
2842 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2843 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2844 will need to explicitly call either of:
2846 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2848 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2850 as appropriate. Even if either of those is used, or the application
2851 explicitly uses the version-specific SSLv2_method() or its client and
2852 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2853 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2854 ciphers, and SSLv2 56-bit DES are no longer available.
2858 *) Fix a double-free in DSA code
2860 A double free bug was discovered when OpenSSL parses malformed DSA private
2861 keys and could lead to a DoS attack or memory corruption for applications
2862 that receive DSA private keys from untrusted sources. This scenario is
2865 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2870 *) Disable SRP fake user seed to address a server memory leak.
2872 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2874 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2875 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2876 was changed to ignore the "fake user" SRP seed, even if the seed
2879 Users should use SRP_VBASE_get1_by_user instead. Note that in
2880 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2881 also that even though configuring the SRP seed attempts to hide
2882 invalid usernames by continuing the handshake with fake
2883 credentials, this behaviour is not constant time and no strong
2884 guarantees are made that the handshake is indistinguishable from
2885 that of a valid user.
2889 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2891 In the BN_hex2bn function the number of hex digits is calculated using an
2892 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2893 large values of |i| this can result in |bn_expand| not allocating any
2894 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2895 field as NULL leading to a subsequent NULL ptr deref. For very large values
2896 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2897 In this case memory is allocated to the internal BIGNUM data field, but it
2898 is insufficiently sized leading to heap corruption. A similar issue exists
2899 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2900 is ever called by user applications with very large untrusted hex/dec data.
2901 This is anticipated to be a rare occurrence.
2903 All OpenSSL internal usage of these functions use data that is not expected
2904 to be untrusted, e.g. config file data or application command line
2905 arguments. If user developed applications generate config file data based
2906 on untrusted data then it is possible that this could also lead to security
2907 consequences. This is also anticipated to be rare.
2909 This issue was reported to OpenSSL by Guido Vranken.
2913 *) Fix memory issues in BIO_*printf functions
2915 The internal |fmtstr| function used in processing a "%s" format string in
2916 the BIO_*printf functions could overflow while calculating the length of a
2917 string and cause an OOB read when printing very long strings.
2919 Additionally the internal |doapr_outch| function can attempt to write to an
2920 OOB memory location (at an offset from the NULL pointer) in the event of a
2921 memory allocation failure. In 1.0.2 and below this could be caused where
2922 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2923 could be in processing a very long "%s" format string. Memory leaks can
2926 The first issue may mask the second issue dependent on compiler behaviour.
2927 These problems could enable attacks where large amounts of untrusted data
2928 is passed to the BIO_*printf functions. If applications use these functions
2929 in this way then they could be vulnerable. OpenSSL itself uses these
2930 functions when printing out human-readable dumps of ASN.1 data. Therefore
2931 applications that print this data could be vulnerable if the data is from
2932 untrusted sources. OpenSSL command line applications could also be
2933 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2934 as command line arguments.
2936 Libssl is not considered directly vulnerable. Additionally certificates etc
2937 received via remote connections via libssl are also unlikely to be able to
2938 trigger these issues because of message size limits enforced within libssl.
2940 This issue was reported to OpenSSL Guido Vranken.
2944 *) Side channel attack on modular exponentiation
2946 A side-channel attack was found which makes use of cache-bank conflicts on
2947 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2948 of RSA keys. The ability to exploit this issue is limited as it relies on
2949 an attacker who has control of code in a thread running on the same
2950 hyper-threaded core as the victim thread which is performing decryptions.
2952 This issue was reported to OpenSSL by Yuval Yarom, The University of
2953 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2954 Nadia Heninger, University of Pennsylvania with more information at
2955 http://cachebleed.info.
2959 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2960 if no keysize is specified with default_bits. This fixes an
2961 omission in an earlier change that changed all RSA/DSA key generation
2962 apps to use 2048 bits by default.
2965 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2966 *) DH small subgroups
2968 Historically OpenSSL only ever generated DH parameters based on "safe"
2969 primes. More recently (in version 1.0.2) support was provided for
2970 generating X9.42 style parameter files such as those required for RFC 5114
2971 support. The primes used in such files may not be "safe". Where an
2972 application is using DH configured with parameters based on primes that are
2973 not "safe" then an attacker could use this fact to find a peer's private
2974 DH exponent. This attack requires that the attacker complete multiple
2975 handshakes in which the peer uses the same private DH exponent. For example
2976 this could be used to discover a TLS server's private DH exponent if it's
2977 reusing the private DH exponent or it's using a static DH ciphersuite.
2979 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2980 TLS. It is not on by default. If the option is not set then the server
2981 reuses the same private DH exponent for the life of the server process and
2982 would be vulnerable to this attack. It is believed that many popular
2983 applications do set this option and would therefore not be at risk.
2985 The fix for this issue adds an additional check where a "q" parameter is
2986 available (as is the case in X9.42 based parameters). This detects the
2987 only known attack, and is the only possible defense for static DH
2988 ciphersuites. This could have some performance impact.
2990 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2991 default and cannot be disabled. This could have some performance impact.
2993 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2997 *) SSLv2 doesn't block disabled ciphers
2999 A malicious client can negotiate SSLv2 ciphers that have been disabled on
3000 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
3001 been disabled, provided that the SSLv2 protocol was not also disabled via
3004 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
3005 and Sebastian Schinzel.
3009 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
3011 *) BN_mod_exp may produce incorrect results on x86_64
3013 There is a carry propagating bug in the x86_64 Montgomery squaring
3014 procedure. No EC algorithms are affected. Analysis suggests that attacks
3015 against RSA and DSA as a result of this defect would be very difficult to
3016 perform and are not believed likely. Attacks against DH are considered just
3017 feasible (although very difficult) because most of the work necessary to
3018 deduce information about a private key may be performed offline. The amount
3019 of resources required for such an attack would be very significant and
3020 likely only accessible to a limited number of attackers. An attacker would
3021 additionally need online access to an unpatched system using the target
3022 private key in a scenario with persistent DH parameters and a private
3023 key that is shared between multiple clients. For example this can occur by
3024 default in OpenSSL DHE based SSL/TLS ciphersuites.
3026 This issue was reported to OpenSSL by Hanno Böck.
3030 *) Certificate verify crash with missing PSS parameter
3032 The signature verification routines will crash with a NULL pointer
3033 dereference if presented with an ASN.1 signature using the RSA PSS
3034 algorithm and absent mask generation function parameter. Since these
3035 routines are used to verify certificate signature algorithms this can be
3036 used to crash any certificate verification operation and exploited in a
3037 DoS attack. Any application which performs certificate verification is
3038 vulnerable including OpenSSL clients and servers which enable client
3041 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
3045 *) X509_ATTRIBUTE memory leak
3047 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
3048 memory. This structure is used by the PKCS#7 and CMS routines so any
3049 application which reads PKCS#7 or CMS data from untrusted sources is
3050 affected. SSL/TLS is not affected.
3052 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
3057 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3058 This changes the decoding behaviour for some invalid messages,
3059 though the change is mostly in the more lenient direction, and
3060 legacy behaviour is preserved as much as possible.
3063 *) In DSA_generate_parameters_ex, if the provided seed is too short,
3065 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
3067 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
3069 *) Alternate chains certificate forgery
3071 During certificate verification, OpenSSL will attempt to find an
3072 alternative certificate chain if the first attempt to build such a chain
3073 fails. An error in the implementation of this logic can mean that an
3074 attacker could cause certain checks on untrusted certificates to be
3075 bypassed, such as the CA flag, enabling them to use a valid leaf
3076 certificate to act as a CA and "issue" an invalid certificate.
3078 This issue was reported to OpenSSL by Adam Langley/David Benjamin
3082 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
3084 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
3085 incompatibility in the handling of HMAC. The previous ABI has now been
3089 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
3091 *) Malformed ECParameters causes infinite loop
3093 When processing an ECParameters structure OpenSSL enters an infinite loop
3094 if the curve specified is over a specially malformed binary polynomial
3097 This can be used to perform denial of service against any
3098 system which processes public keys, certificate requests or
3099 certificates. This includes TLS clients and TLS servers with
3100 client authentication enabled.
3102 This issue was reported to OpenSSL by Joseph Barr-Pixton.
3106 *) Exploitable out-of-bounds read in X509_cmp_time
3108 X509_cmp_time does not properly check the length of the ASN1_TIME
3109 string and can read a few bytes out of bounds. In addition,
3110 X509_cmp_time accepts an arbitrary number of fractional seconds in the
3113 An attacker can use this to craft malformed certificates and CRLs of
3114 various sizes and potentially cause a segmentation fault, resulting in
3115 a DoS on applications that verify certificates or CRLs. TLS clients
3116 that verify CRLs are affected. TLS clients and servers with client
3117 authentication enabled may be affected if they use custom verification
3120 This issue was reported to OpenSSL by Robert Swiecki (Google), and
3121 independently by Hanno Böck.
3125 *) PKCS7 crash with missing EnvelopedContent
3127 The PKCS#7 parsing code does not handle missing inner EncryptedContent
3128 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
3129 with missing content and trigger a NULL pointer dereference on parsing.
3131 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
3132 structures from untrusted sources are affected. OpenSSL clients and
3133 servers are not affected.
3135 This issue was reported to OpenSSL by Michal Zalewski (Google).
3139 *) CMS verify infinite loop with unknown hash function
3141 When verifying a signedData message the CMS code can enter an infinite loop
3142 if presented with an unknown hash function OID. This can be used to perform
3143 denial of service against any system which verifies signedData messages using
3145 This issue was reported to OpenSSL by Johannes Bauer.
3149 *) Race condition handling NewSessionTicket
3151 If a NewSessionTicket is received by a multi-threaded client when attempting to
3152 reuse a previous ticket then a race condition can occur potentially leading to
3153 a double free of the ticket data.
3157 *) Only support 256-bit or stronger elliptic curves with the
3158 'ecdh_auto' setting (server) or by default (client). Of supported
3159 curves, prefer P-256 (both).
3162 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
3164 *) ClientHello sigalgs DoS fix
3166 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
3167 invalid signature algorithms extension a NULL pointer dereference will
3168 occur. This can be exploited in a DoS attack against the server.
3170 This issue was was reported to OpenSSL by David Ramos of Stanford
3173 [Stephen Henson and Matt Caswell]
3175 *) Multiblock corrupted pointer fix
3177 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
3178 feature only applies on 64 bit x86 architecture platforms that support AES
3179 NI instructions. A defect in the implementation of "multiblock" can cause
3180 OpenSSL's internal write buffer to become incorrectly set to NULL when
3181 using non-blocking IO. Typically, when the user application is using a
3182 socket BIO for writing, this will only result in a failed connection.
3183 However if some other BIO is used then it is likely that a segmentation
3184 fault will be triggered, thus enabling a potential DoS attack.
3186 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
3190 *) Segmentation fault in DTLSv1_listen fix
3192 The DTLSv1_listen function is intended to be stateless and processes the
3193 initial ClientHello from many peers. It is common for user code to loop
3194 over the call to DTLSv1_listen until a valid ClientHello is received with
3195 an associated cookie. A defect in the implementation of DTLSv1_listen means
3196 that state is preserved in the SSL object from one invocation to the next
3197 that can lead to a segmentation fault. Errors processing the initial
3198 ClientHello can trigger this scenario. An example of such an error could be
3199 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
3202 This issue was reported to OpenSSL by Per Allansson.
3206 *) Segmentation fault in ASN1_TYPE_cmp fix
3208 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
3209 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
3210 certificate signature algorithm consistency this can be used to crash any
3211 certificate verification operation and exploited in a DoS attack. Any
3212 application which performs certificate verification is vulnerable including
3213 OpenSSL clients and servers which enable client authentication.
3217 *) Segmentation fault for invalid PSS parameters fix
3219 The signature verification routines will crash with a NULL pointer
3220 dereference if presented with an ASN.1 signature using the RSA PSS
3221 algorithm and invalid parameters. Since these routines are used to verify
3222 certificate signature algorithms this can be used to crash any
3223 certificate verification operation and exploited in a DoS attack. Any
3224 application which performs certificate verification is vulnerable including
3225 OpenSSL clients and servers which enable client authentication.
3227 This issue was was reported to OpenSSL by Brian Carpenter.
3231 *) ASN.1 structure reuse memory corruption fix
3233 Reusing a structure in ASN.1 parsing may allow an attacker to cause
3234 memory corruption via an invalid write. Such reuse is and has been
3235 strongly discouraged and is believed to be rare.
3237 Applications that parse structures containing CHOICE or ANY DEFINED BY
3238 components may be affected. Certificate parsing (d2i_X509 and related
3239 functions) are however not affected. OpenSSL clients and servers are
3244 *) PKCS7 NULL pointer dereferences fix
3246 The PKCS#7 parsing code does not handle missing outer ContentInfo
3247 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
3248 missing content and trigger a NULL pointer dereference on parsing.
3250 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
3251 otherwise parse PKCS#7 structures from untrusted sources are
3252 affected. OpenSSL clients and servers are not affected.
3254 This issue was reported to OpenSSL by Michal Zalewski (Google).
3258 *) DoS via reachable assert in SSLv2 servers fix
3260 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
3261 servers that both support SSLv2 and enable export cipher suites by sending
3262 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
3264 This issue was discovered by Sean Burford (Google) and Emilia Käsper
3265 (OpenSSL development team).
3269 *) Empty CKE with client auth and DHE fix
3271 If client auth is used then a server can seg fault in the event of a DHE
3272 ciphersuite being selected and a zero length ClientKeyExchange message
3273 being sent by the client. This could be exploited in a DoS attack.
3277 *) Handshake with unseeded PRNG fix
3279 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
3280 with an unseeded PRNG. The conditions are:
3281 - The client is on a platform where the PRNG has not been seeded
3282 automatically, and the user has not seeded manually
3283 - A protocol specific client method version has been used (i.e. not
3284 SSL_client_methodv23)
3285 - A ciphersuite is used that does not require additional random data from
3286 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
3288 If the handshake succeeds then the client random that has been used will
3289 have been generated from a PRNG with insufficient entropy and therefore the
3290 output may be predictable.
3292 For example using the following command with an unseeded openssl will
3293 succeed on an unpatched platform:
3295 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
3299 *) Use After Free following d2i_ECPrivatekey error fix
3301 A malformed EC private key file consumed via the d2i_ECPrivateKey function
3302 could cause a use after free condition. This, in turn, could cause a double
3303 free in several private key parsing functions (such as d2i_PrivateKey
3304 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
3305 for applications that receive EC private keys from untrusted
3306 sources. This scenario is considered rare.
3308 This issue was discovered by the BoringSSL project and fixed in their
3313 *) X509_to_X509_REQ NULL pointer deref fix
3315 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
3316 the certificate key is invalid. This function is rarely used in practice.
3318 This issue was discovered by Brian Carpenter.
3322 *) Removed the export ciphers from the DEFAULT ciphers
3325 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
3327 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
3328 ARMv5 through ARMv8, as opposite to "locking" it to single one.
3329 So far those who have to target multiple platforms would compromise
3330 and argue that binary targeting say ARMv5 would still execute on
3331 ARMv8. "Universal" build resolves this compromise by providing
3332 near-optimal performance even on newer platforms.
3335 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
3336 (other platforms pending).
3337 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
3339 *) Add support for the SignedCertificateTimestampList certificate and
3340 OCSP response extensions from RFC6962.
3343 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3344 for corner cases. (Certain input points at infinity could lead to
3345 bogus results, with non-infinity inputs mapped to infinity too.)
3348 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
3349 This covers AES, SHA256/512 and GHASH. "Initial" means that most
3350 common cases are optimized and there still is room for further
3351 improvements. Vector Permutation AES for Altivec is also added.
3354 *) Add support for little-endian ppc64 Linux target.
3355 [Marcelo Cerri (IBM)]
3357 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
3358 SHA1, SHA256 and GHASH. "Initial" means that most common cases
3359 are optimized and there still is room for further improvements.
3360 Both 32- and 64-bit modes are supported.
3361 [Andy Polyakov, Ard Biesheuvel (Linaro)]
3363 *) Improved ARMv7 NEON support.
3366 *) Support for SPARC Architecture 2011 crypto extensions, first
3367 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
3368 SHA256/512, MD5, GHASH and modular exponentiation.
3369 [Andy Polyakov, David Miller]
3371 *) Accelerated modular exponentiation for Intel processors, a.k.a.
3373 [Shay Gueron & Vlad Krasnov (Intel Corp)]
3375 *) Support for new and upcoming Intel processors, including AVX2,
3376 BMI and SHA ISA extensions. This includes additional "stitched"
3377 implementations, AESNI-SHA256 and GCM, and multi-buffer support
3380 This work was sponsored by Intel Corp.
3383 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
3384 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
3385 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
3388 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3389 this fixes a limitation in previous versions of OpenSSL.
3392 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3393 MGF1 digest and OAEP label.
3396 *) Add EVP support for key wrapping algorithms, to avoid problems with
3397 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3398 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3399 algorithms and include tests cases.
3402 *) Add functions to allocate and set the fields of an ECDSA_METHOD
3404 [Douglas E. Engert, Steve Henson]
3406 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
3407 difference in days and seconds between two tm or ASN1_TIME structures.
3410 *) Add -rev test option to s_server to just reverse order of characters
3411 received by client and send back to server. Also prints an abbreviated
3412 summary of the connection parameters.
3415 *) New option -brief for s_client and s_server to print out a brief summary
3416 of connection parameters.
3419 *) Add callbacks for arbitrary TLS extensions.
3420 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
3422 *) New option -crl_download in several openssl utilities to download CRLs
3423 from CRLDP extension in certificates.
3426 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
3429 *) New function X509_CRL_diff to generate a delta CRL from the difference
3430 of two full CRLs. Add support to "crl" utility.
3433 *) New functions to set lookup_crls function and to retrieve
3434 X509_STORE from X509_STORE_CTX.
3437 *) Print out deprecated issuer and subject unique ID fields in
3441 *) Extend OCSP I/O functions so they can be used for simple general purpose
3442 HTTP as well as OCSP. New wrapper function which can be used to download
3443 CRLs using the OCSP API.
3446 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
3449 *) SSL_CONF* functions. These provide a common framework for application
3450 configuration using configuration files or command lines.
3453 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
3454 message callback and prints the results. Needs compile time option
3455 "enable-ssl-trace". New options to s_client and s_server to enable
3459 *) New ctrl and macro to retrieve supported points extensions.
3460 Print out extension in s_server and s_client.
3463 *) New functions to retrieve certificate signature and signature
3467 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
3471 *) New Suite B modes for TLS code. These use and enforce the requirements
3472 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
3473 only use Suite B curves. The Suite B modes can be set by using the
3474 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
3477 *) New chain verification flags for Suite B levels of security. Check
3478 algorithms are acceptable when flags are set in X509_verify_cert.
3481 *) Make tls1_check_chain return a set of flags indicating checks passed
3482 by a certificate chain. Add additional tests to handle client
3483 certificates: checks for matching certificate type and issuer name
3487 *) If an attempt is made to use a signature algorithm not in the peer
3488 preference list abort the handshake. If client has no suitable
3489 signature algorithms in response to a certificate request do not
3490 use the certificate.
3493 *) If server EC tmp key is not in client preference list abort handshake.
3496 *) Add support for certificate stores in CERT structure. This makes it
3497 possible to have different stores per SSL structure or one store in
3498 the parent SSL_CTX. Include distinct stores for certificate chain
3499 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3500 to build and store a certificate chain in CERT structure: returning
3501 an error if the chain cannot be built: this will allow applications
3502 to test if a chain is correctly configured.
3504 Note: if the CERT based stores are not set then the parent SSL_CTX
3505 store is used to retain compatibility with existing behaviour.
3509 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3510 mask based on the current session, check mask when sending client
3511 hello and checking the requested ciphersuite.
3514 *) New ctrls to retrieve and set certificate types in a certificate
3515 request message. Print out received values in s_client. If certificate
3516 types is not set with custom values set sensible values based on
3517 supported signature algorithms.
3520 *) Support for distinct client and server supported signature algorithms.
3523 *) Add certificate callback. If set this is called whenever a certificate
3524 is required by client or server. An application can decide which
3525 certificate chain to present based on arbitrary criteria: for example
3526 supported signature algorithms. Add very simple example to s_server.
3527 This fixes many of the problems and restrictions of the existing client
3528 certificate callback: for example you can now clear an existing
3529 certificate and specify the whole chain.
3532 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3533 the certificate can be used for (if anything). Set valid_flags field
3534 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3535 to have similar checks in it.
3537 Add new "cert_flags" field to CERT structure and include a "strict mode".
3538 This enforces some TLS certificate requirements (such as only permitting
3539 certificate signature algorithms contained in the supported algorithms
3540 extension) which some implementations ignore: this option should be used
3541 with caution as it could cause interoperability issues.
3544 *) Update and tidy signature algorithm extension processing. Work out
3545 shared signature algorithms based on preferences and peer algorithms
3546 and print them out in s_client and s_server. Abort handshake if no
3547 shared signature algorithms.
3550 *) Add new functions to allow customised supported signature algorithms
3551 for SSL and SSL_CTX structures. Add options to s_client and s_server
3555 *) New function SSL_certs_clear() to delete all references to certificates
3556 from an SSL structure. Before this once a certificate had been added
3557 it couldn't be removed.
3560 *) Integrate hostname, email address and IP address checking with certificate
3561 verification. New verify options supporting checking in openssl utility.
3564 *) Fixes and wildcard matching support to hostname and email checking
3565 functions. Add manual page.
3566 [Florian Weimer (Red Hat Product Security Team)]
3568 *) New functions to check a hostname email or IP address against a
3569 certificate. Add options x509 utility to print results of checks against
3573 *) Fix OCSP checking.
3574 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3576 *) Initial experimental support for explicitly trusted non-root CAs.
3577 OpenSSL still tries to build a complete chain to a root but if an
3578 intermediate CA has a trust setting included that is used. The first
3579 setting is used: whether to trust (e.g., -addtrust option to the x509
3583 *) Add -trusted_first option which attempts to find certificates in the
3584 trusted store even if an untrusted chain is also supplied.
3587 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3588 platform support for Linux and Android.
3591 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3594 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3595 When in FIPS mode the approved implementations are used as normal,
3596 when not in FIPS mode the internal unapproved versions are used instead.
3597 This means that the FIPS capable OpenSSL isn't forced to use the
3598 (often lower performance) FIPS implementations outside FIPS mode.
3601 *) Transparently support X9.42 DH parameters when calling
3602 PEM_read_bio_DHparameters. This means existing applications can handle
3603 the new parameter format automatically.
3606 *) Initial experimental support for X9.42 DH parameter format: mainly
3607 to support use of 'q' parameter for RFC5114 parameters.
3610 *) Add DH parameters from RFC5114 including test data to dhtest.
3613 *) Support for automatic EC temporary key parameter selection. If enabled
3614 the most preferred EC parameters are automatically used instead of
3615 hardcoded fixed parameters. Now a server just has to call:
3616 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3617 support ECDH and use the most appropriate parameters.
3620 *) Enhance and tidy EC curve and point format TLS extension code. Use
3621 static structures instead of allocation if default values are used.
3622 New ctrls to set curves we wish to support and to retrieve shared curves.
3623 Print out shared curves in s_server. New options to s_server and s_client
3624 to set list of supported curves.
3627 *) New ctrls to retrieve supported signature algorithms and
3628 supported curve values as an array of NIDs. Extend openssl utility
3629 to print out received values.
3632 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3633 between NIDs and the more common NIST names such as "P-256". Enhance
3634 ecparam utility and ECC method to recognise the NIST names for curves.
3637 *) Enhance SSL/TLS certificate chain handling to support different
3638 chains for each certificate instead of one chain in the parent SSL_CTX.
3641 *) Support for fixed DH ciphersuite client authentication: where both
3642 server and client use DH certificates with common parameters.
3645 *) Support for fixed DH ciphersuites: those requiring DH server
3649 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3651 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3652 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3653 X509_CINF_get_signature were reverted post internal team review.
3655 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3657 *) Build fixes for the Windows and OpenVMS platforms
3658 [Matt Caswell and Richard Levitte]
3660 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3662 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3663 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3664 dereference. This could lead to a Denial Of Service attack. Thanks to
3665 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3669 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3670 dtls1_buffer_record function under certain conditions. In particular this
3671 could occur if an attacker sent repeated DTLS records with the same
3672 sequence number but for the next epoch. The memory leak could be exploited
3673 by an attacker in a Denial of Service attack through memory exhaustion.
3674 Thanks to Chris Mueller for reporting this issue.
3678 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3679 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3680 method would be set to NULL which could later result in a NULL pointer
3681 dereference. Thanks to Frank Schmirler for reporting this issue.
3685 *) Abort handshake if server key exchange message is omitted for ephemeral
3688 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3689 reporting this issue.
3693 *) Remove non-export ephemeral RSA code on client and server. This code
3694 violated the TLS standard by allowing the use of temporary RSA keys in
3695 non-export ciphersuites and could be used by a server to effectively
3696 downgrade the RSA key length used to a value smaller than the server
3697 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3698 INRIA or reporting this issue.
3702 *) Fixed issue where DH client certificates are accepted without verification.
3703 An OpenSSL server will accept a DH certificate for client authentication
3704 without the certificate verify message. This effectively allows a client to
3705 authenticate without the use of a private key. This only affects servers
3706 which trust a client certificate authority which issues certificates
3707 containing DH keys: these are extremely rare and hardly ever encountered.
3708 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3713 *) Ensure that the session ID context of an SSL is updated when its
3714 SSL_CTX is updated via SSL_set_SSL_CTX.
3716 The session ID context is typically set from the parent SSL_CTX,
3717 and can vary with the CTX.
3720 *) Fix various certificate fingerprint issues.
3722 By using non-DER or invalid encodings outside the signed portion of a
3723 certificate the fingerprint can be changed without breaking the signature.
3724 Although no details of the signed portion of the certificate can be changed
3725 this can cause problems with some applications: e.g. those using the
3726 certificate fingerprint for blacklists.
3728 1. Reject signatures with non zero unused bits.
3730 If the BIT STRING containing the signature has non zero unused bits reject
3731 the signature. All current signature algorithms require zero unused bits.
3733 2. Check certificate algorithm consistency.
3735 Check the AlgorithmIdentifier inside TBS matches the one in the
3736 certificate signature. NB: this will result in signature failure
3737 errors for some broken certificates.
3739 Thanks to Konrad Kraszewski from Google for reporting this issue.
3741 3. Check DSA/ECDSA signatures use DER.
3743 Re-encode DSA/ECDSA signatures and compare with the original received
3744 signature. Return an error if there is a mismatch.
3746 This will reject various cases including garbage after signature
3747 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3748 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3749 (negative or with leading zeroes).
3751 Further analysis was conducted and fixes were developed by Stephen Henson
3752 of the OpenSSL core team.
3757 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3758 results on some platforms, including x86_64. This bug occurs at random
3759 with a very low probability, and is not known to be exploitable in any
3760 way, though its exact impact is difficult to determine. Thanks to Pieter
3761 Wuille (Blockstream) who reported this issue and also suggested an initial
3762 fix. Further analysis was conducted by the OpenSSL development team and
3763 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3764 the OpenSSL core team.
3768 *) Do not resume sessions on the server if the negotiated protocol
3769 version does not match the session's version. Resuming with a different
3770 version, while not strictly forbidden by the RFC, is of questionable
3771 sanity and breaks all known clients.
3772 [David Benjamin, Emilia Käsper]
3774 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3775 early CCS messages during renegotiation. (Note that because
3776 renegotiation is encrypted, this early CCS was not exploitable.)
3779 *) Tighten client-side session ticket handling during renegotiation:
3780 ensure that the client only accepts a session ticket if the server sends
3781 the extension anew in the ServerHello. Previously, a TLS client would
3782 reuse the old extension state and thus accept a session ticket if one was
3783 announced in the initial ServerHello.
3785 Similarly, ensure that the client requires a session ticket if one
3786 was advertised in the ServerHello. Previously, a TLS client would
3787 ignore a missing NewSessionTicket message.
3790 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3792 *) SRTP Memory Leak.
3794 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3795 sends a carefully crafted handshake message, to cause OpenSSL to fail
3796 to free up to 64k of memory causing a memory leak. This could be
3797 exploited in a Denial Of Service attack. This issue affects OpenSSL
3798 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3799 whether SRTP is used or configured. Implementations of OpenSSL that
3800 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3802 The fix was developed by the OpenSSL team.
3806 *) Session Ticket Memory Leak.
3808 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3809 integrity of that ticket is first verified. In the event of a session
3810 ticket integrity check failing, OpenSSL will fail to free memory
3811 causing a memory leak. By sending a large number of invalid session
3812 tickets an attacker could exploit this issue in a Denial Of Service
3817 *) Build option no-ssl3 is incomplete.
3819 When OpenSSL is configured with "no-ssl3" as a build option, servers
3820 could accept and complete a SSL 3.0 handshake, and clients could be
3821 configured to send them.
3823 [Akamai and the OpenSSL team]
3825 *) Add support for TLS_FALLBACK_SCSV.
3826 Client applications doing fallback retries should call
3827 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3829 [Adam Langley, Bodo Moeller]
3831 *) Add additional DigestInfo checks.
3833 Re-encode DigestInto in DER and check against the original when
3834 verifying RSA signature: this will reject any improperly encoded
3835 DigestInfo structures.
3837 Note: this is a precautionary measure and no attacks are currently known.
3841 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3843 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3844 SRP code can be overrun an internal buffer. Add sanity check that
3845 g, A, B < N to SRP code.
3847 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3848 Group for discovering this issue.
3852 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3853 TLS 1.0 instead of higher protocol versions when the ClientHello message
3854 is badly fragmented. This allows a man-in-the-middle attacker to force a
3855 downgrade to TLS 1.0 even if both the server and the client support a
3856 higher protocol version, by modifying the client's TLS records.
3858 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3859 researching this issue.
3863 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3864 to a denial of service attack. A malicious server can crash the client
3865 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3866 ciphersuite and sending carefully crafted handshake messages.
3868 Thanks to Felix Gröbert (Google) for discovering and researching this
3873 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3874 to leak memory. This can be exploited through a Denial of Service attack.
3875 Thanks to Adam Langley for discovering and researching this issue.
3879 *) An attacker can force openssl to consume large amounts of memory whilst
3880 processing DTLS handshake messages. This can be exploited through a
3881 Denial of Service attack.
3882 Thanks to Adam Langley for discovering and researching this issue.
3886 *) An attacker can force an error condition which causes openssl to crash
3887 whilst processing DTLS packets due to memory being freed twice. This
3888 can be exploited through a Denial of Service attack.
3889 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3894 *) If a multithreaded client connects to a malicious server using a resumed
3895 session and the server sends an ec point format extension it could write
3896 up to 255 bytes to freed memory.
3898 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3903 *) A malicious server can crash an OpenSSL client with a null pointer
3904 dereference (read) by specifying an SRP ciphersuite even though it was not
3905 properly negotiated with the client. This can be exploited through a
3906 Denial of Service attack.
3908 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3909 discovering and researching this issue.
3913 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3914 X509_name_oneline, X509_name_print_ex et al. to leak some information
3915 from the stack. Applications may be affected if they echo pretty printing
3916 output to the attacker.
3918 Thanks to Ivan Fratric (Google) for discovering this issue.
3920 [Emilia Käsper, and Steve Henson]
3922 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3923 for corner cases. (Certain input points at infinity could lead to
3924 bogus results, with non-infinity inputs mapped to infinity too.)
3927 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3929 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3930 handshake can force the use of weak keying material in OpenSSL
3931 SSL/TLS clients and servers.
3933 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3934 researching this issue. (CVE-2014-0224)
3935 [KIKUCHI Masashi, Steve Henson]
3937 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3938 OpenSSL DTLS client the code can be made to recurse eventually crashing
3941 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3943 [Imre Rad, Steve Henson]
3945 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3946 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3947 client or server. This is potentially exploitable to run arbitrary
3948 code on a vulnerable client or server.
3950 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3951 [Jüri Aedla, Steve Henson]
3953 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3954 are subject to a denial of service attack.
3956 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3957 this issue. (CVE-2014-3470)
3958 [Felix Gröbert, Ivan Fratric, Steve Henson]
3960 *) Harmonize version and its documentation. -f flag is used to display
3962 [mancha <mancha1@zoho.com>]
3964 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3965 in i2d_ECPrivateKey.
3966 [mancha <mancha1@zoho.com>]
3968 *) Fix some double frees. These are not thought to be exploitable.
3969 [mancha <mancha1@zoho.com>]
3971 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3973 *) A missing bounds check in the handling of the TLS heartbeat extension
3974 can be used to reveal up to 64k of memory to a connected client or
3977 Thanks for Neel Mehta of Google Security for discovering this bug and to
3978 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3979 preparing the fix (CVE-2014-0160)
3980 [Adam Langley, Bodo Moeller]
3982 *) Fix for the attack described in the paper "Recovering OpenSSL
3983 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3984 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3985 http://eprint.iacr.org/2014/140
3987 Thanks to Yuval Yarom and Naomi Benger for discovering this
3988 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3989 [Yuval Yarom and Naomi Benger]
3991 *) TLS pad extension: draft-agl-tls-padding-03
3993 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3994 TLS client Hello record length value would otherwise be > 255 and
3995 less that 512 pad with a dummy extension containing zeroes so it
3996 is at least 512 bytes long.
3998 [Adam Langley, Steve Henson]
4000 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
4002 *) Fix for TLS record tampering bug. A carefully crafted invalid
4003 handshake could crash OpenSSL with a NULL pointer exception.
4004 Thanks to Anton Johansson for reporting this issues.
4007 *) Keep original DTLS digest and encryption contexts in retransmission
4008 structures so we can use the previous session parameters if they need
4009 to be resent. (CVE-2013-6450)
4012 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
4013 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
4014 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
4015 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
4016 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
4017 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
4018 [Rob Stradling, Adam Langley]
4020 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
4022 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
4023 supporting platforms or when small records were transferred.
4024 [Andy Polyakov, Steve Henson]
4026 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
4028 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
4030 This addresses the flaw in CBC record processing discovered by
4031 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
4032 at: http://www.isg.rhul.ac.uk/tls/
4034 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4035 Security Group at Royal Holloway, University of London
4036 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
4037 Emilia Käsper for the initial patch.
4039 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
4041 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
4042 ciphersuites which can be exploited in a denial of service attack.
4043 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
4044 and detecting this bug and to Wolfgang Ettlinger
4045 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
4049 *) Return an error when checking OCSP signatures when key is NULL.
4050 This fixes a DoS attack. (CVE-2013-0166)
4053 *) Make openssl verify return errors.
4054 [Chris Palmer <palmer@google.com> and Ben Laurie]
4056 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
4057 the right response is stapled. Also change SSL_get_certificate()
4058 so it returns the certificate actually sent.
4059 See http://rt.openssl.org/Ticket/Display.html?id=2836.
4060 [Rob Stradling <rob.stradling@comodo.com>]
4062 *) Fix possible deadlock when decoding public keys.
4065 *) Don't use TLS 1.0 record version number in initial client hello
4069 Changes between 1.0.1b and 1.0.1c [10 May 2012]
4071 *) Sanity check record length before skipping explicit IV in TLS
4072 1.2, 1.1 and DTLS to fix DoS attack.
4074 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
4075 fuzzing as a service testing platform.
4079 *) Initialise tkeylen properly when encrypting CMS messages.
4080 Thanks to Solar Designer of Openwall for reporting this issue.
4083 *) In FIPS mode don't try to use composite ciphers as they are not
4087 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
4089 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
4090 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
4091 mean any application compiled against OpenSSL 1.0.0 headers setting
4092 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
4093 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
4094 0x10000000L Any application which was previously compiled against
4095 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
4096 will need to be recompiled as a result. Letting be results in
4097 inability to disable specifically TLS 1.1 and in client context,
4098 in unlike event, limit maximum offered version to TLS 1.0 [see below].
4101 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
4102 disable just protocol X, but all protocols above X *if* there are
4103 protocols *below* X still enabled. In more practical terms it means
4104 that if application wants to disable TLS1.0 in favor of TLS1.1 and
4105 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
4106 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
4110 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
4112 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
4113 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
4114 in CRYPTO_realloc_clean.
4116 Thanks to Tavis Ormandy, Google Security Team, for discovering this
4117 issue and to Adam Langley <agl@chromium.org> for fixing it.
4119 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
4121 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
4124 *) Workarounds for some broken servers that "hang" if a client hello
4125 record length exceeds 255 bytes.
4127 1. Do not use record version number > TLS 1.0 in initial client
4128 hello: some (but not all) hanging servers will now work.
4129 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
4130 the number of ciphers sent in the client hello. This should be
4131 set to an even number, such as 50, for example by passing:
4132 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
4133 Most broken servers should now work.
4134 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
4135 TLS 1.2 client support entirely.
4138 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
4141 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
4143 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
4144 STRING form instead of a DigestInfo.
4147 *) The format used for MDC2 RSA signatures is inconsistent between EVP
4148 and the RSA_sign/RSA_verify functions. This was made more apparent when
4149 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
4150 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
4151 the correct format in RSA_verify so both forms transparently work.
4154 *) Some servers which support TLS 1.0 can choke if we initially indicate
4155 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
4156 encrypted premaster secret. As a workaround use the maximum permitted
4157 client version in client hello, this should keep such servers happy
4158 and still work with previous versions of OpenSSL.
4161 *) Add support for TLS/DTLS heartbeats.
4162 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4164 *) Add support for SCTP.
4165 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4167 *) Improved PRNG seeding for VOS.
4168 [Paul Green <Paul.Green@stratus.com>]
4170 *) Extensive assembler packs updates, most notably:
4172 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
4173 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
4174 - x86_64: bit-sliced AES implementation;
4175 - ARM: NEON support, contemporary platforms optimizations;
4176 - s390x: z196 support;
4177 - *: GHASH and GF(2^m) multiplication implementations;
4181 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
4182 (removal of unnecessary code)
4183 [Peter Sylvester <peter.sylvester@edelweb.fr>]
4185 *) Add TLS key material exporter from RFC 5705.
4188 *) Add DTLS-SRTP negotiation from RFC 5764.
4191 *) Add Next Protocol Negotiation,
4192 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
4193 disabled with a no-npn flag to config or Configure. Code donated
4195 [Adam Langley <agl@google.com> and Ben Laurie]
4197 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
4198 NIST-P256, NIST-P521, with constant-time single point multiplication on
4199 typical inputs. Compiler support for the nonstandard type __uint128_t is
4200 required to use this (present in gcc 4.4 and later, for 64-bit builds).
4201 Code made available under Apache License version 2.0.
4203 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
4204 line to include this in your build of OpenSSL, and run "make depend" (or
4205 "make update"). This enables the following EC_METHODs:
4207 EC_GFp_nistp224_method()
4208 EC_GFp_nistp256_method()
4209 EC_GFp_nistp521_method()
4211 EC_GROUP_new_by_curve_name() will automatically use these (while
4212 EC_GROUP_new_curve_GFp() currently prefers the more flexible
4214 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
4216 *) Use type ossl_ssize_t instead of ssize_t which isn't available on
4217 all platforms. Move ssize_t definition from e_os.h to the public
4218 header file e_os2.h as it now appears in public header file cms.h
4221 *) New -sigopt option to the ca, req and x509 utilities. Additional
4222 signature parameters can be passed using this option and in
4226 *) Add RSA PSS signing function. This will generate and set the
4227 appropriate AlgorithmIdentifiers for PSS based on those in the
4228 corresponding EVP_MD_CTX structure. No application support yet.
4231 *) Support for companion algorithm specific ASN1 signing routines.
4232 New function ASN1_item_sign_ctx() signs a pre-initialised
4233 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
4234 the appropriate parameters.
4237 *) Add new algorithm specific ASN1 verification initialisation function
4238 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
4239 handling will be the same no matter what EVP_PKEY_METHOD is used.
4240 Add a PSS handler to support verification of PSS signatures: checked
4241 against a number of sample certificates.
4244 *) Add signature printing for PSS. Add PSS OIDs.
4245 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
4247 *) Add algorithm specific signature printing. An individual ASN1 method
4248 can now print out signatures instead of the standard hex dump.
4250 More complex signatures (e.g. PSS) can print out more meaningful
4251 information. Include DSA version that prints out the signature
4255 *) Password based recipient info support for CMS library: implementing
4259 *) Split password based encryption into PBES2 and PBKDF2 functions. This
4260 neatly separates the code into cipher and PBE sections and is required
4261 for some algorithms that split PBES2 into separate pieces (such as
4262 password based CMS).
4265 *) Session-handling fixes:
4266 - Fix handling of connections that are resuming with a session ID,
4267 but also support Session Tickets.
4268 - Fix a bug that suppressed issuing of a new ticket if the client
4269 presented a ticket with an expired session.
4270 - Try to set the ticket lifetime hint to something reasonable.
4271 - Make tickets shorter by excluding irrelevant information.
4272 - On the client side, don't ignore renewed tickets.
4273 [Adam Langley, Bodo Moeller (Google)]
4275 *) Fix PSK session representation.
4278 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
4280 This work was sponsored by Intel.
4283 *) Add GCM support to TLS library. Some custom code is needed to split
4284 the IV between the fixed (from PRF) and explicit (from TLS record)
4285 portions. This adds all GCM ciphersuites supported by RFC5288 and
4286 RFC5289. Generalise some AES* cipherstrings to include GCM and
4287 add a special AESGCM string for GCM only.
4290 *) Expand range of ctrls for AES GCM. Permit setting invocation
4291 field on decrypt and retrieval of invocation field only on encrypt.
4294 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
4295 As required by RFC5289 these ciphersuites cannot be used if for
4296 versions of TLS earlier than 1.2.
4299 *) For FIPS capable OpenSSL interpret a NULL default public key method
4300 as unset and return the appropriate default but do *not* set the default.
4301 This means we can return the appropriate method in applications that
4302 switch between FIPS and non-FIPS modes.
4305 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
4306 ENGINE is used then we cannot handle that in the FIPS module so we
4307 keep original code iff non-FIPS operations are allowed.
4310 *) Add -attime option to openssl utilities.
4311 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
4313 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
4316 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
4317 FIPS EC methods unconditionally for now.
4320 *) New build option no-ec2m to disable characteristic 2 code.
4323 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
4324 all cases can be covered as some introduce binary incompatibilities.
4327 *) Redirect RSA operations to FIPS module including keygen,
4328 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
4331 *) Add similar low level API blocking to ciphers.
4334 *) Low level digest APIs are not approved in FIPS mode: any attempt
4335 to use these will cause a fatal error. Applications that *really* want
4336 to use them can use the private_* version instead.
4339 *) Redirect cipher operations to FIPS module for FIPS builds.
4342 *) Redirect digest operations to FIPS module for FIPS builds.
4345 *) Update build system to add "fips" flag which will link in fipscanister.o
4346 for static and shared library builds embedding a signature if needed.
4349 *) Output TLS supported curves in preference order instead of numerical
4350 order. This is currently hardcoded for the highest order curves first.
4351 This should be configurable so applications can judge speed vs strength.
4354 *) Add TLS v1.2 server support for client authentication.
4357 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
4361 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
4362 FIPS modules versions.
4365 *) Add TLS v1.2 client side support for client authentication. Keep cache
4366 of handshake records longer as we don't know the hash algorithm to use
4367 until after the certificate request message is received.
4370 *) Initial TLS v1.2 client support. Add a default signature algorithms
4371 extension including all the algorithms we support. Parse new signature
4372 format in client key exchange. Relax some ECC signing restrictions for
4373 TLS v1.2 as indicated in RFC5246.
4376 *) Add server support for TLS v1.2 signature algorithms extension. Switch
4377 to new signature format when needed using client digest preference.
4378 All server ciphersuites should now work correctly in TLS v1.2. No client
4379 support yet and no support for client certificates.
4382 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
4383 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
4384 ciphersuites. At present only RSA key exchange ciphersuites work with
4385 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
4386 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
4387 and version checking.
4390 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
4391 with this defined it will not be affected by any changes to ssl internal
4392 structures. Add several utility functions to allow openssl application
4393 to work with OPENSSL_NO_SSL_INTERN defined.
4396 *) A long standing patch to add support for SRP from EdelWeb (Peter
4397 Sylvester and Christophe Renou) was integrated.
4398 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
4399 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
4402 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
4405 *) Permit abbreviated handshakes when renegotiating using the function
4406 SSL_renegotiate_abbreviated().
4407 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4409 *) Add call to ENGINE_register_all_complete() to
4410 ENGINE_load_builtin_engines(), so some implementations get used
4411 automatically instead of needing explicit application support.
4414 *) Add support for TLS key exporter as described in RFC5705.
4415 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
4417 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
4418 a few changes are required:
4420 Add SSL_OP_NO_TLSv1_1 flag.
4421 Add TLSv1_1 methods.
4422 Update version checking logic to handle version 1.1.
4423 Add explicit IV handling (ported from DTLS code).
4424 Add command line options to s_client/s_server.
4427 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
4429 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
4430 in CMS and PKCS7 code. When RSA decryption fails use a random key for
4431 content decryption and always return the same error. Note: this attack
4432 needs on average 2^20 messages so it only affects automated senders. The
4433 old behaviour can be re-enabled in the CMS code by setting the
4434 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
4435 an MMA defence is not necessary.
4436 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
4437 this issue. (CVE-2012-0884)
4440 *) Fix CVE-2011-4619: make sure we really are receiving a
4441 client hello before rejecting multiple SGC restarts. Thanks to
4442 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
4445 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
4447 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
4448 Thanks to Antonio Martin, Enterprise Secure Access Research and
4449 Development, Cisco Systems, Inc. for discovering this bug and
4450 preparing a fix. (CVE-2012-0050)
4453 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
4455 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
4456 of the Vaudenay padding oracle attack on CBC mode encryption
4457 which enables an efficient plaintext recovery attack against
4458 the OpenSSL implementation of DTLS. Their attack exploits timing
4459 differences arising during decryption processing. A research
4460 paper describing this attack can be found at:
4461 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
4462 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4463 Security Group at Royal Holloway, University of London
4464 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
4465 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
4466 for preparing the fix. (CVE-2011-4108)
4467 [Robin Seggelmann, Michael Tuexen]
4469 *) Clear bytes used for block padding of SSL 3.0 records.
4471 [Adam Langley (Google)]
4473 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
4474 Kadianakis <desnacked@gmail.com> for discovering this issue and
4475 Adam Langley for preparing the fix. (CVE-2011-4619)
4476 [Adam Langley (Google)]
4478 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
4479 [Andrey Kulikov <amdeich@gmail.com>]
4481 *) Prevent malformed RFC3779 data triggering an assertion failure.
4482 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
4483 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
4484 [Rob Austein <sra@hactrn.net>]
4486 *) Improved PRNG seeding for VOS.
4487 [Paul Green <Paul.Green@stratus.com>]
4489 *) Fix ssl_ciph.c set-up race.
4490 [Adam Langley (Google)]
4492 *) Fix spurious failures in ecdsatest.c.
4493 [Emilia Käsper (Google)]
4495 *) Fix the BIO_f_buffer() implementation (which was mixing different
4496 interpretations of the '..._len' fields).
4497 [Adam Langley (Google)]
4499 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4500 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4501 threads won't reuse the same blinding coefficients.
4503 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4504 lock to call BN_BLINDING_invert_ex, and avoids one use of
4505 BN_BLINDING_update for each BN_BLINDING structure (previously,
4506 the last update always remained unused).
4507 [Emilia Käsper (Google)]
4509 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4510 [Bob Buckholz (Google)]
4512 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4514 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4515 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4516 [Kaspar Brand <ossl@velox.ch>]
4518 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4519 for multi-threaded use of ECDH. (CVE-2011-3210)
4520 [Adam Langley (Google)]
4522 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4525 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4526 signature public key algorithm by using OID xref utilities instead.
4527 Before this you could only use some ECC ciphersuites with SHA1 only.
4530 *) Add protection against ECDSA timing attacks as mentioned in the paper
4531 by Billy Bob Brumley and Nicola Tuveri, see:
4533 http://eprint.iacr.org/2011/232.pdf
4535 [Billy Bob Brumley and Nicola Tuveri]
4537 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4539 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4540 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4542 *) Fix bug in string printing code: if *any* escaping is enabled we must
4543 escape the escape character (backslash) or the resulting string is
4547 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4549 *) Disable code workaround for ancient and obsolete Netscape browsers
4550 and servers: an attacker can use it in a ciphersuite downgrade attack.
4551 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4554 *) Fixed J-PAKE implementation error, originally discovered by
4555 Sebastien Martini, further info and confirmation from Stefan
4556 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4559 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4561 *) Fix extension code to avoid race conditions which can result in a buffer
4562 overrun vulnerability: resumed sessions must not be modified as they can
4563 be shared by multiple threads. CVE-2010-3864
4566 *) Fix WIN32 build system to correctly link an ENGINE directory into
4570 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4572 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4574 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4576 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4578 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4579 context. The operation can be customised via the ctrl mechanism in
4580 case ENGINEs want to include additional functionality.
4583 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4586 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4587 output hashes compatible with older versions of OpenSSL.
4588 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4590 *) Fix compression algorithm handling: if resuming a session use the
4591 compression algorithm of the resumed session instead of determining
4592 it from client hello again. Don't allow server to change algorithm.
4595 *) Add load_crls() function to apps tidying load_certs() too. Add option
4596 to verify utility to allow additional CRLs to be included.
4599 *) Update OCSP request code to permit adding custom headers to the request:
4600 some responders need this.
4603 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4605 [Julia Lawall <julia@diku.dk>]
4607 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4608 needlessly dereferenced structures, used obsolete functions and
4609 didn't handle all updated verify codes correctly.
4612 *) Disable MD2 in the default configuration.
4615 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4616 indicate the initial BIO being pushed or popped. This makes it possible
4617 to determine whether the BIO is the one explicitly called or as a result
4618 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4619 it handles reference counts correctly and doesn't zero out the I/O bio
4620 when it is not being explicitly popped. WARNING: applications which
4621 included workarounds for the old buggy behaviour will need to be modified
4622 or they could free up already freed BIOs.
4625 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4626 renaming to all platforms (within the 0.9.8 branch, this was
4627 done conditionally on Netware platforms to avoid a name clash).
4628 [Guenter <lists@gknw.net>]
4630 *) Add ECDHE and PSK support to DTLS.
4631 [Michael Tuexen <tuexen@fh-muenster.de>]
4633 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4637 *) Add "missing" function EVP_MD_flags() (without this the only way to
4638 retrieve a digest flags is by accessing the structure directly. Update
4639 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4640 or cipher is registered as in the "from" argument. Print out all
4641 registered digests in the dgst usage message instead of manually
4642 attempting to work them out.
4645 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4646 this allows the use of compression and extensions. Change default cipher
4647 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4648 by default unless an application cipher string requests it.
4651 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4652 key ids to find matching certificates and keys but some PKCS#12 files
4653 don't follow the (somewhat unwritten) rules and this strategy fails.
4654 Now just gather all certificates together and the first private key
4655 then look for the first certificate that matches the key.
4658 *) Support use of registered digest and cipher names for dgst and cipher
4659 commands instead of having to add each one as a special case. So now
4666 openssl dgst -sha256 foo
4668 and this works for ENGINE based algorithms too.
4672 *) Update Gost ENGINE to support parameter files.
4673 [Victor B. Wagner <vitus@cryptocom.ru>]
4675 *) Support GeneralizedTime in ca utility.
4676 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4678 *) Enhance the hash format used for certificate directory links. The new
4679 form uses the canonical encoding (meaning equivalent names will work
4680 even if they aren't identical) and uses SHA1 instead of MD5. This form
4681 is incompatible with the older format and as a result c_rehash should
4682 be used to rebuild symbolic links.
4685 *) Make PKCS#8 the default write format for private keys, replacing the
4686 traditional format. This form is standardised, more secure and doesn't
4687 include an implicit MD5 dependency.
4690 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4691 committed to OpenSSL should pass this lot as a minimum.
4694 *) Add session ticket override functionality for use by EAP-FAST.
4695 [Jouni Malinen <j@w1.fi>]
4697 *) Modify HMAC functions to return a value. Since these can be implemented
4698 in an ENGINE errors can occur.
4701 *) Type-checked OBJ_bsearch_ex.
4704 *) Type-checked OBJ_bsearch. Also some constification necessitated
4705 by type-checking. Still to come: TXT_DB, bsearch(?),
4706 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4710 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4711 seconds to a tm structure directly, instead of going through OS
4712 specific date routines. This avoids any issues with OS routines such
4713 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4714 and X509_time_adj_ex() to cover the extended range. The existing
4715 X509_time_adj() is still usable and will no longer have any date issues.
4718 *) Delta CRL support. New use deltas option which will attempt to locate
4719 and search any appropriate delta CRLs available.
4721 This work was sponsored by Google.
4724 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4725 code and add additional score elements. Validate alternate CRL paths
4726 as part of the CRL checking and indicate a new error "CRL path validation
4727 error" in this case. Applications wanting additional details can use
4728 the verify callback and check the new "parent" field. If this is not
4729 NULL CRL path validation is taking place. Existing applications won't
4730 see this because it requires extended CRL support which is off by
4733 This work was sponsored by Google.
4736 *) Support for freshest CRL extension.
4738 This work was sponsored by Google.
4741 *) Initial indirect CRL support. Currently only supported in the CRLs
4742 passed directly and not via lookup. Process certificate issuer
4743 CRL entry extension and lookup CRL entries by bother issuer name
4744 and serial number. Check and process CRL issuer entry in IDP extension.
4746 This work was sponsored by Google.
4749 *) Add support for distinct certificate and CRL paths. The CRL issuer
4750 certificate is validated separately in this case. Only enabled if
4751 an extended CRL support flag is set: this flag will enable additional
4752 CRL functionality in future.
4754 This work was sponsored by Google.
4757 *) Add support for policy mappings extension.
4759 This work was sponsored by Google.
4762 *) Fixes to pathlength constraint, self issued certificate handling,
4763 policy processing to align with RFC3280 and PKITS tests.
4765 This work was sponsored by Google.
4768 *) Support for name constraints certificate extension. DN, email, DNS
4769 and URI types are currently supported.
4771 This work was sponsored by Google.
4774 *) To cater for systems that provide a pointer-based thread ID rather
4775 than numeric, deprecate the current numeric thread ID mechanism and
4776 replace it with a structure and associated callback type. This
4777 mechanism allows a numeric "hash" to be extracted from a thread ID in
4778 either case, and on platforms where pointers are larger than 'long',
4779 mixing is done to help ensure the numeric 'hash' is usable even if it
4780 can't be guaranteed unique. The default mechanism is to use "&errno"
4781 as a pointer-based thread ID to distinguish between threads.
4783 Applications that want to provide their own thread IDs should now use
4784 CRYPTO_THREADID_set_callback() to register a callback that will call
4785 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4787 Note that ERR_remove_state() is now deprecated, because it is tied
4788 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4789 to free the current thread's error state should be replaced by
4790 ERR_remove_thread_state(NULL).
4792 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4793 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4794 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4795 application was previously providing a numeric thread callback that
4796 was inappropriate for distinguishing threads, then uniqueness might
4797 have been obtained with &errno that happened immediately in the
4798 intermediate development versions of OpenSSL; this is no longer the
4799 case, the numeric thread callback will now override the automatic use
4801 [Geoff Thorpe, with help from Bodo Moeller]
4803 *) Initial support for different CRL issuing certificates. This covers a
4804 simple case where the self issued certificates in the chain exist and
4805 the real CRL issuer is higher in the existing chain.
4807 This work was sponsored by Google.
4810 *) Removed effectively defunct crypto/store from the build.
4813 *) Revamp of STACK to provide stronger type-checking. Still to come:
4814 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4815 ASN1_STRING, CONF_VALUE.
4818 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4819 RAM on SSL connections. This option can save about 34k per idle SSL.
4822 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4823 STACK, TXT_DB, bsearch, qsort.
4826 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4827 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4828 support for data, signedData, compressedData, digestedData and
4829 encryptedData, envelopedData types included. Scripts to check against
4830 RFC4134 examples draft and interop and consistency checks of many
4831 content types and variants.
4834 *) Add options to enc utility to support use of zlib compression BIO.
4837 *) Extend mk1mf to support importing of options and assembly language
4838 files from Configure script, currently only included in VC-WIN32.
4839 The assembly language rules can now optionally generate the source
4840 files from the associated perl scripts.
4843 *) Implement remaining functionality needed to support GOST ciphersuites.
4844 Interop testing has been performed using CryptoPro implementations.
4845 [Victor B. Wagner <vitus@cryptocom.ru>]
4847 *) s390x assembler pack.
4850 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4854 *) Implement Opaque PRF Input TLS extension as specified in
4855 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4856 official specification yet and no extension type assignment by
4857 IANA exists, this extension (for now) will have to be explicitly
4858 enabled when building OpenSSL by providing the extension number
4859 to use. For example, specify an option
4861 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4863 to the "config" or "Configure" script to enable the extension,
4864 assuming extension number 0x9527 (which is a completely arbitrary
4865 and unofficial assignment based on the MD5 hash of the Internet
4866 Draft). Note that by doing so, you potentially lose
4867 interoperability with other TLS implementations since these might
4868 be using the same extension number for other purposes.
4870 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4871 opaque PRF input value to use in the handshake. This will create
4872 an internal copy of the length-'len' string at 'src', and will
4873 return non-zero for success.
4875 To get more control and flexibility, provide a callback function
4878 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4879 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4883 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4886 Callback function 'cb' will be called in handshakes, and is
4887 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4888 Argument 'arg' is for application purposes (the value as given to
4889 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4890 be provided to the callback function). The callback function
4891 has to return non-zero to report success: usually 1 to use opaque
4892 PRF input just if possible, or 2 to enforce use of the opaque PRF
4893 input. In the latter case, the library will abort the handshake
4894 if opaque PRF input is not successfully negotiated.
4896 Arguments 'peerinput' and 'len' given to the callback function
4897 will always be NULL and 0 in the case of a client. A server will
4898 see the client's opaque PRF input through these variables if
4899 available (NULL and 0 otherwise). Note that if the server
4900 provides an opaque PRF input, the length must be the same as the
4901 length of the client's opaque PRF input.
4903 Note that the callback function will only be called when creating
4904 a new session (session resumption can resume whatever was
4905 previously negotiated), and will not be called in SSL 2.0
4906 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4907 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4908 for applications that need to enforce opaque PRF input.
4912 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4915 [Victor B. Wagner <vitus@cryptocom.ru>]
4917 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4918 RFC4507bis. The encrypted ticket format is an encrypted encoded
4919 SSL_SESSION structure, that way new session features are automatically
4922 If a client application caches session in an SSL_SESSION structure
4923 support is transparent because tickets are now stored in the encoded
4926 The SSL_CTX structure automatically generates keys for ticket
4927 protection in servers so again support should be possible
4928 with no application modification.
4930 If a client or server wishes to disable RFC4507 support then the option
4931 SSL_OP_NO_TICKET can be set.
4933 Add a TLS extension debugging callback to allow the contents of any client
4934 or server extensions to be examined.
4936 This work was sponsored by Google.
4939 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4940 OpenSSL should now compile cleanly on gcc 4.2
4941 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4943 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4944 support including streaming MAC support: this is required for GOST
4945 ciphersuite support.
4946 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4948 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4949 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4950 to output in BER and PEM format.
4953 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4954 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4955 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4956 ENGINE support for HMAC keys which are unextractable. New -mac and
4957 -macopt options to dgst utility.
4960 *) New option -sigopt to dgst utility. Update dgst to use
4961 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4962 alternative signing parameters such as X9.31 or PSS in the dgst
4966 *) Change ssl_cipher_apply_rule(), the internal function that does
4967 the work each time a ciphersuite string requests enabling
4968 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4969 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4970 the order of disabled ciphersuites such that those ciphersuites
4971 that most recently went from enabled to disabled not only stay
4972 in order with respect to each other, but also have higher priority
4973 than other disabled ciphersuites the next time ciphersuites are
4976 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4977 the same ciphersuites as with "HIGH" alone, but in a specific
4978 order where the PSK ciphersuites come first (since they are the
4979 most recently disabled ciphersuites when "HIGH" is parsed).
4981 Also, change ssl_create_cipher_list() (using this new
4982 functionality) such that between otherwise identical
4983 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4987 *) Change ssl_create_cipher_list() so that it automatically
4988 arranges the ciphersuites in reasonable order before starting
4989 to process the rule string. Thus, the definition for "DEFAULT"
4990 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4991 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4992 This makes it much easier to arrive at a reasonable default order
4993 in applications for which anonymous ciphers are OK (meaning
4994 that you can't actually use DEFAULT).
4995 [Bodo Moeller; suggested by Victor Duchovni]
4997 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4998 processing) into multiple integers instead of setting
4999 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
5000 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
5001 (These masks as well as the individual bit definitions are hidden
5002 away into the non-exported interface ssl/ssl_locl.h, so this
5003 change to the definition of the SSL_CIPHER structure shouldn't
5004 affect applications.) This give us more bits for each of these
5005 categories, so there is no longer a need to coagulate AES128 and
5006 AES256 into a single algorithm bit, and to coagulate Camellia128
5007 and Camellia256 into a single algorithm bit, which has led to all
5010 Thus, among other things, the kludge introduced in 0.9.7m and
5011 0.9.8e for masking out AES256 independently of AES128 or masking
5012 out Camellia256 independently of AES256 is not needed here in 0.9.9.
5014 With the change, we also introduce new ciphersuite aliases that
5015 so far were missing: "AES128", "AES256", "CAMELLIA128", and
5019 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
5020 Use the leftmost N bytes of the signature input if the input is
5021 larger than the prime q (with N being the size in bytes of q).
5024 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
5025 it yet and it is largely untested.
5028 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
5031 *) Initial incomplete changes to avoid need for function casts in OpenSSL
5032 some compilers (gcc 4.2 and later) reject their use. Safestack is
5033 reimplemented. Update ASN1 to avoid use of legacy functions.
5036 *) Win32/64 targets are linked with Winsock2.
5039 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
5040 to external functions. This can be used to increase CRL handling
5041 efficiency especially when CRLs are very large by (for example) storing
5042 the CRL revoked certificates in a database.
5045 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
5046 new CRLs added to a directory can be used. New command line option
5047 -verify_return_error to s_client and s_server. This causes real errors
5048 to be returned by the verify callback instead of carrying on no matter
5049 what. This reflects the way a "real world" verify callback would behave.
5052 *) GOST engine, supporting several GOST algorithms and public key formats.
5053 Kindly donated by Cryptocom.
5056 *) Partial support for Issuing Distribution Point CRL extension. CRLs
5057 partitioned by DP are handled but no indirect CRL or reason partitioning
5058 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
5059 selected via a scoring technique which handles IDP and AKID in CRLs.
5062 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
5063 will ultimately be used for all verify operations: this will remove the
5064 X509_STORE dependency on certificate verification and allow alternative
5065 lookup methods. X509_STORE based implementations of these two callbacks.
5068 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
5069 Modify get_crl() to find a valid (unexpired) CRL if possible.
5072 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
5073 this would be called X509_CRL_cmp() but that name is already used by
5074 a function that just compares CRL issuer names. Cache several CRL
5075 extensions in X509_CRL structure and cache CRLDP in X509.
5078 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
5079 this maps equivalent X509_NAME structures into a consistent structure.
5080 Name comparison can then be performed rapidly using memcmp().
5083 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
5087 *) Allow digests to supply their own micalg string for S/MIME type using
5088 the ctrl EVP_MD_CTRL_MICALG.
5091 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
5092 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
5093 ctrl. It can then customise the structure before and/or after signing
5097 *) New function OBJ_add_sigid() to allow application defined signature OIDs
5098 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
5099 to free up any added signature OIDs.
5102 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
5103 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
5104 digest and cipher tables. New options added to openssl utility:
5105 list-message-digest-algorithms and list-cipher-algorithms.
5108 *) Change the array representation of binary polynomials: the list
5109 of degrees of non-zero coefficients is now terminated with -1.
5110 Previously it was terminated with 0, which was also part of the
5111 value; thus, the array representation was not applicable to
5112 polynomials where t^0 has coefficient zero. This change makes
5113 the array representation useful in a more general context.
5116 *) Various modifications and fixes to SSL/TLS cipher string
5117 handling. For ECC, the code now distinguishes between fixed ECDH
5118 with RSA certificates on the one hand and with ECDSA certificates
5119 on the other hand, since these are separate ciphersuites. The
5120 unused code for Fortezza ciphersuites has been removed.
5122 For consistency with EDH, ephemeral ECDH is now called "EECDH"
5123 (not "ECDHE"). For consistency with the code for DH
5124 certificates, use of ECDH certificates is now considered ECDH
5125 authentication, not RSA or ECDSA authentication (the latter is
5126 merely the CA's signing algorithm and not actively used in the
5129 The temporary ciphersuite alias "ECCdraft" is no longer
5130 available, and ECC ciphersuites are no longer excluded from "ALL"
5131 and "DEFAULT". The following aliases now exist for RFC 4492
5132 ciphersuites, most of these by analogy with the DH case:
5134 kECDHr - ECDH cert, signed with RSA
5135 kECDHe - ECDH cert, signed with ECDSA
5136 kECDH - ECDH cert (signed with either RSA or ECDSA)
5137 kEECDH - ephemeral ECDH
5138 ECDH - ECDH cert or ephemeral ECDH
5144 AECDH - anonymous ECDH
5145 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
5149 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
5150 Use correct micalg parameters depending on digest(s) in signed message.
5153 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
5154 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
5157 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
5158 an engine to register a method. Add ENGINE lookups for methods and
5159 functional reference processing.
5162 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
5163 EVP_{Sign,Verify}* which allow an application to customise the signature
5167 *) New -resign option to smime utility. This adds one or more signers
5168 to an existing PKCS#7 signedData structure. Also -md option to use an
5169 alternative message digest algorithm for signing.
5172 *) Tidy up PKCS#7 routines and add new functions to make it easier to
5173 create PKCS7 structures containing multiple signers. Update smime
5174 application to support multiple signers.
5177 *) New -macalg option to pkcs12 utility to allow setting of an alternative
5181 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
5182 Reorganize PBE internals to lookup from a static table using NIDs,
5183 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
5184 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
5185 PRF which will be automatically used with PBES2.
5188 *) Replace the algorithm specific calls to generate keys in "req" with the
5192 *) Update PKCS#7 enveloped data routines to use new API. This is now
5193 supported by any public key method supporting the encrypt operation. A
5194 ctrl is added to allow the public key algorithm to examine or modify
5195 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
5199 *) Add a ctrl to asn1 method to allow a public key algorithm to express
5200 a default digest type to use. In most cases this will be SHA1 but some
5201 algorithms (such as GOST) need to specify an alternative digest. The
5202 return value indicates how strong the preference is 1 means optional and
5203 2 is mandatory (that is it is the only supported type). Modify
5204 ASN1_item_sign() to accept a NULL digest argument to indicate it should
5205 use the default md. Update openssl utilities to use the default digest
5206 type for signing if it is not explicitly indicated.
5209 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
5210 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
5211 signing method from the key type. This effectively removes the link
5212 between digests and public key types.
5215 *) Add an OID cross reference table and utility functions. Its purpose is to
5216 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
5217 rsaEncryption. This will allow some of the algorithm specific hackery
5218 needed to use the correct OID to be removed.
5221 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
5222 structures for PKCS7_sign(). They are now set up by the relevant public
5226 *) Add provisional EC pkey method with support for ECDSA and ECDH.
5229 *) Add support for key derivation (agreement) in the API, DH method and
5233 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
5234 public and private key formats. As a side effect these add additional
5235 command line functionality not previously available: DSA signatures can be
5236 generated and verified using pkeyutl and DH key support and generation in
5241 [Oliver Tappe <zooey@hirschkaefer.de>]
5243 *) New make target "install_html_docs" installs HTML renditions of the
5245 [Oliver Tappe <zooey@hirschkaefer.de>]
5247 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
5248 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
5249 support key and parameter generation and add initial key generation
5250 functionality for RSA.
5253 *) Add functions for main EVP_PKEY_method operations. The undocumented
5254 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
5255 EVP_PKEY_{encrypt,decrypt}_old.
5258 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
5259 key API, doesn't do much yet.
5262 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
5263 public key algorithms. New option to openssl utility:
5264 "list-public-key-algorithms" to print out info.
5267 *) Implement the Supported Elliptic Curves Extension for
5268 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5271 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
5272 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
5275 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
5276 utilities such as rsa, dsa, dsaparam etc except they process any key
5280 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
5281 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
5282 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
5286 *) Initial support for pluggable public key ASN1.
5287 De-spaghettify the public key ASN1 handling. Move public and private
5288 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
5289 algorithm specific handling to a single module within the relevant
5290 algorithm directory. Add functions to allow (near) opaque processing
5291 of public and private key structures.
5294 *) Implement the Supported Point Formats Extension for
5295 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5298 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
5299 for the psk identity [hint] and the psk callback functions to the
5300 SSL_SESSION, SSL and SSL_CTX structure.
5303 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
5307 SSL_CTX_use_psk_identity_hint
5308 SSL_get_psk_identity_hint
5309 SSL_get_psk_identity
5310 SSL_use_psk_identity_hint
5312 [Mika Kousa and Pasi Eronen of Nokia Corporation]
5314 *) Add RFC 3161 compliant time stamp request creation, response generation
5315 and response verification functionality.
5316 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
5318 *) Add initial support for TLS extensions, specifically for the server_name
5319 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5320 have new members for a host name. The SSL data structure has an
5321 additional member SSL_CTX *initial_ctx so that new sessions can be
5322 stored in that context to allow for session resumption, even after the
5323 SSL has been switched to a new SSL_CTX in reaction to a client's
5324 server_name extension.
5326 New functions (subject to change):
5328 SSL_get_servername()
5329 SSL_get_servername_type()
5332 New CTRL codes and macros (subject to change):
5334 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5335 - SSL_CTX_set_tlsext_servername_callback()
5336 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5337 - SSL_CTX_set_tlsext_servername_arg()
5338 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5340 openssl s_client has a new '-servername ...' option.
5342 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5343 '-key2 ...', '-servername_fatal' (subject to change). This allows
5344 testing the HostName extension for a specific single host name ('-cert'
5345 and '-key' remain fallbacks for handshakes without HostName
5346 negotiation). If the unrecognized_name alert has to be sent, this by
5347 default is a warning; it becomes fatal with the '-servername_fatal'
5350 [Peter Sylvester, Remy Allais, Christophe Renou]
5352 *) Whirlpool hash implementation is added.
5355 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
5356 bn(64,32). Because of instruction set limitations it doesn't have
5357 any negative impact on performance. This was done mostly in order
5358 to make it possible to share assembler modules, such as bn_mul_mont
5359 implementations, between 32- and 64-bit builds without hassle.
5362 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
5363 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
5367 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
5368 dedicated Montgomery multiplication procedure, is introduced.
5369 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
5370 "64-bit" performance on certain 32-bit targets.
5373 *) New option SSL_OP_NO_COMP to disable use of compression selectively
5374 in SSL structures. New SSL ctrl to set maximum send fragment size.
5375 Save memory by setting the I/O buffer sizes dynamically instead of
5376 using the maximum available value.
5379 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
5380 in addition to the text details.
5383 *) Very, very preliminary EXPERIMENTAL support for printing of general
5384 ASN1 structures. This currently produces rather ugly output and doesn't
5385 handle several customised structures at all.
5388 *) Integrated support for PVK file format and some related formats such
5389 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
5390 these in the 'rsa' and 'dsa' utilities.
5393 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
5396 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
5397 place for the (very old) "NETSCAPE" format certificates which are now
5398 handled using new ASN1 code equivalents.
5401 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
5402 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
5403 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
5406 *) Modify CRL distribution points extension code to print out previously
5407 unsupported fields. Enhance extension setting code to allow setting of
5411 *) Add print and set support for Issuing Distribution Point CRL extension.
5414 *) Change 'Configure' script to enable Camellia by default.
5417 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
5419 *) When rejecting SSL/TLS records due to an incorrect version number, never
5420 update s->server with a new major version number. As of
5421 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
5422 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
5423 the previous behavior could result in a read attempt at NULL when
5424 receiving specific incorrect SSL/TLS records once record payload
5425 protection is active. (CVE-2010-0740)
5426 [Bodo Moeller, Adam Langley <agl@chromium.org>]
5428 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
5429 could be crashed if the relevant tables were not present (e.g. chrooted).
5430 [Tomas Hoger <thoger@redhat.com>]
5432 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
5434 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
5435 [Martin Olsson, Neel Mehta]
5437 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
5438 accommodate for stack sorting, always a write lock!).
5441 *) On some versions of WIN32 Heap32Next is very slow. This can cause
5442 excessive delays in the RAND_poll(): over a minute. As a workaround
5443 include a time check in the inner Heap32Next loop too.
5446 *) The code that handled flushing of data in SSL/TLS originally used the
5447 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
5448 the problem outlined in PR#1949. The fix suggested there however can
5449 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
5450 of Apache). So instead simplify the code to flush unconditionally.
5451 This should be fine since flushing with no data to flush is a no op.
5454 *) Handle TLS versions 2.0 and later properly and correctly use the
5455 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
5456 off ancient servers have a habit of sticking around for a while...
5459 *) Modify compression code so it frees up structures without using the
5460 ex_data callbacks. This works around a problem where some applications
5461 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
5462 restarting) then use compression (e.g. SSL with compression) later.
5463 This results in significant per-connection memory leaks and
5464 has caused some security issues including CVE-2008-1678 and
5468 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
5469 change when encrypting or decrypting.
5472 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
5473 connect and renegotiate with servers which do not support RI.
5474 Until RI is more widely deployed this option is enabled by default.
5477 *) Add "missing" ssl ctrls to clear options and mode.
5480 *) If client attempts to renegotiate and doesn't support RI respond with
5481 a no_renegotiation alert as required by RFC5746. Some renegotiating
5482 TLS clients will continue a connection gracefully when they receive
5483 the alert. Unfortunately OpenSSL mishandled this alert and would hang
5484 waiting for a server hello which it will never receive. Now we treat a
5485 received no_renegotiation alert as a fatal error. This is because
5486 applications requesting a renegotiation might well expect it to succeed
5487 and would have no code in place to handle the server denying it so the
5488 only safe thing to do is to terminate the connection.
5491 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
5492 peer supports secure renegotiation and 0 otherwise. Print out peer
5493 renegotiation support in s_client/s_server.
5496 *) Replace the highly broken and deprecated SPKAC certification method with
5497 the updated NID creation version. This should correctly handle UTF8.
5500 *) Implement RFC5746. Re-enable renegotiation but require the extension
5501 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
5502 turns out to be a bad idea. It has been replaced by
5503 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5504 SSL_CTX_set_options(). This is really not recommended unless you
5505 know what you are doing.
5506 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5508 *) Fixes to stateless session resumption handling. Use initial_ctx when
5509 issuing and attempting to decrypt tickets in case it has changed during
5510 servername handling. Use a non-zero length session ID when attempting
5511 stateless session resumption: this makes it possible to determine if
5512 a resumption has occurred immediately after receiving server hello
5513 (several places in OpenSSL subtly assume this) instead of later in
5517 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5518 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5519 fixes for a few places where the return code is not checked
5521 [Julia Lawall <julia@diku.dk>]
5523 *) Add --strict-warnings option to Configure script to include devteam
5524 warnings in other configurations.
5527 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5528 makes it possible to install openssl libraries in locations which
5529 have names other than "lib", for example "/usr/lib64" which some
5531 [Steve Henson, based on patch from Jeremy Utley]
5533 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5534 X690 8.9.12 and can produce some misleading textual output of OIDs.
5535 [Steve Henson, reported by Dan Kaminsky]
5537 *) Delete MD2 from algorithm tables. This follows the recommendation in
5538 several standards that it is not used in new applications due to
5539 several cryptographic weaknesses. For binary compatibility reasons
5540 the MD2 API is still compiled in by default.
5543 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5547 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5548 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5550 [Guenter <lists@gknw.net>]
5552 *) Fix the server certificate chain building code to use X509_verify_cert(),
5553 it used to have an ad-hoc builder which was unable to cope with anything
5554 other than a simple chain.
5555 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5557 *) Don't check self signed certificate signatures in X509_verify_cert()
5558 by default (a flag can override this): it just wastes time without
5559 adding any security. As a useful side effect self signed root CAs
5560 with non-FIPS digests are now usable in FIPS mode.
5563 *) In dtls1_process_out_of_seq_message() the check if the current message
5564 is already buffered was missing. For every new message was memory
5565 allocated, allowing an attacker to perform an denial of service attack
5566 with sending out of seq handshake messages until there is no memory
5567 left. Additionally every future message was buffered, even if the
5568 sequence number made no sense and would be part of another handshake.
5569 So only messages with sequence numbers less than 10 in advance will be
5570 buffered. (CVE-2009-1378)
5571 [Robin Seggelmann, discovered by Daniel Mentz]
5573 *) Records are buffered if they arrive with a future epoch to be
5574 processed after finishing the corresponding handshake. There is
5575 currently no limitation to this buffer allowing an attacker to perform
5576 a DOS attack with sending records with future epochs until there is no
5577 memory left. This patch adds the pqueue_size() function to determine
5578 the size of a buffer and limits the record buffer to 100 entries.
5580 [Robin Seggelmann, discovered by Daniel Mentz]
5582 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5583 parent structure is freed. (CVE-2009-1379)
5586 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5587 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5590 [Ilya O. <vrghost@gmail.com>]
5592 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5594 *) Disable renegotiation completely - this fixes a severe security
5595 problem (CVE-2009-3555) at the cost of breaking all
5596 renegotiation. Renegotiation can be re-enabled by setting
5597 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5598 run-time. This is really not recommended unless you know what
5602 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5604 *) Don't set val to NULL when freeing up structures, it is freed up by
5605 underlying code. If sizeof(void *) > sizeof(long) this can result in
5606 zeroing past the valid field. (CVE-2009-0789)
5607 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5609 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5610 checked correctly. This would allow some invalid signed attributes to
5611 appear to verify correctly. (CVE-2009-0591)
5612 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5614 *) Reject UniversalString and BMPString types with invalid lengths. This
5615 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5616 a legal length. (CVE-2009-0590)
5619 *) Set S/MIME signing as the default purpose rather than setting it
5620 unconditionally. This allows applications to override it at the store
5624 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5625 to handle some structures.
5628 *) Improve efficiency of mem_gets: don't search whole buffer each time
5630 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5632 *) New -hex option for openssl rand.
5635 *) Print out UTF8String and NumericString when parsing ASN1.
5638 *) Support NumericString type for name components.
5641 *) Allow CC in the environment to override the automatically chosen
5642 compiler. Note that nothing is done to ensure flags work with the
5646 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5648 *) Properly check EVP_VerifyFinal() and similar return values
5650 [Ben Laurie, Bodo Moeller, Google Security Team]
5652 *) Enable TLS extensions by default.
5655 *) Allow the CHIL engine to be loaded, whether the application is
5656 multithreaded or not. (This does not release the developer from the
5657 obligation to set up the dynamic locking callbacks.)
5658 [Sander Temme <sander@temme.net>]
5660 *) Use correct exit code if there is an error in dgst command.
5661 [Steve Henson; problem pointed out by Roland Dirlewanger]
5663 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5664 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5667 *) Add experimental JPAKE support, including demo authentication in
5668 s_client and s_server.
5671 *) Set the comparison function in v3_addr_canonize().
5672 [Rob Austein <sra@hactrn.net>]
5674 *) Add support for XMPP STARTTLS in s_client.
5675 [Philip Paeps <philip@freebsd.org>]
5677 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5678 to ensure that even with this option, only ciphersuites in the
5679 server's preference list will be accepted. (Note that the option
5680 applies only when resuming a session, so the earlier behavior was
5681 just about the algorithm choice for symmetric cryptography.)
5684 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5686 *) Fix NULL pointer dereference if a DTLS server received
5687 ChangeCipherSpec as first record (CVE-2009-1386).
5690 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5691 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5694 *) The fix in 0.9.8c that supposedly got rid of unsafe
5695 double-checked locking was incomplete for RSA blinding,
5696 addressing just one layer of what turns out to have been
5697 doubly unsafe triple-checked locking.
5699 So now fix this for real by retiring the MONT_HELPER macro
5700 in crypto/rsa/rsa_eay.c.
5702 [Bodo Moeller; problem pointed out by Marius Schilder]
5704 *) Various precautionary measures:
5706 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5708 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5709 (NB: This would require knowledge of the secret session ticket key
5710 to exploit, in which case you'd be SOL either way.)
5712 - Change bn_nist.c so that it will properly handle input BIGNUMs
5713 outside the expected range.
5715 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5718 [Neel Mehta, Bodo Moeller]
5720 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5721 the load fails. Useful for distros.
5722 [Ben Laurie and the FreeBSD team]
5724 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5727 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5730 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5732 This work was sponsored by Logica.
5735 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5736 keystores. Support for SSL/TLS client authentication too.
5737 Not compiled unless enable-capieng specified to Configure.
5739 This work was sponsored by Logica.
5742 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5743 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5744 attribute creation routines such as certificate requests and PKCS#12
5748 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5750 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5751 handshake which could lead to a client crash as found using the
5752 Codenomicon TLS test suite (CVE-2008-1672)
5753 [Steve Henson, Mark Cox]
5755 *) Fix double free in TLS server name extensions which could lead to
5756 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5759 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5761 Clear the error queue to ensure that error entries left from
5762 older function calls do not interfere with the correct operation.
5763 [Lutz Jaenicke, Erik de Castro Lopo]
5765 *) Remove root CA certificates of commercial CAs:
5767 The OpenSSL project does not recommend any specific CA and does not
5768 have any policy with respect to including or excluding any CA.
5769 Therefore it does not make any sense to ship an arbitrary selection
5770 of root CA certificates with the OpenSSL software.
5773 *) RSA OAEP patches to fix two separate invalid memory reads.
5774 The first one involves inputs when 'lzero' is greater than
5775 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5776 before the beginning of from). The second one involves inputs where
5777 the 'db' section contains nothing but zeroes (there is a one-byte
5778 invalid read after the end of 'db').
5779 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5781 *) Partial backport from 0.9.9-dev:
5783 Introduce bn_mul_mont (dedicated Montgomery multiplication
5784 procedure) as a candidate for BIGNUM assembler implementation.
5785 While 0.9.9-dev uses assembler for various architectures, only
5786 x86_64 is available by default here in the 0.9.8 branch, and
5787 32-bit x86 is available through a compile-time setting.
5789 To try the 32-bit x86 assembler implementation, use Configure
5790 option "enable-montasm" (which exists only for this backport).
5792 As "enable-montasm" for 32-bit x86 disclaims code stability
5793 anyway, in this constellation we activate additional code
5794 backported from 0.9.9-dev for further performance improvements,
5795 namely BN_from_montgomery_word. (To enable this otherwise,
5796 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5798 [Andy Polyakov (backport partially by Bodo Moeller)]
5800 *) Add TLS session ticket callback. This allows an application to set
5801 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5802 values. This is useful for key rollover for example where several key
5803 sets may exist with different names.
5806 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5807 This was broken until now in 0.9.8 releases, such that the only way
5808 a registered ENGINE could be used (assuming it initialises
5809 successfully on the host) was to explicitly set it as the default
5810 for the relevant algorithms. This is in contradiction with 0.9.7
5811 behaviour and the documentation. With this fix, when an ENGINE is
5812 registered into a given algorithm's table of implementations, the
5813 'uptodate' flag is reset so that auto-discovery will be used next
5814 time a new context for that algorithm attempts to select an
5816 [Ian Lister (tweaked by Geoff Thorpe)]
5818 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5819 implementation in the following ways:
5821 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5824 Lack of BER streaming support means one pass streaming processing is
5825 only supported if data is detached: setting the streaming flag is
5826 ignored for embedded content.
5828 CMS support is disabled by default and must be explicitly enabled
5829 with the enable-cms configuration option.
5832 *) Update the GMP engine glue to do direct copies between BIGNUM and
5833 mpz_t when openssl and GMP use the same limb size. Otherwise the
5834 existing "conversion via a text string export" trick is still used.
5835 [Paul Sheer <paulsheer@gmail.com>]
5837 *) Zlib compression BIO. This is a filter BIO which compressed and
5838 uncompresses any data passed through it.
5841 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5842 RFC3394 compatible AES key wrapping.
5845 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5846 sets string data without copying. X509_ALGOR_set0() and
5847 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5848 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5849 from an X509_ATTRIBUTE structure optionally checking it occurs only
5850 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5854 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5855 to get the expected BN_FLG_CONSTTIME behavior.
5856 [Bodo Moeller (Google)]
5860 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5861 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5862 - added some more tests to do_tests.pl
5863 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5864 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5865 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5866 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5867 - various changes to netware.pl to enable gcc-cross builds on Win32
5869 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5870 - various changes to fix missing prototype warnings
5871 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5872 - added AES, WHIRLPOOL and CPUID assembler code to build files
5873 - added missing AES assembler make rules to mk1mf.pl
5874 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5875 [Guenter Knauf <eflash@gmx.net>]
5877 *) Implement certificate status request TLS extension defined in RFC3546.
5878 A client can set the appropriate parameters and receive the encoded
5879 OCSP response via a callback. A server can query the supplied parameters
5880 and set the encoded OCSP response in the callback. Add simplified examples
5881 to s_client and s_server.
5884 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5886 *) Fix various bugs:
5887 + Binary incompatibility of ssl_ctx_st structure
5888 + DTLS interoperation with non-compliant servers
5889 + Don't call get_session_cb() without proposed session
5890 + Fix ia64 assembler code
5891 [Andy Polyakov, Steve Henson]
5893 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5895 *) DTLS Handshake overhaul. There were longstanding issues with
5896 OpenSSL DTLS implementation, which were making it impossible for
5897 RFC 4347 compliant client to communicate with OpenSSL server.
5898 Unfortunately just fixing these incompatibilities would "cut off"
5899 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5900 server keeps tolerating non RFC compliant syntax. The opposite is
5901 not true, 0.9.8f client can not communicate with earlier server.
5902 This update even addresses CVE-2007-4995.
5905 *) Changes to avoid need for function casts in OpenSSL: some compilers
5906 (gcc 4.2 and later) reject their use.
5907 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5910 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5911 RFC4507bis. The encrypted ticket format is an encrypted encoded
5912 SSL_SESSION structure, that way new session features are automatically
5915 If a client application caches session in an SSL_SESSION structure
5916 support is transparent because tickets are now stored in the encoded
5919 The SSL_CTX structure automatically generates keys for ticket
5920 protection in servers so again support should be possible
5921 with no application modification.
5923 If a client or server wishes to disable RFC4507 support then the option
5924 SSL_OP_NO_TICKET can be set.
5926 Add a TLS extension debugging callback to allow the contents of any client
5927 or server extensions to be examined.
5929 This work was sponsored by Google.
5932 *) Add initial support for TLS extensions, specifically for the server_name
5933 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5934 have new members for a host name. The SSL data structure has an
5935 additional member SSL_CTX *initial_ctx so that new sessions can be
5936 stored in that context to allow for session resumption, even after the
5937 SSL has been switched to a new SSL_CTX in reaction to a client's
5938 server_name extension.
5940 New functions (subject to change):
5942 SSL_get_servername()
5943 SSL_get_servername_type()
5946 New CTRL codes and macros (subject to change):
5948 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5949 - SSL_CTX_set_tlsext_servername_callback()
5950 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5951 - SSL_CTX_set_tlsext_servername_arg()
5952 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5954 openssl s_client has a new '-servername ...' option.
5956 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5957 '-key2 ...', '-servername_fatal' (subject to change). This allows
5958 testing the HostName extension for a specific single host name ('-cert'
5959 and '-key' remain fallbacks for handshakes without HostName
5960 negotiation). If the unrecognized_name alert has to be sent, this by
5961 default is a warning; it becomes fatal with the '-servername_fatal'
5964 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5966 *) Add AES and SSE2 assembly language support to VC++ build.
5969 *) Mitigate attack on final subtraction in Montgomery reduction.
5972 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5973 (which previously caused an internal error).
5976 *) Squeeze another 10% out of IGE mode when in != out.
5979 *) AES IGE mode speedup.
5980 [Dean Gaudet (Google)]
5982 *) Add the Korean symmetric 128-bit cipher SEED (see
5983 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5984 add SEED ciphersuites from RFC 4162:
5986 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5987 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5988 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5989 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5991 To minimize changes between patchlevels in the OpenSSL 0.9.8
5992 series, SEED remains excluded from compilation unless OpenSSL
5993 is configured with 'enable-seed'.
5994 [KISA, Bodo Moeller]
5996 *) Mitigate branch prediction attacks, which can be practical if a
5997 single processor is shared, allowing a spy process to extract
5998 information. For detailed background information, see
5999 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
6000 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
6001 and Necessary Software Countermeasures"). The core of the change
6002 are new versions BN_div_no_branch() and
6003 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
6004 respectively, which are slower, but avoid the security-relevant
6005 conditional branches. These are automatically called by BN_div()
6006 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
6007 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
6008 remove a conditional branch.
6010 BN_FLG_CONSTTIME is the new name for the previous
6011 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
6012 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
6013 in the exponent causes BN_mod_exp_mont() to use the alternative
6014 implementation in BN_mod_exp_mont_consttime().) The old name
6015 remains as a deprecated alias.
6017 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
6018 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
6019 constant-time implementations for more than just exponentiation.
6020 Here too the old name is kept as a deprecated alias.
6022 BN_BLINDING_new() will now use BN_dup() for the modulus so that
6023 the BN_BLINDING structure gets an independent copy of the
6024 modulus. This means that the previous "BIGNUM *m" argument to
6025 BN_BLINDING_new() and to BN_BLINDING_create_param() now
6026 essentially becomes "const BIGNUM *m", although we can't actually
6027 change this in the header file before 0.9.9. It allows
6028 RSA_setup_blinding() to use BN_with_flags() on the modulus to
6029 enable BN_FLG_CONSTTIME.
6031 [Matthew D Wood (Intel Corp)]
6033 *) In the SSL/TLS server implementation, be strict about session ID
6034 context matching (which matters if an application uses a single
6035 external cache for different purposes). Previously,
6036 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
6037 set. This did ensure strict client verification, but meant that,
6038 with applications using a single external cache for quite
6039 different requirements, clients could circumvent ciphersuite
6040 restrictions for a given session ID context by starting a session
6041 in a different context.
6044 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6045 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6046 authentication-only ciphersuites.
6049 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
6050 not complete and could lead to a possible single byte overflow
6051 (CVE-2007-5135) [Ben Laurie]
6053 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
6055 *) Since AES128 and AES256 (and similarly Camellia128 and
6056 Camellia256) share a single mask bit in the logic of
6057 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6058 kludge to work properly if AES128 is available and AES256 isn't
6059 (or if Camellia128 is available and Camellia256 isn't).
6062 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
6063 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
6064 When a point or a seed is encoded in a BIT STRING, we need to
6065 prevent the removal of trailing zero bits to get the proper DER
6066 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
6067 of a NamedBitList, for which trailing 0 bits need to be removed.)
6070 *) Have SSL/TLS server implementation tolerate "mismatched" record
6071 protocol version while receiving ClientHello even if the
6072 ClientHello is fragmented. (The server can't insist on the
6073 particular protocol version it has chosen before the ServerHello
6074 message has informed the client about his choice.)
6077 *) Add RFC 3779 support.
6078 [Rob Austein for ARIN, Ben Laurie]
6080 *) Load error codes if they are not already present instead of using a
6081 static variable. This allows them to be cleanly unloaded and reloaded.
6082 Improve header file function name parsing.
6085 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
6086 or CAPABILITY handshake as required by RFCs.
6089 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
6091 *) Introduce limits to prevent malicious keys being able to
6092 cause a denial of service. (CVE-2006-2940)
6093 [Steve Henson, Bodo Moeller]
6095 *) Fix ASN.1 parsing of certain invalid structures that can result
6096 in a denial of service. (CVE-2006-2937) [Steve Henson]
6098 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6099 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6101 *) Fix SSL client code which could crash if connecting to a
6102 malicious SSLv2 server. (CVE-2006-4343)
6103 [Tavis Ormandy and Will Drewry, Google Security Team]
6105 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
6106 match only those. Before that, "AES256-SHA" would be interpreted
6107 as a pattern and match "AES128-SHA" too (since AES128-SHA got
6108 the same strength classification in 0.9.7h) as we currently only
6109 have a single AES bit in the ciphersuite description bitmap.
6110 That change, however, also applied to ciphersuite strings such as
6111 "RC4-MD5" that intentionally matched multiple ciphersuites --
6112 namely, SSL 2.0 ciphersuites in addition to the more common ones
6113 from SSL 3.0/TLS 1.0.
6115 So we change the selection algorithm again: Naming an explicit
6116 ciphersuite selects this one ciphersuite, and any other similar
6117 ciphersuite (same bitmap) from *other* protocol versions.
6118 Thus, "RC4-MD5" again will properly select both the SSL 2.0
6119 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
6121 Since SSL 2.0 does not have any ciphersuites for which the
6122 128/256 bit distinction would be relevant, this works for now.
6123 The proper fix will be to use different bits for AES128 and
6124 AES256, which would have avoided the problems from the beginning;
6125 however, bits are scarce, so we can only do this in a new release
6126 (not just a patchlevel) when we can change the SSL_CIPHER
6127 definition to split the single 'unsigned long mask' bitmap into
6128 multiple values to extend the available space.
6132 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
6134 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6135 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6137 *) Add AES IGE and biIGE modes.
6140 *) Change the Unix randomness entropy gathering to use poll() when
6141 possible instead of select(), since the latter has some
6142 undesirable limitations.
6143 [Darryl Miles via Richard Levitte and Bodo Moeller]
6145 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
6146 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
6147 cannot be implicitly activated as part of, e.g., the "AES" alias.
6148 However, please upgrade to OpenSSL 0.9.9[-dev] for
6149 non-experimental use of the ECC ciphersuites to get TLS extension
6150 support, which is required for curve and point format negotiation
6151 to avoid potential handshake problems.
6154 *) Disable rogue ciphersuites:
6156 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6157 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6158 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6160 The latter two were purportedly from
6161 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6164 Also deactivate the remaining ciphersuites from
6165 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6166 unofficial, and the ID has long expired.
6169 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6170 dual-core machines) and other potential thread-safety issues.
6173 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
6174 versions), which is now available for royalty-free use
6175 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
6176 Also, add Camellia TLS ciphersuites from RFC 4132.
6178 To minimize changes between patchlevels in the OpenSSL 0.9.8
6179 series, Camellia remains excluded from compilation unless OpenSSL
6180 is configured with 'enable-camellia'.
6183 *) Disable the padding bug check when compression is in use. The padding
6184 bug check assumes the first packet is of even length, this is not
6185 necessarily true if compression is enabled and can result in false
6186 positives causing handshake failure. The actual bug test is ancient
6187 code so it is hoped that implementations will either have fixed it by
6188 now or any which still have the bug do not support compression.
6191 Changes between 0.9.8a and 0.9.8b [04 May 2006]
6193 *) When applying a cipher rule check to see if string match is an explicit
6194 cipher suite and only match that one cipher suite if it is.
6197 *) Link in manifests for VC++ if needed.
6198 [Austin Ziegler <halostatue@gmail.com>]
6200 *) Update support for ECC-based TLS ciphersuites according to
6201 draft-ietf-tls-ecc-12.txt with proposed changes (but without
6202 TLS extensions, which are supported starting with the 0.9.9
6203 branch, not in the OpenSSL 0.9.8 branch).
6206 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
6207 opaque EVP_CIPHER_CTX handling.
6210 *) Fixes and enhancements to zlib compression code. We now only use
6211 "zlib1.dll" and use the default __cdecl calling convention on Win32
6212 to conform with the standards mentioned here:
6213 http://www.zlib.net/DLL_FAQ.txt
6214 Static zlib linking now works on Windows and the new --with-zlib-include
6215 --with-zlib-lib options to Configure can be used to supply the location
6216 of the headers and library. Gracefully handle case where zlib library
6220 *) Several fixes and enhancements to the OID generation code. The old code
6221 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
6222 handle numbers larger than ULONG_MAX, truncated printing and had a
6223 non standard OBJ_obj2txt() behaviour.
6226 *) Add support for building of engines under engine/ as shared libraries
6227 under VC++ build system.
6230 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
6231 Hopefully, we will not see any false combination of paths any more.
6234 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
6236 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6237 (part of SSL_OP_ALL). This option used to disable the
6238 countermeasure against man-in-the-middle protocol-version
6239 rollback in the SSL 2.0 server implementation, which is a bad
6240 idea. (CVE-2005-2969)
6242 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6243 for Information Security, National Institute of Advanced Industrial
6244 Science and Technology [AIST], Japan)]
6246 *) Add two function to clear and return the verify parameter flags.
6249 *) Keep cipherlists sorted in the source instead of sorting them at
6250 runtime, thus removing the need for a lock.
6253 *) Avoid some small subgroup attacks in Diffie-Hellman.
6254 [Nick Mathewson and Ben Laurie]
6256 *) Add functions for well-known primes.
6259 *) Extended Windows CE support.
6260 [Satoshi Nakamura and Andy Polyakov]
6262 *) Initialize SSL_METHOD structures at compile time instead of during
6263 runtime, thus removing the need for a lock.
6266 *) Make PKCS7_decrypt() work even if no certificate is supplied by
6267 attempting to decrypt each encrypted key in turn. Add support to
6271 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
6273 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
6276 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
6279 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
6280 key into the same file any more.
6283 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
6286 *) Add -utf8 command line and config file option to 'ca'.
6287 [Stefan <stf@udoma.org]
6289 *) Removed the macro des_crypt(), as it seems to conflict with some
6290 libraries. Use DES_crypt().
6293 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
6294 involves renaming the source and generated shared-libs for
6295 both. The engines will accept the corrected or legacy ids
6296 ('ncipher' and '4758_cca' respectively) when binding. NB,
6297 this only applies when building 'shared'.
6298 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
6300 *) Add attribute functions to EVP_PKEY structure. Modify
6301 PKCS12_create() to recognize a CSP name attribute and
6302 use it. Make -CSP option work again in pkcs12 utility.
6305 *) Add new functionality to the bn blinding code:
6306 - automatic re-creation of the BN_BLINDING parameters after
6307 a fixed number of uses (currently 32)
6308 - add new function for parameter creation
6309 - introduce flags to control the update behaviour of the
6310 BN_BLINDING parameters
6311 - hide BN_BLINDING structure
6312 Add a second BN_BLINDING slot to the RSA structure to improve
6313 performance when a single RSA object is shared among several
6317 *) Add support for DTLS.
6318 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
6320 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
6321 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
6324 *) Remove buggy and incomplete DH cert support from
6325 ssl/ssl_rsa.c and ssl/s3_both.c
6328 *) Use SHA-1 instead of MD5 as the default digest algorithm for
6329 the apps/openssl applications.
6332 *) Compile clean with "-Wall -Wmissing-prototypes
6333 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
6334 DEBUG_SAFESTACK must also be set.
6337 *) Change ./Configure so that certain algorithms can be disabled by default.
6338 The new counterpiece to "no-xxx" is "enable-xxx".
6340 The patented RC5 and MDC2 algorithms will now be disabled unless
6341 "enable-rc5" and "enable-mdc2", respectively, are specified.
6343 (IDEA remains enabled despite being patented. This is because IDEA
6344 is frequently required for interoperability, and there is no license
6345 fee for non-commercial use. As before, "no-idea" can be used to
6346 avoid this algorithm.)
6350 *) Add processing of proxy certificates (see RFC 3820). This work was
6351 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
6352 EGEE (Enabling Grids for E-science in Europe).
6355 *) RC4 performance overhaul on modern architectures/implementations, such
6356 as Intel P4, IA-64 and AMD64.
6359 *) New utility extract-section.pl. This can be used specify an alternative
6360 section number in a pod file instead of having to treat each file as
6361 a separate case in Makefile. This can be done by adding two lines to the
6364 =for comment openssl_section:XXX
6366 The blank line is mandatory.
6370 *) New arguments -certform, -keyform and -pass for s_client and s_server
6371 to allow alternative format key and certificate files and passphrase
6375 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
6376 update associated structures and add various utility functions.
6378 Add new policy related verify parameters, include policy checking in
6379 standard verify code. Enhance 'smime' application with extra parameters
6380 to support policy checking and print out.
6383 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
6384 Nehemiah processors. These extensions support AES encryption in hardware
6385 as well as RNG (though RNG support is currently disabled).
6386 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
6388 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
6391 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
6392 [Andy Polyakov and a number of other people]
6394 *) Improved PowerPC platform support. Most notably BIGNUM assembler
6395 implementation contributed by IBM.
6396 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
6398 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
6399 exponent rather than 'unsigned long'. There is a corresponding change to
6400 the new 'rsa_keygen' element of the RSA_METHOD structure.
6401 [Jelte Jansen, Geoff Thorpe]
6403 *) Functionality for creating the initial serial number file is now
6404 moved from CA.pl to the 'ca' utility with a new option -create_serial.
6406 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
6407 number file to 1, which is bound to cause problems. To avoid
6408 the problems while respecting compatibility between different 0.9.7
6409 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
6410 CA.pl for serial number initialization. With the new release 0.9.8,
6411 we can fix the problem directly in the 'ca' utility.)
6414 *) Reduced header interdependencies by declaring more opaque objects in
6415 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
6416 give fewer recursive includes, which could break lazy source code - so
6417 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
6418 developers should define this symbol when building and using openssl to
6419 ensure they track the recommended behaviour, interfaces, [etc], but
6420 backwards-compatible behaviour prevails when this isn't defined.
6423 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
6426 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
6427 This will generate a random key of the appropriate length based on the
6428 cipher context. The EVP_CIPHER can provide its own random key generation
6429 routine to support keys of a specific form. This is used in the des and
6430 3des routines to generate a key of the correct parity. Update S/MIME
6431 code to use new functions and hence generate correct parity DES keys.
6432 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
6433 valid (weak or incorrect parity).
6436 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
6437 as looking them up. This is useful when the verified structure may contain
6438 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
6439 present unless the new PKCS7_NO_CRL flag is asserted.
6442 *) Extend ASN1 oid configuration module. It now additionally accepts the
6445 shortName = some long name, 1.2.3.4
6448 *) Reimplemented the BN_CTX implementation. There is now no more static
6449 limitation on the number of variables it can handle nor the depth of the
6450 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
6451 information can now expand as required, and rather than having a single
6452 static array of bignums, BN_CTX now uses a linked-list of such arrays
6453 allowing it to expand on demand whilst maintaining the usefulness of
6454 BN_CTX's "bundling".
6457 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
6458 to allow all RSA operations to function using a single BN_CTX.
6461 *) Preliminary support for certificate policy evaluation and checking. This
6462 is initially intended to pass the tests outlined in "Conformance Testing
6463 of Relying Party Client Certificate Path Processing Logic" v1.07.
6466 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
6467 remained unused and not that useful. A variety of other little bignum
6468 tweaks and fixes have also been made continuing on from the audit (see
6472 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
6473 associated ASN1, EVP and SSL functions and old ASN1 macros.
6476 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
6477 and this should never fail. So the return value from the use of
6478 BN_set_word() (which can fail due to needless expansion) is now deprecated;
6479 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
6482 *) BN_CTX_get() should return zero-valued bignums, providing the same
6483 initialised value as BN_new().
6484 [Geoff Thorpe, suggested by Ulf Möller]
6486 *) Support for inhibitAnyPolicy certificate extension.
6489 *) An audit of the BIGNUM code is underway, for which debugging code is
6490 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
6491 is considered valid when processing BIGNUMs, and causes execution to
6492 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
6493 further steps are taken to deliberately pollute unused data in BIGNUM
6494 structures to try and expose faulty code further on. For now, openssl will
6495 (in its default mode of operation) continue to tolerate the inconsistent
6496 forms that it has tolerated in the past, but authors and packagers should
6497 consider trying openssl and their own applications when compiled with
6498 these debugging symbols defined. It will help highlight potential bugs in
6499 their own code, and will improve the test coverage for OpenSSL itself. At
6500 some point, these tighter rules will become openssl's default to improve
6501 maintainability, though the assert()s and other overheads will remain only
6502 in debugging configurations. See bn.h for more details.
6503 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6505 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6506 that can only be obtained through BN_CTX_new() (which implicitly
6507 initialises it). The presence of this function only made it possible
6508 to overwrite an existing structure (and cause memory leaks).
6511 *) Because of the callback-based approach for implementing LHASH as a
6512 template type, lh_insert() adds opaque objects to hash-tables and
6513 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6514 to clean up those corresponding objects before destroying the hash table
6515 (and losing the object pointers). So some over-zealous constifications in
6516 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6517 objects as "const" and the lh_doall[_arg] callback wrappers are not
6518 prototyped to have "const" restrictions on the object pointers they are
6519 given (and so aren't required to cast them away any more).
6522 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6523 (speed) prefers to use its own implementation. The two implementations
6524 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6525 its object type properly exposed (MS_TM) instead of casting to/from "char
6526 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6527 aren't necessarily the greatest nomenclatures - but this is what was used
6528 internally to the implementation so I've used that for now.
6531 *) Ensure that deprecated functions do not get compiled when
6532 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6533 the self-tests were still using deprecated key-generation functions so
6534 these have been updated also.
6537 *) Reorganise PKCS#7 code to separate the digest location functionality
6538 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6539 New function PKCS7_set_digest() to set the digest type for PKCS#7
6540 digestedData type. Add additional code to correctly generate the
6541 digestedData type and add support for this type in PKCS7 initialization
6545 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6546 structure of type "other".
6549 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6550 sure the loop does correctly stop and breaking ("division by zero")
6551 modulus operations are not performed. The (pre-generated) prime
6552 table crypto/bn/bn_prime.h was already correct, but it could not be
6553 re-generated on some platforms because of the "division by zero"
6554 situation in the script.
6555 [Ralf S. Engelschall]
6557 *) Update support for ECC-based TLS ciphersuites according to
6558 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6559 SHA-1 now is only used for "small" curves (where the
6560 representation of a field element takes up to 24 bytes); for
6561 larger curves, the field element resulting from ECDH is directly
6562 used as premaster secret.
6563 [Douglas Stebila (Sun Microsystems Laboratories)]
6565 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6566 curve secp160r1 to the tests.
6567 [Douglas Stebila (Sun Microsystems Laboratories)]
6569 *) Add the possibility to load symbols globally with DSO.
6570 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6572 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6573 control of the error stack.
6576 *) Add support for STORE in ENGINE.
6579 *) Add the STORE type. The intention is to provide a common interface
6580 to certificate and key stores, be they simple file-based stores, or
6581 HSM-type store, or LDAP stores, or...
6582 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6585 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6586 pass a list of arguments to any function as well as provide a way
6587 for a function to pass data back to the caller.
6590 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6591 works like BUF_strdup() but can be used to duplicate a portion of
6592 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6596 *) Add the function sk_find_ex() which works like sk_find(), but will
6597 return an index to an element even if an exact match couldn't be
6598 found. The index is guaranteed to point at the element where the
6599 searched-for key would be inserted to preserve sorting order.
6602 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6603 takes an extra flags argument for optional functionality. Currently,
6604 the following flags are defined:
6606 OBJ_BSEARCH_VALUE_ON_NOMATCH
6607 This one gets OBJ_bsearch_ex() to return a pointer to the first
6608 element where the comparing function returns a negative or zero
6611 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6612 This one gets OBJ_bsearch_ex() to return a pointer to the first
6613 element where the comparing function returns zero. This is useful
6614 if there are more than one element where the comparing function
6618 *) Make it possible to create self-signed certificates with 'openssl ca'
6619 in such a way that the self-signed certificate becomes part of the
6620 CA database and uses the same mechanisms for serial number generation
6621 as all other certificate signing. The new flag '-selfsign' enables
6622 this functionality. Adapt CA.sh and CA.pl.in.
6625 *) Add functionality to check the public key of a certificate request
6626 against a given private. This is useful to check that a certificate
6627 request can be signed by that key (self-signing).
6630 *) Make it possible to have multiple active certificates with the same
6631 subject in the CA index file. This is done only if the keyword
6632 'unique_subject' is set to 'no' in the main CA section (default
6633 if 'CA_default') of the configuration file. The value is saved
6634 with the database itself in a separate index attribute file,
6635 named like the index file with '.attr' appended to the name.
6638 *) Generate multi-valued AVAs using '+' notation in config files for
6642 *) Support for nameConstraints certificate extension.
6645 *) Support for policyConstraints certificate extension.
6648 *) Support for policyMappings certificate extension.
6651 *) Make sure the default DSA_METHOD implementation only uses its
6652 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6653 and change its own handlers to be NULL so as to remove unnecessary
6654 indirection. This lets alternative implementations fallback to the
6655 default implementation more easily.
6658 *) Support for directoryName in GeneralName related extensions
6662 *) Make it possible to link applications using Makefile.shared.
6663 Make that possible even when linking against static libraries!
6666 *) Support for single pass processing for S/MIME signing. This now
6667 means that S/MIME signing can be done from a pipe, in addition
6668 cleartext signing (multipart/signed type) is effectively streaming
6669 and the signed data does not need to be all held in memory.
6671 This is done with a new flag PKCS7_STREAM. When this flag is set
6672 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6673 is done after the data is output (and digests calculated) in
6674 SMIME_write_PKCS7().
6677 *) Add full support for -rpath/-R, both in shared libraries and
6678 applications, at least on the platforms where it's known how
6682 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6683 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6684 will now compute a table of multiples of the generator that
6685 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6686 faster (notably in the case of a single point multiplication,
6687 scalar * generator).
6688 [Nils Larsch, Bodo Moeller]
6690 *) IPv6 support for certificate extensions. The various extensions
6691 which use the IP:a.b.c.d can now take IPv6 addresses using the
6692 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6696 *) Added an ENGINE that implements RSA by performing private key
6697 exponentiations with the GMP library. The conversions to and from
6698 GMP's mpz_t format aren't optimised nor are any montgomery forms
6699 cached, and on x86 it appears OpenSSL's own performance has caught up.
6700 However there are likely to be other architectures where GMP could
6701 provide a boost. This ENGINE is not built in by default, but it can be
6702 specified at Configure time and should be accompanied by the necessary
6703 linker additions, eg;
6704 ./config -DOPENSSL_USE_GMP -lgmp
6707 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6708 testing availability of engines with "-t" - the old behaviour is
6709 produced by increasing the feature's verbosity with "-tt".
6712 *) ECDSA routines: under certain error conditions uninitialized BN objects
6713 could be freed. Solution: make sure initialization is performed early
6714 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6718 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6719 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6720 software implementations. For DSA and DH, parameter generation can
6721 also be overridden by providing the appropriate method callbacks.
6724 *) Change the "progress" mechanism used in key-generation and
6725 primality testing to functions that take a new BN_GENCB pointer in
6726 place of callback/argument pairs. The new API functions have "_ex"
6727 postfixes and the older functions are reimplemented as wrappers for
6728 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6729 declarations of the old functions to help (graceful) attempts to
6730 migrate to the new functions. Also, the new key-generation API
6731 functions operate on a caller-supplied key-structure and return
6732 success/failure rather than returning a key or NULL - this is to
6733 help make "keygen" another member function of RSA_METHOD etc.
6735 Example for using the new callback interface:
6737 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6741 BN_GENCB_set(&my_cb, my_callback, my_arg);
6743 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6744 /* For the meaning of a, b in calls to my_callback(), see the
6745 * documentation of the function that calls the callback.
6746 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6747 * my_callback should return 1 if it wants BN_is_prime_ex()
6748 * to continue, or 0 to stop.
6753 *) Change the ZLIB compression method to be stateful, and make it
6754 available to TLS with the number defined in
6755 draft-ietf-tls-compression-04.txt.
6758 *) Add the ASN.1 structures and functions for CertificatePair, which
6759 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6761 CertificatePair ::= SEQUENCE {
6762 forward [0] Certificate OPTIONAL,
6763 reverse [1] Certificate OPTIONAL,
6764 -- at least one of the pair shall be present -- }
6766 Also implement the PEM functions to read and write certificate
6767 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6769 This needed to be defined, mostly for the sake of the LDAP
6770 attribute crossCertificatePair, but may prove useful elsewhere as
6774 *) Make it possible to inhibit symlinking of shared libraries in
6775 Makefile.shared, for Cygwin's sake.
6778 *) Extend the BIGNUM API by creating a function
6779 void BN_set_negative(BIGNUM *a, int neg);
6780 and a macro that behave like
6781 int BN_is_negative(const BIGNUM *a);
6783 to avoid the need to access 'a->neg' directly in applications.
6786 *) Implement fast modular reduction for pseudo-Mersenne primes
6787 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6788 EC_GROUP_new_curve_GFp() will now automatically use this
6790 [Nils Larsch <nla@trustcenter.de>]
6792 *) Add new lock type (CRYPTO_LOCK_BN).
6795 *) Change the ENGINE framework to automatically load engines
6796 dynamically from specific directories unless they could be
6797 found to already be built in or loaded. Move all the
6798 current engines except for the cryptodev one to a new
6800 The engines in engines/ are built as shared libraries if
6801 the "shared" options was given to ./Configure or ./config.
6802 Otherwise, they are inserted in libcrypto.a.
6803 /usr/local/ssl/engines is the default directory for dynamic
6804 engines, but that can be overridden at configure time through
6805 the usual use of --prefix and/or --openssldir, and at run
6806 time with the environment variable OPENSSL_ENGINES.
6807 [Geoff Thorpe and Richard Levitte]
6809 *) Add Makefile.shared, a helper makefile to build shared
6810 libraries. Adapt Makefile.org.
6813 *) Add version info to Win32 DLLs.
6814 [Peter 'Luna' Runestig" <peter@runestig.com>]
6816 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6817 can be added using this API to created arbitrary PKCS#12
6818 files while avoiding the low level API.
6820 New options to PKCS12_create(), key or cert can be NULL and
6821 will then be omitted from the output file. The encryption
6822 algorithm NIDs can be set to -1 for no encryption, the mac
6823 iteration count can be set to 0 to omit the mac.
6825 Enhance pkcs12 utility by making the -nokeys and -nocerts
6826 options work when creating a PKCS#12 file. New option -nomac
6827 to omit the mac, NONE can be set for an encryption algorithm.
6828 New code is modified to use the enhanced PKCS12_create()
6829 instead of the low level API.
6832 *) Extend ASN1 encoder to support indefinite length constructed
6833 encoding. This can output sequences tags and octet strings in
6834 this form. Modify pk7_asn1.c to support indefinite length
6835 encoding. This is experimental and needs additional code to
6836 be useful, such as an ASN1 bio and some enhanced streaming
6839 Extend template encode functionality so that tagging is passed
6840 down to the template encoder.
6843 *) Let 'openssl req' fail if an argument to '-newkey' is not
6844 recognized instead of using RSA as a default.
6847 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6848 As these are not official, they are not included in "ALL";
6849 the "ECCdraft" ciphersuite group alias can be used to select them.
6850 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6852 *) Add ECDH engine support.
6853 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6855 *) Add ECDH in new directory crypto/ecdh/.
6856 [Douglas Stebila (Sun Microsystems Laboratories)]
6858 *) Let BN_rand_range() abort with an error after 100 iterations
6859 without success (which indicates a broken PRNG).
6862 *) Change BN_mod_sqrt() so that it verifies that the input value
6863 is really the square of the return value. (Previously,
6864 BN_mod_sqrt would show GIGO behaviour.)
6867 *) Add named elliptic curves over binary fields from X9.62, SECG,
6868 and WAP/WTLS; add OIDs that were still missing.
6870 [Sheueling Chang Shantz and Douglas Stebila
6871 (Sun Microsystems Laboratories)]
6873 *) Extend the EC library for elliptic curves over binary fields
6874 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6877 EC_GF2m_simple_method
6881 EC_GROUP_new_curve_GF2m
6882 EC_GROUP_set_curve_GF2m
6883 EC_GROUP_get_curve_GF2m
6884 EC_POINT_set_affine_coordinates_GF2m
6885 EC_POINT_get_affine_coordinates_GF2m
6886 EC_POINT_set_compressed_coordinates_GF2m
6888 Point compression for binary fields is disabled by default for
6889 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6892 As binary polynomials are represented as BIGNUMs, various members
6893 of the EC_GROUP and EC_POINT data structures can be shared
6894 between the implementations for prime fields and binary fields;
6895 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6896 are essentially identical to their ..._GFp counterparts.
6897 (For simplicity, the '..._GFp' prefix has been dropped from
6898 various internal method names.)
6900 An internal 'field_div' method (similar to 'field_mul' and
6901 'field_sqr') has been added; this is used only for binary fields.
6903 [Sheueling Chang Shantz and Douglas Stebila
6904 (Sun Microsystems Laboratories)]
6906 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6907 through methods ('mul', 'precompute_mult').
6909 The generic implementations (now internally called 'ec_wNAF_mul'
6910 and 'ec_wNAF_precomputed_mult') remain the default if these
6911 methods are undefined.
6913 [Sheueling Chang Shantz and Douglas Stebila
6914 (Sun Microsystems Laboratories)]
6916 *) New function EC_GROUP_get_degree, which is defined through
6917 EC_METHOD. For curves over prime fields, this returns the bit
6918 length of the modulus.
6920 [Sheueling Chang Shantz and Douglas Stebila
6921 (Sun Microsystems Laboratories)]
6923 *) New functions EC_GROUP_dup, EC_POINT_dup.
6924 (These simply call ..._new and ..._copy).
6926 [Sheueling Chang Shantz and Douglas Stebila
6927 (Sun Microsystems Laboratories)]
6929 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6930 Polynomials are represented as BIGNUMs (where the sign bit is not
6931 used) in the following functions [macros]:
6934 BN_GF2m_sub [= BN_GF2m_add]
6935 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6936 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6937 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6939 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6940 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6941 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6942 BN_GF2m_cmp [= BN_ucmp]
6944 (Note that only the 'mod' functions are actually for fields GF(2^m).
6945 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6947 For some functions, an the irreducible polynomial defining a
6948 field can be given as an 'unsigned int[]' with strictly
6949 decreasing elements giving the indices of those bits that are set;
6950 i.e., p[] represents the polynomial
6951 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6953 p[0] > p[1] > ... > p[k] = 0.
6954 This applies to the following functions:
6959 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6960 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6962 BN_GF2m_mod_sqrt_arr
6963 BN_GF2m_mod_solve_quad_arr
6967 Conversion can be performed by the following functions:
6972 bntest.c has additional tests for binary polynomial arithmetic.
6974 Two implementations for BN_GF2m_mod_div() are available.
6975 The default algorithm simply uses BN_GF2m_mod_inv() and
6976 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6977 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6978 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6980 [Sheueling Chang Shantz and Douglas Stebila
6981 (Sun Microsystems Laboratories)]
6983 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6984 functionality is disabled at compile-time.
6985 [Douglas Stebila <douglas.stebila@sun.com>]
6987 *) Change default behaviour of 'openssl asn1parse' so that more
6988 information is visible when viewing, e.g., a certificate:
6990 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6991 mode the content of non-printable OCTET STRINGs is output in a
6992 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6993 avoid the appearance of a printable string.
6994 [Nils Larsch <nla@trustcenter.de>]
6996 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6998 EC_GROUP_set_asn1_flag()
6999 EC_GROUP_get_asn1_flag()
7000 EC_GROUP_set_point_conversion_form()
7001 EC_GROUP_get_point_conversion_form()
7002 These control ASN1 encoding details:
7003 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
7004 has been set to OPENSSL_EC_NAMED_CURVE.
7005 - Points are encoded in uncompressed form by default; options for
7006 asn1_for are as for point2oct, namely
7007 POINT_CONVERSION_COMPRESSED
7008 POINT_CONVERSION_UNCOMPRESSED
7009 POINT_CONVERSION_HYBRID
7011 Also add 'seed' and 'seed_len' members to EC_GROUP with access
7014 EC_GROUP_get0_seed()
7015 EC_GROUP_get_seed_len()
7016 This is used only for ASN1 purposes (so far).
7017 [Nils Larsch <nla@trustcenter.de>]
7019 *) Add 'field_type' member to EC_METHOD, which holds the NID
7020 of the appropriate field type OID. The new function
7021 EC_METHOD_get_field_type() returns this value.
7022 [Nils Larsch <nla@trustcenter.de>]
7027 EC_POINT_point2hex()
7028 EC_POINT_hex2point()
7029 providing useful interfaces to EC_POINT_point2oct() and
7030 EC_POINT_oct2point().
7031 [Nils Larsch <nla@trustcenter.de>]
7033 *) Change internals of the EC library so that the functions
7034 EC_GROUP_set_generator()
7035 EC_GROUP_get_generator()
7036 EC_GROUP_get_order()
7037 EC_GROUP_get_cofactor()
7038 are implemented directly in crypto/ec/ec_lib.c and not dispatched
7039 to methods, which would lead to unnecessary code duplication when
7040 adding different types of curves.
7041 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
7043 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
7044 arithmetic, and such that modified wNAFs are generated
7045 (which avoid length expansion in many cases).
7048 *) Add a function EC_GROUP_check_discriminant() (defined via
7049 EC_METHOD) that verifies that the curve discriminant is non-zero.
7051 Add a function EC_GROUP_check() that makes some sanity tests
7052 on a EC_GROUP, its generator and order. This includes
7053 EC_GROUP_check_discriminant().
7054 [Nils Larsch <nla@trustcenter.de>]
7056 *) Add ECDSA in new directory crypto/ecdsa/.
7058 Add applications 'openssl ecparam' and 'openssl ecdsa'
7059 (these are based on 'openssl dsaparam' and 'openssl dsa').
7061 ECDSA support is also included in various other files across the
7062 library. Most notably,
7063 - 'openssl req' now has a '-newkey ecdsa:file' option;
7064 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
7065 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
7066 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
7067 them suitable for ECDSA where domain parameters must be
7068 extracted before the specific public key;
7069 - ECDSA engine support has been added.
7070 [Nils Larsch <nla@trustcenter.de>]
7072 *) Include some named elliptic curves, and add OIDs from X9.62,
7073 SECG, and WAP/WTLS. Each curve can be obtained from the new
7075 EC_GROUP_new_by_curve_name(),
7076 and the list of available named curves can be obtained with
7077 EC_get_builtin_curves().
7078 Also add a 'curve_name' member to EC_GROUP objects, which can be
7080 EC_GROUP_set_curve_name()
7081 EC_GROUP_get_curve_name()
7082 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
7084 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7085 was actually never needed) and in BN_mul(). The removal in BN_mul()
7086 required a small change in bn_mul_part_recursive() and the addition
7087 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7088 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7089 bn_sub_words() and bn_add_words() except they take arrays with
7093 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
7095 *) Cleanse PEM buffers before freeing them since they may contain
7097 [Benjamin Bennett <ben@psc.edu>]
7099 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
7100 a ciphersuite string such as "DEFAULT:RSA" cannot enable
7101 authentication-only ciphersuites.
7104 *) Since AES128 and AES256 share a single mask bit in the logic of
7105 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
7106 kludge to work properly if AES128 is available and AES256 isn't.
7109 *) Expand security boundary to match 1.1.1 module.
7112 *) Remove redundant features: hash file source, editing of test vectors
7113 modify fipsld to use external fips_premain.c signature.
7116 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
7117 run algorithm test programs.
7120 *) Make algorithm test programs more tolerant of whitespace.
7123 *) Have SSL/TLS server implementation tolerate "mismatched" record
7124 protocol version while receiving ClientHello even if the
7125 ClientHello is fragmented. (The server can't insist on the
7126 particular protocol version it has chosen before the ServerHello
7127 message has informed the client about his choice.)
7130 *) Load error codes if they are not already present instead of using a
7131 static variable. This allows them to be cleanly unloaded and reloaded.
7134 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
7136 *) Introduce limits to prevent malicious keys being able to
7137 cause a denial of service. (CVE-2006-2940)
7138 [Steve Henson, Bodo Moeller]
7140 *) Fix ASN.1 parsing of certain invalid structures that can result
7141 in a denial of service. (CVE-2006-2937) [Steve Henson]
7143 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
7144 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
7146 *) Fix SSL client code which could crash if connecting to a
7147 malicious SSLv2 server. (CVE-2006-4343)
7148 [Tavis Ormandy and Will Drewry, Google Security Team]
7150 *) Change ciphersuite string processing so that an explicit
7151 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
7152 will no longer include "AES128-SHA"), and any other similar
7153 ciphersuite (same bitmap) from *other* protocol versions (so that
7154 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
7155 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
7156 changes from 0.9.8b and 0.9.8d.
7159 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
7161 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
7162 (CVE-2006-4339) [Ben Laurie and Google Security Team]
7164 *) Change the Unix randomness entropy gathering to use poll() when
7165 possible instead of select(), since the latter has some
7166 undesirable limitations.
7167 [Darryl Miles via Richard Levitte and Bodo Moeller]
7169 *) Disable rogue ciphersuites:
7171 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
7172 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
7173 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
7175 The latter two were purportedly from
7176 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
7179 Also deactivate the remaining ciphersuites from
7180 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
7181 unofficial, and the ID has long expired.
7184 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
7185 dual-core machines) and other potential thread-safety issues.
7188 Changes between 0.9.7i and 0.9.7j [04 May 2006]
7190 *) Adapt fipsld and the build system to link against the validated FIPS
7191 module in FIPS mode.
7194 *) Fixes for VC++ 2005 build under Windows.
7197 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
7198 from a Windows bash shell such as MSYS. It is autodetected from the
7199 "config" script when run from a VC++ environment. Modify standard VC++
7200 build to use fipscanister.o from the GNU make build.
7203 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
7205 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
7206 The value now differs depending on if you build for FIPS or not.
7207 BEWARE! A program linked with a shared FIPSed libcrypto can't be
7208 safely run with a non-FIPSed libcrypto, as it may crash because of
7209 the difference induced by this change.
7212 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
7214 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
7215 (part of SSL_OP_ALL). This option used to disable the
7216 countermeasure against man-in-the-middle protocol-version
7217 rollback in the SSL 2.0 server implementation, which is a bad
7218 idea. (CVE-2005-2969)
7220 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
7221 for Information Security, National Institute of Advanced Industrial
7222 Science and Technology [AIST], Japan)]
7224 *) Minimal support for X9.31 signatures and PSS padding modes. This is
7225 mainly for FIPS compliance and not fully integrated at this stage.
7228 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
7229 the exponentiation using a fixed-length exponent. (Otherwise,
7230 the information leaked through timing could expose the secret key
7231 after many signatures; cf. Bleichenbacher's attack on DSA with
7235 *) Make a new fixed-window mod_exp implementation the default for
7236 RSA, DSA, and DH private-key operations so that the sequence of
7237 squares and multiplies and the memory access pattern are
7238 independent of the particular secret key. This will mitigate
7239 cache-timing and potential related attacks.
7241 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
7242 and this is automatically used by BN_mod_exp_mont() if the new flag
7243 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
7244 will use this BN flag for private exponents unless the flag
7245 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
7246 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
7248 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
7250 *) Change the client implementation for SSLv23_method() and
7251 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
7252 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
7253 (Previously, the SSL 2.0 backwards compatible Client Hello
7254 message format would be used even with SSL_OP_NO_SSLv2.)
7257 *) Add support for smime-type MIME parameter in S/MIME messages which some
7261 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
7262 a threadsafe manner. Modify rsa code to use new function and add calls
7263 to dsa and dh code (which had race conditions before).
7266 *) Include the fixed error library code in the C error file definitions
7267 instead of fixing them up at runtime. This keeps the error code
7268 structures constant.
7271 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
7273 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
7276 *) Fixes for newer kerberos headers. NB: the casts are needed because
7277 the 'length' field is signed on one version and unsigned on another
7278 with no (?) obvious way to tell the difference, without these VC++
7279 complains. Also the "definition" of FAR (blank) is no longer included
7280 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
7281 some needed definitions.
7284 *) Undo Cygwin change.
7287 *) Added support for proxy certificates according to RFC 3820.
7288 Because they may be a security thread to unaware applications,
7289 they must be explicitly allowed in run-time. See
7290 docs/HOWTO/proxy_certificates.txt for further information.
7293 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
7295 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
7296 server and client random values. Previously
7297 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
7298 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
7300 This change has negligible security impact because:
7302 1. Server and client random values still have 24 bytes of pseudo random
7305 2. Server and client random values are sent in the clear in the initial
7308 3. The master secret is derived using the premaster secret (48 bytes in
7309 size for static RSA ciphersuites) as well as client server and random
7312 The OpenSSL team would like to thank the UK NISCC for bringing this issue
7315 [Stephen Henson, reported by UK NISCC]
7317 *) Use Windows randomness collection on Cygwin.
7320 *) Fix hang in EGD/PRNGD query when communication socket is closed
7321 prematurely by EGD/PRNGD.
7322 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
7324 *) Prompt for pass phrases when appropriate for PKCS12 input format.
7327 *) Back-port of selected performance improvements from development
7328 branch, as well as improved support for PowerPC platforms.
7331 *) Add lots of checks for memory allocation failure, error codes to indicate
7332 failure and freeing up memory if a failure occurs.
7333 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
7335 *) Add new -passin argument to dgst.
7338 *) Perform some character comparisons of different types in X509_NAME_cmp:
7339 this is needed for some certificates that re-encode DNs into UTF8Strings
7340 (in violation of RFC3280) and can't or won't issue name rollover
7344 *) Make an explicit check during certificate validation to see that
7345 the CA setting in each certificate on the chain is correct. As a
7346 side effect always do the following basic checks on extensions,
7347 not just when there's an associated purpose to the check:
7349 - if there is an unhandled critical extension (unless the user
7350 has chosen to ignore this fault)
7351 - if the path length has been exceeded (if one is set at all)
7352 - that certain extensions fit the associated purpose (if one has
7356 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
7358 *) Avoid a race condition when CRLs are checked in a multi threaded
7359 environment. This would happen due to the reordering of the revoked
7360 entries during signature checking and serial number lookup. Now the
7361 encoding is cached and the serial number sort performed under a lock.
7362 Add new STACK function sk_is_sorted().
7365 *) Add Delta CRL to the extension code.
7368 *) Various fixes to s3_pkt.c so alerts are sent properly.
7369 [David Holmes <d.holmes@f5.com>]
7371 *) Reduce the chances of duplicate issuer name and serial numbers (in
7372 violation of RFC3280) using the OpenSSL certificate creation utilities.
7373 This is done by creating a random 64 bit value for the initial serial
7374 number when a serial number file is created or when a self signed
7375 certificate is created using 'openssl req -x509'. The initial serial
7376 number file is created using 'openssl x509 -next_serial' in CA.pl
7377 rather than being initialized to 1.
7380 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
7382 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7383 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7384 [Joe Orton, Steve Henson]
7386 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
7388 [Joe Orton, Steve Henson]
7390 *) Make it possible to have multiple active certificates with the same
7391 subject in the CA index file. This is done only if the keyword
7392 'unique_subject' is set to 'no' in the main CA section (default
7393 if 'CA_default') of the configuration file. The value is saved
7394 with the database itself in a separate index attribute file,
7395 named like the index file with '.attr' appended to the name.
7398 *) X509 verify fixes. Disable broken certificate workarounds when
7399 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
7400 keyUsage extension present. Don't accept CRLs with unhandled critical
7401 extensions: since verify currently doesn't process CRL extensions this
7402 rejects a CRL with *any* critical extensions. Add new verify error codes
7406 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
7407 A clarification of RFC2560 will require the use of OCTET STRINGs and
7408 some implementations cannot handle the current raw format. Since OpenSSL
7409 copies and compares OCSP nonces as opaque blobs without any attempt at
7410 parsing them this should not create any compatibility issues.
7413 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
7414 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
7415 this HMAC (and other) operations are several times slower than OpenSSL
7419 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
7420 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7422 *) Use the correct content when signing type "other".
7425 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
7427 *) Fix various bugs revealed by running the NISCC test suite:
7429 Stop out of bounds reads in the ASN1 code when presented with
7430 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7432 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
7434 If verify callback ignores invalid public key errors don't try to check
7435 certificate signature with the NULL public key.
7439 *) New -ignore_err option in ocsp application to stop the server
7440 exiting on the first error in a request.
7443 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7444 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7448 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7449 extra data after the compression methods not only for TLS 1.0
7450 but also for SSL 3.0 (as required by the specification).
7451 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7453 *) Change X509_certificate_type() to mark the key as exported/exportable
7454 when it's 512 *bits* long, not 512 bytes.
7457 *) Change AES_cbc_encrypt() so it outputs exact multiple of
7458 blocks during encryption.
7461 *) Various fixes to base64 BIO and non blocking I/O. On write
7462 flushes were not handled properly if the BIO retried. On read
7463 data was not being buffered properly and had various logic bugs.
7464 This also affects blocking I/O when the data being decoded is a
7468 *) Various S/MIME bugfixes and compatibility changes:
7469 output correct application/pkcs7 MIME type if
7470 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
7471 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
7472 of files as .eml work). Correctly handle very long lines in MIME
7476 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
7478 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7479 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7480 a protocol version number mismatch like a decryption error
7481 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7484 *) Turn on RSA blinding by default in the default implementation
7485 to avoid a timing attack. Applications that don't want it can call
7486 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7487 They would be ill-advised to do so in most cases.
7488 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7490 *) Change RSA blinding code so that it works when the PRNG is not
7491 seeded (in this case, the secret RSA exponent is abused as
7492 an unpredictable seed -- if it is not unpredictable, there
7493 is no point in blinding anyway). Make RSA blinding thread-safe
7494 by remembering the creator's thread ID in rsa->blinding and
7495 having all other threads use local one-time blinding factors
7496 (this requires more computation than sharing rsa->blinding, but
7497 avoids excessive locking; and if an RSA object is not shared
7498 between threads, blinding will still be very fast).
7501 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
7502 ENGINE as defaults for all supported algorithms irrespective of
7503 the 'flags' parameter. 'flags' is now honoured, so applications
7504 should make sure they are passing it correctly.
7507 *) Target "mingw" now allows native Windows code to be generated in
7508 the Cygwin environment as well as with the MinGW compiler.
7511 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7513 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7514 via timing by performing a MAC computation even if incorrect
7515 block cipher padding has been found. This is a countermeasure
7516 against active attacks where the attacker has to distinguish
7517 between bad padding and a MAC verification error. (CVE-2003-0078)
7519 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7520 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7521 Martin Vuagnoux (EPFL, Ilion)]
7523 *) Make the no-err option work as intended. The intention with no-err
7524 is not to have the whole error stack handling routines removed from
7525 libcrypto, it's only intended to remove all the function name and
7526 reason texts, thereby removing some of the footprint that may not
7527 be interesting if those errors aren't displayed anyway.
7529 NOTE: it's still possible for any application or module to have its
7530 own set of error texts inserted. The routines are there, just not
7531 used by default when no-err is given.
7534 *) Add support for FreeBSD on IA64.
7535 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7537 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7538 Kerberos function mit_des_cbc_cksum(). Before this change,
7539 the value returned by DES_cbc_cksum() was like the one from
7540 mit_des_cbc_cksum(), except the bytes were swapped.
7541 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7543 *) Allow an application to disable the automatic SSL chain building.
7544 Before this a rather primitive chain build was always performed in
7545 ssl3_output_cert_chain(): an application had no way to send the
7546 correct chain if the automatic operation produced an incorrect result.
7548 Now the chain builder is disabled if either:
7550 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7552 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7554 The reasoning behind this is that an application would not want the
7555 auto chain building to take place if extra chain certificates are
7556 present and it might also want a means of sending no additional
7557 certificates (for example the chain has two certificates and the
7561 *) Add the possibility to build without the ENGINE framework.
7562 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7564 *) Under Win32 gmtime() can return NULL: check return value in
7565 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7568 *) DSA routines: under certain error conditions uninitialized BN objects
7569 could be freed. Solution: make sure initialization is performed early
7570 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7571 Nils Larsch <nla@trustcenter.de> via PR#459)
7574 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7575 checked on reconnect on the client side, therefore session resumption
7576 could still fail with a "ssl session id is different" error. This
7577 behaviour is masked when SSL_OP_ALL is used due to
7578 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7579 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7580 followup to PR #377.
7583 *) IA-32 assembler support enhancements: unified ELF targets, support
7584 for SCO/Caldera platforms, fix for Cygwin shared build.
7587 *) Add support for FreeBSD on sparc64. As a consequence, support for
7588 FreeBSD on non-x86 processors is separate from x86 processors on
7589 the config script, much like the NetBSD support.
7590 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7592 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7594 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7597 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7598 code (06) was taken as the first octet of the session ID and the last
7599 octet was ignored consequently. As a result SSLv2 client side session
7600 caching could not have worked due to the session ID mismatch between
7602 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7606 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7607 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7611 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7612 seems that in spite of existing for more than a year, many application
7613 author have done nothing to provide the necessary callbacks, which
7614 means that this particular engine will not work properly anywhere.
7615 This is a very unfortunate situation which forces us, in the name
7616 of usability, to give the hw_ncipher.c a static lock, which is part
7618 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7619 appear in 0.9.8 or later. We EXPECT application authors to have
7620 dealt properly with this when 0.9.8 is released (unless we actually
7621 make such changes in the libcrypto locking code that changes will
7622 have to be made anyway).
7625 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7626 octets have been read, EOF or an error occurs. Without this change
7627 some truncated ASN1 structures will not produce an error.
7630 *) Disable Heimdal support, since it hasn't been fully implemented.
7631 Still give the possibility to force the use of Heimdal, but with
7632 warnings and a request that patches get sent to openssl-dev.
7635 *) Add the VC-CE target, introduce the WINCE sysname, and add
7636 INSTALL.WCE and appropriate conditionals to make it build.
7637 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7639 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7640 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7641 edit numbers of the version.
7642 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7644 *) Introduce safe string copy and catenation functions
7645 (BUF_strlcpy() and BUF_strlcat()).
7646 [Ben Laurie (CHATS) and Richard Levitte]
7648 *) Avoid using fixed-size buffers for one-line DNs.
7649 [Ben Laurie (CHATS)]
7651 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7652 resizing buffers containing secrets, and use where appropriate.
7653 [Ben Laurie (CHATS)]
7655 *) Avoid using fixed size buffers for configuration file location.
7656 [Ben Laurie (CHATS)]
7658 *) Avoid filename truncation for various CA files.
7659 [Ben Laurie (CHATS)]
7661 *) Use sizeof in preference to magic numbers.
7662 [Ben Laurie (CHATS)]
7664 *) Avoid filename truncation in cert requests.
7665 [Ben Laurie (CHATS)]
7667 *) Add assertions to check for (supposedly impossible) buffer
7669 [Ben Laurie (CHATS)]
7671 *) Don't cache truncated DNS entries in the local cache (this could
7672 potentially lead to a spoofing attack).
7673 [Ben Laurie (CHATS)]
7675 *) Fix various buffers to be large enough for hex/decimal
7676 representations in a platform independent manner.
7677 [Ben Laurie (CHATS)]
7679 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7680 resizing buffers containing secrets, and use where appropriate.
7681 [Ben Laurie (CHATS)]
7683 *) Add BIO_indent() to avoid much slightly worrying code to do
7685 [Ben Laurie (CHATS)]
7687 *) Convert sprintf()/BIO_puts() to BIO_printf().
7688 [Ben Laurie (CHATS)]
7690 *) buffer_gets() could terminate with the buffer only half
7692 [Ben Laurie (CHATS)]
7694 *) Add assertions to prevent user-supplied crypto functions from
7695 overflowing internal buffers by having large block sizes, etc.
7696 [Ben Laurie (CHATS)]
7698 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7700 [Ben Laurie (CHATS)]
7702 *) Eliminate unused copy of key in RC4.
7703 [Ben Laurie (CHATS)]
7705 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7706 [Ben Laurie (CHATS)]
7708 *) Fix off-by-one error in EGD path.
7709 [Ben Laurie (CHATS)]
7711 *) If RANDFILE path is too long, ignore instead of truncating.
7712 [Ben Laurie (CHATS)]
7714 *) Eliminate unused and incorrectly sized X.509 structure
7716 [Ben Laurie (CHATS)]
7718 *) Eliminate unused and dangerous function knumber().
7719 [Ben Laurie (CHATS)]
7721 *) Eliminate unused and dangerous structure, KSSL_ERR.
7722 [Ben Laurie (CHATS)]
7724 *) Protect against overlong session ID context length in an encoded
7725 session object. Since these are local, this does not appear to be
7727 [Ben Laurie (CHATS)]
7729 *) Change from security patch (see 0.9.6e below) that did not affect
7730 the 0.9.6 release series:
7732 Remote buffer overflow in SSL3 protocol - an attacker could
7733 supply an oversized master key in Kerberos-enabled versions.
7735 [Ben Laurie (CHATS)]
7737 *) Change the SSL kerb5 codes to match RFC 2712.
7740 *) Make -nameopt work fully for req and add -reqopt switch.
7741 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7743 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7744 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7746 *) Make sure tests can be performed even if the corresponding algorithms
7747 have been removed entirely. This was also the last step to make
7748 OpenSSL compilable with DJGPP under all reasonable conditions.
7749 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7751 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7752 to allow version independent disabling of normally unselected ciphers,
7753 which may be activated as a side-effect of selecting a single cipher.
7755 (E.g., cipher list string "RSA" enables ciphersuites that are left
7756 out of "ALL" because they do not provide symmetric encryption.
7757 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7758 [Lutz Jaenicke, Bodo Moeller]
7760 *) Add appropriate support for separate platform-dependent build
7761 directories. The recommended way to make a platform-dependent
7762 build directory is the following (tested on Linux), maybe with
7765 # Place yourself outside of the OpenSSL source tree. In
7766 # this example, the environment variable OPENSSL_SOURCE
7767 # is assumed to contain the absolute OpenSSL source directory.
7768 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7769 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7770 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7771 mkdir -p `dirname $F`
7772 ln -s $OPENSSL_SOURCE/$F $F
7775 To be absolutely sure not to disturb the source tree, a "make clean"
7776 is a good thing. If it isn't successful, don't worry about it,
7777 it probably means the source directory is very clean.
7780 *) Make sure any ENGINE control commands make local copies of string
7781 pointers passed to them whenever necessary. Otherwise it is possible
7782 the caller may have overwritten (or deallocated) the original string
7783 data when a later ENGINE operation tries to use the stored values.
7784 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7786 *) Improve diagnostics in file reading and command-line digests.
7787 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7789 *) Add AES modes CFB and OFB to the object database. Correct an
7790 error in AES-CFB decryption.
7793 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7794 allows existing EVP_CIPHER_CTX structures to be reused after
7795 calling EVP_*Final(). This behaviour is used by encryption
7796 BIOs and some applications. This has the side effect that
7797 applications must explicitly clean up cipher contexts with
7798 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7801 *) Check the values of dna and dnb in bn_mul_recursive before calling
7802 bn_mul_comba (a non zero value means the a or b arrays do not contain
7803 n2 elements) and fallback to bn_mul_normal if either is not zero.
7806 *) Fix escaping of non-ASCII characters when using the -subj option
7807 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7810 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7811 form for "surname", serialNumber has no short form.
7812 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7813 therefore remove "mail" short name for "internet 7".
7814 The OID for unique identifiers in X509 certificates is
7815 x500UniqueIdentifier, not uniqueIdentifier.
7816 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7819 *) Add an "init" command to the ENGINE config module and auto initialize
7820 ENGINEs. Without any "init" command the ENGINE will be initialized
7821 after all ctrl commands have been executed on it. If init=1 the
7822 ENGINE is initialized at that point (ctrls before that point are run
7823 on the uninitialized ENGINE and after on the initialized one). If
7824 init=0 then the ENGINE will not be initialized at all.
7827 *) Fix the 'app_verify_callback' interface so that the user-defined
7828 argument is actually passed to the callback: In the
7829 SSL_CTX_set_cert_verify_callback() prototype, the callback
7830 declaration has been changed from
7833 int (*cb)(X509_STORE_CTX *,void *);
7834 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7835 i=s->ctx->app_verify_callback(&ctx)
7836 has been changed into
7837 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7839 To update applications using SSL_CTX_set_cert_verify_callback(),
7840 a dummy argument can be added to their callback functions.
7841 [D. K. Smetters <smetters@parc.xerox.com>]
7843 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7844 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7846 *) Add and OPENSSL_LOAD_CONF define which will cause
7847 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7848 This allows older applications to transparently support certain
7849 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7850 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7851 load the config file and OPENSSL_add_all_algorithms_conf() which will
7852 always load it have also been added.
7855 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7856 Adjust NIDs and EVP layer.
7857 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7859 *) Config modules support in openssl utility.
7861 Most commands now load modules from the config file,
7862 though in a few (such as version) this isn't done
7863 because it couldn't be used for anything.
7865 In the case of ca and req the config file used is
7866 the same as the utility itself: that is the -config
7867 command line option can be used to specify an
7871 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7872 use "openssl_conf" if filename is NULL use default openssl config file.
7875 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7876 config section name. Add a new flag to tolerate a missing config file
7877 and move code to CONF_modules_load_file().
7880 *) Support for crypto accelerator cards from Accelerated Encryption
7881 Processing, www.aep.ie. (Use engine 'aep')
7882 The support was copied from 0.9.6c [engine] and adapted/corrected
7883 to work with the new engine framework.
7884 [AEP Inc. and Richard Levitte]
7886 *) Support for SureWare crypto accelerator cards from Baltimore
7887 Technologies. (Use engine 'sureware')
7888 The support was copied from 0.9.6c [engine] and adapted
7889 to work with the new engine framework.
7892 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7893 make the newer ENGINE framework commands for the CHIL engine work.
7894 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7896 *) Make it possible to produce shared libraries on ReliantUNIX.
7897 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7899 *) Add the configuration target debug-linux-ppro.
7900 Make 'openssl rsa' use the general key loading routines
7901 implemented in apps.c, and make those routines able to
7902 handle the key format FORMAT_NETSCAPE and the variant
7904 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7906 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7907 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7909 *) Add -keyform to rsautl, and document -engine.
7910 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7912 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7913 BIO_R_NO_SUCH_FILE error code rather than the generic
7914 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7917 *) Add new functions
7919 ERR_peek_last_error_line
7920 ERR_peek_last_error_line_data.
7921 These are similar to
7924 ERR_peek_error_line_data,
7925 but report on the latest error recorded rather than the first one
7926 still in the error queue.
7927 [Ben Laurie, Bodo Moeller]
7929 *) default_algorithms option in ENGINE config module. This allows things
7931 default_algorithms = ALL
7932 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7935 *) Preliminary ENGINE config module.
7938 *) New experimental application configuration code.
7941 *) Change the AES code to follow the same name structure as all other
7942 symmetric ciphers, and behave the same way. Move everything to
7943 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7944 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7946 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7947 [Ben Laurie and Theo de Raadt]
7949 *) Add option to output public keys in req command.
7950 [Massimiliano Pala madwolf@openca.org]
7952 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7953 (up to about 10% better than before for P-192 and P-224).
7956 *) New functions/macros
7958 SSL_CTX_set_msg_callback(ctx, cb)
7959 SSL_CTX_set_msg_callback_arg(ctx, arg)
7960 SSL_set_msg_callback(ssl, cb)
7961 SSL_set_msg_callback_arg(ssl, arg)
7963 to request calling a callback function
7965 void cb(int write_p, int version, int content_type,
7966 const void *buf, size_t len, SSL *ssl, void *arg)
7968 whenever a protocol message has been completely received
7969 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7970 protocol version according to which the SSL library interprets
7971 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7972 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7973 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7974 specification (change_cipher_spec(20), alert(21), handshake(22)).
7975 'buf' and 'len' point to the actual message, 'ssl' to the
7976 SSL object, and 'arg' is the application-defined value set by
7977 SSL[_CTX]_set_msg_callback_arg().
7979 'openssl s_client' and 'openssl s_server' have new '-msg' options
7980 to enable a callback that displays all protocol messages.
7983 *) Change the shared library support so shared libraries are built as
7984 soon as the corresponding static library is finished, and thereby get
7985 openssl and the test programs linked against the shared library.
7986 This still only happens when the keyword "shard" has been given to
7987 the configuration scripts.
7989 NOTE: shared library support is still an experimental thing, and
7990 backward binary compatibility is still not guaranteed.
7991 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7993 *) Add support for Subject Information Access extension.
7994 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7996 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7997 additional bytes when new memory had to be allocated, not just
7998 when reusing an existing buffer.
8001 *) New command line and configuration option 'utf8' for the req command.
8002 This allows field values to be specified as UTF8 strings.
8005 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
8006 runs for the former and machine-readable output for the latter.
8009 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
8010 of the e-mail address in the DN (i.e., it will go into a certificate
8011 extension only). The new configuration file option 'email_in_dn = no'
8012 has the same effect.
8013 [Massimiliano Pala madwolf@openca.org]
8015 *) Change all functions with names starting with des_ to be starting
8016 with DES_ instead. Add wrappers that are compatible with libdes,
8017 but are named _ossl_old_des_*. Finally, add macros that map the
8018 des_* symbols to the corresponding _ossl_old_des_* if libdes
8019 compatibility is desired. If OpenSSL 0.9.6c compatibility is
8020 desired, the des_* symbols will be mapped to DES_*, with one
8023 Since we provide two compatibility mappings, the user needs to
8024 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
8025 compatibility is desired. The default (i.e., when that macro
8026 isn't defined) is OpenSSL 0.9.6c compatibility.
8028 There are also macros that enable and disable the support of old
8029 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
8030 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
8031 are defined, the default will apply: to support the old des routines.
8033 In either case, one must include openssl/des.h to get the correct
8034 definitions. Do not try to just include openssl/des_old.h, that
8037 NOTE: This is a major break of an old API into a new one. Software
8038 authors are encouraged to switch to the DES_ style functions. Some
8039 time in the future, des_old.h and the libdes compatibility functions
8040 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
8041 default), and then completely removed.
8044 *) Test for certificates which contain unsupported critical extensions.
8045 If such a certificate is found during a verify operation it is
8046 rejected by default: this behaviour can be overridden by either
8047 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
8048 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
8049 X509_supported_extension() has also been added which returns 1 if a
8050 particular extension is supported.
8053 *) Modify the behaviour of EVP cipher functions in similar way to digests
8054 to retain compatibility with existing code.
8057 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
8058 compatibility with existing code. In particular the 'ctx' parameter does
8059 not have to be to be initialized before the call to EVP_DigestInit() and
8060 it is tidied up after a call to EVP_DigestFinal(). New function
8061 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
8062 EVP_MD_CTX_copy() changed to not require the destination to be
8063 initialized valid and new function EVP_MD_CTX_copy_ex() added which
8064 requires the destination to be valid.
8066 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
8067 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
8070 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
8071 so that complete 'Handshake' protocol structures are kept in memory
8072 instead of overwriting 'msg_type' and 'length' with 'body' data.
8075 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
8076 [Massimo Santin via Richard Levitte]
8078 *) Major restructuring to the underlying ENGINE code. This includes
8079 reduction of linker bloat, separation of pure "ENGINE" manipulation
8080 (initialisation, etc) from functionality dealing with implementations
8081 of specific crypto interfaces. This change also introduces integrated
8082 support for symmetric ciphers and digest implementations - so ENGINEs
8083 can now accelerate these by providing EVP_CIPHER and EVP_MD
8084 implementations of their own. This is detailed in crypto/engine/README
8085 as it couldn't be adequately described here. However, there are a few
8086 API changes worth noting - some RSA, DSA, DH, and RAND functions that
8087 were changed in the original introduction of ENGINE code have now
8088 reverted back - the hooking from this code to ENGINE is now a good
8089 deal more passive and at run-time, operations deal directly with
8090 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
8091 dereferencing through an ENGINE pointer any more. Also, the ENGINE
8092 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
8093 they were not being used by the framework as there is no concept of a
8094 BIGNUM_METHOD and they could not be generalised to the new
8095 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
8096 ENGINE_cpy() has been removed as it cannot be consistently defined in
8100 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
8103 *) Change mkdef.pl to sort symbols that get the same entry number,
8104 and make sure the automatically generated functions ERR_load_*
8105 become part of libeay.num as well.
8108 *) New function SSL_renegotiate_pending(). This returns true once
8109 renegotiation has been requested (either SSL_renegotiate() call
8110 or HelloRequest/ClientHello received from the peer) and becomes
8111 false once a handshake has been completed.
8112 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
8113 sends a HelloRequest, but does not ensure that a handshake takes
8114 place. SSL_renegotiate_pending() is useful for checking if the
8115 client has followed the request.)
8118 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
8119 By default, clients may request session resumption even during
8120 renegotiation (if session ID contexts permit); with this option,
8121 session resumption is possible only in the first handshake.
8123 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
8124 more bits available for options that should not be part of
8125 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
8128 *) Add some demos for certificate and certificate request creation.
8131 *) Make maximum certificate chain size accepted from the peer application
8132 settable (SSL*_get/set_max_cert_list()), as proposed by
8133 "Douglas E. Engert" <deengert@anl.gov>.
8136 *) Add support for shared libraries for Unixware-7
8137 (Boyd Lynn Gerber <gerberb@zenez.com>).
8140 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
8141 be done prior to destruction. Use this to unload error strings from
8142 ENGINEs that load their own error strings. NB: This adds two new API
8143 functions to "get" and "set" this destroy handler in an ENGINE.
8146 *) Alter all existing ENGINE implementations (except "openssl" and
8147 "openbsd") to dynamically instantiate their own error strings. This
8148 makes them more flexible to be built both as statically-linked ENGINEs
8149 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
8150 Also, add stub code to each that makes building them as self-contained
8151 shared-libraries easier (see README.ENGINE).
8154 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
8155 implementations into applications that are completely implemented in
8156 self-contained shared-libraries. The "dynamic" ENGINE exposes control
8157 commands that can be used to configure what shared-library to load and
8158 to control aspects of the way it is handled. Also, made an update to
8159 the README.ENGINE file that brings its information up-to-date and
8160 provides some information and instructions on the "dynamic" ENGINE
8161 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
8164 *) Make it possible to unload ranges of ERR strings with a new
8165 "ERR_unload_strings" function.
8168 *) Add a copy() function to EVP_MD.
8171 *) Make EVP_MD routines take a context pointer instead of just the
8172 md_data void pointer.
8175 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
8176 that the digest can only process a single chunk of data
8177 (typically because it is provided by a piece of
8178 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
8179 is only going to provide a single chunk of data, and hence the
8180 framework needn't accumulate the data for oneshot drivers.
8183 *) As with "ERR", make it possible to replace the underlying "ex_data"
8184 functions. This change also alters the storage and management of global
8185 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
8186 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
8187 index counters. The API functions that use this state have been changed
8188 to take a "class_index" rather than pointers to the class's local STACK
8189 and counter, and there is now an API function to dynamically create new
8190 classes. This centralisation allows us to (a) plug a lot of the
8191 thread-safety problems that existed, and (b) makes it possible to clean
8192 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
8193 such data would previously have always leaked in application code and
8194 workarounds were in place to make the memory debugging turn a blind eye
8195 to it. Application code that doesn't use this new function will still
8196 leak as before, but their memory debugging output will announce it now
8197 rather than letting it slide.
8199 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
8200 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
8201 has a return value to indicate success or failure.
8204 *) Make it possible to replace the underlying "ERR" functions such that the
8205 global state (2 LHASH tables and 2 locks) is only used by the "default"
8206 implementation. This change also adds two functions to "get" and "set"
8207 the implementation prior to it being automatically set the first time
8208 any other ERR function takes place. Ie. an application can call "get",
8209 pass the return value to a module it has just loaded, and that module
8210 can call its own "set" function using that value. This means the
8211 module's "ERR" operations will use (and modify) the error state in the
8212 application and not in its own statically linked copy of OpenSSL code.
8215 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
8216 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
8217 the operation, and provides a more encapsulated way for external code
8218 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
8219 to use these functions rather than manually incrementing the counts.
8221 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
8224 *) Add EVP test program.
8227 *) Add symmetric cipher support to ENGINE. Expect the API to change!
8230 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
8231 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
8232 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
8233 These allow a CRL to be built without having to access X509_CRL fields
8234 directly. Modify 'ca' application to use new functions.
8237 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
8238 bug workarounds. Rollback attack detection is a security feature.
8239 The problem will only arise on OpenSSL servers when TLSv1 is not
8240 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
8241 Software authors not wanting to support TLSv1 will have special reasons
8242 for their choice and can explicitly enable this option.
8243 [Bodo Moeller, Lutz Jaenicke]
8245 *) Rationalise EVP so it can be extended: don't include a union of
8246 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
8247 (similar to those existing for EVP_CIPHER_CTX).
8252 EVP_MD_CTX_init(&md); /* new function call */
8253 EVP_DigestInit(&md, EVP_sha1());
8254 EVP_DigestUpdate(&md, in, len);
8255 EVP_DigestFinal(&md, out, NULL);
8256 EVP_MD_CTX_cleanup(&md); /* new function call */
8260 *) Make DES key schedule conform to the usual scheme, as well as
8261 correcting its structure. This means that calls to DES functions
8262 now have to pass a pointer to a des_key_schedule instead of a
8263 plain des_key_schedule (which was actually always a pointer
8266 des_key_schedule ks;
8268 des_set_key_checked(..., &ks);
8269 des_ncbc_encrypt(..., &ks, ...);
8271 (Note that a later change renames 'des_...' into 'DES_...'.)
8274 *) Initial reduction of linker bloat: the use of some functions, such as
8275 PEM causes large amounts of unused functions to be linked in due to
8276 poor organisation. For example pem_all.c contains every PEM function
8277 which has a knock on effect of linking in large amounts of (unused)
8278 ASN1 code. Grouping together similar functions and splitting unrelated
8279 functions prevents this.
8282 *) Cleanup of EVP macros.
8285 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
8286 correct _ecb suffix.
8289 *) Add initial OCSP responder support to ocsp application. The
8290 revocation information is handled using the text based index
8291 use by the ca application. The responder can either handle
8292 requests generated internally, supplied in files (for example
8293 via a CGI script) or using an internal minimal server.
8296 *) Add configuration choices to get zlib compression for TLS.
8299 *) Changes to Kerberos SSL for RFC 2712 compliance:
8300 1. Implemented real KerberosWrapper, instead of just using
8301 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
8302 2. Implemented optional authenticator field of KerberosWrapper.
8304 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
8305 and authenticator structs; see crypto/krb5/.
8307 Generalized Kerberos calls to support multiple Kerberos libraries.
8308 [Vern Staats <staatsvr@asc.hpc.mil>,
8309 Jeffrey Altman <jaltman@columbia.edu>
8310 via Richard Levitte]
8312 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
8313 already does with RSA. testdsa.h now has 'priv_key/pub_key'
8314 values for each of the key sizes rather than having just
8315 parameters (and 'speed' generating keys each time).
8318 *) Speed up EVP routines.
8321 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
8322 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
8323 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
8324 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
8326 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
8327 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
8328 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
8331 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
8333 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
8336 *) Added the OS2-EMX target.
8337 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
8339 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
8340 to support NCONF routines in extension code. New function CONF_set_nconf()
8341 to allow functions which take an NCONF to also handle the old LHASH
8342 structure: this means that the old CONF compatible routines can be
8343 retained (in particular wrt extensions) without having to duplicate the
8344 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
8347 *) Enhance the general user interface with mechanisms for inner control
8348 and with possibilities to have yes/no kind of prompts.
8351 *) Change all calls to low level digest routines in the library and
8352 applications to use EVP. Add missing calls to HMAC_cleanup() and
8353 don't assume HMAC_CTX can be copied using memcpy().
8354 [Verdon Walker <VWalker@novell.com>, Steve Henson]
8356 *) Add the possibility to control engines through control names but with
8357 arbitrary arguments instead of just a string.
8358 Change the key loaders to take a UI_METHOD instead of a callback
8359 function pointer. NOTE: this breaks binary compatibility with earlier
8360 versions of OpenSSL [engine].
8361 Adapt the nCipher code for these new conditions and add a card insertion
8365 *) Enhance the general user interface with mechanisms to better support
8366 dialog box interfaces, application-defined prompts, the possibility
8367 to use defaults (for example default passwords from somewhere else)
8368 and interrupts/cancellations.
8371 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
8372 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
8375 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
8376 tidy up some unnecessarily weird code in 'sk_new()').
8377 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
8379 *) Change the key loading routines for ENGINEs to use the same kind
8380 callback (pem_password_cb) as all other routines that need this
8384 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
8385 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
8386 than this minimum value is recommended.
8389 *) New random seeder for OpenVMS, using the system process statistics
8390 that are easily reachable.
8393 *) Windows apparently can't transparently handle global
8394 variables defined in DLLs. Initialisations such as:
8396 const ASN1_ITEM *it = &ASN1_INTEGER_it;
8398 won't compile. This is used by the any applications that need to
8399 declare their own ASN1 modules. This was fixed by adding the option
8400 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
8401 needed for static libraries under Win32.
8404 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
8405 setting of purpose and trust fields. New X509_STORE trust and
8406 purpose functions and tidy up setting in other SSL functions.
8409 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
8410 structure. These are inherited by X509_STORE_CTX when it is
8411 initialised. This allows various defaults to be set in the
8412 X509_STORE structure (such as flags for CRL checking and custom
8413 purpose or trust settings) for functions which only use X509_STORE_CTX
8414 internally such as S/MIME.
8416 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
8417 trust settings if they are not set in X509_STORE. This allows X509_STORE
8418 purposes and trust (in S/MIME for example) to override any set by default.
8420 Add command line options for CRL checking to smime, s_client and s_server
8424 *) Initial CRL based revocation checking. If the CRL checking flag(s)
8425 are set then the CRL is looked up in the X509_STORE structure and
8426 its validity and signature checked, then if the certificate is found
8427 in the CRL the verify fails with a revoked error.
8429 Various new CRL related callbacks added to X509_STORE_CTX structure.
8431 Command line options added to 'verify' application to support this.
8433 This needs some additional work, such as being able to handle multiple
8434 CRLs with different times, extension based lookup (rather than just
8435 by subject name) and ultimately more complete V2 CRL extension
8439 *) Add a general user interface API (crypto/ui/). This is designed
8440 to replace things like des_read_password and friends (backward
8441 compatibility functions using this new API are provided).
8442 The purpose is to remove prompting functions from the DES code
8443 section as well as provide for prompting through dialog boxes in
8444 a window system and the like.
8447 *) Add "ex_data" support to ENGINE so implementations can add state at a
8448 per-structure level rather than having to store it globally.
8451 *) Make it possible for ENGINE structures to be copied when retrieved by
8452 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
8453 This causes the "original" ENGINE structure to act like a template,
8454 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
8455 operational state can be localised to each ENGINE structure, despite the
8456 fact they all share the same "methods". New ENGINE structures returned in
8457 this case have no functional references and the return value is the single
8458 structural reference. This matches the single structural reference returned
8459 by ENGINE_by_id() normally, when it is incremented on the pre-existing
8463 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
8464 needs to match any other type at all we need to manually clear the
8468 *) Changes to the "openssl engine" utility to include;
8469 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
8470 about an ENGINE's available control commands.
8471 - executing control commands from command line arguments using the
8472 '-pre' and '-post' switches. '-post' is only used if '-t' is
8473 specified and the ENGINE is successfully initialised. The syntax for
8474 the individual commands are colon-separated, for example;
8475 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
8478 *) New dynamic control command support for ENGINEs. ENGINEs can now
8479 declare their own commands (numbers), names (strings), descriptions,
8480 and input types for run-time discovery by calling applications. A
8481 subset of these commands are implicitly classed as "executable"
8482 depending on their input type, and only these can be invoked through
8483 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
8484 can be based on user input, config files, etc). The distinction is
8485 that "executable" commands cannot return anything other than a boolean
8486 result and can only support numeric or string input, whereas some
8487 discoverable commands may only be for direct use through
8488 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
8489 pointers, or other custom uses. The "executable" commands are to
8490 support parameterisations of ENGINE behaviour that can be
8491 unambiguously defined by ENGINEs and used consistently across any
8492 OpenSSL-based application. Commands have been added to all the
8493 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
8494 control over shared-library paths without source code alterations.
8497 *) Changed all ENGINE implementations to dynamically allocate their
8498 ENGINEs rather than declaring them statically. Apart from this being
8499 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
8500 this also allows the implementations to compile without using the
8501 internal engine_int.h header.
8504 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8505 'const' value. Any code that should be able to modify a RAND_METHOD
8506 should already have non-const pointers to it (ie. they should only
8507 modify their own ones).
8510 *) Made a variety of little tweaks to the ENGINE code.
8511 - "atalla" and "ubsec" string definitions were moved from header files
8512 to C code. "nuron" string definitions were placed in variables
8513 rather than hard-coded - allowing parameterisation of these values
8514 later on via ctrl() commands.
8515 - Removed unused "#if 0"'d code.
8516 - Fixed engine list iteration code so it uses ENGINE_free() to release
8517 structural references.
8518 - Constified the RAND_METHOD element of ENGINE structures.
8519 - Constified various get/set functions as appropriate and added
8520 missing functions (including a catch-all ENGINE_cpy that duplicates
8521 all ENGINE values onto a new ENGINE except reference counts/state).
8522 - Removed NULL parameter checks in get/set functions. Setting a method
8523 or function to NULL is a way of cancelling out a previously set
8524 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8525 and doesn't justify the extra error symbols and code.
8526 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8527 flags from engine_int.h to engine.h.
8528 - Changed prototypes for ENGINE handler functions (init(), finish(),
8529 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8532 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8533 to the algorithm using long division. The binary algorithm can be
8534 used only if the modulus is odd. On 32-bit systems, it is faster
8535 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8536 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8537 up to 450 bits. In 64-bit environments, the binary algorithm
8538 appears to be advantageous for much longer moduli; here we use it
8539 for moduli up to 2048 bits.
8542 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8543 could not support the combine flag in choice fields.
8546 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8547 extensions from a certificate request to the certificate.
8550 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8551 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8552 file: this allows the display of the certificate about to be
8553 signed to be customised, to allow certain fields to be included
8554 or excluded and extension details. The old system didn't display
8555 multicharacter strings properly, omitted fields not in the policy
8556 and couldn't display additional details such as extensions.
8559 *) Function EC_POINTs_mul for multiple scalar multiplication
8560 of an arbitrary number of elliptic curve points
8561 \sum scalars[i]*points[i],
8562 optionally including the generator defined for the EC_GROUP:
8563 scalar*generator + \sum scalars[i]*points[i].
8565 EC_POINT_mul is a simple wrapper function for the typical case
8566 that the point list has just one item (besides the optional
8570 *) First EC_METHODs for curves over GF(p):
8572 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8573 operations and provides various method functions that can also
8574 operate with faster implementations of modular arithmetic.
8576 EC_GFp_mont_method() reuses most functions that are part of
8577 EC_GFp_simple_method, but uses Montgomery arithmetic.
8579 [Bodo Moeller; point addition and point doubling
8580 implementation directly derived from source code provided by
8581 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8583 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8584 crypto/ec/ec_lib.c):
8586 Curves are EC_GROUP objects (with an optional group generator)
8587 based on EC_METHODs that are built into the library.
8589 Points are EC_POINT objects based on EC_GROUP objects.
8591 Most of the framework would be able to handle curves over arbitrary
8592 finite fields, but as there are no obvious types for fields other
8593 than GF(p), some functions are limited to that for now.
8596 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8597 that the file contains a complete HTTP response.
8600 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8601 change the def and num file printf format specifier from "%-40sXXX"
8602 to "%-39s XXX". The latter will always guarantee a space after the
8603 field while the former will cause them to run together if the field
8604 is 40 of more characters long.
8607 *) Constify the cipher and digest 'method' functions and structures
8608 and modify related functions to take constant EVP_MD and EVP_CIPHER
8612 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8613 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8616 *) Modify EVP_Digest*() routines so they now return values. Although the
8617 internal software routines can never fail additional hardware versions
8621 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8623 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8624 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8627 ERR_R_NESTED_ASN1_ERROR
8629 ERR_R_MISSING_ASN1_EOS
8630 were 4 .. 9, conflicting with
8631 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8633 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8634 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8636 Add new error code 'ERR_R_INTERNAL_ERROR'.
8639 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8643 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8644 sets the subject name for a new request or supersedes the
8645 subject name in a given request. Formats that can be parsed are
8646 'CN=Some Name, OU=myOU, C=IT'
8648 'CN=Some Name/OU=myOU/C=IT'.
8650 Add options '-batch' and '-verbose' to 'openssl req'.
8651 [Massimiliano Pala <madwolf@hackmasters.net>]
8653 *) Introduce the possibility to access global variables through
8654 functions on platform were that's the best way to handle exporting
8655 global variables in shared libraries. To enable this functionality,
8656 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8657 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8658 is normally done by Configure or something similar).
8660 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8661 in the source file (foo.c) like this:
8663 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8664 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8666 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8667 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8669 OPENSSL_DECLARE_GLOBAL(int,foo);
8670 #define foo OPENSSL_GLOBAL_REF(foo)
8671 OPENSSL_DECLARE_GLOBAL(double,bar);
8672 #define bar OPENSSL_GLOBAL_REF(bar)
8674 The #defines are very important, and therefore so is including the
8675 header file everywhere where the defined globals are used.
8677 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8678 of ASN.1 items, but that structure is a bit different.
8680 The largest change is in util/mkdef.pl which has been enhanced with
8681 better and easier to understand logic to choose which symbols should
8682 go into the Windows .def files as well as a number of fixes and code
8683 cleanup (among others, algorithm keywords are now sorted
8684 lexicographically to avoid constant rewrites).
8687 *) In BN_div() keep a copy of the sign of 'num' before writing the
8688 result to 'rm' because if rm==num the value will be overwritten
8689 and produce the wrong result if 'num' is negative: this caused
8690 problems with BN_mod() and BN_nnmod().
8693 *) Function OCSP_request_verify(). This checks the signature on an
8694 OCSP request and verifies the signer certificate. The signer
8695 certificate is just checked for a generic purpose and OCSP request
8699 *) Add OCSP_check_validity() function to check the validity of OCSP
8700 responses. OCSP responses are prepared in real time and may only
8701 be a few seconds old. Simply checking that the current time lies
8702 between thisUpdate and nextUpdate max reject otherwise valid responses
8703 caused by either OCSP responder or client clock inaccuracy. Instead
8704 we allow thisUpdate and nextUpdate to fall within a certain period of
8705 the current time. The age of the response can also optionally be
8706 checked. Two new options -validity_period and -status_age added to
8710 *) If signature or public key algorithm is unrecognized print out its
8711 OID rather that just UNKNOWN.
8714 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8715 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8716 ID to be generated from the issuer certificate alone which can then be
8717 passed to OCSP_id_issuer_cmp().
8720 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8721 ASN1 modules to export functions returning ASN1_ITEM pointers
8722 instead of the ASN1_ITEM structures themselves. This adds several
8723 new macros which allow the underlying ASN1 function/structure to
8724 be accessed transparently. As a result code should not use ASN1_ITEM
8725 references directly (such as &X509_it) but instead use the relevant
8726 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8727 use of the new ASN1 code on platforms where exporting structures
8728 is problematical (for example in shared libraries) but exporting
8729 functions returning pointers to structures is not.
8732 *) Add support for overriding the generation of SSL/TLS session IDs.
8733 These callbacks can be registered either in an SSL_CTX or per SSL.
8734 The purpose of this is to allow applications to control, if they wish,
8735 the arbitrary values chosen for use as session IDs, particularly as it
8736 can be useful for session caching in multiple-server environments. A
8737 command-line switch for testing this (and any client code that wishes
8738 to use such a feature) has been added to "s_server".
8739 [Geoff Thorpe, Lutz Jaenicke]
8741 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8742 of the form '#if defined(...) || defined(...) || ...' and
8743 '#if !defined(...) && !defined(...) && ...'. This also avoids
8744 the growing number of special cases it was previously handling.
8747 *) Make all configuration macros available for application by making
8748 sure they are available in opensslconf.h, by giving them names starting
8749 with "OPENSSL_" to avoid conflicts with other packages and by making
8750 sure e_os2.h will cover all platform-specific cases together with
8752 Additionally, it is now possible to define configuration/platform-
8753 specific names (called "system identities"). In the C code, these
8754 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8755 macro with the name beginning with "OPENSSL_SYS_", which is determined
8756 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8760 *) New option -set_serial to 'req' and 'x509' this allows the serial
8761 number to use to be specified on the command line. Previously self
8762 signed certificates were hard coded with serial number 0 and the
8763 CA options of 'x509' had to use a serial number in a file which was
8767 *) New options to 'ca' utility to support V2 CRL entry extensions.
8768 Currently CRL reason, invalidity date and hold instruction are
8769 supported. Add new CRL extensions to V3 code and some new objects.
8772 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8773 disable standard block padding (aka PKCS#5 padding) in the EVP
8774 API, which was previously mandatory. This means that the data is
8775 not padded in any way and so the total length much be a multiple
8776 of the block size, otherwise an error occurs.
8779 *) Initial (incomplete) OCSP SSL support.
8782 *) New function OCSP_parse_url(). This splits up a URL into its host,
8783 port and path components: primarily to parse OCSP URLs. New -url
8784 option to ocsp utility.
8787 *) New nonce behavior. The return value of OCSP_check_nonce() now
8788 reflects the various checks performed. Applications can decide
8789 whether to tolerate certain situations such as an absent nonce
8790 in a response when one was present in a request: the ocsp application
8791 just prints out a warning. New function OCSP_add1_basic_nonce()
8792 this is to allow responders to include a nonce in a response even if
8793 the request is nonce-less.
8796 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8797 skipped when using openssl x509 multiple times on a single input file,
8798 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8801 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8802 set string type: to handle setting ASN1_TIME structures. Fix ca
8803 utility to correctly initialize revocation date of CRLs.
8806 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8807 the clients preferred ciphersuites and rather use its own preferences.
8808 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8809 Internet Explorer by ensuring unchanged hash method during stepup.
8810 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8813 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8814 to aes and add a new 'exist' option to print out symbols that don't
8818 *) Additional options to ocsp utility to allow flags to be set and
8819 additional certificates supplied.
8822 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8823 OCSP client a number of certificate to only verify the response
8827 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8828 handle the new API. Currently only ECB, CBC modes supported. Add new
8831 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8832 Encryption Standard (AES) Ciphersuites for Transport Layer
8833 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8834 not enabled by default and were not part of the "ALL" ciphersuite
8835 alias because they were not yet official; they could be
8836 explicitly requested by specifying the "AESdraft" ciphersuite
8837 group alias. In the final release of OpenSSL 0.9.7, the group
8838 alias is called "AES" and is part of "ALL".)
8839 [Ben Laurie, Steve Henson, Bodo Moeller]
8841 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8842 request to response.
8845 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8846 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8847 extract information from a certificate request. OCSP_response_create()
8848 creates a response and optionally adds a basic response structure.
8849 OCSP_basic_add1_status() adds a complete single response to a basic
8850 response and returns the OCSP_SINGLERESP structure just added (to allow
8851 extensions to be included for example). OCSP_basic_add1_cert() adds a
8852 certificate to a basic response and OCSP_basic_sign() signs a basic
8853 response with various flags. New helper functions ASN1_TIME_check()
8854 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8855 (converts ASN1_TIME to GeneralizedTime).
8858 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8859 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8860 structure from a certificate. X509_pubkey_digest() digests the public_key
8861 contents: this is used in various key identifiers.
8864 *) Make sk_sort() tolerate a NULL argument.
8865 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8867 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8868 passed by the function are trusted implicitly. If any of them signed the
8869 response then it is assumed to be valid and is not verified.
8872 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8873 to data. This was previously part of the PKCS7 ASN1 code. This
8874 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8875 [Steve Henson, reported by Kenneth R. Robinette
8876 <support@securenetterm.com>]
8878 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8879 routines: without these tracing memory leaks is very painful.
8880 Fix leaks in PKCS12 and PKCS7 routines.
8883 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8884 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8885 effectively meant GeneralizedTime would never be used. Now it
8886 is initialised to -1 but X509_time_adj() now has to check the value
8887 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8888 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8889 [Steve Henson, reported by Kenneth R. Robinette
8890 <support@securenetterm.com>]
8892 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8893 result in a zero length in the ASN1_INTEGER structure which was
8894 not consistent with the structure when d2i_ASN1_INTEGER() was used
8895 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8896 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8897 where it did not print out a minus for negative ASN1_INTEGER.
8900 *) Add summary printout to ocsp utility. The various functions which
8901 convert status values to strings have been renamed to:
8902 OCSP_response_status_str(), OCSP_cert_status_str() and
8903 OCSP_crl_reason_str() and are no longer static. New options
8904 to verify nonce values and to disable verification. OCSP response
8905 printout format cleaned up.
8908 *) Add additional OCSP certificate checks. These are those specified
8909 in RFC2560. This consists of two separate checks: the CA of the
8910 certificate being checked must either be the OCSP signer certificate
8911 or the issuer of the OCSP signer certificate. In the latter case the
8912 OCSP signer certificate must contain the OCSP signing extended key
8913 usage. This check is performed by attempting to match the OCSP
8914 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8915 in the OCSP_CERTID structures of the response.
8918 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8919 and related routines. This uses the standard OpenSSL certificate
8920 verify routines to perform initial checks (just CA validity) and
8921 to obtain the certificate chain. Then additional checks will be
8922 performed on the chain. Currently the root CA is checked to see
8923 if it is explicitly trusted for OCSP signing. This is used to set
8924 a root CA as a global signing root: that is any certificate that
8925 chains to that CA is an acceptable OCSP signing certificate.
8928 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8929 extensions from a separate configuration file.
8930 As when reading extensions from the main configuration file,
8931 the '-extensions ...' option may be used for specifying the
8933 [Massimiliano Pala <madwolf@comune.modena.it>]
8935 *) New OCSP utility. Allows OCSP requests to be generated or
8936 read. The request can be sent to a responder and the output
8937 parsed, outputted or printed in text form. Not complete yet:
8938 still needs to check the OCSP response validity.
8941 *) New subcommands for 'openssl ca':
8942 'openssl ca -status <serial>' prints the status of the cert with
8943 the given serial number (according to the index file).
8944 'openssl ca -updatedb' updates the expiry status of certificates
8946 [Massimiliano Pala <madwolf@comune.modena.it>]
8948 *) New '-newreq-nodes' command option to CA.pl. This is like
8949 '-newreq', but calls 'openssl req' with the '-nodes' option
8950 so that the resulting key is not encrypted.
8951 [Damien Miller <djm@mindrot.org>]
8953 *) New configuration for the GNU Hurd.
8954 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8956 *) Initial code to implement OCSP basic response verify. This
8957 is currently incomplete. Currently just finds the signer's
8958 certificate and verifies the signature on the response.
8961 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8962 value of OPENSSLDIR. This is available via the new '-d' option
8963 to 'openssl version', and is also included in 'openssl version -a'.
8966 *) Allowing defining memory allocation callbacks that will be given
8967 file name and line number information in additional arguments
8968 (a const char* and an int). The basic functionality remains, as
8969 well as the original possibility to just replace malloc(),
8970 realloc() and free() by functions that do not know about these
8971 additional arguments. To register and find out the current
8972 settings for extended allocation functions, the following
8973 functions are provided:
8975 CRYPTO_set_mem_ex_functions
8976 CRYPTO_set_locked_mem_ex_functions
8977 CRYPTO_get_mem_ex_functions
8978 CRYPTO_get_locked_mem_ex_functions
8980 These work the same way as CRYPTO_set_mem_functions and friends.
8981 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8982 extended allocation function is enabled.
8983 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8984 a conventional allocation function is enabled.
8985 [Richard Levitte, Bodo Moeller]
8987 *) Finish off removing the remaining LHASH function pointer casts.
8988 There should no longer be any prototype-casting required when using
8989 the LHASH abstraction, and any casts that remain are "bugs". See
8990 the callback types and macros at the head of lhash.h for details
8991 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8994 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8995 If /dev/[u]random devices are not available or do not return enough
8996 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8998 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8999 /etc/entropy will be queried once each in this sequence, querying stops
9000 when enough entropy was collected without querying more sockets.
9003 *) Change the Unix RAND_poll() variant to be able to poll several
9004 random devices, as specified by DEVRANDOM, until a sufficient amount
9005 of data has been collected. We spend at most 10 ms on each file
9006 (select timeout) and read in non-blocking mode. DEVRANDOM now
9007 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
9008 (previously it was just the string "/dev/urandom"), so on typical
9009 platforms the 10 ms delay will never occur.
9010 Also separate out the Unix variant to its own file, rand_unix.c.
9011 For VMS, there's a currently-empty rand_vms.c.
9014 *) Move OCSP client related routines to ocsp_cl.c. These
9015 provide utility functions which an application needing
9016 to issue a request to an OCSP responder and analyse the
9017 response will typically need: as opposed to those which an
9018 OCSP responder itself would need which will be added later.
9020 OCSP_request_sign() signs an OCSP request with an API similar
9021 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
9022 response. OCSP_response_get1_basic() extracts basic response
9023 from response. OCSP_resp_find_status(): finds and extracts status
9024 information from an OCSP_CERTID structure (which will be created
9025 when the request structure is built). These are built from lower
9026 level functions which work on OCSP_SINGLERESP structures but
9027 won't normally be used unless the application wishes to examine
9028 extensions in the OCSP response for example.
9030 Replace nonce routines with a pair of functions.
9031 OCSP_request_add1_nonce() adds a nonce value and optionally
9032 generates a random value. OCSP_check_nonce() checks the
9033 validity of the nonce in an OCSP response.
9036 *) Change function OCSP_request_add() to OCSP_request_add0_id().
9037 This doesn't copy the supplied OCSP_CERTID and avoids the
9038 need to free up the newly created id. Change return type
9039 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
9040 This can then be used to add extensions to the request.
9041 Deleted OCSP_request_new(), since most of its functionality
9042 is now in OCSP_REQUEST_new() (and the case insensitive name
9043 clash) apart from the ability to set the request name which
9044 will be added elsewhere.
9047 *) Update OCSP API. Remove obsolete extensions argument from
9048 various functions. Extensions are now handled using the new
9049 OCSP extension code. New simple OCSP HTTP function which
9050 can be used to send requests and parse the response.
9053 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
9054 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
9055 uses the special reorder version of SET OF to sort the attributes
9056 and reorder them to match the encoded order. This resolves a long
9057 standing problem: a verify on a PKCS7 structure just after signing
9058 it used to fail because the attribute order did not match the
9059 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
9060 it uses the received order. This is necessary to tolerate some broken
9061 software that does not order SET OF. This is handled by encoding
9062 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
9063 to produce the required SET OF.
9066 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
9067 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
9068 files to get correct declarations of the ASN.1 item variables.
9071 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
9072 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
9073 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
9074 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
9075 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
9076 ASN1_ITEM and no wrapper functions.
9079 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
9080 replace the old function pointer based I/O routines. Change most of
9081 the *_d2i_bio() and *_d2i_fp() functions to use these.
9084 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
9085 lines, recognize more "algorithms" that can be deselected, and make
9086 it complain about algorithm deselection that isn't recognised.
9089 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
9090 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
9091 to use new functions. Add NO_ASN1_OLD which can be set to remove
9092 some old style ASN1 functions: this can be used to determine if old
9093 code will still work when these eventually go away.
9096 *) New extension functions for OCSP structures, these follow the
9097 same conventions as certificates and CRLs.
9100 *) New function X509V3_add1_i2d(). This automatically encodes and
9101 adds an extension. Its behaviour can be customised with various
9102 flags to append, replace or delete. Various wrappers added for
9103 certificates and CRLs.
9106 *) Fix to avoid calling the underlying ASN1 print routine when
9107 an extension cannot be parsed. Correct a typo in the
9108 OCSP_SERVICELOC extension. Tidy up print OCSP format.
9111 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
9112 entries for variables.
9115 *) Add functionality to apps/openssl.c for detecting locking
9116 problems: As the program is single-threaded, all we have
9117 to do is register a locking callback using an array for
9118 storing which locks are currently held by the program.
9121 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
9122 SSL_get_ex_data_X509_STORE_idx(), which is used in
9123 ssl_verify_cert_chain() and thus can be called at any time
9124 during TLS/SSL handshakes so that thread-safety is essential.
9125 Unfortunately, the ex_data design is not at all suited
9126 for multi-threaded use, so it probably should be abolished.
9129 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
9130 [Broadcom, tweaked and integrated by Geoff Thorpe]
9132 *) Move common extension printing code to new function
9133 X509V3_print_extensions(). Reorganise OCSP print routines and
9134 implement some needed OCSP ASN1 functions. Add OCSP extensions.
9137 *) New function X509_signature_print() to remove duplication in some
9141 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
9142 set (this was treated exactly the same as SET OF previously). This
9143 is used to reorder the STACK representing the structure to match the
9144 encoding. This will be used to get round a problem where a PKCS7
9145 structure which was signed could not be verified because the STACK
9146 order did not reflect the encoded order.
9149 *) Reimplement the OCSP ASN1 module using the new code.
9152 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
9153 for its ASN1 operations. The old style function pointers still exist
9154 for now but they will eventually go away.
9157 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
9158 completely replaces the old ASN1 functionality with a table driven
9159 encoder and decoder which interprets an ASN1_ITEM structure describing
9160 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
9161 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
9162 has also been converted to the new form.
9165 *) Change BN_mod_exp_recp so that negative moduli are tolerated
9166 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
9167 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
9168 for negative moduli.
9171 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
9172 of not touching the result's sign bit.
9175 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
9179 *) Changed the LHASH code to use prototypes for callbacks, and created
9180 macros to declare and implement thin (optionally static) functions
9181 that provide type-safety and avoid function pointer casting for the
9182 type-specific callbacks.
9185 *) Added Kerberos Cipher Suites to be used with TLS, as written in
9187 [Veers Staats <staatsvr@asc.hpc.mil>,
9188 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
9190 *) Reformat the FAQ so the different questions and answers can be divided
9191 in sections depending on the subject.
9194 *) Have the zlib compression code load ZLIB.DLL dynamically under
9198 *) New function BN_mod_sqrt for computing square roots modulo a prime
9199 (using the probabilistic Tonelli-Shanks algorithm unless
9200 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
9201 be handled deterministically).
9202 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9204 *) Make BN_mod_inverse faster by explicitly handling small quotients
9205 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
9206 512 bits], about 30% for larger ones [1024 or 2048 bits].)
9209 *) New function BN_kronecker.
9212 *) Fix BN_gcd so that it works on negative inputs; the result is
9213 positive unless both parameters are zero.
9214 Previously something reasonably close to an infinite loop was
9215 possible because numbers could be growing instead of shrinking
9216 in the implementation of Euclid's algorithm.
9219 *) Fix BN_is_word() and BN_is_one() macros to take into account the
9220 sign of the number in question.
9222 Fix BN_is_word(a,w) to work correctly for w == 0.
9224 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
9225 because its test if the absolute value of 'a' equals 'w'.
9226 Note that BN_abs_is_word does *not* handle w == 0 reliably;
9227 it exists mostly for use in the implementations of BN_is_zero(),
9228 BN_is_one(), and BN_is_word().
9231 *) New function BN_swap.
9234 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
9235 the exponentiation functions are more likely to produce reasonable
9236 results on negative inputs.
9239 *) Change BN_mod_mul so that the result is always non-negative.
9240 Previously, it could be negative if one of the factors was negative;
9241 I don't think anyone really wanted that behaviour.
9244 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
9245 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
9246 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
9247 and add new functions:
9256 BN_mod_lshift1_quick
9260 These functions always generate non-negative results.
9262 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
9263 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
9265 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
9266 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
9267 be reduced modulo m.
9268 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9271 The following entry accidentally appeared in the CHANGES file
9272 distributed with OpenSSL 0.9.7. The modifications described in
9273 it do *not* apply to OpenSSL 0.9.7.
9275 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
9276 was actually never needed) and in BN_mul(). The removal in BN_mul()
9277 required a small change in bn_mul_part_recursive() and the addition
9278 of the functions bn_cmp_part_words(), bn_sub_part_words() and
9279 bn_add_part_words(), which do the same thing as bn_cmp_words(),
9280 bn_sub_words() and bn_add_words() except they take arrays with
9285 *) In 'openssl passwd', verify passwords read from the terminal
9286 unless the '-salt' option is used (which usually means that
9287 verification would just waste user's time since the resulting
9288 hash is going to be compared with some given password hash)
9289 or the new '-noverify' option is used.
9291 This is an incompatible change, but it does not affect
9292 non-interactive use of 'openssl passwd' (passwords on the command
9293 line, '-stdin' option, '-in ...' option) and thus should not
9297 *) Remove all references to RSAref, since there's no more need for it.
9300 *) Make DSO load along a path given through an environment variable
9301 (SHLIB_PATH) with shl_load().
9304 *) Constify the ENGINE code as a result of BIGNUM constification.
9305 Also constify the RSA code and most things related to it. In a
9306 few places, most notable in the depth of the ASN.1 code, ugly
9307 casts back to non-const were required (to be solved at a later
9311 *) Make it so the openssl application has all engines loaded by default.
9314 *) Constify the BIGNUM routines a little more.
9317 *) Add the following functions:
9319 ENGINE_load_cswift()
9321 ENGINE_load_atalla()
9323 ENGINE_load_builtin_engines()
9325 That way, an application can itself choose if external engines that
9326 are built-in in OpenSSL shall ever be used or not. The benefit is
9327 that applications won't have to be linked with libdl or other dso
9328 libraries unless it's really needed.
9330 Changed 'openssl engine' to load all engines on demand.
9331 Changed the engine header files to avoid the duplication of some
9332 declarations (they differed!).
9335 *) 'openssl engine' can now list capabilities.
9338 *) Better error reporting in 'openssl engine'.
9341 *) Never call load_dh_param(NULL) in s_server.
9344 *) Add engine application. It can currently list engines by name and
9345 identity, and test if they are actually available.
9348 *) Improve RPM specification file by forcing symbolic linking and making
9349 sure the installed documentation is also owned by root.root.
9350 [Damien Miller <djm@mindrot.org>]
9352 *) Give the OpenSSL applications more possibilities to make use of
9353 keys (public as well as private) handled by engines.
9356 *) Add OCSP code that comes from CertCo.
9359 *) Add VMS support for the Rijndael code.
9362 *) Added untested support for Nuron crypto accelerator.
9365 *) Add support for external cryptographic devices. This code was
9366 previously distributed separately as the "engine" branch.
9367 [Geoff Thorpe, Richard Levitte]
9369 *) Rework the filename-translation in the DSO code. It is now possible to
9370 have far greater control over how a "name" is turned into a filename
9371 depending on the operating environment and any oddities about the
9372 different shared library filenames on each system.
9375 *) Support threads on FreeBSD-elf in Configure.
9378 *) Fix for SHA1 assembly problem with MASM: it produces
9379 warnings about corrupt line number information when assembling
9380 with debugging information. This is caused by the overlapping
9382 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
9385 NCONF_get_number() has no error checking at all. As a replacement,
9386 NCONF_get_number_e() is defined (_e for "error checking") and is
9387 promoted strongly. The old NCONF_get_number is kept around for
9388 binary backward compatibility.
9389 Make it possible for methods to load from something other than a BIO,
9390 by providing a function pointer that is given a name instead of a BIO.
9391 For example, this could be used to load configuration data from an
9395 *) Fix for non blocking accept BIOs. Added new I/O special reason
9396 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
9397 with non blocking I/O was not possible because no retry code was
9398 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
9402 *) Added the beginnings of Rijndael support.
9405 *) Fix for bug in DirectoryString mask setting. Add support for
9406 X509_NAME_print_ex() in 'req' and X509_print_ex() function
9407 to allow certificate printing to more controllable, additional
9408 'certopt' option to 'x509' to allow new printing options to be
9412 *) Clean old EAY MD5 hack from e_os.h.
9415 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
9417 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
9418 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
9419 [Joe Orton, Steve Henson]
9421 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
9423 *) Fix additional bug revealed by the NISCC test suite:
9425 Stop bug triggering large recursion when presented with
9426 certain ASN.1 tags (CVE-2003-0851)
9429 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
9431 *) Fix various bugs revealed by running the NISCC test suite:
9433 Stop out of bounds reads in the ASN1 code when presented with
9434 invalid tags (CVE-2003-0543 and CVE-2003-0544).
9436 If verify callback ignores invalid public key errors don't try to check
9437 certificate signature with the NULL public key.
9441 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
9442 if the server requested one: as stated in TLS 1.0 and SSL 3.0
9446 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
9447 extra data after the compression methods not only for TLS 1.0
9448 but also for SSL 3.0 (as required by the specification).
9449 [Bodo Moeller; problem pointed out by Matthias Loepfe]
9451 *) Change X509_certificate_type() to mark the key as exported/exportable
9452 when it's 512 *bits* long, not 512 bytes.
9455 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
9457 *) Countermeasure against the Klima-Pokorny-Rosa extension of
9458 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
9459 a protocol version number mismatch like a decryption error
9460 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
9463 *) Turn on RSA blinding by default in the default implementation
9464 to avoid a timing attack. Applications that don't want it can call
9465 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
9466 They would be ill-advised to do so in most cases.
9467 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
9469 *) Change RSA blinding code so that it works when the PRNG is not
9470 seeded (in this case, the secret RSA exponent is abused as
9471 an unpredictable seed -- if it is not unpredictable, there
9472 is no point in blinding anyway). Make RSA blinding thread-safe
9473 by remembering the creator's thread ID in rsa->blinding and
9474 having all other threads use local one-time blinding factors
9475 (this requires more computation than sharing rsa->blinding, but
9476 avoids excessive locking; and if an RSA object is not shared
9477 between threads, blinding will still be very fast).
9480 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
9482 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
9483 via timing by performing a MAC computation even if incorrect
9484 block cipher padding has been found. This is a countermeasure
9485 against active attacks where the attacker has to distinguish
9486 between bad padding and a MAC verification error. (CVE-2003-0078)
9488 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
9489 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
9490 Martin Vuagnoux (EPFL, Ilion)]
9492 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
9494 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
9495 memory from its contents. This is done with a counter that will
9496 place alternating values in each byte. This can be used to solve
9497 two issues: 1) the removal of calls to memset() by highly optimizing
9498 compilers, and 2) cleansing with other values than 0, since those can
9499 be read through on certain media, for example a swap space on disk.
9502 *) Bugfix: client side session caching did not work with external caching,
9503 because the session->cipher setting was not restored when reloading
9504 from the external cache. This problem was masked, when
9505 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9506 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9509 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9510 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9511 [Zeev Lieber <zeev-l@yahoo.com>]
9513 *) Undo an undocumented change introduced in 0.9.6e which caused
9514 repeated calls to OpenSSL_add_all_ciphers() and
9515 OpenSSL_add_all_digests() to be ignored, even after calling
9519 *) Change the default configuration reader to deal with last line not
9520 being properly terminated.
9523 *) Change X509_NAME_cmp() so it applies the special rules on handling
9524 DN values that are of type PrintableString, as well as RDNs of type
9525 emailAddress where the value has the type ia5String.
9526 [stefank@valicert.com via Richard Levitte]
9528 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9529 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9530 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9531 the bitwise-OR of the two for use by the majority of applications
9532 wanting this behaviour, and update the docs. The documented
9533 behaviour and actual behaviour were inconsistent and had been
9534 changing anyway, so this is more a bug-fix than a behavioural
9536 [Geoff Thorpe, diagnosed by Nadav Har'El]
9538 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9539 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9542 *) Fix initialization code race conditions in
9543 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9544 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9545 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9546 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9547 ssl2_get_cipher_by_char(),
9548 ssl3_get_cipher_by_char().
9549 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9551 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9552 the cached sessions are flushed, as the remove_cb() might use ex_data
9553 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9554 (see [openssl.org #212]).
9555 [Geoff Thorpe, Lutz Jaenicke]
9557 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9558 length, instead of the encoding length to d2i_ASN1_OBJECT.
9561 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9563 *) [In 0.9.6g-engine release:]
9564 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9565 [Lynn Gazis <lgazis@rainbow.com>]
9567 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9569 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9570 and get fix the header length calculation.
9571 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9572 Alon Kantor <alonk@checkpoint.com> (and others),
9575 *) Use proper error handling instead of 'assertions' in buffer
9576 overflow checks added in 0.9.6e. This prevents DoS (the
9577 assertions could call abort()).
9578 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9580 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9582 *) Add various sanity checks to asn1_get_length() to reject
9583 the ASN1 length bytes if they exceed sizeof(long), will appear
9584 negative or the content length exceeds the length of the
9586 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9588 *) Fix cipher selection routines: ciphers without encryption had no flags
9589 for the cipher strength set and where therefore not handled correctly
9590 by the selection routines (PR #130).
9593 *) Fix EVP_dsa_sha macro.
9597 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9598 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9599 that was added in OpenSSL 0.9.6d.
9601 As the countermeasure turned out to be incompatible with some
9602 broken SSL implementations, the new option is part of SSL_OP_ALL.
9603 SSL_OP_ALL is usually employed when compatibility with weird SSL
9604 implementations is desired (e.g. '-bugs' option to 's_client' and
9605 's_server'), so the new option is automatically set in many
9609 *) Changes in security patch:
9611 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9612 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9613 Air Force Materiel Command, USAF, under agreement number
9616 *) Add various sanity checks to asn1_get_length() to reject
9617 the ASN1 length bytes if they exceed sizeof(long), will appear
9618 negative or the content length exceeds the length of the
9619 supplied buffer. (CVE-2002-0659)
9620 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9622 *) Assertions for various potential buffer overflows, not known to
9624 [Ben Laurie (CHATS)]
9626 *) Various temporary buffers to hold ASCII versions of integers were
9627 too small for 64 bit platforms. (CVE-2002-0655)
9628 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9630 *) Remote buffer overflow in SSL3 protocol - an attacker could
9631 supply an oversized session ID to a client. (CVE-2002-0656)
9632 [Ben Laurie (CHATS)]
9634 *) Remote buffer overflow in SSL2 protocol - an attacker could
9635 supply an oversized client master key. (CVE-2002-0656)
9636 [Ben Laurie (CHATS)]
9638 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9640 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9641 encoded as NULL) with id-dsa-with-sha1.
9642 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9644 *) Check various X509_...() return values in apps/req.c.
9645 [Nils Larsch <nla@trustcenter.de>]
9647 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9648 an end-of-file condition would erroneously be flagged, when the CRLF
9649 was just at the end of a processed block. The bug was discovered when
9650 processing data through a buffering memory BIO handing the data to a
9651 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9652 <ptsekov@syntrex.com> and Nedelcho Stanev.
9655 *) Implement a countermeasure against a vulnerability recently found
9656 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9657 before application data chunks to avoid the use of known IVs
9658 with data potentially chosen by the attacker.
9661 *) Fix length checks in ssl3_get_client_hello().
9664 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9665 to prevent ssl3_read_internal() from incorrectly assuming that
9666 ssl3_read_bytes() found application data while handshake
9667 processing was enabled when in fact s->s3->in_read_app_data was
9668 merely automatically cleared during the initial handshake.
9669 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9671 *) Fix object definitions for Private and Enterprise: they were not
9672 recognized in their shortname (=lowercase) representation. Extend
9673 obj_dat.pl to issue an error when using undefined keywords instead
9674 of silently ignoring the problem (Svenning Sorensen
9675 <sss@sss.dnsalias.net>).
9678 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9679 generators, i.e. generators other than 2 and 5. (Previously, the
9680 code did not properly initialise the 'add' and 'rem' values to
9681 BN_generate_prime().)
9683 In the new general case, we do not insist that 'generator' is
9684 actually a primitive root: This requirement is rather pointless;
9685 a generator of the order-q subgroup is just as good, if not
9689 *) Map new X509 verification errors to alerts. Discovered and submitted by
9690 Tom Wu <tom@arcot.com>.
9693 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9694 returning non-zero before the data has been completely received
9695 when using non-blocking I/O.
9696 [Bodo Moeller; problem pointed out by John Hughes]
9698 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9699 [Ben Laurie, Lutz Jaenicke]
9701 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9702 Yoram Zahavi <YoramZ@gilian.com>).
9705 *) Add information about CygWin 1.3 and on, and preserve proper
9706 configuration for the versions before that.
9707 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9709 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9710 check whether we deal with a copy of a session and do not delete from
9711 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9712 <izhar@checkpoint.com>.
9715 *) Do not store session data into the internal session cache, if it
9716 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9717 flag is set). Proposed by Aslam <aslam@funk.com>.
9720 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9724 *) [In 0.9.6d-engine release:]
9725 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9726 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9728 *) Add the configuration target linux-s390x.
9729 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9731 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9732 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9733 variable as an indication that a ClientHello message has been
9734 received. As the flag value will be lost between multiple
9735 invocations of ssl3_accept when using non-blocking I/O, the
9736 function may not be aware that a handshake has actually taken
9737 place, thus preventing a new session from being added to the
9740 To avoid this problem, we now set s->new_session to 2 instead of
9741 using a local variable.
9742 [Lutz Jaenicke, Bodo Moeller]
9744 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9745 if the SSL_R_LENGTH_MISMATCH error is detected.
9746 [Geoff Thorpe, Bodo Moeller]
9748 *) New 'shared_ldflag' column in Configure platform table.
9751 *) Fix EVP_CIPHER_mode macro.
9752 ["Dan S. Camper" <dan@bti.net>]
9754 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9755 type, we must throw them away by setting rr->length to 0.
9756 [D P Chang <dpc@qualys.com>]
9758 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9760 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9761 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9762 worked incorrectly for those cases where range = 10..._2 and
9763 3*range is two bits longer than range.)
9766 *) Only add signing time to PKCS7 structures if it is not already
9770 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9771 OBJ_ld_ce should be OBJ_id_ce.
9772 Also some ip-pda OIDs in crypto/objects/objects.txt were
9773 incorrect (cf. RFC 3039).
9774 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9776 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9777 returns early because it has nothing to do.
9778 [Andy Schneider <andy.schneider@bjss.co.uk>]
9780 *) [In 0.9.6c-engine release:]
9781 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9782 [Andy Schneider <andy.schneider@bjss.co.uk>]
9784 *) [In 0.9.6c-engine release:]
9785 Add support for Cryptographic Appliance's keyserver technology.
9786 (Use engine 'keyclient')
9787 [Cryptographic Appliances and Geoff Thorpe]
9789 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9790 is called via tools/c89.sh because arguments have to be
9791 rearranged (all '-L' options must appear before the first object
9793 [Richard Shapiro <rshapiro@abinitio.com>]
9795 *) [In 0.9.6c-engine release:]
9796 Add support for Broadcom crypto accelerator cards, backported
9798 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9800 *) [In 0.9.6c-engine release:]
9801 Add support for SureWare crypto accelerator cards from
9802 Baltimore Technologies. (Use engine 'sureware')
9803 [Baltimore Technologies and Mark Cox]
9805 *) [In 0.9.6c-engine release:]
9806 Add support for crypto accelerator cards from Accelerated
9807 Encryption Processing, www.aep.ie. (Use engine 'aep')
9808 [AEP Inc. and Mark Cox]
9810 *) Add a configuration entry for gcc on UnixWare.
9811 [Gary Benson <gbenson@redhat.com>]
9813 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9814 messages are stored in a single piece (fixed-length part and
9815 variable-length part combined) and fix various bugs found on the way.
9818 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9819 instead. BIO_gethostbyname() does not know what timeouts are
9820 appropriate, so entries would stay in cache even when they have
9822 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9824 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9825 faced with a pathologically small ClientHello fragment that does
9826 not contain client_version: Instead of aborting with an error,
9827 simply choose the highest available protocol version (i.e.,
9828 TLS 1.0 unless it is disabled). In practice, ClientHello
9829 messages are never sent like this, but this change gives us
9830 strictly correct behaviour at least for TLS.
9833 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9834 never resets s->method to s->ctx->method when called from within
9835 one of the SSL handshake functions.
9836 [Bodo Moeller; problem pointed out by Niko Baric]
9838 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9839 (sent using the client's version number) if client_version is
9840 smaller than the protocol version in use. Also change
9841 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9842 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9843 the client will at least see that alert.
9846 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9850 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9851 client receives HelloRequest while in a handshake.
9852 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9854 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9855 should end in 'break', not 'goto end' which circumvents various
9856 cleanups done in state SSL_ST_OK. But session related stuff
9857 must be disabled for SSL_ST_OK in the case that we just sent a
9860 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9861 before just sending a HelloRequest.
9862 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9864 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9865 reveal whether illegal block cipher padding was found or a MAC
9866 verification error occurred. (Neither SSLerr() codes nor alerts
9867 are directly visible to potential attackers, but the information
9868 may leak via logfiles.)
9870 Similar changes are not required for the SSL 2.0 implementation
9871 because the number of padding bytes is sent in clear for SSL 2.0,
9872 and the extra bytes are just ignored. However ssl/s2_pkt.c
9873 failed to verify that the purported number of padding bytes is in
9877 *) Add OpenUNIX-8 support including shared libraries
9878 (Boyd Lynn Gerber <gerberb@zenez.com>).
9881 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9882 'wristwatch attack' using huge encoding parameters (cf.
9883 James H. Manger's CRYPTO 2001 paper). Note that the
9884 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9885 encoding parameters and hence was not vulnerable.
9888 *) BN_sqr() bug fix.
9889 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9891 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9892 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9893 followed by modular reduction.
9894 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9896 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9897 equivalent based on BN_pseudo_rand() instead of BN_rand().
9900 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9901 This function was broken, as the check for a new client hello message
9902 to handle SGC did not allow these large messages.
9903 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9906 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9909 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9910 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9913 *) Rework the configuration and shared library support for Tru64 Unix.
9914 The configuration part makes use of modern compiler features and
9915 still retains old compiler behavior for those that run older versions
9916 of the OS. The shared library support part includes a variant that
9917 uses the RPATH feature, and is available through the special
9918 configuration target "alpha-cc-rpath", which will never be selected
9920 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9922 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9923 with the same message size as in ssl3_get_certificate_request().
9924 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9925 messages might inadvertently be reject as too long.
9926 [Petr Lampa <lampa@fee.vutbr.cz>]
9928 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9931 *) Modified SSL library such that the verify_callback that has been set
9932 specifically for an SSL object with SSL_set_verify() is actually being
9933 used. Before the change, a verify_callback set with this function was
9934 ignored and the verify_callback() set in the SSL_CTX at the time of
9935 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9936 to allow the necessary settings.
9939 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9940 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9941 done automatically (in contradiction to the requirements of the C
9942 standard). This made problems when used from OpenSSH.
9945 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9946 dh->length and always used
9948 BN_rand_range(priv_key, dh->p).
9950 BN_rand_range() is not necessary for Diffie-Hellman, and this
9951 specific range makes Diffie-Hellman unnecessarily inefficient if
9952 dh->length (recommended exponent length) is much smaller than the
9953 length of dh->p. We could use BN_rand_range() if the order of
9954 the subgroup was stored in the DH structure, but we only have
9959 BN_rand(priv_key, l, ...)
9961 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9967 RSA_eay_public_encrypt
9968 RSA_eay_private_decrypt
9969 RSA_eay_private_encrypt (signing)
9970 RSA_eay_public_decrypt (signature verification)
9972 (default implementations for RSA_public_encrypt,
9973 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9974 always reject numbers >= n.
9977 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9978 to synchronize access to 'locking_thread'. This is necessary on
9979 systems where access to 'locking_thread' (an 'unsigned long'
9980 variable) is not atomic.
9983 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9984 *before* setting the 'crypto_lock_rand' flag. The previous code had
9985 a race condition if 0 is a valid thread ID.
9986 [Travis Vitek <vitek@roguewave.com>]
9988 *) Add support for shared libraries under Irix.
9989 [Albert Chin-A-Young <china@thewrittenword.com>]
9991 *) Add configuration option to build on Linux on both big-endian and
9993 [Ralf Baechle <ralf@uni-koblenz.de>]
9995 *) Add the possibility to create shared libraries on HP-UX.
9998 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
10000 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
10001 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
10002 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
10003 PRNG state recovery was possible based on the output of
10004 one PRNG request appropriately sized to gain knowledge on
10005 'md' followed by enough consecutive 1-byte PRNG requests
10006 to traverse all of 'state'.
10008 1. When updating 'md_local' (the current thread's copy of 'md')
10009 during PRNG output generation, hash all of the previous
10010 'md_local' value, not just the half used for PRNG output.
10012 2. Make the number of bytes from 'state' included into the hash
10013 independent from the number of PRNG bytes requested.
10015 The first measure alone would be sufficient to avoid
10016 Markku-Juhani's attack. (Actually it had never occurred
10017 to me that the half of 'md_local' used for chaining was the
10018 half from which PRNG output bytes were taken -- I had always
10019 assumed that the secret half would be used.) The second
10020 measure makes sure that additional data from 'state' is never
10021 mixed into 'md_local' in small portions; this heuristically
10022 further strengthens the PRNG.
10025 *) Fix crypto/bn/asm/mips3.s.
10028 *) When only the key is given to "enc", the IV is undefined. Print out
10029 an error message in this case.
10032 *) Handle special case when X509_NAME is empty in X509 printing routines.
10035 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
10036 positive and less than q.
10039 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
10040 used: it isn't thread safe and the add_lock_callback should handle
10042 [Paul Rose <Paul.Rose@bridge.com>]
10044 *) Verify that incoming data obeys the block size in
10045 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
10049 [Ulf Möller, Bodo Möller]
10051 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
10052 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
10053 when fixing the server behaviour for backwards-compatible 'client
10054 hello' messages. (Note that the attack is impractical against
10055 SSL 3.0 and TLS 1.0 anyway because length and version checking
10056 means that the probability of guessing a valid ciphertext is
10057 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
10060 Before 0.9.5, the countermeasure (hide the error by generating a
10061 random 'decryption result') did not work properly because
10062 ERR_clear_error() was missing, meaning that SSL_get_error() would
10063 detect the supposedly ignored error.
10065 Both problems are now fixed.
10068 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
10069 (previously it was 1024).
10072 *) Fix for compatibility mode trust settings: ignore trust settings
10073 unless some valid trust or reject settings are present.
10076 *) Fix for blowfish EVP: its a variable length cipher.
10079 *) Fix various bugs related to DSA S/MIME verification. Handle missing
10080 parameters in DSA public key structures and return an error in the
10081 DSA routines if parameters are absent.
10084 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
10085 in the current directory if neither $RANDFILE nor $HOME was set.
10086 RAND_file_name() in 0.9.6a returned NULL in this case. This has
10087 caused some confusion to Windows users who haven't defined $HOME.
10088 Thus RAND_file_name() is changed again: e_os.h can define a
10089 DEFAULT_HOME, which will be used if $HOME is not set.
10090 For Windows, we use "C:"; on other platforms, we still require
10091 environment variables.
10093 *) Move 'if (!initialized) RAND_poll()' into regions protected by
10094 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
10095 having multiple threads call RAND_poll() concurrently.
10098 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
10099 combination of a flag and a thread ID variable.
10100 Otherwise while one thread is in ssleay_rand_bytes (which sets the
10101 flag), *other* threads can enter ssleay_add_bytes without obeying
10102 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
10103 that they do not hold after the first thread unsets add_do_not_lock).
10106 *) Change bctest again: '-x' expressions are not available in all
10107 versions of 'test'.
10110 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
10112 *) Fix a couple of memory leaks in PKCS7_dataDecode()
10113 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
10115 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
10116 the default extension for executables, if any. Also, make the perl
10117 scripts that use symlink() to test if it really exists and use "cp"
10118 if it doesn't. All this made OpenSSL compilable and installable in
10122 *) Fix for asn1_GetSequence() for indefinite length constructed data.
10123 If SEQUENCE is length is indefinite just set c->slen to the total
10124 amount of data available.
10125 [Steve Henson, reported by shige@FreeBSD.org]
10126 [This change does not apply to 0.9.7.]
10128 *) Change bctest to avoid here-documents inside command substitution
10129 (workaround for FreeBSD /bin/sh bug).
10130 For compatibility with Ultrix, avoid shell functions (introduced
10131 in the bctest version that searches along $PATH).
10134 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
10135 with des_encrypt() defined on some operating systems, like Solaris
10139 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
10140 On the Importance of Eliminating Errors in Cryptographic
10141 Computations, J. Cryptology 14 (2001) 2, 101-119,
10142 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
10145 *) MIPS assembler BIGNUM division bug fix.
10148 *) Disabled incorrect Alpha assembler code.
10151 *) Fix PKCS#7 decode routines so they correctly update the length
10152 after reading an EOC for the EXPLICIT tag.
10154 [This change does not apply to 0.9.7.]
10156 *) Fix bug in PKCS#12 key generation routines. This was triggered
10157 if a 3DES key was generated with a 0 initial byte. Include
10158 PKCS12_BROKEN_KEYGEN compilation option to retain the old
10159 (but broken) behaviour.
10162 *) Enhance bctest to search for a working bc along $PATH and print
10164 [Tim Rice <tim@multitalents.net> via Richard Levitte]
10166 *) Fix memory leaks in err.c: free err_data string if necessary;
10167 don't write to the wrong index in ERR_set_error_data.
10170 *) Implement ssl23_peek (analogous to ssl23_read), which previously
10174 *) Replace rdtsc with _emit statements for VC++ version 5.
10175 [Jeremy Cooper <jeremy@baymoo.org>]
10177 *) Make it possible to reuse SSLv2 sessions.
10180 *) In copy_email() check for >= 0 as a return value for
10181 X509_NAME_get_index_by_NID() since 0 is a valid index.
10182 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
10184 *) Avoid coredump with unsupported or invalid public keys by checking if
10185 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
10186 PKCS7_verify() fails with non detached data.
10189 *) Don't use getenv in library functions when run as setuid/setgid.
10190 New function OPENSSL_issetugid().
10193 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
10194 due to incorrect handling of multi-threading:
10196 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
10198 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
10200 3. Count how many times MemCheck_off() has been called so that
10201 nested use can be treated correctly. This also avoids
10202 inband-signalling in the previous code (which relied on the
10203 assumption that thread ID 0 is impossible).
10206 *) Add "-rand" option also to s_client and s_server.
10209 *) Fix CPU detection on Irix 6.x.
10210 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
10211 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10213 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
10216 [This change does not apply to 0.9.7.]
10218 *) Use the cached encoding of an X509_NAME structure rather than
10219 copying it. This is apparently the reason for the libsafe "errors"
10220 but the code is actually correct.
10223 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
10224 Bleichenbacher's DSA attack.
10225 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
10226 to be set and top=0 forces the highest bit to be set; top=-1 is new
10227 and leaves the highest bit random.
10228 [Ulf Moeller, Bodo Moeller]
10230 *) In the NCONF_...-based implementations for CONF_... queries
10231 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
10232 a temporary CONF structure with the data component set to NULL
10233 (which gives segmentation faults in lh_retrieve).
10234 Instead, use NULL for the CONF pointer in CONF_get_string and
10235 CONF_get_number (which may use environment variables) and directly
10236 return NULL from CONF_get_section.
10239 *) Fix potential buffer overrun for EBCDIC.
10242 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
10243 keyUsage if basicConstraints absent for a CA.
10246 *) Make SMIME_write_PKCS7() write mail header values with a format that
10247 is more generally accepted (no spaces before the semicolon), since
10248 some programs can't parse those values properly otherwise. Also make
10249 sure BIO's that break lines after each write do not create invalid
10253 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
10254 macros previously used would not encode an empty SEQUENCE OF
10255 and break the signature.
10257 [This change does not apply to 0.9.7.]
10259 *) Zero the premaster secret after deriving the master secret in
10263 *) Add some EVP_add_digest_alias registrations (as found in
10264 OpenSSL_add_all_digests()) to SSL_library_init()
10265 aka OpenSSL_add_ssl_algorithms(). This provides improved
10266 compatibility with peers using X.509 certificates
10267 with unconventional AlgorithmIdentifier OIDs.
10270 *) Fix for Irix with NO_ASM.
10271 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10273 *) ./config script fixes.
10274 [Ulf Moeller, Richard Levitte]
10276 *) Fix 'openssl passwd -1'.
10279 *) Change PKCS12_key_gen_asc() so it can cope with non null
10280 terminated strings whose length is passed in the passlen
10281 parameter, for example from PEM callbacks. This was done
10282 by adding an extra length parameter to asc2uni().
10283 [Steve Henson, reported by <oddissey@samsung.co.kr>]
10285 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
10286 call failed, free the DSA structure.
10289 *) Fix to uni2asc() to cope with zero length Unicode strings.
10290 These are present in some PKCS#12 files.
10293 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
10294 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
10295 when writing a 32767 byte record.
10296 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
10298 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
10299 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
10301 (RSA objects have a reference count access to which is protected
10302 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
10303 so they are meant to be shared between threads.)
10304 [Bodo Moeller, Geoff Thorpe; original patch submitted by
10305 "Reddie, Steven" <Steven.Reddie@ca.com>]
10307 *) Fix a deadlock in CRYPTO_mem_leaks().
10310 *) Use better test patterns in bntest.
10313 *) rand_win.c fix for Borland C.
10316 *) BN_rshift bugfix for n == 0.
10319 *) Add a 'bctest' script that checks for some known 'bc' bugs
10320 so that 'make test' does not abort just because 'bc' is broken.
10323 *) Store verify_result within SSL_SESSION also for client side to
10324 avoid potential security hole. (Re-used sessions on the client side
10325 always resulted in verify_result==X509_V_OK, not using the original
10326 result of the server certificate verification.)
10329 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
10330 SSL3_RT_APPLICATION_DATA, return 0.
10331 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
10335 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
10336 releases, have been re-implemented by renaming the previous
10337 implementations of ssl2_read and ssl3_read to ssl2_read_internal
10338 and ssl3_read_internal, respectively, and adding 'peek' parameters
10339 to them. The new ssl[23]_{read,peek} functions are calls to
10340 ssl[23]_read_internal with the 'peek' flag set appropriately.
10341 A 'peek' parameter has also been added to ssl3_read_bytes, which
10342 does the actual work for ssl3_read_internal.
10345 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
10346 the method-specific "init()" handler. Also clean up ex_data after
10347 calling the method-specific "finish()" handler. Previously, this was
10348 happening the other way round.
10351 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
10352 The previous value, 12, was not always sufficient for BN_mod_exp().
10355 *) Make sure that shared libraries get the internal name engine with
10356 the full version number and not just 0. This should mark the
10357 shared libraries as not backward compatible. Of course, this should
10358 be changed again when we can guarantee backward binary compatibility.
10361 *) Fix typo in get_cert_by_subject() in by_dir.c
10362 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
10364 *) Rework the system to generate shared libraries:
10366 - Make note of the expected extension for the shared libraries and
10367 if there is a need for symbolic links from for example libcrypto.so.0
10368 to libcrypto.so.0.9.7. There is extended info in Configure for
10371 - Make as few rebuilds of the shared libraries as possible.
10373 - Still avoid linking the OpenSSL programs with the shared libraries.
10375 - When installing, install the shared libraries separately from the
10379 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
10381 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
10382 and not in SSL_clear because the latter is also used by the
10383 accept/connect functions; previously, the settings made by
10384 SSL_set_read_ahead would be lost during the handshake.
10385 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
10387 *) Correct util/mkdef.pl to be selective about disabled algorithms.
10388 Previously, it would create entries for disabled algorithms no
10392 *) Added several new manual pages for SSL_* function.
10395 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
10397 *) In ssl23_get_client_hello, generate an error message when faced
10398 with an initial SSL 3.0/TLS record that is too small to contain the
10399 first two bytes of the ClientHello message, i.e. client_version.
10400 (Note that this is a pathologic case that probably has never happened
10401 in real life.) The previous approach was to use the version number
10402 from the record header as a substitute; but our protocol choice
10403 should not depend on that one because it is not authenticated
10404 by the Finished messages.
10407 *) More robust randomness gathering functions for Windows.
10408 [Jeffrey Altman <jaltman@columbia.edu>]
10410 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
10411 not set then we don't setup the error code for issuer check errors
10412 to avoid possibly overwriting other errors which the callback does
10413 handle. If an application does set the flag then we assume it knows
10414 what it is doing and can handle the new informational codes
10418 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
10419 a general "ANY" type, as such it should be able to decode anything
10420 including tagged types. However it didn't check the class so it would
10421 wrongly interpret tagged types in the same way as their universal
10422 counterpart and unknown types were just rejected. Changed so that the
10423 tagged and unknown types are handled in the same way as a SEQUENCE:
10424 that is the encoding is stored intact. There is also a new type
10425 "V_ASN1_OTHER" which is used when the class is not universal, in this
10426 case we have no idea what the actual type is so we just lump them all
10430 *) On VMS, stdout may very well lead to a file that is written to
10431 in a record-oriented fashion. That means that every write() will
10432 write a separate record, which will be read separately by the
10433 programs trying to read from it. This can be very confusing.
10435 The solution is to put a BIO filter in the way that will buffer
10436 text until a linefeed is reached, and then write everything a
10437 line at a time, so every record written will be an actual line,
10438 not chunks of lines and not (usually doesn't happen, but I've
10439 seen it once) several lines in one record. BIO_f_linebuffer() is
10442 Currently, it's a VMS-only method, because that's where it has
10443 been tested well enough.
10446 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
10447 it can return incorrect results.
10448 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
10449 but it was in 0.9.6-beta[12].)
10452 *) Disable the check for content being present when verifying detached
10453 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
10454 include zero length content when signing messages.
10457 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
10458 BIO_ctrl (for BIO pairs).
10461 *) Add DSO method for VMS.
10464 *) Bug fix: Montgomery multiplication could produce results with the
10468 *) Add RPM specification openssl.spec and modify it to build three
10469 packages. The default package contains applications, application
10470 documentation and run-time libraries. The devel package contains
10471 include files, static libraries and function documentation. The
10472 doc package contains the contents of the doc directory. The original
10473 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
10476 *) Add a large number of documentation files for many SSL routines.
10477 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10479 *) Add a configuration entry for Sony News 4.
10480 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
10482 *) Don't set the two most significant bits to one when generating a
10483 random number < q in the DSA library.
10486 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
10487 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
10488 the underlying transport is blocking) if a handshake took place.
10489 (The default behaviour is needed by applications such as s_client
10490 and s_server that use select() to determine when to use SSL_read;
10491 but for applications that know in advance when to expect data, it
10492 just makes things more complicated.)
10495 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
10499 *) Add a few more EBCDIC conditionals that make `req' and `x509'
10500 work better on such systems.
10501 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10503 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10504 Update PKCS12_parse() so it copies the friendlyName and the
10505 keyid to the certificates aux info.
10508 *) Fix bug in PKCS7_verify() which caused an infinite loop
10509 if there was more than one signature.
10510 [Sven Uszpelkat <su@celocom.de>]
10512 *) Major change in util/mkdef.pl to include extra information
10513 about each symbol, as well as presenting variables as well
10514 as functions. This change means that there's n more need
10515 to rebuild the .num files when some algorithms are excluded.
10518 *) Allow the verify time to be set by an application,
10519 rather than always using the current time.
10522 *) Phase 2 verify code reorganisation. The certificate
10523 verify code now looks up an issuer certificate by a
10524 number of criteria: subject name, authority key id
10525 and key usage. It also verifies self signed certificates
10526 by the same criteria. The main comparison function is
10527 X509_check_issued() which performs these checks.
10529 Lot of changes were necessary in order to support this
10530 without completely rewriting the lookup code.
10532 Authority and subject key identifier are now cached.
10534 The LHASH 'certs' is X509_STORE has now been replaced
10535 by a STACK_OF(X509_OBJECT). This is mainly because an
10536 LHASH can't store or retrieve multiple objects with
10537 the same hash value.
10539 As a result various functions (which were all internal
10540 use only) have changed to handle the new X509_STORE
10541 structure. This will break anything that messed round
10542 with X509_STORE internally.
10544 The functions X509_STORE_add_cert() now checks for an
10545 exact match, rather than just subject name.
10547 The X509_STORE API doesn't directly support the retrieval
10548 of multiple certificates matching a given criteria, however
10549 this can be worked round by performing a lookup first
10550 (which will fill the cache with candidate certificates)
10551 and then examining the cache for matches. This is probably
10552 the best we can do without throwing out X509_LOOKUP
10553 entirely (maybe later...).
10555 The X509_VERIFY_CTX structure has been enhanced considerably.
10557 All certificate lookup operations now go via a get_issuer()
10558 callback. Although this currently uses an X509_STORE it
10559 can be replaced by custom lookups. This is a simple way
10560 to bypass the X509_STORE hackery necessary to make this
10561 work and makes it possible to use more efficient techniques
10562 in future. A very simple version which uses a simple
10563 STACK for its trusted certificate store is also provided
10564 using X509_STORE_CTX_trusted_stack().
10566 The verify_cb() and verify() callbacks now have equivalents
10567 in the X509_STORE_CTX structure.
10569 X509_STORE_CTX also has a 'flags' field which can be used
10570 to customise the verify behaviour.
10573 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10574 excludes S/MIME capabilities.
10577 *) When a certificate request is read in keep a copy of the
10578 original encoding of the signed data and use it when outputting
10579 again. Signatures then use the original encoding rather than
10580 a decoded, encoded version which may cause problems if the
10581 request is improperly encoded.
10584 *) For consistency with other BIO_puts implementations, call
10585 buffer_write(b, ...) directly in buffer_puts instead of calling
10588 In BIO_puts, increment b->num_write as in BIO_write.
10589 [Peter.Sylvester@EdelWeb.fr]
10591 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10592 BN_zero, we may not return a BIGNUM with an array consisting of
10593 words set to zero.)
10596 *) Avoid calling abort() from within the library when problems are
10597 detected, except if preprocessor symbols have been defined
10598 (such as REF_CHECK, BN_DEBUG etc.).
10601 *) New openssl application 'rsautl'. This utility can be
10602 used for low level RSA operations. DER public key
10603 BIO/fp routines also added.
10606 *) New Configure entry and patches for compiling on QNX 4.
10607 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10609 *) A demo state-machine implementation was sponsored by
10610 Nuron (http://www.nuron.com/) and is now available in
10611 demos/state_machine.
10614 *) New options added to the 'dgst' utility for signature
10615 generation and verification.
10618 *) Unrecognized PKCS#7 content types are now handled via a
10619 catch all ASN1_TYPE structure. This allows unsupported
10620 types to be stored as a "blob" and an application can
10621 encode and decode it manually.
10624 *) Fix various signed/unsigned issues to make a_strex.c
10625 compile under VC++.
10626 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10628 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10629 length if passed a buffer. ASN1_INTEGER_to_BN failed
10630 if passed a NULL BN and its argument was negative.
10631 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10633 *) Modification to PKCS#7 encoding routines to output definite
10634 length encoding. Since currently the whole structures are in
10635 memory there's not real point in using indefinite length
10636 constructed encoding. However if OpenSSL is compiled with
10637 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10640 *) Added BIO_vprintf() and BIO_vsnprintf().
10643 *) Added more prefixes to parse for in the strings written
10644 through a logging bio, to cover all the levels that are available
10645 through syslog. The prefixes are now:
10647 PANIC, EMERG, EMR => LOG_EMERG
10648 ALERT, ALR => LOG_ALERT
10649 CRIT, CRI => LOG_CRIT
10650 ERROR, ERR => LOG_ERR
10651 WARNING, WARN, WAR => LOG_WARNING
10652 NOTICE, NOTE, NOT => LOG_NOTICE
10653 INFO, INF => LOG_INFO
10654 DEBUG, DBG => LOG_DEBUG
10656 and as before, if none of those prefixes are present at the
10657 beginning of the string, LOG_ERR is chosen.
10659 On Win32, the LOG_* levels are mapped according to this:
10661 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10662 LOG_WARNING => EVENTLOG_WARNING_TYPE
10663 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10667 *) Made it possible to reconfigure with just the configuration
10668 argument "reconf" or "reconfigure". The command line arguments
10669 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10670 and are retrieved from there when reconfiguring.
10673 *) MD4 implemented.
10674 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10676 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10679 *) The obj_dat.pl script was messing up the sorting of object
10680 names. The reason was that it compared the quoted version
10681 of strings as a result "OCSP" > "OCSP Signing" because
10682 " > SPACE. Changed script to store unquoted versions of
10683 names and add quotes on output. It was also omitting some
10684 names from the lookup table if they were given a default
10685 value (that is if SN is missing it is given the same
10686 value as LN and vice versa), these are now added on the
10687 grounds that if an object has a name we should be able to
10688 look it up. Finally added warning output when duplicate
10689 short or long names are found.
10692 *) Changes needed for Tandem NSK.
10693 [Scott Uroff <scott@xypro.com>]
10695 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10696 RSA_padding_check_SSLv23(), special padding was never detected
10697 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10698 version rollback attacks was not effective.
10700 In s23_clnt.c, don't use special rollback-attack detection padding
10701 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10702 client; similarly, in s23_srvr.c, don't do the rollback check if
10703 SSL 2.0 is the only protocol enabled in the server.
10706 *) Make it possible to get hexdumps of unprintable data with 'openssl
10707 asn1parse'. By implication, the functions ASN1_parse_dump() and
10708 BIO_dump_indent() are added.
10711 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10712 these print out strings and name structures based on various
10713 flags including RFC2253 support and proper handling of
10714 multibyte characters. Added options to the 'x509' utility
10715 to allow the various flags to be set.
10718 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10719 Also change the functions X509_cmp_current_time() and
10720 X509_gmtime_adj() work with an ASN1_TIME structure,
10721 this will enable certificates using GeneralizedTime in validity
10722 dates to be checked.
10725 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10726 negative public key encodings) on by default,
10727 NO_NEG_PUBKEY_BUG can be set to disable it.
10730 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10731 content octets. An i2c_ASN1_OBJECT is unnecessary because
10732 the encoding can be trivially obtained from the structure.
10735 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10736 not read locks (CRYPTO_r_[un]lock).
10739 *) A first attempt at creating official support for shared
10740 libraries through configuration. I've kept it so the
10741 default is static libraries only, and the OpenSSL programs
10742 are always statically linked for now, but there are
10743 preparations for dynamic linking in place.
10744 This has been tested on Linux and Tru64.
10747 *) Randomness polling function for Win9x, as described in:
10748 Peter Gutmann, Software Generation of Practically Strong
10752 *) Fix so PRNG is seeded in req if using an already existing
10756 *) New options to smime application. -inform and -outform
10757 allow alternative formats for the S/MIME message including
10758 PEM and DER. The -content option allows the content to be
10759 specified separately. This should allow things like Netscape
10760 form signing output easier to verify.
10763 *) Fix the ASN1 encoding of tags using the 'long form'.
10766 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10767 STRING types. These convert content octets to and from the
10768 underlying type. The actual tag and length octets are
10769 already assumed to have been read in and checked. These
10770 are needed because all other string types have virtually
10771 identical handling apart from the tag. By having versions
10772 of the ASN1 functions that just operate on content octets
10773 IMPLICIT tagging can be handled properly. It also allows
10774 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10775 and ASN1_INTEGER are identical apart from the tag.
10778 *) Change the handling of OID objects as follows:
10780 - New object identifiers are inserted in objects.txt, following
10781 the syntax given in objects.README.
10782 - objects.pl is used to process obj_mac.num and create a new
10784 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10787 This is currently kind of a hack, and the perl code in objects.pl
10788 isn't very elegant, but it works as I intended. The simplest way
10789 to check that it worked correctly is to look in obj_dat.h and
10790 check the array nid_objs and make sure the objects haven't moved
10791 around (this is important!). Additions are OK, as well as
10792 consistent name changes.
10795 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10798 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10799 The given file adds to whatever has already been seeded into the
10800 random pool through the RANDFILE configuration file option or
10801 environment variable, or the default random state file.
10804 *) mkstack.pl now sorts each macro group into lexical order.
10805 Previously the output order depended on the order the files
10806 appeared in the directory, resulting in needless rewriting
10810 *) Patches to make OpenSSL compile under Win32 again. Mostly
10811 work arounds for the VC++ problem that it treats func() as
10812 func(void). Also stripped out the parts of mkdef.pl that
10813 added extra typesafe functions: these no longer exist.
10816 *) Reorganisation of the stack code. The macros are now all
10817 collected in safestack.h . Each macro is defined in terms of
10818 a "stack macro" of the form SKM_<name>(type, a, b). The
10819 DEBUG_SAFESTACK is now handled in terms of function casts,
10820 this has the advantage of retaining type safety without the
10821 use of additional functions. If DEBUG_SAFESTACK is not defined
10822 then the non typesafe macros are used instead. Also modified the
10823 mkstack.pl script to handle the new form. Needs testing to see
10824 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10825 the default if no major problems. Similar behaviour for ASN1_SET_OF
10826 and PKCS12_STACK_OF.
10829 *) When some versions of IIS use the 'NET' form of private key the
10830 key derivation algorithm is different. Normally MD5(password) is
10831 used as a 128 bit RC4 key. In the modified case
10832 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10833 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10834 as the old Netscape_RSA functions except they have an additional
10835 'sgckey' parameter which uses the modified algorithm. Also added
10836 an -sgckey command line option to the rsa utility. Thanks to
10837 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10838 algorithm to openssl-dev.
10841 *) The evp_local.h macros were using 'c.##kname' which resulted in
10842 invalid expansion on some systems (SCO 5.0.5 for example).
10843 Corrected to 'c.kname'.
10844 [Phillip Porch <root@theporch.com>]
10846 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10847 a STACK of email addresses from a certificate or request, these look
10848 in the subject name and the subject alternative name extensions and
10849 omit any duplicate addresses.
10852 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10853 This makes DSA verification about 2 % faster.
10856 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10857 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10858 plus overhead for 1024 bit moduli).
10859 This makes exponentiations about 0.5 % faster for 1024 bit
10860 exponents (as measured by "openssl speed rsa2048").
10863 *) Rename memory handling macros to avoid conflicts with other
10865 Malloc => OPENSSL_malloc
10866 Malloc_locked => OPENSSL_malloc_locked
10867 Realloc => OPENSSL_realloc
10868 Free => OPENSSL_free
10871 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10872 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10875 *) CygWin32 support.
10876 [John Jarvie <jjarvie@newsguy.com>]
10878 *) The type-safe stack code has been rejigged. It is now only compiled
10879 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10880 by default all type-specific stack functions are "#define"d back to
10881 standard stack functions. This results in more streamlined output
10882 but retains the type-safety checking possibilities of the original
10886 *) The STACK code has been cleaned up, and certain type declarations
10887 that didn't make a lot of sense have been brought in line. This has
10888 also involved a cleanup of sorts in safestack.h to more correctly
10889 map type-safe stack functions onto their plain stack counterparts.
10890 This work has also resulted in a variety of "const"ifications of
10891 lots of the code, especially "_cmp" operations which should normally
10892 be prototyped with "const" parameters anyway.
10895 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10896 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10897 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10898 where all of 'md' is used each time the PRNG is used, but 'state'
10899 is used only indexed by a cyclic counter. As entropy may not be
10900 well distributed from the beginning, 'md' is important as a
10901 chaining variable. However, the output function chains only half
10902 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10903 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10904 in all of 'state' being rewritten, with the new values depending
10905 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10908 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10909 the handshake is continued after ssl_verify_cert_chain();
10910 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10911 can lead to 'unexplainable' connection aborts later.
10912 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10914 *) Major EVP API cipher revision.
10915 Add hooks for extra EVP features. This allows various cipher
10916 parameters to be set in the EVP interface. Support added for variable
10917 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10918 setting of RC2 and RC5 parameters.
10920 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10923 Remove lots of duplicated code from the EVP library. For example *every*
10924 cipher init() function handles the 'iv' in the same way according to the
10925 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10926 for CFB and OFB modes they zero ctx->num.
10928 New functionality allows removal of S/MIME code RC2 hack.
10930 Most of the routines have the same form and so can be declared in terms
10933 By shifting this to the top level EVP_CipherInit() it can be removed from
10934 all individual ciphers. If the cipher wants to handle IVs or keys
10935 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10938 Change lots of functions like EVP_EncryptUpdate() to now return a
10939 value: although software versions of the algorithms cannot fail
10940 any installed hardware versions can.
10943 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10944 this option is set, tolerate broken clients that send the negotiated
10945 protocol version number instead of the requested protocol version
10949 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10950 i.e. non-zero for export ciphersuites, zero otherwise.
10951 Previous versions had this flag inverted, inconsistent with
10952 rsa_tmp_cb (..._TMP_RSA_CB).
10953 [Bodo Moeller; problem reported by Amit Chopra]
10955 *) Add missing DSA library text string. Work around for some IIS
10956 key files with invalid SEQUENCE encoding.
10959 *) Add a document (doc/standards.txt) that list all kinds of standards
10960 and so on that are implemented in OpenSSL.
10963 *) Enhance c_rehash script. Old version would mishandle certificates
10964 with the same subject name hash and wouldn't handle CRLs at all.
10965 Added -fingerprint option to crl utility, to support new c_rehash
10969 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10972 *) Fix for SSL server purpose checking. Server checking was
10973 rejecting certificates which had extended key usage present
10974 but no ssl client purpose.
10975 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10977 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10978 is a little unclear about how a blank password is handled.
10979 Since the password in encoded as a BMPString with terminating
10980 double NULL a zero length password would end up as just the
10981 double NULL. However no password at all is different and is
10982 handled differently in the PKCS#12 key generation code. NS
10983 treats a blank password as zero length. MSIE treats it as no
10984 password on export: but it will try both on import. We now do
10985 the same: PKCS12_parse() tries zero length and no password if
10986 the password is set to "" or NULL (NULL is now a valid password:
10987 it wasn't before) as does the pkcs12 application.
10990 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10991 perror when PEM_read_bio_X509_REQ fails, the error message must
10992 be obtained from the error queue.
10995 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10996 it in ERR_remove_state if appropriate, and change ERR_get_state
10997 accordingly to avoid race conditions (this is necessary because
10998 thread_hash is no longer constant once set).
11001 *) Bugfix for linux-elf makefile.one.
11004 *) RSA_get_default_method() will now cause a default
11005 RSA_METHOD to be chosen if one doesn't exist already.
11006 Previously this was only set during a call to RSA_new()
11007 or RSA_new_method(NULL) meaning it was possible for
11008 RSA_get_default_method() to return NULL.
11011 *) Added native name translation to the existing DSO code
11012 that will convert (if the flag to do so is set) filenames
11013 that are sufficiently small and have no path information
11014 into a canonical native form. Eg. "blah" converted to
11015 "libblah.so" or "blah.dll" etc.
11018 *) New function ERR_error_string_n(e, buf, len) which is like
11019 ERR_error_string(e, buf), but writes at most 'len' bytes
11020 including the 0 terminator. For ERR_error_string_n, 'buf'
11022 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
11024 *) CONF library reworked to become more general. A new CONF
11025 configuration file reader "class" is implemented as well as a
11026 new functions (NCONF_*, for "New CONF") to handle it. The now
11027 old CONF_* functions are still there, but are reimplemented to
11028 work in terms of the new functions. Also, a set of functions
11029 to handle the internal storage of the configuration data is
11030 provided to make it easier to write new configuration file
11031 reader "classes" (I can definitely see something reading a
11032 configuration file in XML format, for example), called _CONF_*,
11033 or "the configuration storage API"...
11035 The new configuration file reading functions are:
11037 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
11038 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
11040 NCONF_default, NCONF_WIN32
11042 NCONF_dump_fp, NCONF_dump_bio
11044 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
11045 NCONF_new creates a new CONF object. This works in the same way
11046 as other interfaces in OpenSSL, like the BIO interface.
11047 NCONF_dump_* dump the internal storage of the configuration file,
11048 which is useful for debugging. All other functions take the same
11049 arguments as the old CONF_* functions with the exception of the
11050 first that must be a `CONF *' instead of a `LHASH *'.
11052 To make it easier to use the new classes with the old CONF_* functions,
11053 the function CONF_set_default_method is provided.
11056 *) Add '-tls1' option to 'openssl ciphers', which was already
11057 mentioned in the documentation but had not been implemented.
11058 (This option is not yet really useful because even the additional
11059 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
11062 *) Initial DSO code added into libcrypto for letting OpenSSL (and
11063 OpenSSL-based applications) load shared libraries and bind to
11064 them in a portable way.
11065 [Geoff Thorpe, with contributions from Richard Levitte]
11067 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
11069 *) Make sure _lrotl and _lrotr are only used with MSVC.
11071 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
11072 (the default implementation of RAND_status).
11074 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
11075 to '-clrext' (= clear extensions), as intended and documented.
11076 [Bodo Moeller; inconsistency pointed out by Michael Attili
11077 <attili@amaxo.com>]
11079 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
11080 was larger than the MD block size.
11081 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
11083 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
11084 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
11085 using the passed key: if the passed key was a private key the result
11086 of X509_print(), for example, would be to print out all the private key
11090 *) des_quad_cksum() byte order bug fix.
11091 [Ulf Möller, using the problem description in krb4-0.9.7, where
11092 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
11094 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
11096 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
11098 *) For easily testing in shell scripts whether some command
11099 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
11100 returns with exit code 0 iff no command of the given name is available.
11101 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
11102 the output goes to stdout and nothing is printed to stderr.
11103 Additional arguments are always ignored.
11105 Since for each cipher there is a command of the same name,
11106 the 'no-cipher' compilation switches can be tested this way.
11108 ('openssl no-XXX' is not able to detect pseudo-commands such
11109 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
11112 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
11115 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
11116 is set; it will be thrown away anyway because each handshake creates
11118 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
11119 to parameters -- in previous versions (since OpenSSL 0.9.3) the
11120 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
11121 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
11124 *) New s_client option -ign_eof: EOF at stdin is ignored, and
11125 'Q' and 'R' lose their special meanings (quit/renegotiate).
11126 This is part of what -quiet does; unlike -quiet, -ign_eof
11127 does not suppress any output.
11130 *) Add compatibility options to the purpose and trust code. The
11131 purpose X509_PURPOSE_ANY is "any purpose" which automatically
11132 accepts a certificate or CA, this was the previous behaviour,
11133 with all the associated security issues.
11135 X509_TRUST_COMPAT is the old trust behaviour: only and
11136 automatically trust self signed roots in certificate store. A
11137 new trust setting X509_TRUST_DEFAULT is used to specify that
11138 a purpose has no associated trust setting and it should instead
11139 use the value in the default purpose.
11142 *) Fix the PKCS#8 DSA private key code so it decodes keys again
11143 and fix a memory leak.
11146 *) In util/mkerr.pl (which implements 'make errors'), preserve
11147 reason strings from the previous version of the .c file, as
11148 the default to have only downcase letters (and digits) in
11149 automatically generated reasons codes is not always appropriate.
11152 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
11153 using strerror. Previously, ERR_reason_error_string() returned
11154 library names as reason strings for SYSerr; but SYSerr is a special
11155 case where small numbers are errno values, not library numbers.
11158 *) Add '-dsaparam' option to 'openssl dhparam' application. This
11159 converts DSA parameters into DH parameters. (When creating parameters,
11160 DSA_generate_parameters is used.)
11163 *) Include 'length' (recommended exponent length) in C code generated
11164 by 'openssl dhparam -C'.
11167 *) The second argument to set_label in perlasm was already being used
11168 so couldn't be used as a "file scope" flag. Moved to third argument
11172 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
11173 instead of RAND_bytes for encryption IVs and salts.
11176 *) Include RAND_status() into RAND_METHOD instead of implementing
11177 it only for md_rand.c Otherwise replacing the PRNG by calling
11178 RAND_set_rand_method would be impossible.
11181 *) Don't let DSA_generate_key() enter an infinite loop if the random
11182 number generation fails.
11185 *) New 'rand' application for creating pseudo-random output.
11188 *) Added configuration support for Linux/IA64
11189 [Rolf Haberrecker <rolf@suse.de>]
11191 *) Assembler module support for Mingw32.
11194 *) Shared library support for HPUX (in shlib/).
11195 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
11197 *) Shared library support for Solaris gcc.
11198 [Lutz Behnke <behnke@trustcenter.de>]
11200 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
11202 *) PKCS7_encrypt() was adding text MIME headers twice because they
11203 were added manually and by SMIME_crlf_copy().
11206 *) In bntest.c don't call BN_rand with zero bits argument.
11207 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
11209 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
11210 case was implemented. This caused BN_div_recp() to fail occasionally.
11213 *) Add an optional second argument to the set_label() in the perl
11214 assembly language builder. If this argument exists and is set
11215 to 1 it signals that the assembler should use a symbol whose
11216 scope is the entire file, not just the current function. This
11217 is needed with MASM which uses the format label:: for this scope.
11218 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
11220 *) Change the ASN1 types so they are typedefs by default. Before
11221 almost all types were #define'd to ASN1_STRING which was causing
11222 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
11226 *) Change names of new functions to the new get1/get0 naming
11227 convention: After 'get1', the caller owns a reference count
11228 and has to call ..._free; 'get0' returns a pointer to some
11229 data structure without incrementing reference counters.
11230 (Some of the existing 'get' functions increment a reference
11231 counter, some don't.)
11232 Similarly, 'set1' and 'add1' functions increase reference
11233 counters or duplicate objects.
11236 *) Allow for the possibility of temp RSA key generation failure:
11237 the code used to assume it always worked and crashed on failure.
11240 *) Fix potential buffer overrun problem in BIO_printf().
11241 [Ulf Möller, using public domain code by Patrick Powell; problem
11242 pointed out by David Sacerdote <das33@cornell.edu>]
11244 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
11245 RAND_egd() and RAND_status(). In the command line application,
11246 the EGD socket can be specified like a seed file using RANDFILE
11250 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
11251 Some CAs (e.g. Verisign) distribute certificates in this form.
11254 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
11255 list to exclude them. This means that no special compilation option
11256 is needed to use anonymous DH: it just needs to be included in the
11260 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
11261 EVP_MD_type. The old functionality is available in a new macro called
11262 EVP_MD_md(). Change code that uses it and update docs.
11265 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
11266 where the 'void *' argument is replaced by a function pointer argument.
11267 Previously 'void *' was abused to point to functions, which works on
11268 many platforms, but is not correct. As these functions are usually
11269 called by macros defined in OpenSSL header files, most source code
11270 should work without changes.
11273 *) <openssl/opensslconf.h> (which is created by Configure) now contains
11274 sections with information on -D... compiler switches used for
11275 compiling the library so that applications can see them. To enable
11276 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
11277 must be defined. E.g.,
11278 #define OPENSSL_ALGORITHM_DEFINES
11279 #include <openssl/opensslconf.h>
11280 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
11281 [Richard Levitte, Ulf and Bodo Möller]
11283 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
11287 *) Change the 'other' type in certificate aux info to a STACK_OF
11288 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
11289 the required ASN1 format: arbitrary types determined by an OID.
11292 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
11293 argument to 'req'. This is not because the function is newer or
11294 better than others it just uses the work 'NEW' in the certificate
11295 request header lines. Some software needs this.
11298 *) Reorganise password command line arguments: now passwords can be
11299 obtained from various sources. Delete the PEM_cb function and make
11300 it the default behaviour: i.e. if the callback is NULL and the
11301 usrdata argument is not NULL interpret it as a null terminated pass
11302 phrase. If usrdata and the callback are NULL then the pass phrase
11303 is prompted for as usual.
11306 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
11307 the support is automatically enabled. The resulting binaries will
11308 autodetect the card and use it if present.
11309 [Ben Laurie and Compaq Inc.]
11311 *) Work around for Netscape hang bug. This sends certificate request
11312 and server done in one record. Since this is perfectly legal in the
11313 SSL/TLS protocol it isn't a "bug" option and is on by default. See
11314 the bugs/SSLv3 entry for more info.
11317 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
11320 *) Add -rand argument to smime and pkcs12 applications and read/write
11324 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
11327 *) Add command line password options to the remaining applications.
11330 *) Bug fix for BN_div_recp() for numerators with an even number of
11334 *) More tests in bntest.c, and changed test_bn output.
11337 *) ./config recognizes MacOS X now.
11340 *) Bug fix for BN_div() when the first words of num and divisor are
11341 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
11344 *) Add support for various broken PKCS#8 formats, and command line
11345 options to produce them.
11348 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
11349 get temporary BIGNUMs from a BN_CTX.
11352 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
11356 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
11357 include a #define from the old name to the new. The original intent
11358 was that statically linked binaries could for example just call
11359 SSLeay_add_all_ciphers() to just add ciphers to the table and not
11360 link with digests. This never worked because SSLeay_add_all_digests()
11361 and SSLeay_add_all_ciphers() were in the same source file so calling
11362 one would link with the other. They are now in separate source files.
11365 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
11368 *) Use a less unusual form of the Miller-Rabin primality test (it used
11369 a binary algorithm for exponentiation integrated into the Miller-Rabin
11370 loop, our standard modexp algorithms are faster).
11373 *) Support for the EBCDIC character set completed.
11374 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
11376 *) Source code cleanups: use const where appropriate, eliminate casts,
11377 use void * instead of char * in lhash.
11380 *) Bugfix: ssl3_send_server_key_exchange was not restartable
11381 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
11382 this the server could overwrite ephemeral keys that the client
11386 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
11387 using 50 iterations of the Rabin-Miller test.
11389 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
11390 iterations of the Rabin-Miller test as required by the appendix
11391 to FIPS PUB 186[-1]) instead of DSA_is_prime.
11392 As BN_is_prime_fasttest includes trial division, DSA parameter
11393 generation becomes much faster.
11395 This implies a change for the callback functions in DSA_is_prime
11396 and DSA_generate_parameters: The callback function is called once
11397 for each positive witness in the Rabin-Miller test, not just
11398 occasionally in the inner loop; and the parameters to the
11399 callback function now provide an iteration count for the outer
11400 loop rather than for the current invocation of the inner loop.
11401 DSA_generate_parameters additionally can call the callback
11402 function with an 'iteration count' of -1, meaning that a
11403 candidate has passed the trial division test (when q is generated
11404 from an application-provided seed, trial division is skipped).
11407 *) New function BN_is_prime_fasttest that optionally does trial
11408 division before starting the Rabin-Miller test and has
11409 an additional BN_CTX * argument (whereas BN_is_prime always
11410 has to allocate at least one BN_CTX).
11411 'callback(1, -1, cb_arg)' is called when a number has passed the
11412 trial division stage.
11415 *) Fix for bug in CRL encoding. The validity dates weren't being handled
11419 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
11422 *) New function BN_pseudo_rand().
11425 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
11426 bignum version of BN_from_montgomery() with the working code from
11427 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
11431 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
11432 made it impossible to use the same SSL_SESSION data structure in
11433 SSL2 clients in multiple threads.
11436 *) The return value of RAND_load_file() no longer counts bytes obtained
11437 by stat(). RAND_load_file(..., -1) is new and uses the complete file
11438 to seed the PRNG (previously an explicit byte count was required).
11439 [Ulf Möller, Bodo Möller]
11441 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
11442 used (char *) instead of (void *) and had casts all over the place.
11445 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
11448 *) Retain source code compatibility for BN_prime_checks macro:
11449 BN_is_prime(..., BN_prime_checks, ...) now uses
11450 BN_prime_checks_for_size to determine the appropriate number of
11451 Rabin-Miller iterations.
11454 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
11455 DH_CHECK_P_NOT_SAFE_PRIME.
11456 (Check if this is true? OpenPGP calls them "strong".)
11459 *) Merge the functionality of "dh" and "gendh" programs into a new program
11460 "dhparam". The old programs are retained for now but will handle DH keys
11461 (instead of parameters) in future.
11464 *) Make the ciphers, s_server and s_client programs check the return values
11465 when a new cipher list is set.
11468 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
11469 ciphers. Before when the 56bit ciphers were enabled the sorting was
11472 The syntax for the cipher sorting has been extended to support sorting by
11473 cipher-strength (using the strength_bits hard coded in the tables).
11474 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
11476 Fix a bug in the cipher-command parser: when supplying a cipher command
11477 string with an "undefined" symbol (neither command nor alphanumeric
11478 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
11479 an error is flagged.
11481 Due to the strength-sorting extension, the code of the
11482 ssl_create_cipher_list() function was completely rearranged. I hope that
11483 the readability was also increased :-)
11484 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
11486 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
11487 for the first serial number and places 2 in the serial number file. This
11488 avoids problems when the root CA is created with serial number zero and
11489 the first user certificate has the same issuer name and serial number
11493 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
11494 the new code. Add documentation for this stuff.
11497 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
11498 X509_*() to X509at_*() on the grounds that they don't handle X509
11499 structures and behave in an analogous way to the X509v3 functions:
11500 they shouldn't be called directly but wrapper functions should be used
11503 So we also now have some wrapper functions that call the X509at functions
11504 when passed certificate requests. (TO DO: similar things can be done with
11505 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11506 things. Some of these need some d2i or i2d and print functionality
11507 because they handle more complex structures.)
11510 *) Add missing #ifndefs that caused missing symbols when building libssl
11511 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11512 NO_RSA in ssl/s2*.c.
11513 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11515 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11516 has a return value which indicates the quality of the random data
11517 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11518 error queue. New function RAND_pseudo_bytes() generates output that is
11519 guaranteed to be unique but not unpredictable. RAND_add is like
11520 RAND_seed, but takes an extra argument for an entropy estimate
11521 (RAND_seed always assumes full entropy).
11524 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11525 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11526 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11527 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11528 false-positive rate of at most 2^-80 for random input.
11531 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11534 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11535 in the 0.9.5 release), this returns the chain
11536 from an X509_CTX structure with a dup of the stack and all
11537 the X509 reference counts upped: so the stack will exist
11538 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11541 Also make SSL_SESSION_print() print out the verify return
11545 *) Add manpage for the pkcs12 command. Also change the default
11546 behaviour so MAC iteration counts are used unless the new
11547 -nomaciter option is used. This improves file security and
11548 only older versions of MSIE (4.0 for example) need it.
11551 *) Honor the no-xxx Configure options when creating .DEF files.
11554 *) Add PKCS#10 attributes to field table: challengePassword,
11555 unstructuredName and unstructuredAddress. These are taken from
11556 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11557 international characters are used.
11559 More changes to X509_ATTRIBUTE code: allow the setting of types
11560 based on strings. Remove the 'loc' parameter when adding
11561 attributes because these will be a SET OF encoding which is sorted
11565 *) Initial changes to the 'req' utility to allow request generation
11566 automation. This will allow an application to just generate a template
11567 file containing all the field values and have req construct the
11570 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11571 used all over the place including certificate requests and PKCS#7
11572 structures. They are currently handled manually where necessary with
11573 some primitive wrappers for PKCS#7. The new functions behave in a
11574 manner analogous to the X509 extension functions: they allow
11575 attributes to be looked up by NID and added.
11577 Later something similar to the X509V3 code would be desirable to
11578 automatically handle the encoding, decoding and printing of the
11579 more complex types. The string types like challengePassword can
11580 be handled by the string table functions.
11582 Also modified the multi byte string table handling. Now there is
11583 a 'global mask' which masks out certain types. The table itself
11584 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11585 is useful when for example there is only one permissible type
11586 (as in countryName) and using the mask might result in no valid
11590 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11591 SSL_get_peer_finished to allow applications to obtain the latest
11592 Finished messages sent to the peer or expected from the peer,
11593 respectively. (SSL_get_peer_finished is usually the Finished message
11594 actually received from the peer, otherwise the protocol will be aborted.)
11596 As the Finished message are message digests of the complete handshake
11597 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11598 be used for external authentication procedures when the authentication
11599 provided by SSL/TLS is not desired or is not enough.
11602 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11603 the host supports BWX extension and if Compaq C is present on the
11604 $PATH. Just exploiting of the BWX extension results in 20-30%
11605 performance kick for some algorithms, e.g. DES and RC4 to mention
11606 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11610 *) Add support for MS "fast SGC". This is arguably a violation of the
11611 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11612 weak crypto and after checking the certificate is SGC a second one
11613 with strong crypto. MS SGC stops the first handshake after receiving
11614 the server certificate message and sends a second client hello. Since
11615 a server will typically do all the time consuming operations before
11616 expecting any further messages from the client (server key exchange
11617 is the most expensive) there is little difference between the two.
11619 To get OpenSSL to support MS SGC we have to permit a second client
11620 hello message after we have sent server done. In addition we have to
11621 reset the MAC if we do get this second client hello.
11624 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11625 if a DER encoded private key is RSA or DSA traditional format. Changed
11626 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11627 format DER encoded private key. Newer code should use PKCS#8 format which
11628 has the key type encoded in the ASN1 structure. Added DER private key
11629 support to pkcs8 application.
11632 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11633 ciphersuites has been selected (as required by the SSL 3/TLS 1
11634 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11635 is set, we interpret this as a request to violate the specification
11636 (the worst that can happen is a handshake failure, and 'correct'
11637 behaviour would result in a handshake failure anyway).
11640 *) In SSL_CTX_add_session, take into account that there might be multiple
11641 SSL_SESSION structures with the same session ID (e.g. when two threads
11642 concurrently obtain them from an external cache).
11643 The internal cache can handle only one SSL_SESSION with a given ID,
11644 so if there's a conflict, we now throw out the old one to achieve
11648 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11649 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11650 some routines that use cipher OIDs: some ciphers do not have OIDs
11651 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11655 *) Simplify the trust setting structure and code. Now we just have
11656 two sequences of OIDs for trusted and rejected settings. These will
11657 typically have values the same as the extended key usage extension
11658 and any application specific purposes.
11660 The trust checking code now has a default behaviour: it will just
11661 check for an object with the same NID as the passed id. Functions can
11662 be provided to override either the default behaviour or the behaviour
11663 for a given id. SSL client, server and email already have functions
11664 in place for compatibility: they check the NID and also return "trusted"
11665 if the certificate is self signed.
11668 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11669 traditional format into an EVP_PKEY structure.
11672 *) Add a password callback function PEM_cb() which either prompts for
11673 a password if usr_data is NULL or otherwise assumes it is a null
11674 terminated password. Allow passwords to be passed on command line
11675 environment or config files in a few more utilities.
11678 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11679 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11680 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11681 Update documentation.
11684 *) Support for ASN1 "NULL" type. This could be handled before by using
11685 ASN1_TYPE but there wasn't any function that would try to read a NULL
11686 and produce an error if it couldn't. For compatibility we also have
11687 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11688 don't allocate anything because they don't need to.
11691 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11693 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11695 *) Rebuild of the memory allocation routines used by OpenSSL code and
11696 possibly others as well. The purpose is to make an interface that
11697 provide hooks so anyone can build a separate set of allocation and
11698 deallocation routines to be used by OpenSSL, for example memory
11699 pool implementations, or something else, which was previously hard
11700 since Malloc(), Realloc() and Free() were defined as macros having
11701 the values malloc, realloc and free, respectively (except for Win32
11702 compilations). The same is provided for memory debugging code.
11703 OpenSSL already comes with functionality to find memory leaks, but
11704 this gives people a chance to debug other memory problems.
11706 With these changes, a new set of functions and macros have appeared:
11708 CRYPTO_set_mem_debug_functions() [F]
11709 CRYPTO_get_mem_debug_functions() [F]
11710 CRYPTO_dbg_set_options() [F]
11711 CRYPTO_dbg_get_options() [F]
11712 CRYPTO_malloc_debug_init() [M]
11714 The memory debug functions are NULL by default, unless the library
11715 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11716 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11717 gives the standard debugging functions that come with OpenSSL) or
11718 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11719 provided by the library user) must be used. When the standard
11720 debugging functions are used, CRYPTO_dbg_set_options can be used to
11721 request additional information:
11722 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11723 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11725 Also, things like CRYPTO_set_mem_functions will always give the
11726 expected result (the new set of functions is used for allocation
11727 and deallocation) at all times, regardless of platform and compiler
11730 To finish it up, some functions that were never use in any other
11731 way than through macros have a new API and new semantic:
11733 CRYPTO_dbg_malloc()
11734 CRYPTO_dbg_realloc()
11737 All macros of value have retained their old syntax.
11738 [Richard Levitte and Bodo Moeller]
11740 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11741 ordering of SMIMECapabilities wasn't in "strength order" and there
11742 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11746 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11747 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11748 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11750 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11751 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11752 functionality to handle multipart/signed properly) and a utility
11753 called 'smime' to call all this stuff. This is based on code I
11754 originally wrote for Celo who have kindly allowed it to be
11755 included in OpenSSL.
11758 *) Add variants des_set_key_checked and des_set_key_unchecked of
11759 des_set_key (aka des_key_sched). Global variable des_check_key
11760 decides which of these is called by des_set_key; this way
11761 des_check_key behaves as it always did, but applications and
11762 the library itself, which was buggy for des_check_key == 1,
11763 have a cleaner way to pick the version they need.
11766 *) New function PKCS12_newpass() which changes the password of a
11770 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11771 dynamic mix. In both cases the ids can be used as an index into the
11772 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11773 functions so they accept a list of the field values and the
11774 application doesn't need to directly manipulate the X509_TRUST
11778 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11782 *) Modify the way the V3 extension code looks up extensions. This now
11783 works in a similar way to the object code: we have some "standard"
11784 extensions in a static table which is searched with OBJ_bsearch()
11785 and the application can add dynamic ones if needed. The file
11786 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11787 updated whenever a new extension is added to the core code and kept
11788 in ext_nid order. There is a simple program 'tabtest.c' which checks
11789 this. New extensions are not added too often so this file can readily
11790 be maintained manually.
11792 There are two big advantages in doing things this way. The extensions
11793 can be looked up immediately and no longer need to be "added" using
11794 X509V3_add_standard_extensions(): this function now does nothing.
11795 [Side note: I get *lots* of email saying the extension code doesn't
11796 work because people forget to call this function]
11797 Also no dynamic allocation is done unless new extensions are added:
11798 so if we don't add custom extensions there is no need to call
11799 X509V3_EXT_cleanup().
11802 *) Modify enc utility's salting as follows: make salting the default. Add a
11803 magic header, so unsalted files fail gracefully instead of just decrypting
11804 to garbage. This is because not salting is a big security hole, so people
11805 should be discouraged from doing it.
11808 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11809 digest to be passed on the command line but it only used this
11810 parameter when signing a certificate. Modified so all relevant
11811 operations are affected by the digest parameter including the
11812 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11813 DSA key was used because it didn't fix the digest.
11816 *) Initial certificate chain verify code. Currently tests the untrusted
11817 certificates for consistency with the verify purpose (which is set
11818 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11820 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11821 this is because it will reject chains with invalid extensions whereas
11822 every previous version of OpenSSL and SSLeay made no checks at all.
11824 Trust code: checks the root CA for the relevant trust settings. Trust
11825 settings have an initial value consistent with the verify purpose: e.g.
11826 if the verify purpose is for SSL client use it expects the CA to be
11827 trusted for SSL client use. However the default value can be changed to
11828 permit custom trust settings: one example of this would be to only trust
11829 certificates from a specific "secure" set of CAs.
11831 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11832 which should be used for version portability: especially since the
11833 verify structure is likely to change more often now.
11835 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11836 to set them. If not set then assume SSL clients will verify SSL servers
11839 Two new options to the verify program: -untrusted allows a set of
11840 untrusted certificates to be passed in and -purpose which sets the
11841 intended purpose of the certificate. If a purpose is set then the
11842 new chain verify code is used to check extension consistency.
11845 *) Support for the authority information access extension.
11848 *) Modify RSA and DSA PEM read routines to transparently handle
11849 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11850 public keys in a format compatible with certificate
11851 SubjectPublicKeyInfo structures. Unfortunately there were already
11852 functions called *_PublicKey_* which used various odd formats so
11853 these are retained for compatibility: however the DSA variants were
11854 never in a public release so they have been deleted. Changed dsa/rsa
11855 utilities to handle the new format: note no releases ever handled public
11856 keys so we should be OK.
11858 The primary motivation for this change is to avoid the same fiasco
11859 that dogs private keys: there are several incompatible private key
11860 formats some of which are standard and some OpenSSL specific and
11861 require various evil hacks to allow partial transparent handling and
11862 even then it doesn't work with DER formats. Given the option anything
11863 other than PKCS#8 should be dumped: but the other formats have to
11864 stay in the name of compatibility.
11866 With public keys and the benefit of hindsight one standard format
11867 is used which works with EVP_PKEY, RSA or DSA structures: though
11868 it clearly returns an error if you try to read the wrong kind of key.
11870 Added a -pubkey option to the 'x509' utility to output the public key.
11871 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11872 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11873 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11874 that do the same as the EVP_PKEY_assign_*() except they up the
11875 reference count of the added key (they don't "swallow" the
11879 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11880 CRLs would fail if the file contained no certificates or no CRLs:
11881 added a new function to read in both types and return the number
11882 read: this means that if none are read it will be an error. The
11883 DER versions of the certificate and CRL reader would always fail
11884 because it isn't possible to mix certificates and CRLs in DER format
11885 without choking one or the other routine. Changed this to just read
11886 a certificate: this is the best we can do. Also modified the code
11887 in apps/verify.c to take notice of return codes: it was previously
11888 attempting to read in certificates from NULL pointers and ignoring
11889 any errors: this is one reason why the cert and CRL reader seemed
11890 to work. It doesn't check return codes from the default certificate
11891 routines: these may well fail if the certificates aren't installed.
11894 *) Code to support otherName option in GeneralName.
11897 *) First update to verify code. Change the verify utility
11898 so it warns if it is passed a self signed certificate:
11899 for consistency with the normal behaviour. X509_verify
11900 has been modified to it will now verify a self signed
11901 certificate if *exactly* the same certificate appears
11902 in the store: it was previously impossible to trust a
11903 single self signed certificate. This means that:
11904 openssl verify ss.pem
11905 now gives a warning about a self signed certificate but
11906 openssl verify -CAfile ss.pem ss.pem
11910 *) For servers, store verify_result in SSL_SESSION data structure
11911 (and add it to external session representation).
11912 This is needed when client certificate verifications fails,
11913 but an application-provided verification callback (set by
11914 SSL_CTX_set_cert_verify_callback) allows accepting the session
11915 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11916 but returns 1): When the session is reused, we have to set
11917 ssl->verify_result to the appropriate error code to avoid
11919 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11921 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11922 case in PKCS7_dataInit() where the signed PKCS7 structure
11923 didn't contain any existing data because it was being created.
11924 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11926 *) Add a salt to the key derivation routines in enc.c. This
11927 forms the first 8 bytes of the encrypted file. Also add a
11928 -S option to allow a salt to be input on the command line.
11931 *) New function X509_cmp(). Oddly enough there wasn't a function
11932 to compare two certificates. We do this by working out the SHA1
11933 hash and comparing that. X509_cmp() will be needed by the trust
11937 *) SSL_get1_session() is like SSL_get_session(), but increments
11938 the reference count in the SSL_SESSION returned.
11939 [Geoff Thorpe <geoff@eu.c2.net>]
11941 *) Fix for 'req': it was adding a null to request attributes.
11942 Also change the X509_LOOKUP and X509_INFO code to handle
11943 certificate auxiliary information.
11946 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11950 *) Add the possibility to add extra information to the memory leak
11951 detecting output, to form tracebacks, showing from where each
11952 allocation was originated: CRYPTO_push_info("constant string") adds
11953 the string plus current file name and line number to a per-thread
11954 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11955 is like calling CYRPTO_pop_info() until the stack is empty.
11956 Also updated memory leak detection code to be multi-thread-safe.
11959 *) Add options -text and -noout to pkcs7 utility and delete the
11960 encryption options which never did anything. Update docs.
11963 *) Add options to some of the utilities to allow the pass phrase
11964 to be included on either the command line (not recommended on
11965 OSes like Unix) or read from the environment. Update the
11966 manpages and fix a few bugs.
11969 *) Add a few manpages for some of the openssl commands.
11972 *) Fix the -revoke option in ca. It was freeing up memory twice,
11973 leaking and not finding already revoked certificates.
11976 *) Extensive changes to support certificate auxiliary information.
11977 This involves the use of X509_CERT_AUX structure and X509_AUX
11978 functions. An X509_AUX function such as PEM_read_X509_AUX()
11979 can still read in a certificate file in the usual way but it
11980 will also read in any additional "auxiliary information". By
11981 doing things this way a fair degree of compatibility can be
11982 retained: existing certificates can have this information added
11983 using the new 'x509' options.
11985 Current auxiliary information includes an "alias" and some trust
11986 settings. The trust settings will ultimately be used in enhanced
11987 certificate chain verification routines: currently a certificate
11988 can only be trusted if it is self signed and then it is trusted
11992 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11993 The problem was that one of the replacement routines had not been working
11994 since SSLeay releases. For now the offending routine has been replaced
11995 with non-optimised assembler. Even so, this now gives around 95%
11996 performance improvement for 1024 bit RSA signs.
11999 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
12000 handling. Most clients have the effective key size in bits equal to
12001 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
12002 A few however don't do this and instead use the size of the decrypted key
12003 to determine the RC2 key length and the AlgorithmIdentifier to determine
12004 the effective key length. In this case the effective key length can still
12005 be 40 bits but the key length can be 168 bits for example. This is fixed
12006 by manually forcing an RC2 key into the EVP_PKEY structure because the
12007 EVP code can't currently handle unusual RC2 key sizes: it always assumes
12008 the key length and effective key length are equal.
12011 *) Add a bunch of functions that should simplify the creation of
12012 X509_NAME structures. Now you should be able to do:
12013 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
12014 and have it automatically work out the correct field type and fill in
12015 the structures. The more adventurous can try:
12016 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
12017 and it will (hopefully) work out the correct multibyte encoding.
12020 *) Change the 'req' utility to use the new field handling and multibyte
12021 copy routines. Before the DN field creation was handled in an ad hoc
12022 way in req, ca, and x509 which was rather broken and didn't support
12023 BMPStrings or UTF8Strings. Since some software doesn't implement
12024 BMPStrings or UTF8Strings yet, they can be enabled using the config file
12025 using the dirstring_type option. See the new comment in the default
12026 openssl.cnf for more info.
12029 *) Make crypto/rand/md_rand.c more robust:
12030 - Assure unique random numbers after fork().
12031 - Make sure that concurrent threads access the global counter and
12032 md serializably so that we never lose entropy in them
12033 or use exactly the same state in multiple threads.
12034 Access to the large state is not always serializable because
12035 the additional locking could be a performance killer, and
12036 md should be large enough anyway.
12039 *) New file apps/app_rand.c with commonly needed functionality
12040 for handling the random seed file.
12042 Use the random seed file in some applications that previously did not:
12044 dsaparam -genkey (which also ignored its '-rand' option),
12047 x509 (when signing).
12048 Except on systems with /dev/urandom, it is crucial to have a random
12049 seed file at least for key creation, DSA signing, and for DH exchanges;
12050 for RSA signatures we could do without one.
12052 gendh and gendsa (unlike genrsa) used to read only the first byte
12053 of each file listed in the '-rand' option. The function as previously
12054 found in genrsa is now in app_rand.c and is used by all programs
12055 that support '-rand'.
12058 *) In RAND_write_file, use mode 0600 for creating files;
12059 don't just chmod when it may be too late.
12062 *) Report an error from X509_STORE_load_locations
12063 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
12066 *) New function ASN1_mbstring_copy() this copies a string in either
12067 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
12068 into an ASN1_STRING type. A mask of permissible types is passed
12069 and it chooses the "minimal" type to use or an error if not type
12073 *) Add function equivalents to the various macros in asn1.h. The old
12074 macros are retained with an M_ prefix. Code inside the library can
12075 use the M_ macros. External code (including the openssl utility)
12076 should *NOT* in order to be "shared library friendly".
12079 *) Add various functions that can check a certificate's extensions
12080 to see if it usable for various purposes such as SSL client,
12081 server or S/MIME and CAs of these types. This is currently
12082 VERY EXPERIMENTAL but will ultimately be used for certificate chain
12083 verification. Also added a -purpose flag to x509 utility to
12084 print out all the purposes.
12087 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
12091 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
12092 for, obtain and decode and extension and obtain its critical flag.
12093 This allows all the necessary extension code to be handled in a
12094 single function call.
12097 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
12098 platforms. See crypto/rc4/rc4_enc.c for further details.
12101 *) New -noout option to asn1parse. This causes no output to be produced
12102 its main use is when combined with -strparse and -out to extract data
12103 from a file (which may not be in ASN.1 format).
12106 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
12107 when producing the local key id.
12108 [Richard Levitte <levitte@stacken.kth.se>]
12110 *) New option -dhparam in s_server. This allows a DH parameter file to be
12111 stated explicitly. If it is not stated then it tries the first server
12112 certificate file. The previous behaviour hard coded the filename
12116 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
12117 a public key to be input or output. For example:
12118 openssl rsa -in key.pem -pubout -out pubkey.pem
12119 Also added necessary DSA public key functions to handle this.
12122 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
12123 in the message. This was handled by allowing
12124 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
12125 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
12127 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
12128 to the end of the strings whereas this didn't. This would cause problems
12129 if strings read with d2i_ASN1_bytes() were later modified.
12130 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
12132 *) Fix for base64 decode bug. When a base64 bio reads only one line of
12133 data and it contains EOF it will end up returning an error. This is
12134 caused by input 46 bytes long. The cause is due to the way base64
12135 BIOs find the start of base64 encoded data. They do this by trying a
12136 trial decode on each line until they find one that works. When they
12137 do a flag is set and it starts again knowing it can pass all the
12138 data directly through the decoder. Unfortunately it doesn't reset
12139 the context it uses. This means that if EOF is reached an attempt
12140 is made to pass two EOFs through the context and this causes the
12141 resulting error. This can also cause other problems as well. As is
12142 usual with these problems it takes *ages* to find and the fix is
12143 trivial: move one line.
12144 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
12146 *) Ugly workaround to get s_client and s_server working under Windows. The
12147 old code wouldn't work because it needed to select() on sockets and the
12148 tty (for keypresses and to see if data could be written). Win32 only
12149 supports select() on sockets so we select() with a 1s timeout on the
12150 sockets and then see if any characters are waiting to be read, if none
12151 are present then we retry, we also assume we can always write data to
12152 the tty. This isn't nice because the code then blocks until we've
12153 received a complete line of data and it is effectively polling the
12154 keyboard at 1s intervals: however it's quite a bit better than not
12155 working at all :-) A dedicated Windows application might handle this
12156 with an event loop for example.
12159 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
12160 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
12161 will be called when RSA_sign() and RSA_verify() are used. This is useful
12162 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
12163 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
12164 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
12165 This necessitated the support of an extra signature type NID_md5_sha1
12166 for SSL signatures and modifications to the SSL library to use it instead
12167 of calling RSA_public_decrypt() and RSA_private_encrypt().
12170 *) Add new -verify -CAfile and -CApath options to the crl program, these
12171 will lookup a CRL issuers certificate and verify the signature in a
12172 similar way to the verify program. Tidy up the crl program so it
12173 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
12174 less strict. It will now permit CRL extensions even if it is not
12175 a V2 CRL: this will allow it to tolerate some broken CRLs.
12178 *) Initialize all non-automatic variables each time one of the openssl
12179 sub-programs is started (this is necessary as they may be started
12180 multiple times from the "OpenSSL>" prompt).
12181 [Lennart Bang, Bodo Moeller]
12183 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
12184 removing all other RSA functionality (this is what NO_RSA does). This
12185 is so (for example) those in the US can disable those operations covered
12186 by the RSA patent while allowing storage and parsing of RSA keys and RSA
12190 *) Non-copying interface to BIO pairs.
12191 (still largely untested)
12194 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
12195 ASCII string. This was handled independently in various places before.
12198 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
12199 UTF8 strings a character at a time.
12202 *) Use client_version from client hello to select the protocol
12203 (s23_srvr.c) and for RSA client key exchange verification
12204 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
12207 *) Add various utility functions to handle SPKACs, these were previously
12208 handled by poking round in the structure internals. Added new function
12209 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
12210 print, verify and generate SPKACs. Based on an original idea from
12211 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
12214 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
12217 *) Allow the config file extension section to be overwritten on the
12218 command line. Based on an original idea from Massimiliano Pala
12219 <madwolf@comune.modena.it>. The new option is called -extensions
12220 and can be applied to ca, req and x509. Also -reqexts to override
12221 the request extensions in req and -crlexts to override the crl extensions
12225 *) Add new feature to the SPKAC handling in ca. Now you can include
12226 the same field multiple times by preceding it by "XXXX." for example:
12229 this is the same syntax as used in the req config file.
12232 *) Allow certificate extensions to be added to certificate requests. These
12233 are specified in a 'req_extensions' option of the req section of the
12234 config file. They can be printed out with the -text option to req but
12235 are otherwise ignored at present.
12238 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
12239 data read consists of only the final block it would not decrypted because
12240 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
12241 A misplaced 'break' also meant the decrypted final block might not be
12242 copied until the next read.
12245 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
12246 a few extra parameters to the DH structure: these will be useful if
12247 for example we want the value of 'q' or implement X9.42 DH.
12250 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
12251 provides hooks that allow the default DSA functions or functions on a
12252 "per key" basis to be replaced. This allows hardware acceleration and
12253 hardware key storage to be handled without major modification to the
12254 library. Also added low level modexp hooks and CRYPTO_EX structure and
12255 associated functions.
12258 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
12259 as "read only": it can't be written to and the buffer it points to will
12260 not be freed. Reading from a read only BIO is much more efficient than
12261 a normal memory BIO. This was added because there are several times when
12262 an area of memory needs to be read from a BIO. The previous method was
12263 to create a memory BIO and write the data to it, this results in two
12264 copies of the data and an O(n^2) reading algorithm. There is a new
12265 function BIO_new_mem_buf() which creates a read only memory BIO from
12266 an area of memory. Also modified the PKCS#7 routines to use read only
12270 *) Bugfix: ssl23_get_client_hello did not work properly when called in
12271 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
12272 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
12273 but a retry condition occurred while trying to read the rest.
12276 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
12277 NID_pkcs7_encrypted by default: this was wrong since this should almost
12278 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
12279 the encrypted data type: this is a more sensible place to put it and it
12280 allows the PKCS#12 code to be tidied up that duplicated this
12284 *) Changed obj_dat.pl script so it takes its input and output files on
12285 the command line. This should avoid shell escape redirection problems
12289 *) Initial support for certificate extension requests, these are included
12290 in things like Xenroll certificate requests. Included functions to allow
12291 extensions to be obtained and added.
12294 *) -crlf option to s_client and s_server for sending newlines as
12295 CRLF (as required by many protocols).
12298 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
12300 *) Install libRSAglue.a when OpenSSL is built with RSAref.
12301 [Ralf S. Engelschall]
12303 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
12304 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
12306 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
12310 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
12311 DH parameters/keys (q is lost during that conversion, but the resulting
12312 DH parameters contain its length).
12314 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
12315 much faster than DH_generate_parameters (which creates parameters
12316 where p = 2*q + 1), and also the smaller q makes DH computations
12317 much more efficient (160-bit exponentiation instead of 1024-bit
12318 exponentiation); so this provides a convenient way to support DHE
12319 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
12320 utter importance to use
12321 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12323 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12324 when such DH parameters are used, because otherwise small subgroup
12325 attacks may become possible!
12328 *) Avoid memory leak in i2d_DHparams.
12331 *) Allow the -k option to be used more than once in the enc program:
12332 this allows the same encrypted message to be read by multiple recipients.
12335 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
12336 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
12337 it will always use the numerical form of the OID, even if it has a short
12341 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
12342 method only got called if p,q,dmp1,dmq1,iqmp components were present,
12343 otherwise bn_mod_exp was called. In the case of hardware keys for example
12344 no private key components need be present and it might store extra data
12345 in the RSA structure, which cannot be accessed from bn_mod_exp.
12346 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
12347 private key operations.
12350 *) Added support for SPARC Linux.
12353 *) pem_password_cb function type incompatibly changed from
12354 typedef int pem_password_cb(char *buf, int size, int rwflag);
12356 ....(char *buf, int size, int rwflag, void *userdata);
12357 so that applications can pass data to their callbacks:
12358 The PEM[_ASN1]_{read,write}... functions and macros now take an
12359 additional void * argument, which is just handed through whenever
12360 the password callback is called.
12361 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
12363 New function SSL_CTX_set_default_passwd_cb_userdata.
12365 Compatibility note: As many C implementations push function arguments
12366 onto the stack in reverse order, the new library version is likely to
12367 interoperate with programs that have been compiled with the old
12368 pem_password_cb definition (PEM_whatever takes some data that
12369 happens to be on the stack as its last argument, and the callback
12370 just ignores this garbage); but there is no guarantee whatsoever that
12373 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
12374 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
12375 problems not only on Windows, but also on some Unix platforms.
12376 To avoid problematic command lines, these definitions are now in an
12377 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
12378 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
12381 *) MIPS III/IV assembler module is reimplemented.
12384 *) More DES library cleanups: remove references to srand/rand and
12385 delete an unused file.
12388 *) Add support for the free Netwide assembler (NASM) under Win32,
12389 since not many people have MASM (ml) and it can be hard to obtain.
12390 This is currently experimental but it seems to work OK and pass all
12391 the tests. Check out INSTALL.W32 for info.
12394 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
12395 without temporary keys kept an extra copy of the server key,
12396 and connections with temporary keys did not free everything in case
12400 *) New function RSA_check_key and new openssl rsa option -check
12401 for verifying the consistency of RSA keys.
12402 [Ulf Moeller, Bodo Moeller]
12404 *) Various changes to make Win32 compile work:
12405 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
12406 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
12407 comparison" warnings.
12408 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
12411 *) Add a debugging option to PKCS#5 v2 key generation function: when
12412 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
12413 derived keys are printed to stderr.
12416 *) Copy the flags in ASN1_STRING_dup().
12417 [Roman E. Pavlov <pre@mo.msk.ru>]
12419 *) The x509 application mishandled signing requests containing DSA
12420 keys when the signing key was also DSA and the parameters didn't match.
12422 It was supposed to omit the parameters when they matched the signing key:
12423 the verifying software was then supposed to automatically use the CA's
12424 parameters if they were absent from the end user certificate.
12426 Omitting parameters is no longer recommended. The test was also
12427 the wrong way round! This was probably due to unusual behaviour in
12428 EVP_cmp_parameters() which returns 1 if the parameters match.
12429 This meant that parameters were omitted when they *didn't* match and
12430 the certificate was useless. Certificates signed with 'ca' didn't have
12432 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
12434 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
12435 The interface is as follows:
12436 Applications can use
12437 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
12438 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
12439 "off" is now the default.
12440 The library internally uses
12441 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
12442 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
12443 to disable memory-checking temporarily.
12445 Some inconsistent states that previously were possible (and were
12446 even the default) are now avoided.
12448 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
12449 with each memory chunk allocated; this is occasionally more helpful
12450 than just having a counter.
12452 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
12454 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
12458 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
12459 which largely parallels "options", but is for changing API behaviour,
12460 whereas "options" are about protocol behaviour.
12461 Initial "mode" flags are:
12463 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
12464 a single record has been written.
12465 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
12466 retries use the same buffer location.
12467 (But all of the contents must be
12471 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
12474 *) Fix problems with no-hmac etc.
12475 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
12477 *) New functions RSA_get_default_method(), RSA_set_method() and
12478 RSA_get_method(). These allows replacement of RSA_METHODs without having
12479 to mess around with the internals of an RSA structure.
12482 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
12483 Also really enable memory leak checks in openssl.c and in some
12485 [Chad C. Mulligan, Bodo Moeller]
12487 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
12488 up the length of negative integers. This has now been simplified to just
12489 store the length when it is first determined and use it later, rather
12490 than trying to keep track of where data is copied and updating it to
12492 [Steve Henson, reported by Brien Wheeler
12493 <bwheeler@authentica-security.com>]
12495 *) Add a new function PKCS7_signatureVerify. This allows the verification
12496 of a PKCS#7 signature but with the signing certificate passed to the
12497 function itself. This contrasts with PKCS7_dataVerify which assumes the
12498 certificate is present in the PKCS#7 structure. This isn't always the
12499 case: certificates can be omitted from a PKCS#7 structure and be
12500 distributed by "out of band" means (such as a certificate database).
12503 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12504 function prototypes in pem.h, also change util/mkdef.pl to add the
12505 necessary function names.
12508 *) mk1mf.pl (used by Windows builds) did not properly read the
12509 options set by Configure in the top level Makefile, and Configure
12510 was not even able to write more than one option correctly.
12511 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12514 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12515 file to be loaded from a BIO or FILE pointer. The BIO version will
12516 for example allow memory BIOs to contain config info.
12519 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12520 Whoever hopes to achieve shared-library compatibility across versions
12521 must use this, not the compile-time macro.
12522 (Exercise 0.9.4: Which is the minimum library version required by
12524 Note: All this applies only to multi-threaded programs, others don't
12528 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12529 through a BIO pair triggered the default case, i.e.
12530 SSLerr(...,SSL_R_UNKNOWN_STATE).
12533 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12534 can use the SSL library even if none of the specific BIOs is
12538 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12539 for the encoded length.
12540 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12542 *) Add initial documentation of the X509V3 functions.
12545 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12546 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12547 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12548 secure PKCS#8 private key format with a high iteration count.
12551 *) Fix determination of Perl interpreter: A perl or perl5
12552 _directory_ in $PATH was also accepted as the interpreter.
12553 [Ralf S. Engelschall]
12555 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12556 wrong with it but it was very old and did things like calling
12557 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12558 unusual formatting.
12561 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12562 to use the new extension code.
12565 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12566 with macros. This should make it easier to change their form, add extra
12567 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12571 *) Add to configuration table a new entry that can specify an alternative
12572 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12573 according to Mark Crispin <MRC@Panda.COM>.
12577 *) DES CBC did not update the IV. Weird.
12580 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12581 Changing the behaviour of the former might break existing programs --
12582 where IV updating is needed, des_ncbc_encrypt can be used.
12585 *) When bntest is run from "make test" it drives bc to check its
12586 calculations, as well as internally checking them. If an internal check
12587 fails, it needs to cause bc to give a non-zero result or make test carries
12588 on without noticing the failure. Fixed.
12591 *) DES library cleanups.
12594 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12595 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12596 ciphers. NOTE: although the key derivation function has been verified
12597 against some published test vectors it has not been extensively tested
12598 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12602 *) Instead of "mkdir -p", which is not fully portable, use new
12603 Perl script "util/mkdir-p.pl".
12606 *) Rewrite the way password based encryption (PBE) is handled. It used to
12607 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12608 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12609 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12610 the 'parameter' field of the AlgorithmIdentifier is passed to the
12611 underlying key generation function so it must do its own ASN1 parsing.
12612 This has also changed the EVP_PBE_CipherInit() function which now has a
12613 'parameter' argument instead of literal salt and iteration count values
12614 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12617 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12618 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12619 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12620 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12621 value was just used as a "magic string" and not used directly its
12622 value doesn't matter.
12625 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12629 *) "linux-sparc64" configuration (ultrapenguin).
12630 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12631 "linux-sparc" configuration.
12632 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12634 *) config now generates no-xxx options for missing ciphers.
12637 *) Support the EBCDIC character set (work in progress).
12638 File ebcdic.c not yet included because it has a different license.
12639 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12641 *) Support BS2000/OSD-POSIX.
12642 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12644 *) Make callbacks for key generation use void * instead of char *.
12647 *) Make S/MIME samples compile (not yet tested).
12650 *) Additional typesafe stacks.
12653 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12657 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12659 *) New configuration variant "sco5-gcc".
12661 *) Updated some demos.
12662 [Sean O Riordain, Wade Scholine]
12664 *) Add missing BIO_free at exit of pkcs12 application.
12667 *) Fix memory leak in conf.c.
12670 *) Updates for Win32 to assembler version of MD5.
12673 *) Set #! path to perl in apps/der_chop to where we found it
12674 instead of using a fixed path.
12677 *) SHA library changes for irix64-mips4-cc.
12680 *) Improvements for VMS support.
12684 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12686 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12687 This also avoids the problems with SC4.2 and unpatched SC5.
12688 [Andy Polyakov <appro@fy.chalmers.se>]
12690 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12691 These are required because of the typesafe stack would otherwise break
12692 existing code. If old code used a structure member which used to be STACK
12693 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12694 sk_num or sk_value it would produce an error because the num, data members
12695 are not present in STACK_OF. Now it just produces a warning. sk_set
12696 replaces the old method of assigning a value to sk_value
12697 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12698 that does this will no longer work (and should use sk_set instead) but
12699 this could be regarded as a "questionable" behaviour anyway.
12702 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12703 correctly handle encrypted S/MIME data.
12706 *) Change type of various DES function arguments from des_cblock
12707 (which means, in function argument declarations, pointer to char)
12708 to des_cblock * (meaning pointer to array with 8 char elements),
12709 which allows the compiler to do more typechecking; it was like
12710 that back in SSLeay, but with lots of ugly casts.
12712 Introduce new type const_des_cblock.
12715 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12716 problems: find RecipientInfo structure that matches recipient certificate
12717 and initialise the ASN1 structures properly based on passed cipher.
12720 *) Belatedly make the BN tests actually check the results.
12723 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12724 to and from BNs: it was completely broken. New compilation option
12725 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12726 key elements as negative integers.
12729 *) Reorganize and speed up MD5.
12730 [Andy Polyakov <appro@fy.chalmers.se>]
12733 [Richard Levitte <richard@levitte.org>]
12735 *) New option -out to asn1parse to allow the parsed structure to be
12736 output to a file. This is most useful when combined with the -strparse
12737 option to examine the output of things like OCTET STRINGS.
12740 *) Make SSL library a little more fool-proof by not requiring any longer
12741 that SSL_set_{accept,connect}_state be called before
12742 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12743 in many applications because usually everything *appeared* to work as
12744 intended anyway -- now it really works as intended).
12747 *) Move openssl.cnf out of lib/.
12750 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12751 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12752 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12753 [Ralf S. Engelschall]
12755 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12756 handle PKCS#7 enveloped data properly.
12757 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12759 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12760 copying pointers. The cert_st handling is changed by this in
12761 various ways (and thus what used to be known as ctx->default_cert
12762 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12763 any longer when s->cert does not give us what we need).
12764 ssl_cert_instantiate becomes obsolete by this change.
12765 As soon as we've got the new code right (possibly it already is?),
12766 we have solved a couple of bugs of the earlier code where s->cert
12767 was used as if it could not have been shared with other SSL structures.
12769 Note that using the SSL API in certain dirty ways now will result
12770 in different behaviour than observed with earlier library versions:
12771 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12772 does not influence s as it used to.
12774 In order to clean up things more thoroughly, inside SSL_SESSION
12775 we don't use CERT any longer, but a new structure SESS_CERT
12776 that holds per-session data (if available); currently, this is
12777 the peer's certificate chain and, for clients, the server's certificate
12778 and temporary key. CERT holds only those values that can have
12779 meaningful defaults in an SSL_CTX.
12782 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12783 from the internal representation. Various PKCS#7 fixes: remove some
12784 evil casts and set the enc_dig_alg field properly based on the signing
12788 *) Allow PKCS#12 password to be set from the command line or the
12789 environment. Let 'ca' get its config file name from the environment
12790 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12794 *) Allow certificate policies extension to use an IA5STRING for the
12795 organization field. This is contrary to the PKIX definition but
12796 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12800 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12801 without disallowing inline assembler and the like for non-pedantic builds.
12804 *) Support Borland C++ builder.
12805 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12807 *) Support Mingw32.
12810 *) SHA-1 cleanups and performance enhancements.
12811 [Andy Polyakov <appro@fy.chalmers.se>]
12813 *) Sparc v8plus assembler for the bignum library.
12814 [Andy Polyakov <appro@fy.chalmers.se>]
12816 *) Accept any -xxx and +xxx compiler options in Configure.
12819 *) Update HPUX configuration.
12822 *) Add missing sk_<type>_unshift() function to safestack.h
12823 [Ralf S. Engelschall]
12825 *) New function SSL_CTX_use_certificate_chain_file that sets the
12826 "extra_cert"s in addition to the certificate. (This makes sense
12827 only for "PEM" format files, as chains as a whole are not
12831 *) Support verify_depth from the SSL API.
12832 x509_vfy.c had what can be considered an off-by-one-error:
12833 Its depth (which was not part of the external interface)
12834 was actually counting the number of certificates in a chain;
12835 now it really counts the depth.
12838 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12839 instead of X509err, which often resulted in confusing error
12840 messages since the error codes are not globally unique
12841 (e.g. an alleged error in ssl3_accept when a certificate
12842 didn't match the private key).
12844 *) New function SSL_CTX_set_session_id_context that allows to set a default
12845 value (so that you don't need SSL_set_session_id_context for each
12846 connection using the SSL_CTX).
12849 *) OAEP decoding bug fix.
12852 *) Support INSTALL_PREFIX for package builders, as proposed by
12856 *) New Configure options "threads" and "no-threads". For systems
12857 where the proper compiler options are known (currently Solaris
12858 and Linux), "threads" is the default.
12861 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12864 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12865 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12866 such as /usr/local/bin.
12869 *) "make linux-shared" to build shared libraries.
12870 [Niels Poppe <niels@netbox.org>]
12872 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12875 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12876 extension adding in x509 utility.
12879 *) Remove NOPROTO sections and error code comments.
12882 *) Partial rewrite of the DEF file generator to now parse the ANSI
12886 *) New Configure options --prefix=DIR and --openssldir=DIR.
12889 *) Complete rewrite of the error code script(s). It is all now handled
12890 by one script at the top level which handles error code gathering,
12891 header rewriting and C source file generation. It should be much better
12892 than the old method: it now uses a modified version of Ulf's parser to
12893 read the ANSI prototypes in all header files (thus the old K&R definitions
12894 aren't needed for error creation any more) and do a better job of
12895 translating function codes into names. The old 'ASN1 error code embedded
12896 in a comment' is no longer necessary and it doesn't use .err files which
12897 have now been deleted. Also the error code call doesn't have to appear all
12898 on one line (which resulted in some large lines...).
12901 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12904 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12905 0 (which usually indicates a closed connection), but continue reading.
12908 *) Fix some race conditions.
12911 *) Add support for CRL distribution points extension. Add Certificate
12912 Policies and CRL distribution points documentation.
12915 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12918 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12919 8 of keying material. Merlin has also confirmed interop with this fix
12920 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12921 [Merlin Hughes <merlin@baltimore.ie>]
12923 *) Fix lots of warnings.
12924 [Richard Levitte <levitte@stacken.kth.se>]
12926 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12927 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12928 [Richard Levitte <levitte@stacken.kth.se>]
12930 *) Fix problems with sizeof(long) == 8.
12931 [Andy Polyakov <appro@fy.chalmers.se>]
12933 *) Change functions to ANSI C.
12936 *) Fix typos in error codes.
12937 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12939 *) Remove defunct assembler files from Configure.
12942 *) SPARC v8 assembler BIGNUM implementation.
12943 [Andy Polyakov <appro@fy.chalmers.se>]
12945 *) Support for Certificate Policies extension: both print and set.
12946 Various additions to support the r2i method this uses.
12949 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12950 return a const string when you are expecting an allocated buffer.
12953 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12954 types DirectoryString and DisplayText.
12957 *) Add code to allow r2i extensions to access the configuration database,
12958 add an LHASH database driver and add several ctx helper functions.
12961 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12962 fail when they extended the size of a BIGNUM.
12965 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12966 support typesafe stack.
12969 *) Fix typo in SSL_[gs]et_options().
12970 [Nils Frostberg <nils@medcom.se>]
12972 *) Delete various functions and files that belonged to the (now obsolete)
12973 old X509V3 handling code.
12976 *) New Configure option "rsaref".
12979 *) Don't auto-generate pem.h.
12982 *) Introduce type-safe ASN.1 SETs.
12985 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12986 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12988 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12989 that links with OpenSSL (well at least cause lots of warnings), but fear
12990 not: the conversion is trivial, and it eliminates loads of evil casts. A
12991 few STACKed things have been converted already. Feel free to convert more.
12992 In the fullness of time, I'll do away with the STACK type altogether.
12995 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12996 specified in <certfile> by updating the entry in the index.txt file.
12997 This way one no longer has to edit the index.txt file manually for
12998 revoking a certificate. The -revoke option does the gory details now.
12999 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
13001 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
13002 `-text' option at all and this way the `-noout -text' combination was
13003 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
13004 [Ralf S. Engelschall]
13006 *) Make sure a corresponding plain text error message exists for the
13007 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
13008 verify callback function determined that a certificate was revoked.
13009 [Ralf S. Engelschall]
13011 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
13012 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
13013 all available ciphers including rc5, which was forgotten until now.
13014 In order to let the testing shell script know which algorithms
13015 are available, a new (up to now undocumented) command
13016 "openssl list-cipher-commands" is used.
13019 *) Bugfix: s_client occasionally would sleep in select() when
13020 it should have checked SSL_pending() first.
13023 *) New functions DSA_do_sign and DSA_do_verify to provide access to
13024 the raw DSA values prior to ASN.1 encoding.
13027 *) Tweaks to Configure
13028 [Niels Poppe <niels@netbox.org>]
13030 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
13034 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
13037 *) New config option to avoid instructions that are illegal on the 80386.
13038 The default code is faster, but requires at least a 486.
13041 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
13042 SSL2_SERVER_VERSION (not used at all) macros, which are now the
13043 same as SSL2_VERSION anyway.
13046 *) New "-showcerts" option for s_client.
13049 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
13050 application. Various cleanups and fixes.
13053 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
13054 modify error routines to work internally. Add error codes and PBE init
13055 to library startup routines.
13058 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
13059 packing functions to asn1 and evp. Changed function names and error
13060 codes along the way.
13063 *) PKCS12 integration: and so it begins... First of several patches to
13064 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
13065 objects to objects.h
13068 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
13069 and display support for Thawte strong extranet extension.
13072 *) Add LinuxPPC support.
13073 [Jeff Dubrule <igor@pobox.org>]
13075 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
13076 bn_div_words in alpha.s.
13077 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
13079 *) Make sure the RSA OAEP test is skipped under -DRSAref because
13080 OAEP isn't supported when OpenSSL is built with RSAref.
13081 [Ulf Moeller <ulf@fitug.de>]
13083 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
13084 so they no longer are missing under -DNOPROTO.
13085 [Soren S. Jorvang <soren@t.dk>]
13088 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
13090 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
13091 doesn't work when the session is reused. Coming soon!
13094 *) Fix a security hole, that allows sessions to be reused in the wrong
13095 context thus bypassing client cert protection! All software that uses
13096 client certs and session caches in multiple contexts NEEDS PATCHING to
13097 allow session reuse! A fuller solution is in the works.
13098 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
13100 *) Some more source tree cleanups (removed obsolete files
13101 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
13102 permission on "config" script to be executable) and a fix for the INSTALL
13104 [Ulf Moeller <ulf@fitug.de>]
13106 *) Remove some legacy and erroneous uses of malloc, free instead of
13108 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
13110 *) Make rsa_oaep_test return non-zero on error.
13111 [Ulf Moeller <ulf@fitug.de>]
13113 *) Add support for native Solaris shared libraries. Configure
13114 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
13115 if someone would make that last step automatic.
13116 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
13118 *) ctx_size was not built with the right compiler during "make links". Fixed.
13121 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
13122 except NULL ciphers". This means the default cipher list will no longer
13123 enable NULL ciphers. They need to be specifically enabled e.g. with
13124 the string "DEFAULT:eNULL".
13127 *) Fix to RSA private encryption routines: if p < q then it would
13128 occasionally produce an invalid result. This will only happen with
13129 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
13132 *) Be less restrictive and allow also `perl util/perlpath.pl
13133 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
13134 because this way one can also use an interpreter named `perl5' (which is
13135 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
13136 installed as `perl').
13137 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13139 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
13140 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13142 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
13143 advapi32.lib to Win32 build and change the pem test comparison
13144 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
13145 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
13146 and crypto/des/ede_cbcm_enc.c.
13149 *) DES quad checksum was broken on big-endian architectures. Fixed.
13152 *) Comment out two functions in bio.h that aren't implemented. Fix up the
13153 Win32 test batch file so it (might) work again. The Win32 test batch file
13154 is horrible: I feel ill....
13157 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
13158 in e_os.h. Audit of header files to check ANSI and non ANSI
13159 sections: 10 functions were absent from non ANSI section and not exported
13160 from Windows DLLs. Fixed up libeay.num for new functions.
13163 *) Make `openssl version' output lines consistent.
13164 [Ralf S. Engelschall]
13166 *) Fix Win32 symbol export lists for BIO functions: Added
13167 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
13168 to ms/libeay{16,32}.def.
13169 [Ralf S. Engelschall]
13171 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
13172 fine under Unix and passes some trivial tests I've now added. But the
13173 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
13174 added to make sure no one expects that this stuff really works in the
13175 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
13176 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
13178 [Ralf S. Engelschall]
13180 *) Fix the generation of two part addresses in perl.
13181 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
13183 *) Add config entry for Linux on MIPS.
13184 [John Tobey <jtobey@channel1.com>]
13186 *) Make links whenever Configure is run, unless we are on Windoze.
13189 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
13190 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
13194 *) Add a useful kludge to allow package maintainers to specify compiler and
13195 other platforms details on the command line without having to patch the
13196 Configure script every time: One now can use ``perl Configure
13197 <id>:<details>'', i.e. platform ids are allowed to have details appended
13198 to them (separated by colons). This is treated as there would be a static
13199 pre-configured entry in Configure's %table under key <id> with value
13200 <details> and ``perl Configure <id>'' is called. So, when you want to
13201 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
13202 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
13203 now, which overrides the FreeBSD-elf entry on-the-fly.
13204 [Ralf S. Engelschall]
13206 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
13209 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
13210 on the `perl Configure ...' command line. This way one can compile
13211 OpenSSL libraries with Position Independent Code (PIC) which is needed
13212 for linking it into DSOs.
13213 [Ralf S. Engelschall]
13215 *) Remarkably, export ciphers were totally broken and no-one had noticed!
13219 *) Cleaned up the LICENSE document: The official contact for any license
13220 questions now is the OpenSSL core team under openssl-core@openssl.org.
13221 And add a paragraph about the dual-license situation to make sure people
13222 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
13223 to the OpenSSL toolkit.
13224 [Ralf S. Engelschall]
13226 *) General source tree makefile cleanups: Made `making xxx in yyy...'
13227 display consistent in the source tree and replaced `/bin/rm' by `rm'.
13228 Additionally cleaned up the `make links' target: Remove unnecessary
13229 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
13230 to speed processing and no longer clutter the display with confusing
13231 stuff. Instead only the actually done links are displayed.
13232 [Ralf S. Engelschall]
13234 *) Permit null encryption ciphersuites, used for authentication only. It used
13235 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
13236 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
13240 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
13241 signed attributes when verifying signatures (this would break them),
13242 the detached data encoding was wrong and public keys obtained using
13243 X509_get_pubkey() weren't freed.
13246 *) Add text documentation for the BUFFER functions. Also added a work around
13247 to a Win95 console bug. This was triggered by the password read stuff: the
13248 last character typed gets carried over to the next fread(). If you were
13249 generating a new cert request using 'req' for example then the last
13250 character of the passphrase would be CR which would then enter the first
13254 *) Added the new `Includes OpenSSL Cryptography Software' button as
13255 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
13256 button and can be used by applications based on OpenSSL to show the
13257 relationship to the OpenSSL project.
13258 [Ralf S. Engelschall]
13260 *) Remove confusing variables in function signatures in files
13261 ssl/ssl_lib.c and ssl/ssl.h.
13262 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13264 *) Don't install bss_file.c under PREFIX/include/
13265 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13267 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
13268 functions that return function pointers and has support for NT specific
13269 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
13270 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
13271 unsigned to signed types: this was killing the Win32 compile.
13274 *) Add new certificate file to stack functions,
13275 SSL_add_dir_cert_subjects_to_stack() and
13276 SSL_add_file_cert_subjects_to_stack(). These largely supplant
13277 SSL_load_client_CA_file(), and can be used to add multiple certs easily
13278 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
13279 This means that Apache-SSL and similar packages don't have to mess around
13280 to add as many CAs as they want to the preferred list.
13283 *) Experiment with doxygen documentation. Currently only partially applied to
13285 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
13286 openssl.doxy as the configuration file.
13289 *) Get rid of remaining C++-style comments which strict C compilers hate.
13290 [Ralf S. Engelschall, pointed out by Carlos Amengual]
13292 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
13293 compiled in by default: it has problems with large keys.
13296 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
13297 DH private keys and/or callback functions which directly correspond to
13298 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
13299 is needed for applications which have to configure certificates on a
13300 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
13302 For the RSA certificate situation is makes no difference, but
13303 for the DSA certificate situation this fixes the "no shared cipher"
13304 problem where the OpenSSL cipher selection procedure failed because the
13305 temporary keys were not overtaken from the context and the API provided
13306 no way to reconfigure them.
13307 The new functions now let applications reconfigure the stuff and they
13308 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
13309 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
13310 non-public-API function ssl_cert_instantiate() is used as a helper
13311 function and also to reduce code redundancy inside ssl_rsa.c.
13312 [Ralf S. Engelschall]
13314 *) Move s_server -dcert and -dkey options out of the undocumented feature
13315 area because they are useful for the DSA situation and should be
13316 recognized by the users.
13317 [Ralf S. Engelschall]
13319 *) Fix the cipher decision scheme for export ciphers: the export bits are
13320 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
13321 SSL_EXP_MASK. So, the original variable has to be used instead of the
13322 already masked variable.
13323 [Richard Levitte <levitte@stacken.kth.se>]
13325 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
13326 [Richard Levitte <levitte@stacken.kth.se>]
13328 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
13329 from `int' to `unsigned int' because it's a length and initialized by
13330 EVP_DigestFinal() which expects an `unsigned int *'.
13331 [Richard Levitte <levitte@stacken.kth.se>]
13333 *) Don't hard-code path to Perl interpreter on shebang line of Configure
13334 script. Instead use the usual Shell->Perl transition trick.
13335 [Ralf S. Engelschall]
13337 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
13338 (in addition to RSA certificates) to match the behaviour of `openssl dsa
13339 -noout -modulus' as it's already the case for `openssl rsa -noout
13340 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
13341 currently the public key is printed (a decision which was already done by
13342 `openssl dsa -modulus' in the past) which serves a similar purpose.
13343 Additionally the NO_RSA no longer completely removes the whole -modulus
13344 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
13346 [Ralf S. Engelschall]
13348 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
13349 BIO. See the source (crypto/evp/bio_ok.c) for more info.
13350 [Arne Ansper <arne@ats.cyber.ee>]
13352 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
13353 to be added. Now both 'req' and 'ca' can use new objects defined in the
13357 *) Add cool BIO that does syslog (or event log on NT).
13358 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
13360 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
13361 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
13362 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
13363 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
13366 *) Add preliminary config info for new extension code.
13369 *) Make RSA_NO_PADDING really use no padding.
13370 [Ulf Moeller <ulf@fitug.de>]
13372 *) Generate errors when private/public key check is done.
13375 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
13376 for some CRL extensions and new objects added.
13379 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
13380 key usage extension and fuller support for authority key id.
13383 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
13384 padding method for RSA, which is recommended for new applications in PKCS
13385 #1 v2.0 (RFC 2437, October 1998).
13386 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
13387 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
13388 against Bleichbacher's attack on RSA.
13389 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
13392 *) Updates to the new SSL compression code
13393 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13395 *) Fix so that the version number in the master secret, when passed
13396 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
13397 (because the server will not accept higher), that the version number
13398 is 0x03,0x01, not 0x03,0x00
13399 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13401 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
13402 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
13403 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
13406 *) Support for RAW extensions where an arbitrary extension can be
13407 created by including its DER encoding. See apps/openssl.cnf for
13411 *) Make sure latest Perl versions don't interpret some generated C array
13412 code as Perl array code in the crypto/err/err_genc.pl script.
13413 [Lars Weber <3weber@informatik.uni-hamburg.de>]
13415 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
13416 not many people have the assembler. Various Win32 compilation fixes and
13417 update to the INSTALL.W32 file with (hopefully) more accurate Win32
13418 build instructions.
13421 *) Modify configure script 'Configure' to automatically create crypto/date.h
13422 file under Win32 and also build pem.h from pem.org. New script
13423 util/mkfiles.pl to create the MINFO file on environments that can't do a
13424 'make files': perl util/mkfiles.pl >MINFO should work.
13427 *) Major rework of DES function declarations, in the pursuit of correctness
13428 and purity. As a result, many evil casts evaporated, and some weirdness,
13429 too. You may find this causes warnings in your code. Zapping your evil
13430 casts will probably fix them. Mostly.
13433 *) Fix for a typo in asn1.h. Bug fix to object creation script
13434 obj_dat.pl. It considered a zero in an object definition to mean
13435 "end of object": none of the objects in objects.h have any zeros
13436 so it wasn't spotted.
13437 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
13439 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
13440 Masking (CBCM). In the absence of test vectors, the best I have been able
13441 to do is check that the decrypt undoes the encrypt, so far. Send me test
13442 vectors if you have them.
13445 *) Correct calculation of key length for export ciphers (too much space was
13446 allocated for null ciphers). This has not been tested!
13449 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
13450 message is now correct (it understands "crypto" and "ssl" on its
13451 command line). There is also now an "update" option. This will update
13452 the util/ssleay.num and util/libeay.num files with any new functions.
13454 perl util/mkdef.pl crypto ssl update
13455 it will update them.
13458 *) Overhauled the Perl interface (perl/*):
13459 - ported BN stuff to OpenSSL's different BN library
13460 - made the perl/ source tree CVS-aware
13461 - renamed the package from SSLeay to OpenSSL (the files still contain
13462 their history because I've copied them in the repository)
13463 - removed obsolete files (the test scripts will be replaced
13464 by better Test::Harness variants in the future)
13465 [Ralf S. Engelschall]
13467 *) First cut for a very conservative source tree cleanup:
13468 1. merge various obsolete readme texts into doc/ssleay.txt
13469 where we collect the old documents and readme texts.
13470 2. remove the first part of files where I'm already sure that we no
13471 longer need them because of three reasons: either they are just temporary
13472 files which were left by Eric or they are preserved original files where
13473 I've verified that the diff is also available in the CVS via "cvs diff
13474 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
13475 the crypto/md/ stuff).
13476 [Ralf S. Engelschall]
13478 *) More extension code. Incomplete support for subject and issuer alt
13479 name, issuer and authority key id. Change the i2v function parameters
13480 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
13481 what that's for :-) Fix to ASN1 macro which messed up
13482 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
13485 *) Preliminary support for ENUMERATED type. This is largely copied from the
13489 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
13490 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13492 *) Make sure `make rehash' target really finds the `openssl' program.
13493 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13495 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
13496 like to hear about it if this slows down other processors.
13499 *) Add CygWin32 platform information to Configure script.
13500 [Alan Batie <batie@aahz.jf.intel.com>]
13502 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
13503 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13505 *) New program nseq to manipulate netscape certificate sequences
13508 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13512 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13513 but the BN code had some problems that would cause failures when
13514 doing certificate verification and some other functions.
13515 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13517 *) Add ASN1 and PEM code to support netscape certificate sequences.
13520 *) Add ASN1 and PEM code to support netscape certificate sequences.
13523 *) Add several PKIX and private extended key usage OIDs.
13526 *) Modify the 'ca' program to handle the new extension code. Modify
13527 openssl.cnf for new extension format, add comments.
13530 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13531 and add a sample to openssl.cnf so req -x509 now adds appropriate
13535 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13536 error code, add initial support to X509_print() and x509 application.
13539 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13540 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13541 stuff is currently isolated and isn't even compiled yet.
13544 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13545 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13546 Removed the versions check from X509 routines when loading extensions:
13547 this allows certain broken certificates that don't set the version
13548 properly to be processed.
13551 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13552 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13553 can still be regenerated with "make depend".
13556 *) Spelling mistake in C version of CAST-128.
13557 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13559 *) Changes to the error generation code. The perl script err-code.pl
13560 now reads in the old error codes and retains the old numbers, only
13561 adding new ones if necessary. It also only changes the .err files if new
13562 codes are added. The makefiles have been modified to only insert errors
13563 when needed (to avoid needlessly modifying header files). This is done
13564 by only inserting errors if the .err file is newer than the auto generated
13565 C file. To rebuild all the error codes from scratch (the old behaviour)
13566 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13567 or delete all the .err files.
13570 *) CAST-128 was incorrectly implemented for short keys. The C version has
13571 been fixed, but is untested. The assembler versions are also fixed, but
13572 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13573 to regenerate it if needed.
13574 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13575 Hagino <itojun@kame.net>]
13577 *) File was opened incorrectly in randfile.c.
13578 [Ulf Möller <ulf@fitug.de>]
13580 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13581 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13582 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13583 al: it's just almost always a UTCTime. Note this patch adds new error
13584 codes so do a "make errors" if there are problems.
13587 *) Correct Linux 1 recognition in config.
13588 [Ulf Möller <ulf@fitug.de>]
13590 *) Remove pointless MD5 hash when using DSA keys in ca.
13591 [Anonymous <nobody@replay.com>]
13593 *) Generate an error if given an empty string as a cert directory. Also
13594 generate an error if handed NULL (previously returned 0 to indicate an
13595 error, but didn't set one).
13596 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13598 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13601 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13602 parameters. This was causing a warning which killed off the Win32 compile.
13605 *) Remove C++ style comments from crypto/bn/bn_local.h.
13606 [Neil Costigan <neil.costigan@celocom.com>]
13608 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13609 based on a text string, looking up short and long names and finally
13610 "dot" format. The "dot" format stuff didn't work. Added new function
13611 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13612 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13613 OID is not part of the table.
13616 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13617 X509_LOOKUP_by_alias().
13620 *) Sort openssl functions by name.
13623 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13624 encryption from sample DSA keys (in case anyone is interested the password
13628 *) Make _all_ *_free functions accept a NULL pointer.
13629 [Frans Heymans <fheymans@isaserver.be>]
13631 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13633 [Anonymous <nobody@replay.com>]
13635 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13636 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13638 *) Don't blow it for numeric -newkey arguments to apps/req.
13639 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13641 *) Temp key "for export" tests were wrong in s3_srvr.c.
13642 [Anonymous <nobody@replay.com>]
13644 *) Add prototype for temp key callback functions
13645 SSL_CTX_set_tmp_{rsa,dh}_callback().
13648 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13649 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13652 *) X509_name_add_entry() freed the wrong thing after an error.
13653 [Arne Ansper <arne@ats.cyber.ee>]
13655 *) rsa_eay.c would attempt to free a NULL context.
13656 [Arne Ansper <arne@ats.cyber.ee>]
13658 *) BIO_s_socket() had a broken should_retry() on Windoze.
13659 [Arne Ansper <arne@ats.cyber.ee>]
13661 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13662 [Arne Ansper <arne@ats.cyber.ee>]
13664 *) Make sure the already existing X509_STORE->depth variable is initialized
13665 in X509_STORE_new(), but document the fact that this variable is still
13666 unused in the certificate verification process.
13667 [Ralf S. Engelschall]
13669 *) Fix the various library and apps files to free up pkeys obtained from
13670 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13673 *) Fix reference counting in X509_PUBKEY_get(). This makes
13674 demos/maurice/example2.c work, amongst others, probably.
13675 [Steve Henson and Ben Laurie]
13677 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13678 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13679 are no longer created. This way we have a single and consistent command
13680 line interface `openssl <command>', similar to `cvs <command>'.
13681 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13683 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13684 BIT STRING wrapper always have zero unused bits.
13687 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13690 *) Make the top-level INSTALL documentation easier to understand.
13693 *) Makefiles updated to exit if an error occurs in a sub-directory
13694 make (including if user presses ^C) [Paul Sutton]
13696 *) Make Montgomery context stuff explicit in RSA data structure.
13699 *) Fix build order of pem and err to allow for generated pem.h.
13702 *) Fix renumbering bug in X509_NAME_delete_entry().
13705 *) Enhanced the err-ins.pl script so it makes the error library number
13706 global and can add a library name. This is needed for external ASN1 and
13707 other error libraries.
13710 *) Fixed sk_insert which never worked properly.
13713 *) Fix ASN1 macros so they can handle indefinite length constructed
13714 EXPLICIT tags. Some non standard certificates use these: they can now
13718 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13719 into a single doc/ssleay.txt bundle. This way the information is still
13720 preserved but no longer messes up this directory. Now it's new room for
13721 the new set of documentation files.
13722 [Ralf S. Engelschall]
13724 *) SETs were incorrectly DER encoded. This was a major pain, because they
13725 shared code with SEQUENCEs, which aren't coded the same. This means that
13726 almost everything to do with SETs or SEQUENCEs has either changed name or
13727 number of arguments.
13728 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13730 *) Fix test data to work with the above.
13733 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13734 was already fixed by Eric for 0.9.1 it seems.
13735 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13737 *) Autodetect FreeBSD3.
13740 *) Fix various bugs in Configure. This affects the following platforms:
13744 unixware-2.0-pentium
13748 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13749 before they are needed.
13752 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13756 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13758 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13759 changed SSLeay to OpenSSL in version strings.
13760 [Ralf S. Engelschall]
13762 *) Some fixups to the top-level documents.
13765 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13766 because the symlink to include/ was missing.
13767 [Ralf S. Engelschall]
13769 *) Incorporated the popular no-RSA/DSA-only patches
13770 which allow to compile a RSA-free SSLeay.
13771 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13773 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13774 when "ssleay" is still not found.
13775 [Ralf S. Engelschall]
13777 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13778 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13780 *) Updated the README file.
13781 [Ralf S. Engelschall]
13783 *) Added various .cvsignore files in the CVS repository subdirs
13784 to make a "cvs update" really silent.
13785 [Ralf S. Engelschall]
13787 *) Recompiled the error-definition header files and added
13788 missing symbols to the Win32 linker tables.
13789 [Ralf S. Engelschall]
13791 *) Cleaned up the top-level documents;
13792 o new files: CHANGES and LICENSE
13793 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13794 o merged COPYRIGHT into LICENSE
13795 o removed obsolete TODO file
13796 o renamed MICROSOFT to INSTALL.W32
13797 [Ralf S. Engelschall]
13799 *) Removed dummy files from the 0.9.1b source tree:
13800 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13801 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13802 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13803 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13804 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13805 [Ralf S. Engelschall]
13807 *) Added various platform portability fixes.
13810 *) The Genesis of the OpenSSL rpject:
13811 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13812 Young and Tim J. Hudson created while they were working for C2Net until
13814 [The OpenSSL Project]
13817 Changes between 0.9.0b and 0.9.1b [not released]
13819 *) Updated a few CA certificates under certs/
13822 *) Changed some BIGNUM api stuff.
13825 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13826 DGUX x86, Linux Alpha, etc.
13829 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13830 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13834 *) Add -strparse option to asn1pars program which parses nested
13836 [Dr Stephen Henson <shenson@bigfoot.com>]
13838 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13841 *) DSA fix for "ca" program.
13844 *) Added "-genkey" option to "dsaparam" program.
13847 *) Added RIPE MD160 (rmd160) message digest.
13850 *) Added -a (all) option to "ssleay version" command.
13853 *) Added PLATFORM define which is the id given to Configure.
13856 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13859 *) Extended the ASN.1 parser routines.
13862 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13865 *) Added a BN_CTX to the BN library.
13868 *) Fixed the weak key values in DES library
13871 *) Changed API in EVP library for cipher aliases.
13874 *) Added support for RC2/64bit cipher.
13877 *) Converted the lhash library to the crypto/mem.c functions.
13880 *) Added more recognized ASN.1 object ids.
13883 *) Added more RSA padding checks for SSL/TLS.
13886 *) Added BIO proxy/filter functionality.
13889 *) Added extra_certs to SSL_CTX which can be used
13890 send extra CA certificates to the client in the CA cert chain sending
13891 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13894 *) Now Fortezza is denied in the authentication phase because
13895 this is key exchange mechanism is not supported by SSLeay at all.
13898 *) Additional PKCS1 checks.
13901 *) Support the string "TLSv1" for all TLS v1 ciphers.
13904 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13905 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13908 *) Fixed a few memory leaks.
13911 *) Fixed various code and comment typos.
13914 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13915 bytes sent in the client random.
13916 [Edward Bishop <ebishop@spyglass.com>]