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 [30 May 2023]
12 *) Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
13 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
15 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
16 numeric text form. For gigantic sub-identifiers, this would take a very
17 long time, the time complexity being O(n^2) where n is the size of that
18 sub-identifier. (CVE-2023-2650)
20 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
21 IDENTIFIER to canonical numeric text form if the size of that OBJECT
22 IDENTIFIER is 586 bytes or less, and fail otherwise.
24 The basis for this restriction is RFC 2578 (STD 58), section 3.5. OBJECT
25 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
26 most 128 sub-identifiers, and that the maximum value that each sub-
27 identifier may have is 2^32-1 (4294967295 decimal).
29 For each byte of every sub-identifier, only the 7 lower bits are part of
30 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
31 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
34 Ref: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
38 *) Reworked the Fix for the Timing Oracle in RSA Decryption (CVE-2022-4304).
39 The previous fix for this timing side channel turned out to cause
40 a severe 2-3x performance regression in the typical use case
41 compared to 1.1.1s. The new fix uses existing constant time
42 code paths, and restores the previous performance level while
43 fully eliminating all existing timing side channels.
44 The fix was developed by Bernd Edlinger with testing support
48 *) Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
49 that it does not enable policy checking. Thanks to
50 David Benjamin for discovering this issue. (CVE-2023-0466)
53 *) Fixed an issue where invalid certificate policies in leaf certificates are
54 silently ignored by OpenSSL and other certificate policy checks are skipped
55 for that certificate. A malicious CA could use this to deliberately assert
56 invalid certificate policies in order to circumvent policy checking on the
57 certificate altogether. (CVE-2023-0465)
60 *) Limited the number of nodes created in a policy tree to mitigate
61 against CVE-2023-0464. The default limit is set to 1000 nodes, which
62 should be sufficient for most installations. If required, the limit
63 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
64 time define to a desired maximum number of nodes or zero to allow
65 unlimited growth. (CVE-2023-0464)
68 Changes between 1.1.1s and 1.1.1t [7 Feb 2023]
70 *) Fixed X.400 address type confusion in X.509 GeneralName.
72 There is a type confusion vulnerability relating to X.400 address processing
73 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
74 but subsequently interpreted by GENERAL_NAME_cmp as an ASN1_TYPE. This
75 vulnerability may allow an attacker who can provide a certificate chain and
76 CRL (neither of which need have a valid signature) to pass arbitrary
77 pointers to a memcmp call, creating a possible read primitive, subject to
78 some constraints. Refer to the advisory for more information. Thanks to
79 David Benjamin for discovering this issue. (CVE-2023-0286)
81 This issue has been fixed by changing the public header file definition of
82 GENERAL_NAME so that x400Address reflects the implementation. It was not
83 possible for any existing application to successfully use the existing
84 definition; however, if any application references the x400Address field
85 (e.g. in dead code), note that the type of this field has changed. There is
89 *) Fixed Use-after-free following BIO_new_NDEF.
91 The public API function BIO_new_NDEF is a helper function used for
92 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
93 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
94 be called directly by end user applications.
96 The function receives a BIO from the caller, prepends a new BIO_f_asn1
97 filter BIO onto the front of it to form a BIO chain, and then returns
98 the new head of the BIO chain to the caller. Under certain conditions,
99 for example if a CMS recipient public key is invalid, the new filter BIO
100 is freed and the function returns a NULL result indicating a failure.
101 However, in this case, the BIO chain is not properly cleaned up and the
102 BIO passed by the caller still retains internal pointers to the previously
103 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
104 then a use-after-free will occur. This will most likely result in a crash.
106 [Viktor Dukhovni, Matt Caswell]
108 *) Fixed Double free after calling PEM_read_bio_ex.
110 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
111 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
112 data. If the function succeeds then the "name_out", "header" and "data"
113 arguments are populated with pointers to buffers containing the relevant
114 decoded data. The caller is responsible for freeing those buffers. It is
115 possible to construct a PEM file that results in 0 bytes of payload data.
116 In this case PEM_read_bio_ex() will return a failure code but will populate
117 the header argument with a pointer to a buffer that has already been freed.
118 If the caller also frees this buffer then a double free will occur. This
119 will most likely lead to a crash.
121 The functions PEM_read_bio() and PEM_read() are simple wrappers around
122 PEM_read_bio_ex() and therefore these functions are also directly affected.
124 These functions are also called indirectly by a number of other OpenSSL
125 functions including PEM_X509_INFO_read_bio_ex() and
126 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
127 internal uses of these functions are not vulnerable because the caller does
128 not free the header argument if PEM_read_bio_ex() returns a failure code.
130 [Kurt Roeckx, Matt Caswell]
132 *) Fixed Timing Oracle in RSA Decryption.
134 A timing based side channel exists in the OpenSSL RSA Decryption
135 implementation which could be sufficient to recover a plaintext across
136 a network in a Bleichenbacher style attack. To achieve a successful
137 decryption an attacker would have to be able to send a very large number
138 of trial messages for decryption. The vulnerability affects all RSA padding
139 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
141 [Dmitry Belyavsky, Hubert Kario]
143 Changes between 1.1.1r and 1.1.1s [1 Nov 2022]
145 *) Fixed a regression introduced in 1.1.1r version not refreshing the
146 certificate data to be signed before signing the certificate.
149 Changes between 1.1.1q and 1.1.1r [11 Oct 2022]
151 *) Fixed the linux-mips64 Configure target which was missing the
152 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
156 *) Fixed a strict aliasing problem in bn_nist. Clang-14 optimisation was
157 causing incorrect results in some cases as a result.
160 *) Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
161 report correct results in some cases
164 *) Fixed a regression introduced in 1.1.1o for re-signing certificates with
168 *) Added the loongarch64 target
171 *) Fixed a DRBG seed propagation thread safety issue
174 *) Fixed a memory leak in tls13_generate_secret
177 *) Fixed reported performance degradation on aarch64. Restored the
178 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
179 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
180 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
181 The new algorithm is still used for 32 bit targets.
184 *) Added a missing header for memcmp that caused compilation failure on some
188 Changes between 1.1.1p and 1.1.1q [5 Jul 2022]
190 *) AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
191 implementation would not encrypt the entirety of the data under some
192 circumstances. This could reveal sixteen bytes of data that was
193 preexisting in the memory that wasn't written. In the special case of
194 "in place" encryption, sixteen bytes of the plaintext would be revealed.
196 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
197 they are both unaffected.
199 [Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño]
201 Changes between 1.1.1o and 1.1.1p [21 Jun 2022]
203 *) In addition to the c_rehash shell command injection identified in
204 CVE-2022-1292, further bugs where the c_rehash script does not
205 properly sanitise shell metacharacters to prevent command injection have been
208 When the CVE-2022-1292 was fixed it was not discovered that there
209 are other places in the script where the file names of certificates
210 being hashed were possibly passed to a command executed through the shell.
212 This script is distributed by some operating systems in a manner where
213 it is automatically executed. On such operating systems, an attacker
214 could execute arbitrary commands with the privileges of the script.
216 Use of the c_rehash script is considered obsolete and should be replaced
217 by the OpenSSL rehash command line tool.
219 [Daniel Fiala, Tomáš Mráz]
221 *) When OpenSSL TLS client is connecting without any supported elliptic
222 curves and TLS-1.3 protocol is disabled the connection will no longer fail
223 if a ciphersuite that does not use a key exchange based on elliptic
224 curves can be negotiated.
227 Changes between 1.1.1n and 1.1.1o [3 May 2022]
229 *) Fixed a bug in the c_rehash script which was not properly sanitising shell
230 metacharacters to prevent command injection. This script is distributed
231 by some operating systems in a manner where it is automatically executed.
232 On such operating systems, an attacker could execute arbitrary commands
233 with the privileges of the script.
235 Use of the c_rehash script is considered obsolete and should be replaced
236 by the OpenSSL rehash command line tool.
240 Changes between 1.1.1m and 1.1.1n [15 Mar 2022]
242 *) Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
243 for non-prime moduli.
245 Internally this function is used when parsing certificates that contain
246 elliptic curve public keys in compressed form or explicit elliptic curve
247 parameters with a base point encoded in compressed form.
249 It is possible to trigger the infinite loop by crafting a certificate that
250 has invalid explicit curve parameters.
252 Since certificate parsing happens prior to verification of the certificate
253 signature, any process that parses an externally supplied certificate may
254 thus be subject to a denial of service attack. The infinite loop can also
255 be reached when parsing crafted private keys as they can contain explicit
256 elliptic curve parameters.
258 Thus vulnerable situations include:
260 - TLS clients consuming server certificates
261 - TLS servers consuming client certificates
262 - Hosting providers taking certificates or private keys from customers
263 - Certificate authorities parsing certification requests from subscribers
264 - Anything else which parses ASN.1 elliptic curve parameters
266 Also any other applications that use the BN_mod_sqrt() where the attacker
267 can control the parameter values are vulnerable to this DoS issue.
271 *) Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
272 to the list of ciphersuites providing Perfect Forward Secrecy as
273 required by SECLEVEL >= 3.
275 [Dmitry Belyavskiy, Nicola Tuveri]
277 Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
279 *) Avoid loading of a dynamic engine twice.
283 *) Fixed building on Debian with kfreebsd kernels
287 *) Prioritise DANE TLSA issuer certs over peer certs
291 *) Fixed random API for MacOS prior to 10.12
293 These MacOS versions don't support the CommonCrypto APIs
297 Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
299 *) Fixed an SM2 Decryption Buffer Overflow.
301 In order to decrypt SM2 encrypted data an application is expected to call the
302 API function EVP_PKEY_decrypt(). Typically an application will call this
303 function twice. The first time, on entry, the "out" parameter can be NULL and,
304 on exit, the "outlen" parameter is populated with the buffer size required to
305 hold the decrypted plaintext. The application can then allocate a sufficiently
306 sized buffer and call EVP_PKEY_decrypt() again, but this time passing a non-NULL
307 value for the "out" parameter.
309 A bug in the implementation of the SM2 decryption code means that the
310 calculation of the buffer size required to hold the plaintext returned by the
311 first call to EVP_PKEY_decrypt() can be smaller than the actual size required by
312 the second call. This can lead to a buffer overflow when EVP_PKEY_decrypt() is
313 called by the application a second time with a buffer that is too small.
315 A malicious attacker who is able present SM2 content for decryption to an
316 application could cause attacker chosen data to overflow the buffer by up to a
317 maximum of 62 bytes altering the contents of other data held after the
318 buffer, possibly changing application behaviour or causing the application to
319 crash. The location of the buffer is application dependent but is typically
324 *) Fixed various read buffer overruns processing ASN.1 strings
326 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
327 structure which contains a buffer holding the string data and a field holding
328 the buffer length. This contrasts with normal C strings which are repesented as
329 a buffer for the string data which is terminated with a NUL (0) byte.
331 Although not a strict requirement, ASN.1 strings that are parsed using OpenSSL's
332 own "d2i" functions (and other similar parsing functions) as well as any string
333 whose value has been set with the ASN1_STRING_set() function will additionally
334 NUL terminate the byte array in the ASN1_STRING structure.
336 However, it is possible for applications to directly construct valid ASN1_STRING
337 structures which do not NUL terminate the byte array by directly setting the
338 "data" and "length" fields in the ASN1_STRING array. This can also happen by
339 using the ASN1_STRING_set0() function.
341 Numerous OpenSSL functions that print ASN.1 data have been found to assume that
342 the ASN1_STRING byte array will be NUL terminated, even though this is not
343 guaranteed for strings that have been directly constructed. Where an application
344 requests an ASN.1 structure to be printed, and where that ASN.1 structure
345 contains ASN1_STRINGs that have been directly constructed by the application
346 without NUL terminating the "data" field, then a read buffer overrun can occur.
348 The same thing can also occur during name constraints processing of certificates
349 (for example if a certificate has been directly constructed by the application
350 instead of loading it via the OpenSSL parsing functions, and the certificate
351 contains non NUL terminated ASN1_STRING structures). It can also occur in the
352 X509_get1_email(), X509_REQ_get1_email() and X509_get1_ocsp() functions.
354 If a malicious actor can cause an application to directly construct an
355 ASN1_STRING and then process it through one of the affected OpenSSL functions
356 then this issue could be hit. This might result in a crash (causing a Denial of
357 Service attack). It could also result in the disclosure of private memory
358 contents (such as private keys, or sensitive plaintext).
362 Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
364 *) Fixed a problem with verifying a certificate chain when using the
365 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks
366 of the certificates present in a certificate chain. It is not set by
369 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
370 the chain that have explicitly encoded elliptic curve parameters was added
371 as an additional strict check.
373 An error in the implementation of this check meant that the result of a
374 previous check to confirm that certificates in the chain are valid CA
375 certificates was overwritten. This effectively bypasses the check
376 that non-CA certificates must not be able to issue other certificates.
378 If a "purpose" has been configured then there is a subsequent opportunity
379 for checks that the certificate is a valid CA. All of the named "purpose"
380 values implemented in libcrypto perform this check. Therefore, where
381 a purpose is set the certificate chain will still be rejected even when the
382 strict flag has been used. A purpose is set by default in libssl client and
383 server certificate verification routines, but it can be overridden or
384 removed by an application.
386 In order to be affected, an application must explicitly set the
387 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
388 for the certificate verification or, in the case of TLS client or server
389 applications, override the default purpose.
393 *) Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
394 crafted renegotiation ClientHello message from a client. If a TLSv1.2
395 renegotiation ClientHello omits the signature_algorithms extension (where
396 it was present in the initial ClientHello), but includes a
397 signature_algorithms_cert extension then a NULL pointer dereference will
398 result, leading to a crash and a denial of service attack.
400 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
401 (which is the default configuration). OpenSSL TLS clients are not impacted
404 [Peter Kästle and Samuel Sapalski]
406 Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
408 *) Fixed the X509_issuer_and_serial_hash() function. It attempts to
409 create a unique hash value based on the issuer and serial number data
410 contained within an X509 certificate. However it was failing to correctly
411 handle any errors that may occur while parsing the issuer field (which might
412 occur if the issuer field is maliciously constructed). This may subsequently
413 result in a NULL pointer deref and a crash leading to a potential denial of
418 *) Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
419 padding mode to correctly check for rollback attacks. This is considered a
420 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
424 *) Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
425 functions. Previously they could overflow the output length argument in some
426 cases where the input length is close to the maximum permissable length for
427 an integer on the platform. In such cases the return value from the function
428 call would be 1 (indicating success), but the output length value would be
429 negative. This could cause applications to behave incorrectly or crash.
433 *) Fixed SRP_Calc_client_key so that it runs in constant time. The previous
434 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
435 could be exploited in a side channel attack to recover the password. Since
436 the attack is local host only this is outside of the current OpenSSL
437 threat model and therefore no CVE is assigned.
439 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
443 Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
445 *) Fixed NULL pointer deref in the GENERAL_NAME_cmp function
446 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
447 If an attacker can control both items being compared then this could lead
448 to a possible denial of service attack. OpenSSL itself uses the
449 GENERAL_NAME_cmp function for two purposes:
450 1) Comparing CRL distribution point names between an available CRL and a
451 CRL distribution point embedded in an X509 certificate
452 2) When verifying that a timestamp response token signer matches the
453 timestamp authority name (exposed via the API functions
454 TS_RESP_verify_response and TS_RESP_verify_token)
458 *) Add support for Apple Silicon M1 Macs with the darwin64-arm64-cc target.
461 *) The security callback, which can be customised by application code, supports
462 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
463 in the "other" parameter. In most places this is what is passed. All these
464 places occur server side. However there was one client side call of this
465 security operation and it passed a DH object instead. This is incorrect
466 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
467 of the other locations. Therefore this client side call has been changed to
468 pass an EVP_PKEY instead.
471 *) In 1.1.1h, an expired trusted (root) certificate was not anymore rejected
472 when validating a certificate path. This check is restored in 1.1.1i.
475 Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
477 *) Certificates with explicit curve parameters are now disallowed in
478 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
481 *) The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
482 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
483 conversely, silently ignore DTLS protocol version bounds when configuring
484 TLS-based contexts. The commands can be repeated to set bounds of both
485 types. The same applies with the corresponding "min_protocol" and
486 "max_protocol" command-line switches, in case some application uses both TLS
489 SSL_CTX instances that are created for a fixed protocol version (e.g.
490 TLSv1_server_method()) also silently ignore version bounds. Previously
491 attempts to apply bounds to these protocol versions would result in an
492 error. Now only the "version-flexible" SSL_CTX instances are subject to
493 limits in configuration files in command-line options.
496 *) Handshake now fails if Extended Master Secret extension is dropped
500 *) Accidentally, an expired trusted (root) certificate is not anymore rejected
501 when validating a certificate path.
504 *) The Oracle Developer Studio compiler will start reporting deprecated APIs
506 Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
508 *) Fixed segmentation fault in SSL_check_chain()
509 Server or client applications that call the SSL_check_chain() function
510 during or after a TLS 1.3 handshake may crash due to a NULL pointer
511 dereference as a result of incorrect handling of the
512 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
513 or unrecognised signature algorithm is received from the peer. This could
514 be exploited by a malicious peer in a Denial of Service attack.
518 *) Added AES consttime code for no-asm configurations
519 an optional constant time support for AES was added
520 when building openssl for no-asm.
521 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
522 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
523 At this time this feature is by default disabled.
524 It will be enabled by default in 3.0.
527 Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
529 *) Revert the change of EOF detection while reading in libssl to avoid
530 regressions in applications depending on the current way of reporting
531 the EOF. As the existing method is not fully accurate the change to
532 reporting the EOF via SSL_ERROR_SSL is kept on the current development
533 branch and will be present in the 3.0 release.
536 *) Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
537 when primes for RSA keys are computed.
538 Since we previously always generated primes == 2 (mod 3) for RSA keys,
539 the 2-prime and 3-prime RSA modules were easy to distinguish, since
540 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
541 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
542 This avoids possible fingerprinting of newly generated RSA modules.
545 Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
546 *) Properly detect EOF while reading in libssl. Previously if we hit an EOF
547 while reading in libssl then we would report an error back to the
548 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
549 an error to the stack (which means we instead return SSL_ERROR_SSL) and
550 therefore give a hint as to what went wrong.
553 *) Check that ed25519 and ed448 are allowed by the security level. Previously
554 signature algorithms not using an MD were not being checked that they were
555 allowed by the security level.
558 *) Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
559 was not quite right. The behaviour was not consistent between resumption
560 and normal handshakes, and also not quite consistent with historical
561 behaviour. The behaviour in various scenarios has been clarified and
562 it has been updated to make it match historical behaviour as closely as
566 *) [VMS only] The header files that the VMS compilers include automatically,
567 __DECC_INCLUDE_PROLOGUE.H and __DECC_INCLUDE_EPILOGUE.H, use pragmas that
568 the C++ compiler doesn't understand. This is a shortcoming in the
569 compiler, but can be worked around with __cplusplus guards.
571 C++ applications that use OpenSSL libraries must be compiled using the
572 qualifier '/NAMES=(AS_IS,SHORTENED)' to be able to use all the OpenSSL
573 functions. Otherwise, only functions with symbols of less than 31
574 characters can be used, as the linker will not be able to successfully
575 resolve symbols with longer names.
578 *) Corrected the documentation of the return values from the EVP_DigestSign*
579 set of functions. The documentation mentioned negative values for some
580 errors, but this was never the case, so the mention of negative values
583 Code that followed the documentation and thereby check with something
584 like 'EVP_DigestSignInit(...) <= 0' will continue to work undisturbed.
587 *) Fixed an an overflow bug in the x64_64 Montgomery squaring procedure
588 used in exponentiation with 512-bit moduli. No EC algorithms are
589 affected. Analysis suggests that attacks against 2-prime RSA1024,
590 3-prime RSA1536, and DSA1024 as a result of this defect would be very
591 difficult to perform and are not believed likely. Attacks against DH512
592 are considered just feasible. However, for an attack the target would
593 have to re-use the DH512 private key, which is not recommended anyway.
594 Also applications directly using the low level API BN_mod_exp may be
595 affected if they use BN_FLG_CONSTTIME.
599 *) Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
600 The presence of this system service is determined at run-time.
603 *) Added newline escaping functionality to a filename when using openssl dgst.
604 This output format is to replicate the output format found in the '*sum'
605 checksum programs. This aims to preserve backward compatibility.
606 [Matt Eaton, Richard Levitte, and Paul Dale]
608 *) Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
612 Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
614 *) Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
615 number generator (RNG). This was intended to include protection in the
616 event of a fork() system call in order to ensure that the parent and child
617 processes did not share the same RNG state. However this protection was not
618 being used in the default case.
620 A partial mitigation for this issue is that the output from a high
621 precision timer is mixed into the RNG state so the likelihood of a parent
622 and child process sharing state is significantly reduced.
624 If an application already calls OPENSSL_init_crypto() explicitly using
625 OPENSSL_INIT_ATFORK then this problem does not occur at all.
627 [Matthias St. Pierre]
629 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is
630 used even when parsing explicit parameters, when loading a serialized key
631 or calling `EC_GROUP_new_from_ecpkparameters()`/
632 `EC_GROUP_new_from_ecparameters()`.
633 This prevents bypass of security hardening and performance gains,
634 especially for curves with specialized EC_METHODs.
635 By default, if a key encoded with explicit parameters is loaded and later
636 serialized, the output is still encoded with explicit parameters, even if
637 internally a "named" EC_GROUP is used for computation.
640 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before
641 this change, EC_GROUP_set_generator would accept order and/or cofactor as
642 NULL. After this change, only the cofactor parameter can be NULL. It also
643 does some minimal sanity checks on the passed order.
647 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
648 An attack is simple, if the first CMS_recipientInfo is valid but the
649 second CMS_recipientInfo is chosen ciphertext. If the second
650 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
651 encryption key will be replaced by garbage, and the message cannot be
652 decoded, but if the RSA decryption fails, the correct encryption key is
653 used and the recipient will not notice the attack.
654 As a work around for this potential attack the length of the decrypted
655 key must be equal to the cipher default key length, in case the
656 certifiate is not given and all recipientInfo are tried out.
657 The old behaviour can be re-enabled in the CMS code by setting the
658 CMS_DEBUG_DECRYPT flag.
662 *) Early start up entropy quality from the DEVRANDOM seed source has been
663 improved for older Linux systems. The RAND subsystem will wait for
664 /dev/random to be producing output before seeding from /dev/urandom.
665 The seeded state is stored for future library initialisations using
666 a system global shared memory segment. The shared memory identifier
667 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
668 the desired value. The default identifier is 114.
671 *) Correct the extended master secret constant on EBCDIC systems. Without this
672 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
673 negotiate EMS will fail. Unfortunately this also means that TLS connections
674 between EBCDIC systems with this fix, and EBCDIC systems without this
675 fix will fail if they negotiate EMS.
678 *) Use Windows installation paths in the mingw builds
680 Mingw isn't a POSIX environment per se, which means that Windows
681 paths should be used for installation.
685 *) Changed DH_check to accept parameters with order q and 2q subgroups.
686 With order 2q subgroups the bit 0 of the private key is not secret
687 but DH_generate_key works around that by clearing bit 0 of the
688 private key for those. This avoids leaking bit 0 of the private key.
691 *) Significantly reduce secure memory usage by the randomness pools.
694 *) Revert the DEVRANDOM_WAIT feature for Linux systems
696 The DEVRANDOM_WAIT feature added a select() call to wait for the
697 /dev/random device to become readable before reading from the
700 It turned out that this change had negative side effects on
701 performance which were not acceptable. After some discussion it
702 was decided to revert this feature and leave it up to the OS
703 resp. the platform maintainer to ensure a proper initialization
704 during early boot time.
705 [Matthias St. Pierre]
707 Changes between 1.1.1b and 1.1.1c [28 May 2019]
709 *) Add build tests for C++. These are generated files that only do one
710 thing, to include one public OpenSSL head file each. This tests that
711 the public header files can be usefully included in a C++ application.
713 This test isn't enabled by default. It can be enabled with the option
714 'enable-buildtest-c++'.
717 *) Enable SHA3 pre-hashing for ECDSA and DSA.
720 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
721 This changes the size when using the genpkey app when no size is given. It
722 fixes an omission in earlier changes that changed all RSA, DSA and DH
723 generation apps to use 2048 bits by default.
726 *) Reorganize the manual pages to consistently have RETURN VALUES,
727 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
728 util/fix-doc-nits accordingly.
729 [Paul Yang, Joshua Lock]
731 *) Add the missing accessor EVP_PKEY_get0_engine()
734 *) Have apps like 's_client' and 's_server' output the signature scheme
735 along with other cipher suite parameters when debugging.
738 *) Make OPENSSL_config() error agnostic again.
741 *) Do the error handling in RSA decryption constant time.
744 *) Prevent over long nonces in ChaCha20-Poly1305.
746 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
747 for every encryption operation. RFC 7539 specifies that the nonce value
748 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
749 and front pads the nonce with 0 bytes if it is less than 12
750 bytes. However it also incorrectly allows a nonce to be set of up to 16
751 bytes. In this case only the last 12 bytes are significant and any
752 additional leading bytes are ignored.
754 It is a requirement of using this cipher that nonce values are
755 unique. Messages encrypted using a reused nonce value are susceptible to
756 serious confidentiality and integrity attacks. If an application changes
757 the default nonce length to be longer than 12 bytes and then makes a
758 change to the leading bytes of the nonce expecting the new value to be a
759 new unique nonce then such an application could inadvertently encrypt
760 messages with a reused nonce.
762 Additionally the ignored bytes in a long nonce are not covered by the
763 integrity guarantee of this cipher. Any application that relies on the
764 integrity of these ignored leading bytes of a long nonce may be further
765 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
766 is safe because no such use sets such a long nonce value. However user
767 applications that use this cipher directly and set a non-default nonce
768 length to be longer than 12 bytes may be vulnerable.
770 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
775 *) Add DEVRANDOM_WAIT feature for Linux systems
777 On older Linux systems where the getrandom() system call is not available,
778 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
779 Contrary to getrandom(), the /dev/urandom device will not block during
780 early boot when the kernel CSPRNG has not been seeded yet.
782 To mitigate this known weakness, use select() to wait for /dev/random to
783 become readable before reading from /dev/urandom.
785 *) Ensure that SM2 only uses SM3 as digest algorithm
788 Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
790 *) Added SCA hardening for modular field inversion in EC_GROUP through
791 a new dedicated field_inv() pointer in EC_METHOD.
792 This also addresses a leakage affecting conversions from projective
793 to affine coordinates.
794 [Billy Bob Brumley, Nicola Tuveri]
796 *) Change the info callback signals for the start and end of a post-handshake
797 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
798 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
799 confused by this and assume that a TLSv1.2 renegotiation has started. This
800 can break KeyUpdate handling. Instead we no longer signal the start and end
801 of a post handshake message exchange (although the messages themselves are
802 still signalled). This could break some applications that were expecting
803 the old signals. However without this KeyUpdate is not usable for many
807 *) Fix a bug in the computation of the endpoint-pair shared secret used
808 by DTLS over SCTP. This breaks interoperability with older versions
809 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
810 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
811 interoperability with such broken implementations. However, enabling
812 this switch breaks interoperability with correct implementations.
814 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
815 re-used X509_PUBKEY object if the second PUBKEY is malformed.
818 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
821 *) Remove the 'dist' target and add a tarball building script. The
822 'dist' target has fallen out of use, and it shouldn't be
823 necessary to configure just to create a source distribution.
826 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
828 *) Timing vulnerability in DSA signature generation
830 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
831 timing side channel attack. An attacker could use variations in the signing
832 algorithm to recover the private key.
834 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
838 *) Timing vulnerability in ECDSA signature generation
840 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
841 timing side channel attack. An attacker could use variations in the signing
842 algorithm to recover the private key.
844 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
848 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
849 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
850 are retained for backwards compatibility.
853 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
854 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
855 of two gigabytes and the error handling improved.
857 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
858 categorized as a normal bug, not a security issue, because the DRBG reseeds
859 automatically and is fully functional even without additional randomness
860 provided by the application.
862 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
864 *) Add a new ClientHello callback. Provides a callback interface that gives
865 the application the ability to adjust the nascent SSL object at the
866 earliest stage of ClientHello processing, immediately after extensions have
867 been collected but before they have been processed. In particular, this
868 callback can adjust the supported TLS versions in response to the contents
872 *) Add SM2 base algorithm support.
875 *) s390x assembly pack: add (improved) hardware-support for the following
876 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
877 aes-cfb/cfb8, aes-ecb.
880 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
881 parameter is no longer accepted, as it leads to a corrupt table. NULL
882 pem_str is reserved for alias entries only.
885 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
886 step for prime curves. The new implementation is based on formulae from
887 differential addition-and-doubling in homogeneous projective coordinates
888 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
889 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
890 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
891 to work in projective coordinates.
892 [Billy Bob Brumley, Nicola Tuveri]
894 *) Change generating and checking of primes so that the error rate of not
895 being prime depends on the intended use based on the size of the input.
896 For larger primes this will result in more rounds of Miller-Rabin.
897 The maximal error rate for primes with more than 1080 bits is lowered
899 [Kurt Roeckx, Annie Yousar]
901 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
904 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
905 moving between systems, and to avoid confusion when a Windows build is
906 done with mingw vs with MSVC. For POSIX installs, there's still a
907 symlink or copy named 'tsget' to avoid that confusion as well.
910 *) Revert blinding in ECDSA sign and instead make problematic addition
911 length-invariant. Switch even to fixed-length Montgomery multiplication.
914 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
915 step for binary curves. The new implementation is based on formulae from
916 differential addition-and-doubling in mixed Lopez-Dahab projective
917 coordinates, modified to independently blind the operands.
918 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
920 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
921 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
922 EC_METHODs to implement their own specialized "ladder step", to take
923 advantage of more favorable coordinate systems or more efficient
924 differential addition-and-doubling algorithms.
925 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
927 *) Modified the random device based seed sources to keep the relevant
928 file descriptors open rather than reopening them on each access.
929 This allows such sources to operate in a chroot() jail without
930 the associated device nodes being available. This behaviour can be
931 controlled using RAND_keep_random_devices_open().
934 *) Numerous side-channel attack mitigations have been applied. This may have
935 performance impacts for some algorithms for the benefit of improved
936 security. Specific changes are noted in this change log by their respective
940 *) AIX shared library support overhaul. Switch to AIX "natural" way of
941 handling shared libraries, which means collecting shared objects of
942 different versions and bitnesses in one common archive. This allows to
943 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
944 doesn't affect the way 3rd party applications are linked, only how
945 multi-version installation is managed.
948 *) Make ec_group_do_inverse_ord() more robust and available to other
949 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
950 mitigations are applied to the fallback BN_mod_inverse().
951 When using this function rather than BN_mod_inverse() directly, new
952 EC cryptosystem implementations are then safer-by-default.
955 *) Add coordinate blinding for EC_POINT and implement projective
956 coordinate blinding for generic prime curves as a countermeasure to
957 chosen point SCA attacks.
958 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
960 *) Add blinding to ECDSA and DSA signatures to protect against side channel
961 attacks discovered by Keegan Ryan (NCC Group).
964 *) Enforce checking in the pkeyutl command line app to ensure that the input
965 length does not exceed the maximum supported digest length when performing
966 a sign, verify or verifyrecover operation.
969 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
970 I/O in combination with something like select() or poll() will hang. This
971 can be turned off again using SSL_CTX_clear_mode().
972 Many applications do not properly handle non-application data records, and
973 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
974 around the problems in those applications, but can also break some.
975 It's recommended to read the manpages about SSL_read(), SSL_write(),
976 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
977 SSL_CTX_set_read_ahead() again.
980 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
981 now allow empty (zero character) pass phrases.
984 *) Apply blinding to binary field modular inversion and remove patent
985 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
988 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
989 binary and prime elliptic curves.
992 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
993 constant time fixed point multiplication.
996 *) Revise elliptic curve scalar multiplication with timing attack
997 defenses: ec_wNAF_mul redirects to a constant time implementation
998 when computing fixed point and variable point multiplication (which
999 in OpenSSL are mostly used with secret scalars in keygen, sign,
1000 ECDH derive operations).
1001 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
1004 *) Updated CONTRIBUTING
1007 *) Updated DRBG / RAND to request nonce and additional low entropy
1008 randomness from the system.
1009 [Matthias St. Pierre]
1011 *) Updated 'openssl rehash' to use OpenSSL consistent default.
1014 *) Moved the load of the ssl_conf module to libcrypto, which helps
1015 loading engines that libssl uses before libssl is initialised.
1018 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1021 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1022 [Ingo Schwarze, Rich Salz]
1024 *) Added output of accepting IP address and port for 'openssl s_server'
1027 *) Added a new API for TLSv1.3 ciphersuites:
1028 SSL_CTX_set_ciphersuites()
1029 SSL_set_ciphersuites()
1032 *) Memory allocation failures consistently add an error to the error
1036 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1037 in libcrypto when run as setuid/setgid.
1040 *) Load any config file by default when libssl is used.
1043 *) Added new public header file <openssl/rand_drbg.h> and documentation
1044 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1045 [Matthias St. Pierre]
1047 *) QNX support removed (cannot find contributors to get their approval
1048 for the license change).
1051 *) TLSv1.3 replay protection for early data has been implemented. See the
1052 SSL_read_early_data() man page for further details.
1055 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1056 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1057 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1058 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1059 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1060 configuration has been separated out. See the ciphers man page or the
1061 SSL_CTX_set_ciphersuites() man page for more information.
1064 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1065 in responder mode now supports the new "-multi" option, which
1066 spawns the specified number of child processes to handle OCSP
1067 requests. The "-timeout" option now also limits the OCSP
1068 responder's patience to wait to receive the full client request
1069 on a newly accepted connection. Child processes are respawned
1070 as needed, and the CA index file is automatically reloaded
1071 when changed. This makes it possible to run the "ocsp" responder
1072 as a long-running service, making the OpenSSL CA somewhat more
1073 feature-complete. In this mode, most diagnostic messages logged
1074 after entering the event loop are logged via syslog(3) rather than
1078 *) Added support for X448 and Ed448. Heavily based on original work by
1082 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
1083 objects loaded. This adds the functions OSSL_STORE_expect() and
1084 OSSL_STORE_find() as well as needed tools to construct searches and
1085 get the search data out of them.
1088 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
1089 version of OpenSSL should review their configuration settings to ensure
1090 that they are still appropriate for TLSv1.3. For further information see:
1091 https://wiki.openssl.org/index.php/TLS1.3
1094 *) Grand redesign of the OpenSSL random generator
1096 The default RAND method now utilizes an AES-CTR DRBG according to
1097 NIST standard SP 800-90Ar1. The new random generator is essentially
1098 a port of the default random generator from the OpenSSL FIPS 2.0
1099 object module. It is a hybrid deterministic random bit generator
1100 using an AES-CTR bit stream and which seeds and reseeds itself
1101 automatically using trusted system entropy sources.
1103 Some of its new features are:
1104 o Support for multiple DRBG instances with seed chaining.
1105 o The default RAND method makes use of a DRBG.
1106 o There is a public and private DRBG instance.
1107 o The DRBG instances are fork-safe.
1108 o Keep all global DRBG instances on the secure heap if it is enabled.
1109 o The public and private DRBG instance are per thread for lock free
1111 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
1113 *) Changed Configure so it only says what it does and doesn't dump
1114 so much data. Instead, ./configdata.pm should be used as a script
1115 to display all sorts of configuration data.
1118 *) Added processing of "make variables" to Configure.
1121 *) Added SHA512/224 and SHA512/256 algorithm support.
1124 *) The last traces of Netware support, first removed in 1.1.0, have
1128 *) Get rid of Makefile.shared, and in the process, make the processing
1129 of certain files (rc.obj, or the .def/.map/.opt files produced from
1130 the ordinal files) more visible and hopefully easier to trace and
1131 debug (or make silent).
1134 *) Make it possible to have environment variable assignments as
1135 arguments to config / Configure.
1138 *) Add multi-prime RSA (RFC 8017) support.
1141 *) Add SM3 implemented according to GB/T 32905-2016
1142 [ Jack Lloyd <jack.lloyd@ribose.com>,
1143 Ronald Tse <ronald.tse@ribose.com>,
1144 Erick Borsboom <erick.borsboom@ribose.com> ]
1146 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
1147 as documented in RFC6066.
1148 Based on a patch from Tomasz Moń
1149 [Filipe Raimundo da Silva]
1151 *) Add SM4 implemented according to GB/T 32907-2016.
1152 [ Jack Lloyd <jack.lloyd@ribose.com>,
1153 Ronald Tse <ronald.tse@ribose.com>,
1154 Erick Borsboom <erick.borsboom@ribose.com> ]
1156 *) Reimplement -newreq-nodes and ERR_error_string_n; the
1157 original author does not agree with the license change.
1160 *) Add ARIA AEAD TLS support.
1163 *) Some macro definitions to support VS6 have been removed. Visual
1164 Studio 6 has not worked since 1.1.0
1167 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
1168 without clearing the errors.
1171 *) Add "atfork" functions. If building on a system that without
1172 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1173 requirements. The RAND facility now uses/requires this.
1179 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
1180 not possible to disable entirely. However, it's still possible to
1181 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1184 To disable, configure with 'no-ui-console'. 'no-ui' is still
1185 possible to use as an alias. Check at compile time with the
1186 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1187 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1190 *) Add a STORE module, which implements a uniform and URI based reader of
1191 stores that can contain keys, certificates, CRLs and numerous other
1192 objects. The main API is loosely based on a few stdio functions,
1193 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1194 OSSL_STORE_error and OSSL_STORE_close.
1195 The implementation uses backends called "loaders" to implement arbitrary
1196 URI schemes. There is one built in "loader" for the 'file' scheme.
1199 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
1200 then adjusted to work on FreeBSD 8.4 as well.
1201 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1202 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1205 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
1206 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1207 error code calls like this:
1209 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1211 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1212 that can be encoded in C. For the foreseeable future, this will only
1214 [Richard Levitte and Tim Hudson]
1216 *) Removed BSD cryptodev engine.
1219 *) Add a build target 'build_all_generated', to build all generated files
1220 and only that. This can be used to prepare everything that requires
1221 things like perl for a system that lacks perl and then move everything
1222 to that system and do the rest of the build there.
1225 *) In the UI interface, make it possible to duplicate the user data. This
1226 can be used by engines that need to retain the data for a longer time
1227 than just the call where this user data is passed.
1230 *) Ignore the '-named_curve auto' value for compatibility of applications
1232 [Tomas Mraz <tmraz@fedoraproject.org>]
1234 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1235 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1236 alerts across multiple records (some of which could be empty). In practice
1237 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1238 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1239 support this at all. Supporting it adds significant complexity to the
1240 record layer, and its removal is unlikely to cause interoperability
1244 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1245 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1246 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1250 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1251 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1252 [Richard Levitte, Andy Polyakov]
1254 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1258 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1259 platform rather than 'mingw'.
1262 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
1263 success if they are asked to add an object which already exists
1264 in the store. This change cascades to other functions which load
1265 certificates and CRLs.
1268 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
1269 facilitate stack unwinding even from assembly subroutines.
1272 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1273 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1276 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1277 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1278 which is the minimum version we support.
1281 *) Certificate time validation (X509_cmp_time) enforces stricter
1282 compliance with RFC 5280. Fractional seconds and timezone offsets
1283 are no longer allowed.
1286 *) Add support for ARIA
1289 *) s_client will now send the Server Name Indication (SNI) extension by
1290 default unless the new "-noservername" option is used. The server name is
1291 based on the host provided to the "-connect" option unless overridden by
1292 using "-servername".
1295 *) Add support for SipHash
1298 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1299 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1300 prevent issues where no progress is being made and the peer continually
1301 sends unrecognised record types, using up resources processing them.
1304 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
1305 using the algorithm defined in
1306 https://www.akkadia.org/drepper/SHA-crypt.txt
1309 *) Heartbeat support has been removed; the ABI is changed for now.
1310 [Richard Levitte, Rich Salz]
1312 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1315 *) The RSA "null" method, which was partially supported to avoid patent
1316 issues, has been replaced to always returns NULL.
1320 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
1322 *) Client DoS due to large DH parameter
1324 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1325 malicious server can send a very large prime value to the client. This will
1326 cause the client to spend an unreasonably long period of time generating a
1327 key for this prime resulting in a hang until the client has finished. This
1328 could be exploited in a Denial Of Service attack.
1330 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1334 *) Cache timing vulnerability in RSA Key Generation
1336 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1337 a cache timing side channel attack. An attacker with sufficient access to
1338 mount cache timing attacks during the RSA key generation process could
1339 recover the private key.
1341 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1342 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1346 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1347 parameter is no longer accepted, as it leads to a corrupt table. NULL
1348 pem_str is reserved for alias entries only.
1351 *) Revert blinding in ECDSA sign and instead make problematic addition
1352 length-invariant. Switch even to fixed-length Montgomery multiplication.
1355 *) Change generating and checking of primes so that the error rate of not
1356 being prime depends on the intended use based on the size of the input.
1357 For larger primes this will result in more rounds of Miller-Rabin.
1358 The maximal error rate for primes with more than 1080 bits is lowered
1360 [Kurt Roeckx, Annie Yousar]
1362 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1365 *) Add blinding to ECDSA and DSA signatures to protect against side channel
1366 attacks discovered by Keegan Ryan (NCC Group).
1369 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1370 now allow empty (zero character) pass phrases.
1373 *) Certificate time validation (X509_cmp_time) enforces stricter
1374 compliance with RFC 5280. Fractional seconds and timezone offsets
1375 are no longer allowed.
1378 *) Fixed a text canonicalisation bug in CMS
1380 Where a CMS detached signature is used with text content the text goes
1381 through a canonicalisation process first prior to signing or verifying a
1382 signature. This process strips trailing space at the end of lines, converts
1383 line terminators to CRLF and removes additional trailing line terminators
1384 at the end of a file. A bug in the canonicalisation process meant that
1385 some characters, such as form-feed, were incorrectly treated as whitespace
1386 and removed. This is contrary to the specification (RFC5485). This fix
1387 could mean that detached text data signed with an earlier version of
1388 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
1389 signed with a fixed OpenSSL may fail to verify with an earlier version of
1390 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
1391 and use the "-binary" flag (for the "cms" command line application) or set
1392 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
1395 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
1397 *) Constructed ASN.1 types with a recursive definition could exceed the stack
1399 Constructed ASN.1 types with a recursive definition (such as can be found
1400 in PKCS7) could eventually exceed the stack given malicious input with
1401 excessive recursion. This could result in a Denial Of Service attack. There
1402 are no such structures used within SSL/TLS that come from untrusted sources
1403 so this is considered safe.
1405 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
1410 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
1412 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
1413 effectively reduced to only comparing the least significant bit of each
1414 byte. This allows an attacker to forge messages that would be considered as
1415 authenticated in an amount of tries lower than that guaranteed by the
1416 security claims of the scheme. The module can only be compiled by the
1417 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
1419 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
1424 *) Add a build target 'build_all_generated', to build all generated files
1425 and only that. This can be used to prepare everything that requires
1426 things like perl for a system that lacks perl and then move everything
1427 to that system and do the rest of the build there.
1430 *) Backport SSL_OP_NO_RENGOTIATION
1432 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
1433 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
1434 changes this is no longer possible in 1.1.0. Therefore the new
1435 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
1436 1.1.0 to provide equivalent functionality.
1438 Note that if an application built against 1.1.0h headers (or above) is run
1439 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
1440 accepted but nothing will happen, i.e. renegotiation will not be prevented.
1443 *) Removed the OS390-Unix config target. It relied on a script that doesn't
1447 *) rsaz_1024_mul_avx2 overflow bug on x86_64
1449 There is an overflow bug in the AVX2 Montgomery multiplication procedure
1450 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
1451 Analysis suggests that attacks against RSA and DSA as a result of this
1452 defect would be very difficult to perform and are not believed likely.
1453 Attacks against DH1024 are considered just feasible, because most of the
1454 work necessary to deduce information about a private key may be performed
1455 offline. The amount of resources required for such an attack would be
1456 significant. However, for an attack on TLS to be meaningful, the server
1457 would have to share the DH1024 private key among multiple clients, which is
1458 no longer an option since CVE-2016-0701.
1460 This only affects processors that support the AVX2 but not ADX extensions
1461 like Intel Haswell (4th generation).
1463 This issue was reported to OpenSSL by David Benjamin (Google). The issue
1464 was originally found via the OSS-Fuzz project.
1468 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
1470 *) bn_sqrx8x_internal carry bug on x86_64
1472 There is a carry propagating bug in the x86_64 Montgomery squaring
1473 procedure. No EC algorithms are affected. Analysis suggests that attacks
1474 against RSA and DSA as a result of this defect would be very difficult to
1475 perform and are not believed likely. Attacks against DH are considered just
1476 feasible (although very difficult) because most of the work necessary to
1477 deduce information about a private key may be performed offline. The amount
1478 of resources required for such an attack would be very significant and
1479 likely only accessible to a limited number of attackers. An attacker would
1480 additionally need online access to an unpatched system using the target
1481 private key in a scenario with persistent DH parameters and a private
1482 key that is shared between multiple clients.
1484 This only affects processors that support the BMI1, BMI2 and ADX extensions
1485 like Intel Broadwell (5th generation) and later or AMD Ryzen.
1487 This issue was reported to OpenSSL by the OSS-Fuzz project.
1491 *) Malformed X.509 IPAddressFamily could cause OOB read
1493 If an X.509 certificate has a malformed IPAddressFamily extension,
1494 OpenSSL could do a one-byte buffer overread. The most likely result
1495 would be an erroneous display of the certificate in text format.
1497 This issue was reported to OpenSSL by the OSS-Fuzz project.
1501 Changes between 1.1.0e and 1.1.0f [25 May 2017]
1503 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1504 platform rather than 'mingw'.
1507 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1508 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1509 which is the minimum version we support.
1512 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
1514 *) Encrypt-Then-Mac renegotiation crash
1516 During a renegotiation handshake if the Encrypt-Then-Mac extension is
1517 negotiated where it was not in the original handshake (or vice-versa) then
1518 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
1519 and servers are affected.
1521 This issue was reported to OpenSSL by Joe Orton (Red Hat).
1525 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
1527 *) Truncated packet could crash via OOB read
1529 If one side of an SSL/TLS path is running on a 32-bit host and a specific
1530 cipher is being used, then a truncated packet can cause that host to
1531 perform an out-of-bounds read, usually resulting in a crash.
1533 This issue was reported to OpenSSL by Robert Święcki of Google.
1537 *) Bad (EC)DHE parameters cause a client crash
1539 If a malicious server supplies bad parameters for a DHE or ECDHE key
1540 exchange then this can result in the client attempting to dereference a
1541 NULL pointer leading to a client crash. This could be exploited in a Denial
1544 This issue was reported to OpenSSL by Guido Vranken.
1548 *) BN_mod_exp may produce incorrect results on x86_64
1550 There is a carry propagating bug in the x86_64 Montgomery squaring
1551 procedure. No EC algorithms are affected. Analysis suggests that attacks
1552 against RSA and DSA as a result of this defect would be very difficult to
1553 perform and are not believed likely. Attacks against DH are considered just
1554 feasible (although very difficult) because most of the work necessary to
1555 deduce information about a private key may be performed offline. The amount
1556 of resources required for such an attack would be very significant and
1557 likely only accessible to a limited number of attackers. An attacker would
1558 additionally need online access to an unpatched system using the target
1559 private key in a scenario with persistent DH parameters and a private
1560 key that is shared between multiple clients. For example this can occur by
1561 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1562 similar to CVE-2015-3193 but must be treated as a separate problem.
1564 This issue was reported to OpenSSL by the OSS-Fuzz project.
1568 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1570 *) ChaCha20/Poly1305 heap-buffer-overflow
1572 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1573 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1574 crash. This issue is not considered to be exploitable beyond a DoS.
1576 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1580 *) CMS Null dereference
1582 Applications parsing invalid CMS structures can crash with a NULL pointer
1583 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1584 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1585 structure callback if an attempt is made to free certain invalid encodings.
1586 Only CHOICE structures using a callback which do not handle NULL value are
1589 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1593 *) Montgomery multiplication may produce incorrect results
1595 There is a carry propagating bug in the Broadwell-specific Montgomery
1596 multiplication procedure that handles input lengths divisible by, but
1597 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1598 and DH private keys are impossible. This is because the subroutine in
1599 question is not used in operations with the private key itself and an input
1600 of the attacker's direct choice. Otherwise the bug can manifest itself as
1601 transient authentication and key negotiation failures or reproducible
1602 erroneous outcome of public-key operations with specially crafted input.
1603 Among EC algorithms only Brainpool P-512 curves are affected and one
1604 presumably can attack ECDH key negotiation. Impact was not analyzed in
1605 detail, because pre-requisites for attack are considered unlikely. Namely
1606 multiple clients have to choose the curve in question and the server has to
1607 share the private key among them, neither of which is default behaviour.
1608 Even then only clients that chose the curve will be affected.
1610 This issue was publicly reported as transient failures and was not
1611 initially recognized as a security issue. Thanks to Richard Morgan for
1612 providing reproducible case.
1616 *) Removed automatic addition of RPATH in shared libraries and executables,
1617 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1620 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1622 *) Fix Use After Free for large message sizes
1624 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1625 message larger than approx 16k is received then the underlying buffer to
1626 store the incoming message is reallocated and moved. Unfortunately a
1627 dangling pointer to the old location is left which results in an attempt to
1628 write to the previously freed location. This is likely to result in a
1629 crash, however it could potentially lead to execution of arbitrary code.
1631 This issue only affects OpenSSL 1.1.0a.
1633 This issue was reported to OpenSSL by Robert Święcki.
1637 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1639 *) OCSP Status Request extension unbounded memory growth
1641 A malicious client can send an excessively large OCSP Status Request
1642 extension. If that client continually requests renegotiation, sending a
1643 large OCSP Status Request extension each time, then there will be unbounded
1644 memory growth on the server. This will eventually lead to a Denial Of
1645 Service attack through memory exhaustion. Servers with a default
1646 configuration are vulnerable even if they do not support OCSP. Builds using
1647 the "no-ocsp" build time option are not affected.
1649 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1653 *) SSL_peek() hang on empty record
1655 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1656 sends an empty record. This could be exploited by a malicious peer in a
1657 Denial Of Service attack.
1659 This issue was reported to OpenSSL by Alex Gaynor.
1663 *) Excessive allocation of memory in tls_get_message_header() and
1664 dtls1_preprocess_fragment()
1666 A (D)TLS message includes 3 bytes for its length in the header for the
1667 message. This would allow for messages up to 16Mb in length. Messages of
1668 this length are excessive and OpenSSL includes a check to ensure that a
1669 peer is sending reasonably sized messages in order to avoid too much memory
1670 being consumed to service a connection. A flaw in the logic of version
1671 1.1.0 means that memory for the message is allocated too early, prior to
1672 the excessive message length check. Due to way memory is allocated in
1673 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1674 to service a connection. This could lead to a Denial of Service through
1675 memory exhaustion. However, the excessive message length check still takes
1676 place, and this would cause the connection to immediately fail. Assuming
1677 that the application calls SSL_free() on the failed connection in a timely
1678 manner then the 21Mb of allocated memory will then be immediately freed
1679 again. Therefore the excessive memory allocation will be transitory in
1680 nature. This then means that there is only a security impact if:
1682 1) The application does not call SSL_free() in a timely manner in the event
1683 that the connection fails
1685 2) The application is working in a constrained environment where there is
1686 very little free memory
1688 3) The attacker initiates multiple connection attempts such that there are
1689 multiple connections in a state where memory has been allocated for the
1690 connection; SSL_free() has not yet been called; and there is insufficient
1691 memory to service the multiple requests.
1693 Except in the instance of (1) above any Denial Of Service is likely to be
1694 transitory because as soon as the connection fails the memory is
1695 subsequently freed again in the SSL_free() call. However there is an
1696 increased risk during this period of application crashes due to the lack of
1697 memory - which would then mean a more serious Denial of Service.
1699 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1700 (CVE-2016-6307 and CVE-2016-6308)
1703 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1704 had to be removed. Primary reason is that vendor assembler can't
1705 assemble our modules with -KPIC flag. As result it, assembly
1706 support, was not even available as option. But its lack means
1707 lack of side-channel resistant code, which is incompatible with
1708 security by todays standards. Fortunately gcc is readily available
1709 prepackaged option, which we firmly point at...
1712 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1714 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1715 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1716 (to any value) allows Windows user to access PKCS#12 file generated
1717 with Windows CryptoAPI and protected with non-ASCII password, as well
1718 as files generated under UTF-8 locale on Linux also protected with
1722 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1723 have been disabled by default and removed from DEFAULT, just like RC4.
1724 See the RC4 item below to re-enable both.
1727 *) The method for finding the storage location for the Windows RAND seed file
1728 has changed. First we check %RANDFILE%. If that is not set then we check
1729 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1730 all else fails we fall back to C:\.
1733 *) The EVP_EncryptUpdate() function has had its return type changed from void
1734 to int. A return of 0 indicates and error while a return of 1 indicates
1738 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1739 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1740 off the constant time implementation for RSA, DSA and DH have been made
1741 no-ops and deprecated.
1744 *) Windows RAND implementation was simplified to only get entropy by
1745 calling CryptGenRandom(). Various other RAND-related tickets
1747 [Joseph Wylie Yandle, Rich Salz]
1749 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1750 and OPENSSL_LH_, respectively. The old names are available
1751 with API compatibility. They new names are now completely documented.
1754 *) Unify TYPE_up_ref(obj) methods signature.
1755 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1756 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1757 int (instead of void) like all others TYPE_up_ref() methods.
1758 So now these methods also check the return value of CRYPTO_atomic_add(),
1759 and the validity of object reference counter.
1760 [fdasilvayy@gmail.com]
1762 *) With Windows Visual Studio builds, the .pdb files are installed
1763 alongside the installed libraries and executables. For a static
1764 library installation, ossl_static.pdb is the associate compiler
1765 generated .pdb file to be used when linking programs.
1768 *) Remove openssl.spec. Packaging files belong with the packagers.
1771 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1772 recognise x86_64 architectures automatically. You can still decide
1773 to build for a different bitness with the environment variable
1774 KERNEL_BITS (can be 32 or 64), for example:
1776 KERNEL_BITS=32 ./config
1780 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1781 256 bit AES and HMAC with SHA256.
1784 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1787 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1790 *) To enable users to have their own config files and build file templates,
1791 Configure looks in the directory indicated by the environment variable
1792 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1793 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1794 name and is used as is.
1797 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1798 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1799 X509_CERT_FILE_CTX was removed.
1802 *) "shared" builds are now the default. To create only static libraries use
1803 the "no-shared" Configure option.
1806 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1807 All of these option have not worked for some while and are fundamental
1811 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1812 global cleanup functions are no longer required because they are handled
1813 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1814 Explicitly de-initing can cause problems (e.g. where a library that uses
1815 OpenSSL de-inits, but an application is still using it). The affected
1816 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1817 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1818 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1819 COMP_zlib_cleanup().
1822 *) --strict-warnings no longer enables runtime debugging options
1823 such as REF_DEBUG. Instead, debug options are automatically
1824 enabled with '--debug' builds.
1825 [Andy Polyakov, Emilia Käsper]
1827 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1828 have been moved out of the public header files. New functions for managing
1829 these have been added.
1832 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1833 objects have been moved out of the public header files. New
1834 functions for managing these have been added.
1837 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1838 have been moved out of the public header files. New functions for managing
1839 these have been added.
1842 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1843 moved out of the public header files. New functions for managing these
1847 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1850 *) Removed the mk1mf build scripts.
1853 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1854 it is always safe to #include a header now.
1857 *) Removed the aged BC-32 config and all its supporting scripts
1860 *) Removed support for Ultrix, Netware, and OS/2.
1863 *) Add support for HKDF.
1864 [Alessandro Ghedini]
1866 *) Add support for blake2b and blake2s
1869 *) Added support for "pipelining". Ciphers that have the
1870 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1871 encryptions/decryptions simultaneously. There are currently no built-in
1872 ciphers with this property but the expectation is that engines will be able
1873 to offer it to significantly improve throughput. Support has been extended
1874 into libssl so that multiple records for a single connection can be
1875 processed in one go (for >=TLS 1.1).
1878 *) Added the AFALG engine. This is an async capable engine which is able to
1879 offload work to the Linux kernel. In this initial version it only supports
1880 AES128-CBC. The kernel must be version 4.1.0 or greater.
1883 *) OpenSSL now uses a new threading API. It is no longer necessary to
1884 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1885 are two supported threading models: pthreads and windows threads. It is
1886 also possible to configure OpenSSL at compile time for "no-threads". The
1887 old threading API should no longer be used. The functions have been
1888 replaced with "no-op" compatibility macros.
1889 [Alessandro Ghedini, Matt Caswell]
1891 *) Modify behavior of ALPN to invoke callback after SNI/servername
1892 callback, such that updates to the SSL_CTX affect ALPN.
1895 *) Add SSL_CIPHER queries for authentication and key-exchange.
1898 *) Changes to the DEFAULT cipherlist:
1899 - Prefer (EC)DHE handshakes over plain RSA.
1900 - Prefer AEAD ciphers over legacy ciphers.
1901 - Prefer ECDSA over RSA when both certificates are available.
1902 - Prefer TLSv1.2 ciphers/PRF.
1903 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1907 *) Change the ECC default curve list to be this, in order: x25519,
1908 secp256r1, secp521r1, secp384r1.
1911 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1912 disabled by default. They can be re-enabled using the
1913 enable-weak-ssl-ciphers option to Configure.
1916 *) If the server has ALPN configured, but supports no protocols that the
1917 client advertises, send a fatal "no_application_protocol" alert.
1918 This behaviour is SHALL in RFC 7301, though it isn't universally
1919 implemented by other servers.
1922 *) Add X25519 support.
1923 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1924 for public and private key encoding using the format documented in
1925 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1926 key generation and key derivation.
1928 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1932 *) Deprecate SRP_VBASE_get_by_user.
1933 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1934 In order to fix an unavoidable memory leak (CVE-2016-0798),
1935 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1936 seed, even if the seed is configured.
1938 Users should use SRP_VBASE_get1_by_user instead. Note that in
1939 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1940 also that even though configuring the SRP seed attempts to hide
1941 invalid usernames by continuing the handshake with fake
1942 credentials, this behaviour is not constant time and no strong
1943 guarantees are made that the handshake is indistinguishable from
1944 that of a valid user.
1947 *) Configuration change; it's now possible to build dynamic engines
1948 without having to build shared libraries and vice versa. This
1949 only applies to the engines in engines/, those in crypto/engine/
1950 will always be built into libcrypto (i.e. "static").
1952 Building dynamic engines is enabled by default; to disable, use
1953 the configuration option "disable-dynamic-engine".
1955 The only requirements for building dynamic engines are the
1956 presence of the DSO module and building with position independent
1957 code, so they will also automatically be disabled if configuring
1958 with "disable-dso" or "disable-pic".
1960 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1961 are also taken away from openssl/opensslconf.h, as they are
1965 *) Configuration change; if there is a known flag to compile
1966 position independent code, it will always be applied on the
1967 libcrypto and libssl object files, and never on the application
1968 object files. This means other libraries that use routines from
1969 libcrypto / libssl can be made into shared libraries regardless
1970 of how OpenSSL was configured.
1972 If this isn't desirable, the configuration options "disable-pic"
1973 or "no-pic" can be used to disable the use of PIC. This will
1974 also disable building shared libraries and dynamic engines.
1977 *) Removed JPAKE code. It was experimental and has no wide use.
1980 *) The INSTALL_PREFIX Makefile variable has been renamed to
1981 DESTDIR. That makes for less confusion on what this variable
1982 is for. Also, the configuration option --install_prefix is
1986 *) Heartbeat for TLS has been removed and is disabled by default
1987 for DTLS; configure with enable-heartbeats. Code that uses the
1988 old #define's might need to be updated.
1989 [Emilia Käsper, Rich Salz]
1991 *) Rename REF_CHECK to REF_DEBUG.
1994 *) New "unified" build system
1996 The "unified" build system is aimed to be a common system for all
1997 platforms we support. With it comes new support for VMS.
1999 This system builds supports building in a different directory tree
2000 than the source tree. It produces one Makefile (for unix family
2001 or lookalikes), or one descrip.mms (for VMS).
2003 The source of information to make the Makefile / descrip.mms is
2004 small files called 'build.info', holding the necessary
2005 information for each directory with source to compile, and a
2006 template in Configurations, like unix-Makefile.tmpl or
2009 With this change, the library names were also renamed on Windows
2010 and on VMS. They now have names that are closer to the standard
2011 on Unix, and include the major version number, and in certain
2012 cases, the architecture they are built for. See "Notes on shared
2013 libraries" in INSTALL.
2015 We rely heavily on the perl module Text::Template.
2018 *) Added support for auto-initialisation and de-initialisation of the library.
2019 OpenSSL no longer requires explicit init or deinit routines to be called,
2020 except in certain circumstances. See the OPENSSL_init_crypto() and
2021 OPENSSL_init_ssl() man pages for further information.
2024 *) The arguments to the DTLSv1_listen function have changed. Specifically the
2025 "peer" argument is now expected to be a BIO_ADDR object.
2027 *) Rewrite of BIO networking library. The BIO library lacked consistent
2028 support of IPv6, and adding it required some more extensive
2029 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2030 which hold all types of addresses and chains of address information.
2031 It also introduces a new API, with functions like BIO_socket,
2032 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2033 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2034 have been adapted accordingly.
2037 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2041 *) CRIME protection: disable compression by default, even if OpenSSL is
2042 compiled with zlib enabled. Applications can still enable compression
2043 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2044 using the SSL_CONF library to configure compression.
2047 *) The signature of the session callback configured with
2048 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2049 was explicitly marked as 'const unsigned char*' instead of
2053 *) Always DPURIFY. Remove the use of uninitialized memory in the
2054 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2057 *) Removed many obsolete configuration items, including
2058 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2059 MD2_CHAR, MD2_INT, MD2_LONG
2061 IDEA_SHORT, IDEA_LONG
2062 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2063 [Rich Salz, with advice from Andy Polyakov]
2065 *) Many BN internals have been moved to an internal header file.
2066 [Rich Salz with help from Andy Polyakov]
2068 *) Configuration and writing out the results from it has changed.
2069 Files such as Makefile include/openssl/opensslconf.h and are now
2070 produced through general templates, such as Makefile.in and
2071 crypto/opensslconf.h.in and some help from the perl module
2074 Also, the center of configuration information is no longer
2075 Makefile. Instead, Configure produces a perl module in
2076 configdata.pm which holds most of the config data (in the hash
2077 table %config), the target data that comes from the target
2078 configuration in one of the Configurations/*.conf files (in
2082 *) To clarify their intended purposes, the Configure options
2083 --prefix and --openssldir change their semantics, and become more
2084 straightforward and less interdependent.
2086 --prefix shall be used exclusively to give the location INSTALLTOP
2087 where programs, scripts, libraries, include files and manuals are
2088 going to be installed. The default is now /usr/local.
2090 --openssldir shall be used exclusively to give the default
2091 location OPENSSLDIR where certificates, private keys, CRLs are
2092 managed. This is also where the default openssl.cnf gets
2094 If the directory given with this option is a relative path, the
2095 values of both the --prefix value and the --openssldir value will
2096 be combined to become OPENSSLDIR.
2097 The default for --openssldir is INSTALLTOP/ssl.
2099 Anyone who uses --openssldir to specify where OpenSSL is to be
2100 installed MUST change to use --prefix instead.
2103 *) The GOST engine was out of date and therefore it has been removed. An up
2104 to date GOST engine is now being maintained in an external repository.
2105 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
2106 support for GOST ciphersuites (these are only activated if a GOST engine
2110 *) EGD is no longer supported by default; use enable-egd when
2112 [Ben Kaduk and Rich Salz]
2114 *) The distribution now has Makefile.in files, which are used to
2115 create Makefile's when Configure is run. *Configure must be run
2116 before trying to build now.*
2119 *) The return value for SSL_CIPHER_description() for error conditions
2123 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2125 Obtaining and performing DNSSEC validation of TLSA records is
2126 the application's responsibility. The application provides
2127 the TLSA records of its choice to OpenSSL, and these are then
2128 used to authenticate the peer.
2130 The TLSA records need not even come from DNS. They can, for
2131 example, be used to implement local end-entity certificate or
2132 trust-anchor "pinning", where the "pin" data takes the form
2133 of TLSA records, which can augment or replace verification
2134 based on the usual WebPKI public certification authorities.
2137 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2138 continues to support deprecated interfaces in default builds.
2139 However, applications are strongly advised to compile their
2140 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2141 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2142 or the 1.1.0 releases.
2144 In environments in which all applications have been ported to
2145 not use any deprecated interfaces OpenSSL's Configure script
2146 should be used with the --api=1.1.0 option to entirely remove
2147 support for the deprecated features from the library and
2148 unconditionally disable them in the installed headers.
2149 Essentially the same effect can be achieved with the "no-deprecated"
2150 argument to Configure, except that this will always restrict
2151 the build to just the latest API, rather than a fixed API
2154 As applications are ported to future revisions of the API,
2155 they should update their compile-time OPENSSL_API_COMPAT define
2156 accordingly, but in most cases should be able to continue to
2157 compile with later releases.
2159 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2160 0x10000000L and 0x00908000L, respectively. However those
2161 versions did not support the OPENSSL_API_COMPAT feature, and
2162 so applications are not typically tested for explicit support
2163 of just the undeprecated features of either release.
2166 *) Add support for setting the minimum and maximum supported protocol.
2167 It can bet set via the SSL_set_min_proto_version() and
2168 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2169 MaxProtocol. It's recommended to use the new APIs to disable
2170 protocols instead of disabling individual protocols using
2171 SSL_set_options() or SSL_CONF's Protocol. This change also
2172 removes support for disabling TLS 1.2 in the OpenSSL TLS
2173 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2176 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2179 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2180 and integrates ECDSA and ECDH functionality into EC. Implementations can
2181 now redirect key generation and no longer need to convert to or from
2184 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2185 include the ec.h header file instead.
2188 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
2189 ciphers who are no longer supported and drops support the ephemeral RSA key
2190 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2193 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2194 opaque. For HMAC_CTX, the following constructors and destructors
2197 HMAC_CTX *HMAC_CTX_new(void);
2198 void HMAC_CTX_free(HMAC_CTX *ctx);
2200 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2201 destroy such methods has been added. See EVP_MD_meth_new(3) and
2202 EVP_CIPHER_meth_new(3) for documentation.
2205 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2206 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2207 EVP_MD_CTX_reset() should be called instead to reinitialise
2208 an already created structure.
2209 2) For consistency with the majority of our object creators and
2210 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2211 EVP_MD_CTX_(new|free). The old names are retained as macros
2212 for deprecated builds.
2215 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
2216 cryptographic operations to be performed asynchronously as long as an
2217 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2218 further details. Libssl has also had this capability integrated with the
2219 introduction of the new mode SSL_MODE_ASYNC and associated error
2220 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2221 pages. This work was developed in partnership with Intel Corp.
2224 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2225 always enabled now. If you want to disable the support you should
2226 exclude it using the list of supported ciphers. This also means that the
2227 "-no_ecdhe" option has been removed from s_server.
2230 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2231 SSL_{CTX_}set1_curves() which can set a list.
2234 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2235 curve you want to support using SSL_{CTX_}set1_curves().
2238 *) State machine rewrite. The state machine code has been significantly
2239 refactored in order to remove much duplication of code and solve issues
2240 with the old code (see ssl/statem/README for further details). This change
2241 does have some associated API changes. Notably the SSL_state() function
2242 has been removed and replaced by SSL_get_state which now returns an
2243 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
2244 altogether. The previous handshake states defined in ssl.h and ssl3.h have
2248 *) All instances of the string "ssleay" in the public API were replaced
2249 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
2250 Some error codes related to internal RSA_eay API's were renamed.
2253 *) The demo files in crypto/threads were moved to demo/threads.
2256 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
2258 [Matt Caswell, Rich Salz]
2260 *) New ASN.1 embed macro.
2262 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
2263 structure is not allocated: it is part of the parent. That is instead of
2271 This reduces memory fragmentation and make it impossible to accidentally
2272 set a mandatory field to NULL.
2274 This currently only works for some fields specifically a SEQUENCE, CHOICE,
2275 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
2276 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
2280 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
2283 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
2284 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
2285 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
2286 DES and RC4 ciphersuites.
2289 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2290 This changes the decoding behaviour for some invalid messages,
2291 though the change is mostly in the more lenient direction, and
2292 legacy behaviour is preserved as much as possible.
2295 *) Fix no-stdio build.
2296 [ David Woodhouse <David.Woodhouse@intel.com> and also
2297 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
2299 *) New testing framework
2300 The testing framework has been largely rewritten and is now using
2301 perl and the perl modules Test::Harness and an extended variant of
2302 Test::More called OpenSSL::Test to do its work. All test scripts in
2303 test/ have been rewritten into test recipes, and all direct calls to
2304 executables in test/Makefile have become individual recipes using the
2305 simplified testing OpenSSL::Test::Simple.
2307 For documentation on our testing modules, do:
2309 perldoc test/testlib/OpenSSL/Test/Simple.pm
2310 perldoc test/testlib/OpenSSL/Test.pm
2314 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
2315 are used; the latter aborts on memory leaks (usually checked on exit).
2316 Some undocumented "set malloc, etc., hooks" functions were removed
2317 and others were changed. All are now documented.
2320 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2322 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2324 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
2325 from RFC4279, RFC4785, RFC5487, RFC5489.
2327 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
2328 original RSA_PSK patch.
2331 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
2332 era flag was never set throughout the codebase (only read). Also removed
2333 SSL3_FLAGS_POP_BUFFER which was only used if
2334 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
2337 *) Changed the default name options in the "ca", "crl", "req" and "x509"
2338 to be "oneline" instead of "compat".
2341 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
2342 not aware of clients that still exhibit this bug, and the workaround
2343 hasn't been working properly for a while.
2346 *) The return type of BIO_number_read() and BIO_number_written() as well as
2347 the corresponding num_read and num_write members in the BIO structure has
2348 changed from unsigned long to uint64_t. On platforms where an unsigned
2349 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
2353 *) Given the pervasive nature of TLS extensions it is inadvisable to run
2354 OpenSSL without support for them. It also means that maintaining
2355 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
2356 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
2359 *) Removed support for the two export grade static DH ciphersuites
2360 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
2361 were newly added (along with a number of other static DH ciphersuites) to
2362 1.0.2. However the two export ones have *never* worked since they were
2363 introduced. It seems strange in any case to be adding new export
2364 ciphersuites, and given "logjam" it also does not seem correct to fix them.
2367 *) Version negotiation has been rewritten. In particular SSLv23_method(),
2368 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
2369 and turned into macros which simply call the new preferred function names
2370 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
2371 should use the new names instead. Also as part of this change the ssl23.h
2372 header file has been removed.
2375 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
2376 code and the associated standard is no longer considered fit-for-purpose.
2379 *) RT2547 was closed. When generating a private key, try to make the
2380 output file readable only by the owner. This behavior change might
2381 be noticeable when interacting with other software.
2383 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
2387 *) Added HTTP GET support to the ocsp command.
2390 *) Changed default digest for the dgst and enc commands from MD5 to
2394 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
2397 *) Added support for TLS extended master secret from
2398 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
2399 initial patch which was a great help during development.
2402 *) All libssl internal structures have been removed from the public header
2403 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
2404 now redundant). Users should not attempt to access internal structures
2405 directly. Instead they should use the provided API functions.
2408 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
2409 Access to deprecated functions can be re-enabled by running config with
2410 "enable-deprecated". In addition applications wishing to use deprecated
2411 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
2412 will, by default, disable some transitive includes that previously existed
2413 in the header files (e.g. ec.h will no longer, by default, include bn.h)
2416 *) Added support for OCB mode. OpenSSL has been granted a patent license
2417 compatible with the OpenSSL license for use of OCB. Details are available
2418 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
2419 for OCB can be removed by calling config with no-ocb.
2422 *) SSLv2 support has been removed. It still supports receiving a SSLv2
2423 compatible client hello.
2426 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
2427 done while fixing the error code for the key-too-small case.
2428 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
2430 *) CA.sh has been removed; use CA.pl instead.
2433 *) Removed old DES API.
2436 *) Remove various unsupported platforms:
2442 Sinix/ReliantUNIX RM400
2447 16-bit platforms such as WIN16
2450 *) Clean up OPENSSL_NO_xxx #define's
2451 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
2452 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
2453 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
2454 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
2455 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
2456 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
2457 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
2458 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
2459 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
2460 Remove MS_STATIC; it's a relic from platforms <32 bits.
2463 *) Cleaned up dead code
2464 Remove all but one '#ifdef undef' which is to be looked at.
2467 *) Clean up calling of xxx_free routines.
2468 Just like free(), fix most of the xxx_free routines to accept
2469 NULL. Remove the non-null checks from callers. Save much code.
2472 *) Add secure heap for storage of private keys (when possible).
2473 Add BIO_s_secmem(), CBIGNUM, etc.
2474 Contributed by Akamai Technologies under our Corporate CLA.
2477 *) Experimental support for a new, fast, unbiased prime candidate generator,
2478 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
2479 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
2481 *) New output format NSS in the sess_id command line tool. This allows
2482 exporting the session id and the master key in NSS keylog format.
2483 [Martin Kaiser <martin@kaiser.cx>]
2485 *) Harmonize version and its documentation. -f flag is used to display
2487 [mancha <mancha1@zoho.com>]
2489 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2490 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
2491 [mancha <mancha1@zoho.com>]
2493 *) Fix some double frees. These are not thought to be exploitable.
2494 [mancha <mancha1@zoho.com>]
2496 *) A missing bounds check in the handling of the TLS heartbeat extension
2497 can be used to reveal up to 64k of memory to a connected client or
2500 Thanks for Neel Mehta of Google Security for discovering this bug and to
2501 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2502 preparing the fix (CVE-2014-0160)
2503 [Adam Langley, Bodo Moeller]
2505 *) Fix for the attack described in the paper "Recovering OpenSSL
2506 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2507 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2508 http://eprint.iacr.org/2014/140
2510 Thanks to Yuval Yarom and Naomi Benger for discovering this
2511 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2512 [Yuval Yarom and Naomi Benger]
2514 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2515 this fixes a limitation in previous versions of OpenSSL.
2518 *) Experimental encrypt-then-mac support.
2520 Experimental support for encrypt then mac from
2521 draft-gutmann-tls-encrypt-then-mac-02.txt
2523 To enable it set the appropriate extension number (0x42 for the test
2524 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
2526 For non-compliant peers (i.e. just about everything) this should have no
2529 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
2533 *) Add EVP support for key wrapping algorithms, to avoid problems with
2534 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2535 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2536 algorithms and include tests cases.
2539 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2543 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2544 MGF1 digest and OAEP label.
2547 *) Make openssl verify return errors.
2548 [Chris Palmer <palmer@google.com> and Ben Laurie]
2550 *) New function ASN1_TIME_diff to calculate the difference between two
2551 ASN1_TIME structures or one structure and the current time.
2554 *) Update fips_test_suite to support multiple command line options. New
2555 test to induce all self test errors in sequence and check expected
2559 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2560 sign or verify all in one operation.
2563 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2564 test programs and fips_test_suite. Includes functionality to parse
2565 the minimal script output of fipsalgest.pl directly.
2568 *) Add authorisation parameter to FIPS_module_mode_set().
2571 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2574 *) Use separate DRBG fields for internal and external flags. New function
2575 FIPS_drbg_health_check() to perform on demand health checking. Add
2576 generation tests to fips_test_suite with reduced health check interval to
2577 demonstrate periodic health checking. Add "nodh" option to
2578 fips_test_suite to skip very slow DH test.
2581 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2585 *) More extensive health check for DRBG checking many more failure modes.
2586 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2587 combination: call this in fips_test_suite.
2590 *) Add support for canonical generation of DSA parameter 'g'. See
2593 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2594 POST to handle HMAC cases.
2597 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2598 to return numerical and string versions of the FIPS module number.
2601 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2602 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2603 outside the validated module in the FIPS capable OpenSSL.
2606 *) Minor change to DRBG entropy callback semantics. In some cases
2607 there is no multiple of the block length between min_len and
2608 max_len. Allow the callback to return more than max_len bytes
2609 of entropy but discard any extra: it is the callback's responsibility
2610 to ensure that the extra data discarded does not impact the
2611 requested amount of entropy.
2614 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2615 information in FIPS186-3, SP800-57 and SP800-131A.
2618 *) CCM support via EVP. Interface is very similar to GCM case except we
2619 must supply all data in one chunk (i.e. no update, final) and the
2620 message length must be supplied if AAD is used. Add algorithm test
2624 *) Initial version of POST overhaul. Add POST callback to allow the status
2625 of POST to be monitored and/or failures induced. Modify fips_test_suite
2626 to use callback. Always run all selftests even if one fails.
2629 *) XTS support including algorithm test driver in the fips_gcmtest program.
2630 Note: this does increase the maximum key length from 32 to 64 bytes but
2631 there should be no binary compatibility issues as existing applications
2632 will never use XTS mode.
2635 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2636 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2637 performs algorithm blocking for unapproved PRNG types. Also do not
2638 set PRNG type in FIPS_mode_set(): leave this to the application.
2639 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2640 the standard OpenSSL PRNG: set additional data to a date time vector.
2643 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2644 This shouldn't present any incompatibility problems because applications
2645 shouldn't be using these directly and any that are will need to rethink
2646 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2649 *) Extensive self tests and health checking required by SP800-90 DRBG.
2650 Remove strength parameter from FIPS_drbg_instantiate and always
2651 instantiate at maximum supported strength.
2654 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2657 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2660 *) New function DH_compute_key_padded() to compute a DH key and pad with
2661 leading zeroes if needed: this complies with SP800-56A et al.
2664 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2665 anything, incomplete, subject to change and largely untested at present.
2668 *) Modify fipscanisteronly build option to only build the necessary object
2669 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2672 *) Add experimental option FIPSSYMS to give all symbols in
2673 fipscanister.o and FIPS or fips prefix. This will avoid
2674 conflicts with future versions of OpenSSL. Add perl script
2675 util/fipsas.pl to preprocess assembly language source files
2676 and rename any affected symbols.
2679 *) Add selftest checks and algorithm block of non-fips algorithms in
2680 FIPS mode. Remove DES2 from selftests.
2683 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2684 return internal method without any ENGINE dependencies. Add new
2685 tiny fips sign and verify functions.
2688 *) New build option no-ec2m to disable characteristic 2 code.
2691 *) New build option "fipscanisteronly". This only builds fipscanister.o
2692 and (currently) associated fips utilities. Uses the file Makefile.fips
2693 instead of Makefile.org as the prototype.
2696 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2697 Update fips_gcmtest to use IV generator.
2700 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2701 setting output buffer to NULL. The *Final function must be
2702 called although it will not retrieve any additional data. The tag
2703 can be set or retrieved with a ctrl. The IV length is by default 12
2704 bytes (96 bits) but can be set to an alternative value. If the IV
2705 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2709 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2710 underlying do_cipher function handles all cipher semantics itself
2711 including padding and finalisation. This is useful if (for example)
2712 an ENGINE cipher handles block padding itself. The behaviour of
2713 do_cipher is subtly changed if this flag is set: the return value
2714 is the number of characters written to the output buffer (zero is
2715 no longer an error code) or a negative error code. Also if the
2716 input buffer is NULL and length 0 finalisation should be performed.
2719 *) If a candidate issuer certificate is already part of the constructed
2720 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2723 *) Improve forward-security support: add functions
2725 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2726 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2728 for use by SSL/TLS servers; the callback function will be called whenever a
2729 new session is created, and gets to decide whether the session may be
2730 cached to make it resumable (return 0) or not (return 1). (As by the
2731 SSL/TLS protocol specifications, the session_id sent by the server will be
2732 empty to indicate that the session is not resumable; also, the server will
2733 not generate RFC 4507 (RFC 5077) session tickets.)
2735 A simple reasonable callback implementation is to return is_forward_secure.
2736 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2737 by the SSL/TLS server library, indicating whether it can provide forward
2739 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2741 *) New -verify_name option in command line utilities to set verification
2745 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2746 Add CMAC pkey methods.
2749 *) Experimental renegotiation in s_server -www mode. If the client
2750 browses /reneg connection is renegotiated. If /renegcert it is
2751 renegotiated requesting a certificate.
2754 *) Add an "external" session cache for debugging purposes to s_server. This
2755 should help trace issues which normally are only apparent in deployed
2756 multi-process servers.
2759 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2760 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2761 BIO_set_cipher() and some obscure PEM functions were changed so they
2762 can now return an error. The RAND changes required a change to the
2763 RAND_METHOD structure.
2766 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2767 a gcc attribute to warn if the result of a function is ignored. This
2768 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2769 whose return value is often ignored.
2772 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2773 These allow SCTs (signed certificate timestamps) to be requested and
2774 validated when establishing a connection.
2775 [Rob Percival <robpercival@google.com>]
2777 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2779 *) Prevent padding oracle in AES-NI CBC MAC check
2781 A MITM attacker can use a padding oracle attack to decrypt traffic
2782 when the connection uses an AES CBC cipher and the server support
2785 This issue was introduced as part of the fix for Lucky 13 padding
2786 attack (CVE-2013-0169). The padding check was rewritten to be in
2787 constant time by making sure that always the same bytes are read and
2788 compared against either the MAC or padding bytes. But it no longer
2789 checked that there was enough data to have both the MAC and padding
2792 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2796 *) Fix EVP_EncodeUpdate overflow
2798 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2799 Base64 encoding of binary data. If an attacker is able to supply very large
2800 amounts of input data then a length check can overflow resulting in a heap
2803 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2804 the PEM_write_bio* family of functions. These are mainly used within the
2805 OpenSSL command line applications, so any application which processes data
2806 from an untrusted source and outputs it as a PEM file should be considered
2807 vulnerable to this issue. User applications that call these APIs directly
2808 with large amounts of untrusted data may also be vulnerable.
2810 This issue was reported by Guido Vranken.
2814 *) Fix EVP_EncryptUpdate overflow
2816 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2817 is able to supply very large amounts of input data after a previous call to
2818 EVP_EncryptUpdate() with a partial block then a length check can overflow
2819 resulting in a heap corruption. Following an analysis of all OpenSSL
2820 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2821 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2822 the first called function after an EVP_EncryptInit(), and therefore that
2823 specific call must be safe. The second form is where the length passed to
2824 EVP_EncryptUpdate() can be seen from the code to be some small value and
2825 therefore there is no possibility of an overflow. Since all instances are
2826 one of these two forms, it is believed that there can be no overflows in
2827 internal code due to this problem. It should be noted that
2828 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2829 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2830 of these calls have also been analysed too and it is believed there are no
2831 instances in internal usage where an overflow could occur.
2833 This issue was reported by Guido Vranken.
2837 *) Prevent ASN.1 BIO excessive memory allocation
2839 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2840 a short invalid encoding can cause allocation of large amounts of memory
2841 potentially consuming excessive resources or exhausting memory.
2843 Any application parsing untrusted data through d2i BIO functions is
2844 affected. The memory based functions such as d2i_X509() are *not* affected.
2845 Since the memory based functions are used by the TLS library, TLS
2846 applications are not affected.
2848 This issue was reported by Brian Carpenter.
2854 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2855 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2856 in arbitrary stack data being returned in the buffer.
2858 This issue was reported by Guido Vranken.
2862 *) Modify behavior of ALPN to invoke callback after SNI/servername
2863 callback, such that updates to the SSL_CTX affect ALPN.
2866 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2870 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2871 methods are enabled and ssl2 is disabled the methods return NULL.
2874 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2876 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2877 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2878 provide any "EXPORT" or "LOW" strength ciphers.
2881 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2882 is by default disabled at build-time. Builds that are not configured with
2883 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2884 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2885 will need to explicitly call either of:
2887 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2889 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2891 as appropriate. Even if either of those is used, or the application
2892 explicitly uses the version-specific SSLv2_method() or its client and
2893 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2894 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2895 ciphers, and SSLv2 56-bit DES are no longer available.
2899 *) Fix a double-free in DSA code
2901 A double free bug was discovered when OpenSSL parses malformed DSA private
2902 keys and could lead to a DoS attack or memory corruption for applications
2903 that receive DSA private keys from untrusted sources. This scenario is
2906 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2911 *) Disable SRP fake user seed to address a server memory leak.
2913 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2915 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2916 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2917 was changed to ignore the "fake user" SRP seed, even if the seed
2920 Users should use SRP_VBASE_get1_by_user instead. Note that in
2921 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2922 also that even though configuring the SRP seed attempts to hide
2923 invalid usernames by continuing the handshake with fake
2924 credentials, this behaviour is not constant time and no strong
2925 guarantees are made that the handshake is indistinguishable from
2926 that of a valid user.
2930 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2932 In the BN_hex2bn function the number of hex digits is calculated using an
2933 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2934 large values of |i| this can result in |bn_expand| not allocating any
2935 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2936 field as NULL leading to a subsequent NULL ptr deref. For very large values
2937 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2938 In this case memory is allocated to the internal BIGNUM data field, but it
2939 is insufficiently sized leading to heap corruption. A similar issue exists
2940 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2941 is ever called by user applications with very large untrusted hex/dec data.
2942 This is anticipated to be a rare occurrence.
2944 All OpenSSL internal usage of these functions use data that is not expected
2945 to be untrusted, e.g. config file data or application command line
2946 arguments. If user developed applications generate config file data based
2947 on untrusted data then it is possible that this could also lead to security
2948 consequences. This is also anticipated to be rare.
2950 This issue was reported to OpenSSL by Guido Vranken.
2954 *) Fix memory issues in BIO_*printf functions
2956 The internal |fmtstr| function used in processing a "%s" format string in
2957 the BIO_*printf functions could overflow while calculating the length of a
2958 string and cause an OOB read when printing very long strings.
2960 Additionally the internal |doapr_outch| function can attempt to write to an
2961 OOB memory location (at an offset from the NULL pointer) in the event of a
2962 memory allocation failure. In 1.0.2 and below this could be caused where
2963 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2964 could be in processing a very long "%s" format string. Memory leaks can
2967 The first issue may mask the second issue dependent on compiler behaviour.
2968 These problems could enable attacks where large amounts of untrusted data
2969 is passed to the BIO_*printf functions. If applications use these functions
2970 in this way then they could be vulnerable. OpenSSL itself uses these
2971 functions when printing out human-readable dumps of ASN.1 data. Therefore
2972 applications that print this data could be vulnerable if the data is from
2973 untrusted sources. OpenSSL command line applications could also be
2974 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2975 as command line arguments.
2977 Libssl is not considered directly vulnerable. Additionally certificates etc
2978 received via remote connections via libssl are also unlikely to be able to
2979 trigger these issues because of message size limits enforced within libssl.
2981 This issue was reported to OpenSSL Guido Vranken.
2985 *) Side channel attack on modular exponentiation
2987 A side-channel attack was found which makes use of cache-bank conflicts on
2988 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2989 of RSA keys. The ability to exploit this issue is limited as it relies on
2990 an attacker who has control of code in a thread running on the same
2991 hyper-threaded core as the victim thread which is performing decryptions.
2993 This issue was reported to OpenSSL by Yuval Yarom, The University of
2994 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2995 Nadia Heninger, University of Pennsylvania with more information at
2996 http://cachebleed.info.
3000 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
3001 if no keysize is specified with default_bits. This fixes an
3002 omission in an earlier change that changed all RSA/DSA key generation
3003 apps to use 2048 bits by default.
3006 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
3007 *) DH small subgroups
3009 Historically OpenSSL only ever generated DH parameters based on "safe"
3010 primes. More recently (in version 1.0.2) support was provided for
3011 generating X9.42 style parameter files such as those required for RFC 5114
3012 support. The primes used in such files may not be "safe". Where an
3013 application is using DH configured with parameters based on primes that are
3014 not "safe" then an attacker could use this fact to find a peer's private
3015 DH exponent. This attack requires that the attacker complete multiple
3016 handshakes in which the peer uses the same private DH exponent. For example
3017 this could be used to discover a TLS server's private DH exponent if it's
3018 reusing the private DH exponent or it's using a static DH ciphersuite.
3020 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
3021 TLS. It is not on by default. If the option is not set then the server
3022 reuses the same private DH exponent for the life of the server process and
3023 would be vulnerable to this attack. It is believed that many popular
3024 applications do set this option and would therefore not be at risk.
3026 The fix for this issue adds an additional check where a "q" parameter is
3027 available (as is the case in X9.42 based parameters). This detects the
3028 only known attack, and is the only possible defense for static DH
3029 ciphersuites. This could have some performance impact.
3031 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
3032 default and cannot be disabled. This could have some performance impact.
3034 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
3038 *) SSLv2 doesn't block disabled ciphers
3040 A malicious client can negotiate SSLv2 ciphers that have been disabled on
3041 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
3042 been disabled, provided that the SSLv2 protocol was not also disabled via
3045 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
3046 and Sebastian Schinzel.
3050 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
3052 *) BN_mod_exp may produce incorrect results on x86_64
3054 There is a carry propagating bug in the x86_64 Montgomery squaring
3055 procedure. No EC algorithms are affected. Analysis suggests that attacks
3056 against RSA and DSA as a result of this defect would be very difficult to
3057 perform and are not believed likely. Attacks against DH are considered just
3058 feasible (although very difficult) because most of the work necessary to
3059 deduce information about a private key may be performed offline. The amount
3060 of resources required for such an attack would be very significant and
3061 likely only accessible to a limited number of attackers. An attacker would
3062 additionally need online access to an unpatched system using the target
3063 private key in a scenario with persistent DH parameters and a private
3064 key that is shared between multiple clients. For example this can occur by
3065 default in OpenSSL DHE based SSL/TLS ciphersuites.
3067 This issue was reported to OpenSSL by Hanno Böck.
3071 *) Certificate verify crash with missing PSS parameter
3073 The signature verification routines will crash with a NULL pointer
3074 dereference if presented with an ASN.1 signature using the RSA PSS
3075 algorithm and absent mask generation function parameter. Since these
3076 routines are used to verify certificate signature algorithms this can be
3077 used to crash any certificate verification operation and exploited in a
3078 DoS attack. Any application which performs certificate verification is
3079 vulnerable including OpenSSL clients and servers which enable client
3082 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
3086 *) X509_ATTRIBUTE memory leak
3088 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
3089 memory. This structure is used by the PKCS#7 and CMS routines so any
3090 application which reads PKCS#7 or CMS data from untrusted sources is
3091 affected. SSL/TLS is not affected.
3093 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
3098 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3099 This changes the decoding behaviour for some invalid messages,
3100 though the change is mostly in the more lenient direction, and
3101 legacy behaviour is preserved as much as possible.
3104 *) In DSA_generate_parameters_ex, if the provided seed is too short,
3106 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
3108 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
3110 *) Alternate chains certificate forgery
3112 During certificate verification, OpenSSL will attempt to find an
3113 alternative certificate chain if the first attempt to build such a chain
3114 fails. An error in the implementation of this logic can mean that an
3115 attacker could cause certain checks on untrusted certificates to be
3116 bypassed, such as the CA flag, enabling them to use a valid leaf
3117 certificate to act as a CA and "issue" an invalid certificate.
3119 This issue was reported to OpenSSL by Adam Langley/David Benjamin
3123 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
3125 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
3126 incompatibility in the handling of HMAC. The previous ABI has now been
3130 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
3132 *) Malformed ECParameters causes infinite loop
3134 When processing an ECParameters structure OpenSSL enters an infinite loop
3135 if the curve specified is over a specially malformed binary polynomial
3138 This can be used to perform denial of service against any
3139 system which processes public keys, certificate requests or
3140 certificates. This includes TLS clients and TLS servers with
3141 client authentication enabled.
3143 This issue was reported to OpenSSL by Joseph Barr-Pixton.
3147 *) Exploitable out-of-bounds read in X509_cmp_time
3149 X509_cmp_time does not properly check the length of the ASN1_TIME
3150 string and can read a few bytes out of bounds. In addition,
3151 X509_cmp_time accepts an arbitrary number of fractional seconds in the
3154 An attacker can use this to craft malformed certificates and CRLs of
3155 various sizes and potentially cause a segmentation fault, resulting in
3156 a DoS on applications that verify certificates or CRLs. TLS clients
3157 that verify CRLs are affected. TLS clients and servers with client
3158 authentication enabled may be affected if they use custom verification
3161 This issue was reported to OpenSSL by Robert Swiecki (Google), and
3162 independently by Hanno Böck.
3166 *) PKCS7 crash with missing EnvelopedContent
3168 The PKCS#7 parsing code does not handle missing inner EncryptedContent
3169 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
3170 with missing content and trigger a NULL pointer dereference on parsing.
3172 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
3173 structures from untrusted sources are affected. OpenSSL clients and
3174 servers are not affected.
3176 This issue was reported to OpenSSL by Michal Zalewski (Google).
3180 *) CMS verify infinite loop with unknown hash function
3182 When verifying a signedData message the CMS code can enter an infinite loop
3183 if presented with an unknown hash function OID. This can be used to perform
3184 denial of service against any system which verifies signedData messages using
3186 This issue was reported to OpenSSL by Johannes Bauer.
3190 *) Race condition handling NewSessionTicket
3192 If a NewSessionTicket is received by a multi-threaded client when attempting to
3193 reuse a previous ticket then a race condition can occur potentially leading to
3194 a double free of the ticket data.
3198 *) Only support 256-bit or stronger elliptic curves with the
3199 'ecdh_auto' setting (server) or by default (client). Of supported
3200 curves, prefer P-256 (both).
3203 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
3205 *) ClientHello sigalgs DoS fix
3207 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
3208 invalid signature algorithms extension a NULL pointer dereference will
3209 occur. This can be exploited in a DoS attack against the server.
3211 This issue was was reported to OpenSSL by David Ramos of Stanford
3214 [Stephen Henson and Matt Caswell]
3216 *) Multiblock corrupted pointer fix
3218 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
3219 feature only applies on 64 bit x86 architecture platforms that support AES
3220 NI instructions. A defect in the implementation of "multiblock" can cause
3221 OpenSSL's internal write buffer to become incorrectly set to NULL when
3222 using non-blocking IO. Typically, when the user application is using a
3223 socket BIO for writing, this will only result in a failed connection.
3224 However if some other BIO is used then it is likely that a segmentation
3225 fault will be triggered, thus enabling a potential DoS attack.
3227 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
3231 *) Segmentation fault in DTLSv1_listen fix
3233 The DTLSv1_listen function is intended to be stateless and processes the
3234 initial ClientHello from many peers. It is common for user code to loop
3235 over the call to DTLSv1_listen until a valid ClientHello is received with
3236 an associated cookie. A defect in the implementation of DTLSv1_listen means
3237 that state is preserved in the SSL object from one invocation to the next
3238 that can lead to a segmentation fault. Errors processing the initial
3239 ClientHello can trigger this scenario. An example of such an error could be
3240 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
3243 This issue was reported to OpenSSL by Per Allansson.
3247 *) Segmentation fault in ASN1_TYPE_cmp fix
3249 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
3250 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
3251 certificate signature algorithm consistency this can be used to crash any
3252 certificate verification operation and exploited in a DoS attack. Any
3253 application which performs certificate verification is vulnerable including
3254 OpenSSL clients and servers which enable client authentication.
3258 *) Segmentation fault for invalid PSS parameters fix
3260 The signature verification routines will crash with a NULL pointer
3261 dereference if presented with an ASN.1 signature using the RSA PSS
3262 algorithm and invalid parameters. Since these routines are used to verify
3263 certificate signature algorithms this can be used to crash any
3264 certificate verification operation and exploited in a DoS attack. Any
3265 application which performs certificate verification is vulnerable including
3266 OpenSSL clients and servers which enable client authentication.
3268 This issue was was reported to OpenSSL by Brian Carpenter.
3272 *) ASN.1 structure reuse memory corruption fix
3274 Reusing a structure in ASN.1 parsing may allow an attacker to cause
3275 memory corruption via an invalid write. Such reuse is and has been
3276 strongly discouraged and is believed to be rare.
3278 Applications that parse structures containing CHOICE or ANY DEFINED BY
3279 components may be affected. Certificate parsing (d2i_X509 and related
3280 functions) are however not affected. OpenSSL clients and servers are
3285 *) PKCS7 NULL pointer dereferences fix
3287 The PKCS#7 parsing code does not handle missing outer ContentInfo
3288 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
3289 missing content and trigger a NULL pointer dereference on parsing.
3291 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
3292 otherwise parse PKCS#7 structures from untrusted sources are
3293 affected. OpenSSL clients and servers are not affected.
3295 This issue was reported to OpenSSL by Michal Zalewski (Google).
3299 *) DoS via reachable assert in SSLv2 servers fix
3301 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
3302 servers that both support SSLv2 and enable export cipher suites by sending
3303 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
3305 This issue was discovered by Sean Burford (Google) and Emilia Käsper
3306 (OpenSSL development team).
3310 *) Empty CKE with client auth and DHE fix
3312 If client auth is used then a server can seg fault in the event of a DHE
3313 ciphersuite being selected and a zero length ClientKeyExchange message
3314 being sent by the client. This could be exploited in a DoS attack.
3318 *) Handshake with unseeded PRNG fix
3320 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
3321 with an unseeded PRNG. The conditions are:
3322 - The client is on a platform where the PRNG has not been seeded
3323 automatically, and the user has not seeded manually
3324 - A protocol specific client method version has been used (i.e. not
3325 SSL_client_methodv23)
3326 - A ciphersuite is used that does not require additional random data from
3327 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
3329 If the handshake succeeds then the client random that has been used will
3330 have been generated from a PRNG with insufficient entropy and therefore the
3331 output may be predictable.
3333 For example using the following command with an unseeded openssl will
3334 succeed on an unpatched platform:
3336 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
3340 *) Use After Free following d2i_ECPrivatekey error fix
3342 A malformed EC private key file consumed via the d2i_ECPrivateKey function
3343 could cause a use after free condition. This, in turn, could cause a double
3344 free in several private key parsing functions (such as d2i_PrivateKey
3345 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
3346 for applications that receive EC private keys from untrusted
3347 sources. This scenario is considered rare.
3349 This issue was discovered by the BoringSSL project and fixed in their
3354 *) X509_to_X509_REQ NULL pointer deref fix
3356 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
3357 the certificate key is invalid. This function is rarely used in practice.
3359 This issue was discovered by Brian Carpenter.
3363 *) Removed the export ciphers from the DEFAULT ciphers
3366 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
3368 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
3369 ARMv5 through ARMv8, as opposite to "locking" it to single one.
3370 So far those who have to target multiple platforms would compromise
3371 and argue that binary targeting say ARMv5 would still execute on
3372 ARMv8. "Universal" build resolves this compromise by providing
3373 near-optimal performance even on newer platforms.
3376 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
3377 (other platforms pending).
3378 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
3380 *) Add support for the SignedCertificateTimestampList certificate and
3381 OCSP response extensions from RFC6962.
3384 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3385 for corner cases. (Certain input points at infinity could lead to
3386 bogus results, with non-infinity inputs mapped to infinity too.)
3389 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
3390 This covers AES, SHA256/512 and GHASH. "Initial" means that most
3391 common cases are optimized and there still is room for further
3392 improvements. Vector Permutation AES for Altivec is also added.
3395 *) Add support for little-endian ppc64 Linux target.
3396 [Marcelo Cerri (IBM)]
3398 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
3399 SHA1, SHA256 and GHASH. "Initial" means that most common cases
3400 are optimized and there still is room for further improvements.
3401 Both 32- and 64-bit modes are supported.
3402 [Andy Polyakov, Ard Biesheuvel (Linaro)]
3404 *) Improved ARMv7 NEON support.
3407 *) Support for SPARC Architecture 2011 crypto extensions, first
3408 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
3409 SHA256/512, MD5, GHASH and modular exponentiation.
3410 [Andy Polyakov, David Miller]
3412 *) Accelerated modular exponentiation for Intel processors, a.k.a.
3414 [Shay Gueron & Vlad Krasnov (Intel Corp)]
3416 *) Support for new and upcoming Intel processors, including AVX2,
3417 BMI and SHA ISA extensions. This includes additional "stitched"
3418 implementations, AESNI-SHA256 and GCM, and multi-buffer support
3421 This work was sponsored by Intel Corp.
3424 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
3425 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
3426 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
3429 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3430 this fixes a limitation in previous versions of OpenSSL.
3433 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3434 MGF1 digest and OAEP label.
3437 *) Add EVP support for key wrapping algorithms, to avoid problems with
3438 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3439 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3440 algorithms and include tests cases.
3443 *) Add functions to allocate and set the fields of an ECDSA_METHOD
3445 [Douglas E. Engert, Steve Henson]
3447 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
3448 difference in days and seconds between two tm or ASN1_TIME structures.
3451 *) Add -rev test option to s_server to just reverse order of characters
3452 received by client and send back to server. Also prints an abbreviated
3453 summary of the connection parameters.
3456 *) New option -brief for s_client and s_server to print out a brief summary
3457 of connection parameters.
3460 *) Add callbacks for arbitrary TLS extensions.
3461 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
3463 *) New option -crl_download in several openssl utilities to download CRLs
3464 from CRLDP extension in certificates.
3467 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
3470 *) New function X509_CRL_diff to generate a delta CRL from the difference
3471 of two full CRLs. Add support to "crl" utility.
3474 *) New functions to set lookup_crls function and to retrieve
3475 X509_STORE from X509_STORE_CTX.
3478 *) Print out deprecated issuer and subject unique ID fields in
3482 *) Extend OCSP I/O functions so they can be used for simple general purpose
3483 HTTP as well as OCSP. New wrapper function which can be used to download
3484 CRLs using the OCSP API.
3487 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
3490 *) SSL_CONF* functions. These provide a common framework for application
3491 configuration using configuration files or command lines.
3494 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
3495 message callback and prints the results. Needs compile time option
3496 "enable-ssl-trace". New options to s_client and s_server to enable
3500 *) New ctrl and macro to retrieve supported points extensions.
3501 Print out extension in s_server and s_client.
3504 *) New functions to retrieve certificate signature and signature
3508 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
3512 *) New Suite B modes for TLS code. These use and enforce the requirements
3513 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
3514 only use Suite B curves. The Suite B modes can be set by using the
3515 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
3518 *) New chain verification flags for Suite B levels of security. Check
3519 algorithms are acceptable when flags are set in X509_verify_cert.
3522 *) Make tls1_check_chain return a set of flags indicating checks passed
3523 by a certificate chain. Add additional tests to handle client
3524 certificates: checks for matching certificate type and issuer name
3528 *) If an attempt is made to use a signature algorithm not in the peer
3529 preference list abort the handshake. If client has no suitable
3530 signature algorithms in response to a certificate request do not
3531 use the certificate.
3534 *) If server EC tmp key is not in client preference list abort handshake.
3537 *) Add support for certificate stores in CERT structure. This makes it
3538 possible to have different stores per SSL structure or one store in
3539 the parent SSL_CTX. Include distinct stores for certificate chain
3540 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3541 to build and store a certificate chain in CERT structure: returning
3542 an error if the chain cannot be built: this will allow applications
3543 to test if a chain is correctly configured.
3545 Note: if the CERT based stores are not set then the parent SSL_CTX
3546 store is used to retain compatibility with existing behaviour.
3550 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3551 mask based on the current session, check mask when sending client
3552 hello and checking the requested ciphersuite.
3555 *) New ctrls to retrieve and set certificate types in a certificate
3556 request message. Print out received values in s_client. If certificate
3557 types is not set with custom values set sensible values based on
3558 supported signature algorithms.
3561 *) Support for distinct client and server supported signature algorithms.
3564 *) Add certificate callback. If set this is called whenever a certificate
3565 is required by client or server. An application can decide which
3566 certificate chain to present based on arbitrary criteria: for example
3567 supported signature algorithms. Add very simple example to s_server.
3568 This fixes many of the problems and restrictions of the existing client
3569 certificate callback: for example you can now clear an existing
3570 certificate and specify the whole chain.
3573 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3574 the certificate can be used for (if anything). Set valid_flags field
3575 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3576 to have similar checks in it.
3578 Add new "cert_flags" field to CERT structure and include a "strict mode".
3579 This enforces some TLS certificate requirements (such as only permitting
3580 certificate signature algorithms contained in the supported algorithms
3581 extension) which some implementations ignore: this option should be used
3582 with caution as it could cause interoperability issues.
3585 *) Update and tidy signature algorithm extension processing. Work out
3586 shared signature algorithms based on preferences and peer algorithms
3587 and print them out in s_client and s_server. Abort handshake if no
3588 shared signature algorithms.
3591 *) Add new functions to allow customised supported signature algorithms
3592 for SSL and SSL_CTX structures. Add options to s_client and s_server
3596 *) New function SSL_certs_clear() to delete all references to certificates
3597 from an SSL structure. Before this once a certificate had been added
3598 it couldn't be removed.
3601 *) Integrate hostname, email address and IP address checking with certificate
3602 verification. New verify options supporting checking in openssl utility.
3605 *) Fixes and wildcard matching support to hostname and email checking
3606 functions. Add manual page.
3607 [Florian Weimer (Red Hat Product Security Team)]
3609 *) New functions to check a hostname email or IP address against a
3610 certificate. Add options x509 utility to print results of checks against
3614 *) Fix OCSP checking.
3615 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3617 *) Initial experimental support for explicitly trusted non-root CAs.
3618 OpenSSL still tries to build a complete chain to a root but if an
3619 intermediate CA has a trust setting included that is used. The first
3620 setting is used: whether to trust (e.g., -addtrust option to the x509
3624 *) Add -trusted_first option which attempts to find certificates in the
3625 trusted store even if an untrusted chain is also supplied.
3628 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3629 platform support for Linux and Android.
3632 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3635 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3636 When in FIPS mode the approved implementations are used as normal,
3637 when not in FIPS mode the internal unapproved versions are used instead.
3638 This means that the FIPS capable OpenSSL isn't forced to use the
3639 (often lower performance) FIPS implementations outside FIPS mode.
3642 *) Transparently support X9.42 DH parameters when calling
3643 PEM_read_bio_DHparameters. This means existing applications can handle
3644 the new parameter format automatically.
3647 *) Initial experimental support for X9.42 DH parameter format: mainly
3648 to support use of 'q' parameter for RFC5114 parameters.
3651 *) Add DH parameters from RFC5114 including test data to dhtest.
3654 *) Support for automatic EC temporary key parameter selection. If enabled
3655 the most preferred EC parameters are automatically used instead of
3656 hardcoded fixed parameters. Now a server just has to call:
3657 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3658 support ECDH and use the most appropriate parameters.
3661 *) Enhance and tidy EC curve and point format TLS extension code. Use
3662 static structures instead of allocation if default values are used.
3663 New ctrls to set curves we wish to support and to retrieve shared curves.
3664 Print out shared curves in s_server. New options to s_server and s_client
3665 to set list of supported curves.
3668 *) New ctrls to retrieve supported signature algorithms and
3669 supported curve values as an array of NIDs. Extend openssl utility
3670 to print out received values.
3673 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3674 between NIDs and the more common NIST names such as "P-256". Enhance
3675 ecparam utility and ECC method to recognise the NIST names for curves.
3678 *) Enhance SSL/TLS certificate chain handling to support different
3679 chains for each certificate instead of one chain in the parent SSL_CTX.
3682 *) Support for fixed DH ciphersuite client authentication: where both
3683 server and client use DH certificates with common parameters.
3686 *) Support for fixed DH ciphersuites: those requiring DH server
3690 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3692 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3693 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3694 X509_CINF_get_signature were reverted post internal team review.
3696 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3698 *) Build fixes for the Windows and OpenVMS platforms
3699 [Matt Caswell and Richard Levitte]
3701 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3703 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3704 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3705 dereference. This could lead to a Denial Of Service attack. Thanks to
3706 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3710 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3711 dtls1_buffer_record function under certain conditions. In particular this
3712 could occur if an attacker sent repeated DTLS records with the same
3713 sequence number but for the next epoch. The memory leak could be exploited
3714 by an attacker in a Denial of Service attack through memory exhaustion.
3715 Thanks to Chris Mueller for reporting this issue.
3719 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3720 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3721 method would be set to NULL which could later result in a NULL pointer
3722 dereference. Thanks to Frank Schmirler for reporting this issue.
3726 *) Abort handshake if server key exchange message is omitted for ephemeral
3729 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3730 reporting this issue.
3734 *) Remove non-export ephemeral RSA code on client and server. This code
3735 violated the TLS standard by allowing the use of temporary RSA keys in
3736 non-export ciphersuites and could be used by a server to effectively
3737 downgrade the RSA key length used to a value smaller than the server
3738 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3739 INRIA or reporting this issue.
3743 *) Fixed issue where DH client certificates are accepted without verification.
3744 An OpenSSL server will accept a DH certificate for client authentication
3745 without the certificate verify message. This effectively allows a client to
3746 authenticate without the use of a private key. This only affects servers
3747 which trust a client certificate authority which issues certificates
3748 containing DH keys: these are extremely rare and hardly ever encountered.
3749 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3754 *) Ensure that the session ID context of an SSL is updated when its
3755 SSL_CTX is updated via SSL_set_SSL_CTX.
3757 The session ID context is typically set from the parent SSL_CTX,
3758 and can vary with the CTX.
3761 *) Fix various certificate fingerprint issues.
3763 By using non-DER or invalid encodings outside the signed portion of a
3764 certificate the fingerprint can be changed without breaking the signature.
3765 Although no details of the signed portion of the certificate can be changed
3766 this can cause problems with some applications: e.g. those using the
3767 certificate fingerprint for blacklists.
3769 1. Reject signatures with non zero unused bits.
3771 If the BIT STRING containing the signature has non zero unused bits reject
3772 the signature. All current signature algorithms require zero unused bits.
3774 2. Check certificate algorithm consistency.
3776 Check the AlgorithmIdentifier inside TBS matches the one in the
3777 certificate signature. NB: this will result in signature failure
3778 errors for some broken certificates.
3780 Thanks to Konrad Kraszewski from Google for reporting this issue.
3782 3. Check DSA/ECDSA signatures use DER.
3784 Re-encode DSA/ECDSA signatures and compare with the original received
3785 signature. Return an error if there is a mismatch.
3787 This will reject various cases including garbage after signature
3788 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3789 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3790 (negative or with leading zeroes).
3792 Further analysis was conducted and fixes were developed by Stephen Henson
3793 of the OpenSSL core team.
3798 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3799 results on some platforms, including x86_64. This bug occurs at random
3800 with a very low probability, and is not known to be exploitable in any
3801 way, though its exact impact is difficult to determine. Thanks to Pieter
3802 Wuille (Blockstream) who reported this issue and also suggested an initial
3803 fix. Further analysis was conducted by the OpenSSL development team and
3804 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3805 the OpenSSL core team.
3809 *) Do not resume sessions on the server if the negotiated protocol
3810 version does not match the session's version. Resuming with a different
3811 version, while not strictly forbidden by the RFC, is of questionable
3812 sanity and breaks all known clients.
3813 [David Benjamin, Emilia Käsper]
3815 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3816 early CCS messages during renegotiation. (Note that because
3817 renegotiation is encrypted, this early CCS was not exploitable.)
3820 *) Tighten client-side session ticket handling during renegotiation:
3821 ensure that the client only accepts a session ticket if the server sends
3822 the extension anew in the ServerHello. Previously, a TLS client would
3823 reuse the old extension state and thus accept a session ticket if one was
3824 announced in the initial ServerHello.
3826 Similarly, ensure that the client requires a session ticket if one
3827 was advertised in the ServerHello. Previously, a TLS client would
3828 ignore a missing NewSessionTicket message.
3831 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3833 *) SRTP Memory Leak.
3835 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3836 sends a carefully crafted handshake message, to cause OpenSSL to fail
3837 to free up to 64k of memory causing a memory leak. This could be
3838 exploited in a Denial Of Service attack. This issue affects OpenSSL
3839 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3840 whether SRTP is used or configured. Implementations of OpenSSL that
3841 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3843 The fix was developed by the OpenSSL team.
3847 *) Session Ticket Memory Leak.
3849 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3850 integrity of that ticket is first verified. In the event of a session
3851 ticket integrity check failing, OpenSSL will fail to free memory
3852 causing a memory leak. By sending a large number of invalid session
3853 tickets an attacker could exploit this issue in a Denial Of Service
3858 *) Build option no-ssl3 is incomplete.
3860 When OpenSSL is configured with "no-ssl3" as a build option, servers
3861 could accept and complete a SSL 3.0 handshake, and clients could be
3862 configured to send them.
3864 [Akamai and the OpenSSL team]
3866 *) Add support for TLS_FALLBACK_SCSV.
3867 Client applications doing fallback retries should call
3868 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3870 [Adam Langley, Bodo Moeller]
3872 *) Add additional DigestInfo checks.
3874 Re-encode DigestInto in DER and check against the original when
3875 verifying RSA signature: this will reject any improperly encoded
3876 DigestInfo structures.
3878 Note: this is a precautionary measure and no attacks are currently known.
3882 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3884 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3885 SRP code can be overrun an internal buffer. Add sanity check that
3886 g, A, B < N to SRP code.
3888 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3889 Group for discovering this issue.
3893 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3894 TLS 1.0 instead of higher protocol versions when the ClientHello message
3895 is badly fragmented. This allows a man-in-the-middle attacker to force a
3896 downgrade to TLS 1.0 even if both the server and the client support a
3897 higher protocol version, by modifying the client's TLS records.
3899 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3900 researching this issue.
3904 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3905 to a denial of service attack. A malicious server can crash the client
3906 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3907 ciphersuite and sending carefully crafted handshake messages.
3909 Thanks to Felix Gröbert (Google) for discovering and researching this
3914 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3915 to leak memory. This can be exploited through a Denial of Service attack.
3916 Thanks to Adam Langley for discovering and researching this issue.
3920 *) An attacker can force openssl to consume large amounts of memory whilst
3921 processing DTLS handshake messages. This can be exploited through a
3922 Denial of Service attack.
3923 Thanks to Adam Langley for discovering and researching this issue.
3927 *) An attacker can force an error condition which causes openssl to crash
3928 whilst processing DTLS packets due to memory being freed twice. This
3929 can be exploited through a Denial of Service attack.
3930 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3935 *) If a multithreaded client connects to a malicious server using a resumed
3936 session and the server sends an ec point format extension it could write
3937 up to 255 bytes to freed memory.
3939 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3944 *) A malicious server can crash an OpenSSL client with a null pointer
3945 dereference (read) by specifying an SRP ciphersuite even though it was not
3946 properly negotiated with the client. This can be exploited through a
3947 Denial of Service attack.
3949 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3950 discovering and researching this issue.
3954 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3955 X509_name_oneline, X509_name_print_ex et al. to leak some information
3956 from the stack. Applications may be affected if they echo pretty printing
3957 output to the attacker.
3959 Thanks to Ivan Fratric (Google) for discovering this issue.
3961 [Emilia Käsper, and Steve Henson]
3963 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3964 for corner cases. (Certain input points at infinity could lead to
3965 bogus results, with non-infinity inputs mapped to infinity too.)
3968 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3970 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3971 handshake can force the use of weak keying material in OpenSSL
3972 SSL/TLS clients and servers.
3974 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3975 researching this issue. (CVE-2014-0224)
3976 [KIKUCHI Masashi, Steve Henson]
3978 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3979 OpenSSL DTLS client the code can be made to recurse eventually crashing
3982 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3984 [Imre Rad, Steve Henson]
3986 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3987 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3988 client or server. This is potentially exploitable to run arbitrary
3989 code on a vulnerable client or server.
3991 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3992 [Jüri Aedla, Steve Henson]
3994 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3995 are subject to a denial of service attack.
3997 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3998 this issue. (CVE-2014-3470)
3999 [Felix Gröbert, Ivan Fratric, Steve Henson]
4001 *) Harmonize version and its documentation. -f flag is used to display
4003 [mancha <mancha1@zoho.com>]
4005 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
4006 in i2d_ECPrivateKey.
4007 [mancha <mancha1@zoho.com>]
4009 *) Fix some double frees. These are not thought to be exploitable.
4010 [mancha <mancha1@zoho.com>]
4012 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
4014 *) A missing bounds check in the handling of the TLS heartbeat extension
4015 can be used to reveal up to 64k of memory to a connected client or
4018 Thanks for Neel Mehta of Google Security for discovering this bug and to
4019 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4020 preparing the fix (CVE-2014-0160)
4021 [Adam Langley, Bodo Moeller]
4023 *) Fix for the attack described in the paper "Recovering OpenSSL
4024 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4025 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4026 http://eprint.iacr.org/2014/140
4028 Thanks to Yuval Yarom and Naomi Benger for discovering this
4029 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
4030 [Yuval Yarom and Naomi Benger]
4032 *) TLS pad extension: draft-agl-tls-padding-03
4034 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
4035 TLS client Hello record length value would otherwise be > 255 and
4036 less that 512 pad with a dummy extension containing zeroes so it
4037 is at least 512 bytes long.
4039 [Adam Langley, Steve Henson]
4041 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
4043 *) Fix for TLS record tampering bug. A carefully crafted invalid
4044 handshake could crash OpenSSL with a NULL pointer exception.
4045 Thanks to Anton Johansson for reporting this issues.
4048 *) Keep original DTLS digest and encryption contexts in retransmission
4049 structures so we can use the previous session parameters if they need
4050 to be resent. (CVE-2013-6450)
4053 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
4054 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
4055 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
4056 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
4057 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
4058 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
4059 [Rob Stradling, Adam Langley]
4061 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
4063 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
4064 supporting platforms or when small records were transferred.
4065 [Andy Polyakov, Steve Henson]
4067 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
4069 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
4071 This addresses the flaw in CBC record processing discovered by
4072 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
4073 at: http://www.isg.rhul.ac.uk/tls/
4075 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4076 Security Group at Royal Holloway, University of London
4077 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
4078 Emilia Käsper for the initial patch.
4080 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
4082 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
4083 ciphersuites which can be exploited in a denial of service attack.
4084 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
4085 and detecting this bug and to Wolfgang Ettlinger
4086 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
4090 *) Return an error when checking OCSP signatures when key is NULL.
4091 This fixes a DoS attack. (CVE-2013-0166)
4094 *) Make openssl verify return errors.
4095 [Chris Palmer <palmer@google.com> and Ben Laurie]
4097 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
4098 the right response is stapled. Also change SSL_get_certificate()
4099 so it returns the certificate actually sent.
4100 See http://rt.openssl.org/Ticket/Display.html?id=2836.
4101 [Rob Stradling <rob.stradling@comodo.com>]
4103 *) Fix possible deadlock when decoding public keys.
4106 *) Don't use TLS 1.0 record version number in initial client hello
4110 Changes between 1.0.1b and 1.0.1c [10 May 2012]
4112 *) Sanity check record length before skipping explicit IV in TLS
4113 1.2, 1.1 and DTLS to fix DoS attack.
4115 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
4116 fuzzing as a service testing platform.
4120 *) Initialise tkeylen properly when encrypting CMS messages.
4121 Thanks to Solar Designer of Openwall for reporting this issue.
4124 *) In FIPS mode don't try to use composite ciphers as they are not
4128 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
4130 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
4131 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
4132 mean any application compiled against OpenSSL 1.0.0 headers setting
4133 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
4134 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
4135 0x10000000L Any application which was previously compiled against
4136 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
4137 will need to be recompiled as a result. Letting be results in
4138 inability to disable specifically TLS 1.1 and in client context,
4139 in unlike event, limit maximum offered version to TLS 1.0 [see below].
4142 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
4143 disable just protocol X, but all protocols above X *if* there are
4144 protocols *below* X still enabled. In more practical terms it means
4145 that if application wants to disable TLS1.0 in favor of TLS1.1 and
4146 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
4147 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
4151 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
4153 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
4154 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
4155 in CRYPTO_realloc_clean.
4157 Thanks to Tavis Ormandy, Google Security Team, for discovering this
4158 issue and to Adam Langley <agl@chromium.org> for fixing it.
4160 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
4162 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
4165 *) Workarounds for some broken servers that "hang" if a client hello
4166 record length exceeds 255 bytes.
4168 1. Do not use record version number > TLS 1.0 in initial client
4169 hello: some (but not all) hanging servers will now work.
4170 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
4171 the number of ciphers sent in the client hello. This should be
4172 set to an even number, such as 50, for example by passing:
4173 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
4174 Most broken servers should now work.
4175 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
4176 TLS 1.2 client support entirely.
4179 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
4182 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
4184 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
4185 STRING form instead of a DigestInfo.
4188 *) The format used for MDC2 RSA signatures is inconsistent between EVP
4189 and the RSA_sign/RSA_verify functions. This was made more apparent when
4190 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
4191 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
4192 the correct format in RSA_verify so both forms transparently work.
4195 *) Some servers which support TLS 1.0 can choke if we initially indicate
4196 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
4197 encrypted premaster secret. As a workaround use the maximum permitted
4198 client version in client hello, this should keep such servers happy
4199 and still work with previous versions of OpenSSL.
4202 *) Add support for TLS/DTLS heartbeats.
4203 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4205 *) Add support for SCTP.
4206 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4208 *) Improved PRNG seeding for VOS.
4209 [Paul Green <Paul.Green@stratus.com>]
4211 *) Extensive assembler packs updates, most notably:
4213 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
4214 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
4215 - x86_64: bit-sliced AES implementation;
4216 - ARM: NEON support, contemporary platforms optimizations;
4217 - s390x: z196 support;
4218 - *: GHASH and GF(2^m) multiplication implementations;
4222 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
4223 (removal of unnecessary code)
4224 [Peter Sylvester <peter.sylvester@edelweb.fr>]
4226 *) Add TLS key material exporter from RFC 5705.
4229 *) Add DTLS-SRTP negotiation from RFC 5764.
4232 *) Add Next Protocol Negotiation,
4233 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
4234 disabled with a no-npn flag to config or Configure. Code donated
4236 [Adam Langley <agl@google.com> and Ben Laurie]
4238 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
4239 NIST-P256, NIST-P521, with constant-time single point multiplication on
4240 typical inputs. Compiler support for the nonstandard type __uint128_t is
4241 required to use this (present in gcc 4.4 and later, for 64-bit builds).
4242 Code made available under Apache License version 2.0.
4244 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
4245 line to include this in your build of OpenSSL, and run "make depend" (or
4246 "make update"). This enables the following EC_METHODs:
4248 EC_GFp_nistp224_method()
4249 EC_GFp_nistp256_method()
4250 EC_GFp_nistp521_method()
4252 EC_GROUP_new_by_curve_name() will automatically use these (while
4253 EC_GROUP_new_curve_GFp() currently prefers the more flexible
4255 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
4257 *) Use type ossl_ssize_t instead of ssize_t which isn't available on
4258 all platforms. Move ssize_t definition from e_os.h to the public
4259 header file e_os2.h as it now appears in public header file cms.h
4262 *) New -sigopt option to the ca, req and x509 utilities. Additional
4263 signature parameters can be passed using this option and in
4267 *) Add RSA PSS signing function. This will generate and set the
4268 appropriate AlgorithmIdentifiers for PSS based on those in the
4269 corresponding EVP_MD_CTX structure. No application support yet.
4272 *) Support for companion algorithm specific ASN1 signing routines.
4273 New function ASN1_item_sign_ctx() signs a pre-initialised
4274 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
4275 the appropriate parameters.
4278 *) Add new algorithm specific ASN1 verification initialisation function
4279 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
4280 handling will be the same no matter what EVP_PKEY_METHOD is used.
4281 Add a PSS handler to support verification of PSS signatures: checked
4282 against a number of sample certificates.
4285 *) Add signature printing for PSS. Add PSS OIDs.
4286 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
4288 *) Add algorithm specific signature printing. An individual ASN1 method
4289 can now print out signatures instead of the standard hex dump.
4291 More complex signatures (e.g. PSS) can print out more meaningful
4292 information. Include DSA version that prints out the signature
4296 *) Password based recipient info support for CMS library: implementing
4300 *) Split password based encryption into PBES2 and PBKDF2 functions. This
4301 neatly separates the code into cipher and PBE sections and is required
4302 for some algorithms that split PBES2 into separate pieces (such as
4303 password based CMS).
4306 *) Session-handling fixes:
4307 - Fix handling of connections that are resuming with a session ID,
4308 but also support Session Tickets.
4309 - Fix a bug that suppressed issuing of a new ticket if the client
4310 presented a ticket with an expired session.
4311 - Try to set the ticket lifetime hint to something reasonable.
4312 - Make tickets shorter by excluding irrelevant information.
4313 - On the client side, don't ignore renewed tickets.
4314 [Adam Langley, Bodo Moeller (Google)]
4316 *) Fix PSK session representation.
4319 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
4321 This work was sponsored by Intel.
4324 *) Add GCM support to TLS library. Some custom code is needed to split
4325 the IV between the fixed (from PRF) and explicit (from TLS record)
4326 portions. This adds all GCM ciphersuites supported by RFC5288 and
4327 RFC5289. Generalise some AES* cipherstrings to include GCM and
4328 add a special AESGCM string for GCM only.
4331 *) Expand range of ctrls for AES GCM. Permit setting invocation
4332 field on decrypt and retrieval of invocation field only on encrypt.
4335 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
4336 As required by RFC5289 these ciphersuites cannot be used if for
4337 versions of TLS earlier than 1.2.
4340 *) For FIPS capable OpenSSL interpret a NULL default public key method
4341 as unset and return the appropriate default but do *not* set the default.
4342 This means we can return the appropriate method in applications that
4343 switch between FIPS and non-FIPS modes.
4346 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
4347 ENGINE is used then we cannot handle that in the FIPS module so we
4348 keep original code iff non-FIPS operations are allowed.
4351 *) Add -attime option to openssl utilities.
4352 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
4354 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
4357 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
4358 FIPS EC methods unconditionally for now.
4361 *) New build option no-ec2m to disable characteristic 2 code.
4364 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
4365 all cases can be covered as some introduce binary incompatibilities.
4368 *) Redirect RSA operations to FIPS module including keygen,
4369 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
4372 *) Add similar low level API blocking to ciphers.
4375 *) Low level digest APIs are not approved in FIPS mode: any attempt
4376 to use these will cause a fatal error. Applications that *really* want
4377 to use them can use the private_* version instead.
4380 *) Redirect cipher operations to FIPS module for FIPS builds.
4383 *) Redirect digest operations to FIPS module for FIPS builds.
4386 *) Update build system to add "fips" flag which will link in fipscanister.o
4387 for static and shared library builds embedding a signature if needed.
4390 *) Output TLS supported curves in preference order instead of numerical
4391 order. This is currently hardcoded for the highest order curves first.
4392 This should be configurable so applications can judge speed vs strength.
4395 *) Add TLS v1.2 server support for client authentication.
4398 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
4402 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
4403 FIPS modules versions.
4406 *) Add TLS v1.2 client side support for client authentication. Keep cache
4407 of handshake records longer as we don't know the hash algorithm to use
4408 until after the certificate request message is received.
4411 *) Initial TLS v1.2 client support. Add a default signature algorithms
4412 extension including all the algorithms we support. Parse new signature
4413 format in client key exchange. Relax some ECC signing restrictions for
4414 TLS v1.2 as indicated in RFC5246.
4417 *) Add server support for TLS v1.2 signature algorithms extension. Switch
4418 to new signature format when needed using client digest preference.
4419 All server ciphersuites should now work correctly in TLS v1.2. No client
4420 support yet and no support for client certificates.
4423 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
4424 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
4425 ciphersuites. At present only RSA key exchange ciphersuites work with
4426 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
4427 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
4428 and version checking.
4431 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
4432 with this defined it will not be affected by any changes to ssl internal
4433 structures. Add several utility functions to allow openssl application
4434 to work with OPENSSL_NO_SSL_INTERN defined.
4437 *) A long standing patch to add support for SRP from EdelWeb (Peter
4438 Sylvester and Christophe Renou) was integrated.
4439 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
4440 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
4443 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
4446 *) Permit abbreviated handshakes when renegotiating using the function
4447 SSL_renegotiate_abbreviated().
4448 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4450 *) Add call to ENGINE_register_all_complete() to
4451 ENGINE_load_builtin_engines(), so some implementations get used
4452 automatically instead of needing explicit application support.
4455 *) Add support for TLS key exporter as described in RFC5705.
4456 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
4458 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
4459 a few changes are required:
4461 Add SSL_OP_NO_TLSv1_1 flag.
4462 Add TLSv1_1 methods.
4463 Update version checking logic to handle version 1.1.
4464 Add explicit IV handling (ported from DTLS code).
4465 Add command line options to s_client/s_server.
4468 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
4470 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
4471 in CMS and PKCS7 code. When RSA decryption fails use a random key for
4472 content decryption and always return the same error. Note: this attack
4473 needs on average 2^20 messages so it only affects automated senders. The
4474 old behaviour can be re-enabled in the CMS code by setting the
4475 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
4476 an MMA defence is not necessary.
4477 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
4478 this issue. (CVE-2012-0884)
4481 *) Fix CVE-2011-4619: make sure we really are receiving a
4482 client hello before rejecting multiple SGC restarts. Thanks to
4483 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
4486 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
4488 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
4489 Thanks to Antonio Martin, Enterprise Secure Access Research and
4490 Development, Cisco Systems, Inc. for discovering this bug and
4491 preparing a fix. (CVE-2012-0050)
4494 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
4496 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
4497 of the Vaudenay padding oracle attack on CBC mode encryption
4498 which enables an efficient plaintext recovery attack against
4499 the OpenSSL implementation of DTLS. Their attack exploits timing
4500 differences arising during decryption processing. A research
4501 paper describing this attack can be found at:
4502 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
4503 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4504 Security Group at Royal Holloway, University of London
4505 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
4506 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
4507 for preparing the fix. (CVE-2011-4108)
4508 [Robin Seggelmann, Michael Tuexen]
4510 *) Clear bytes used for block padding of SSL 3.0 records.
4512 [Adam Langley (Google)]
4514 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
4515 Kadianakis <desnacked@gmail.com> for discovering this issue and
4516 Adam Langley for preparing the fix. (CVE-2011-4619)
4517 [Adam Langley (Google)]
4519 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
4520 [Andrey Kulikov <amdeich@gmail.com>]
4522 *) Prevent malformed RFC3779 data triggering an assertion failure.
4523 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
4524 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
4525 [Rob Austein <sra@hactrn.net>]
4527 *) Improved PRNG seeding for VOS.
4528 [Paul Green <Paul.Green@stratus.com>]
4530 *) Fix ssl_ciph.c set-up race.
4531 [Adam Langley (Google)]
4533 *) Fix spurious failures in ecdsatest.c.
4534 [Emilia Käsper (Google)]
4536 *) Fix the BIO_f_buffer() implementation (which was mixing different
4537 interpretations of the '..._len' fields).
4538 [Adam Langley (Google)]
4540 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4541 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4542 threads won't reuse the same blinding coefficients.
4544 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4545 lock to call BN_BLINDING_invert_ex, and avoids one use of
4546 BN_BLINDING_update for each BN_BLINDING structure (previously,
4547 the last update always remained unused).
4548 [Emilia Käsper (Google)]
4550 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4551 [Bob Buckholz (Google)]
4553 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4555 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4556 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4557 [Kaspar Brand <ossl@velox.ch>]
4559 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4560 for multi-threaded use of ECDH. (CVE-2011-3210)
4561 [Adam Langley (Google)]
4563 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4566 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4567 signature public key algorithm by using OID xref utilities instead.
4568 Before this you could only use some ECC ciphersuites with SHA1 only.
4571 *) Add protection against ECDSA timing attacks as mentioned in the paper
4572 by Billy Bob Brumley and Nicola Tuveri, see:
4574 http://eprint.iacr.org/2011/232.pdf
4576 [Billy Bob Brumley and Nicola Tuveri]
4578 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4580 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4581 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4583 *) Fix bug in string printing code: if *any* escaping is enabled we must
4584 escape the escape character (backslash) or the resulting string is
4588 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4590 *) Disable code workaround for ancient and obsolete Netscape browsers
4591 and servers: an attacker can use it in a ciphersuite downgrade attack.
4592 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4595 *) Fixed J-PAKE implementation error, originally discovered by
4596 Sebastien Martini, further info and confirmation from Stefan
4597 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4600 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4602 *) Fix extension code to avoid race conditions which can result in a buffer
4603 overrun vulnerability: resumed sessions must not be modified as they can
4604 be shared by multiple threads. CVE-2010-3864
4607 *) Fix WIN32 build system to correctly link an ENGINE directory into
4611 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4613 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4615 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4617 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4619 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4620 context. The operation can be customised via the ctrl mechanism in
4621 case ENGINEs want to include additional functionality.
4624 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4627 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4628 output hashes compatible with older versions of OpenSSL.
4629 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4631 *) Fix compression algorithm handling: if resuming a session use the
4632 compression algorithm of the resumed session instead of determining
4633 it from client hello again. Don't allow server to change algorithm.
4636 *) Add load_crls() function to apps tidying load_certs() too. Add option
4637 to verify utility to allow additional CRLs to be included.
4640 *) Update OCSP request code to permit adding custom headers to the request:
4641 some responders need this.
4644 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4646 [Julia Lawall <julia@diku.dk>]
4648 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4649 needlessly dereferenced structures, used obsolete functions and
4650 didn't handle all updated verify codes correctly.
4653 *) Disable MD2 in the default configuration.
4656 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4657 indicate the initial BIO being pushed or popped. This makes it possible
4658 to determine whether the BIO is the one explicitly called or as a result
4659 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4660 it handles reference counts correctly and doesn't zero out the I/O bio
4661 when it is not being explicitly popped. WARNING: applications which
4662 included workarounds for the old buggy behaviour will need to be modified
4663 or they could free up already freed BIOs.
4666 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4667 renaming to all platforms (within the 0.9.8 branch, this was
4668 done conditionally on Netware platforms to avoid a name clash).
4669 [Guenter <lists@gknw.net>]
4671 *) Add ECDHE and PSK support to DTLS.
4672 [Michael Tuexen <tuexen@fh-muenster.de>]
4674 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4678 *) Add "missing" function EVP_MD_flags() (without this the only way to
4679 retrieve a digest flags is by accessing the structure directly. Update
4680 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4681 or cipher is registered as in the "from" argument. Print out all
4682 registered digests in the dgst usage message instead of manually
4683 attempting to work them out.
4686 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4687 this allows the use of compression and extensions. Change default cipher
4688 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4689 by default unless an application cipher string requests it.
4692 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4693 key ids to find matching certificates and keys but some PKCS#12 files
4694 don't follow the (somewhat unwritten) rules and this strategy fails.
4695 Now just gather all certificates together and the first private key
4696 then look for the first certificate that matches the key.
4699 *) Support use of registered digest and cipher names for dgst and cipher
4700 commands instead of having to add each one as a special case. So now
4707 openssl dgst -sha256 foo
4709 and this works for ENGINE based algorithms too.
4713 *) Update Gost ENGINE to support parameter files.
4714 [Victor B. Wagner <vitus@cryptocom.ru>]
4716 *) Support GeneralizedTime in ca utility.
4717 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4719 *) Enhance the hash format used for certificate directory links. The new
4720 form uses the canonical encoding (meaning equivalent names will work
4721 even if they aren't identical) and uses SHA1 instead of MD5. This form
4722 is incompatible with the older format and as a result c_rehash should
4723 be used to rebuild symbolic links.
4726 *) Make PKCS#8 the default write format for private keys, replacing the
4727 traditional format. This form is standardised, more secure and doesn't
4728 include an implicit MD5 dependency.
4731 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4732 committed to OpenSSL should pass this lot as a minimum.
4735 *) Add session ticket override functionality for use by EAP-FAST.
4736 [Jouni Malinen <j@w1.fi>]
4738 *) Modify HMAC functions to return a value. Since these can be implemented
4739 in an ENGINE errors can occur.
4742 *) Type-checked OBJ_bsearch_ex.
4745 *) Type-checked OBJ_bsearch. Also some constification necessitated
4746 by type-checking. Still to come: TXT_DB, bsearch(?),
4747 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4751 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4752 seconds to a tm structure directly, instead of going through OS
4753 specific date routines. This avoids any issues with OS routines such
4754 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4755 and X509_time_adj_ex() to cover the extended range. The existing
4756 X509_time_adj() is still usable and will no longer have any date issues.
4759 *) Delta CRL support. New use deltas option which will attempt to locate
4760 and search any appropriate delta CRLs available.
4762 This work was sponsored by Google.
4765 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4766 code and add additional score elements. Validate alternate CRL paths
4767 as part of the CRL checking and indicate a new error "CRL path validation
4768 error" in this case. Applications wanting additional details can use
4769 the verify callback and check the new "parent" field. If this is not
4770 NULL CRL path validation is taking place. Existing applications won't
4771 see this because it requires extended CRL support which is off by
4774 This work was sponsored by Google.
4777 *) Support for freshest CRL extension.
4779 This work was sponsored by Google.
4782 *) Initial indirect CRL support. Currently only supported in the CRLs
4783 passed directly and not via lookup. Process certificate issuer
4784 CRL entry extension and lookup CRL entries by bother issuer name
4785 and serial number. Check and process CRL issuer entry in IDP extension.
4787 This work was sponsored by Google.
4790 *) Add support for distinct certificate and CRL paths. The CRL issuer
4791 certificate is validated separately in this case. Only enabled if
4792 an extended CRL support flag is set: this flag will enable additional
4793 CRL functionality in future.
4795 This work was sponsored by Google.
4798 *) Add support for policy mappings extension.
4800 This work was sponsored by Google.
4803 *) Fixes to pathlength constraint, self issued certificate handling,
4804 policy processing to align with RFC3280 and PKITS tests.
4806 This work was sponsored by Google.
4809 *) Support for name constraints certificate extension. DN, email, DNS
4810 and URI types are currently supported.
4812 This work was sponsored by Google.
4815 *) To cater for systems that provide a pointer-based thread ID rather
4816 than numeric, deprecate the current numeric thread ID mechanism and
4817 replace it with a structure and associated callback type. This
4818 mechanism allows a numeric "hash" to be extracted from a thread ID in
4819 either case, and on platforms where pointers are larger than 'long',
4820 mixing is done to help ensure the numeric 'hash' is usable even if it
4821 can't be guaranteed unique. The default mechanism is to use "&errno"
4822 as a pointer-based thread ID to distinguish between threads.
4824 Applications that want to provide their own thread IDs should now use
4825 CRYPTO_THREADID_set_callback() to register a callback that will call
4826 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4828 Note that ERR_remove_state() is now deprecated, because it is tied
4829 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4830 to free the current thread's error state should be replaced by
4831 ERR_remove_thread_state(NULL).
4833 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4834 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4835 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4836 application was previously providing a numeric thread callback that
4837 was inappropriate for distinguishing threads, then uniqueness might
4838 have been obtained with &errno that happened immediately in the
4839 intermediate development versions of OpenSSL; this is no longer the
4840 case, the numeric thread callback will now override the automatic use
4842 [Geoff Thorpe, with help from Bodo Moeller]
4844 *) Initial support for different CRL issuing certificates. This covers a
4845 simple case where the self issued certificates in the chain exist and
4846 the real CRL issuer is higher in the existing chain.
4848 This work was sponsored by Google.
4851 *) Removed effectively defunct crypto/store from the build.
4854 *) Revamp of STACK to provide stronger type-checking. Still to come:
4855 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4856 ASN1_STRING, CONF_VALUE.
4859 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4860 RAM on SSL connections. This option can save about 34k per idle SSL.
4863 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4864 STACK, TXT_DB, bsearch, qsort.
4867 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4868 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4869 support for data, signedData, compressedData, digestedData and
4870 encryptedData, envelopedData types included. Scripts to check against
4871 RFC4134 examples draft and interop and consistency checks of many
4872 content types and variants.
4875 *) Add options to enc utility to support use of zlib compression BIO.
4878 *) Extend mk1mf to support importing of options and assembly language
4879 files from Configure script, currently only included in VC-WIN32.
4880 The assembly language rules can now optionally generate the source
4881 files from the associated perl scripts.
4884 *) Implement remaining functionality needed to support GOST ciphersuites.
4885 Interop testing has been performed using CryptoPro implementations.
4886 [Victor B. Wagner <vitus@cryptocom.ru>]
4888 *) s390x assembler pack.
4891 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4895 *) Implement Opaque PRF Input TLS extension as specified in
4896 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4897 official specification yet and no extension type assignment by
4898 IANA exists, this extension (for now) will have to be explicitly
4899 enabled when building OpenSSL by providing the extension number
4900 to use. For example, specify an option
4902 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4904 to the "config" or "Configure" script to enable the extension,
4905 assuming extension number 0x9527 (which is a completely arbitrary
4906 and unofficial assignment based on the MD5 hash of the Internet
4907 Draft). Note that by doing so, you potentially lose
4908 interoperability with other TLS implementations since these might
4909 be using the same extension number for other purposes.
4911 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4912 opaque PRF input value to use in the handshake. This will create
4913 an internal copy of the length-'len' string at 'src', and will
4914 return non-zero for success.
4916 To get more control and flexibility, provide a callback function
4919 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4920 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4924 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4927 Callback function 'cb' will be called in handshakes, and is
4928 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4929 Argument 'arg' is for application purposes (the value as given to
4930 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4931 be provided to the callback function). The callback function
4932 has to return non-zero to report success: usually 1 to use opaque
4933 PRF input just if possible, or 2 to enforce use of the opaque PRF
4934 input. In the latter case, the library will abort the handshake
4935 if opaque PRF input is not successfully negotiated.
4937 Arguments 'peerinput' and 'len' given to the callback function
4938 will always be NULL and 0 in the case of a client. A server will
4939 see the client's opaque PRF input through these variables if
4940 available (NULL and 0 otherwise). Note that if the server
4941 provides an opaque PRF input, the length must be the same as the
4942 length of the client's opaque PRF input.
4944 Note that the callback function will only be called when creating
4945 a new session (session resumption can resume whatever was
4946 previously negotiated), and will not be called in SSL 2.0
4947 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4948 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4949 for applications that need to enforce opaque PRF input.
4953 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4956 [Victor B. Wagner <vitus@cryptocom.ru>]
4958 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4959 RFC4507bis. The encrypted ticket format is an encrypted encoded
4960 SSL_SESSION structure, that way new session features are automatically
4963 If a client application caches session in an SSL_SESSION structure
4964 support is transparent because tickets are now stored in the encoded
4967 The SSL_CTX structure automatically generates keys for ticket
4968 protection in servers so again support should be possible
4969 with no application modification.
4971 If a client or server wishes to disable RFC4507 support then the option
4972 SSL_OP_NO_TICKET can be set.
4974 Add a TLS extension debugging callback to allow the contents of any client
4975 or server extensions to be examined.
4977 This work was sponsored by Google.
4980 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4981 OpenSSL should now compile cleanly on gcc 4.2
4982 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4984 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4985 support including streaming MAC support: this is required for GOST
4986 ciphersuite support.
4987 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4989 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4990 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4991 to output in BER and PEM format.
4994 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4995 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4996 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4997 ENGINE support for HMAC keys which are unextractable. New -mac and
4998 -macopt options to dgst utility.
5001 *) New option -sigopt to dgst utility. Update dgst to use
5002 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
5003 alternative signing parameters such as X9.31 or PSS in the dgst
5007 *) Change ssl_cipher_apply_rule(), the internal function that does
5008 the work each time a ciphersuite string requests enabling
5009 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
5010 removing ("!foo+bar") a class of ciphersuites: Now it maintains
5011 the order of disabled ciphersuites such that those ciphersuites
5012 that most recently went from enabled to disabled not only stay
5013 in order with respect to each other, but also have higher priority
5014 than other disabled ciphersuites the next time ciphersuites are
5017 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
5018 the same ciphersuites as with "HIGH" alone, but in a specific
5019 order where the PSK ciphersuites come first (since they are the
5020 most recently disabled ciphersuites when "HIGH" is parsed).
5022 Also, change ssl_create_cipher_list() (using this new
5023 functionality) such that between otherwise identical
5024 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
5028 *) Change ssl_create_cipher_list() so that it automatically
5029 arranges the ciphersuites in reasonable order before starting
5030 to process the rule string. Thus, the definition for "DEFAULT"
5031 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
5032 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
5033 This makes it much easier to arrive at a reasonable default order
5034 in applications for which anonymous ciphers are OK (meaning
5035 that you can't actually use DEFAULT).
5036 [Bodo Moeller; suggested by Victor Duchovni]
5038 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
5039 processing) into multiple integers instead of setting
5040 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
5041 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
5042 (These masks as well as the individual bit definitions are hidden
5043 away into the non-exported interface ssl/ssl_locl.h, so this
5044 change to the definition of the SSL_CIPHER structure shouldn't
5045 affect applications.) This give us more bits for each of these
5046 categories, so there is no longer a need to coagulate AES128 and
5047 AES256 into a single algorithm bit, and to coagulate Camellia128
5048 and Camellia256 into a single algorithm bit, which has led to all
5051 Thus, among other things, the kludge introduced in 0.9.7m and
5052 0.9.8e for masking out AES256 independently of AES128 or masking
5053 out Camellia256 independently of AES256 is not needed here in 0.9.9.
5055 With the change, we also introduce new ciphersuite aliases that
5056 so far were missing: "AES128", "AES256", "CAMELLIA128", and
5060 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
5061 Use the leftmost N bytes of the signature input if the input is
5062 larger than the prime q (with N being the size in bytes of q).
5065 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
5066 it yet and it is largely untested.
5069 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
5072 *) Initial incomplete changes to avoid need for function casts in OpenSSL
5073 some compilers (gcc 4.2 and later) reject their use. Safestack is
5074 reimplemented. Update ASN1 to avoid use of legacy functions.
5077 *) Win32/64 targets are linked with Winsock2.
5080 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
5081 to external functions. This can be used to increase CRL handling
5082 efficiency especially when CRLs are very large by (for example) storing
5083 the CRL revoked certificates in a database.
5086 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
5087 new CRLs added to a directory can be used. New command line option
5088 -verify_return_error to s_client and s_server. This causes real errors
5089 to be returned by the verify callback instead of carrying on no matter
5090 what. This reflects the way a "real world" verify callback would behave.
5093 *) GOST engine, supporting several GOST algorithms and public key formats.
5094 Kindly donated by Cryptocom.
5097 *) Partial support for Issuing Distribution Point CRL extension. CRLs
5098 partitioned by DP are handled but no indirect CRL or reason partitioning
5099 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
5100 selected via a scoring technique which handles IDP and AKID in CRLs.
5103 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
5104 will ultimately be used for all verify operations: this will remove the
5105 X509_STORE dependency on certificate verification and allow alternative
5106 lookup methods. X509_STORE based implementations of these two callbacks.
5109 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
5110 Modify get_crl() to find a valid (unexpired) CRL if possible.
5113 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
5114 this would be called X509_CRL_cmp() but that name is already used by
5115 a function that just compares CRL issuer names. Cache several CRL
5116 extensions in X509_CRL structure and cache CRLDP in X509.
5119 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
5120 this maps equivalent X509_NAME structures into a consistent structure.
5121 Name comparison can then be performed rapidly using memcmp().
5124 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
5128 *) Allow digests to supply their own micalg string for S/MIME type using
5129 the ctrl EVP_MD_CTRL_MICALG.
5132 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
5133 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
5134 ctrl. It can then customise the structure before and/or after signing
5138 *) New function OBJ_add_sigid() to allow application defined signature OIDs
5139 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
5140 to free up any added signature OIDs.
5143 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
5144 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
5145 digest and cipher tables. New options added to openssl utility:
5146 list-message-digest-algorithms and list-cipher-algorithms.
5149 *) Change the array representation of binary polynomials: the list
5150 of degrees of non-zero coefficients is now terminated with -1.
5151 Previously it was terminated with 0, which was also part of the
5152 value; thus, the array representation was not applicable to
5153 polynomials where t^0 has coefficient zero. This change makes
5154 the array representation useful in a more general context.
5157 *) Various modifications and fixes to SSL/TLS cipher string
5158 handling. For ECC, the code now distinguishes between fixed ECDH
5159 with RSA certificates on the one hand and with ECDSA certificates
5160 on the other hand, since these are separate ciphersuites. The
5161 unused code for Fortezza ciphersuites has been removed.
5163 For consistency with EDH, ephemeral ECDH is now called "EECDH"
5164 (not "ECDHE"). For consistency with the code for DH
5165 certificates, use of ECDH certificates is now considered ECDH
5166 authentication, not RSA or ECDSA authentication (the latter is
5167 merely the CA's signing algorithm and not actively used in the
5170 The temporary ciphersuite alias "ECCdraft" is no longer
5171 available, and ECC ciphersuites are no longer excluded from "ALL"
5172 and "DEFAULT". The following aliases now exist for RFC 4492
5173 ciphersuites, most of these by analogy with the DH case:
5175 kECDHr - ECDH cert, signed with RSA
5176 kECDHe - ECDH cert, signed with ECDSA
5177 kECDH - ECDH cert (signed with either RSA or ECDSA)
5178 kEECDH - ephemeral ECDH
5179 ECDH - ECDH cert or ephemeral ECDH
5185 AECDH - anonymous ECDH
5186 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
5190 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
5191 Use correct micalg parameters depending on digest(s) in signed message.
5194 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
5195 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
5198 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
5199 an engine to register a method. Add ENGINE lookups for methods and
5200 functional reference processing.
5203 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
5204 EVP_{Sign,Verify}* which allow an application to customise the signature
5208 *) New -resign option to smime utility. This adds one or more signers
5209 to an existing PKCS#7 signedData structure. Also -md option to use an
5210 alternative message digest algorithm for signing.
5213 *) Tidy up PKCS#7 routines and add new functions to make it easier to
5214 create PKCS7 structures containing multiple signers. Update smime
5215 application to support multiple signers.
5218 *) New -macalg option to pkcs12 utility to allow setting of an alternative
5222 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
5223 Reorganize PBE internals to lookup from a static table using NIDs,
5224 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
5225 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
5226 PRF which will be automatically used with PBES2.
5229 *) Replace the algorithm specific calls to generate keys in "req" with the
5233 *) Update PKCS#7 enveloped data routines to use new API. This is now
5234 supported by any public key method supporting the encrypt operation. A
5235 ctrl is added to allow the public key algorithm to examine or modify
5236 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
5240 *) Add a ctrl to asn1 method to allow a public key algorithm to express
5241 a default digest type to use. In most cases this will be SHA1 but some
5242 algorithms (such as GOST) need to specify an alternative digest. The
5243 return value indicates how strong the preference is 1 means optional and
5244 2 is mandatory (that is it is the only supported type). Modify
5245 ASN1_item_sign() to accept a NULL digest argument to indicate it should
5246 use the default md. Update openssl utilities to use the default digest
5247 type for signing if it is not explicitly indicated.
5250 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
5251 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
5252 signing method from the key type. This effectively removes the link
5253 between digests and public key types.
5256 *) Add an OID cross reference table and utility functions. Its purpose is to
5257 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
5258 rsaEncryption. This will allow some of the algorithm specific hackery
5259 needed to use the correct OID to be removed.
5262 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
5263 structures for PKCS7_sign(). They are now set up by the relevant public
5267 *) Add provisional EC pkey method with support for ECDSA and ECDH.
5270 *) Add support for key derivation (agreement) in the API, DH method and
5274 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
5275 public and private key formats. As a side effect these add additional
5276 command line functionality not previously available: DSA signatures can be
5277 generated and verified using pkeyutl and DH key support and generation in
5282 [Oliver Tappe <zooey@hirschkaefer.de>]
5284 *) New make target "install_html_docs" installs HTML renditions of the
5286 [Oliver Tappe <zooey@hirschkaefer.de>]
5288 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
5289 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
5290 support key and parameter generation and add initial key generation
5291 functionality for RSA.
5294 *) Add functions for main EVP_PKEY_method operations. The undocumented
5295 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
5296 EVP_PKEY_{encrypt,decrypt}_old.
5299 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
5300 key API, doesn't do much yet.
5303 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
5304 public key algorithms. New option to openssl utility:
5305 "list-public-key-algorithms" to print out info.
5308 *) Implement the Supported Elliptic Curves Extension for
5309 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5312 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
5313 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
5316 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
5317 utilities such as rsa, dsa, dsaparam etc except they process any key
5321 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
5322 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
5323 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
5327 *) Initial support for pluggable public key ASN1.
5328 De-spaghettify the public key ASN1 handling. Move public and private
5329 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
5330 algorithm specific handling to a single module within the relevant
5331 algorithm directory. Add functions to allow (near) opaque processing
5332 of public and private key structures.
5335 *) Implement the Supported Point Formats Extension for
5336 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5339 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
5340 for the psk identity [hint] and the psk callback functions to the
5341 SSL_SESSION, SSL and SSL_CTX structure.
5344 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
5348 SSL_CTX_use_psk_identity_hint
5349 SSL_get_psk_identity_hint
5350 SSL_get_psk_identity
5351 SSL_use_psk_identity_hint
5353 [Mika Kousa and Pasi Eronen of Nokia Corporation]
5355 *) Add RFC 3161 compliant time stamp request creation, response generation
5356 and response verification functionality.
5357 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
5359 *) Add initial support for TLS extensions, specifically for the server_name
5360 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5361 have new members for a host name. The SSL data structure has an
5362 additional member SSL_CTX *initial_ctx so that new sessions can be
5363 stored in that context to allow for session resumption, even after the
5364 SSL has been switched to a new SSL_CTX in reaction to a client's
5365 server_name extension.
5367 New functions (subject to change):
5369 SSL_get_servername()
5370 SSL_get_servername_type()
5373 New CTRL codes and macros (subject to change):
5375 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5376 - SSL_CTX_set_tlsext_servername_callback()
5377 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5378 - SSL_CTX_set_tlsext_servername_arg()
5379 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5381 openssl s_client has a new '-servername ...' option.
5383 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5384 '-key2 ...', '-servername_fatal' (subject to change). This allows
5385 testing the HostName extension for a specific single host name ('-cert'
5386 and '-key' remain fallbacks for handshakes without HostName
5387 negotiation). If the unrecognized_name alert has to be sent, this by
5388 default is a warning; it becomes fatal with the '-servername_fatal'
5391 [Peter Sylvester, Remy Allais, Christophe Renou]
5393 *) Whirlpool hash implementation is added.
5396 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
5397 bn(64,32). Because of instruction set limitations it doesn't have
5398 any negative impact on performance. This was done mostly in order
5399 to make it possible to share assembler modules, such as bn_mul_mont
5400 implementations, between 32- and 64-bit builds without hassle.
5403 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
5404 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
5408 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
5409 dedicated Montgomery multiplication procedure, is introduced.
5410 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
5411 "64-bit" performance on certain 32-bit targets.
5414 *) New option SSL_OP_NO_COMP to disable use of compression selectively
5415 in SSL structures. New SSL ctrl to set maximum send fragment size.
5416 Save memory by setting the I/O buffer sizes dynamically instead of
5417 using the maximum available value.
5420 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
5421 in addition to the text details.
5424 *) Very, very preliminary EXPERIMENTAL support for printing of general
5425 ASN1 structures. This currently produces rather ugly output and doesn't
5426 handle several customised structures at all.
5429 *) Integrated support for PVK file format and some related formats such
5430 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
5431 these in the 'rsa' and 'dsa' utilities.
5434 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
5437 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
5438 place for the (very old) "NETSCAPE" format certificates which are now
5439 handled using new ASN1 code equivalents.
5442 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
5443 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
5444 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
5447 *) Modify CRL distribution points extension code to print out previously
5448 unsupported fields. Enhance extension setting code to allow setting of
5452 *) Add print and set support for Issuing Distribution Point CRL extension.
5455 *) Change 'Configure' script to enable Camellia by default.
5458 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
5460 *) When rejecting SSL/TLS records due to an incorrect version number, never
5461 update s->server with a new major version number. As of
5462 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
5463 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
5464 the previous behavior could result in a read attempt at NULL when
5465 receiving specific incorrect SSL/TLS records once record payload
5466 protection is active. (CVE-2010-0740)
5467 [Bodo Moeller, Adam Langley <agl@chromium.org>]
5469 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
5470 could be crashed if the relevant tables were not present (e.g. chrooted).
5471 [Tomas Hoger <thoger@redhat.com>]
5473 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
5475 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
5476 [Martin Olsson, Neel Mehta]
5478 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
5479 accommodate for stack sorting, always a write lock!).
5482 *) On some versions of WIN32 Heap32Next is very slow. This can cause
5483 excessive delays in the RAND_poll(): over a minute. As a workaround
5484 include a time check in the inner Heap32Next loop too.
5487 *) The code that handled flushing of data in SSL/TLS originally used the
5488 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
5489 the problem outlined in PR#1949. The fix suggested there however can
5490 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
5491 of Apache). So instead simplify the code to flush unconditionally.
5492 This should be fine since flushing with no data to flush is a no op.
5495 *) Handle TLS versions 2.0 and later properly and correctly use the
5496 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
5497 off ancient servers have a habit of sticking around for a while...
5500 *) Modify compression code so it frees up structures without using the
5501 ex_data callbacks. This works around a problem where some applications
5502 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
5503 restarting) then use compression (e.g. SSL with compression) later.
5504 This results in significant per-connection memory leaks and
5505 has caused some security issues including CVE-2008-1678 and
5509 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
5510 change when encrypting or decrypting.
5513 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
5514 connect and renegotiate with servers which do not support RI.
5515 Until RI is more widely deployed this option is enabled by default.
5518 *) Add "missing" ssl ctrls to clear options and mode.
5521 *) If client attempts to renegotiate and doesn't support RI respond with
5522 a no_renegotiation alert as required by RFC5746. Some renegotiating
5523 TLS clients will continue a connection gracefully when they receive
5524 the alert. Unfortunately OpenSSL mishandled this alert and would hang
5525 waiting for a server hello which it will never receive. Now we treat a
5526 received no_renegotiation alert as a fatal error. This is because
5527 applications requesting a renegotiation might well expect it to succeed
5528 and would have no code in place to handle the server denying it so the
5529 only safe thing to do is to terminate the connection.
5532 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
5533 peer supports secure renegotiation and 0 otherwise. Print out peer
5534 renegotiation support in s_client/s_server.
5537 *) Replace the highly broken and deprecated SPKAC certification method with
5538 the updated NID creation version. This should correctly handle UTF8.
5541 *) Implement RFC5746. Re-enable renegotiation but require the extension
5542 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
5543 turns out to be a bad idea. It has been replaced by
5544 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5545 SSL_CTX_set_options(). This is really not recommended unless you
5546 know what you are doing.
5547 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5549 *) Fixes to stateless session resumption handling. Use initial_ctx when
5550 issuing and attempting to decrypt tickets in case it has changed during
5551 servername handling. Use a non-zero length session ID when attempting
5552 stateless session resumption: this makes it possible to determine if
5553 a resumption has occurred immediately after receiving server hello
5554 (several places in OpenSSL subtly assume this) instead of later in
5558 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5559 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5560 fixes for a few places where the return code is not checked
5562 [Julia Lawall <julia@diku.dk>]
5564 *) Add --strict-warnings option to Configure script to include devteam
5565 warnings in other configurations.
5568 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5569 makes it possible to install openssl libraries in locations which
5570 have names other than "lib", for example "/usr/lib64" which some
5572 [Steve Henson, based on patch from Jeremy Utley]
5574 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5575 X690 8.9.12 and can produce some misleading textual output of OIDs.
5576 [Steve Henson, reported by Dan Kaminsky]
5578 *) Delete MD2 from algorithm tables. This follows the recommendation in
5579 several standards that it is not used in new applications due to
5580 several cryptographic weaknesses. For binary compatibility reasons
5581 the MD2 API is still compiled in by default.
5584 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5588 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5589 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5591 [Guenter <lists@gknw.net>]
5593 *) Fix the server certificate chain building code to use X509_verify_cert(),
5594 it used to have an ad-hoc builder which was unable to cope with anything
5595 other than a simple chain.
5596 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5598 *) Don't check self signed certificate signatures in X509_verify_cert()
5599 by default (a flag can override this): it just wastes time without
5600 adding any security. As a useful side effect self signed root CAs
5601 with non-FIPS digests are now usable in FIPS mode.
5604 *) In dtls1_process_out_of_seq_message() the check if the current message
5605 is already buffered was missing. For every new message was memory
5606 allocated, allowing an attacker to perform an denial of service attack
5607 with sending out of seq handshake messages until there is no memory
5608 left. Additionally every future message was buffered, even if the
5609 sequence number made no sense and would be part of another handshake.
5610 So only messages with sequence numbers less than 10 in advance will be
5611 buffered. (CVE-2009-1378)
5612 [Robin Seggelmann, discovered by Daniel Mentz]
5614 *) Records are buffered if they arrive with a future epoch to be
5615 processed after finishing the corresponding handshake. There is
5616 currently no limitation to this buffer allowing an attacker to perform
5617 a DOS attack with sending records with future epochs until there is no
5618 memory left. This patch adds the pqueue_size() function to determine
5619 the size of a buffer and limits the record buffer to 100 entries.
5621 [Robin Seggelmann, discovered by Daniel Mentz]
5623 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5624 parent structure is freed. (CVE-2009-1379)
5627 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5628 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5631 [Ilya O. <vrghost@gmail.com>]
5633 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5635 *) Disable renegotiation completely - this fixes a severe security
5636 problem (CVE-2009-3555) at the cost of breaking all
5637 renegotiation. Renegotiation can be re-enabled by setting
5638 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5639 run-time. This is really not recommended unless you know what
5643 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5645 *) Don't set val to NULL when freeing up structures, it is freed up by
5646 underlying code. If sizeof(void *) > sizeof(long) this can result in
5647 zeroing past the valid field. (CVE-2009-0789)
5648 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5650 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5651 checked correctly. This would allow some invalid signed attributes to
5652 appear to verify correctly. (CVE-2009-0591)
5653 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5655 *) Reject UniversalString and BMPString types with invalid lengths. This
5656 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5657 a legal length. (CVE-2009-0590)
5660 *) Set S/MIME signing as the default purpose rather than setting it
5661 unconditionally. This allows applications to override it at the store
5665 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5666 to handle some structures.
5669 *) Improve efficiency of mem_gets: don't search whole buffer each time
5671 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5673 *) New -hex option for openssl rand.
5676 *) Print out UTF8String and NumericString when parsing ASN1.
5679 *) Support NumericString type for name components.
5682 *) Allow CC in the environment to override the automatically chosen
5683 compiler. Note that nothing is done to ensure flags work with the
5687 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5689 *) Properly check EVP_VerifyFinal() and similar return values
5691 [Ben Laurie, Bodo Moeller, Google Security Team]
5693 *) Enable TLS extensions by default.
5696 *) Allow the CHIL engine to be loaded, whether the application is
5697 multithreaded or not. (This does not release the developer from the
5698 obligation to set up the dynamic locking callbacks.)
5699 [Sander Temme <sander@temme.net>]
5701 *) Use correct exit code if there is an error in dgst command.
5702 [Steve Henson; problem pointed out by Roland Dirlewanger]
5704 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5705 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5708 *) Add experimental JPAKE support, including demo authentication in
5709 s_client and s_server.
5712 *) Set the comparison function in v3_addr_canonize().
5713 [Rob Austein <sra@hactrn.net>]
5715 *) Add support for XMPP STARTTLS in s_client.
5716 [Philip Paeps <philip@freebsd.org>]
5718 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5719 to ensure that even with this option, only ciphersuites in the
5720 server's preference list will be accepted. (Note that the option
5721 applies only when resuming a session, so the earlier behavior was
5722 just about the algorithm choice for symmetric cryptography.)
5725 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5727 *) Fix NULL pointer dereference if a DTLS server received
5728 ChangeCipherSpec as first record (CVE-2009-1386).
5731 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5732 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5735 *) The fix in 0.9.8c that supposedly got rid of unsafe
5736 double-checked locking was incomplete for RSA blinding,
5737 addressing just one layer of what turns out to have been
5738 doubly unsafe triple-checked locking.
5740 So now fix this for real by retiring the MONT_HELPER macro
5741 in crypto/rsa/rsa_eay.c.
5743 [Bodo Moeller; problem pointed out by Marius Schilder]
5745 *) Various precautionary measures:
5747 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5749 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5750 (NB: This would require knowledge of the secret session ticket key
5751 to exploit, in which case you'd be SOL either way.)
5753 - Change bn_nist.c so that it will properly handle input BIGNUMs
5754 outside the expected range.
5756 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5759 [Neel Mehta, Bodo Moeller]
5761 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5762 the load fails. Useful for distros.
5763 [Ben Laurie and the FreeBSD team]
5765 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5768 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5771 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5773 This work was sponsored by Logica.
5776 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5777 keystores. Support for SSL/TLS client authentication too.
5778 Not compiled unless enable-capieng specified to Configure.
5780 This work was sponsored by Logica.
5783 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5784 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5785 attribute creation routines such as certificate requests and PKCS#12
5789 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5791 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5792 handshake which could lead to a client crash as found using the
5793 Codenomicon TLS test suite (CVE-2008-1672)
5794 [Steve Henson, Mark Cox]
5796 *) Fix double free in TLS server name extensions which could lead to
5797 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5800 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5802 Clear the error queue to ensure that error entries left from
5803 older function calls do not interfere with the correct operation.
5804 [Lutz Jaenicke, Erik de Castro Lopo]
5806 *) Remove root CA certificates of commercial CAs:
5808 The OpenSSL project does not recommend any specific CA and does not
5809 have any policy with respect to including or excluding any CA.
5810 Therefore it does not make any sense to ship an arbitrary selection
5811 of root CA certificates with the OpenSSL software.
5814 *) RSA OAEP patches to fix two separate invalid memory reads.
5815 The first one involves inputs when 'lzero' is greater than
5816 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5817 before the beginning of from). The second one involves inputs where
5818 the 'db' section contains nothing but zeroes (there is a one-byte
5819 invalid read after the end of 'db').
5820 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5822 *) Partial backport from 0.9.9-dev:
5824 Introduce bn_mul_mont (dedicated Montgomery multiplication
5825 procedure) as a candidate for BIGNUM assembler implementation.
5826 While 0.9.9-dev uses assembler for various architectures, only
5827 x86_64 is available by default here in the 0.9.8 branch, and
5828 32-bit x86 is available through a compile-time setting.
5830 To try the 32-bit x86 assembler implementation, use Configure
5831 option "enable-montasm" (which exists only for this backport).
5833 As "enable-montasm" for 32-bit x86 disclaims code stability
5834 anyway, in this constellation we activate additional code
5835 backported from 0.9.9-dev for further performance improvements,
5836 namely BN_from_montgomery_word. (To enable this otherwise,
5837 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5839 [Andy Polyakov (backport partially by Bodo Moeller)]
5841 *) Add TLS session ticket callback. This allows an application to set
5842 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5843 values. This is useful for key rollover for example where several key
5844 sets may exist with different names.
5847 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5848 This was broken until now in 0.9.8 releases, such that the only way
5849 a registered ENGINE could be used (assuming it initialises
5850 successfully on the host) was to explicitly set it as the default
5851 for the relevant algorithms. This is in contradiction with 0.9.7
5852 behaviour and the documentation. With this fix, when an ENGINE is
5853 registered into a given algorithm's table of implementations, the
5854 'uptodate' flag is reset so that auto-discovery will be used next
5855 time a new context for that algorithm attempts to select an
5857 [Ian Lister (tweaked by Geoff Thorpe)]
5859 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5860 implementation in the following ways:
5862 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5865 Lack of BER streaming support means one pass streaming processing is
5866 only supported if data is detached: setting the streaming flag is
5867 ignored for embedded content.
5869 CMS support is disabled by default and must be explicitly enabled
5870 with the enable-cms configuration option.
5873 *) Update the GMP engine glue to do direct copies between BIGNUM and
5874 mpz_t when openssl and GMP use the same limb size. Otherwise the
5875 existing "conversion via a text string export" trick is still used.
5876 [Paul Sheer <paulsheer@gmail.com>]
5878 *) Zlib compression BIO. This is a filter BIO which compressed and
5879 uncompresses any data passed through it.
5882 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5883 RFC3394 compatible AES key wrapping.
5886 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5887 sets string data without copying. X509_ALGOR_set0() and
5888 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5889 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5890 from an X509_ATTRIBUTE structure optionally checking it occurs only
5891 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5895 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5896 to get the expected BN_FLG_CONSTTIME behavior.
5897 [Bodo Moeller (Google)]
5901 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5902 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5903 - added some more tests to do_tests.pl
5904 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5905 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5906 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5907 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5908 - various changes to netware.pl to enable gcc-cross builds on Win32
5910 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5911 - various changes to fix missing prototype warnings
5912 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5913 - added AES, WHIRLPOOL and CPUID assembler code to build files
5914 - added missing AES assembler make rules to mk1mf.pl
5915 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5916 [Guenter Knauf <eflash@gmx.net>]
5918 *) Implement certificate status request TLS extension defined in RFC3546.
5919 A client can set the appropriate parameters and receive the encoded
5920 OCSP response via a callback. A server can query the supplied parameters
5921 and set the encoded OCSP response in the callback. Add simplified examples
5922 to s_client and s_server.
5925 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5927 *) Fix various bugs:
5928 + Binary incompatibility of ssl_ctx_st structure
5929 + DTLS interoperation with non-compliant servers
5930 + Don't call get_session_cb() without proposed session
5931 + Fix ia64 assembler code
5932 [Andy Polyakov, Steve Henson]
5934 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5936 *) DTLS Handshake overhaul. There were longstanding issues with
5937 OpenSSL DTLS implementation, which were making it impossible for
5938 RFC 4347 compliant client to communicate with OpenSSL server.
5939 Unfortunately just fixing these incompatibilities would "cut off"
5940 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5941 server keeps tolerating non RFC compliant syntax. The opposite is
5942 not true, 0.9.8f client can not communicate with earlier server.
5943 This update even addresses CVE-2007-4995.
5946 *) Changes to avoid need for function casts in OpenSSL: some compilers
5947 (gcc 4.2 and later) reject their use.
5948 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5951 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5952 RFC4507bis. The encrypted ticket format is an encrypted encoded
5953 SSL_SESSION structure, that way new session features are automatically
5956 If a client application caches session in an SSL_SESSION structure
5957 support is transparent because tickets are now stored in the encoded
5960 The SSL_CTX structure automatically generates keys for ticket
5961 protection in servers so again support should be possible
5962 with no application modification.
5964 If a client or server wishes to disable RFC4507 support then the option
5965 SSL_OP_NO_TICKET can be set.
5967 Add a TLS extension debugging callback to allow the contents of any client
5968 or server extensions to be examined.
5970 This work was sponsored by Google.
5973 *) Add initial support for TLS extensions, specifically for the server_name
5974 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5975 have new members for a host name. The SSL data structure has an
5976 additional member SSL_CTX *initial_ctx so that new sessions can be
5977 stored in that context to allow for session resumption, even after the
5978 SSL has been switched to a new SSL_CTX in reaction to a client's
5979 server_name extension.
5981 New functions (subject to change):
5983 SSL_get_servername()
5984 SSL_get_servername_type()
5987 New CTRL codes and macros (subject to change):
5989 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5990 - SSL_CTX_set_tlsext_servername_callback()
5991 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5992 - SSL_CTX_set_tlsext_servername_arg()
5993 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5995 openssl s_client has a new '-servername ...' option.
5997 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5998 '-key2 ...', '-servername_fatal' (subject to change). This allows
5999 testing the HostName extension for a specific single host name ('-cert'
6000 and '-key' remain fallbacks for handshakes without HostName
6001 negotiation). If the unrecognized_name alert has to be sent, this by
6002 default is a warning; it becomes fatal with the '-servername_fatal'
6005 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
6007 *) Add AES and SSE2 assembly language support to VC++ build.
6010 *) Mitigate attack on final subtraction in Montgomery reduction.
6013 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
6014 (which previously caused an internal error).
6017 *) Squeeze another 10% out of IGE mode when in != out.
6020 *) AES IGE mode speedup.
6021 [Dean Gaudet (Google)]
6023 *) Add the Korean symmetric 128-bit cipher SEED (see
6024 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
6025 add SEED ciphersuites from RFC 4162:
6027 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
6028 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
6029 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
6030 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
6032 To minimize changes between patchlevels in the OpenSSL 0.9.8
6033 series, SEED remains excluded from compilation unless OpenSSL
6034 is configured with 'enable-seed'.
6035 [KISA, Bodo Moeller]
6037 *) Mitigate branch prediction attacks, which can be practical if a
6038 single processor is shared, allowing a spy process to extract
6039 information. For detailed background information, see
6040 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
6041 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
6042 and Necessary Software Countermeasures"). The core of the change
6043 are new versions BN_div_no_branch() and
6044 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
6045 respectively, which are slower, but avoid the security-relevant
6046 conditional branches. These are automatically called by BN_div()
6047 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
6048 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
6049 remove a conditional branch.
6051 BN_FLG_CONSTTIME is the new name for the previous
6052 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
6053 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
6054 in the exponent causes BN_mod_exp_mont() to use the alternative
6055 implementation in BN_mod_exp_mont_consttime().) The old name
6056 remains as a deprecated alias.
6058 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
6059 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
6060 constant-time implementations for more than just exponentiation.
6061 Here too the old name is kept as a deprecated alias.
6063 BN_BLINDING_new() will now use BN_dup() for the modulus so that
6064 the BN_BLINDING structure gets an independent copy of the
6065 modulus. This means that the previous "BIGNUM *m" argument to
6066 BN_BLINDING_new() and to BN_BLINDING_create_param() now
6067 essentially becomes "const BIGNUM *m", although we can't actually
6068 change this in the header file before 0.9.9. It allows
6069 RSA_setup_blinding() to use BN_with_flags() on the modulus to
6070 enable BN_FLG_CONSTTIME.
6072 [Matthew D Wood (Intel Corp)]
6074 *) In the SSL/TLS server implementation, be strict about session ID
6075 context matching (which matters if an application uses a single
6076 external cache for different purposes). Previously,
6077 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
6078 set. This did ensure strict client verification, but meant that,
6079 with applications using a single external cache for quite
6080 different requirements, clients could circumvent ciphersuite
6081 restrictions for a given session ID context by starting a session
6082 in a different context.
6085 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6086 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6087 authentication-only ciphersuites.
6090 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
6091 not complete and could lead to a possible single byte overflow
6092 (CVE-2007-5135) [Ben Laurie]
6094 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
6096 *) Since AES128 and AES256 (and similarly Camellia128 and
6097 Camellia256) share a single mask bit in the logic of
6098 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6099 kludge to work properly if AES128 is available and AES256 isn't
6100 (or if Camellia128 is available and Camellia256 isn't).
6103 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
6104 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
6105 When a point or a seed is encoded in a BIT STRING, we need to
6106 prevent the removal of trailing zero bits to get the proper DER
6107 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
6108 of a NamedBitList, for which trailing 0 bits need to be removed.)
6111 *) Have SSL/TLS server implementation tolerate "mismatched" record
6112 protocol version while receiving ClientHello even if the
6113 ClientHello is fragmented. (The server can't insist on the
6114 particular protocol version it has chosen before the ServerHello
6115 message has informed the client about his choice.)
6118 *) Add RFC 3779 support.
6119 [Rob Austein for ARIN, Ben Laurie]
6121 *) Load error codes if they are not already present instead of using a
6122 static variable. This allows them to be cleanly unloaded and reloaded.
6123 Improve header file function name parsing.
6126 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
6127 or CAPABILITY handshake as required by RFCs.
6130 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
6132 *) Introduce limits to prevent malicious keys being able to
6133 cause a denial of service. (CVE-2006-2940)
6134 [Steve Henson, Bodo Moeller]
6136 *) Fix ASN.1 parsing of certain invalid structures that can result
6137 in a denial of service. (CVE-2006-2937) [Steve Henson]
6139 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6140 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6142 *) Fix SSL client code which could crash if connecting to a
6143 malicious SSLv2 server. (CVE-2006-4343)
6144 [Tavis Ormandy and Will Drewry, Google Security Team]
6146 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
6147 match only those. Before that, "AES256-SHA" would be interpreted
6148 as a pattern and match "AES128-SHA" too (since AES128-SHA got
6149 the same strength classification in 0.9.7h) as we currently only
6150 have a single AES bit in the ciphersuite description bitmap.
6151 That change, however, also applied to ciphersuite strings such as
6152 "RC4-MD5" that intentionally matched multiple ciphersuites --
6153 namely, SSL 2.0 ciphersuites in addition to the more common ones
6154 from SSL 3.0/TLS 1.0.
6156 So we change the selection algorithm again: Naming an explicit
6157 ciphersuite selects this one ciphersuite, and any other similar
6158 ciphersuite (same bitmap) from *other* protocol versions.
6159 Thus, "RC4-MD5" again will properly select both the SSL 2.0
6160 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
6162 Since SSL 2.0 does not have any ciphersuites for which the
6163 128/256 bit distinction would be relevant, this works for now.
6164 The proper fix will be to use different bits for AES128 and
6165 AES256, which would have avoided the problems from the beginning;
6166 however, bits are scarce, so we can only do this in a new release
6167 (not just a patchlevel) when we can change the SSL_CIPHER
6168 definition to split the single 'unsigned long mask' bitmap into
6169 multiple values to extend the available space.
6173 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
6175 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6176 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6178 *) Add AES IGE and biIGE modes.
6181 *) Change the Unix randomness entropy gathering to use poll() when
6182 possible instead of select(), since the latter has some
6183 undesirable limitations.
6184 [Darryl Miles via Richard Levitte and Bodo Moeller]
6186 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
6187 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
6188 cannot be implicitly activated as part of, e.g., the "AES" alias.
6189 However, please upgrade to OpenSSL 0.9.9[-dev] for
6190 non-experimental use of the ECC ciphersuites to get TLS extension
6191 support, which is required for curve and point format negotiation
6192 to avoid potential handshake problems.
6195 *) Disable rogue ciphersuites:
6197 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6198 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6199 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6201 The latter two were purportedly from
6202 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6205 Also deactivate the remaining ciphersuites from
6206 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6207 unofficial, and the ID has long expired.
6210 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6211 dual-core machines) and other potential thread-safety issues.
6214 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
6215 versions), which is now available for royalty-free use
6216 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
6217 Also, add Camellia TLS ciphersuites from RFC 4132.
6219 To minimize changes between patchlevels in the OpenSSL 0.9.8
6220 series, Camellia remains excluded from compilation unless OpenSSL
6221 is configured with 'enable-camellia'.
6224 *) Disable the padding bug check when compression is in use. The padding
6225 bug check assumes the first packet is of even length, this is not
6226 necessarily true if compression is enabled and can result in false
6227 positives causing handshake failure. The actual bug test is ancient
6228 code so it is hoped that implementations will either have fixed it by
6229 now or any which still have the bug do not support compression.
6232 Changes between 0.9.8a and 0.9.8b [04 May 2006]
6234 *) When applying a cipher rule check to see if string match is an explicit
6235 cipher suite and only match that one cipher suite if it is.
6238 *) Link in manifests for VC++ if needed.
6239 [Austin Ziegler <halostatue@gmail.com>]
6241 *) Update support for ECC-based TLS ciphersuites according to
6242 draft-ietf-tls-ecc-12.txt with proposed changes (but without
6243 TLS extensions, which are supported starting with the 0.9.9
6244 branch, not in the OpenSSL 0.9.8 branch).
6247 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
6248 opaque EVP_CIPHER_CTX handling.
6251 *) Fixes and enhancements to zlib compression code. We now only use
6252 "zlib1.dll" and use the default __cdecl calling convention on Win32
6253 to conform with the standards mentioned here:
6254 http://www.zlib.net/DLL_FAQ.txt
6255 Static zlib linking now works on Windows and the new --with-zlib-include
6256 --with-zlib-lib options to Configure can be used to supply the location
6257 of the headers and library. Gracefully handle case where zlib library
6261 *) Several fixes and enhancements to the OID generation code. The old code
6262 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
6263 handle numbers larger than ULONG_MAX, truncated printing and had a
6264 non standard OBJ_obj2txt() behaviour.
6267 *) Add support for building of engines under engine/ as shared libraries
6268 under VC++ build system.
6271 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
6272 Hopefully, we will not see any false combination of paths any more.
6275 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
6277 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6278 (part of SSL_OP_ALL). This option used to disable the
6279 countermeasure against man-in-the-middle protocol-version
6280 rollback in the SSL 2.0 server implementation, which is a bad
6281 idea. (CVE-2005-2969)
6283 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6284 for Information Security, National Institute of Advanced Industrial
6285 Science and Technology [AIST], Japan)]
6287 *) Add two function to clear and return the verify parameter flags.
6290 *) Keep cipherlists sorted in the source instead of sorting them at
6291 runtime, thus removing the need for a lock.
6294 *) Avoid some small subgroup attacks in Diffie-Hellman.
6295 [Nick Mathewson and Ben Laurie]
6297 *) Add functions for well-known primes.
6300 *) Extended Windows CE support.
6301 [Satoshi Nakamura and Andy Polyakov]
6303 *) Initialize SSL_METHOD structures at compile time instead of during
6304 runtime, thus removing the need for a lock.
6307 *) Make PKCS7_decrypt() work even if no certificate is supplied by
6308 attempting to decrypt each encrypted key in turn. Add support to
6312 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
6314 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
6317 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
6320 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
6321 key into the same file any more.
6324 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
6327 *) Add -utf8 command line and config file option to 'ca'.
6328 [Stefan <stf@udoma.org]
6330 *) Removed the macro des_crypt(), as it seems to conflict with some
6331 libraries. Use DES_crypt().
6334 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
6335 involves renaming the source and generated shared-libs for
6336 both. The engines will accept the corrected or legacy ids
6337 ('ncipher' and '4758_cca' respectively) when binding. NB,
6338 this only applies when building 'shared'.
6339 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
6341 *) Add attribute functions to EVP_PKEY structure. Modify
6342 PKCS12_create() to recognize a CSP name attribute and
6343 use it. Make -CSP option work again in pkcs12 utility.
6346 *) Add new functionality to the bn blinding code:
6347 - automatic re-creation of the BN_BLINDING parameters after
6348 a fixed number of uses (currently 32)
6349 - add new function for parameter creation
6350 - introduce flags to control the update behaviour of the
6351 BN_BLINDING parameters
6352 - hide BN_BLINDING structure
6353 Add a second BN_BLINDING slot to the RSA structure to improve
6354 performance when a single RSA object is shared among several
6358 *) Add support for DTLS.
6359 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
6361 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
6362 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
6365 *) Remove buggy and incomplete DH cert support from
6366 ssl/ssl_rsa.c and ssl/s3_both.c
6369 *) Use SHA-1 instead of MD5 as the default digest algorithm for
6370 the apps/openssl applications.
6373 *) Compile clean with "-Wall -Wmissing-prototypes
6374 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
6375 DEBUG_SAFESTACK must also be set.
6378 *) Change ./Configure so that certain algorithms can be disabled by default.
6379 The new counterpiece to "no-xxx" is "enable-xxx".
6381 The patented RC5 and MDC2 algorithms will now be disabled unless
6382 "enable-rc5" and "enable-mdc2", respectively, are specified.
6384 (IDEA remains enabled despite being patented. This is because IDEA
6385 is frequently required for interoperability, and there is no license
6386 fee for non-commercial use. As before, "no-idea" can be used to
6387 avoid this algorithm.)
6391 *) Add processing of proxy certificates (see RFC 3820). This work was
6392 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
6393 EGEE (Enabling Grids for E-science in Europe).
6396 *) RC4 performance overhaul on modern architectures/implementations, such
6397 as Intel P4, IA-64 and AMD64.
6400 *) New utility extract-section.pl. This can be used specify an alternative
6401 section number in a pod file instead of having to treat each file as
6402 a separate case in Makefile. This can be done by adding two lines to the
6405 =for comment openssl_section:XXX
6407 The blank line is mandatory.
6411 *) New arguments -certform, -keyform and -pass for s_client and s_server
6412 to allow alternative format key and certificate files and passphrase
6416 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
6417 update associated structures and add various utility functions.
6419 Add new policy related verify parameters, include policy checking in
6420 standard verify code. Enhance 'smime' application with extra parameters
6421 to support policy checking and print out.
6424 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
6425 Nehemiah processors. These extensions support AES encryption in hardware
6426 as well as RNG (though RNG support is currently disabled).
6427 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
6429 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
6432 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
6433 [Andy Polyakov and a number of other people]
6435 *) Improved PowerPC platform support. Most notably BIGNUM assembler
6436 implementation contributed by IBM.
6437 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
6439 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
6440 exponent rather than 'unsigned long'. There is a corresponding change to
6441 the new 'rsa_keygen' element of the RSA_METHOD structure.
6442 [Jelte Jansen, Geoff Thorpe]
6444 *) Functionality for creating the initial serial number file is now
6445 moved from CA.pl to the 'ca' utility with a new option -create_serial.
6447 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
6448 number file to 1, which is bound to cause problems. To avoid
6449 the problems while respecting compatibility between different 0.9.7
6450 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
6451 CA.pl for serial number initialization. With the new release 0.9.8,
6452 we can fix the problem directly in the 'ca' utility.)
6455 *) Reduced header interdependencies by declaring more opaque objects in
6456 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
6457 give fewer recursive includes, which could break lazy source code - so
6458 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
6459 developers should define this symbol when building and using openssl to
6460 ensure they track the recommended behaviour, interfaces, [etc], but
6461 backwards-compatible behaviour prevails when this isn't defined.
6464 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
6467 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
6468 This will generate a random key of the appropriate length based on the
6469 cipher context. The EVP_CIPHER can provide its own random key generation
6470 routine to support keys of a specific form. This is used in the des and
6471 3des routines to generate a key of the correct parity. Update S/MIME
6472 code to use new functions and hence generate correct parity DES keys.
6473 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
6474 valid (weak or incorrect parity).
6477 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
6478 as looking them up. This is useful when the verified structure may contain
6479 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
6480 present unless the new PKCS7_NO_CRL flag is asserted.
6483 *) Extend ASN1 oid configuration module. It now additionally accepts the
6486 shortName = some long name, 1.2.3.4
6489 *) Reimplemented the BN_CTX implementation. There is now no more static
6490 limitation on the number of variables it can handle nor the depth of the
6491 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
6492 information can now expand as required, and rather than having a single
6493 static array of bignums, BN_CTX now uses a linked-list of such arrays
6494 allowing it to expand on demand whilst maintaining the usefulness of
6495 BN_CTX's "bundling".
6498 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
6499 to allow all RSA operations to function using a single BN_CTX.
6502 *) Preliminary support for certificate policy evaluation and checking. This
6503 is initially intended to pass the tests outlined in "Conformance Testing
6504 of Relying Party Client Certificate Path Processing Logic" v1.07.
6507 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
6508 remained unused and not that useful. A variety of other little bignum
6509 tweaks and fixes have also been made continuing on from the audit (see
6513 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
6514 associated ASN1, EVP and SSL functions and old ASN1 macros.
6517 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
6518 and this should never fail. So the return value from the use of
6519 BN_set_word() (which can fail due to needless expansion) is now deprecated;
6520 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
6523 *) BN_CTX_get() should return zero-valued bignums, providing the same
6524 initialised value as BN_new().
6525 [Geoff Thorpe, suggested by Ulf Möller]
6527 *) Support for inhibitAnyPolicy certificate extension.
6530 *) An audit of the BIGNUM code is underway, for which debugging code is
6531 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
6532 is considered valid when processing BIGNUMs, and causes execution to
6533 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
6534 further steps are taken to deliberately pollute unused data in BIGNUM
6535 structures to try and expose faulty code further on. For now, openssl will
6536 (in its default mode of operation) continue to tolerate the inconsistent
6537 forms that it has tolerated in the past, but authors and packagers should
6538 consider trying openssl and their own applications when compiled with
6539 these debugging symbols defined. It will help highlight potential bugs in
6540 their own code, and will improve the test coverage for OpenSSL itself. At
6541 some point, these tighter rules will become openssl's default to improve
6542 maintainability, though the assert()s and other overheads will remain only
6543 in debugging configurations. See bn.h for more details.
6544 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6546 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6547 that can only be obtained through BN_CTX_new() (which implicitly
6548 initialises it). The presence of this function only made it possible
6549 to overwrite an existing structure (and cause memory leaks).
6552 *) Because of the callback-based approach for implementing LHASH as a
6553 template type, lh_insert() adds opaque objects to hash-tables and
6554 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6555 to clean up those corresponding objects before destroying the hash table
6556 (and losing the object pointers). So some over-zealous constifications in
6557 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6558 objects as "const" and the lh_doall[_arg] callback wrappers are not
6559 prototyped to have "const" restrictions on the object pointers they are
6560 given (and so aren't required to cast them away any more).
6563 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6564 (speed) prefers to use its own implementation. The two implementations
6565 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6566 its object type properly exposed (MS_TM) instead of casting to/from "char
6567 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6568 aren't necessarily the greatest nomenclatures - but this is what was used
6569 internally to the implementation so I've used that for now.
6572 *) Ensure that deprecated functions do not get compiled when
6573 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6574 the self-tests were still using deprecated key-generation functions so
6575 these have been updated also.
6578 *) Reorganise PKCS#7 code to separate the digest location functionality
6579 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6580 New function PKCS7_set_digest() to set the digest type for PKCS#7
6581 digestedData type. Add additional code to correctly generate the
6582 digestedData type and add support for this type in PKCS7 initialization
6586 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6587 structure of type "other".
6590 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6591 sure the loop does correctly stop and breaking ("division by zero")
6592 modulus operations are not performed. The (pre-generated) prime
6593 table crypto/bn/bn_prime.h was already correct, but it could not be
6594 re-generated on some platforms because of the "division by zero"
6595 situation in the script.
6596 [Ralf S. Engelschall]
6598 *) Update support for ECC-based TLS ciphersuites according to
6599 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6600 SHA-1 now is only used for "small" curves (where the
6601 representation of a field element takes up to 24 bytes); for
6602 larger curves, the field element resulting from ECDH is directly
6603 used as premaster secret.
6604 [Douglas Stebila (Sun Microsystems Laboratories)]
6606 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6607 curve secp160r1 to the tests.
6608 [Douglas Stebila (Sun Microsystems Laboratories)]
6610 *) Add the possibility to load symbols globally with DSO.
6611 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6613 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6614 control of the error stack.
6617 *) Add support for STORE in ENGINE.
6620 *) Add the STORE type. The intention is to provide a common interface
6621 to certificate and key stores, be they simple file-based stores, or
6622 HSM-type store, or LDAP stores, or...
6623 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6626 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6627 pass a list of arguments to any function as well as provide a way
6628 for a function to pass data back to the caller.
6631 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6632 works like BUF_strdup() but can be used to duplicate a portion of
6633 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6637 *) Add the function sk_find_ex() which works like sk_find(), but will
6638 return an index to an element even if an exact match couldn't be
6639 found. The index is guaranteed to point at the element where the
6640 searched-for key would be inserted to preserve sorting order.
6643 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6644 takes an extra flags argument for optional functionality. Currently,
6645 the following flags are defined:
6647 OBJ_BSEARCH_VALUE_ON_NOMATCH
6648 This one gets OBJ_bsearch_ex() to return a pointer to the first
6649 element where the comparing function returns a negative or zero
6652 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6653 This one gets OBJ_bsearch_ex() to return a pointer to the first
6654 element where the comparing function returns zero. This is useful
6655 if there are more than one element where the comparing function
6659 *) Make it possible to create self-signed certificates with 'openssl ca'
6660 in such a way that the self-signed certificate becomes part of the
6661 CA database and uses the same mechanisms for serial number generation
6662 as all other certificate signing. The new flag '-selfsign' enables
6663 this functionality. Adapt CA.sh and CA.pl.in.
6666 *) Add functionality to check the public key of a certificate request
6667 against a given private. This is useful to check that a certificate
6668 request can be signed by that key (self-signing).
6671 *) Make it possible to have multiple active certificates with the same
6672 subject in the CA index file. This is done only if the keyword
6673 'unique_subject' is set to 'no' in the main CA section (default
6674 if 'CA_default') of the configuration file. The value is saved
6675 with the database itself in a separate index attribute file,
6676 named like the index file with '.attr' appended to the name.
6679 *) Generate multi-valued AVAs using '+' notation in config files for
6683 *) Support for nameConstraints certificate extension.
6686 *) Support for policyConstraints certificate extension.
6689 *) Support for policyMappings certificate extension.
6692 *) Make sure the default DSA_METHOD implementation only uses its
6693 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6694 and change its own handlers to be NULL so as to remove unnecessary
6695 indirection. This lets alternative implementations fallback to the
6696 default implementation more easily.
6699 *) Support for directoryName in GeneralName related extensions
6703 *) Make it possible to link applications using Makefile.shared.
6704 Make that possible even when linking against static libraries!
6707 *) Support for single pass processing for S/MIME signing. This now
6708 means that S/MIME signing can be done from a pipe, in addition
6709 cleartext signing (multipart/signed type) is effectively streaming
6710 and the signed data does not need to be all held in memory.
6712 This is done with a new flag PKCS7_STREAM. When this flag is set
6713 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6714 is done after the data is output (and digests calculated) in
6715 SMIME_write_PKCS7().
6718 *) Add full support for -rpath/-R, both in shared libraries and
6719 applications, at least on the platforms where it's known how
6723 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6724 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6725 will now compute a table of multiples of the generator that
6726 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6727 faster (notably in the case of a single point multiplication,
6728 scalar * generator).
6729 [Nils Larsch, Bodo Moeller]
6731 *) IPv6 support for certificate extensions. The various extensions
6732 which use the IP:a.b.c.d can now take IPv6 addresses using the
6733 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6737 *) Added an ENGINE that implements RSA by performing private key
6738 exponentiations with the GMP library. The conversions to and from
6739 GMP's mpz_t format aren't optimised nor are any montgomery forms
6740 cached, and on x86 it appears OpenSSL's own performance has caught up.
6741 However there are likely to be other architectures where GMP could
6742 provide a boost. This ENGINE is not built in by default, but it can be
6743 specified at Configure time and should be accompanied by the necessary
6744 linker additions, eg;
6745 ./config -DOPENSSL_USE_GMP -lgmp
6748 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6749 testing availability of engines with "-t" - the old behaviour is
6750 produced by increasing the feature's verbosity with "-tt".
6753 *) ECDSA routines: under certain error conditions uninitialized BN objects
6754 could be freed. Solution: make sure initialization is performed early
6755 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6759 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6760 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6761 software implementations. For DSA and DH, parameter generation can
6762 also be overridden by providing the appropriate method callbacks.
6765 *) Change the "progress" mechanism used in key-generation and
6766 primality testing to functions that take a new BN_GENCB pointer in
6767 place of callback/argument pairs. The new API functions have "_ex"
6768 postfixes and the older functions are reimplemented as wrappers for
6769 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6770 declarations of the old functions to help (graceful) attempts to
6771 migrate to the new functions. Also, the new key-generation API
6772 functions operate on a caller-supplied key-structure and return
6773 success/failure rather than returning a key or NULL - this is to
6774 help make "keygen" another member function of RSA_METHOD etc.
6776 Example for using the new callback interface:
6778 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6782 BN_GENCB_set(&my_cb, my_callback, my_arg);
6784 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6785 /* For the meaning of a, b in calls to my_callback(), see the
6786 * documentation of the function that calls the callback.
6787 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6788 * my_callback should return 1 if it wants BN_is_prime_ex()
6789 * to continue, or 0 to stop.
6794 *) Change the ZLIB compression method to be stateful, and make it
6795 available to TLS with the number defined in
6796 draft-ietf-tls-compression-04.txt.
6799 *) Add the ASN.1 structures and functions for CertificatePair, which
6800 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6802 CertificatePair ::= SEQUENCE {
6803 forward [0] Certificate OPTIONAL,
6804 reverse [1] Certificate OPTIONAL,
6805 -- at least one of the pair shall be present -- }
6807 Also implement the PEM functions to read and write certificate
6808 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6810 This needed to be defined, mostly for the sake of the LDAP
6811 attribute crossCertificatePair, but may prove useful elsewhere as
6815 *) Make it possible to inhibit symlinking of shared libraries in
6816 Makefile.shared, for Cygwin's sake.
6819 *) Extend the BIGNUM API by creating a function
6820 void BN_set_negative(BIGNUM *a, int neg);
6821 and a macro that behave like
6822 int BN_is_negative(const BIGNUM *a);
6824 to avoid the need to access 'a->neg' directly in applications.
6827 *) Implement fast modular reduction for pseudo-Mersenne primes
6828 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6829 EC_GROUP_new_curve_GFp() will now automatically use this
6831 [Nils Larsch <nla@trustcenter.de>]
6833 *) Add new lock type (CRYPTO_LOCK_BN).
6836 *) Change the ENGINE framework to automatically load engines
6837 dynamically from specific directories unless they could be
6838 found to already be built in or loaded. Move all the
6839 current engines except for the cryptodev one to a new
6841 The engines in engines/ are built as shared libraries if
6842 the "shared" options was given to ./Configure or ./config.
6843 Otherwise, they are inserted in libcrypto.a.
6844 /usr/local/ssl/engines is the default directory for dynamic
6845 engines, but that can be overridden at configure time through
6846 the usual use of --prefix and/or --openssldir, and at run
6847 time with the environment variable OPENSSL_ENGINES.
6848 [Geoff Thorpe and Richard Levitte]
6850 *) Add Makefile.shared, a helper makefile to build shared
6851 libraries. Adapt Makefile.org.
6854 *) Add version info to Win32 DLLs.
6855 [Peter 'Luna' Runestig" <peter@runestig.com>]
6857 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6858 can be added using this API to created arbitrary PKCS#12
6859 files while avoiding the low level API.
6861 New options to PKCS12_create(), key or cert can be NULL and
6862 will then be omitted from the output file. The encryption
6863 algorithm NIDs can be set to -1 for no encryption, the mac
6864 iteration count can be set to 0 to omit the mac.
6866 Enhance pkcs12 utility by making the -nokeys and -nocerts
6867 options work when creating a PKCS#12 file. New option -nomac
6868 to omit the mac, NONE can be set for an encryption algorithm.
6869 New code is modified to use the enhanced PKCS12_create()
6870 instead of the low level API.
6873 *) Extend ASN1 encoder to support indefinite length constructed
6874 encoding. This can output sequences tags and octet strings in
6875 this form. Modify pk7_asn1.c to support indefinite length
6876 encoding. This is experimental and needs additional code to
6877 be useful, such as an ASN1 bio and some enhanced streaming
6880 Extend template encode functionality so that tagging is passed
6881 down to the template encoder.
6884 *) Let 'openssl req' fail if an argument to '-newkey' is not
6885 recognized instead of using RSA as a default.
6888 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6889 As these are not official, they are not included in "ALL";
6890 the "ECCdraft" ciphersuite group alias can be used to select them.
6891 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6893 *) Add ECDH engine support.
6894 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6896 *) Add ECDH in new directory crypto/ecdh/.
6897 [Douglas Stebila (Sun Microsystems Laboratories)]
6899 *) Let BN_rand_range() abort with an error after 100 iterations
6900 without success (which indicates a broken PRNG).
6903 *) Change BN_mod_sqrt() so that it verifies that the input value
6904 is really the square of the return value. (Previously,
6905 BN_mod_sqrt would show GIGO behaviour.)
6908 *) Add named elliptic curves over binary fields from X9.62, SECG,
6909 and WAP/WTLS; add OIDs that were still missing.
6911 [Sheueling Chang Shantz and Douglas Stebila
6912 (Sun Microsystems Laboratories)]
6914 *) Extend the EC library for elliptic curves over binary fields
6915 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6918 EC_GF2m_simple_method
6922 EC_GROUP_new_curve_GF2m
6923 EC_GROUP_set_curve_GF2m
6924 EC_GROUP_get_curve_GF2m
6925 EC_POINT_set_affine_coordinates_GF2m
6926 EC_POINT_get_affine_coordinates_GF2m
6927 EC_POINT_set_compressed_coordinates_GF2m
6929 Point compression for binary fields is disabled by default for
6930 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6933 As binary polynomials are represented as BIGNUMs, various members
6934 of the EC_GROUP and EC_POINT data structures can be shared
6935 between the implementations for prime fields and binary fields;
6936 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6937 are essentially identical to their ..._GFp counterparts.
6938 (For simplicity, the '..._GFp' prefix has been dropped from
6939 various internal method names.)
6941 An internal 'field_div' method (similar to 'field_mul' and
6942 'field_sqr') has been added; this is used only for binary fields.
6944 [Sheueling Chang Shantz and Douglas Stebila
6945 (Sun Microsystems Laboratories)]
6947 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6948 through methods ('mul', 'precompute_mult').
6950 The generic implementations (now internally called 'ec_wNAF_mul'
6951 and 'ec_wNAF_precomputed_mult') remain the default if these
6952 methods are undefined.
6954 [Sheueling Chang Shantz and Douglas Stebila
6955 (Sun Microsystems Laboratories)]
6957 *) New function EC_GROUP_get_degree, which is defined through
6958 EC_METHOD. For curves over prime fields, this returns the bit
6959 length of the modulus.
6961 [Sheueling Chang Shantz and Douglas Stebila
6962 (Sun Microsystems Laboratories)]
6964 *) New functions EC_GROUP_dup, EC_POINT_dup.
6965 (These simply call ..._new and ..._copy).
6967 [Sheueling Chang Shantz and Douglas Stebila
6968 (Sun Microsystems Laboratories)]
6970 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6971 Polynomials are represented as BIGNUMs (where the sign bit is not
6972 used) in the following functions [macros]:
6975 BN_GF2m_sub [= BN_GF2m_add]
6976 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6977 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6978 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6980 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6981 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6982 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6983 BN_GF2m_cmp [= BN_ucmp]
6985 (Note that only the 'mod' functions are actually for fields GF(2^m).
6986 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6988 For some functions, an the irreducible polynomial defining a
6989 field can be given as an 'unsigned int[]' with strictly
6990 decreasing elements giving the indices of those bits that are set;
6991 i.e., p[] represents the polynomial
6992 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6994 p[0] > p[1] > ... > p[k] = 0.
6995 This applies to the following functions:
7000 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
7001 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
7003 BN_GF2m_mod_sqrt_arr
7004 BN_GF2m_mod_solve_quad_arr
7008 Conversion can be performed by the following functions:
7013 bntest.c has additional tests for binary polynomial arithmetic.
7015 Two implementations for BN_GF2m_mod_div() are available.
7016 The default algorithm simply uses BN_GF2m_mod_inv() and
7017 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
7018 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
7019 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
7021 [Sheueling Chang Shantz and Douglas Stebila
7022 (Sun Microsystems Laboratories)]
7024 *) Add new error code 'ERR_R_DISABLED' that can be used when some
7025 functionality is disabled at compile-time.
7026 [Douglas Stebila <douglas.stebila@sun.com>]
7028 *) Change default behaviour of 'openssl asn1parse' so that more
7029 information is visible when viewing, e.g., a certificate:
7031 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
7032 mode the content of non-printable OCTET STRINGs is output in a
7033 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
7034 avoid the appearance of a printable string.
7035 [Nils Larsch <nla@trustcenter.de>]
7037 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
7039 EC_GROUP_set_asn1_flag()
7040 EC_GROUP_get_asn1_flag()
7041 EC_GROUP_set_point_conversion_form()
7042 EC_GROUP_get_point_conversion_form()
7043 These control ASN1 encoding details:
7044 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
7045 has been set to OPENSSL_EC_NAMED_CURVE.
7046 - Points are encoded in uncompressed form by default; options for
7047 asn1_for are as for point2oct, namely
7048 POINT_CONVERSION_COMPRESSED
7049 POINT_CONVERSION_UNCOMPRESSED
7050 POINT_CONVERSION_HYBRID
7052 Also add 'seed' and 'seed_len' members to EC_GROUP with access
7055 EC_GROUP_get0_seed()
7056 EC_GROUP_get_seed_len()
7057 This is used only for ASN1 purposes (so far).
7058 [Nils Larsch <nla@trustcenter.de>]
7060 *) Add 'field_type' member to EC_METHOD, which holds the NID
7061 of the appropriate field type OID. The new function
7062 EC_METHOD_get_field_type() returns this value.
7063 [Nils Larsch <nla@trustcenter.de>]
7068 EC_POINT_point2hex()
7069 EC_POINT_hex2point()
7070 providing useful interfaces to EC_POINT_point2oct() and
7071 EC_POINT_oct2point().
7072 [Nils Larsch <nla@trustcenter.de>]
7074 *) Change internals of the EC library so that the functions
7075 EC_GROUP_set_generator()
7076 EC_GROUP_get_generator()
7077 EC_GROUP_get_order()
7078 EC_GROUP_get_cofactor()
7079 are implemented directly in crypto/ec/ec_lib.c and not dispatched
7080 to methods, which would lead to unnecessary code duplication when
7081 adding different types of curves.
7082 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
7084 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
7085 arithmetic, and such that modified wNAFs are generated
7086 (which avoid length expansion in many cases).
7089 *) Add a function EC_GROUP_check_discriminant() (defined via
7090 EC_METHOD) that verifies that the curve discriminant is non-zero.
7092 Add a function EC_GROUP_check() that makes some sanity tests
7093 on a EC_GROUP, its generator and order. This includes
7094 EC_GROUP_check_discriminant().
7095 [Nils Larsch <nla@trustcenter.de>]
7097 *) Add ECDSA in new directory crypto/ecdsa/.
7099 Add applications 'openssl ecparam' and 'openssl ecdsa'
7100 (these are based on 'openssl dsaparam' and 'openssl dsa').
7102 ECDSA support is also included in various other files across the
7103 library. Most notably,
7104 - 'openssl req' now has a '-newkey ecdsa:file' option;
7105 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
7106 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
7107 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
7108 them suitable for ECDSA where domain parameters must be
7109 extracted before the specific public key;
7110 - ECDSA engine support has been added.
7111 [Nils Larsch <nla@trustcenter.de>]
7113 *) Include some named elliptic curves, and add OIDs from X9.62,
7114 SECG, and WAP/WTLS. Each curve can be obtained from the new
7116 EC_GROUP_new_by_curve_name(),
7117 and the list of available named curves can be obtained with
7118 EC_get_builtin_curves().
7119 Also add a 'curve_name' member to EC_GROUP objects, which can be
7121 EC_GROUP_set_curve_name()
7122 EC_GROUP_get_curve_name()
7123 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
7125 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7126 was actually never needed) and in BN_mul(). The removal in BN_mul()
7127 required a small change in bn_mul_part_recursive() and the addition
7128 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7129 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7130 bn_sub_words() and bn_add_words() except they take arrays with
7134 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
7136 *) Cleanse PEM buffers before freeing them since they may contain
7138 [Benjamin Bennett <ben@psc.edu>]
7140 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
7141 a ciphersuite string such as "DEFAULT:RSA" cannot enable
7142 authentication-only ciphersuites.
7145 *) Since AES128 and AES256 share a single mask bit in the logic of
7146 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
7147 kludge to work properly if AES128 is available and AES256 isn't.
7150 *) Expand security boundary to match 1.1.1 module.
7153 *) Remove redundant features: hash file source, editing of test vectors
7154 modify fipsld to use external fips_premain.c signature.
7157 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
7158 run algorithm test programs.
7161 *) Make algorithm test programs more tolerant of whitespace.
7164 *) Have SSL/TLS server implementation tolerate "mismatched" record
7165 protocol version while receiving ClientHello even if the
7166 ClientHello is fragmented. (The server can't insist on the
7167 particular protocol version it has chosen before the ServerHello
7168 message has informed the client about his choice.)
7171 *) Load error codes if they are not already present instead of using a
7172 static variable. This allows them to be cleanly unloaded and reloaded.
7175 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
7177 *) Introduce limits to prevent malicious keys being able to
7178 cause a denial of service. (CVE-2006-2940)
7179 [Steve Henson, Bodo Moeller]
7181 *) Fix ASN.1 parsing of certain invalid structures that can result
7182 in a denial of service. (CVE-2006-2937) [Steve Henson]
7184 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
7185 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
7187 *) Fix SSL client code which could crash if connecting to a
7188 malicious SSLv2 server. (CVE-2006-4343)
7189 [Tavis Ormandy and Will Drewry, Google Security Team]
7191 *) Change ciphersuite string processing so that an explicit
7192 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
7193 will no longer include "AES128-SHA"), and any other similar
7194 ciphersuite (same bitmap) from *other* protocol versions (so that
7195 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
7196 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
7197 changes from 0.9.8b and 0.9.8d.
7200 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
7202 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
7203 (CVE-2006-4339) [Ben Laurie and Google Security Team]
7205 *) Change the Unix randomness entropy gathering to use poll() when
7206 possible instead of select(), since the latter has some
7207 undesirable limitations.
7208 [Darryl Miles via Richard Levitte and Bodo Moeller]
7210 *) Disable rogue ciphersuites:
7212 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
7213 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
7214 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
7216 The latter two were purportedly from
7217 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
7220 Also deactivate the remaining ciphersuites from
7221 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
7222 unofficial, and the ID has long expired.
7225 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
7226 dual-core machines) and other potential thread-safety issues.
7229 Changes between 0.9.7i and 0.9.7j [04 May 2006]
7231 *) Adapt fipsld and the build system to link against the validated FIPS
7232 module in FIPS mode.
7235 *) Fixes for VC++ 2005 build under Windows.
7238 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
7239 from a Windows bash shell such as MSYS. It is autodetected from the
7240 "config" script when run from a VC++ environment. Modify standard VC++
7241 build to use fipscanister.o from the GNU make build.
7244 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
7246 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
7247 The value now differs depending on if you build for FIPS or not.
7248 BEWARE! A program linked with a shared FIPSed libcrypto can't be
7249 safely run with a non-FIPSed libcrypto, as it may crash because of
7250 the difference induced by this change.
7253 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
7255 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
7256 (part of SSL_OP_ALL). This option used to disable the
7257 countermeasure against man-in-the-middle protocol-version
7258 rollback in the SSL 2.0 server implementation, which is a bad
7259 idea. (CVE-2005-2969)
7261 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
7262 for Information Security, National Institute of Advanced Industrial
7263 Science and Technology [AIST], Japan)]
7265 *) Minimal support for X9.31 signatures and PSS padding modes. This is
7266 mainly for FIPS compliance and not fully integrated at this stage.
7269 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
7270 the exponentiation using a fixed-length exponent. (Otherwise,
7271 the information leaked through timing could expose the secret key
7272 after many signatures; cf. Bleichenbacher's attack on DSA with
7276 *) Make a new fixed-window mod_exp implementation the default for
7277 RSA, DSA, and DH private-key operations so that the sequence of
7278 squares and multiplies and the memory access pattern are
7279 independent of the particular secret key. This will mitigate
7280 cache-timing and potential related attacks.
7282 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
7283 and this is automatically used by BN_mod_exp_mont() if the new flag
7284 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
7285 will use this BN flag for private exponents unless the flag
7286 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
7287 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
7289 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
7291 *) Change the client implementation for SSLv23_method() and
7292 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
7293 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
7294 (Previously, the SSL 2.0 backwards compatible Client Hello
7295 message format would be used even with SSL_OP_NO_SSLv2.)
7298 *) Add support for smime-type MIME parameter in S/MIME messages which some
7302 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
7303 a threadsafe manner. Modify rsa code to use new function and add calls
7304 to dsa and dh code (which had race conditions before).
7307 *) Include the fixed error library code in the C error file definitions
7308 instead of fixing them up at runtime. This keeps the error code
7309 structures constant.
7312 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
7314 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
7317 *) Fixes for newer kerberos headers. NB: the casts are needed because
7318 the 'length' field is signed on one version and unsigned on another
7319 with no (?) obvious way to tell the difference, without these VC++
7320 complains. Also the "definition" of FAR (blank) is no longer included
7321 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
7322 some needed definitions.
7325 *) Undo Cygwin change.
7328 *) Added support for proxy certificates according to RFC 3820.
7329 Because they may be a security thread to unaware applications,
7330 they must be explicitly allowed in run-time. See
7331 docs/HOWTO/proxy_certificates.txt for further information.
7334 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
7336 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
7337 server and client random values. Previously
7338 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
7339 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
7341 This change has negligible security impact because:
7343 1. Server and client random values still have 24 bytes of pseudo random
7346 2. Server and client random values are sent in the clear in the initial
7349 3. The master secret is derived using the premaster secret (48 bytes in
7350 size for static RSA ciphersuites) as well as client server and random
7353 The OpenSSL team would like to thank the UK NISCC for bringing this issue
7356 [Stephen Henson, reported by UK NISCC]
7358 *) Use Windows randomness collection on Cygwin.
7361 *) Fix hang in EGD/PRNGD query when communication socket is closed
7362 prematurely by EGD/PRNGD.
7363 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
7365 *) Prompt for pass phrases when appropriate for PKCS12 input format.
7368 *) Back-port of selected performance improvements from development
7369 branch, as well as improved support for PowerPC platforms.
7372 *) Add lots of checks for memory allocation failure, error codes to indicate
7373 failure and freeing up memory if a failure occurs.
7374 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
7376 *) Add new -passin argument to dgst.
7379 *) Perform some character comparisons of different types in X509_NAME_cmp:
7380 this is needed for some certificates that re-encode DNs into UTF8Strings
7381 (in violation of RFC3280) and can't or won't issue name rollover
7385 *) Make an explicit check during certificate validation to see that
7386 the CA setting in each certificate on the chain is correct. As a
7387 side effect always do the following basic checks on extensions,
7388 not just when there's an associated purpose to the check:
7390 - if there is an unhandled critical extension (unless the user
7391 has chosen to ignore this fault)
7392 - if the path length has been exceeded (if one is set at all)
7393 - that certain extensions fit the associated purpose (if one has
7397 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
7399 *) Avoid a race condition when CRLs are checked in a multi threaded
7400 environment. This would happen due to the reordering of the revoked
7401 entries during signature checking and serial number lookup. Now the
7402 encoding is cached and the serial number sort performed under a lock.
7403 Add new STACK function sk_is_sorted().
7406 *) Add Delta CRL to the extension code.
7409 *) Various fixes to s3_pkt.c so alerts are sent properly.
7410 [David Holmes <d.holmes@f5.com>]
7412 *) Reduce the chances of duplicate issuer name and serial numbers (in
7413 violation of RFC3280) using the OpenSSL certificate creation utilities.
7414 This is done by creating a random 64 bit value for the initial serial
7415 number when a serial number file is created or when a self signed
7416 certificate is created using 'openssl req -x509'. The initial serial
7417 number file is created using 'openssl x509 -next_serial' in CA.pl
7418 rather than being initialized to 1.
7421 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
7423 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7424 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7425 [Joe Orton, Steve Henson]
7427 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
7429 [Joe Orton, Steve Henson]
7431 *) Make it possible to have multiple active certificates with the same
7432 subject in the CA index file. This is done only if the keyword
7433 'unique_subject' is set to 'no' in the main CA section (default
7434 if 'CA_default') of the configuration file. The value is saved
7435 with the database itself in a separate index attribute file,
7436 named like the index file with '.attr' appended to the name.
7439 *) X509 verify fixes. Disable broken certificate workarounds when
7440 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
7441 keyUsage extension present. Don't accept CRLs with unhandled critical
7442 extensions: since verify currently doesn't process CRL extensions this
7443 rejects a CRL with *any* critical extensions. Add new verify error codes
7447 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
7448 A clarification of RFC2560 will require the use of OCTET STRINGs and
7449 some implementations cannot handle the current raw format. Since OpenSSL
7450 copies and compares OCSP nonces as opaque blobs without any attempt at
7451 parsing them this should not create any compatibility issues.
7454 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
7455 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
7456 this HMAC (and other) operations are several times slower than OpenSSL
7460 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
7461 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7463 *) Use the correct content when signing type "other".
7466 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
7468 *) Fix various bugs revealed by running the NISCC test suite:
7470 Stop out of bounds reads in the ASN1 code when presented with
7471 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7473 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
7475 If verify callback ignores invalid public key errors don't try to check
7476 certificate signature with the NULL public key.
7480 *) New -ignore_err option in ocsp application to stop the server
7481 exiting on the first error in a request.
7484 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7485 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7489 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7490 extra data after the compression methods not only for TLS 1.0
7491 but also for SSL 3.0 (as required by the specification).
7492 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7494 *) Change X509_certificate_type() to mark the key as exported/exportable
7495 when it's 512 *bits* long, not 512 bytes.
7498 *) Change AES_cbc_encrypt() so it outputs exact multiple of
7499 blocks during encryption.
7502 *) Various fixes to base64 BIO and non blocking I/O. On write
7503 flushes were not handled properly if the BIO retried. On read
7504 data was not being buffered properly and had various logic bugs.
7505 This also affects blocking I/O when the data being decoded is a
7509 *) Various S/MIME bugfixes and compatibility changes:
7510 output correct application/pkcs7 MIME type if
7511 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
7512 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
7513 of files as .eml work). Correctly handle very long lines in MIME
7517 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
7519 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7520 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7521 a protocol version number mismatch like a decryption error
7522 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7525 *) Turn on RSA blinding by default in the default implementation
7526 to avoid a timing attack. Applications that don't want it can call
7527 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7528 They would be ill-advised to do so in most cases.
7529 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7531 *) Change RSA blinding code so that it works when the PRNG is not
7532 seeded (in this case, the secret RSA exponent is abused as
7533 an unpredictable seed -- if it is not unpredictable, there
7534 is no point in blinding anyway). Make RSA blinding thread-safe
7535 by remembering the creator's thread ID in rsa->blinding and
7536 having all other threads use local one-time blinding factors
7537 (this requires more computation than sharing rsa->blinding, but
7538 avoids excessive locking; and if an RSA object is not shared
7539 between threads, blinding will still be very fast).
7542 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
7543 ENGINE as defaults for all supported algorithms irrespective of
7544 the 'flags' parameter. 'flags' is now honoured, so applications
7545 should make sure they are passing it correctly.
7548 *) Target "mingw" now allows native Windows code to be generated in
7549 the Cygwin environment as well as with the MinGW compiler.
7552 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7554 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7555 via timing by performing a MAC computation even if incorrect
7556 block cipher padding has been found. This is a countermeasure
7557 against active attacks where the attacker has to distinguish
7558 between bad padding and a MAC verification error. (CVE-2003-0078)
7560 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7561 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7562 Martin Vuagnoux (EPFL, Ilion)]
7564 *) Make the no-err option work as intended. The intention with no-err
7565 is not to have the whole error stack handling routines removed from
7566 libcrypto, it's only intended to remove all the function name and
7567 reason texts, thereby removing some of the footprint that may not
7568 be interesting if those errors aren't displayed anyway.
7570 NOTE: it's still possible for any application or module to have its
7571 own set of error texts inserted. The routines are there, just not
7572 used by default when no-err is given.
7575 *) Add support for FreeBSD on IA64.
7576 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7578 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7579 Kerberos function mit_des_cbc_cksum(). Before this change,
7580 the value returned by DES_cbc_cksum() was like the one from
7581 mit_des_cbc_cksum(), except the bytes were swapped.
7582 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7584 *) Allow an application to disable the automatic SSL chain building.
7585 Before this a rather primitive chain build was always performed in
7586 ssl3_output_cert_chain(): an application had no way to send the
7587 correct chain if the automatic operation produced an incorrect result.
7589 Now the chain builder is disabled if either:
7591 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7593 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7595 The reasoning behind this is that an application would not want the
7596 auto chain building to take place if extra chain certificates are
7597 present and it might also want a means of sending no additional
7598 certificates (for example the chain has two certificates and the
7602 *) Add the possibility to build without the ENGINE framework.
7603 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7605 *) Under Win32 gmtime() can return NULL: check return value in
7606 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7609 *) DSA routines: under certain error conditions uninitialized BN objects
7610 could be freed. Solution: make sure initialization is performed early
7611 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7612 Nils Larsch <nla@trustcenter.de> via PR#459)
7615 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7616 checked on reconnect on the client side, therefore session resumption
7617 could still fail with a "ssl session id is different" error. This
7618 behaviour is masked when SSL_OP_ALL is used due to
7619 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7620 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7621 followup to PR #377.
7624 *) IA-32 assembler support enhancements: unified ELF targets, support
7625 for SCO/Caldera platforms, fix for Cygwin shared build.
7628 *) Add support for FreeBSD on sparc64. As a consequence, support for
7629 FreeBSD on non-x86 processors is separate from x86 processors on
7630 the config script, much like the NetBSD support.
7631 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7633 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7635 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7638 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7639 code (06) was taken as the first octet of the session ID and the last
7640 octet was ignored consequently. As a result SSLv2 client side session
7641 caching could not have worked due to the session ID mismatch between
7643 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7647 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7648 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7652 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7653 seems that in spite of existing for more than a year, many application
7654 author have done nothing to provide the necessary callbacks, which
7655 means that this particular engine will not work properly anywhere.
7656 This is a very unfortunate situation which forces us, in the name
7657 of usability, to give the hw_ncipher.c a static lock, which is part
7659 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7660 appear in 0.9.8 or later. We EXPECT application authors to have
7661 dealt properly with this when 0.9.8 is released (unless we actually
7662 make such changes in the libcrypto locking code that changes will
7663 have to be made anyway).
7666 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7667 octets have been read, EOF or an error occurs. Without this change
7668 some truncated ASN1 structures will not produce an error.
7671 *) Disable Heimdal support, since it hasn't been fully implemented.
7672 Still give the possibility to force the use of Heimdal, but with
7673 warnings and a request that patches get sent to openssl-dev.
7676 *) Add the VC-CE target, introduce the WINCE sysname, and add
7677 INSTALL.WCE and appropriate conditionals to make it build.
7678 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7680 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7681 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7682 edit numbers of the version.
7683 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7685 *) Introduce safe string copy and catenation functions
7686 (BUF_strlcpy() and BUF_strlcat()).
7687 [Ben Laurie (CHATS) and Richard Levitte]
7689 *) Avoid using fixed-size buffers for one-line DNs.
7690 [Ben Laurie (CHATS)]
7692 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7693 resizing buffers containing secrets, and use where appropriate.
7694 [Ben Laurie (CHATS)]
7696 *) Avoid using fixed size buffers for configuration file location.
7697 [Ben Laurie (CHATS)]
7699 *) Avoid filename truncation for various CA files.
7700 [Ben Laurie (CHATS)]
7702 *) Use sizeof in preference to magic numbers.
7703 [Ben Laurie (CHATS)]
7705 *) Avoid filename truncation in cert requests.
7706 [Ben Laurie (CHATS)]
7708 *) Add assertions to check for (supposedly impossible) buffer
7710 [Ben Laurie (CHATS)]
7712 *) Don't cache truncated DNS entries in the local cache (this could
7713 potentially lead to a spoofing attack).
7714 [Ben Laurie (CHATS)]
7716 *) Fix various buffers to be large enough for hex/decimal
7717 representations in a platform independent manner.
7718 [Ben Laurie (CHATS)]
7720 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7721 resizing buffers containing secrets, and use where appropriate.
7722 [Ben Laurie (CHATS)]
7724 *) Add BIO_indent() to avoid much slightly worrying code to do
7726 [Ben Laurie (CHATS)]
7728 *) Convert sprintf()/BIO_puts() to BIO_printf().
7729 [Ben Laurie (CHATS)]
7731 *) buffer_gets() could terminate with the buffer only half
7733 [Ben Laurie (CHATS)]
7735 *) Add assertions to prevent user-supplied crypto functions from
7736 overflowing internal buffers by having large block sizes, etc.
7737 [Ben Laurie (CHATS)]
7739 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7741 [Ben Laurie (CHATS)]
7743 *) Eliminate unused copy of key in RC4.
7744 [Ben Laurie (CHATS)]
7746 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7747 [Ben Laurie (CHATS)]
7749 *) Fix off-by-one error in EGD path.
7750 [Ben Laurie (CHATS)]
7752 *) If RANDFILE path is too long, ignore instead of truncating.
7753 [Ben Laurie (CHATS)]
7755 *) Eliminate unused and incorrectly sized X.509 structure
7757 [Ben Laurie (CHATS)]
7759 *) Eliminate unused and dangerous function knumber().
7760 [Ben Laurie (CHATS)]
7762 *) Eliminate unused and dangerous structure, KSSL_ERR.
7763 [Ben Laurie (CHATS)]
7765 *) Protect against overlong session ID context length in an encoded
7766 session object. Since these are local, this does not appear to be
7768 [Ben Laurie (CHATS)]
7770 *) Change from security patch (see 0.9.6e below) that did not affect
7771 the 0.9.6 release series:
7773 Remote buffer overflow in SSL3 protocol - an attacker could
7774 supply an oversized master key in Kerberos-enabled versions.
7776 [Ben Laurie (CHATS)]
7778 *) Change the SSL kerb5 codes to match RFC 2712.
7781 *) Make -nameopt work fully for req and add -reqopt switch.
7782 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7784 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7785 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7787 *) Make sure tests can be performed even if the corresponding algorithms
7788 have been removed entirely. This was also the last step to make
7789 OpenSSL compilable with DJGPP under all reasonable conditions.
7790 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7792 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7793 to allow version independent disabling of normally unselected ciphers,
7794 which may be activated as a side-effect of selecting a single cipher.
7796 (E.g., cipher list string "RSA" enables ciphersuites that are left
7797 out of "ALL" because they do not provide symmetric encryption.
7798 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7799 [Lutz Jaenicke, Bodo Moeller]
7801 *) Add appropriate support for separate platform-dependent build
7802 directories. The recommended way to make a platform-dependent
7803 build directory is the following (tested on Linux), maybe with
7806 # Place yourself outside of the OpenSSL source tree. In
7807 # this example, the environment variable OPENSSL_SOURCE
7808 # is assumed to contain the absolute OpenSSL source directory.
7809 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7810 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7811 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7812 mkdir -p `dirname $F`
7813 ln -s $OPENSSL_SOURCE/$F $F
7816 To be absolutely sure not to disturb the source tree, a "make clean"
7817 is a good thing. If it isn't successful, don't worry about it,
7818 it probably means the source directory is very clean.
7821 *) Make sure any ENGINE control commands make local copies of string
7822 pointers passed to them whenever necessary. Otherwise it is possible
7823 the caller may have overwritten (or deallocated) the original string
7824 data when a later ENGINE operation tries to use the stored values.
7825 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7827 *) Improve diagnostics in file reading and command-line digests.
7828 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7830 *) Add AES modes CFB and OFB to the object database. Correct an
7831 error in AES-CFB decryption.
7834 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7835 allows existing EVP_CIPHER_CTX structures to be reused after
7836 calling EVP_*Final(). This behaviour is used by encryption
7837 BIOs and some applications. This has the side effect that
7838 applications must explicitly clean up cipher contexts with
7839 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7842 *) Check the values of dna and dnb in bn_mul_recursive before calling
7843 bn_mul_comba (a non zero value means the a or b arrays do not contain
7844 n2 elements) and fallback to bn_mul_normal if either is not zero.
7847 *) Fix escaping of non-ASCII characters when using the -subj option
7848 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7851 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7852 form for "surname", serialNumber has no short form.
7853 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7854 therefore remove "mail" short name for "internet 7".
7855 The OID for unique identifiers in X509 certificates is
7856 x500UniqueIdentifier, not uniqueIdentifier.
7857 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7860 *) Add an "init" command to the ENGINE config module and auto initialize
7861 ENGINEs. Without any "init" command the ENGINE will be initialized
7862 after all ctrl commands have been executed on it. If init=1 the
7863 ENGINE is initialized at that point (ctrls before that point are run
7864 on the uninitialized ENGINE and after on the initialized one). If
7865 init=0 then the ENGINE will not be initialized at all.
7868 *) Fix the 'app_verify_callback' interface so that the user-defined
7869 argument is actually passed to the callback: In the
7870 SSL_CTX_set_cert_verify_callback() prototype, the callback
7871 declaration has been changed from
7874 int (*cb)(X509_STORE_CTX *,void *);
7875 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7876 i=s->ctx->app_verify_callback(&ctx)
7877 has been changed into
7878 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7880 To update applications using SSL_CTX_set_cert_verify_callback(),
7881 a dummy argument can be added to their callback functions.
7882 [D. K. Smetters <smetters@parc.xerox.com>]
7884 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7885 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7887 *) Add and OPENSSL_LOAD_CONF define which will cause
7888 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7889 This allows older applications to transparently support certain
7890 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7891 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7892 load the config file and OPENSSL_add_all_algorithms_conf() which will
7893 always load it have also been added.
7896 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7897 Adjust NIDs and EVP layer.
7898 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7900 *) Config modules support in openssl utility.
7902 Most commands now load modules from the config file,
7903 though in a few (such as version) this isn't done
7904 because it couldn't be used for anything.
7906 In the case of ca and req the config file used is
7907 the same as the utility itself: that is the -config
7908 command line option can be used to specify an
7912 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7913 use "openssl_conf" if filename is NULL use default openssl config file.
7916 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7917 config section name. Add a new flag to tolerate a missing config file
7918 and move code to CONF_modules_load_file().
7921 *) Support for crypto accelerator cards from Accelerated Encryption
7922 Processing, www.aep.ie. (Use engine 'aep')
7923 The support was copied from 0.9.6c [engine] and adapted/corrected
7924 to work with the new engine framework.
7925 [AEP Inc. and Richard Levitte]
7927 *) Support for SureWare crypto accelerator cards from Baltimore
7928 Technologies. (Use engine 'sureware')
7929 The support was copied from 0.9.6c [engine] and adapted
7930 to work with the new engine framework.
7933 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7934 make the newer ENGINE framework commands for the CHIL engine work.
7935 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7937 *) Make it possible to produce shared libraries on ReliantUNIX.
7938 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7940 *) Add the configuration target debug-linux-ppro.
7941 Make 'openssl rsa' use the general key loading routines
7942 implemented in apps.c, and make those routines able to
7943 handle the key format FORMAT_NETSCAPE and the variant
7945 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7947 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7948 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7950 *) Add -keyform to rsautl, and document -engine.
7951 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7953 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7954 BIO_R_NO_SUCH_FILE error code rather than the generic
7955 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7958 *) Add new functions
7960 ERR_peek_last_error_line
7961 ERR_peek_last_error_line_data.
7962 These are similar to
7965 ERR_peek_error_line_data,
7966 but report on the latest error recorded rather than the first one
7967 still in the error queue.
7968 [Ben Laurie, Bodo Moeller]
7970 *) default_algorithms option in ENGINE config module. This allows things
7972 default_algorithms = ALL
7973 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7976 *) Preliminary ENGINE config module.
7979 *) New experimental application configuration code.
7982 *) Change the AES code to follow the same name structure as all other
7983 symmetric ciphers, and behave the same way. Move everything to
7984 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7985 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7987 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7988 [Ben Laurie and Theo de Raadt]
7990 *) Add option to output public keys in req command.
7991 [Massimiliano Pala madwolf@openca.org]
7993 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7994 (up to about 10% better than before for P-192 and P-224).
7997 *) New functions/macros
7999 SSL_CTX_set_msg_callback(ctx, cb)
8000 SSL_CTX_set_msg_callback_arg(ctx, arg)
8001 SSL_set_msg_callback(ssl, cb)
8002 SSL_set_msg_callback_arg(ssl, arg)
8004 to request calling a callback function
8006 void cb(int write_p, int version, int content_type,
8007 const void *buf, size_t len, SSL *ssl, void *arg)
8009 whenever a protocol message has been completely received
8010 (write_p == 0) or sent (write_p == 1). Here 'version' is the
8011 protocol version according to which the SSL library interprets
8012 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
8013 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
8014 the content type as defined in the SSL 3.0/TLS 1.0 protocol
8015 specification (change_cipher_spec(20), alert(21), handshake(22)).
8016 'buf' and 'len' point to the actual message, 'ssl' to the
8017 SSL object, and 'arg' is the application-defined value set by
8018 SSL[_CTX]_set_msg_callback_arg().
8020 'openssl s_client' and 'openssl s_server' have new '-msg' options
8021 to enable a callback that displays all protocol messages.
8024 *) Change the shared library support so shared libraries are built as
8025 soon as the corresponding static library is finished, and thereby get
8026 openssl and the test programs linked against the shared library.
8027 This still only happens when the keyword "shard" has been given to
8028 the configuration scripts.
8030 NOTE: shared library support is still an experimental thing, and
8031 backward binary compatibility is still not guaranteed.
8032 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
8034 *) Add support for Subject Information Access extension.
8035 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
8037 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
8038 additional bytes when new memory had to be allocated, not just
8039 when reusing an existing buffer.
8042 *) New command line and configuration option 'utf8' for the req command.
8043 This allows field values to be specified as UTF8 strings.
8046 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
8047 runs for the former and machine-readable output for the latter.
8050 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
8051 of the e-mail address in the DN (i.e., it will go into a certificate
8052 extension only). The new configuration file option 'email_in_dn = no'
8053 has the same effect.
8054 [Massimiliano Pala madwolf@openca.org]
8056 *) Change all functions with names starting with des_ to be starting
8057 with DES_ instead. Add wrappers that are compatible with libdes,
8058 but are named _ossl_old_des_*. Finally, add macros that map the
8059 des_* symbols to the corresponding _ossl_old_des_* if libdes
8060 compatibility is desired. If OpenSSL 0.9.6c compatibility is
8061 desired, the des_* symbols will be mapped to DES_*, with one
8064 Since we provide two compatibility mappings, the user needs to
8065 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
8066 compatibility is desired. The default (i.e., when that macro
8067 isn't defined) is OpenSSL 0.9.6c compatibility.
8069 There are also macros that enable and disable the support of old
8070 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
8071 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
8072 are defined, the default will apply: to support the old des routines.
8074 In either case, one must include openssl/des.h to get the correct
8075 definitions. Do not try to just include openssl/des_old.h, that
8078 NOTE: This is a major break of an old API into a new one. Software
8079 authors are encouraged to switch to the DES_ style functions. Some
8080 time in the future, des_old.h and the libdes compatibility functions
8081 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
8082 default), and then completely removed.
8085 *) Test for certificates which contain unsupported critical extensions.
8086 If such a certificate is found during a verify operation it is
8087 rejected by default: this behaviour can be overridden by either
8088 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
8089 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
8090 X509_supported_extension() has also been added which returns 1 if a
8091 particular extension is supported.
8094 *) Modify the behaviour of EVP cipher functions in similar way to digests
8095 to retain compatibility with existing code.
8098 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
8099 compatibility with existing code. In particular the 'ctx' parameter does
8100 not have to be to be initialized before the call to EVP_DigestInit() and
8101 it is tidied up after a call to EVP_DigestFinal(). New function
8102 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
8103 EVP_MD_CTX_copy() changed to not require the destination to be
8104 initialized valid and new function EVP_MD_CTX_copy_ex() added which
8105 requires the destination to be valid.
8107 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
8108 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
8111 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
8112 so that complete 'Handshake' protocol structures are kept in memory
8113 instead of overwriting 'msg_type' and 'length' with 'body' data.
8116 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
8117 [Massimo Santin via Richard Levitte]
8119 *) Major restructuring to the underlying ENGINE code. This includes
8120 reduction of linker bloat, separation of pure "ENGINE" manipulation
8121 (initialisation, etc) from functionality dealing with implementations
8122 of specific crypto interfaces. This change also introduces integrated
8123 support for symmetric ciphers and digest implementations - so ENGINEs
8124 can now accelerate these by providing EVP_CIPHER and EVP_MD
8125 implementations of their own. This is detailed in crypto/engine/README
8126 as it couldn't be adequately described here. However, there are a few
8127 API changes worth noting - some RSA, DSA, DH, and RAND functions that
8128 were changed in the original introduction of ENGINE code have now
8129 reverted back - the hooking from this code to ENGINE is now a good
8130 deal more passive and at run-time, operations deal directly with
8131 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
8132 dereferencing through an ENGINE pointer any more. Also, the ENGINE
8133 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
8134 they were not being used by the framework as there is no concept of a
8135 BIGNUM_METHOD and they could not be generalised to the new
8136 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
8137 ENGINE_cpy() has been removed as it cannot be consistently defined in
8141 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
8144 *) Change mkdef.pl to sort symbols that get the same entry number,
8145 and make sure the automatically generated functions ERR_load_*
8146 become part of libeay.num as well.
8149 *) New function SSL_renegotiate_pending(). This returns true once
8150 renegotiation has been requested (either SSL_renegotiate() call
8151 or HelloRequest/ClientHello received from the peer) and becomes
8152 false once a handshake has been completed.
8153 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
8154 sends a HelloRequest, but does not ensure that a handshake takes
8155 place. SSL_renegotiate_pending() is useful for checking if the
8156 client has followed the request.)
8159 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
8160 By default, clients may request session resumption even during
8161 renegotiation (if session ID contexts permit); with this option,
8162 session resumption is possible only in the first handshake.
8164 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
8165 more bits available for options that should not be part of
8166 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
8169 *) Add some demos for certificate and certificate request creation.
8172 *) Make maximum certificate chain size accepted from the peer application
8173 settable (SSL*_get/set_max_cert_list()), as proposed by
8174 "Douglas E. Engert" <deengert@anl.gov>.
8177 *) Add support for shared libraries for Unixware-7
8178 (Boyd Lynn Gerber <gerberb@zenez.com>).
8181 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
8182 be done prior to destruction. Use this to unload error strings from
8183 ENGINEs that load their own error strings. NB: This adds two new API
8184 functions to "get" and "set" this destroy handler in an ENGINE.
8187 *) Alter all existing ENGINE implementations (except "openssl" and
8188 "openbsd") to dynamically instantiate their own error strings. This
8189 makes them more flexible to be built both as statically-linked ENGINEs
8190 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
8191 Also, add stub code to each that makes building them as self-contained
8192 shared-libraries easier (see README.ENGINE).
8195 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
8196 implementations into applications that are completely implemented in
8197 self-contained shared-libraries. The "dynamic" ENGINE exposes control
8198 commands that can be used to configure what shared-library to load and
8199 to control aspects of the way it is handled. Also, made an update to
8200 the README.ENGINE file that brings its information up-to-date and
8201 provides some information and instructions on the "dynamic" ENGINE
8202 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
8205 *) Make it possible to unload ranges of ERR strings with a new
8206 "ERR_unload_strings" function.
8209 *) Add a copy() function to EVP_MD.
8212 *) Make EVP_MD routines take a context pointer instead of just the
8213 md_data void pointer.
8216 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
8217 that the digest can only process a single chunk of data
8218 (typically because it is provided by a piece of
8219 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
8220 is only going to provide a single chunk of data, and hence the
8221 framework needn't accumulate the data for oneshot drivers.
8224 *) As with "ERR", make it possible to replace the underlying "ex_data"
8225 functions. This change also alters the storage and management of global
8226 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
8227 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
8228 index counters. The API functions that use this state have been changed
8229 to take a "class_index" rather than pointers to the class's local STACK
8230 and counter, and there is now an API function to dynamically create new
8231 classes. This centralisation allows us to (a) plug a lot of the
8232 thread-safety problems that existed, and (b) makes it possible to clean
8233 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
8234 such data would previously have always leaked in application code and
8235 workarounds were in place to make the memory debugging turn a blind eye
8236 to it. Application code that doesn't use this new function will still
8237 leak as before, but their memory debugging output will announce it now
8238 rather than letting it slide.
8240 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
8241 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
8242 has a return value to indicate success or failure.
8245 *) Make it possible to replace the underlying "ERR" functions such that the
8246 global state (2 LHASH tables and 2 locks) is only used by the "default"
8247 implementation. This change also adds two functions to "get" and "set"
8248 the implementation prior to it being automatically set the first time
8249 any other ERR function takes place. Ie. an application can call "get",
8250 pass the return value to a module it has just loaded, and that module
8251 can call its own "set" function using that value. This means the
8252 module's "ERR" operations will use (and modify) the error state in the
8253 application and not in its own statically linked copy of OpenSSL code.
8256 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
8257 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
8258 the operation, and provides a more encapsulated way for external code
8259 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
8260 to use these functions rather than manually incrementing the counts.
8262 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
8265 *) Add EVP test program.
8268 *) Add symmetric cipher support to ENGINE. Expect the API to change!
8271 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
8272 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
8273 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
8274 These allow a CRL to be built without having to access X509_CRL fields
8275 directly. Modify 'ca' application to use new functions.
8278 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
8279 bug workarounds. Rollback attack detection is a security feature.
8280 The problem will only arise on OpenSSL servers when TLSv1 is not
8281 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
8282 Software authors not wanting to support TLSv1 will have special reasons
8283 for their choice and can explicitly enable this option.
8284 [Bodo Moeller, Lutz Jaenicke]
8286 *) Rationalise EVP so it can be extended: don't include a union of
8287 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
8288 (similar to those existing for EVP_CIPHER_CTX).
8293 EVP_MD_CTX_init(&md); /* new function call */
8294 EVP_DigestInit(&md, EVP_sha1());
8295 EVP_DigestUpdate(&md, in, len);
8296 EVP_DigestFinal(&md, out, NULL);
8297 EVP_MD_CTX_cleanup(&md); /* new function call */
8301 *) Make DES key schedule conform to the usual scheme, as well as
8302 correcting its structure. This means that calls to DES functions
8303 now have to pass a pointer to a des_key_schedule instead of a
8304 plain des_key_schedule (which was actually always a pointer
8307 des_key_schedule ks;
8309 des_set_key_checked(..., &ks);
8310 des_ncbc_encrypt(..., &ks, ...);
8312 (Note that a later change renames 'des_...' into 'DES_...'.)
8315 *) Initial reduction of linker bloat: the use of some functions, such as
8316 PEM causes large amounts of unused functions to be linked in due to
8317 poor organisation. For example pem_all.c contains every PEM function
8318 which has a knock on effect of linking in large amounts of (unused)
8319 ASN1 code. Grouping together similar functions and splitting unrelated
8320 functions prevents this.
8323 *) Cleanup of EVP macros.
8326 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
8327 correct _ecb suffix.
8330 *) Add initial OCSP responder support to ocsp application. The
8331 revocation information is handled using the text based index
8332 use by the ca application. The responder can either handle
8333 requests generated internally, supplied in files (for example
8334 via a CGI script) or using an internal minimal server.
8337 *) Add configuration choices to get zlib compression for TLS.
8340 *) Changes to Kerberos SSL for RFC 2712 compliance:
8341 1. Implemented real KerberosWrapper, instead of just using
8342 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
8343 2. Implemented optional authenticator field of KerberosWrapper.
8345 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
8346 and authenticator structs; see crypto/krb5/.
8348 Generalized Kerberos calls to support multiple Kerberos libraries.
8349 [Vern Staats <staatsvr@asc.hpc.mil>,
8350 Jeffrey Altman <jaltman@columbia.edu>
8351 via Richard Levitte]
8353 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
8354 already does with RSA. testdsa.h now has 'priv_key/pub_key'
8355 values for each of the key sizes rather than having just
8356 parameters (and 'speed' generating keys each time).
8359 *) Speed up EVP routines.
8362 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
8363 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
8364 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
8365 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
8367 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
8368 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
8369 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
8372 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
8374 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
8377 *) Added the OS2-EMX target.
8378 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
8380 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
8381 to support NCONF routines in extension code. New function CONF_set_nconf()
8382 to allow functions which take an NCONF to also handle the old LHASH
8383 structure: this means that the old CONF compatible routines can be
8384 retained (in particular wrt extensions) without having to duplicate the
8385 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
8388 *) Enhance the general user interface with mechanisms for inner control
8389 and with possibilities to have yes/no kind of prompts.
8392 *) Change all calls to low level digest routines in the library and
8393 applications to use EVP. Add missing calls to HMAC_cleanup() and
8394 don't assume HMAC_CTX can be copied using memcpy().
8395 [Verdon Walker <VWalker@novell.com>, Steve Henson]
8397 *) Add the possibility to control engines through control names but with
8398 arbitrary arguments instead of just a string.
8399 Change the key loaders to take a UI_METHOD instead of a callback
8400 function pointer. NOTE: this breaks binary compatibility with earlier
8401 versions of OpenSSL [engine].
8402 Adapt the nCipher code for these new conditions and add a card insertion
8406 *) Enhance the general user interface with mechanisms to better support
8407 dialog box interfaces, application-defined prompts, the possibility
8408 to use defaults (for example default passwords from somewhere else)
8409 and interrupts/cancellations.
8412 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
8413 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
8416 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
8417 tidy up some unnecessarily weird code in 'sk_new()').
8418 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
8420 *) Change the key loading routines for ENGINEs to use the same kind
8421 callback (pem_password_cb) as all other routines that need this
8425 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
8426 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
8427 than this minimum value is recommended.
8430 *) New random seeder for OpenVMS, using the system process statistics
8431 that are easily reachable.
8434 *) Windows apparently can't transparently handle global
8435 variables defined in DLLs. Initialisations such as:
8437 const ASN1_ITEM *it = &ASN1_INTEGER_it;
8439 won't compile. This is used by the any applications that need to
8440 declare their own ASN1 modules. This was fixed by adding the option
8441 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
8442 needed for static libraries under Win32.
8445 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
8446 setting of purpose and trust fields. New X509_STORE trust and
8447 purpose functions and tidy up setting in other SSL functions.
8450 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
8451 structure. These are inherited by X509_STORE_CTX when it is
8452 initialised. This allows various defaults to be set in the
8453 X509_STORE structure (such as flags for CRL checking and custom
8454 purpose or trust settings) for functions which only use X509_STORE_CTX
8455 internally such as S/MIME.
8457 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
8458 trust settings if they are not set in X509_STORE. This allows X509_STORE
8459 purposes and trust (in S/MIME for example) to override any set by default.
8461 Add command line options for CRL checking to smime, s_client and s_server
8465 *) Initial CRL based revocation checking. If the CRL checking flag(s)
8466 are set then the CRL is looked up in the X509_STORE structure and
8467 its validity and signature checked, then if the certificate is found
8468 in the CRL the verify fails with a revoked error.
8470 Various new CRL related callbacks added to X509_STORE_CTX structure.
8472 Command line options added to 'verify' application to support this.
8474 This needs some additional work, such as being able to handle multiple
8475 CRLs with different times, extension based lookup (rather than just
8476 by subject name) and ultimately more complete V2 CRL extension
8480 *) Add a general user interface API (crypto/ui/). This is designed
8481 to replace things like des_read_password and friends (backward
8482 compatibility functions using this new API are provided).
8483 The purpose is to remove prompting functions from the DES code
8484 section as well as provide for prompting through dialog boxes in
8485 a window system and the like.
8488 *) Add "ex_data" support to ENGINE so implementations can add state at a
8489 per-structure level rather than having to store it globally.
8492 *) Make it possible for ENGINE structures to be copied when retrieved by
8493 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
8494 This causes the "original" ENGINE structure to act like a template,
8495 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
8496 operational state can be localised to each ENGINE structure, despite the
8497 fact they all share the same "methods". New ENGINE structures returned in
8498 this case have no functional references and the return value is the single
8499 structural reference. This matches the single structural reference returned
8500 by ENGINE_by_id() normally, when it is incremented on the pre-existing
8504 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
8505 needs to match any other type at all we need to manually clear the
8509 *) Changes to the "openssl engine" utility to include;
8510 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
8511 about an ENGINE's available control commands.
8512 - executing control commands from command line arguments using the
8513 '-pre' and '-post' switches. '-post' is only used if '-t' is
8514 specified and the ENGINE is successfully initialised. The syntax for
8515 the individual commands are colon-separated, for example;
8516 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
8519 *) New dynamic control command support for ENGINEs. ENGINEs can now
8520 declare their own commands (numbers), names (strings), descriptions,
8521 and input types for run-time discovery by calling applications. A
8522 subset of these commands are implicitly classed as "executable"
8523 depending on their input type, and only these can be invoked through
8524 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
8525 can be based on user input, config files, etc). The distinction is
8526 that "executable" commands cannot return anything other than a boolean
8527 result and can only support numeric or string input, whereas some
8528 discoverable commands may only be for direct use through
8529 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
8530 pointers, or other custom uses. The "executable" commands are to
8531 support parameterisations of ENGINE behaviour that can be
8532 unambiguously defined by ENGINEs and used consistently across any
8533 OpenSSL-based application. Commands have been added to all the
8534 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
8535 control over shared-library paths without source code alterations.
8538 *) Changed all ENGINE implementations to dynamically allocate their
8539 ENGINEs rather than declaring them statically. Apart from this being
8540 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
8541 this also allows the implementations to compile without using the
8542 internal engine_int.h header.
8545 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8546 'const' value. Any code that should be able to modify a RAND_METHOD
8547 should already have non-const pointers to it (ie. they should only
8548 modify their own ones).
8551 *) Made a variety of little tweaks to the ENGINE code.
8552 - "atalla" and "ubsec" string definitions were moved from header files
8553 to C code. "nuron" string definitions were placed in variables
8554 rather than hard-coded - allowing parameterisation of these values
8555 later on via ctrl() commands.
8556 - Removed unused "#if 0"'d code.
8557 - Fixed engine list iteration code so it uses ENGINE_free() to release
8558 structural references.
8559 - Constified the RAND_METHOD element of ENGINE structures.
8560 - Constified various get/set functions as appropriate and added
8561 missing functions (including a catch-all ENGINE_cpy that duplicates
8562 all ENGINE values onto a new ENGINE except reference counts/state).
8563 - Removed NULL parameter checks in get/set functions. Setting a method
8564 or function to NULL is a way of cancelling out a previously set
8565 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8566 and doesn't justify the extra error symbols and code.
8567 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8568 flags from engine_int.h to engine.h.
8569 - Changed prototypes for ENGINE handler functions (init(), finish(),
8570 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8573 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8574 to the algorithm using long division. The binary algorithm can be
8575 used only if the modulus is odd. On 32-bit systems, it is faster
8576 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8577 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8578 up to 450 bits. In 64-bit environments, the binary algorithm
8579 appears to be advantageous for much longer moduli; here we use it
8580 for moduli up to 2048 bits.
8583 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8584 could not support the combine flag in choice fields.
8587 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8588 extensions from a certificate request to the certificate.
8591 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8592 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8593 file: this allows the display of the certificate about to be
8594 signed to be customised, to allow certain fields to be included
8595 or excluded and extension details. The old system didn't display
8596 multicharacter strings properly, omitted fields not in the policy
8597 and couldn't display additional details such as extensions.
8600 *) Function EC_POINTs_mul for multiple scalar multiplication
8601 of an arbitrary number of elliptic curve points
8602 \sum scalars[i]*points[i],
8603 optionally including the generator defined for the EC_GROUP:
8604 scalar*generator + \sum scalars[i]*points[i].
8606 EC_POINT_mul is a simple wrapper function for the typical case
8607 that the point list has just one item (besides the optional
8611 *) First EC_METHODs for curves over GF(p):
8613 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8614 operations and provides various method functions that can also
8615 operate with faster implementations of modular arithmetic.
8617 EC_GFp_mont_method() reuses most functions that are part of
8618 EC_GFp_simple_method, but uses Montgomery arithmetic.
8620 [Bodo Moeller; point addition and point doubling
8621 implementation directly derived from source code provided by
8622 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8624 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8625 crypto/ec/ec_lib.c):
8627 Curves are EC_GROUP objects (with an optional group generator)
8628 based on EC_METHODs that are built into the library.
8630 Points are EC_POINT objects based on EC_GROUP objects.
8632 Most of the framework would be able to handle curves over arbitrary
8633 finite fields, but as there are no obvious types for fields other
8634 than GF(p), some functions are limited to that for now.
8637 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8638 that the file contains a complete HTTP response.
8641 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8642 change the def and num file printf format specifier from "%-40sXXX"
8643 to "%-39s XXX". The latter will always guarantee a space after the
8644 field while the former will cause them to run together if the field
8645 is 40 of more characters long.
8648 *) Constify the cipher and digest 'method' functions and structures
8649 and modify related functions to take constant EVP_MD and EVP_CIPHER
8653 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8654 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8657 *) Modify EVP_Digest*() routines so they now return values. Although the
8658 internal software routines can never fail additional hardware versions
8662 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8664 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8665 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8668 ERR_R_NESTED_ASN1_ERROR
8670 ERR_R_MISSING_ASN1_EOS
8671 were 4 .. 9, conflicting with
8672 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8674 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8675 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8677 Add new error code 'ERR_R_INTERNAL_ERROR'.
8680 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8684 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8685 sets the subject name for a new request or supersedes the
8686 subject name in a given request. Formats that can be parsed are
8687 'CN=Some Name, OU=myOU, C=IT'
8689 'CN=Some Name/OU=myOU/C=IT'.
8691 Add options '-batch' and '-verbose' to 'openssl req'.
8692 [Massimiliano Pala <madwolf@hackmasters.net>]
8694 *) Introduce the possibility to access global variables through
8695 functions on platform were that's the best way to handle exporting
8696 global variables in shared libraries. To enable this functionality,
8697 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8698 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8699 is normally done by Configure or something similar).
8701 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8702 in the source file (foo.c) like this:
8704 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8705 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8707 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8708 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8710 OPENSSL_DECLARE_GLOBAL(int,foo);
8711 #define foo OPENSSL_GLOBAL_REF(foo)
8712 OPENSSL_DECLARE_GLOBAL(double,bar);
8713 #define bar OPENSSL_GLOBAL_REF(bar)
8715 The #defines are very important, and therefore so is including the
8716 header file everywhere where the defined globals are used.
8718 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8719 of ASN.1 items, but that structure is a bit different.
8721 The largest change is in util/mkdef.pl which has been enhanced with
8722 better and easier to understand logic to choose which symbols should
8723 go into the Windows .def files as well as a number of fixes and code
8724 cleanup (among others, algorithm keywords are now sorted
8725 lexicographically to avoid constant rewrites).
8728 *) In BN_div() keep a copy of the sign of 'num' before writing the
8729 result to 'rm' because if rm==num the value will be overwritten
8730 and produce the wrong result if 'num' is negative: this caused
8731 problems with BN_mod() and BN_nnmod().
8734 *) Function OCSP_request_verify(). This checks the signature on an
8735 OCSP request and verifies the signer certificate. The signer
8736 certificate is just checked for a generic purpose and OCSP request
8740 *) Add OCSP_check_validity() function to check the validity of OCSP
8741 responses. OCSP responses are prepared in real time and may only
8742 be a few seconds old. Simply checking that the current time lies
8743 between thisUpdate and nextUpdate max reject otherwise valid responses
8744 caused by either OCSP responder or client clock inaccuracy. Instead
8745 we allow thisUpdate and nextUpdate to fall within a certain period of
8746 the current time. The age of the response can also optionally be
8747 checked. Two new options -validity_period and -status_age added to
8751 *) If signature or public key algorithm is unrecognized print out its
8752 OID rather that just UNKNOWN.
8755 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8756 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8757 ID to be generated from the issuer certificate alone which can then be
8758 passed to OCSP_id_issuer_cmp().
8761 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8762 ASN1 modules to export functions returning ASN1_ITEM pointers
8763 instead of the ASN1_ITEM structures themselves. This adds several
8764 new macros which allow the underlying ASN1 function/structure to
8765 be accessed transparently. As a result code should not use ASN1_ITEM
8766 references directly (such as &X509_it) but instead use the relevant
8767 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8768 use of the new ASN1 code on platforms where exporting structures
8769 is problematical (for example in shared libraries) but exporting
8770 functions returning pointers to structures is not.
8773 *) Add support for overriding the generation of SSL/TLS session IDs.
8774 These callbacks can be registered either in an SSL_CTX or per SSL.
8775 The purpose of this is to allow applications to control, if they wish,
8776 the arbitrary values chosen for use as session IDs, particularly as it
8777 can be useful for session caching in multiple-server environments. A
8778 command-line switch for testing this (and any client code that wishes
8779 to use such a feature) has been added to "s_server".
8780 [Geoff Thorpe, Lutz Jaenicke]
8782 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8783 of the form '#if defined(...) || defined(...) || ...' and
8784 '#if !defined(...) && !defined(...) && ...'. This also avoids
8785 the growing number of special cases it was previously handling.
8788 *) Make all configuration macros available for application by making
8789 sure they are available in opensslconf.h, by giving them names starting
8790 with "OPENSSL_" to avoid conflicts with other packages and by making
8791 sure e_os2.h will cover all platform-specific cases together with
8793 Additionally, it is now possible to define configuration/platform-
8794 specific names (called "system identities"). In the C code, these
8795 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8796 macro with the name beginning with "OPENSSL_SYS_", which is determined
8797 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8801 *) New option -set_serial to 'req' and 'x509' this allows the serial
8802 number to use to be specified on the command line. Previously self
8803 signed certificates were hard coded with serial number 0 and the
8804 CA options of 'x509' had to use a serial number in a file which was
8808 *) New options to 'ca' utility to support V2 CRL entry extensions.
8809 Currently CRL reason, invalidity date and hold instruction are
8810 supported. Add new CRL extensions to V3 code and some new objects.
8813 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8814 disable standard block padding (aka PKCS#5 padding) in the EVP
8815 API, which was previously mandatory. This means that the data is
8816 not padded in any way and so the total length much be a multiple
8817 of the block size, otherwise an error occurs.
8820 *) Initial (incomplete) OCSP SSL support.
8823 *) New function OCSP_parse_url(). This splits up a URL into its host,
8824 port and path components: primarily to parse OCSP URLs. New -url
8825 option to ocsp utility.
8828 *) New nonce behavior. The return value of OCSP_check_nonce() now
8829 reflects the various checks performed. Applications can decide
8830 whether to tolerate certain situations such as an absent nonce
8831 in a response when one was present in a request: the ocsp application
8832 just prints out a warning. New function OCSP_add1_basic_nonce()
8833 this is to allow responders to include a nonce in a response even if
8834 the request is nonce-less.
8837 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8838 skipped when using openssl x509 multiple times on a single input file,
8839 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8842 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8843 set string type: to handle setting ASN1_TIME structures. Fix ca
8844 utility to correctly initialize revocation date of CRLs.
8847 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8848 the clients preferred ciphersuites and rather use its own preferences.
8849 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8850 Internet Explorer by ensuring unchanged hash method during stepup.
8851 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8854 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8855 to aes and add a new 'exist' option to print out symbols that don't
8859 *) Additional options to ocsp utility to allow flags to be set and
8860 additional certificates supplied.
8863 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8864 OCSP client a number of certificate to only verify the response
8868 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8869 handle the new API. Currently only ECB, CBC modes supported. Add new
8872 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8873 Encryption Standard (AES) Ciphersuites for Transport Layer
8874 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8875 not enabled by default and were not part of the "ALL" ciphersuite
8876 alias because they were not yet official; they could be
8877 explicitly requested by specifying the "AESdraft" ciphersuite
8878 group alias. In the final release of OpenSSL 0.9.7, the group
8879 alias is called "AES" and is part of "ALL".)
8880 [Ben Laurie, Steve Henson, Bodo Moeller]
8882 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8883 request to response.
8886 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8887 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8888 extract information from a certificate request. OCSP_response_create()
8889 creates a response and optionally adds a basic response structure.
8890 OCSP_basic_add1_status() adds a complete single response to a basic
8891 response and returns the OCSP_SINGLERESP structure just added (to allow
8892 extensions to be included for example). OCSP_basic_add1_cert() adds a
8893 certificate to a basic response and OCSP_basic_sign() signs a basic
8894 response with various flags. New helper functions ASN1_TIME_check()
8895 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8896 (converts ASN1_TIME to GeneralizedTime).
8899 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8900 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8901 structure from a certificate. X509_pubkey_digest() digests the public_key
8902 contents: this is used in various key identifiers.
8905 *) Make sk_sort() tolerate a NULL argument.
8906 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8908 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8909 passed by the function are trusted implicitly. If any of them signed the
8910 response then it is assumed to be valid and is not verified.
8913 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8914 to data. This was previously part of the PKCS7 ASN1 code. This
8915 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8916 [Steve Henson, reported by Kenneth R. Robinette
8917 <support@securenetterm.com>]
8919 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8920 routines: without these tracing memory leaks is very painful.
8921 Fix leaks in PKCS12 and PKCS7 routines.
8924 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8925 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8926 effectively meant GeneralizedTime would never be used. Now it
8927 is initialised to -1 but X509_time_adj() now has to check the value
8928 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8929 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8930 [Steve Henson, reported by Kenneth R. Robinette
8931 <support@securenetterm.com>]
8933 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8934 result in a zero length in the ASN1_INTEGER structure which was
8935 not consistent with the structure when d2i_ASN1_INTEGER() was used
8936 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8937 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8938 where it did not print out a minus for negative ASN1_INTEGER.
8941 *) Add summary printout to ocsp utility. The various functions which
8942 convert status values to strings have been renamed to:
8943 OCSP_response_status_str(), OCSP_cert_status_str() and
8944 OCSP_crl_reason_str() and are no longer static. New options
8945 to verify nonce values and to disable verification. OCSP response
8946 printout format cleaned up.
8949 *) Add additional OCSP certificate checks. These are those specified
8950 in RFC2560. This consists of two separate checks: the CA of the
8951 certificate being checked must either be the OCSP signer certificate
8952 or the issuer of the OCSP signer certificate. In the latter case the
8953 OCSP signer certificate must contain the OCSP signing extended key
8954 usage. This check is performed by attempting to match the OCSP
8955 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8956 in the OCSP_CERTID structures of the response.
8959 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8960 and related routines. This uses the standard OpenSSL certificate
8961 verify routines to perform initial checks (just CA validity) and
8962 to obtain the certificate chain. Then additional checks will be
8963 performed on the chain. Currently the root CA is checked to see
8964 if it is explicitly trusted for OCSP signing. This is used to set
8965 a root CA as a global signing root: that is any certificate that
8966 chains to that CA is an acceptable OCSP signing certificate.
8969 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8970 extensions from a separate configuration file.
8971 As when reading extensions from the main configuration file,
8972 the '-extensions ...' option may be used for specifying the
8974 [Massimiliano Pala <madwolf@comune.modena.it>]
8976 *) New OCSP utility. Allows OCSP requests to be generated or
8977 read. The request can be sent to a responder and the output
8978 parsed, outputted or printed in text form. Not complete yet:
8979 still needs to check the OCSP response validity.
8982 *) New subcommands for 'openssl ca':
8983 'openssl ca -status <serial>' prints the status of the cert with
8984 the given serial number (according to the index file).
8985 'openssl ca -updatedb' updates the expiry status of certificates
8987 [Massimiliano Pala <madwolf@comune.modena.it>]
8989 *) New '-newreq-nodes' command option to CA.pl. This is like
8990 '-newreq', but calls 'openssl req' with the '-nodes' option
8991 so that the resulting key is not encrypted.
8992 [Damien Miller <djm@mindrot.org>]
8994 *) New configuration for the GNU Hurd.
8995 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8997 *) Initial code to implement OCSP basic response verify. This
8998 is currently incomplete. Currently just finds the signer's
8999 certificate and verifies the signature on the response.
9002 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
9003 value of OPENSSLDIR. This is available via the new '-d' option
9004 to 'openssl version', and is also included in 'openssl version -a'.
9007 *) Allowing defining memory allocation callbacks that will be given
9008 file name and line number information in additional arguments
9009 (a const char* and an int). The basic functionality remains, as
9010 well as the original possibility to just replace malloc(),
9011 realloc() and free() by functions that do not know about these
9012 additional arguments. To register and find out the current
9013 settings for extended allocation functions, the following
9014 functions are provided:
9016 CRYPTO_set_mem_ex_functions
9017 CRYPTO_set_locked_mem_ex_functions
9018 CRYPTO_get_mem_ex_functions
9019 CRYPTO_get_locked_mem_ex_functions
9021 These work the same way as CRYPTO_set_mem_functions and friends.
9022 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
9023 extended allocation function is enabled.
9024 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
9025 a conventional allocation function is enabled.
9026 [Richard Levitte, Bodo Moeller]
9028 *) Finish off removing the remaining LHASH function pointer casts.
9029 There should no longer be any prototype-casting required when using
9030 the LHASH abstraction, and any casts that remain are "bugs". See
9031 the callback types and macros at the head of lhash.h for details
9032 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
9035 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
9036 If /dev/[u]random devices are not available or do not return enough
9037 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
9039 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
9040 /etc/entropy will be queried once each in this sequence, querying stops
9041 when enough entropy was collected without querying more sockets.
9044 *) Change the Unix RAND_poll() variant to be able to poll several
9045 random devices, as specified by DEVRANDOM, until a sufficient amount
9046 of data has been collected. We spend at most 10 ms on each file
9047 (select timeout) and read in non-blocking mode. DEVRANDOM now
9048 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
9049 (previously it was just the string "/dev/urandom"), so on typical
9050 platforms the 10 ms delay will never occur.
9051 Also separate out the Unix variant to its own file, rand_unix.c.
9052 For VMS, there's a currently-empty rand_vms.c.
9055 *) Move OCSP client related routines to ocsp_cl.c. These
9056 provide utility functions which an application needing
9057 to issue a request to an OCSP responder and analyse the
9058 response will typically need: as opposed to those which an
9059 OCSP responder itself would need which will be added later.
9061 OCSP_request_sign() signs an OCSP request with an API similar
9062 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
9063 response. OCSP_response_get1_basic() extracts basic response
9064 from response. OCSP_resp_find_status(): finds and extracts status
9065 information from an OCSP_CERTID structure (which will be created
9066 when the request structure is built). These are built from lower
9067 level functions which work on OCSP_SINGLERESP structures but
9068 won't normally be used unless the application wishes to examine
9069 extensions in the OCSP response for example.
9071 Replace nonce routines with a pair of functions.
9072 OCSP_request_add1_nonce() adds a nonce value and optionally
9073 generates a random value. OCSP_check_nonce() checks the
9074 validity of the nonce in an OCSP response.
9077 *) Change function OCSP_request_add() to OCSP_request_add0_id().
9078 This doesn't copy the supplied OCSP_CERTID and avoids the
9079 need to free up the newly created id. Change return type
9080 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
9081 This can then be used to add extensions to the request.
9082 Deleted OCSP_request_new(), since most of its functionality
9083 is now in OCSP_REQUEST_new() (and the case insensitive name
9084 clash) apart from the ability to set the request name which
9085 will be added elsewhere.
9088 *) Update OCSP API. Remove obsolete extensions argument from
9089 various functions. Extensions are now handled using the new
9090 OCSP extension code. New simple OCSP HTTP function which
9091 can be used to send requests and parse the response.
9094 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
9095 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
9096 uses the special reorder version of SET OF to sort the attributes
9097 and reorder them to match the encoded order. This resolves a long
9098 standing problem: a verify on a PKCS7 structure just after signing
9099 it used to fail because the attribute order did not match the
9100 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
9101 it uses the received order. This is necessary to tolerate some broken
9102 software that does not order SET OF. This is handled by encoding
9103 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
9104 to produce the required SET OF.
9107 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
9108 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
9109 files to get correct declarations of the ASN.1 item variables.
9112 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
9113 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
9114 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
9115 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
9116 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
9117 ASN1_ITEM and no wrapper functions.
9120 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
9121 replace the old function pointer based I/O routines. Change most of
9122 the *_d2i_bio() and *_d2i_fp() functions to use these.
9125 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
9126 lines, recognize more "algorithms" that can be deselected, and make
9127 it complain about algorithm deselection that isn't recognised.
9130 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
9131 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
9132 to use new functions. Add NO_ASN1_OLD which can be set to remove
9133 some old style ASN1 functions: this can be used to determine if old
9134 code will still work when these eventually go away.
9137 *) New extension functions for OCSP structures, these follow the
9138 same conventions as certificates and CRLs.
9141 *) New function X509V3_add1_i2d(). This automatically encodes and
9142 adds an extension. Its behaviour can be customised with various
9143 flags to append, replace or delete. Various wrappers added for
9144 certificates and CRLs.
9147 *) Fix to avoid calling the underlying ASN1 print routine when
9148 an extension cannot be parsed. Correct a typo in the
9149 OCSP_SERVICELOC extension. Tidy up print OCSP format.
9152 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
9153 entries for variables.
9156 *) Add functionality to apps/openssl.c for detecting locking
9157 problems: As the program is single-threaded, all we have
9158 to do is register a locking callback using an array for
9159 storing which locks are currently held by the program.
9162 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
9163 SSL_get_ex_data_X509_STORE_idx(), which is used in
9164 ssl_verify_cert_chain() and thus can be called at any time
9165 during TLS/SSL handshakes so that thread-safety is essential.
9166 Unfortunately, the ex_data design is not at all suited
9167 for multi-threaded use, so it probably should be abolished.
9170 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
9171 [Broadcom, tweaked and integrated by Geoff Thorpe]
9173 *) Move common extension printing code to new function
9174 X509V3_print_extensions(). Reorganise OCSP print routines and
9175 implement some needed OCSP ASN1 functions. Add OCSP extensions.
9178 *) New function X509_signature_print() to remove duplication in some
9182 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
9183 set (this was treated exactly the same as SET OF previously). This
9184 is used to reorder the STACK representing the structure to match the
9185 encoding. This will be used to get round a problem where a PKCS7
9186 structure which was signed could not be verified because the STACK
9187 order did not reflect the encoded order.
9190 *) Reimplement the OCSP ASN1 module using the new code.
9193 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
9194 for its ASN1 operations. The old style function pointers still exist
9195 for now but they will eventually go away.
9198 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
9199 completely replaces the old ASN1 functionality with a table driven
9200 encoder and decoder which interprets an ASN1_ITEM structure describing
9201 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
9202 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
9203 has also been converted to the new form.
9206 *) Change BN_mod_exp_recp so that negative moduli are tolerated
9207 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
9208 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
9209 for negative moduli.
9212 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
9213 of not touching the result's sign bit.
9216 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
9220 *) Changed the LHASH code to use prototypes for callbacks, and created
9221 macros to declare and implement thin (optionally static) functions
9222 that provide type-safety and avoid function pointer casting for the
9223 type-specific callbacks.
9226 *) Added Kerberos Cipher Suites to be used with TLS, as written in
9228 [Veers Staats <staatsvr@asc.hpc.mil>,
9229 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
9231 *) Reformat the FAQ so the different questions and answers can be divided
9232 in sections depending on the subject.
9235 *) Have the zlib compression code load ZLIB.DLL dynamically under
9239 *) New function BN_mod_sqrt for computing square roots modulo a prime
9240 (using the probabilistic Tonelli-Shanks algorithm unless
9241 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
9242 be handled deterministically).
9243 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9245 *) Make BN_mod_inverse faster by explicitly handling small quotients
9246 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
9247 512 bits], about 30% for larger ones [1024 or 2048 bits].)
9250 *) New function BN_kronecker.
9253 *) Fix BN_gcd so that it works on negative inputs; the result is
9254 positive unless both parameters are zero.
9255 Previously something reasonably close to an infinite loop was
9256 possible because numbers could be growing instead of shrinking
9257 in the implementation of Euclid's algorithm.
9260 *) Fix BN_is_word() and BN_is_one() macros to take into account the
9261 sign of the number in question.
9263 Fix BN_is_word(a,w) to work correctly for w == 0.
9265 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
9266 because its test if the absolute value of 'a' equals 'w'.
9267 Note that BN_abs_is_word does *not* handle w == 0 reliably;
9268 it exists mostly for use in the implementations of BN_is_zero(),
9269 BN_is_one(), and BN_is_word().
9272 *) New function BN_swap.
9275 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
9276 the exponentiation functions are more likely to produce reasonable
9277 results on negative inputs.
9280 *) Change BN_mod_mul so that the result is always non-negative.
9281 Previously, it could be negative if one of the factors was negative;
9282 I don't think anyone really wanted that behaviour.
9285 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
9286 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
9287 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
9288 and add new functions:
9297 BN_mod_lshift1_quick
9301 These functions always generate non-negative results.
9303 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
9304 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
9306 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
9307 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
9308 be reduced modulo m.
9309 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9312 The following entry accidentally appeared in the CHANGES file
9313 distributed with OpenSSL 0.9.7. The modifications described in
9314 it do *not* apply to OpenSSL 0.9.7.
9316 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
9317 was actually never needed) and in BN_mul(). The removal in BN_mul()
9318 required a small change in bn_mul_part_recursive() and the addition
9319 of the functions bn_cmp_part_words(), bn_sub_part_words() and
9320 bn_add_part_words(), which do the same thing as bn_cmp_words(),
9321 bn_sub_words() and bn_add_words() except they take arrays with
9326 *) In 'openssl passwd', verify passwords read from the terminal
9327 unless the '-salt' option is used (which usually means that
9328 verification would just waste user's time since the resulting
9329 hash is going to be compared with some given password hash)
9330 or the new '-noverify' option is used.
9332 This is an incompatible change, but it does not affect
9333 non-interactive use of 'openssl passwd' (passwords on the command
9334 line, '-stdin' option, '-in ...' option) and thus should not
9338 *) Remove all references to RSAref, since there's no more need for it.
9341 *) Make DSO load along a path given through an environment variable
9342 (SHLIB_PATH) with shl_load().
9345 *) Constify the ENGINE code as a result of BIGNUM constification.
9346 Also constify the RSA code and most things related to it. In a
9347 few places, most notable in the depth of the ASN.1 code, ugly
9348 casts back to non-const were required (to be solved at a later
9352 *) Make it so the openssl application has all engines loaded by default.
9355 *) Constify the BIGNUM routines a little more.
9358 *) Add the following functions:
9360 ENGINE_load_cswift()
9362 ENGINE_load_atalla()
9364 ENGINE_load_builtin_engines()
9366 That way, an application can itself choose if external engines that
9367 are built-in in OpenSSL shall ever be used or not. The benefit is
9368 that applications won't have to be linked with libdl or other dso
9369 libraries unless it's really needed.
9371 Changed 'openssl engine' to load all engines on demand.
9372 Changed the engine header files to avoid the duplication of some
9373 declarations (they differed!).
9376 *) 'openssl engine' can now list capabilities.
9379 *) Better error reporting in 'openssl engine'.
9382 *) Never call load_dh_param(NULL) in s_server.
9385 *) Add engine application. It can currently list engines by name and
9386 identity, and test if they are actually available.
9389 *) Improve RPM specification file by forcing symbolic linking and making
9390 sure the installed documentation is also owned by root.root.
9391 [Damien Miller <djm@mindrot.org>]
9393 *) Give the OpenSSL applications more possibilities to make use of
9394 keys (public as well as private) handled by engines.
9397 *) Add OCSP code that comes from CertCo.
9400 *) Add VMS support for the Rijndael code.
9403 *) Added untested support for Nuron crypto accelerator.
9406 *) Add support for external cryptographic devices. This code was
9407 previously distributed separately as the "engine" branch.
9408 [Geoff Thorpe, Richard Levitte]
9410 *) Rework the filename-translation in the DSO code. It is now possible to
9411 have far greater control over how a "name" is turned into a filename
9412 depending on the operating environment and any oddities about the
9413 different shared library filenames on each system.
9416 *) Support threads on FreeBSD-elf in Configure.
9419 *) Fix for SHA1 assembly problem with MASM: it produces
9420 warnings about corrupt line number information when assembling
9421 with debugging information. This is caused by the overlapping
9423 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
9426 NCONF_get_number() has no error checking at all. As a replacement,
9427 NCONF_get_number_e() is defined (_e for "error checking") and is
9428 promoted strongly. The old NCONF_get_number is kept around for
9429 binary backward compatibility.
9430 Make it possible for methods to load from something other than a BIO,
9431 by providing a function pointer that is given a name instead of a BIO.
9432 For example, this could be used to load configuration data from an
9436 *) Fix for non blocking accept BIOs. Added new I/O special reason
9437 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
9438 with non blocking I/O was not possible because no retry code was
9439 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
9443 *) Added the beginnings of Rijndael support.
9446 *) Fix for bug in DirectoryString mask setting. Add support for
9447 X509_NAME_print_ex() in 'req' and X509_print_ex() function
9448 to allow certificate printing to more controllable, additional
9449 'certopt' option to 'x509' to allow new printing options to be
9453 *) Clean old EAY MD5 hack from e_os.h.
9456 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
9458 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
9459 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
9460 [Joe Orton, Steve Henson]
9462 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
9464 *) Fix additional bug revealed by the NISCC test suite:
9466 Stop bug triggering large recursion when presented with
9467 certain ASN.1 tags (CVE-2003-0851)
9470 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
9472 *) Fix various bugs revealed by running the NISCC test suite:
9474 Stop out of bounds reads in the ASN1 code when presented with
9475 invalid tags (CVE-2003-0543 and CVE-2003-0544).
9477 If verify callback ignores invalid public key errors don't try to check
9478 certificate signature with the NULL public key.
9482 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
9483 if the server requested one: as stated in TLS 1.0 and SSL 3.0
9487 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
9488 extra data after the compression methods not only for TLS 1.0
9489 but also for SSL 3.0 (as required by the specification).
9490 [Bodo Moeller; problem pointed out by Matthias Loepfe]
9492 *) Change X509_certificate_type() to mark the key as exported/exportable
9493 when it's 512 *bits* long, not 512 bytes.
9496 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
9498 *) Countermeasure against the Klima-Pokorny-Rosa extension of
9499 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
9500 a protocol version number mismatch like a decryption error
9501 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
9504 *) Turn on RSA blinding by default in the default implementation
9505 to avoid a timing attack. Applications that don't want it can call
9506 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
9507 They would be ill-advised to do so in most cases.
9508 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
9510 *) Change RSA blinding code so that it works when the PRNG is not
9511 seeded (in this case, the secret RSA exponent is abused as
9512 an unpredictable seed -- if it is not unpredictable, there
9513 is no point in blinding anyway). Make RSA blinding thread-safe
9514 by remembering the creator's thread ID in rsa->blinding and
9515 having all other threads use local one-time blinding factors
9516 (this requires more computation than sharing rsa->blinding, but
9517 avoids excessive locking; and if an RSA object is not shared
9518 between threads, blinding will still be very fast).
9521 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
9523 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
9524 via timing by performing a MAC computation even if incorrect
9525 block cipher padding has been found. This is a countermeasure
9526 against active attacks where the attacker has to distinguish
9527 between bad padding and a MAC verification error. (CVE-2003-0078)
9529 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
9530 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
9531 Martin Vuagnoux (EPFL, Ilion)]
9533 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
9535 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
9536 memory from its contents. This is done with a counter that will
9537 place alternating values in each byte. This can be used to solve
9538 two issues: 1) the removal of calls to memset() by highly optimizing
9539 compilers, and 2) cleansing with other values than 0, since those can
9540 be read through on certain media, for example a swap space on disk.
9543 *) Bugfix: client side session caching did not work with external caching,
9544 because the session->cipher setting was not restored when reloading
9545 from the external cache. This problem was masked, when
9546 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9547 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9550 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9551 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9552 [Zeev Lieber <zeev-l@yahoo.com>]
9554 *) Undo an undocumented change introduced in 0.9.6e which caused
9555 repeated calls to OpenSSL_add_all_ciphers() and
9556 OpenSSL_add_all_digests() to be ignored, even after calling
9560 *) Change the default configuration reader to deal with last line not
9561 being properly terminated.
9564 *) Change X509_NAME_cmp() so it applies the special rules on handling
9565 DN values that are of type PrintableString, as well as RDNs of type
9566 emailAddress where the value has the type ia5String.
9567 [stefank@valicert.com via Richard Levitte]
9569 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9570 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9571 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9572 the bitwise-OR of the two for use by the majority of applications
9573 wanting this behaviour, and update the docs. The documented
9574 behaviour and actual behaviour were inconsistent and had been
9575 changing anyway, so this is more a bug-fix than a behavioural
9577 [Geoff Thorpe, diagnosed by Nadav Har'El]
9579 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9580 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9583 *) Fix initialization code race conditions in
9584 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9585 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9586 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9587 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9588 ssl2_get_cipher_by_char(),
9589 ssl3_get_cipher_by_char().
9590 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9592 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9593 the cached sessions are flushed, as the remove_cb() might use ex_data
9594 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9595 (see [openssl.org #212]).
9596 [Geoff Thorpe, Lutz Jaenicke]
9598 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9599 length, instead of the encoding length to d2i_ASN1_OBJECT.
9602 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9604 *) [In 0.9.6g-engine release:]
9605 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9606 [Lynn Gazis <lgazis@rainbow.com>]
9608 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9610 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9611 and get fix the header length calculation.
9612 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9613 Alon Kantor <alonk@checkpoint.com> (and others),
9616 *) Use proper error handling instead of 'assertions' in buffer
9617 overflow checks added in 0.9.6e. This prevents DoS (the
9618 assertions could call abort()).
9619 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9621 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9623 *) Add various sanity checks to asn1_get_length() to reject
9624 the ASN1 length bytes if they exceed sizeof(long), will appear
9625 negative or the content length exceeds the length of the
9627 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9629 *) Fix cipher selection routines: ciphers without encryption had no flags
9630 for the cipher strength set and where therefore not handled correctly
9631 by the selection routines (PR #130).
9634 *) Fix EVP_dsa_sha macro.
9638 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9639 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9640 that was added in OpenSSL 0.9.6d.
9642 As the countermeasure turned out to be incompatible with some
9643 broken SSL implementations, the new option is part of SSL_OP_ALL.
9644 SSL_OP_ALL is usually employed when compatibility with weird SSL
9645 implementations is desired (e.g. '-bugs' option to 's_client' and
9646 's_server'), so the new option is automatically set in many
9650 *) Changes in security patch:
9652 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9653 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9654 Air Force Materiel Command, USAF, under agreement number
9657 *) Add various sanity checks to asn1_get_length() to reject
9658 the ASN1 length bytes if they exceed sizeof(long), will appear
9659 negative or the content length exceeds the length of the
9660 supplied buffer. (CVE-2002-0659)
9661 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9663 *) Assertions for various potential buffer overflows, not known to
9665 [Ben Laurie (CHATS)]
9667 *) Various temporary buffers to hold ASCII versions of integers were
9668 too small for 64 bit platforms. (CVE-2002-0655)
9669 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9671 *) Remote buffer overflow in SSL3 protocol - an attacker could
9672 supply an oversized session ID to a client. (CVE-2002-0656)
9673 [Ben Laurie (CHATS)]
9675 *) Remote buffer overflow in SSL2 protocol - an attacker could
9676 supply an oversized client master key. (CVE-2002-0656)
9677 [Ben Laurie (CHATS)]
9679 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9681 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9682 encoded as NULL) with id-dsa-with-sha1.
9683 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9685 *) Check various X509_...() return values in apps/req.c.
9686 [Nils Larsch <nla@trustcenter.de>]
9688 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9689 an end-of-file condition would erroneously be flagged, when the CRLF
9690 was just at the end of a processed block. The bug was discovered when
9691 processing data through a buffering memory BIO handing the data to a
9692 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9693 <ptsekov@syntrex.com> and Nedelcho Stanev.
9696 *) Implement a countermeasure against a vulnerability recently found
9697 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9698 before application data chunks to avoid the use of known IVs
9699 with data potentially chosen by the attacker.
9702 *) Fix length checks in ssl3_get_client_hello().
9705 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9706 to prevent ssl3_read_internal() from incorrectly assuming that
9707 ssl3_read_bytes() found application data while handshake
9708 processing was enabled when in fact s->s3->in_read_app_data was
9709 merely automatically cleared during the initial handshake.
9710 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9712 *) Fix object definitions for Private and Enterprise: they were not
9713 recognized in their shortname (=lowercase) representation. Extend
9714 obj_dat.pl to issue an error when using undefined keywords instead
9715 of silently ignoring the problem (Svenning Sorensen
9716 <sss@sss.dnsalias.net>).
9719 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9720 generators, i.e. generators other than 2 and 5. (Previously, the
9721 code did not properly initialise the 'add' and 'rem' values to
9722 BN_generate_prime().)
9724 In the new general case, we do not insist that 'generator' is
9725 actually a primitive root: This requirement is rather pointless;
9726 a generator of the order-q subgroup is just as good, if not
9730 *) Map new X509 verification errors to alerts. Discovered and submitted by
9731 Tom Wu <tom@arcot.com>.
9734 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9735 returning non-zero before the data has been completely received
9736 when using non-blocking I/O.
9737 [Bodo Moeller; problem pointed out by John Hughes]
9739 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9740 [Ben Laurie, Lutz Jaenicke]
9742 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9743 Yoram Zahavi <YoramZ@gilian.com>).
9746 *) Add information about CygWin 1.3 and on, and preserve proper
9747 configuration for the versions before that.
9748 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9750 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9751 check whether we deal with a copy of a session and do not delete from
9752 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9753 <izhar@checkpoint.com>.
9756 *) Do not store session data into the internal session cache, if it
9757 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9758 flag is set). Proposed by Aslam <aslam@funk.com>.
9761 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9765 *) [In 0.9.6d-engine release:]
9766 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9767 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9769 *) Add the configuration target linux-s390x.
9770 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9772 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9773 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9774 variable as an indication that a ClientHello message has been
9775 received. As the flag value will be lost between multiple
9776 invocations of ssl3_accept when using non-blocking I/O, the
9777 function may not be aware that a handshake has actually taken
9778 place, thus preventing a new session from being added to the
9781 To avoid this problem, we now set s->new_session to 2 instead of
9782 using a local variable.
9783 [Lutz Jaenicke, Bodo Moeller]
9785 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9786 if the SSL_R_LENGTH_MISMATCH error is detected.
9787 [Geoff Thorpe, Bodo Moeller]
9789 *) New 'shared_ldflag' column in Configure platform table.
9792 *) Fix EVP_CIPHER_mode macro.
9793 ["Dan S. Camper" <dan@bti.net>]
9795 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9796 type, we must throw them away by setting rr->length to 0.
9797 [D P Chang <dpc@qualys.com>]
9799 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9801 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9802 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9803 worked incorrectly for those cases where range = 10..._2 and
9804 3*range is two bits longer than range.)
9807 *) Only add signing time to PKCS7 structures if it is not already
9811 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9812 OBJ_ld_ce should be OBJ_id_ce.
9813 Also some ip-pda OIDs in crypto/objects/objects.txt were
9814 incorrect (cf. RFC 3039).
9815 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9817 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9818 returns early because it has nothing to do.
9819 [Andy Schneider <andy.schneider@bjss.co.uk>]
9821 *) [In 0.9.6c-engine release:]
9822 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9823 [Andy Schneider <andy.schneider@bjss.co.uk>]
9825 *) [In 0.9.6c-engine release:]
9826 Add support for Cryptographic Appliance's keyserver technology.
9827 (Use engine 'keyclient')
9828 [Cryptographic Appliances and Geoff Thorpe]
9830 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9831 is called via tools/c89.sh because arguments have to be
9832 rearranged (all '-L' options must appear before the first object
9834 [Richard Shapiro <rshapiro@abinitio.com>]
9836 *) [In 0.9.6c-engine release:]
9837 Add support for Broadcom crypto accelerator cards, backported
9839 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9841 *) [In 0.9.6c-engine release:]
9842 Add support for SureWare crypto accelerator cards from
9843 Baltimore Technologies. (Use engine 'sureware')
9844 [Baltimore Technologies and Mark Cox]
9846 *) [In 0.9.6c-engine release:]
9847 Add support for crypto accelerator cards from Accelerated
9848 Encryption Processing, www.aep.ie. (Use engine 'aep')
9849 [AEP Inc. and Mark Cox]
9851 *) Add a configuration entry for gcc on UnixWare.
9852 [Gary Benson <gbenson@redhat.com>]
9854 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9855 messages are stored in a single piece (fixed-length part and
9856 variable-length part combined) and fix various bugs found on the way.
9859 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9860 instead. BIO_gethostbyname() does not know what timeouts are
9861 appropriate, so entries would stay in cache even when they have
9863 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9865 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9866 faced with a pathologically small ClientHello fragment that does
9867 not contain client_version: Instead of aborting with an error,
9868 simply choose the highest available protocol version (i.e.,
9869 TLS 1.0 unless it is disabled). In practice, ClientHello
9870 messages are never sent like this, but this change gives us
9871 strictly correct behaviour at least for TLS.
9874 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9875 never resets s->method to s->ctx->method when called from within
9876 one of the SSL handshake functions.
9877 [Bodo Moeller; problem pointed out by Niko Baric]
9879 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9880 (sent using the client's version number) if client_version is
9881 smaller than the protocol version in use. Also change
9882 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9883 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9884 the client will at least see that alert.
9887 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9891 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9892 client receives HelloRequest while in a handshake.
9893 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9895 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9896 should end in 'break', not 'goto end' which circumvents various
9897 cleanups done in state SSL_ST_OK. But session related stuff
9898 must be disabled for SSL_ST_OK in the case that we just sent a
9901 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9902 before just sending a HelloRequest.
9903 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9905 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9906 reveal whether illegal block cipher padding was found or a MAC
9907 verification error occurred. (Neither SSLerr() codes nor alerts
9908 are directly visible to potential attackers, but the information
9909 may leak via logfiles.)
9911 Similar changes are not required for the SSL 2.0 implementation
9912 because the number of padding bytes is sent in clear for SSL 2.0,
9913 and the extra bytes are just ignored. However ssl/s2_pkt.c
9914 failed to verify that the purported number of padding bytes is in
9918 *) Add OpenUNIX-8 support including shared libraries
9919 (Boyd Lynn Gerber <gerberb@zenez.com>).
9922 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9923 'wristwatch attack' using huge encoding parameters (cf.
9924 James H. Manger's CRYPTO 2001 paper). Note that the
9925 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9926 encoding parameters and hence was not vulnerable.
9929 *) BN_sqr() bug fix.
9930 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9932 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9933 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9934 followed by modular reduction.
9935 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9937 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9938 equivalent based on BN_pseudo_rand() instead of BN_rand().
9941 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9942 This function was broken, as the check for a new client hello message
9943 to handle SGC did not allow these large messages.
9944 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9947 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9950 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9951 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9954 *) Rework the configuration and shared library support for Tru64 Unix.
9955 The configuration part makes use of modern compiler features and
9956 still retains old compiler behavior for those that run older versions
9957 of the OS. The shared library support part includes a variant that
9958 uses the RPATH feature, and is available through the special
9959 configuration target "alpha-cc-rpath", which will never be selected
9961 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9963 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9964 with the same message size as in ssl3_get_certificate_request().
9965 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9966 messages might inadvertently be reject as too long.
9967 [Petr Lampa <lampa@fee.vutbr.cz>]
9969 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9972 *) Modified SSL library such that the verify_callback that has been set
9973 specifically for an SSL object with SSL_set_verify() is actually being
9974 used. Before the change, a verify_callback set with this function was
9975 ignored and the verify_callback() set in the SSL_CTX at the time of
9976 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9977 to allow the necessary settings.
9980 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9981 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9982 done automatically (in contradiction to the requirements of the C
9983 standard). This made problems when used from OpenSSH.
9986 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9987 dh->length and always used
9989 BN_rand_range(priv_key, dh->p).
9991 BN_rand_range() is not necessary for Diffie-Hellman, and this
9992 specific range makes Diffie-Hellman unnecessarily inefficient if
9993 dh->length (recommended exponent length) is much smaller than the
9994 length of dh->p. We could use BN_rand_range() if the order of
9995 the subgroup was stored in the DH structure, but we only have
10000 BN_rand(priv_key, l, ...)
10002 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
10008 RSA_eay_public_encrypt
10009 RSA_eay_private_decrypt
10010 RSA_eay_private_encrypt (signing)
10011 RSA_eay_public_decrypt (signature verification)
10013 (default implementations for RSA_public_encrypt,
10014 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
10015 always reject numbers >= n.
10018 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
10019 to synchronize access to 'locking_thread'. This is necessary on
10020 systems where access to 'locking_thread' (an 'unsigned long'
10021 variable) is not atomic.
10024 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
10025 *before* setting the 'crypto_lock_rand' flag. The previous code had
10026 a race condition if 0 is a valid thread ID.
10027 [Travis Vitek <vitek@roguewave.com>]
10029 *) Add support for shared libraries under Irix.
10030 [Albert Chin-A-Young <china@thewrittenword.com>]
10032 *) Add configuration option to build on Linux on both big-endian and
10033 little-endian MIPS.
10034 [Ralf Baechle <ralf@uni-koblenz.de>]
10036 *) Add the possibility to create shared libraries on HP-UX.
10039 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
10041 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
10042 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
10043 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
10044 PRNG state recovery was possible based on the output of
10045 one PRNG request appropriately sized to gain knowledge on
10046 'md' followed by enough consecutive 1-byte PRNG requests
10047 to traverse all of 'state'.
10049 1. When updating 'md_local' (the current thread's copy of 'md')
10050 during PRNG output generation, hash all of the previous
10051 'md_local' value, not just the half used for PRNG output.
10053 2. Make the number of bytes from 'state' included into the hash
10054 independent from the number of PRNG bytes requested.
10056 The first measure alone would be sufficient to avoid
10057 Markku-Juhani's attack. (Actually it had never occurred
10058 to me that the half of 'md_local' used for chaining was the
10059 half from which PRNG output bytes were taken -- I had always
10060 assumed that the secret half would be used.) The second
10061 measure makes sure that additional data from 'state' is never
10062 mixed into 'md_local' in small portions; this heuristically
10063 further strengthens the PRNG.
10066 *) Fix crypto/bn/asm/mips3.s.
10069 *) When only the key is given to "enc", the IV is undefined. Print out
10070 an error message in this case.
10073 *) Handle special case when X509_NAME is empty in X509 printing routines.
10076 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
10077 positive and less than q.
10080 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
10081 used: it isn't thread safe and the add_lock_callback should handle
10083 [Paul Rose <Paul.Rose@bridge.com>]
10085 *) Verify that incoming data obeys the block size in
10086 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
10090 [Ulf Möller, Bodo Möller]
10092 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
10093 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
10094 when fixing the server behaviour for backwards-compatible 'client
10095 hello' messages. (Note that the attack is impractical against
10096 SSL 3.0 and TLS 1.0 anyway because length and version checking
10097 means that the probability of guessing a valid ciphertext is
10098 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
10101 Before 0.9.5, the countermeasure (hide the error by generating a
10102 random 'decryption result') did not work properly because
10103 ERR_clear_error() was missing, meaning that SSL_get_error() would
10104 detect the supposedly ignored error.
10106 Both problems are now fixed.
10109 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
10110 (previously it was 1024).
10113 *) Fix for compatibility mode trust settings: ignore trust settings
10114 unless some valid trust or reject settings are present.
10117 *) Fix for blowfish EVP: its a variable length cipher.
10120 *) Fix various bugs related to DSA S/MIME verification. Handle missing
10121 parameters in DSA public key structures and return an error in the
10122 DSA routines if parameters are absent.
10125 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
10126 in the current directory if neither $RANDFILE nor $HOME was set.
10127 RAND_file_name() in 0.9.6a returned NULL in this case. This has
10128 caused some confusion to Windows users who haven't defined $HOME.
10129 Thus RAND_file_name() is changed again: e_os.h can define a
10130 DEFAULT_HOME, which will be used if $HOME is not set.
10131 For Windows, we use "C:"; on other platforms, we still require
10132 environment variables.
10134 *) Move 'if (!initialized) RAND_poll()' into regions protected by
10135 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
10136 having multiple threads call RAND_poll() concurrently.
10139 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
10140 combination of a flag and a thread ID variable.
10141 Otherwise while one thread is in ssleay_rand_bytes (which sets the
10142 flag), *other* threads can enter ssleay_add_bytes without obeying
10143 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
10144 that they do not hold after the first thread unsets add_do_not_lock).
10147 *) Change bctest again: '-x' expressions are not available in all
10148 versions of 'test'.
10151 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
10153 *) Fix a couple of memory leaks in PKCS7_dataDecode()
10154 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
10156 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
10157 the default extension for executables, if any. Also, make the perl
10158 scripts that use symlink() to test if it really exists and use "cp"
10159 if it doesn't. All this made OpenSSL compilable and installable in
10163 *) Fix for asn1_GetSequence() for indefinite length constructed data.
10164 If SEQUENCE is length is indefinite just set c->slen to the total
10165 amount of data available.
10166 [Steve Henson, reported by shige@FreeBSD.org]
10167 [This change does not apply to 0.9.7.]
10169 *) Change bctest to avoid here-documents inside command substitution
10170 (workaround for FreeBSD /bin/sh bug).
10171 For compatibility with Ultrix, avoid shell functions (introduced
10172 in the bctest version that searches along $PATH).
10175 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
10176 with des_encrypt() defined on some operating systems, like Solaris
10180 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
10181 On the Importance of Eliminating Errors in Cryptographic
10182 Computations, J. Cryptology 14 (2001) 2, 101-119,
10183 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
10186 *) MIPS assembler BIGNUM division bug fix.
10189 *) Disabled incorrect Alpha assembler code.
10192 *) Fix PKCS#7 decode routines so they correctly update the length
10193 after reading an EOC for the EXPLICIT tag.
10195 [This change does not apply to 0.9.7.]
10197 *) Fix bug in PKCS#12 key generation routines. This was triggered
10198 if a 3DES key was generated with a 0 initial byte. Include
10199 PKCS12_BROKEN_KEYGEN compilation option to retain the old
10200 (but broken) behaviour.
10203 *) Enhance bctest to search for a working bc along $PATH and print
10205 [Tim Rice <tim@multitalents.net> via Richard Levitte]
10207 *) Fix memory leaks in err.c: free err_data string if necessary;
10208 don't write to the wrong index in ERR_set_error_data.
10211 *) Implement ssl23_peek (analogous to ssl23_read), which previously
10215 *) Replace rdtsc with _emit statements for VC++ version 5.
10216 [Jeremy Cooper <jeremy@baymoo.org>]
10218 *) Make it possible to reuse SSLv2 sessions.
10221 *) In copy_email() check for >= 0 as a return value for
10222 X509_NAME_get_index_by_NID() since 0 is a valid index.
10223 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
10225 *) Avoid coredump with unsupported or invalid public keys by checking if
10226 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
10227 PKCS7_verify() fails with non detached data.
10230 *) Don't use getenv in library functions when run as setuid/setgid.
10231 New function OPENSSL_issetugid().
10234 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
10235 due to incorrect handling of multi-threading:
10237 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
10239 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
10241 3. Count how many times MemCheck_off() has been called so that
10242 nested use can be treated correctly. This also avoids
10243 inband-signalling in the previous code (which relied on the
10244 assumption that thread ID 0 is impossible).
10247 *) Add "-rand" option also to s_client and s_server.
10250 *) Fix CPU detection on Irix 6.x.
10251 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
10252 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10254 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
10257 [This change does not apply to 0.9.7.]
10259 *) Use the cached encoding of an X509_NAME structure rather than
10260 copying it. This is apparently the reason for the libsafe "errors"
10261 but the code is actually correct.
10264 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
10265 Bleichenbacher's DSA attack.
10266 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
10267 to be set and top=0 forces the highest bit to be set; top=-1 is new
10268 and leaves the highest bit random.
10269 [Ulf Moeller, Bodo Moeller]
10271 *) In the NCONF_...-based implementations for CONF_... queries
10272 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
10273 a temporary CONF structure with the data component set to NULL
10274 (which gives segmentation faults in lh_retrieve).
10275 Instead, use NULL for the CONF pointer in CONF_get_string and
10276 CONF_get_number (which may use environment variables) and directly
10277 return NULL from CONF_get_section.
10280 *) Fix potential buffer overrun for EBCDIC.
10283 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
10284 keyUsage if basicConstraints absent for a CA.
10287 *) Make SMIME_write_PKCS7() write mail header values with a format that
10288 is more generally accepted (no spaces before the semicolon), since
10289 some programs can't parse those values properly otherwise. Also make
10290 sure BIO's that break lines after each write do not create invalid
10294 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
10295 macros previously used would not encode an empty SEQUENCE OF
10296 and break the signature.
10298 [This change does not apply to 0.9.7.]
10300 *) Zero the premaster secret after deriving the master secret in
10304 *) Add some EVP_add_digest_alias registrations (as found in
10305 OpenSSL_add_all_digests()) to SSL_library_init()
10306 aka OpenSSL_add_ssl_algorithms(). This provides improved
10307 compatibility with peers using X.509 certificates
10308 with unconventional AlgorithmIdentifier OIDs.
10311 *) Fix for Irix with NO_ASM.
10312 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10314 *) ./config script fixes.
10315 [Ulf Moeller, Richard Levitte]
10317 *) Fix 'openssl passwd -1'.
10320 *) Change PKCS12_key_gen_asc() so it can cope with non null
10321 terminated strings whose length is passed in the passlen
10322 parameter, for example from PEM callbacks. This was done
10323 by adding an extra length parameter to asc2uni().
10324 [Steve Henson, reported by <oddissey@samsung.co.kr>]
10326 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
10327 call failed, free the DSA structure.
10330 *) Fix to uni2asc() to cope with zero length Unicode strings.
10331 These are present in some PKCS#12 files.
10334 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
10335 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
10336 when writing a 32767 byte record.
10337 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
10339 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
10340 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
10342 (RSA objects have a reference count access to which is protected
10343 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
10344 so they are meant to be shared between threads.)
10345 [Bodo Moeller, Geoff Thorpe; original patch submitted by
10346 "Reddie, Steven" <Steven.Reddie@ca.com>]
10348 *) Fix a deadlock in CRYPTO_mem_leaks().
10351 *) Use better test patterns in bntest.
10354 *) rand_win.c fix for Borland C.
10357 *) BN_rshift bugfix for n == 0.
10360 *) Add a 'bctest' script that checks for some known 'bc' bugs
10361 so that 'make test' does not abort just because 'bc' is broken.
10364 *) Store verify_result within SSL_SESSION also for client side to
10365 avoid potential security hole. (Re-used sessions on the client side
10366 always resulted in verify_result==X509_V_OK, not using the original
10367 result of the server certificate verification.)
10370 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
10371 SSL3_RT_APPLICATION_DATA, return 0.
10372 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
10376 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
10377 releases, have been re-implemented by renaming the previous
10378 implementations of ssl2_read and ssl3_read to ssl2_read_internal
10379 and ssl3_read_internal, respectively, and adding 'peek' parameters
10380 to them. The new ssl[23]_{read,peek} functions are calls to
10381 ssl[23]_read_internal with the 'peek' flag set appropriately.
10382 A 'peek' parameter has also been added to ssl3_read_bytes, which
10383 does the actual work for ssl3_read_internal.
10386 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
10387 the method-specific "init()" handler. Also clean up ex_data after
10388 calling the method-specific "finish()" handler. Previously, this was
10389 happening the other way round.
10392 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
10393 The previous value, 12, was not always sufficient for BN_mod_exp().
10396 *) Make sure that shared libraries get the internal name engine with
10397 the full version number and not just 0. This should mark the
10398 shared libraries as not backward compatible. Of course, this should
10399 be changed again when we can guarantee backward binary compatibility.
10402 *) Fix typo in get_cert_by_subject() in by_dir.c
10403 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
10405 *) Rework the system to generate shared libraries:
10407 - Make note of the expected extension for the shared libraries and
10408 if there is a need for symbolic links from for example libcrypto.so.0
10409 to libcrypto.so.0.9.7. There is extended info in Configure for
10412 - Make as few rebuilds of the shared libraries as possible.
10414 - Still avoid linking the OpenSSL programs with the shared libraries.
10416 - When installing, install the shared libraries separately from the
10420 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
10422 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
10423 and not in SSL_clear because the latter is also used by the
10424 accept/connect functions; previously, the settings made by
10425 SSL_set_read_ahead would be lost during the handshake.
10426 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
10428 *) Correct util/mkdef.pl to be selective about disabled algorithms.
10429 Previously, it would create entries for disabled algorithms no
10433 *) Added several new manual pages for SSL_* function.
10436 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
10438 *) In ssl23_get_client_hello, generate an error message when faced
10439 with an initial SSL 3.0/TLS record that is too small to contain the
10440 first two bytes of the ClientHello message, i.e. client_version.
10441 (Note that this is a pathologic case that probably has never happened
10442 in real life.) The previous approach was to use the version number
10443 from the record header as a substitute; but our protocol choice
10444 should not depend on that one because it is not authenticated
10445 by the Finished messages.
10448 *) More robust randomness gathering functions for Windows.
10449 [Jeffrey Altman <jaltman@columbia.edu>]
10451 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
10452 not set then we don't setup the error code for issuer check errors
10453 to avoid possibly overwriting other errors which the callback does
10454 handle. If an application does set the flag then we assume it knows
10455 what it is doing and can handle the new informational codes
10459 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
10460 a general "ANY" type, as such it should be able to decode anything
10461 including tagged types. However it didn't check the class so it would
10462 wrongly interpret tagged types in the same way as their universal
10463 counterpart and unknown types were just rejected. Changed so that the
10464 tagged and unknown types are handled in the same way as a SEQUENCE:
10465 that is the encoding is stored intact. There is also a new type
10466 "V_ASN1_OTHER" which is used when the class is not universal, in this
10467 case we have no idea what the actual type is so we just lump them all
10471 *) On VMS, stdout may very well lead to a file that is written to
10472 in a record-oriented fashion. That means that every write() will
10473 write a separate record, which will be read separately by the
10474 programs trying to read from it. This can be very confusing.
10476 The solution is to put a BIO filter in the way that will buffer
10477 text until a linefeed is reached, and then write everything a
10478 line at a time, so every record written will be an actual line,
10479 not chunks of lines and not (usually doesn't happen, but I've
10480 seen it once) several lines in one record. BIO_f_linebuffer() is
10483 Currently, it's a VMS-only method, because that's where it has
10484 been tested well enough.
10487 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
10488 it can return incorrect results.
10489 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
10490 but it was in 0.9.6-beta[12].)
10493 *) Disable the check for content being present when verifying detached
10494 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
10495 include zero length content when signing messages.
10498 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
10499 BIO_ctrl (for BIO pairs).
10502 *) Add DSO method for VMS.
10505 *) Bug fix: Montgomery multiplication could produce results with the
10509 *) Add RPM specification openssl.spec and modify it to build three
10510 packages. The default package contains applications, application
10511 documentation and run-time libraries. The devel package contains
10512 include files, static libraries and function documentation. The
10513 doc package contains the contents of the doc directory. The original
10514 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
10517 *) Add a large number of documentation files for many SSL routines.
10518 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10520 *) Add a configuration entry for Sony News 4.
10521 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
10523 *) Don't set the two most significant bits to one when generating a
10524 random number < q in the DSA library.
10527 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
10528 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
10529 the underlying transport is blocking) if a handshake took place.
10530 (The default behaviour is needed by applications such as s_client
10531 and s_server that use select() to determine when to use SSL_read;
10532 but for applications that know in advance when to expect data, it
10533 just makes things more complicated.)
10536 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
10540 *) Add a few more EBCDIC conditionals that make `req' and `x509'
10541 work better on such systems.
10542 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10544 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10545 Update PKCS12_parse() so it copies the friendlyName and the
10546 keyid to the certificates aux info.
10549 *) Fix bug in PKCS7_verify() which caused an infinite loop
10550 if there was more than one signature.
10551 [Sven Uszpelkat <su@celocom.de>]
10553 *) Major change in util/mkdef.pl to include extra information
10554 about each symbol, as well as presenting variables as well
10555 as functions. This change means that there's n more need
10556 to rebuild the .num files when some algorithms are excluded.
10559 *) Allow the verify time to be set by an application,
10560 rather than always using the current time.
10563 *) Phase 2 verify code reorganisation. The certificate
10564 verify code now looks up an issuer certificate by a
10565 number of criteria: subject name, authority key id
10566 and key usage. It also verifies self signed certificates
10567 by the same criteria. The main comparison function is
10568 X509_check_issued() which performs these checks.
10570 Lot of changes were necessary in order to support this
10571 without completely rewriting the lookup code.
10573 Authority and subject key identifier are now cached.
10575 The LHASH 'certs' is X509_STORE has now been replaced
10576 by a STACK_OF(X509_OBJECT). This is mainly because an
10577 LHASH can't store or retrieve multiple objects with
10578 the same hash value.
10580 As a result various functions (which were all internal
10581 use only) have changed to handle the new X509_STORE
10582 structure. This will break anything that messed round
10583 with X509_STORE internally.
10585 The functions X509_STORE_add_cert() now checks for an
10586 exact match, rather than just subject name.
10588 The X509_STORE API doesn't directly support the retrieval
10589 of multiple certificates matching a given criteria, however
10590 this can be worked round by performing a lookup first
10591 (which will fill the cache with candidate certificates)
10592 and then examining the cache for matches. This is probably
10593 the best we can do without throwing out X509_LOOKUP
10594 entirely (maybe later...).
10596 The X509_VERIFY_CTX structure has been enhanced considerably.
10598 All certificate lookup operations now go via a get_issuer()
10599 callback. Although this currently uses an X509_STORE it
10600 can be replaced by custom lookups. This is a simple way
10601 to bypass the X509_STORE hackery necessary to make this
10602 work and makes it possible to use more efficient techniques
10603 in future. A very simple version which uses a simple
10604 STACK for its trusted certificate store is also provided
10605 using X509_STORE_CTX_trusted_stack().
10607 The verify_cb() and verify() callbacks now have equivalents
10608 in the X509_STORE_CTX structure.
10610 X509_STORE_CTX also has a 'flags' field which can be used
10611 to customise the verify behaviour.
10614 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10615 excludes S/MIME capabilities.
10618 *) When a certificate request is read in keep a copy of the
10619 original encoding of the signed data and use it when outputting
10620 again. Signatures then use the original encoding rather than
10621 a decoded, encoded version which may cause problems if the
10622 request is improperly encoded.
10625 *) For consistency with other BIO_puts implementations, call
10626 buffer_write(b, ...) directly in buffer_puts instead of calling
10629 In BIO_puts, increment b->num_write as in BIO_write.
10630 [Peter.Sylvester@EdelWeb.fr]
10632 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10633 BN_zero, we may not return a BIGNUM with an array consisting of
10634 words set to zero.)
10637 *) Avoid calling abort() from within the library when problems are
10638 detected, except if preprocessor symbols have been defined
10639 (such as REF_CHECK, BN_DEBUG etc.).
10642 *) New openssl application 'rsautl'. This utility can be
10643 used for low level RSA operations. DER public key
10644 BIO/fp routines also added.
10647 *) New Configure entry and patches for compiling on QNX 4.
10648 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10650 *) A demo state-machine implementation was sponsored by
10651 Nuron (http://www.nuron.com/) and is now available in
10652 demos/state_machine.
10655 *) New options added to the 'dgst' utility for signature
10656 generation and verification.
10659 *) Unrecognized PKCS#7 content types are now handled via a
10660 catch all ASN1_TYPE structure. This allows unsupported
10661 types to be stored as a "blob" and an application can
10662 encode and decode it manually.
10665 *) Fix various signed/unsigned issues to make a_strex.c
10666 compile under VC++.
10667 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10669 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10670 length if passed a buffer. ASN1_INTEGER_to_BN failed
10671 if passed a NULL BN and its argument was negative.
10672 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10674 *) Modification to PKCS#7 encoding routines to output definite
10675 length encoding. Since currently the whole structures are in
10676 memory there's not real point in using indefinite length
10677 constructed encoding. However if OpenSSL is compiled with
10678 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10681 *) Added BIO_vprintf() and BIO_vsnprintf().
10684 *) Added more prefixes to parse for in the strings written
10685 through a logging bio, to cover all the levels that are available
10686 through syslog. The prefixes are now:
10688 PANIC, EMERG, EMR => LOG_EMERG
10689 ALERT, ALR => LOG_ALERT
10690 CRIT, CRI => LOG_CRIT
10691 ERROR, ERR => LOG_ERR
10692 WARNING, WARN, WAR => LOG_WARNING
10693 NOTICE, NOTE, NOT => LOG_NOTICE
10694 INFO, INF => LOG_INFO
10695 DEBUG, DBG => LOG_DEBUG
10697 and as before, if none of those prefixes are present at the
10698 beginning of the string, LOG_ERR is chosen.
10700 On Win32, the LOG_* levels are mapped according to this:
10702 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10703 LOG_WARNING => EVENTLOG_WARNING_TYPE
10704 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10708 *) Made it possible to reconfigure with just the configuration
10709 argument "reconf" or "reconfigure". The command line arguments
10710 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10711 and are retrieved from there when reconfiguring.
10714 *) MD4 implemented.
10715 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10717 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10720 *) The obj_dat.pl script was messing up the sorting of object
10721 names. The reason was that it compared the quoted version
10722 of strings as a result "OCSP" > "OCSP Signing" because
10723 " > SPACE. Changed script to store unquoted versions of
10724 names and add quotes on output. It was also omitting some
10725 names from the lookup table if they were given a default
10726 value (that is if SN is missing it is given the same
10727 value as LN and vice versa), these are now added on the
10728 grounds that if an object has a name we should be able to
10729 look it up. Finally added warning output when duplicate
10730 short or long names are found.
10733 *) Changes needed for Tandem NSK.
10734 [Scott Uroff <scott@xypro.com>]
10736 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10737 RSA_padding_check_SSLv23(), special padding was never detected
10738 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10739 version rollback attacks was not effective.
10741 In s23_clnt.c, don't use special rollback-attack detection padding
10742 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10743 client; similarly, in s23_srvr.c, don't do the rollback check if
10744 SSL 2.0 is the only protocol enabled in the server.
10747 *) Make it possible to get hexdumps of unprintable data with 'openssl
10748 asn1parse'. By implication, the functions ASN1_parse_dump() and
10749 BIO_dump_indent() are added.
10752 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10753 these print out strings and name structures based on various
10754 flags including RFC2253 support and proper handling of
10755 multibyte characters. Added options to the 'x509' utility
10756 to allow the various flags to be set.
10759 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10760 Also change the functions X509_cmp_current_time() and
10761 X509_gmtime_adj() work with an ASN1_TIME structure,
10762 this will enable certificates using GeneralizedTime in validity
10763 dates to be checked.
10766 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10767 negative public key encodings) on by default,
10768 NO_NEG_PUBKEY_BUG can be set to disable it.
10771 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10772 content octets. An i2c_ASN1_OBJECT is unnecessary because
10773 the encoding can be trivially obtained from the structure.
10776 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10777 not read locks (CRYPTO_r_[un]lock).
10780 *) A first attempt at creating official support for shared
10781 libraries through configuration. I've kept it so the
10782 default is static libraries only, and the OpenSSL programs
10783 are always statically linked for now, but there are
10784 preparations for dynamic linking in place.
10785 This has been tested on Linux and Tru64.
10788 *) Randomness polling function for Win9x, as described in:
10789 Peter Gutmann, Software Generation of Practically Strong
10793 *) Fix so PRNG is seeded in req if using an already existing
10797 *) New options to smime application. -inform and -outform
10798 allow alternative formats for the S/MIME message including
10799 PEM and DER. The -content option allows the content to be
10800 specified separately. This should allow things like Netscape
10801 form signing output easier to verify.
10804 *) Fix the ASN1 encoding of tags using the 'long form'.
10807 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10808 STRING types. These convert content octets to and from the
10809 underlying type. The actual tag and length octets are
10810 already assumed to have been read in and checked. These
10811 are needed because all other string types have virtually
10812 identical handling apart from the tag. By having versions
10813 of the ASN1 functions that just operate on content octets
10814 IMPLICIT tagging can be handled properly. It also allows
10815 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10816 and ASN1_INTEGER are identical apart from the tag.
10819 *) Change the handling of OID objects as follows:
10821 - New object identifiers are inserted in objects.txt, following
10822 the syntax given in objects.README.
10823 - objects.pl is used to process obj_mac.num and create a new
10825 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10828 This is currently kind of a hack, and the perl code in objects.pl
10829 isn't very elegant, but it works as I intended. The simplest way
10830 to check that it worked correctly is to look in obj_dat.h and
10831 check the array nid_objs and make sure the objects haven't moved
10832 around (this is important!). Additions are OK, as well as
10833 consistent name changes.
10836 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10839 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10840 The given file adds to whatever has already been seeded into the
10841 random pool through the RANDFILE configuration file option or
10842 environment variable, or the default random state file.
10845 *) mkstack.pl now sorts each macro group into lexical order.
10846 Previously the output order depended on the order the files
10847 appeared in the directory, resulting in needless rewriting
10851 *) Patches to make OpenSSL compile under Win32 again. Mostly
10852 work arounds for the VC++ problem that it treats func() as
10853 func(void). Also stripped out the parts of mkdef.pl that
10854 added extra typesafe functions: these no longer exist.
10857 *) Reorganisation of the stack code. The macros are now all
10858 collected in safestack.h . Each macro is defined in terms of
10859 a "stack macro" of the form SKM_<name>(type, a, b). The
10860 DEBUG_SAFESTACK is now handled in terms of function casts,
10861 this has the advantage of retaining type safety without the
10862 use of additional functions. If DEBUG_SAFESTACK is not defined
10863 then the non typesafe macros are used instead. Also modified the
10864 mkstack.pl script to handle the new form. Needs testing to see
10865 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10866 the default if no major problems. Similar behaviour for ASN1_SET_OF
10867 and PKCS12_STACK_OF.
10870 *) When some versions of IIS use the 'NET' form of private key the
10871 key derivation algorithm is different. Normally MD5(password) is
10872 used as a 128 bit RC4 key. In the modified case
10873 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10874 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10875 as the old Netscape_RSA functions except they have an additional
10876 'sgckey' parameter which uses the modified algorithm. Also added
10877 an -sgckey command line option to the rsa utility. Thanks to
10878 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10879 algorithm to openssl-dev.
10882 *) The evp_local.h macros were using 'c.##kname' which resulted in
10883 invalid expansion on some systems (SCO 5.0.5 for example).
10884 Corrected to 'c.kname'.
10885 [Phillip Porch <root@theporch.com>]
10887 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10888 a STACK of email addresses from a certificate or request, these look
10889 in the subject name and the subject alternative name extensions and
10890 omit any duplicate addresses.
10893 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10894 This makes DSA verification about 2 % faster.
10897 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10898 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10899 plus overhead for 1024 bit moduli).
10900 This makes exponentiations about 0.5 % faster for 1024 bit
10901 exponents (as measured by "openssl speed rsa2048").
10904 *) Rename memory handling macros to avoid conflicts with other
10906 Malloc => OPENSSL_malloc
10907 Malloc_locked => OPENSSL_malloc_locked
10908 Realloc => OPENSSL_realloc
10909 Free => OPENSSL_free
10912 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10913 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10916 *) CygWin32 support.
10917 [John Jarvie <jjarvie@newsguy.com>]
10919 *) The type-safe stack code has been rejigged. It is now only compiled
10920 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10921 by default all type-specific stack functions are "#define"d back to
10922 standard stack functions. This results in more streamlined output
10923 but retains the type-safety checking possibilities of the original
10927 *) The STACK code has been cleaned up, and certain type declarations
10928 that didn't make a lot of sense have been brought in line. This has
10929 also involved a cleanup of sorts in safestack.h to more correctly
10930 map type-safe stack functions onto their plain stack counterparts.
10931 This work has also resulted in a variety of "const"ifications of
10932 lots of the code, especially "_cmp" operations which should normally
10933 be prototyped with "const" parameters anyway.
10936 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10937 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10938 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10939 where all of 'md' is used each time the PRNG is used, but 'state'
10940 is used only indexed by a cyclic counter. As entropy may not be
10941 well distributed from the beginning, 'md' is important as a
10942 chaining variable. However, the output function chains only half
10943 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10944 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10945 in all of 'state' being rewritten, with the new values depending
10946 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10949 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10950 the handshake is continued after ssl_verify_cert_chain();
10951 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10952 can lead to 'unexplainable' connection aborts later.
10953 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10955 *) Major EVP API cipher revision.
10956 Add hooks for extra EVP features. This allows various cipher
10957 parameters to be set in the EVP interface. Support added for variable
10958 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10959 setting of RC2 and RC5 parameters.
10961 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10964 Remove lots of duplicated code from the EVP library. For example *every*
10965 cipher init() function handles the 'iv' in the same way according to the
10966 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10967 for CFB and OFB modes they zero ctx->num.
10969 New functionality allows removal of S/MIME code RC2 hack.
10971 Most of the routines have the same form and so can be declared in terms
10974 By shifting this to the top level EVP_CipherInit() it can be removed from
10975 all individual ciphers. If the cipher wants to handle IVs or keys
10976 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10979 Change lots of functions like EVP_EncryptUpdate() to now return a
10980 value: although software versions of the algorithms cannot fail
10981 any installed hardware versions can.
10984 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10985 this option is set, tolerate broken clients that send the negotiated
10986 protocol version number instead of the requested protocol version
10990 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10991 i.e. non-zero for export ciphersuites, zero otherwise.
10992 Previous versions had this flag inverted, inconsistent with
10993 rsa_tmp_cb (..._TMP_RSA_CB).
10994 [Bodo Moeller; problem reported by Amit Chopra]
10996 *) Add missing DSA library text string. Work around for some IIS
10997 key files with invalid SEQUENCE encoding.
11000 *) Add a document (doc/standards.txt) that list all kinds of standards
11001 and so on that are implemented in OpenSSL.
11004 *) Enhance c_rehash script. Old version would mishandle certificates
11005 with the same subject name hash and wouldn't handle CRLs at all.
11006 Added -fingerprint option to crl utility, to support new c_rehash
11010 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
11013 *) Fix for SSL server purpose checking. Server checking was
11014 rejecting certificates which had extended key usage present
11015 but no ssl client purpose.
11016 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
11018 *) Make PKCS#12 code work with no password. The PKCS#12 spec
11019 is a little unclear about how a blank password is handled.
11020 Since the password in encoded as a BMPString with terminating
11021 double NULL a zero length password would end up as just the
11022 double NULL. However no password at all is different and is
11023 handled differently in the PKCS#12 key generation code. NS
11024 treats a blank password as zero length. MSIE treats it as no
11025 password on export: but it will try both on import. We now do
11026 the same: PKCS12_parse() tries zero length and no password if
11027 the password is set to "" or NULL (NULL is now a valid password:
11028 it wasn't before) as does the pkcs12 application.
11031 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
11032 perror when PEM_read_bio_X509_REQ fails, the error message must
11033 be obtained from the error queue.
11036 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
11037 it in ERR_remove_state if appropriate, and change ERR_get_state
11038 accordingly to avoid race conditions (this is necessary because
11039 thread_hash is no longer constant once set).
11042 *) Bugfix for linux-elf makefile.one.
11045 *) RSA_get_default_method() will now cause a default
11046 RSA_METHOD to be chosen if one doesn't exist already.
11047 Previously this was only set during a call to RSA_new()
11048 or RSA_new_method(NULL) meaning it was possible for
11049 RSA_get_default_method() to return NULL.
11052 *) Added native name translation to the existing DSO code
11053 that will convert (if the flag to do so is set) filenames
11054 that are sufficiently small and have no path information
11055 into a canonical native form. Eg. "blah" converted to
11056 "libblah.so" or "blah.dll" etc.
11059 *) New function ERR_error_string_n(e, buf, len) which is like
11060 ERR_error_string(e, buf), but writes at most 'len' bytes
11061 including the 0 terminator. For ERR_error_string_n, 'buf'
11063 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
11065 *) CONF library reworked to become more general. A new CONF
11066 configuration file reader "class" is implemented as well as a
11067 new functions (NCONF_*, for "New CONF") to handle it. The now
11068 old CONF_* functions are still there, but are reimplemented to
11069 work in terms of the new functions. Also, a set of functions
11070 to handle the internal storage of the configuration data is
11071 provided to make it easier to write new configuration file
11072 reader "classes" (I can definitely see something reading a
11073 configuration file in XML format, for example), called _CONF_*,
11074 or "the configuration storage API"...
11076 The new configuration file reading functions are:
11078 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
11079 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
11081 NCONF_default, NCONF_WIN32
11083 NCONF_dump_fp, NCONF_dump_bio
11085 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
11086 NCONF_new creates a new CONF object. This works in the same way
11087 as other interfaces in OpenSSL, like the BIO interface.
11088 NCONF_dump_* dump the internal storage of the configuration file,
11089 which is useful for debugging. All other functions take the same
11090 arguments as the old CONF_* functions with the exception of the
11091 first that must be a `CONF *' instead of a `LHASH *'.
11093 To make it easier to use the new classes with the old CONF_* functions,
11094 the function CONF_set_default_method is provided.
11097 *) Add '-tls1' option to 'openssl ciphers', which was already
11098 mentioned in the documentation but had not been implemented.
11099 (This option is not yet really useful because even the additional
11100 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
11103 *) Initial DSO code added into libcrypto for letting OpenSSL (and
11104 OpenSSL-based applications) load shared libraries and bind to
11105 them in a portable way.
11106 [Geoff Thorpe, with contributions from Richard Levitte]
11108 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
11110 *) Make sure _lrotl and _lrotr are only used with MSVC.
11112 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
11113 (the default implementation of RAND_status).
11115 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
11116 to '-clrext' (= clear extensions), as intended and documented.
11117 [Bodo Moeller; inconsistency pointed out by Michael Attili
11118 <attili@amaxo.com>]
11120 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
11121 was larger than the MD block size.
11122 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
11124 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
11125 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
11126 using the passed key: if the passed key was a private key the result
11127 of X509_print(), for example, would be to print out all the private key
11131 *) des_quad_cksum() byte order bug fix.
11132 [Ulf Möller, using the problem description in krb4-0.9.7, where
11133 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
11135 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
11137 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
11139 *) For easily testing in shell scripts whether some command
11140 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
11141 returns with exit code 0 iff no command of the given name is available.
11142 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
11143 the output goes to stdout and nothing is printed to stderr.
11144 Additional arguments are always ignored.
11146 Since for each cipher there is a command of the same name,
11147 the 'no-cipher' compilation switches can be tested this way.
11149 ('openssl no-XXX' is not able to detect pseudo-commands such
11150 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
11153 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
11156 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
11157 is set; it will be thrown away anyway because each handshake creates
11159 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
11160 to parameters -- in previous versions (since OpenSSL 0.9.3) the
11161 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
11162 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
11165 *) New s_client option -ign_eof: EOF at stdin is ignored, and
11166 'Q' and 'R' lose their special meanings (quit/renegotiate).
11167 This is part of what -quiet does; unlike -quiet, -ign_eof
11168 does not suppress any output.
11171 *) Add compatibility options to the purpose and trust code. The
11172 purpose X509_PURPOSE_ANY is "any purpose" which automatically
11173 accepts a certificate or CA, this was the previous behaviour,
11174 with all the associated security issues.
11176 X509_TRUST_COMPAT is the old trust behaviour: only and
11177 automatically trust self signed roots in certificate store. A
11178 new trust setting X509_TRUST_DEFAULT is used to specify that
11179 a purpose has no associated trust setting and it should instead
11180 use the value in the default purpose.
11183 *) Fix the PKCS#8 DSA private key code so it decodes keys again
11184 and fix a memory leak.
11187 *) In util/mkerr.pl (which implements 'make errors'), preserve
11188 reason strings from the previous version of the .c file, as
11189 the default to have only downcase letters (and digits) in
11190 automatically generated reasons codes is not always appropriate.
11193 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
11194 using strerror. Previously, ERR_reason_error_string() returned
11195 library names as reason strings for SYSerr; but SYSerr is a special
11196 case where small numbers are errno values, not library numbers.
11199 *) Add '-dsaparam' option to 'openssl dhparam' application. This
11200 converts DSA parameters into DH parameters. (When creating parameters,
11201 DSA_generate_parameters is used.)
11204 *) Include 'length' (recommended exponent length) in C code generated
11205 by 'openssl dhparam -C'.
11208 *) The second argument to set_label in perlasm was already being used
11209 so couldn't be used as a "file scope" flag. Moved to third argument
11213 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
11214 instead of RAND_bytes for encryption IVs and salts.
11217 *) Include RAND_status() into RAND_METHOD instead of implementing
11218 it only for md_rand.c Otherwise replacing the PRNG by calling
11219 RAND_set_rand_method would be impossible.
11222 *) Don't let DSA_generate_key() enter an infinite loop if the random
11223 number generation fails.
11226 *) New 'rand' application for creating pseudo-random output.
11229 *) Added configuration support for Linux/IA64
11230 [Rolf Haberrecker <rolf@suse.de>]
11232 *) Assembler module support for Mingw32.
11235 *) Shared library support for HPUX (in shlib/).
11236 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
11238 *) Shared library support for Solaris gcc.
11239 [Lutz Behnke <behnke@trustcenter.de>]
11241 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
11243 *) PKCS7_encrypt() was adding text MIME headers twice because they
11244 were added manually and by SMIME_crlf_copy().
11247 *) In bntest.c don't call BN_rand with zero bits argument.
11248 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
11250 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
11251 case was implemented. This caused BN_div_recp() to fail occasionally.
11254 *) Add an optional second argument to the set_label() in the perl
11255 assembly language builder. If this argument exists and is set
11256 to 1 it signals that the assembler should use a symbol whose
11257 scope is the entire file, not just the current function. This
11258 is needed with MASM which uses the format label:: for this scope.
11259 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
11261 *) Change the ASN1 types so they are typedefs by default. Before
11262 almost all types were #define'd to ASN1_STRING which was causing
11263 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
11267 *) Change names of new functions to the new get1/get0 naming
11268 convention: After 'get1', the caller owns a reference count
11269 and has to call ..._free; 'get0' returns a pointer to some
11270 data structure without incrementing reference counters.
11271 (Some of the existing 'get' functions increment a reference
11272 counter, some don't.)
11273 Similarly, 'set1' and 'add1' functions increase reference
11274 counters or duplicate objects.
11277 *) Allow for the possibility of temp RSA key generation failure:
11278 the code used to assume it always worked and crashed on failure.
11281 *) Fix potential buffer overrun problem in BIO_printf().
11282 [Ulf Möller, using public domain code by Patrick Powell; problem
11283 pointed out by David Sacerdote <das33@cornell.edu>]
11285 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
11286 RAND_egd() and RAND_status(). In the command line application,
11287 the EGD socket can be specified like a seed file using RANDFILE
11291 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
11292 Some CAs (e.g. Verisign) distribute certificates in this form.
11295 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
11296 list to exclude them. This means that no special compilation option
11297 is needed to use anonymous DH: it just needs to be included in the
11301 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
11302 EVP_MD_type. The old functionality is available in a new macro called
11303 EVP_MD_md(). Change code that uses it and update docs.
11306 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
11307 where the 'void *' argument is replaced by a function pointer argument.
11308 Previously 'void *' was abused to point to functions, which works on
11309 many platforms, but is not correct. As these functions are usually
11310 called by macros defined in OpenSSL header files, most source code
11311 should work without changes.
11314 *) <openssl/opensslconf.h> (which is created by Configure) now contains
11315 sections with information on -D... compiler switches used for
11316 compiling the library so that applications can see them. To enable
11317 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
11318 must be defined. E.g.,
11319 #define OPENSSL_ALGORITHM_DEFINES
11320 #include <openssl/opensslconf.h>
11321 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
11322 [Richard Levitte, Ulf and Bodo Möller]
11324 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
11328 *) Change the 'other' type in certificate aux info to a STACK_OF
11329 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
11330 the required ASN1 format: arbitrary types determined by an OID.
11333 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
11334 argument to 'req'. This is not because the function is newer or
11335 better than others it just uses the work 'NEW' in the certificate
11336 request header lines. Some software needs this.
11339 *) Reorganise password command line arguments: now passwords can be
11340 obtained from various sources. Delete the PEM_cb function and make
11341 it the default behaviour: i.e. if the callback is NULL and the
11342 usrdata argument is not NULL interpret it as a null terminated pass
11343 phrase. If usrdata and the callback are NULL then the pass phrase
11344 is prompted for as usual.
11347 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
11348 the support is automatically enabled. The resulting binaries will
11349 autodetect the card and use it if present.
11350 [Ben Laurie and Compaq Inc.]
11352 *) Work around for Netscape hang bug. This sends certificate request
11353 and server done in one record. Since this is perfectly legal in the
11354 SSL/TLS protocol it isn't a "bug" option and is on by default. See
11355 the bugs/SSLv3 entry for more info.
11358 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
11361 *) Add -rand argument to smime and pkcs12 applications and read/write
11365 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
11368 *) Add command line password options to the remaining applications.
11371 *) Bug fix for BN_div_recp() for numerators with an even number of
11375 *) More tests in bntest.c, and changed test_bn output.
11378 *) ./config recognizes MacOS X now.
11381 *) Bug fix for BN_div() when the first words of num and divisor are
11382 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
11385 *) Add support for various broken PKCS#8 formats, and command line
11386 options to produce them.
11389 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
11390 get temporary BIGNUMs from a BN_CTX.
11393 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
11397 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
11398 include a #define from the old name to the new. The original intent
11399 was that statically linked binaries could for example just call
11400 SSLeay_add_all_ciphers() to just add ciphers to the table and not
11401 link with digests. This never worked because SSLeay_add_all_digests()
11402 and SSLeay_add_all_ciphers() were in the same source file so calling
11403 one would link with the other. They are now in separate source files.
11406 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
11409 *) Use a less unusual form of the Miller-Rabin primality test (it used
11410 a binary algorithm for exponentiation integrated into the Miller-Rabin
11411 loop, our standard modexp algorithms are faster).
11414 *) Support for the EBCDIC character set completed.
11415 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
11417 *) Source code cleanups: use const where appropriate, eliminate casts,
11418 use void * instead of char * in lhash.
11421 *) Bugfix: ssl3_send_server_key_exchange was not restartable
11422 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
11423 this the server could overwrite ephemeral keys that the client
11427 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
11428 using 50 iterations of the Rabin-Miller test.
11430 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
11431 iterations of the Rabin-Miller test as required by the appendix
11432 to FIPS PUB 186[-1]) instead of DSA_is_prime.
11433 As BN_is_prime_fasttest includes trial division, DSA parameter
11434 generation becomes much faster.
11436 This implies a change for the callback functions in DSA_is_prime
11437 and DSA_generate_parameters: The callback function is called once
11438 for each positive witness in the Rabin-Miller test, not just
11439 occasionally in the inner loop; and the parameters to the
11440 callback function now provide an iteration count for the outer
11441 loop rather than for the current invocation of the inner loop.
11442 DSA_generate_parameters additionally can call the callback
11443 function with an 'iteration count' of -1, meaning that a
11444 candidate has passed the trial division test (when q is generated
11445 from an application-provided seed, trial division is skipped).
11448 *) New function BN_is_prime_fasttest that optionally does trial
11449 division before starting the Rabin-Miller test and has
11450 an additional BN_CTX * argument (whereas BN_is_prime always
11451 has to allocate at least one BN_CTX).
11452 'callback(1, -1, cb_arg)' is called when a number has passed the
11453 trial division stage.
11456 *) Fix for bug in CRL encoding. The validity dates weren't being handled
11460 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
11463 *) New function BN_pseudo_rand().
11466 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
11467 bignum version of BN_from_montgomery() with the working code from
11468 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
11472 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
11473 made it impossible to use the same SSL_SESSION data structure in
11474 SSL2 clients in multiple threads.
11477 *) The return value of RAND_load_file() no longer counts bytes obtained
11478 by stat(). RAND_load_file(..., -1) is new and uses the complete file
11479 to seed the PRNG (previously an explicit byte count was required).
11480 [Ulf Möller, Bodo Möller]
11482 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
11483 used (char *) instead of (void *) and had casts all over the place.
11486 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
11489 *) Retain source code compatibility for BN_prime_checks macro:
11490 BN_is_prime(..., BN_prime_checks, ...) now uses
11491 BN_prime_checks_for_size to determine the appropriate number of
11492 Rabin-Miller iterations.
11495 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
11496 DH_CHECK_P_NOT_SAFE_PRIME.
11497 (Check if this is true? OpenPGP calls them "strong".)
11500 *) Merge the functionality of "dh" and "gendh" programs into a new program
11501 "dhparam". The old programs are retained for now but will handle DH keys
11502 (instead of parameters) in future.
11505 *) Make the ciphers, s_server and s_client programs check the return values
11506 when a new cipher list is set.
11509 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
11510 ciphers. Before when the 56bit ciphers were enabled the sorting was
11513 The syntax for the cipher sorting has been extended to support sorting by
11514 cipher-strength (using the strength_bits hard coded in the tables).
11515 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
11517 Fix a bug in the cipher-command parser: when supplying a cipher command
11518 string with an "undefined" symbol (neither command nor alphanumeric
11519 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
11520 an error is flagged.
11522 Due to the strength-sorting extension, the code of the
11523 ssl_create_cipher_list() function was completely rearranged. I hope that
11524 the readability was also increased :-)
11525 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
11527 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
11528 for the first serial number and places 2 in the serial number file. This
11529 avoids problems when the root CA is created with serial number zero and
11530 the first user certificate has the same issuer name and serial number
11534 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
11535 the new code. Add documentation for this stuff.
11538 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
11539 X509_*() to X509at_*() on the grounds that they don't handle X509
11540 structures and behave in an analogous way to the X509v3 functions:
11541 they shouldn't be called directly but wrapper functions should be used
11544 So we also now have some wrapper functions that call the X509at functions
11545 when passed certificate requests. (TO DO: similar things can be done with
11546 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11547 things. Some of these need some d2i or i2d and print functionality
11548 because they handle more complex structures.)
11551 *) Add missing #ifndefs that caused missing symbols when building libssl
11552 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11553 NO_RSA in ssl/s2*.c.
11554 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11556 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11557 has a return value which indicates the quality of the random data
11558 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11559 error queue. New function RAND_pseudo_bytes() generates output that is
11560 guaranteed to be unique but not unpredictable. RAND_add is like
11561 RAND_seed, but takes an extra argument for an entropy estimate
11562 (RAND_seed always assumes full entropy).
11565 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11566 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11567 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11568 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11569 false-positive rate of at most 2^-80 for random input.
11572 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11575 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11576 in the 0.9.5 release), this returns the chain
11577 from an X509_CTX structure with a dup of the stack and all
11578 the X509 reference counts upped: so the stack will exist
11579 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11582 Also make SSL_SESSION_print() print out the verify return
11586 *) Add manpage for the pkcs12 command. Also change the default
11587 behaviour so MAC iteration counts are used unless the new
11588 -nomaciter option is used. This improves file security and
11589 only older versions of MSIE (4.0 for example) need it.
11592 *) Honor the no-xxx Configure options when creating .DEF files.
11595 *) Add PKCS#10 attributes to field table: challengePassword,
11596 unstructuredName and unstructuredAddress. These are taken from
11597 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11598 international characters are used.
11600 More changes to X509_ATTRIBUTE code: allow the setting of types
11601 based on strings. Remove the 'loc' parameter when adding
11602 attributes because these will be a SET OF encoding which is sorted
11606 *) Initial changes to the 'req' utility to allow request generation
11607 automation. This will allow an application to just generate a template
11608 file containing all the field values and have req construct the
11611 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11612 used all over the place including certificate requests and PKCS#7
11613 structures. They are currently handled manually where necessary with
11614 some primitive wrappers for PKCS#7. The new functions behave in a
11615 manner analogous to the X509 extension functions: they allow
11616 attributes to be looked up by NID and added.
11618 Later something similar to the X509V3 code would be desirable to
11619 automatically handle the encoding, decoding and printing of the
11620 more complex types. The string types like challengePassword can
11621 be handled by the string table functions.
11623 Also modified the multi byte string table handling. Now there is
11624 a 'global mask' which masks out certain types. The table itself
11625 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11626 is useful when for example there is only one permissible type
11627 (as in countryName) and using the mask might result in no valid
11631 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11632 SSL_get_peer_finished to allow applications to obtain the latest
11633 Finished messages sent to the peer or expected from the peer,
11634 respectively. (SSL_get_peer_finished is usually the Finished message
11635 actually received from the peer, otherwise the protocol will be aborted.)
11637 As the Finished message are message digests of the complete handshake
11638 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11639 be used for external authentication procedures when the authentication
11640 provided by SSL/TLS is not desired or is not enough.
11643 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11644 the host supports BWX extension and if Compaq C is present on the
11645 $PATH. Just exploiting of the BWX extension results in 20-30%
11646 performance kick for some algorithms, e.g. DES and RC4 to mention
11647 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11651 *) Add support for MS "fast SGC". This is arguably a violation of the
11652 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11653 weak crypto and after checking the certificate is SGC a second one
11654 with strong crypto. MS SGC stops the first handshake after receiving
11655 the server certificate message and sends a second client hello. Since
11656 a server will typically do all the time consuming operations before
11657 expecting any further messages from the client (server key exchange
11658 is the most expensive) there is little difference between the two.
11660 To get OpenSSL to support MS SGC we have to permit a second client
11661 hello message after we have sent server done. In addition we have to
11662 reset the MAC if we do get this second client hello.
11665 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11666 if a DER encoded private key is RSA or DSA traditional format. Changed
11667 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11668 format DER encoded private key. Newer code should use PKCS#8 format which
11669 has the key type encoded in the ASN1 structure. Added DER private key
11670 support to pkcs8 application.
11673 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11674 ciphersuites has been selected (as required by the SSL 3/TLS 1
11675 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11676 is set, we interpret this as a request to violate the specification
11677 (the worst that can happen is a handshake failure, and 'correct'
11678 behaviour would result in a handshake failure anyway).
11681 *) In SSL_CTX_add_session, take into account that there might be multiple
11682 SSL_SESSION structures with the same session ID (e.g. when two threads
11683 concurrently obtain them from an external cache).
11684 The internal cache can handle only one SSL_SESSION with a given ID,
11685 so if there's a conflict, we now throw out the old one to achieve
11689 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11690 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11691 some routines that use cipher OIDs: some ciphers do not have OIDs
11692 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11696 *) Simplify the trust setting structure and code. Now we just have
11697 two sequences of OIDs for trusted and rejected settings. These will
11698 typically have values the same as the extended key usage extension
11699 and any application specific purposes.
11701 The trust checking code now has a default behaviour: it will just
11702 check for an object with the same NID as the passed id. Functions can
11703 be provided to override either the default behaviour or the behaviour
11704 for a given id. SSL client, server and email already have functions
11705 in place for compatibility: they check the NID and also return "trusted"
11706 if the certificate is self signed.
11709 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11710 traditional format into an EVP_PKEY structure.
11713 *) Add a password callback function PEM_cb() which either prompts for
11714 a password if usr_data is NULL or otherwise assumes it is a null
11715 terminated password. Allow passwords to be passed on command line
11716 environment or config files in a few more utilities.
11719 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11720 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11721 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11722 Update documentation.
11725 *) Support for ASN1 "NULL" type. This could be handled before by using
11726 ASN1_TYPE but there wasn't any function that would try to read a NULL
11727 and produce an error if it couldn't. For compatibility we also have
11728 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11729 don't allocate anything because they don't need to.
11732 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11734 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11736 *) Rebuild of the memory allocation routines used by OpenSSL code and
11737 possibly others as well. The purpose is to make an interface that
11738 provide hooks so anyone can build a separate set of allocation and
11739 deallocation routines to be used by OpenSSL, for example memory
11740 pool implementations, or something else, which was previously hard
11741 since Malloc(), Realloc() and Free() were defined as macros having
11742 the values malloc, realloc and free, respectively (except for Win32
11743 compilations). The same is provided for memory debugging code.
11744 OpenSSL already comes with functionality to find memory leaks, but
11745 this gives people a chance to debug other memory problems.
11747 With these changes, a new set of functions and macros have appeared:
11749 CRYPTO_set_mem_debug_functions() [F]
11750 CRYPTO_get_mem_debug_functions() [F]
11751 CRYPTO_dbg_set_options() [F]
11752 CRYPTO_dbg_get_options() [F]
11753 CRYPTO_malloc_debug_init() [M]
11755 The memory debug functions are NULL by default, unless the library
11756 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11757 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11758 gives the standard debugging functions that come with OpenSSL) or
11759 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11760 provided by the library user) must be used. When the standard
11761 debugging functions are used, CRYPTO_dbg_set_options can be used to
11762 request additional information:
11763 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11764 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11766 Also, things like CRYPTO_set_mem_functions will always give the
11767 expected result (the new set of functions is used for allocation
11768 and deallocation) at all times, regardless of platform and compiler
11771 To finish it up, some functions that were never use in any other
11772 way than through macros have a new API and new semantic:
11774 CRYPTO_dbg_malloc()
11775 CRYPTO_dbg_realloc()
11778 All macros of value have retained their old syntax.
11779 [Richard Levitte and Bodo Moeller]
11781 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11782 ordering of SMIMECapabilities wasn't in "strength order" and there
11783 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11787 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11788 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11789 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11791 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11792 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11793 functionality to handle multipart/signed properly) and a utility
11794 called 'smime' to call all this stuff. This is based on code I
11795 originally wrote for Celo who have kindly allowed it to be
11796 included in OpenSSL.
11799 *) Add variants des_set_key_checked and des_set_key_unchecked of
11800 des_set_key (aka des_key_sched). Global variable des_check_key
11801 decides which of these is called by des_set_key; this way
11802 des_check_key behaves as it always did, but applications and
11803 the library itself, which was buggy for des_check_key == 1,
11804 have a cleaner way to pick the version they need.
11807 *) New function PKCS12_newpass() which changes the password of a
11811 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11812 dynamic mix. In both cases the ids can be used as an index into the
11813 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11814 functions so they accept a list of the field values and the
11815 application doesn't need to directly manipulate the X509_TRUST
11819 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11823 *) Modify the way the V3 extension code looks up extensions. This now
11824 works in a similar way to the object code: we have some "standard"
11825 extensions in a static table which is searched with OBJ_bsearch()
11826 and the application can add dynamic ones if needed. The file
11827 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11828 updated whenever a new extension is added to the core code and kept
11829 in ext_nid order. There is a simple program 'tabtest.c' which checks
11830 this. New extensions are not added too often so this file can readily
11831 be maintained manually.
11833 There are two big advantages in doing things this way. The extensions
11834 can be looked up immediately and no longer need to be "added" using
11835 X509V3_add_standard_extensions(): this function now does nothing.
11836 [Side note: I get *lots* of email saying the extension code doesn't
11837 work because people forget to call this function]
11838 Also no dynamic allocation is done unless new extensions are added:
11839 so if we don't add custom extensions there is no need to call
11840 X509V3_EXT_cleanup().
11843 *) Modify enc utility's salting as follows: make salting the default. Add a
11844 magic header, so unsalted files fail gracefully instead of just decrypting
11845 to garbage. This is because not salting is a big security hole, so people
11846 should be discouraged from doing it.
11849 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11850 digest to be passed on the command line but it only used this
11851 parameter when signing a certificate. Modified so all relevant
11852 operations are affected by the digest parameter including the
11853 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11854 DSA key was used because it didn't fix the digest.
11857 *) Initial certificate chain verify code. Currently tests the untrusted
11858 certificates for consistency with the verify purpose (which is set
11859 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11861 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11862 this is because it will reject chains with invalid extensions whereas
11863 every previous version of OpenSSL and SSLeay made no checks at all.
11865 Trust code: checks the root CA for the relevant trust settings. Trust
11866 settings have an initial value consistent with the verify purpose: e.g.
11867 if the verify purpose is for SSL client use it expects the CA to be
11868 trusted for SSL client use. However the default value can be changed to
11869 permit custom trust settings: one example of this would be to only trust
11870 certificates from a specific "secure" set of CAs.
11872 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11873 which should be used for version portability: especially since the
11874 verify structure is likely to change more often now.
11876 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11877 to set them. If not set then assume SSL clients will verify SSL servers
11880 Two new options to the verify program: -untrusted allows a set of
11881 untrusted certificates to be passed in and -purpose which sets the
11882 intended purpose of the certificate. If a purpose is set then the
11883 new chain verify code is used to check extension consistency.
11886 *) Support for the authority information access extension.
11889 *) Modify RSA and DSA PEM read routines to transparently handle
11890 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11891 public keys in a format compatible with certificate
11892 SubjectPublicKeyInfo structures. Unfortunately there were already
11893 functions called *_PublicKey_* which used various odd formats so
11894 these are retained for compatibility: however the DSA variants were
11895 never in a public release so they have been deleted. Changed dsa/rsa
11896 utilities to handle the new format: note no releases ever handled public
11897 keys so we should be OK.
11899 The primary motivation for this change is to avoid the same fiasco
11900 that dogs private keys: there are several incompatible private key
11901 formats some of which are standard and some OpenSSL specific and
11902 require various evil hacks to allow partial transparent handling and
11903 even then it doesn't work with DER formats. Given the option anything
11904 other than PKCS#8 should be dumped: but the other formats have to
11905 stay in the name of compatibility.
11907 With public keys and the benefit of hindsight one standard format
11908 is used which works with EVP_PKEY, RSA or DSA structures: though
11909 it clearly returns an error if you try to read the wrong kind of key.
11911 Added a -pubkey option to the 'x509' utility to output the public key.
11912 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11913 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11914 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11915 that do the same as the EVP_PKEY_assign_*() except they up the
11916 reference count of the added key (they don't "swallow" the
11920 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11921 CRLs would fail if the file contained no certificates or no CRLs:
11922 added a new function to read in both types and return the number
11923 read: this means that if none are read it will be an error. The
11924 DER versions of the certificate and CRL reader would always fail
11925 because it isn't possible to mix certificates and CRLs in DER format
11926 without choking one or the other routine. Changed this to just read
11927 a certificate: this is the best we can do. Also modified the code
11928 in apps/verify.c to take notice of return codes: it was previously
11929 attempting to read in certificates from NULL pointers and ignoring
11930 any errors: this is one reason why the cert and CRL reader seemed
11931 to work. It doesn't check return codes from the default certificate
11932 routines: these may well fail if the certificates aren't installed.
11935 *) Code to support otherName option in GeneralName.
11938 *) First update to verify code. Change the verify utility
11939 so it warns if it is passed a self signed certificate:
11940 for consistency with the normal behaviour. X509_verify
11941 has been modified to it will now verify a self signed
11942 certificate if *exactly* the same certificate appears
11943 in the store: it was previously impossible to trust a
11944 single self signed certificate. This means that:
11945 openssl verify ss.pem
11946 now gives a warning about a self signed certificate but
11947 openssl verify -CAfile ss.pem ss.pem
11951 *) For servers, store verify_result in SSL_SESSION data structure
11952 (and add it to external session representation).
11953 This is needed when client certificate verifications fails,
11954 but an application-provided verification callback (set by
11955 SSL_CTX_set_cert_verify_callback) allows accepting the session
11956 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11957 but returns 1): When the session is reused, we have to set
11958 ssl->verify_result to the appropriate error code to avoid
11960 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11962 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11963 case in PKCS7_dataInit() where the signed PKCS7 structure
11964 didn't contain any existing data because it was being created.
11965 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11967 *) Add a salt to the key derivation routines in enc.c. This
11968 forms the first 8 bytes of the encrypted file. Also add a
11969 -S option to allow a salt to be input on the command line.
11972 *) New function X509_cmp(). Oddly enough there wasn't a function
11973 to compare two certificates. We do this by working out the SHA1
11974 hash and comparing that. X509_cmp() will be needed by the trust
11978 *) SSL_get1_session() is like SSL_get_session(), but increments
11979 the reference count in the SSL_SESSION returned.
11980 [Geoff Thorpe <geoff@eu.c2.net>]
11982 *) Fix for 'req': it was adding a null to request attributes.
11983 Also change the X509_LOOKUP and X509_INFO code to handle
11984 certificate auxiliary information.
11987 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11991 *) Add the possibility to add extra information to the memory leak
11992 detecting output, to form tracebacks, showing from where each
11993 allocation was originated: CRYPTO_push_info("constant string") adds
11994 the string plus current file name and line number to a per-thread
11995 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11996 is like calling CYRPTO_pop_info() until the stack is empty.
11997 Also updated memory leak detection code to be multi-thread-safe.
12000 *) Add options -text and -noout to pkcs7 utility and delete the
12001 encryption options which never did anything. Update docs.
12004 *) Add options to some of the utilities to allow the pass phrase
12005 to be included on either the command line (not recommended on
12006 OSes like Unix) or read from the environment. Update the
12007 manpages and fix a few bugs.
12010 *) Add a few manpages for some of the openssl commands.
12013 *) Fix the -revoke option in ca. It was freeing up memory twice,
12014 leaking and not finding already revoked certificates.
12017 *) Extensive changes to support certificate auxiliary information.
12018 This involves the use of X509_CERT_AUX structure and X509_AUX
12019 functions. An X509_AUX function such as PEM_read_X509_AUX()
12020 can still read in a certificate file in the usual way but it
12021 will also read in any additional "auxiliary information". By
12022 doing things this way a fair degree of compatibility can be
12023 retained: existing certificates can have this information added
12024 using the new 'x509' options.
12026 Current auxiliary information includes an "alias" and some trust
12027 settings. The trust settings will ultimately be used in enhanced
12028 certificate chain verification routines: currently a certificate
12029 can only be trusted if it is self signed and then it is trusted
12033 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
12034 The problem was that one of the replacement routines had not been working
12035 since SSLeay releases. For now the offending routine has been replaced
12036 with non-optimised assembler. Even so, this now gives around 95%
12037 performance improvement for 1024 bit RSA signs.
12040 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
12041 handling. Most clients have the effective key size in bits equal to
12042 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
12043 A few however don't do this and instead use the size of the decrypted key
12044 to determine the RC2 key length and the AlgorithmIdentifier to determine
12045 the effective key length. In this case the effective key length can still
12046 be 40 bits but the key length can be 168 bits for example. This is fixed
12047 by manually forcing an RC2 key into the EVP_PKEY structure because the
12048 EVP code can't currently handle unusual RC2 key sizes: it always assumes
12049 the key length and effective key length are equal.
12052 *) Add a bunch of functions that should simplify the creation of
12053 X509_NAME structures. Now you should be able to do:
12054 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
12055 and have it automatically work out the correct field type and fill in
12056 the structures. The more adventurous can try:
12057 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
12058 and it will (hopefully) work out the correct multibyte encoding.
12061 *) Change the 'req' utility to use the new field handling and multibyte
12062 copy routines. Before the DN field creation was handled in an ad hoc
12063 way in req, ca, and x509 which was rather broken and didn't support
12064 BMPStrings or UTF8Strings. Since some software doesn't implement
12065 BMPStrings or UTF8Strings yet, they can be enabled using the config file
12066 using the dirstring_type option. See the new comment in the default
12067 openssl.cnf for more info.
12070 *) Make crypto/rand/md_rand.c more robust:
12071 - Assure unique random numbers after fork().
12072 - Make sure that concurrent threads access the global counter and
12073 md serializably so that we never lose entropy in them
12074 or use exactly the same state in multiple threads.
12075 Access to the large state is not always serializable because
12076 the additional locking could be a performance killer, and
12077 md should be large enough anyway.
12080 *) New file apps/app_rand.c with commonly needed functionality
12081 for handling the random seed file.
12083 Use the random seed file in some applications that previously did not:
12085 dsaparam -genkey (which also ignored its '-rand' option),
12088 x509 (when signing).
12089 Except on systems with /dev/urandom, it is crucial to have a random
12090 seed file at least for key creation, DSA signing, and for DH exchanges;
12091 for RSA signatures we could do without one.
12093 gendh and gendsa (unlike genrsa) used to read only the first byte
12094 of each file listed in the '-rand' option. The function as previously
12095 found in genrsa is now in app_rand.c and is used by all programs
12096 that support '-rand'.
12099 *) In RAND_write_file, use mode 0600 for creating files;
12100 don't just chmod when it may be too late.
12103 *) Report an error from X509_STORE_load_locations
12104 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
12107 *) New function ASN1_mbstring_copy() this copies a string in either
12108 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
12109 into an ASN1_STRING type. A mask of permissible types is passed
12110 and it chooses the "minimal" type to use or an error if not type
12114 *) Add function equivalents to the various macros in asn1.h. The old
12115 macros are retained with an M_ prefix. Code inside the library can
12116 use the M_ macros. External code (including the openssl utility)
12117 should *NOT* in order to be "shared library friendly".
12120 *) Add various functions that can check a certificate's extensions
12121 to see if it usable for various purposes such as SSL client,
12122 server or S/MIME and CAs of these types. This is currently
12123 VERY EXPERIMENTAL but will ultimately be used for certificate chain
12124 verification. Also added a -purpose flag to x509 utility to
12125 print out all the purposes.
12128 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
12132 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
12133 for, obtain and decode and extension and obtain its critical flag.
12134 This allows all the necessary extension code to be handled in a
12135 single function call.
12138 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
12139 platforms. See crypto/rc4/rc4_enc.c for further details.
12142 *) New -noout option to asn1parse. This causes no output to be produced
12143 its main use is when combined with -strparse and -out to extract data
12144 from a file (which may not be in ASN.1 format).
12147 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
12148 when producing the local key id.
12149 [Richard Levitte <levitte@stacken.kth.se>]
12151 *) New option -dhparam in s_server. This allows a DH parameter file to be
12152 stated explicitly. If it is not stated then it tries the first server
12153 certificate file. The previous behaviour hard coded the filename
12157 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
12158 a public key to be input or output. For example:
12159 openssl rsa -in key.pem -pubout -out pubkey.pem
12160 Also added necessary DSA public key functions to handle this.
12163 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
12164 in the message. This was handled by allowing
12165 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
12166 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
12168 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
12169 to the end of the strings whereas this didn't. This would cause problems
12170 if strings read with d2i_ASN1_bytes() were later modified.
12171 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
12173 *) Fix for base64 decode bug. When a base64 bio reads only one line of
12174 data and it contains EOF it will end up returning an error. This is
12175 caused by input 46 bytes long. The cause is due to the way base64
12176 BIOs find the start of base64 encoded data. They do this by trying a
12177 trial decode on each line until they find one that works. When they
12178 do a flag is set and it starts again knowing it can pass all the
12179 data directly through the decoder. Unfortunately it doesn't reset
12180 the context it uses. This means that if EOF is reached an attempt
12181 is made to pass two EOFs through the context and this causes the
12182 resulting error. This can also cause other problems as well. As is
12183 usual with these problems it takes *ages* to find and the fix is
12184 trivial: move one line.
12185 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
12187 *) Ugly workaround to get s_client and s_server working under Windows. The
12188 old code wouldn't work because it needed to select() on sockets and the
12189 tty (for keypresses and to see if data could be written). Win32 only
12190 supports select() on sockets so we select() with a 1s timeout on the
12191 sockets and then see if any characters are waiting to be read, if none
12192 are present then we retry, we also assume we can always write data to
12193 the tty. This isn't nice because the code then blocks until we've
12194 received a complete line of data and it is effectively polling the
12195 keyboard at 1s intervals: however it's quite a bit better than not
12196 working at all :-) A dedicated Windows application might handle this
12197 with an event loop for example.
12200 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
12201 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
12202 will be called when RSA_sign() and RSA_verify() are used. This is useful
12203 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
12204 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
12205 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
12206 This necessitated the support of an extra signature type NID_md5_sha1
12207 for SSL signatures and modifications to the SSL library to use it instead
12208 of calling RSA_public_decrypt() and RSA_private_encrypt().
12211 *) Add new -verify -CAfile and -CApath options to the crl program, these
12212 will lookup a CRL issuers certificate and verify the signature in a
12213 similar way to the verify program. Tidy up the crl program so it
12214 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
12215 less strict. It will now permit CRL extensions even if it is not
12216 a V2 CRL: this will allow it to tolerate some broken CRLs.
12219 *) Initialize all non-automatic variables each time one of the openssl
12220 sub-programs is started (this is necessary as they may be started
12221 multiple times from the "OpenSSL>" prompt).
12222 [Lennart Bang, Bodo Moeller]
12224 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
12225 removing all other RSA functionality (this is what NO_RSA does). This
12226 is so (for example) those in the US can disable those operations covered
12227 by the RSA patent while allowing storage and parsing of RSA keys and RSA
12231 *) Non-copying interface to BIO pairs.
12232 (still largely untested)
12235 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
12236 ASCII string. This was handled independently in various places before.
12239 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
12240 UTF8 strings a character at a time.
12243 *) Use client_version from client hello to select the protocol
12244 (s23_srvr.c) and for RSA client key exchange verification
12245 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
12248 *) Add various utility functions to handle SPKACs, these were previously
12249 handled by poking round in the structure internals. Added new function
12250 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
12251 print, verify and generate SPKACs. Based on an original idea from
12252 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
12255 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
12258 *) Allow the config file extension section to be overwritten on the
12259 command line. Based on an original idea from Massimiliano Pala
12260 <madwolf@comune.modena.it>. The new option is called -extensions
12261 and can be applied to ca, req and x509. Also -reqexts to override
12262 the request extensions in req and -crlexts to override the crl extensions
12266 *) Add new feature to the SPKAC handling in ca. Now you can include
12267 the same field multiple times by preceding it by "XXXX." for example:
12270 this is the same syntax as used in the req config file.
12273 *) Allow certificate extensions to be added to certificate requests. These
12274 are specified in a 'req_extensions' option of the req section of the
12275 config file. They can be printed out with the -text option to req but
12276 are otherwise ignored at present.
12279 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
12280 data read consists of only the final block it would not decrypted because
12281 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
12282 A misplaced 'break' also meant the decrypted final block might not be
12283 copied until the next read.
12286 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
12287 a few extra parameters to the DH structure: these will be useful if
12288 for example we want the value of 'q' or implement X9.42 DH.
12291 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
12292 provides hooks that allow the default DSA functions or functions on a
12293 "per key" basis to be replaced. This allows hardware acceleration and
12294 hardware key storage to be handled without major modification to the
12295 library. Also added low level modexp hooks and CRYPTO_EX structure and
12296 associated functions.
12299 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
12300 as "read only": it can't be written to and the buffer it points to will
12301 not be freed. Reading from a read only BIO is much more efficient than
12302 a normal memory BIO. This was added because there are several times when
12303 an area of memory needs to be read from a BIO. The previous method was
12304 to create a memory BIO and write the data to it, this results in two
12305 copies of the data and an O(n^2) reading algorithm. There is a new
12306 function BIO_new_mem_buf() which creates a read only memory BIO from
12307 an area of memory. Also modified the PKCS#7 routines to use read only
12311 *) Bugfix: ssl23_get_client_hello did not work properly when called in
12312 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
12313 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
12314 but a retry condition occurred while trying to read the rest.
12317 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
12318 NID_pkcs7_encrypted by default: this was wrong since this should almost
12319 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
12320 the encrypted data type: this is a more sensible place to put it and it
12321 allows the PKCS#12 code to be tidied up that duplicated this
12325 *) Changed obj_dat.pl script so it takes its input and output files on
12326 the command line. This should avoid shell escape redirection problems
12330 *) Initial support for certificate extension requests, these are included
12331 in things like Xenroll certificate requests. Included functions to allow
12332 extensions to be obtained and added.
12335 *) -crlf option to s_client and s_server for sending newlines as
12336 CRLF (as required by many protocols).
12339 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
12341 *) Install libRSAglue.a when OpenSSL is built with RSAref.
12342 [Ralf S. Engelschall]
12344 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
12345 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
12347 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
12351 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
12352 DH parameters/keys (q is lost during that conversion, but the resulting
12353 DH parameters contain its length).
12355 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
12356 much faster than DH_generate_parameters (which creates parameters
12357 where p = 2*q + 1), and also the smaller q makes DH computations
12358 much more efficient (160-bit exponentiation instead of 1024-bit
12359 exponentiation); so this provides a convenient way to support DHE
12360 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
12361 utter importance to use
12362 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12364 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12365 when such DH parameters are used, because otherwise small subgroup
12366 attacks may become possible!
12369 *) Avoid memory leak in i2d_DHparams.
12372 *) Allow the -k option to be used more than once in the enc program:
12373 this allows the same encrypted message to be read by multiple recipients.
12376 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
12377 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
12378 it will always use the numerical form of the OID, even if it has a short
12382 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
12383 method only got called if p,q,dmp1,dmq1,iqmp components were present,
12384 otherwise bn_mod_exp was called. In the case of hardware keys for example
12385 no private key components need be present and it might store extra data
12386 in the RSA structure, which cannot be accessed from bn_mod_exp.
12387 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
12388 private key operations.
12391 *) Added support for SPARC Linux.
12394 *) pem_password_cb function type incompatibly changed from
12395 typedef int pem_password_cb(char *buf, int size, int rwflag);
12397 ....(char *buf, int size, int rwflag, void *userdata);
12398 so that applications can pass data to their callbacks:
12399 The PEM[_ASN1]_{read,write}... functions and macros now take an
12400 additional void * argument, which is just handed through whenever
12401 the password callback is called.
12402 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
12404 New function SSL_CTX_set_default_passwd_cb_userdata.
12406 Compatibility note: As many C implementations push function arguments
12407 onto the stack in reverse order, the new library version is likely to
12408 interoperate with programs that have been compiled with the old
12409 pem_password_cb definition (PEM_whatever takes some data that
12410 happens to be on the stack as its last argument, and the callback
12411 just ignores this garbage); but there is no guarantee whatsoever that
12414 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
12415 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
12416 problems not only on Windows, but also on some Unix platforms.
12417 To avoid problematic command lines, these definitions are now in an
12418 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
12419 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
12422 *) MIPS III/IV assembler module is reimplemented.
12425 *) More DES library cleanups: remove references to srand/rand and
12426 delete an unused file.
12429 *) Add support for the free Netwide assembler (NASM) under Win32,
12430 since not many people have MASM (ml) and it can be hard to obtain.
12431 This is currently experimental but it seems to work OK and pass all
12432 the tests. Check out INSTALL.W32 for info.
12435 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
12436 without temporary keys kept an extra copy of the server key,
12437 and connections with temporary keys did not free everything in case
12441 *) New function RSA_check_key and new openssl rsa option -check
12442 for verifying the consistency of RSA keys.
12443 [Ulf Moeller, Bodo Moeller]
12445 *) Various changes to make Win32 compile work:
12446 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
12447 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
12448 comparison" warnings.
12449 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
12452 *) Add a debugging option to PKCS#5 v2 key generation function: when
12453 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
12454 derived keys are printed to stderr.
12457 *) Copy the flags in ASN1_STRING_dup().
12458 [Roman E. Pavlov <pre@mo.msk.ru>]
12460 *) The x509 application mishandled signing requests containing DSA
12461 keys when the signing key was also DSA and the parameters didn't match.
12463 It was supposed to omit the parameters when they matched the signing key:
12464 the verifying software was then supposed to automatically use the CA's
12465 parameters if they were absent from the end user certificate.
12467 Omitting parameters is no longer recommended. The test was also
12468 the wrong way round! This was probably due to unusual behaviour in
12469 EVP_cmp_parameters() which returns 1 if the parameters match.
12470 This meant that parameters were omitted when they *didn't* match and
12471 the certificate was useless. Certificates signed with 'ca' didn't have
12473 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
12475 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
12476 The interface is as follows:
12477 Applications can use
12478 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
12479 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
12480 "off" is now the default.
12481 The library internally uses
12482 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
12483 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
12484 to disable memory-checking temporarily.
12486 Some inconsistent states that previously were possible (and were
12487 even the default) are now avoided.
12489 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
12490 with each memory chunk allocated; this is occasionally more helpful
12491 than just having a counter.
12493 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
12495 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
12499 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
12500 which largely parallels "options", but is for changing API behaviour,
12501 whereas "options" are about protocol behaviour.
12502 Initial "mode" flags are:
12504 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
12505 a single record has been written.
12506 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
12507 retries use the same buffer location.
12508 (But all of the contents must be
12512 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
12515 *) Fix problems with no-hmac etc.
12516 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
12518 *) New functions RSA_get_default_method(), RSA_set_method() and
12519 RSA_get_method(). These allows replacement of RSA_METHODs without having
12520 to mess around with the internals of an RSA structure.
12523 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
12524 Also really enable memory leak checks in openssl.c and in some
12526 [Chad C. Mulligan, Bodo Moeller]
12528 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
12529 up the length of negative integers. This has now been simplified to just
12530 store the length when it is first determined and use it later, rather
12531 than trying to keep track of where data is copied and updating it to
12533 [Steve Henson, reported by Brien Wheeler
12534 <bwheeler@authentica-security.com>]
12536 *) Add a new function PKCS7_signatureVerify. This allows the verification
12537 of a PKCS#7 signature but with the signing certificate passed to the
12538 function itself. This contrasts with PKCS7_dataVerify which assumes the
12539 certificate is present in the PKCS#7 structure. This isn't always the
12540 case: certificates can be omitted from a PKCS#7 structure and be
12541 distributed by "out of band" means (such as a certificate database).
12544 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12545 function prototypes in pem.h, also change util/mkdef.pl to add the
12546 necessary function names.
12549 *) mk1mf.pl (used by Windows builds) did not properly read the
12550 options set by Configure in the top level Makefile, and Configure
12551 was not even able to write more than one option correctly.
12552 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12555 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12556 file to be loaded from a BIO or FILE pointer. The BIO version will
12557 for example allow memory BIOs to contain config info.
12560 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12561 Whoever hopes to achieve shared-library compatibility across versions
12562 must use this, not the compile-time macro.
12563 (Exercise 0.9.4: Which is the minimum library version required by
12565 Note: All this applies only to multi-threaded programs, others don't
12569 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12570 through a BIO pair triggered the default case, i.e.
12571 SSLerr(...,SSL_R_UNKNOWN_STATE).
12574 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12575 can use the SSL library even if none of the specific BIOs is
12579 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12580 for the encoded length.
12581 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12583 *) Add initial documentation of the X509V3 functions.
12586 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12587 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12588 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12589 secure PKCS#8 private key format with a high iteration count.
12592 *) Fix determination of Perl interpreter: A perl or perl5
12593 _directory_ in $PATH was also accepted as the interpreter.
12594 [Ralf S. Engelschall]
12596 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12597 wrong with it but it was very old and did things like calling
12598 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12599 unusual formatting.
12602 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12603 to use the new extension code.
12606 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12607 with macros. This should make it easier to change their form, add extra
12608 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12612 *) Add to configuration table a new entry that can specify an alternative
12613 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12614 according to Mark Crispin <MRC@Panda.COM>.
12618 *) DES CBC did not update the IV. Weird.
12621 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12622 Changing the behaviour of the former might break existing programs --
12623 where IV updating is needed, des_ncbc_encrypt can be used.
12626 *) When bntest is run from "make test" it drives bc to check its
12627 calculations, as well as internally checking them. If an internal check
12628 fails, it needs to cause bc to give a non-zero result or make test carries
12629 on without noticing the failure. Fixed.
12632 *) DES library cleanups.
12635 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12636 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12637 ciphers. NOTE: although the key derivation function has been verified
12638 against some published test vectors it has not been extensively tested
12639 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12643 *) Instead of "mkdir -p", which is not fully portable, use new
12644 Perl script "util/mkdir-p.pl".
12647 *) Rewrite the way password based encryption (PBE) is handled. It used to
12648 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12649 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12650 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12651 the 'parameter' field of the AlgorithmIdentifier is passed to the
12652 underlying key generation function so it must do its own ASN1 parsing.
12653 This has also changed the EVP_PBE_CipherInit() function which now has a
12654 'parameter' argument instead of literal salt and iteration count values
12655 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12658 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12659 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12660 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12661 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12662 value was just used as a "magic string" and not used directly its
12663 value doesn't matter.
12666 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12670 *) "linux-sparc64" configuration (ultrapenguin).
12671 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12672 "linux-sparc" configuration.
12673 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12675 *) config now generates no-xxx options for missing ciphers.
12678 *) Support the EBCDIC character set (work in progress).
12679 File ebcdic.c not yet included because it has a different license.
12680 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12682 *) Support BS2000/OSD-POSIX.
12683 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12685 *) Make callbacks for key generation use void * instead of char *.
12688 *) Make S/MIME samples compile (not yet tested).
12691 *) Additional typesafe stacks.
12694 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12698 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12700 *) New configuration variant "sco5-gcc".
12702 *) Updated some demos.
12703 [Sean O Riordain, Wade Scholine]
12705 *) Add missing BIO_free at exit of pkcs12 application.
12708 *) Fix memory leak in conf.c.
12711 *) Updates for Win32 to assembler version of MD5.
12714 *) Set #! path to perl in apps/der_chop to where we found it
12715 instead of using a fixed path.
12718 *) SHA library changes for irix64-mips4-cc.
12721 *) Improvements for VMS support.
12725 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12727 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12728 This also avoids the problems with SC4.2 and unpatched SC5.
12729 [Andy Polyakov <appro@fy.chalmers.se>]
12731 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12732 These are required because of the typesafe stack would otherwise break
12733 existing code. If old code used a structure member which used to be STACK
12734 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12735 sk_num or sk_value it would produce an error because the num, data members
12736 are not present in STACK_OF. Now it just produces a warning. sk_set
12737 replaces the old method of assigning a value to sk_value
12738 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12739 that does this will no longer work (and should use sk_set instead) but
12740 this could be regarded as a "questionable" behaviour anyway.
12743 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12744 correctly handle encrypted S/MIME data.
12747 *) Change type of various DES function arguments from des_cblock
12748 (which means, in function argument declarations, pointer to char)
12749 to des_cblock * (meaning pointer to array with 8 char elements),
12750 which allows the compiler to do more typechecking; it was like
12751 that back in SSLeay, but with lots of ugly casts.
12753 Introduce new type const_des_cblock.
12756 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12757 problems: find RecipientInfo structure that matches recipient certificate
12758 and initialise the ASN1 structures properly based on passed cipher.
12761 *) Belatedly make the BN tests actually check the results.
12764 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12765 to and from BNs: it was completely broken. New compilation option
12766 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12767 key elements as negative integers.
12770 *) Reorganize and speed up MD5.
12771 [Andy Polyakov <appro@fy.chalmers.se>]
12774 [Richard Levitte <richard@levitte.org>]
12776 *) New option -out to asn1parse to allow the parsed structure to be
12777 output to a file. This is most useful when combined with the -strparse
12778 option to examine the output of things like OCTET STRINGS.
12781 *) Make SSL library a little more fool-proof by not requiring any longer
12782 that SSL_set_{accept,connect}_state be called before
12783 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12784 in many applications because usually everything *appeared* to work as
12785 intended anyway -- now it really works as intended).
12788 *) Move openssl.cnf out of lib/.
12791 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12792 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12793 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12794 [Ralf S. Engelschall]
12796 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12797 handle PKCS#7 enveloped data properly.
12798 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12800 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12801 copying pointers. The cert_st handling is changed by this in
12802 various ways (and thus what used to be known as ctx->default_cert
12803 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12804 any longer when s->cert does not give us what we need).
12805 ssl_cert_instantiate becomes obsolete by this change.
12806 As soon as we've got the new code right (possibly it already is?),
12807 we have solved a couple of bugs of the earlier code where s->cert
12808 was used as if it could not have been shared with other SSL structures.
12810 Note that using the SSL API in certain dirty ways now will result
12811 in different behaviour than observed with earlier library versions:
12812 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12813 does not influence s as it used to.
12815 In order to clean up things more thoroughly, inside SSL_SESSION
12816 we don't use CERT any longer, but a new structure SESS_CERT
12817 that holds per-session data (if available); currently, this is
12818 the peer's certificate chain and, for clients, the server's certificate
12819 and temporary key. CERT holds only those values that can have
12820 meaningful defaults in an SSL_CTX.
12823 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12824 from the internal representation. Various PKCS#7 fixes: remove some
12825 evil casts and set the enc_dig_alg field properly based on the signing
12829 *) Allow PKCS#12 password to be set from the command line or the
12830 environment. Let 'ca' get its config file name from the environment
12831 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12835 *) Allow certificate policies extension to use an IA5STRING for the
12836 organization field. This is contrary to the PKIX definition but
12837 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12841 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12842 without disallowing inline assembler and the like for non-pedantic builds.
12845 *) Support Borland C++ builder.
12846 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12848 *) Support Mingw32.
12851 *) SHA-1 cleanups and performance enhancements.
12852 [Andy Polyakov <appro@fy.chalmers.se>]
12854 *) Sparc v8plus assembler for the bignum library.
12855 [Andy Polyakov <appro@fy.chalmers.se>]
12857 *) Accept any -xxx and +xxx compiler options in Configure.
12860 *) Update HPUX configuration.
12863 *) Add missing sk_<type>_unshift() function to safestack.h
12864 [Ralf S. Engelschall]
12866 *) New function SSL_CTX_use_certificate_chain_file that sets the
12867 "extra_cert"s in addition to the certificate. (This makes sense
12868 only for "PEM" format files, as chains as a whole are not
12872 *) Support verify_depth from the SSL API.
12873 x509_vfy.c had what can be considered an off-by-one-error:
12874 Its depth (which was not part of the external interface)
12875 was actually counting the number of certificates in a chain;
12876 now it really counts the depth.
12879 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12880 instead of X509err, which often resulted in confusing error
12881 messages since the error codes are not globally unique
12882 (e.g. an alleged error in ssl3_accept when a certificate
12883 didn't match the private key).
12885 *) New function SSL_CTX_set_session_id_context that allows to set a default
12886 value (so that you don't need SSL_set_session_id_context for each
12887 connection using the SSL_CTX).
12890 *) OAEP decoding bug fix.
12893 *) Support INSTALL_PREFIX for package builders, as proposed by
12897 *) New Configure options "threads" and "no-threads". For systems
12898 where the proper compiler options are known (currently Solaris
12899 and Linux), "threads" is the default.
12902 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12905 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12906 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12907 such as /usr/local/bin.
12910 *) "make linux-shared" to build shared libraries.
12911 [Niels Poppe <niels@netbox.org>]
12913 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12916 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12917 extension adding in x509 utility.
12920 *) Remove NOPROTO sections and error code comments.
12923 *) Partial rewrite of the DEF file generator to now parse the ANSI
12927 *) New Configure options --prefix=DIR and --openssldir=DIR.
12930 *) Complete rewrite of the error code script(s). It is all now handled
12931 by one script at the top level which handles error code gathering,
12932 header rewriting and C source file generation. It should be much better
12933 than the old method: it now uses a modified version of Ulf's parser to
12934 read the ANSI prototypes in all header files (thus the old K&R definitions
12935 aren't needed for error creation any more) and do a better job of
12936 translating function codes into names. The old 'ASN1 error code embedded
12937 in a comment' is no longer necessary and it doesn't use .err files which
12938 have now been deleted. Also the error code call doesn't have to appear all
12939 on one line (which resulted in some large lines...).
12942 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12945 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12946 0 (which usually indicates a closed connection), but continue reading.
12949 *) Fix some race conditions.
12952 *) Add support for CRL distribution points extension. Add Certificate
12953 Policies and CRL distribution points documentation.
12956 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12959 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12960 8 of keying material. Merlin has also confirmed interop with this fix
12961 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12962 [Merlin Hughes <merlin@baltimore.ie>]
12964 *) Fix lots of warnings.
12965 [Richard Levitte <levitte@stacken.kth.se>]
12967 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12968 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12969 [Richard Levitte <levitte@stacken.kth.se>]
12971 *) Fix problems with sizeof(long) == 8.
12972 [Andy Polyakov <appro@fy.chalmers.se>]
12974 *) Change functions to ANSI C.
12977 *) Fix typos in error codes.
12978 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12980 *) Remove defunct assembler files from Configure.
12983 *) SPARC v8 assembler BIGNUM implementation.
12984 [Andy Polyakov <appro@fy.chalmers.se>]
12986 *) Support for Certificate Policies extension: both print and set.
12987 Various additions to support the r2i method this uses.
12990 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12991 return a const string when you are expecting an allocated buffer.
12994 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12995 types DirectoryString and DisplayText.
12998 *) Add code to allow r2i extensions to access the configuration database,
12999 add an LHASH database driver and add several ctx helper functions.
13002 *) Fix an evil bug in bn_expand2() which caused various BN functions to
13003 fail when they extended the size of a BIGNUM.
13006 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
13007 support typesafe stack.
13010 *) Fix typo in SSL_[gs]et_options().
13011 [Nils Frostberg <nils@medcom.se>]
13013 *) Delete various functions and files that belonged to the (now obsolete)
13014 old X509V3 handling code.
13017 *) New Configure option "rsaref".
13020 *) Don't auto-generate pem.h.
13023 *) Introduce type-safe ASN.1 SETs.
13026 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
13027 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
13029 *) Introduce type-safe STACKs. This will almost certainly break lots of code
13030 that links with OpenSSL (well at least cause lots of warnings), but fear
13031 not: the conversion is trivial, and it eliminates loads of evil casts. A
13032 few STACKed things have been converted already. Feel free to convert more.
13033 In the fullness of time, I'll do away with the STACK type altogether.
13036 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
13037 specified in <certfile> by updating the entry in the index.txt file.
13038 This way one no longer has to edit the index.txt file manually for
13039 revoking a certificate. The -revoke option does the gory details now.
13040 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
13042 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
13043 `-text' option at all and this way the `-noout -text' combination was
13044 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
13045 [Ralf S. Engelschall]
13047 *) Make sure a corresponding plain text error message exists for the
13048 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
13049 verify callback function determined that a certificate was revoked.
13050 [Ralf S. Engelschall]
13052 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
13053 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
13054 all available ciphers including rc5, which was forgotten until now.
13055 In order to let the testing shell script know which algorithms
13056 are available, a new (up to now undocumented) command
13057 "openssl list-cipher-commands" is used.
13060 *) Bugfix: s_client occasionally would sleep in select() when
13061 it should have checked SSL_pending() first.
13064 *) New functions DSA_do_sign and DSA_do_verify to provide access to
13065 the raw DSA values prior to ASN.1 encoding.
13068 *) Tweaks to Configure
13069 [Niels Poppe <niels@netbox.org>]
13071 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
13075 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
13078 *) New config option to avoid instructions that are illegal on the 80386.
13079 The default code is faster, but requires at least a 486.
13082 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
13083 SSL2_SERVER_VERSION (not used at all) macros, which are now the
13084 same as SSL2_VERSION anyway.
13087 *) New "-showcerts" option for s_client.
13090 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
13091 application. Various cleanups and fixes.
13094 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
13095 modify error routines to work internally. Add error codes and PBE init
13096 to library startup routines.
13099 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
13100 packing functions to asn1 and evp. Changed function names and error
13101 codes along the way.
13104 *) PKCS12 integration: and so it begins... First of several patches to
13105 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
13106 objects to objects.h
13109 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
13110 and display support for Thawte strong extranet extension.
13113 *) Add LinuxPPC support.
13114 [Jeff Dubrule <igor@pobox.org>]
13116 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
13117 bn_div_words in alpha.s.
13118 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
13120 *) Make sure the RSA OAEP test is skipped under -DRSAref because
13121 OAEP isn't supported when OpenSSL is built with RSAref.
13122 [Ulf Moeller <ulf@fitug.de>]
13124 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
13125 so they no longer are missing under -DNOPROTO.
13126 [Soren S. Jorvang <soren@t.dk>]
13129 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
13131 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
13132 doesn't work when the session is reused. Coming soon!
13135 *) Fix a security hole, that allows sessions to be reused in the wrong
13136 context thus bypassing client cert protection! All software that uses
13137 client certs and session caches in multiple contexts NEEDS PATCHING to
13138 allow session reuse! A fuller solution is in the works.
13139 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
13141 *) Some more source tree cleanups (removed obsolete files
13142 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
13143 permission on "config" script to be executable) and a fix for the INSTALL
13145 [Ulf Moeller <ulf@fitug.de>]
13147 *) Remove some legacy and erroneous uses of malloc, free instead of
13149 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
13151 *) Make rsa_oaep_test return non-zero on error.
13152 [Ulf Moeller <ulf@fitug.de>]
13154 *) Add support for native Solaris shared libraries. Configure
13155 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
13156 if someone would make that last step automatic.
13157 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
13159 *) ctx_size was not built with the right compiler during "make links". Fixed.
13162 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
13163 except NULL ciphers". This means the default cipher list will no longer
13164 enable NULL ciphers. They need to be specifically enabled e.g. with
13165 the string "DEFAULT:eNULL".
13168 *) Fix to RSA private encryption routines: if p < q then it would
13169 occasionally produce an invalid result. This will only happen with
13170 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
13173 *) Be less restrictive and allow also `perl util/perlpath.pl
13174 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
13175 because this way one can also use an interpreter named `perl5' (which is
13176 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
13177 installed as `perl').
13178 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13180 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
13181 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13183 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
13184 advapi32.lib to Win32 build and change the pem test comparison
13185 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
13186 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
13187 and crypto/des/ede_cbcm_enc.c.
13190 *) DES quad checksum was broken on big-endian architectures. Fixed.
13193 *) Comment out two functions in bio.h that aren't implemented. Fix up the
13194 Win32 test batch file so it (might) work again. The Win32 test batch file
13195 is horrible: I feel ill....
13198 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
13199 in e_os.h. Audit of header files to check ANSI and non ANSI
13200 sections: 10 functions were absent from non ANSI section and not exported
13201 from Windows DLLs. Fixed up libeay.num for new functions.
13204 *) Make `openssl version' output lines consistent.
13205 [Ralf S. Engelschall]
13207 *) Fix Win32 symbol export lists for BIO functions: Added
13208 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
13209 to ms/libeay{16,32}.def.
13210 [Ralf S. Engelschall]
13212 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
13213 fine under Unix and passes some trivial tests I've now added. But the
13214 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
13215 added to make sure no one expects that this stuff really works in the
13216 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
13217 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
13219 [Ralf S. Engelschall]
13221 *) Fix the generation of two part addresses in perl.
13222 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
13224 *) Add config entry for Linux on MIPS.
13225 [John Tobey <jtobey@channel1.com>]
13227 *) Make links whenever Configure is run, unless we are on Windoze.
13230 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
13231 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
13235 *) Add a useful kludge to allow package maintainers to specify compiler and
13236 other platforms details on the command line without having to patch the
13237 Configure script every time: One now can use ``perl Configure
13238 <id>:<details>'', i.e. platform ids are allowed to have details appended
13239 to them (separated by colons). This is treated as there would be a static
13240 pre-configured entry in Configure's %table under key <id> with value
13241 <details> and ``perl Configure <id>'' is called. So, when you want to
13242 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
13243 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
13244 now, which overrides the FreeBSD-elf entry on-the-fly.
13245 [Ralf S. Engelschall]
13247 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
13250 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
13251 on the `perl Configure ...' command line. This way one can compile
13252 OpenSSL libraries with Position Independent Code (PIC) which is needed
13253 for linking it into DSOs.
13254 [Ralf S. Engelschall]
13256 *) Remarkably, export ciphers were totally broken and no-one had noticed!
13260 *) Cleaned up the LICENSE document: The official contact for any license
13261 questions now is the OpenSSL core team under openssl-core@openssl.org.
13262 And add a paragraph about the dual-license situation to make sure people
13263 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
13264 to the OpenSSL toolkit.
13265 [Ralf S. Engelschall]
13267 *) General source tree makefile cleanups: Made `making xxx in yyy...'
13268 display consistent in the source tree and replaced `/bin/rm' by `rm'.
13269 Additionally cleaned up the `make links' target: Remove unnecessary
13270 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
13271 to speed processing and no longer clutter the display with confusing
13272 stuff. Instead only the actually done links are displayed.
13273 [Ralf S. Engelschall]
13275 *) Permit null encryption ciphersuites, used for authentication only. It used
13276 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
13277 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
13281 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
13282 signed attributes when verifying signatures (this would break them),
13283 the detached data encoding was wrong and public keys obtained using
13284 X509_get_pubkey() weren't freed.
13287 *) Add text documentation for the BUFFER functions. Also added a work around
13288 to a Win95 console bug. This was triggered by the password read stuff: the
13289 last character typed gets carried over to the next fread(). If you were
13290 generating a new cert request using 'req' for example then the last
13291 character of the passphrase would be CR which would then enter the first
13295 *) Added the new `Includes OpenSSL Cryptography Software' button as
13296 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
13297 button and can be used by applications based on OpenSSL to show the
13298 relationship to the OpenSSL project.
13299 [Ralf S. Engelschall]
13301 *) Remove confusing variables in function signatures in files
13302 ssl/ssl_lib.c and ssl/ssl.h.
13303 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13305 *) Don't install bss_file.c under PREFIX/include/
13306 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13308 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
13309 functions that return function pointers and has support for NT specific
13310 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
13311 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
13312 unsigned to signed types: this was killing the Win32 compile.
13315 *) Add new certificate file to stack functions,
13316 SSL_add_dir_cert_subjects_to_stack() and
13317 SSL_add_file_cert_subjects_to_stack(). These largely supplant
13318 SSL_load_client_CA_file(), and can be used to add multiple certs easily
13319 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
13320 This means that Apache-SSL and similar packages don't have to mess around
13321 to add as many CAs as they want to the preferred list.
13324 *) Experiment with doxygen documentation. Currently only partially applied to
13326 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
13327 openssl.doxy as the configuration file.
13330 *) Get rid of remaining C++-style comments which strict C compilers hate.
13331 [Ralf S. Engelschall, pointed out by Carlos Amengual]
13333 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
13334 compiled in by default: it has problems with large keys.
13337 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
13338 DH private keys and/or callback functions which directly correspond to
13339 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
13340 is needed for applications which have to configure certificates on a
13341 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
13343 For the RSA certificate situation is makes no difference, but
13344 for the DSA certificate situation this fixes the "no shared cipher"
13345 problem where the OpenSSL cipher selection procedure failed because the
13346 temporary keys were not overtaken from the context and the API provided
13347 no way to reconfigure them.
13348 The new functions now let applications reconfigure the stuff and they
13349 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
13350 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
13351 non-public-API function ssl_cert_instantiate() is used as a helper
13352 function and also to reduce code redundancy inside ssl_rsa.c.
13353 [Ralf S. Engelschall]
13355 *) Move s_server -dcert and -dkey options out of the undocumented feature
13356 area because they are useful for the DSA situation and should be
13357 recognized by the users.
13358 [Ralf S. Engelschall]
13360 *) Fix the cipher decision scheme for export ciphers: the export bits are
13361 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
13362 SSL_EXP_MASK. So, the original variable has to be used instead of the
13363 already masked variable.
13364 [Richard Levitte <levitte@stacken.kth.se>]
13366 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
13367 [Richard Levitte <levitte@stacken.kth.se>]
13369 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
13370 from `int' to `unsigned int' because it's a length and initialized by
13371 EVP_DigestFinal() which expects an `unsigned int *'.
13372 [Richard Levitte <levitte@stacken.kth.se>]
13374 *) Don't hard-code path to Perl interpreter on shebang line of Configure
13375 script. Instead use the usual Shell->Perl transition trick.
13376 [Ralf S. Engelschall]
13378 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
13379 (in addition to RSA certificates) to match the behaviour of `openssl dsa
13380 -noout -modulus' as it's already the case for `openssl rsa -noout
13381 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
13382 currently the public key is printed (a decision which was already done by
13383 `openssl dsa -modulus' in the past) which serves a similar purpose.
13384 Additionally the NO_RSA no longer completely removes the whole -modulus
13385 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
13387 [Ralf S. Engelschall]
13389 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
13390 BIO. See the source (crypto/evp/bio_ok.c) for more info.
13391 [Arne Ansper <arne@ats.cyber.ee>]
13393 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
13394 to be added. Now both 'req' and 'ca' can use new objects defined in the
13398 *) Add cool BIO that does syslog (or event log on NT).
13399 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
13401 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
13402 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
13403 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
13404 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
13407 *) Add preliminary config info for new extension code.
13410 *) Make RSA_NO_PADDING really use no padding.
13411 [Ulf Moeller <ulf@fitug.de>]
13413 *) Generate errors when private/public key check is done.
13416 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
13417 for some CRL extensions and new objects added.
13420 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
13421 key usage extension and fuller support for authority key id.
13424 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
13425 padding method for RSA, which is recommended for new applications in PKCS
13426 #1 v2.0 (RFC 2437, October 1998).
13427 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
13428 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
13429 against Bleichbacher's attack on RSA.
13430 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
13433 *) Updates to the new SSL compression code
13434 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13436 *) Fix so that the version number in the master secret, when passed
13437 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
13438 (because the server will not accept higher), that the version number
13439 is 0x03,0x01, not 0x03,0x00
13440 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13442 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
13443 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
13444 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
13447 *) Support for RAW extensions where an arbitrary extension can be
13448 created by including its DER encoding. See apps/openssl.cnf for
13452 *) Make sure latest Perl versions don't interpret some generated C array
13453 code as Perl array code in the crypto/err/err_genc.pl script.
13454 [Lars Weber <3weber@informatik.uni-hamburg.de>]
13456 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
13457 not many people have the assembler. Various Win32 compilation fixes and
13458 update to the INSTALL.W32 file with (hopefully) more accurate Win32
13459 build instructions.
13462 *) Modify configure script 'Configure' to automatically create crypto/date.h
13463 file under Win32 and also build pem.h from pem.org. New script
13464 util/mkfiles.pl to create the MINFO file on environments that can't do a
13465 'make files': perl util/mkfiles.pl >MINFO should work.
13468 *) Major rework of DES function declarations, in the pursuit of correctness
13469 and purity. As a result, many evil casts evaporated, and some weirdness,
13470 too. You may find this causes warnings in your code. Zapping your evil
13471 casts will probably fix them. Mostly.
13474 *) Fix for a typo in asn1.h. Bug fix to object creation script
13475 obj_dat.pl. It considered a zero in an object definition to mean
13476 "end of object": none of the objects in objects.h have any zeros
13477 so it wasn't spotted.
13478 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
13480 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
13481 Masking (CBCM). In the absence of test vectors, the best I have been able
13482 to do is check that the decrypt undoes the encrypt, so far. Send me test
13483 vectors if you have them.
13486 *) Correct calculation of key length for export ciphers (too much space was
13487 allocated for null ciphers). This has not been tested!
13490 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
13491 message is now correct (it understands "crypto" and "ssl" on its
13492 command line). There is also now an "update" option. This will update
13493 the util/ssleay.num and util/libeay.num files with any new functions.
13495 perl util/mkdef.pl crypto ssl update
13496 it will update them.
13499 *) Overhauled the Perl interface (perl/*):
13500 - ported BN stuff to OpenSSL's different BN library
13501 - made the perl/ source tree CVS-aware
13502 - renamed the package from SSLeay to OpenSSL (the files still contain
13503 their history because I've copied them in the repository)
13504 - removed obsolete files (the test scripts will be replaced
13505 by better Test::Harness variants in the future)
13506 [Ralf S. Engelschall]
13508 *) First cut for a very conservative source tree cleanup:
13509 1. merge various obsolete readme texts into doc/ssleay.txt
13510 where we collect the old documents and readme texts.
13511 2. remove the first part of files where I'm already sure that we no
13512 longer need them because of three reasons: either they are just temporary
13513 files which were left by Eric or they are preserved original files where
13514 I've verified that the diff is also available in the CVS via "cvs diff
13515 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
13516 the crypto/md/ stuff).
13517 [Ralf S. Engelschall]
13519 *) More extension code. Incomplete support for subject and issuer alt
13520 name, issuer and authority key id. Change the i2v function parameters
13521 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
13522 what that's for :-) Fix to ASN1 macro which messed up
13523 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
13526 *) Preliminary support for ENUMERATED type. This is largely copied from the
13530 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
13531 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13533 *) Make sure `make rehash' target really finds the `openssl' program.
13534 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13536 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
13537 like to hear about it if this slows down other processors.
13540 *) Add CygWin32 platform information to Configure script.
13541 [Alan Batie <batie@aahz.jf.intel.com>]
13543 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
13544 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13546 *) New program nseq to manipulate netscape certificate sequences
13549 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13553 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13554 but the BN code had some problems that would cause failures when
13555 doing certificate verification and some other functions.
13556 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13558 *) Add ASN1 and PEM code to support netscape certificate sequences.
13561 *) Add ASN1 and PEM code to support netscape certificate sequences.
13564 *) Add several PKIX and private extended key usage OIDs.
13567 *) Modify the 'ca' program to handle the new extension code. Modify
13568 openssl.cnf for new extension format, add comments.
13571 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13572 and add a sample to openssl.cnf so req -x509 now adds appropriate
13576 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13577 error code, add initial support to X509_print() and x509 application.
13580 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13581 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13582 stuff is currently isolated and isn't even compiled yet.
13585 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13586 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13587 Removed the versions check from X509 routines when loading extensions:
13588 this allows certain broken certificates that don't set the version
13589 properly to be processed.
13592 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13593 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13594 can still be regenerated with "make depend".
13597 *) Spelling mistake in C version of CAST-128.
13598 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13600 *) Changes to the error generation code. The perl script err-code.pl
13601 now reads in the old error codes and retains the old numbers, only
13602 adding new ones if necessary. It also only changes the .err files if new
13603 codes are added. The makefiles have been modified to only insert errors
13604 when needed (to avoid needlessly modifying header files). This is done
13605 by only inserting errors if the .err file is newer than the auto generated
13606 C file. To rebuild all the error codes from scratch (the old behaviour)
13607 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13608 or delete all the .err files.
13611 *) CAST-128 was incorrectly implemented for short keys. The C version has
13612 been fixed, but is untested. The assembler versions are also fixed, but
13613 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13614 to regenerate it if needed.
13615 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13616 Hagino <itojun@kame.net>]
13618 *) File was opened incorrectly in randfile.c.
13619 [Ulf Möller <ulf@fitug.de>]
13621 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13622 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13623 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13624 al: it's just almost always a UTCTime. Note this patch adds new error
13625 codes so do a "make errors" if there are problems.
13628 *) Correct Linux 1 recognition in config.
13629 [Ulf Möller <ulf@fitug.de>]
13631 *) Remove pointless MD5 hash when using DSA keys in ca.
13632 [Anonymous <nobody@replay.com>]
13634 *) Generate an error if given an empty string as a cert directory. Also
13635 generate an error if handed NULL (previously returned 0 to indicate an
13636 error, but didn't set one).
13637 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13639 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13642 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13643 parameters. This was causing a warning which killed off the Win32 compile.
13646 *) Remove C++ style comments from crypto/bn/bn_local.h.
13647 [Neil Costigan <neil.costigan@celocom.com>]
13649 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13650 based on a text string, looking up short and long names and finally
13651 "dot" format. The "dot" format stuff didn't work. Added new function
13652 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13653 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13654 OID is not part of the table.
13657 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13658 X509_LOOKUP_by_alias().
13661 *) Sort openssl functions by name.
13664 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13665 encryption from sample DSA keys (in case anyone is interested the password
13669 *) Make _all_ *_free functions accept a NULL pointer.
13670 [Frans Heymans <fheymans@isaserver.be>]
13672 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13674 [Anonymous <nobody@replay.com>]
13676 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13677 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13679 *) Don't blow it for numeric -newkey arguments to apps/req.
13680 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13682 *) Temp key "for export" tests were wrong in s3_srvr.c.
13683 [Anonymous <nobody@replay.com>]
13685 *) Add prototype for temp key callback functions
13686 SSL_CTX_set_tmp_{rsa,dh}_callback().
13689 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13690 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13693 *) X509_name_add_entry() freed the wrong thing after an error.
13694 [Arne Ansper <arne@ats.cyber.ee>]
13696 *) rsa_eay.c would attempt to free a NULL context.
13697 [Arne Ansper <arne@ats.cyber.ee>]
13699 *) BIO_s_socket() had a broken should_retry() on Windoze.
13700 [Arne Ansper <arne@ats.cyber.ee>]
13702 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13703 [Arne Ansper <arne@ats.cyber.ee>]
13705 *) Make sure the already existing X509_STORE->depth variable is initialized
13706 in X509_STORE_new(), but document the fact that this variable is still
13707 unused in the certificate verification process.
13708 [Ralf S. Engelschall]
13710 *) Fix the various library and apps files to free up pkeys obtained from
13711 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13714 *) Fix reference counting in X509_PUBKEY_get(). This makes
13715 demos/maurice/example2.c work, amongst others, probably.
13716 [Steve Henson and Ben Laurie]
13718 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13719 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13720 are no longer created. This way we have a single and consistent command
13721 line interface `openssl <command>', similar to `cvs <command>'.
13722 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13724 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13725 BIT STRING wrapper always have zero unused bits.
13728 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13731 *) Make the top-level INSTALL documentation easier to understand.
13734 *) Makefiles updated to exit if an error occurs in a sub-directory
13735 make (including if user presses ^C) [Paul Sutton]
13737 *) Make Montgomery context stuff explicit in RSA data structure.
13740 *) Fix build order of pem and err to allow for generated pem.h.
13743 *) Fix renumbering bug in X509_NAME_delete_entry().
13746 *) Enhanced the err-ins.pl script so it makes the error library number
13747 global and can add a library name. This is needed for external ASN1 and
13748 other error libraries.
13751 *) Fixed sk_insert which never worked properly.
13754 *) Fix ASN1 macros so they can handle indefinite length constructed
13755 EXPLICIT tags. Some non standard certificates use these: they can now
13759 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13760 into a single doc/ssleay.txt bundle. This way the information is still
13761 preserved but no longer messes up this directory. Now it's new room for
13762 the new set of documentation files.
13763 [Ralf S. Engelschall]
13765 *) SETs were incorrectly DER encoded. This was a major pain, because they
13766 shared code with SEQUENCEs, which aren't coded the same. This means that
13767 almost everything to do with SETs or SEQUENCEs has either changed name or
13768 number of arguments.
13769 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13771 *) Fix test data to work with the above.
13774 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13775 was already fixed by Eric for 0.9.1 it seems.
13776 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13778 *) Autodetect FreeBSD3.
13781 *) Fix various bugs in Configure. This affects the following platforms:
13785 unixware-2.0-pentium
13789 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13790 before they are needed.
13793 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13797 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13799 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13800 changed SSLeay to OpenSSL in version strings.
13801 [Ralf S. Engelschall]
13803 *) Some fixups to the top-level documents.
13806 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13807 because the symlink to include/ was missing.
13808 [Ralf S. Engelschall]
13810 *) Incorporated the popular no-RSA/DSA-only patches
13811 which allow to compile a RSA-free SSLeay.
13812 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13814 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13815 when "ssleay" is still not found.
13816 [Ralf S. Engelschall]
13818 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13819 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13821 *) Updated the README file.
13822 [Ralf S. Engelschall]
13824 *) Added various .cvsignore files in the CVS repository subdirs
13825 to make a "cvs update" really silent.
13826 [Ralf S. Engelschall]
13828 *) Recompiled the error-definition header files and added
13829 missing symbols to the Win32 linker tables.
13830 [Ralf S. Engelschall]
13832 *) Cleaned up the top-level documents;
13833 o new files: CHANGES and LICENSE
13834 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13835 o merged COPYRIGHT into LICENSE
13836 o removed obsolete TODO file
13837 o renamed MICROSOFT to INSTALL.W32
13838 [Ralf S. Engelschall]
13840 *) Removed dummy files from the 0.9.1b source tree:
13841 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13842 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13843 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13844 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13845 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13846 [Ralf S. Engelschall]
13848 *) Added various platform portability fixes.
13851 *) The Genesis of the OpenSSL rpject:
13852 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13853 Young and Tim J. Hudson created while they were working for C2Net until
13855 [The OpenSSL Project]
13858 Changes between 0.9.0b and 0.9.1b [not released]
13860 *) Updated a few CA certificates under certs/
13863 *) Changed some BIGNUM api stuff.
13866 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13867 DGUX x86, Linux Alpha, etc.
13870 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13871 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13875 *) Add -strparse option to asn1pars program which parses nested
13877 [Dr Stephen Henson <shenson@bigfoot.com>]
13879 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13882 *) DSA fix for "ca" program.
13885 *) Added "-genkey" option to "dsaparam" program.
13888 *) Added RIPE MD160 (rmd160) message digest.
13891 *) Added -a (all) option to "ssleay version" command.
13894 *) Added PLATFORM define which is the id given to Configure.
13897 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13900 *) Extended the ASN.1 parser routines.
13903 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13906 *) Added a BN_CTX to the BN library.
13909 *) Fixed the weak key values in DES library
13912 *) Changed API in EVP library for cipher aliases.
13915 *) Added support for RC2/64bit cipher.
13918 *) Converted the lhash library to the crypto/mem.c functions.
13921 *) Added more recognized ASN.1 object ids.
13924 *) Added more RSA padding checks for SSL/TLS.
13927 *) Added BIO proxy/filter functionality.
13930 *) Added extra_certs to SSL_CTX which can be used
13931 send extra CA certificates to the client in the CA cert chain sending
13932 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13935 *) Now Fortezza is denied in the authentication phase because
13936 this is key exchange mechanism is not supported by SSLeay at all.
13939 *) Additional PKCS1 checks.
13942 *) Support the string "TLSv1" for all TLS v1 ciphers.
13945 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13946 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13949 *) Fixed a few memory leaks.
13952 *) Fixed various code and comment typos.
13955 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13956 bytes sent in the client random.
13957 [Edward Bishop <ebishop@spyglass.com>]