155714Skris 255714Skris OpenSSL CHANGES 355714Skris _______________ 455714Skris 5306196Sjkim Changes between 1.0.1t and 1.0.1u [22 Sep 2016] 6306196Sjkim 7306196Sjkim *) OCSP Status Request extension unbounded memory growth 8306196Sjkim 9306196Sjkim A malicious client can send an excessively large OCSP Status Request 10306196Sjkim extension. If that client continually requests renegotiation, sending a 11306196Sjkim large OCSP Status Request extension each time, then there will be unbounded 12306196Sjkim memory growth on the server. This will eventually lead to a Denial Of 13306196Sjkim Service attack through memory exhaustion. Servers with a default 14306196Sjkim configuration are vulnerable even if they do not support OCSP. Builds using 15306196Sjkim the "no-ocsp" build time option are not affected. 16306196Sjkim 17306196Sjkim This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 18306196Sjkim (CVE-2016-6304) 19306196Sjkim [Matt Caswell] 20306196Sjkim 21306196Sjkim *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from 22306196Sjkim HIGH to MEDIUM. 23306196Sjkim 24306196Sjkim This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan 25306196Sjkim Leurent (INRIA) 26306196Sjkim (CVE-2016-2183) 27306196Sjkim [Rich Salz] 28306196Sjkim 29306196Sjkim *) OOB write in MDC2_Update() 30306196Sjkim 31306196Sjkim An overflow can occur in MDC2_Update() either if called directly or 32306196Sjkim through the EVP_DigestUpdate() function using MDC2. If an attacker 33306196Sjkim is able to supply very large amounts of input data after a previous 34306196Sjkim call to EVP_EncryptUpdate() with a partial block then a length check 35306196Sjkim can overflow resulting in a heap corruption. 36306196Sjkim 37306196Sjkim The amount of data needed is comparable to SIZE_MAX which is impractical 38306196Sjkim on most platforms. 39306196Sjkim 40306196Sjkim This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 41306196Sjkim (CVE-2016-6303) 42306196Sjkim [Stephen Henson] 43306196Sjkim 44306196Sjkim *) Malformed SHA512 ticket DoS 45306196Sjkim 46306196Sjkim If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a 47306196Sjkim DoS attack where a malformed ticket will result in an OOB read which will 48306196Sjkim ultimately crash. 49306196Sjkim 50306196Sjkim The use of SHA512 in TLS session tickets is comparatively rare as it requires 51306196Sjkim a custom server callback and ticket lookup mechanism. 52306196Sjkim 53306196Sjkim This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 54306196Sjkim (CVE-2016-6302) 55306196Sjkim [Stephen Henson] 56306196Sjkim 57306196Sjkim *) OOB write in BN_bn2dec() 58306196Sjkim 59306196Sjkim The function BN_bn2dec() does not check the return value of BN_div_word(). 60306196Sjkim This can cause an OOB write if an application uses this function with an 61306196Sjkim overly large BIGNUM. This could be a problem if an overly large certificate 62306196Sjkim or CRL is printed out from an untrusted source. TLS is not affected because 63306196Sjkim record limits will reject an oversized certificate before it is parsed. 64306196Sjkim 65306196Sjkim This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 66306196Sjkim (CVE-2016-2182) 67306196Sjkim [Stephen Henson] 68306196Sjkim 69306196Sjkim *) OOB read in TS_OBJ_print_bio() 70306196Sjkim 71306196Sjkim The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is 72306196Sjkim the total length the OID text representation would use and not the amount 73306196Sjkim of data written. This will result in OOB reads when large OIDs are 74306196Sjkim presented. 75306196Sjkim 76306196Sjkim This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 77306196Sjkim (CVE-2016-2180) 78306196Sjkim [Stephen Henson] 79306196Sjkim 80306196Sjkim *) Pointer arithmetic undefined behaviour 81306196Sjkim 82306196Sjkim Avoid some undefined pointer arithmetic 83306196Sjkim 84306196Sjkim A common idiom in the codebase is to check limits in the following manner: 85306196Sjkim "p + len > limit" 86306196Sjkim 87306196Sjkim Where "p" points to some malloc'd data of SIZE bytes and 88306196Sjkim limit == p + SIZE 89306196Sjkim 90306196Sjkim "len" here could be from some externally supplied data (e.g. from a TLS 91306196Sjkim message). 92306196Sjkim 93306196Sjkim The rules of C pointer arithmetic are such that "p + len" is only well 94306196Sjkim defined where len <= SIZE. Therefore the above idiom is actually 95306196Sjkim undefined behaviour. 96306196Sjkim 97306196Sjkim For example this could cause problems if some malloc implementation 98306196Sjkim provides an address for "p" such that "p + len" actually overflows for 99306196Sjkim values of len that are too big and therefore p + len < limit. 100306196Sjkim 101306196Sjkim This issue was reported to OpenSSL by Guido Vranken 102306196Sjkim (CVE-2016-2177) 103306196Sjkim [Matt Caswell] 104306196Sjkim 105306196Sjkim *) Constant time flag not preserved in DSA signing 106306196Sjkim 107306196Sjkim Operations in the DSA signing algorithm should run in constant time in 108306196Sjkim order to avoid side channel attacks. A flaw in the OpenSSL DSA 109306196Sjkim implementation means that a non-constant time codepath is followed for 110306196Sjkim certain operations. This has been demonstrated through a cache-timing 111306196Sjkim attack to be sufficient for an attacker to recover the private DSA key. 112306196Sjkim 113306196Sjkim This issue was reported by C��sar Pereida (Aalto University), Billy Brumley 114306196Sjkim (Tampere University of Technology), and Yuval Yarom (The University of 115306196Sjkim Adelaide and NICTA). 116306196Sjkim (CVE-2016-2178) 117306196Sjkim [C��sar Pereida] 118306196Sjkim 119306196Sjkim *) DTLS buffered message DoS 120306196Sjkim 121306196Sjkim In a DTLS connection where handshake messages are delivered out-of-order 122306196Sjkim those messages that OpenSSL is not yet ready to process will be buffered 123306196Sjkim for later use. Under certain circumstances, a flaw in the logic means that 124306196Sjkim those messages do not get removed from the buffer even though the handshake 125306196Sjkim has been completed. An attacker could force up to approx. 15 messages to 126306196Sjkim remain in the buffer when they are no longer required. These messages will 127306196Sjkim be cleared when the DTLS connection is closed. The default maximum size for 128306196Sjkim a message is 100k. Therefore the attacker could force an additional 1500k 129306196Sjkim to be consumed per connection. By opening many simulataneous connections an 130306196Sjkim attacker could cause a DoS attack through memory exhaustion. 131306196Sjkim 132306196Sjkim This issue was reported to OpenSSL by Quan Luo. 133306196Sjkim (CVE-2016-2179) 134306196Sjkim [Matt Caswell] 135306196Sjkim 136306196Sjkim *) DTLS replay protection DoS 137306196Sjkim 138306196Sjkim A flaw in the DTLS replay attack protection mechanism means that records 139306196Sjkim that arrive for future epochs update the replay protection "window" before 140306196Sjkim the MAC for the record has been validated. This could be exploited by an 141306196Sjkim attacker by sending a record for the next epoch (which does not have to 142306196Sjkim decrypt or have a valid MAC), with a very large sequence number. This means 143306196Sjkim that all subsequent legitimate packets are dropped causing a denial of 144306196Sjkim service for a specific DTLS connection. 145306196Sjkim 146306196Sjkim This issue was reported to OpenSSL by the OCAP audit team. 147306196Sjkim (CVE-2016-2181) 148306196Sjkim [Matt Caswell] 149306196Sjkim 150306196Sjkim *) Certificate message OOB reads 151306196Sjkim 152306196Sjkim In OpenSSL 1.0.2 and earlier some missing message length checks can result 153306196Sjkim in OOB reads of up to 2 bytes beyond an allocated buffer. There is a 154306196Sjkim theoretical DoS risk but this has not been observed in practice on common 155306196Sjkim platforms. 156306196Sjkim 157306196Sjkim The messages affected are client certificate, client certificate request 158306196Sjkim and server certificate. As a result the attack can only be performed 159306196Sjkim against a client or a server which enables client authentication. 160306196Sjkim 161306196Sjkim This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) 162306196Sjkim (CVE-2016-6306) 163306196Sjkim [Stephen Henson] 164306196Sjkim 165298999Sjkim Changes between 1.0.1s and 1.0.1t [3 May 2016] 166298999Sjkim 167298999Sjkim *) Prevent padding oracle in AES-NI CBC MAC check 168298999Sjkim 169298999Sjkim A MITM attacker can use a padding oracle attack to decrypt traffic 170298999Sjkim when the connection uses an AES CBC cipher and the server support 171298999Sjkim AES-NI. 172298999Sjkim 173298999Sjkim This issue was introduced as part of the fix for Lucky 13 padding 174298999Sjkim attack (CVE-2013-0169). The padding check was rewritten to be in 175298999Sjkim constant time by making sure that always the same bytes are read and 176298999Sjkim compared against either the MAC or padding bytes. But it no longer 177298999Sjkim checked that there was enough data to have both the MAC and padding 178298999Sjkim bytes. 179298999Sjkim 180298999Sjkim This issue was reported by Juraj Somorovsky using TLS-Attacker. 181298999Sjkim (CVE-2016-2107) 182298999Sjkim [Kurt Roeckx] 183298999Sjkim 184298999Sjkim *) Fix EVP_EncodeUpdate overflow 185298999Sjkim 186298999Sjkim An overflow can occur in the EVP_EncodeUpdate() function which is used for 187298999Sjkim Base64 encoding of binary data. If an attacker is able to supply very large 188298999Sjkim amounts of input data then a length check can overflow resulting in a heap 189298999Sjkim corruption. 190298999Sjkim 191298999Sjkim Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by 192298999Sjkim the PEM_write_bio* family of functions. These are mainly used within the 193298999Sjkim OpenSSL command line applications, so any application which processes data 194298999Sjkim from an untrusted source and outputs it as a PEM file should be considered 195298999Sjkim vulnerable to this issue. User applications that call these APIs directly 196298999Sjkim with large amounts of untrusted data may also be vulnerable. 197298999Sjkim 198298999Sjkim This issue was reported by Guido Vranken. 199298999Sjkim (CVE-2016-2105) 200298999Sjkim [Matt Caswell] 201298999Sjkim 202298999Sjkim *) Fix EVP_EncryptUpdate overflow 203298999Sjkim 204298999Sjkim An overflow can occur in the EVP_EncryptUpdate() function. If an attacker 205298999Sjkim is able to supply very large amounts of input data after a previous call to 206298999Sjkim EVP_EncryptUpdate() with a partial block then a length check can overflow 207298999Sjkim resulting in a heap corruption. Following an analysis of all OpenSSL 208298999Sjkim internal usage of the EVP_EncryptUpdate() function all usage is one of two 209298999Sjkim forms. The first form is where the EVP_EncryptUpdate() call is known to be 210298999Sjkim the first called function after an EVP_EncryptInit(), and therefore that 211298999Sjkim specific call must be safe. The second form is where the length passed to 212298999Sjkim EVP_EncryptUpdate() can be seen from the code to be some small value and 213298999Sjkim therefore there is no possibility of an overflow. Since all instances are 214298999Sjkim one of these two forms, it is believed that there can be no overflows in 215298999Sjkim internal code due to this problem. It should be noted that 216298999Sjkim EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths. 217298999Sjkim Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances 218298999Sjkim of these calls have also been analysed too and it is believed there are no 219298999Sjkim instances in internal usage where an overflow could occur. 220298999Sjkim 221298999Sjkim This issue was reported by Guido Vranken. 222298999Sjkim (CVE-2016-2106) 223298999Sjkim [Matt Caswell] 224298999Sjkim 225298999Sjkim *) Prevent ASN.1 BIO excessive memory allocation 226298999Sjkim 227298999Sjkim When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio() 228298999Sjkim a short invalid encoding can casuse allocation of large amounts of memory 229298999Sjkim potentially consuming excessive resources or exhausting memory. 230298999Sjkim 231298999Sjkim Any application parsing untrusted data through d2i BIO functions is 232298999Sjkim affected. The memory based functions such as d2i_X509() are *not* affected. 233298999Sjkim Since the memory based functions are used by the TLS library, TLS 234298999Sjkim applications are not affected. 235298999Sjkim 236298999Sjkim This issue was reported by Brian Carpenter. 237298999Sjkim (CVE-2016-2109) 238298999Sjkim [Stephen Henson] 239298999Sjkim 240298999Sjkim *) EBCDIC overread 241298999Sjkim 242298999Sjkim ASN1 Strings that are over 1024 bytes can cause an overread in applications 243298999Sjkim using the X509_NAME_oneline() function on EBCDIC systems. This could result 244298999Sjkim in arbitrary stack data being returned in the buffer. 245298999Sjkim 246298999Sjkim This issue was reported by Guido Vranken. 247298999Sjkim (CVE-2016-2176) 248298999Sjkim [Matt Caswell] 249298999Sjkim 250298999Sjkim *) Modify behavior of ALPN to invoke callback after SNI/servername 251298999Sjkim callback, such that updates to the SSL_CTX affect ALPN. 252298999Sjkim [Todd Short] 253298999Sjkim 254298999Sjkim *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the 255298999Sjkim default. 256298999Sjkim [Kurt Roeckx] 257298999Sjkim 258298999Sjkim *) Only remove the SSLv2 methods with the no-ssl2-method option. When the 259298999Sjkim methods are enabled and ssl2 is disabled the methods return NULL. 260298999Sjkim [Kurt Roeckx] 261298999Sjkim 262296317Sdelphij Changes between 1.0.1r and 1.0.1s [1 Mar 2016] 263296317Sdelphij 264296317Sdelphij * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL. 265296317Sdelphij Builds that are not configured with "enable-weak-ssl-ciphers" will not 266296317Sdelphij provide any "EXPORT" or "LOW" strength ciphers. 267296317Sdelphij [Viktor Dukhovni] 268296317Sdelphij 269296317Sdelphij * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2 270296317Sdelphij is by default disabled at build-time. Builds that are not configured with 271296317Sdelphij "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used, 272296317Sdelphij users who want to negotiate SSLv2 via the version-flexible SSLv23_method() 273296317Sdelphij will need to explicitly call either of: 274296317Sdelphij 275296317Sdelphij SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2); 276296317Sdelphij or 277296317Sdelphij SSL_clear_options(ssl, SSL_OP_NO_SSLv2); 278296317Sdelphij 279296317Sdelphij as appropriate. Even if either of those is used, or the application 280296317Sdelphij explicitly uses the version-specific SSLv2_method() or its client and 281296317Sdelphij server variants, SSLv2 ciphers vulnerable to exhaustive search key 282296317Sdelphij recovery have been removed. Specifically, the SSLv2 40-bit EXPORT 283296317Sdelphij ciphers, and SSLv2 56-bit DES are no longer available. 284296317Sdelphij (CVE-2016-0800) 285296317Sdelphij [Viktor Dukhovni] 286296317Sdelphij 287296317Sdelphij *) Fix a double-free in DSA code 288296317Sdelphij 289296317Sdelphij A double free bug was discovered when OpenSSL parses malformed DSA private 290296317Sdelphij keys and could lead to a DoS attack or memory corruption for applications 291296317Sdelphij that receive DSA private keys from untrusted sources. This scenario is 292296317Sdelphij considered rare. 293296317Sdelphij 294296317Sdelphij This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using 295296317Sdelphij libFuzzer. 296296317Sdelphij (CVE-2016-0705) 297296317Sdelphij [Stephen Henson] 298296317Sdelphij 299296317Sdelphij *) Disable SRP fake user seed to address a server memory leak. 300296317Sdelphij 301296317Sdelphij Add a new method SRP_VBASE_get1_by_user that handles the seed properly. 302296317Sdelphij 303296317Sdelphij SRP_VBASE_get_by_user had inconsistent memory management behaviour. 304296317Sdelphij In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user 305296317Sdelphij was changed to ignore the "fake user" SRP seed, even if the seed 306296317Sdelphij is configured. 307296317Sdelphij 308296317Sdelphij Users should use SRP_VBASE_get1_by_user instead. Note that in 309296317Sdelphij SRP_VBASE_get1_by_user, caller must free the returned value. Note 310296317Sdelphij also that even though configuring the SRP seed attempts to hide 311296317Sdelphij invalid usernames by continuing the handshake with fake 312296317Sdelphij credentials, this behaviour is not constant time and no strong 313296317Sdelphij guarantees are made that the handshake is indistinguishable from 314296317Sdelphij that of a valid user. 315296317Sdelphij (CVE-2016-0798) 316296317Sdelphij [Emilia K��sper] 317296317Sdelphij 318296317Sdelphij *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption 319296317Sdelphij 320296317Sdelphij In the BN_hex2bn function the number of hex digits is calculated using an 321296317Sdelphij int value |i|. Later |bn_expand| is called with a value of |i * 4|. For 322296317Sdelphij large values of |i| this can result in |bn_expand| not allocating any 323296317Sdelphij memory because |i * 4| is negative. This can leave the internal BIGNUM data 324296317Sdelphij field as NULL leading to a subsequent NULL ptr deref. For very large values 325296317Sdelphij of |i|, the calculation |i * 4| could be a positive value smaller than |i|. 326296317Sdelphij In this case memory is allocated to the internal BIGNUM data field, but it 327296317Sdelphij is insufficiently sized leading to heap corruption. A similar issue exists 328296317Sdelphij in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn 329296317Sdelphij is ever called by user applications with very large untrusted hex/dec data. 330296317Sdelphij This is anticipated to be a rare occurrence. 331296317Sdelphij 332296317Sdelphij All OpenSSL internal usage of these functions use data that is not expected 333296317Sdelphij to be untrusted, e.g. config file data or application command line 334296317Sdelphij arguments. If user developed applications generate config file data based 335296317Sdelphij on untrusted data then it is possible that this could also lead to security 336296317Sdelphij consequences. This is also anticipated to be rare. 337296317Sdelphij 338296317Sdelphij This issue was reported to OpenSSL by Guido Vranken. 339296317Sdelphij (CVE-2016-0797) 340296317Sdelphij [Matt Caswell] 341296317Sdelphij 342296317Sdelphij *) Fix memory issues in BIO_*printf functions 343296317Sdelphij 344296317Sdelphij The internal |fmtstr| function used in processing a "%s" format string in 345296317Sdelphij the BIO_*printf functions could overflow while calculating the length of a 346296317Sdelphij string and cause an OOB read when printing very long strings. 347296317Sdelphij 348296317Sdelphij Additionally the internal |doapr_outch| function can attempt to write to an 349296317Sdelphij OOB memory location (at an offset from the NULL pointer) in the event of a 350296317Sdelphij memory allocation failure. In 1.0.2 and below this could be caused where 351296317Sdelphij the size of a buffer to be allocated is greater than INT_MAX. E.g. this 352296317Sdelphij could be in processing a very long "%s" format string. Memory leaks can 353296317Sdelphij also occur. 354296317Sdelphij 355296317Sdelphij The first issue may mask the second issue dependent on compiler behaviour. 356296317Sdelphij These problems could enable attacks where large amounts of untrusted data 357296317Sdelphij is passed to the BIO_*printf functions. If applications use these functions 358296317Sdelphij in this way then they could be vulnerable. OpenSSL itself uses these 359296317Sdelphij functions when printing out human-readable dumps of ASN.1 data. Therefore 360296317Sdelphij applications that print this data could be vulnerable if the data is from 361296317Sdelphij untrusted sources. OpenSSL command line applications could also be 362296317Sdelphij vulnerable where they print out ASN.1 data, or if untrusted data is passed 363296317Sdelphij as command line arguments. 364296317Sdelphij 365296317Sdelphij Libssl is not considered directly vulnerable. Additionally certificates etc 366296317Sdelphij received via remote connections via libssl are also unlikely to be able to 367296317Sdelphij trigger these issues because of message size limits enforced within libssl. 368296317Sdelphij 369296317Sdelphij This issue was reported to OpenSSL Guido Vranken. 370296317Sdelphij (CVE-2016-0799) 371296317Sdelphij [Matt Caswell] 372296317Sdelphij 373296317Sdelphij *) Side channel attack on modular exponentiation 374296317Sdelphij 375296317Sdelphij A side-channel attack was found which makes use of cache-bank conflicts on 376296317Sdelphij the Intel Sandy-Bridge microarchitecture which could lead to the recovery 377296317Sdelphij of RSA keys. The ability to exploit this issue is limited as it relies on 378296317Sdelphij an attacker who has control of code in a thread running on the same 379296317Sdelphij hyper-threaded core as the victim thread which is performing decryptions. 380296317Sdelphij 381296317Sdelphij This issue was reported to OpenSSL by Yuval Yarom, The University of 382296317Sdelphij Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and 383296317Sdelphij Nadia Heninger, University of Pennsylvania with more information at 384296317Sdelphij http://cachebleed.info. 385296317Sdelphij (CVE-2016-0702) 386296317Sdelphij [Andy Polyakov] 387296317Sdelphij 388296317Sdelphij *) Change the req app to generate a 2048-bit RSA/DSA key by default, 389296317Sdelphij if no keysize is specified with default_bits. This fixes an 390296317Sdelphij omission in an earlier change that changed all RSA/DSA key generation 391296317Sdelphij apps to use 2048 bits by default. 392296317Sdelphij [Emilia K��sper] 393296317Sdelphij 394295016Sjkim Changes between 1.0.1q and 1.0.1r [28 Jan 2016] 395295016Sjkim 396295016Sjkim *) Protection for DH small subgroup attacks 397295016Sjkim 398295016Sjkim As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been 399295016Sjkim switched on by default and cannot be disabled. This could have some 400295016Sjkim performance impact. 401295016Sjkim [Matt Caswell] 402295016Sjkim 403295016Sjkim *) SSLv2 doesn't block disabled ciphers 404295016Sjkim 405295016Sjkim A malicious client can negotiate SSLv2 ciphers that have been disabled on 406295016Sjkim the server and complete SSLv2 handshakes even if all SSLv2 ciphers have 407295016Sjkim been disabled, provided that the SSLv2 protocol was not also disabled via 408295016Sjkim SSL_OP_NO_SSLv2. 409295016Sjkim 410295016Sjkim This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram 411295016Sjkim and Sebastian Schinzel. 412295016Sjkim (CVE-2015-3197) 413295016Sjkim [Viktor Dukhovni] 414295016Sjkim 415295016Sjkim *) Reject DH handshakes with parameters shorter than 1024 bits. 416295016Sjkim [Kurt Roeckx] 417295016Sjkim 418291721Sjkim Changes between 1.0.1p and 1.0.1q [3 Dec 2015] 419291721Sjkim 420291721Sjkim *) Certificate verify crash with missing PSS parameter 421291721Sjkim 422291721Sjkim The signature verification routines will crash with a NULL pointer 423291721Sjkim dereference if presented with an ASN.1 signature using the RSA PSS 424291721Sjkim algorithm and absent mask generation function parameter. Since these 425291721Sjkim routines are used to verify certificate signature algorithms this can be 426291721Sjkim used to crash any certificate verification operation and exploited in a 427291721Sjkim DoS attack. Any application which performs certificate verification is 428291721Sjkim vulnerable including OpenSSL clients and servers which enable client 429291721Sjkim authentication. 430291721Sjkim 431291721Sjkim This issue was reported to OpenSSL by Lo��c Jonas Etienne (Qnective AG). 432291721Sjkim (CVE-2015-3194) 433291721Sjkim [Stephen Henson] 434291721Sjkim 435291721Sjkim *) X509_ATTRIBUTE memory leak 436291721Sjkim 437291721Sjkim When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak 438291721Sjkim memory. This structure is used by the PKCS#7 and CMS routines so any 439291721Sjkim application which reads PKCS#7 or CMS data from untrusted sources is 440291721Sjkim affected. SSL/TLS is not affected. 441291721Sjkim 442291721Sjkim This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using 443291721Sjkim libFuzzer. 444291721Sjkim (CVE-2015-3195) 445291721Sjkim [Stephen Henson] 446291721Sjkim 447291721Sjkim *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs. 448291721Sjkim This changes the decoding behaviour for some invalid messages, 449291721Sjkim though the change is mostly in the more lenient direction, and 450291721Sjkim legacy behaviour is preserved as much as possible. 451291721Sjkim [Emilia K��sper] 452291721Sjkim 453291721Sjkim *) In DSA_generate_parameters_ex, if the provided seed is too short, 454296317Sdelphij use a random seed, as already documented. 455291721Sjkim [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>] 456291721Sjkim 457285330Sjkim Changes between 1.0.1o and 1.0.1p [9 Jul 2015] 458285330Sjkim 459285330Sjkim *) Alternate chains certificate forgery 460285330Sjkim 461285330Sjkim During certificate verfification, OpenSSL will attempt to find an 462285330Sjkim alternative certificate chain if the first attempt to build such a chain 463285330Sjkim fails. An error in the implementation of this logic can mean that an 464285330Sjkim attacker could cause certain checks on untrusted certificates to be 465285330Sjkim bypassed, such as the CA flag, enabling them to use a valid leaf 466285330Sjkim certificate to act as a CA and "issue" an invalid certificate. 467285330Sjkim 468285330Sjkim This issue was reported to OpenSSL by Adam Langley/David Benjamin 469285330Sjkim (Google/BoringSSL). 470291721Sjkim (CVE-2015-1793) 471285330Sjkim [Matt Caswell] 472285330Sjkim 473291721Sjkim *) Race condition handling PSK identify hint 474291721Sjkim 475291721Sjkim If PSK identity hints are received by a multi-threaded client then 476291721Sjkim the values are wrongly updated in the parent SSL_CTX structure. This can 477291721Sjkim result in a race condition potentially leading to a double free of the 478291721Sjkim identify hint data. 479291721Sjkim (CVE-2015-3196) 480291721Sjkim [Stephen Henson] 481291721Sjkim 482284330Sjkim Changes between 1.0.1n and 1.0.1o [12 Jun 2015] 483284330Sjkim *) Fix HMAC ABI incompatibility. The previous version introduced an ABI 484284330Sjkim incompatibility in the handling of HMAC. The previous ABI has now been 485284330Sjkim restored. 486284330Sjkim 487284285Sjkim Changes between 1.0.1m and 1.0.1n [11 Jun 2015] 488284285Sjkim 489284285Sjkim *) Malformed ECParameters causes infinite loop 490284285Sjkim 491284285Sjkim When processing an ECParameters structure OpenSSL enters an infinite loop 492284285Sjkim if the curve specified is over a specially malformed binary polynomial 493284285Sjkim field. 494284285Sjkim 495284285Sjkim This can be used to perform denial of service against any 496284285Sjkim system which processes public keys, certificate requests or 497284285Sjkim certificates. This includes TLS clients and TLS servers with 498284285Sjkim client authentication enabled. 499284285Sjkim 500284285Sjkim This issue was reported to OpenSSL by Joseph Barr-Pixton. 501284285Sjkim (CVE-2015-1788) 502284285Sjkim [Andy Polyakov] 503284285Sjkim 504284285Sjkim *) Exploitable out-of-bounds read in X509_cmp_time 505284285Sjkim 506284285Sjkim X509_cmp_time does not properly check the length of the ASN1_TIME 507284285Sjkim string and can read a few bytes out of bounds. In addition, 508284285Sjkim X509_cmp_time accepts an arbitrary number of fractional seconds in the 509284285Sjkim time string. 510284285Sjkim 511284285Sjkim An attacker can use this to craft malformed certificates and CRLs of 512284285Sjkim various sizes and potentially cause a segmentation fault, resulting in 513284285Sjkim a DoS on applications that verify certificates or CRLs. TLS clients 514284285Sjkim that verify CRLs are affected. TLS clients and servers with client 515284285Sjkim authentication enabled may be affected if they use custom verification 516284285Sjkim callbacks. 517284285Sjkim 518284285Sjkim This issue was reported to OpenSSL by Robert Swiecki (Google), and 519291721Sjkim independently by Hanno B��ck. 520284285Sjkim (CVE-2015-1789) 521291721Sjkim [Emilia K��sper] 522284285Sjkim 523284285Sjkim *) PKCS7 crash with missing EnvelopedContent 524284285Sjkim 525284285Sjkim The PKCS#7 parsing code does not handle missing inner EncryptedContent 526284285Sjkim correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs 527284285Sjkim with missing content and trigger a NULL pointer dereference on parsing. 528284285Sjkim 529284285Sjkim Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 530284285Sjkim structures from untrusted sources are affected. OpenSSL clients and 531284285Sjkim servers are not affected. 532284285Sjkim 533284285Sjkim This issue was reported to OpenSSL by Michal Zalewski (Google). 534284285Sjkim (CVE-2015-1790) 535291721Sjkim [Emilia K��sper] 536284285Sjkim 537284285Sjkim *) CMS verify infinite loop with unknown hash function 538284285Sjkim 539284285Sjkim When verifying a signedData message the CMS code can enter an infinite loop 540284285Sjkim if presented with an unknown hash function OID. This can be used to perform 541284285Sjkim denial of service against any system which verifies signedData messages using 542284285Sjkim the CMS code. 543284285Sjkim This issue was reported to OpenSSL by Johannes Bauer. 544284285Sjkim (CVE-2015-1792) 545284285Sjkim [Stephen Henson] 546284285Sjkim 547284285Sjkim *) Race condition handling NewSessionTicket 548284285Sjkim 549284285Sjkim If a NewSessionTicket is received by a multi-threaded client when attempting to 550284285Sjkim reuse a previous ticket then a race condition can occur potentially leading to 551284285Sjkim a double free of the ticket data. 552284285Sjkim (CVE-2015-1791) 553284285Sjkim [Matt Caswell] 554284285Sjkim 555284285Sjkim *) Reject DH handshakes with parameters shorter than 768 bits. 556284285Sjkim [Kurt Roeckx and Emilia Kasper] 557284285Sjkim 558291721Sjkim *) dhparam: generate 2048-bit parameters by default. 559291721Sjkim [Kurt Roeckx and Emilia Kasper] 560291721Sjkim 561280304Sjkim Changes between 1.0.1l and 1.0.1m [19 Mar 2015] 562280304Sjkim 563280304Sjkim *) Segmentation fault in ASN1_TYPE_cmp fix 564280304Sjkim 565280304Sjkim The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is 566280304Sjkim made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check 567280304Sjkim certificate signature algorithm consistency this can be used to crash any 568280304Sjkim certificate verification operation and exploited in a DoS attack. Any 569280304Sjkim application which performs certificate verification is vulnerable including 570280304Sjkim OpenSSL clients and servers which enable client authentication. 571280304Sjkim (CVE-2015-0286) 572280304Sjkim [Stephen Henson] 573280304Sjkim 574280304Sjkim *) ASN.1 structure reuse memory corruption fix 575280304Sjkim 576280304Sjkim Reusing a structure in ASN.1 parsing may allow an attacker to cause 577280304Sjkim memory corruption via an invalid write. Such reuse is and has been 578280304Sjkim strongly discouraged and is believed to be rare. 579280304Sjkim 580280304Sjkim Applications that parse structures containing CHOICE or ANY DEFINED BY 581280304Sjkim components may be affected. Certificate parsing (d2i_X509 and related 582280304Sjkim functions) are however not affected. OpenSSL clients and servers are 583280304Sjkim not affected. 584280304Sjkim (CVE-2015-0287) 585280304Sjkim [Stephen Henson] 586280304Sjkim 587280304Sjkim *) PKCS7 NULL pointer dereferences fix 588280304Sjkim 589280304Sjkim The PKCS#7 parsing code does not handle missing outer ContentInfo 590280304Sjkim correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with 591280304Sjkim missing content and trigger a NULL pointer dereference on parsing. 592280304Sjkim 593280304Sjkim Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or 594280304Sjkim otherwise parse PKCS#7 structures from untrusted sources are 595280304Sjkim affected. OpenSSL clients and servers are not affected. 596280304Sjkim 597280304Sjkim This issue was reported to OpenSSL by Michal Zalewski (Google). 598280304Sjkim (CVE-2015-0289) 599291721Sjkim [Emilia K��sper] 600280304Sjkim 601280304Sjkim *) DoS via reachable assert in SSLv2 servers fix 602280304Sjkim 603280304Sjkim A malicious client can trigger an OPENSSL_assert (i.e., an abort) in 604280304Sjkim servers that both support SSLv2 and enable export cipher suites by sending 605280304Sjkim a specially crafted SSLv2 CLIENT-MASTER-KEY message. 606280304Sjkim 607291721Sjkim This issue was discovered by Sean Burford (Google) and Emilia K��sper 608280304Sjkim (OpenSSL development team). 609280304Sjkim (CVE-2015-0293) 610291721Sjkim [Emilia K��sper] 611280304Sjkim 612280304Sjkim *) Use After Free following d2i_ECPrivatekey error fix 613280304Sjkim 614280304Sjkim A malformed EC private key file consumed via the d2i_ECPrivateKey function 615280304Sjkim could cause a use after free condition. This, in turn, could cause a double 616280304Sjkim free in several private key parsing functions (such as d2i_PrivateKey 617280304Sjkim or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption 618280304Sjkim for applications that receive EC private keys from untrusted 619280304Sjkim sources. This scenario is considered rare. 620280304Sjkim 621280304Sjkim This issue was discovered by the BoringSSL project and fixed in their 622280304Sjkim commit 517073cd4b. 623280304Sjkim (CVE-2015-0209) 624280304Sjkim [Matt Caswell] 625280304Sjkim 626280304Sjkim *) X509_to_X509_REQ NULL pointer deref fix 627280304Sjkim 628280304Sjkim The function X509_to_X509_REQ will crash with a NULL pointer dereference if 629280304Sjkim the certificate key is invalid. This function is rarely used in practice. 630280304Sjkim 631280304Sjkim This issue was discovered by Brian Carpenter. 632280304Sjkim (CVE-2015-0288) 633280304Sjkim [Stephen Henson] 634280304Sjkim 635280304Sjkim *) Removed the export ciphers from the DEFAULT ciphers 636280304Sjkim [Kurt Roeckx] 637280304Sjkim 638277597Sjkim Changes between 1.0.1k and 1.0.1l [15 Jan 2015] 639277597Sjkim 640277597Sjkim *) Build fixes for the Windows and OpenVMS platforms 641277597Sjkim [Matt Caswell and Richard Levitte] 642277597Sjkim 643276864Sjkim Changes between 1.0.1j and 1.0.1k [8 Jan 2015] 644276864Sjkim 645276864Sjkim *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS 646276864Sjkim message can cause a segmentation fault in OpenSSL due to a NULL pointer 647276864Sjkim dereference. This could lead to a Denial Of Service attack. Thanks to 648276864Sjkim Markus Stenberg of Cisco Systems, Inc. for reporting this issue. 649276864Sjkim (CVE-2014-3571) 650276864Sjkim [Steve Henson] 651276864Sjkim 652276864Sjkim *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the 653276864Sjkim dtls1_buffer_record function under certain conditions. In particular this 654276864Sjkim could occur if an attacker sent repeated DTLS records with the same 655276864Sjkim sequence number but for the next epoch. The memory leak could be exploited 656276864Sjkim by an attacker in a Denial of Service attack through memory exhaustion. 657276864Sjkim Thanks to Chris Mueller for reporting this issue. 658276864Sjkim (CVE-2015-0206) 659276864Sjkim [Matt Caswell] 660276864Sjkim 661276864Sjkim *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is 662276864Sjkim built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl 663276864Sjkim method would be set to NULL which could later result in a NULL pointer 664276864Sjkim dereference. Thanks to Frank Schmirler for reporting this issue. 665276864Sjkim (CVE-2014-3569) 666276864Sjkim [Kurt Roeckx] 667276864Sjkim 668276864Sjkim *) Abort handshake if server key exchange message is omitted for ephemeral 669276864Sjkim ECDH ciphersuites. 670276864Sjkim 671276864Sjkim Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for 672276864Sjkim reporting this issue. 673276864Sjkim (CVE-2014-3572) 674276864Sjkim [Steve Henson] 675276864Sjkim 676276864Sjkim *) Remove non-export ephemeral RSA code on client and server. This code 677276864Sjkim violated the TLS standard by allowing the use of temporary RSA keys in 678276864Sjkim non-export ciphersuites and could be used by a server to effectively 679276864Sjkim downgrade the RSA key length used to a value smaller than the server 680276864Sjkim certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at 681276864Sjkim INRIA or reporting this issue. 682276864Sjkim (CVE-2015-0204) 683276864Sjkim [Steve Henson] 684276864Sjkim 685276864Sjkim *) Fixed issue where DH client certificates are accepted without verification. 686276864Sjkim An OpenSSL server will accept a DH certificate for client authentication 687276864Sjkim without the certificate verify message. This effectively allows a client to 688276864Sjkim authenticate without the use of a private key. This only affects servers 689276864Sjkim which trust a client certificate authority which issues certificates 690276864Sjkim containing DH keys: these are extremely rare and hardly ever encountered. 691276864Sjkim Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting 692276864Sjkim this issue. 693276864Sjkim (CVE-2015-0205) 694276864Sjkim [Steve Henson] 695276864Sjkim 696276864Sjkim *) Ensure that the session ID context of an SSL is updated when its 697276864Sjkim SSL_CTX is updated via SSL_set_SSL_CTX. 698276864Sjkim 699276864Sjkim The session ID context is typically set from the parent SSL_CTX, 700276864Sjkim and can vary with the CTX. 701276864Sjkim [Adam Langley] 702276864Sjkim 703276864Sjkim *) Fix various certificate fingerprint issues. 704276864Sjkim 705276864Sjkim By using non-DER or invalid encodings outside the signed portion of a 706276864Sjkim certificate the fingerprint can be changed without breaking the signature. 707276864Sjkim Although no details of the signed portion of the certificate can be changed 708276864Sjkim this can cause problems with some applications: e.g. those using the 709276864Sjkim certificate fingerprint for blacklists. 710276864Sjkim 711276864Sjkim 1. Reject signatures with non zero unused bits. 712276864Sjkim 713276864Sjkim If the BIT STRING containing the signature has non zero unused bits reject 714276864Sjkim the signature. All current signature algorithms require zero unused bits. 715276864Sjkim 716276864Sjkim 2. Check certificate algorithm consistency. 717276864Sjkim 718276864Sjkim Check the AlgorithmIdentifier inside TBS matches the one in the 719276864Sjkim certificate signature. NB: this will result in signature failure 720276864Sjkim errors for some broken certificates. 721276864Sjkim 722276864Sjkim Thanks to Konrad Kraszewski from Google for reporting this issue. 723276864Sjkim 724276864Sjkim 3. Check DSA/ECDSA signatures use DER. 725276864Sjkim 726276864Sjkim Reencode DSA/ECDSA signatures and compare with the original received 727276864Sjkim signature. Return an error if there is a mismatch. 728276864Sjkim 729276864Sjkim This will reject various cases including garbage after signature 730276864Sjkim (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS 731276864Sjkim program for discovering this case) and use of BER or invalid ASN.1 INTEGERs 732276864Sjkim (negative or with leading zeroes). 733276864Sjkim 734276864Sjkim Further analysis was conducted and fixes were developed by Stephen Henson 735276864Sjkim of the OpenSSL core team. 736276864Sjkim 737276864Sjkim (CVE-2014-8275) 738276864Sjkim [Steve Henson] 739276864Sjkim 740276864Sjkim *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect 741276864Sjkim results on some platforms, including x86_64. This bug occurs at random 742276864Sjkim with a very low probability, and is not known to be exploitable in any 743276864Sjkim way, though its exact impact is difficult to determine. Thanks to Pieter 744276864Sjkim Wuille (Blockstream) who reported this issue and also suggested an initial 745276864Sjkim fix. Further analysis was conducted by the OpenSSL development team and 746276864Sjkim Adam Langley of Google. The final fix was developed by Andy Polyakov of 747276864Sjkim the OpenSSL core team. 748276864Sjkim (CVE-2014-3570) 749276864Sjkim [Andy Polyakov] 750276864Sjkim 751276864Sjkim *) Do not resume sessions on the server if the negotiated protocol 752276864Sjkim version does not match the session's version. Resuming with a different 753276864Sjkim version, while not strictly forbidden by the RFC, is of questionable 754276864Sjkim sanity and breaks all known clients. 755291721Sjkim [David Benjamin, Emilia K��sper] 756276864Sjkim 757276864Sjkim *) Tighten handling of the ChangeCipherSpec (CCS) message: reject 758276864Sjkim early CCS messages during renegotiation. (Note that because 759276864Sjkim renegotiation is encrypted, this early CCS was not exploitable.) 760291721Sjkim [Emilia K��sper] 761276864Sjkim 762276864Sjkim *) Tighten client-side session ticket handling during renegotiation: 763276864Sjkim ensure that the client only accepts a session ticket if the server sends 764276864Sjkim the extension anew in the ServerHello. Previously, a TLS client would 765276864Sjkim reuse the old extension state and thus accept a session ticket if one was 766276864Sjkim announced in the initial ServerHello. 767276864Sjkim 768276864Sjkim Similarly, ensure that the client requires a session ticket if one 769276864Sjkim was advertised in the ServerHello. Previously, a TLS client would 770276864Sjkim ignore a missing NewSessionTicket message. 771291721Sjkim [Emilia K��sper] 772276864Sjkim 773273149Sjkim Changes between 1.0.1i and 1.0.1j [15 Oct 2014] 774273149Sjkim 775273149Sjkim *) SRTP Memory Leak. 776273149Sjkim 777273149Sjkim A flaw in the DTLS SRTP extension parsing code allows an attacker, who 778273149Sjkim sends a carefully crafted handshake message, to cause OpenSSL to fail 779273149Sjkim to free up to 64k of memory causing a memory leak. This could be 780273149Sjkim exploited in a Denial Of Service attack. This issue affects OpenSSL 781273149Sjkim 1.0.1 server implementations for both SSL/TLS and DTLS regardless of 782273149Sjkim whether SRTP is used or configured. Implementations of OpenSSL that 783273149Sjkim have been compiled with OPENSSL_NO_SRTP defined are not affected. 784273149Sjkim 785273149Sjkim The fix was developed by the OpenSSL team. 786273149Sjkim (CVE-2014-3513) 787273149Sjkim [OpenSSL team] 788273149Sjkim 789273149Sjkim *) Session Ticket Memory Leak. 790273149Sjkim 791273149Sjkim When an OpenSSL SSL/TLS/DTLS server receives a session ticket the 792273149Sjkim integrity of that ticket is first verified. In the event of a session 793273149Sjkim ticket integrity check failing, OpenSSL will fail to free memory 794273149Sjkim causing a memory leak. By sending a large number of invalid session 795273149Sjkim tickets an attacker could exploit this issue in a Denial Of Service 796273149Sjkim attack. 797273149Sjkim (CVE-2014-3567) 798273149Sjkim [Steve Henson] 799273149Sjkim 800273149Sjkim *) Build option no-ssl3 is incomplete. 801273149Sjkim 802273149Sjkim When OpenSSL is configured with "no-ssl3" as a build option, servers 803273149Sjkim could accept and complete a SSL 3.0 handshake, and clients could be 804273149Sjkim configured to send them. 805273149Sjkim (CVE-2014-3568) 806273149Sjkim [Akamai and the OpenSSL team] 807273149Sjkim 808273149Sjkim *) Add support for TLS_FALLBACK_SCSV. 809273149Sjkim Client applications doing fallback retries should call 810273149Sjkim SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). 811273149Sjkim (CVE-2014-3566) 812273149Sjkim [Adam Langley, Bodo Moeller] 813273149Sjkim 814273149Sjkim *) Add additional DigestInfo checks. 815273149Sjkim 816273149Sjkim Reencode DigestInto in DER and check against the original when 817273149Sjkim verifying RSA signature: this will reject any improperly encoded 818273149Sjkim DigestInfo structures. 819273149Sjkim 820273149Sjkim Note: this is a precautionary measure and no attacks are currently known. 821273149Sjkim 822273149Sjkim [Steve Henson] 823273149Sjkim 824269686Sjkim Changes between 1.0.1h and 1.0.1i [6 Aug 2014] 825269686Sjkim 826269686Sjkim *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the 827269686Sjkim SRP code can be overrun an internal buffer. Add sanity check that 828269686Sjkim g, A, B < N to SRP code. 829269686Sjkim 830269686Sjkim Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC 831269686Sjkim Group for discovering this issue. 832269686Sjkim (CVE-2014-3512) 833269686Sjkim [Steve Henson] 834269686Sjkim 835269686Sjkim *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate 836269686Sjkim TLS 1.0 instead of higher protocol versions when the ClientHello message 837269686Sjkim is badly fragmented. This allows a man-in-the-middle attacker to force a 838269686Sjkim downgrade to TLS 1.0 even if both the server and the client support a 839269686Sjkim higher protocol version, by modifying the client's TLS records. 840269686Sjkim 841269686Sjkim Thanks to David Benjamin and Adam Langley (Google) for discovering and 842269686Sjkim researching this issue. 843269686Sjkim (CVE-2014-3511) 844269686Sjkim [David Benjamin] 845269686Sjkim 846269686Sjkim *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject 847269686Sjkim to a denial of service attack. A malicious server can crash the client 848269686Sjkim with a null pointer dereference (read) by specifying an anonymous (EC)DH 849269686Sjkim ciphersuite and sending carefully crafted handshake messages. 850269686Sjkim 851291721Sjkim Thanks to Felix Gr��bert (Google) for discovering and researching this 852269686Sjkim issue. 853269686Sjkim (CVE-2014-3510) 854291721Sjkim [Emilia K��sper] 855269686Sjkim 856269686Sjkim *) By sending carefully crafted DTLS packets an attacker could cause openssl 857269686Sjkim to leak memory. This can be exploited through a Denial of Service attack. 858269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 859269686Sjkim (CVE-2014-3507) 860269686Sjkim [Adam Langley] 861269686Sjkim 862269686Sjkim *) An attacker can force openssl to consume large amounts of memory whilst 863269686Sjkim processing DTLS handshake messages. This can be exploited through a 864269686Sjkim Denial of Service attack. 865269686Sjkim Thanks to Adam Langley for discovering and researching this issue. 866269686Sjkim (CVE-2014-3506) 867269686Sjkim [Adam Langley] 868269686Sjkim 869269686Sjkim *) An attacker can force an error condition which causes openssl to crash 870269686Sjkim whilst processing DTLS packets due to memory being freed twice. This 871269686Sjkim can be exploited through a Denial of Service attack. 872269686Sjkim Thanks to Adam Langley and Wan-Teh Chang for discovering and researching 873269686Sjkim this issue. 874269686Sjkim (CVE-2014-3505) 875269686Sjkim [Adam Langley] 876269686Sjkim 877269686Sjkim *) If a multithreaded client connects to a malicious server using a resumed 878269686Sjkim session and the server sends an ec point format extension it could write 879269686Sjkim up to 255 bytes to freed memory. 880269686Sjkim 881269686Sjkim Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this 882269686Sjkim issue. 883269686Sjkim (CVE-2014-3509) 884269686Sjkim [Gabor Tyukasz] 885269686Sjkim 886269686Sjkim *) A malicious server can crash an OpenSSL client with a null pointer 887269686Sjkim dereference (read) by specifying an SRP ciphersuite even though it was not 888269686Sjkim properly negotiated with the client. This can be exploited through a 889269686Sjkim Denial of Service attack. 890269686Sjkim 891291721Sjkim Thanks to Joonas Kuorilehto and Riku Hietam��ki (Codenomicon) for 892269686Sjkim discovering and researching this issue. 893269686Sjkim (CVE-2014-5139) 894269686Sjkim [Steve Henson] 895269686Sjkim 896269686Sjkim *) A flaw in OBJ_obj2txt may cause pretty printing functions such as 897269686Sjkim X509_name_oneline, X509_name_print_ex et al. to leak some information 898269686Sjkim from the stack. Applications may be affected if they echo pretty printing 899269686Sjkim output to the attacker. 900269686Sjkim 901269686Sjkim Thanks to Ivan Fratric (Google) for discovering this issue. 902269686Sjkim (CVE-2014-3508) 903291721Sjkim [Emilia K��sper, and Steve Henson] 904269686Sjkim 905269686Sjkim *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) 906269686Sjkim for corner cases. (Certain input points at infinity could lead to 907269686Sjkim bogus results, with non-infinity inputs mapped to infinity too.) 908269686Sjkim [Bodo Moeller] 909269686Sjkim 910267258Sjkim Changes between 1.0.1g and 1.0.1h [5 Jun 2014] 911267258Sjkim 912267258Sjkim *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted 913267258Sjkim handshake can force the use of weak keying material in OpenSSL 914267258Sjkim SSL/TLS clients and servers. 915267258Sjkim 916267258Sjkim Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and 917267258Sjkim researching this issue. (CVE-2014-0224) 918267258Sjkim [KIKUCHI Masashi, Steve Henson] 919267258Sjkim 920267258Sjkim *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an 921267258Sjkim OpenSSL DTLS client the code can be made to recurse eventually crashing 922267258Sjkim in a DoS attack. 923267258Sjkim 924267258Sjkim Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. 925267258Sjkim (CVE-2014-0221) 926267258Sjkim [Imre Rad, Steve Henson] 927267258Sjkim 928267258Sjkim *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can 929267258Sjkim be triggered by sending invalid DTLS fragments to an OpenSSL DTLS 930267258Sjkim client or server. This is potentially exploitable to run arbitrary 931267258Sjkim code on a vulnerable client or server. 932267258Sjkim 933291721Sjkim Thanks to J��ri Aedla for reporting this issue. (CVE-2014-0195) 934291721Sjkim [J��ri Aedla, Steve Henson] 935267258Sjkim 936267258Sjkim *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites 937267258Sjkim are subject to a denial of service attack. 938267258Sjkim 939291721Sjkim Thanks to Felix Gr��bert and Ivan Fratric at Google for discovering 940267258Sjkim this issue. (CVE-2014-3470) 941291721Sjkim [Felix Gr��bert, Ivan Fratric, Steve Henson] 942267258Sjkim 943267258Sjkim *) Harmonize version and its documentation. -f flag is used to display 944267258Sjkim compilation flags. 945267258Sjkim [mancha <mancha1@zoho.com>] 946267258Sjkim 947267258Sjkim *) Fix eckey_priv_encode so it immediately returns an error upon a failure 948267258Sjkim in i2d_ECPrivateKey. 949267258Sjkim [mancha <mancha1@zoho.com>] 950267258Sjkim 951267258Sjkim *) Fix some double frees. These are not thought to be exploitable. 952267258Sjkim [mancha <mancha1@zoho.com>] 953267258Sjkim 954264331Sjkim Changes between 1.0.1f and 1.0.1g [7 Apr 2014] 955264331Sjkim 956264331Sjkim *) A missing bounds check in the handling of the TLS heartbeat extension 957264331Sjkim can be used to reveal up to 64k of memory to a connected client or 958264331Sjkim server. 959264331Sjkim 960264331Sjkim Thanks for Neel Mehta of Google Security for discovering this bug and to 961264331Sjkim Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for 962264331Sjkim preparing the fix (CVE-2014-0160) 963264331Sjkim [Adam Langley, Bodo Moeller] 964264331Sjkim 965264331Sjkim *) Fix for the attack described in the paper "Recovering OpenSSL 966264331Sjkim ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" 967264331Sjkim by Yuval Yarom and Naomi Benger. Details can be obtained from: 968264331Sjkim http://eprint.iacr.org/2014/140 969264331Sjkim 970264331Sjkim Thanks to Yuval Yarom and Naomi Benger for discovering this 971264331Sjkim flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) 972264331Sjkim [Yuval Yarom and Naomi Benger] 973264331Sjkim 974264331Sjkim *) TLS pad extension: draft-agl-tls-padding-03 975264331Sjkim 976264331Sjkim Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the 977264331Sjkim TLS client Hello record length value would otherwise be > 255 and 978264331Sjkim less that 512 pad with a dummy extension containing zeroes so it 979264331Sjkim is at least 512 bytes long. 980264331Sjkim 981264331Sjkim [Adam Langley, Steve Henson] 982264331Sjkim 983264331Sjkim Changes between 1.0.1e and 1.0.1f [6 Jan 2014] 984264331Sjkim 985264331Sjkim *) Fix for TLS record tampering bug. A carefully crafted invalid 986264331Sjkim handshake could crash OpenSSL with a NULL pointer exception. 987264331Sjkim Thanks to Anton Johansson for reporting this issues. 988264331Sjkim (CVE-2013-4353) 989264331Sjkim 990264331Sjkim *) Keep original DTLS digest and encryption contexts in retransmission 991264331Sjkim structures so we can use the previous session parameters if they need 992264331Sjkim to be resent. (CVE-2013-6450) 993264331Sjkim [Steve Henson] 994264331Sjkim 995264331Sjkim *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which 996264331Sjkim avoids preferring ECDHE-ECDSA ciphers when the client appears to be 997264331Sjkim Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for 998264331Sjkim several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug 999264331Sjkim is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing 1000264331Sjkim 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. 1001264331Sjkim [Rob Stradling, Adam Langley] 1002264331Sjkim 1003246772Sjkim Changes between 1.0.1d and 1.0.1e [11 Feb 2013] 1004246772Sjkim 1005264331Sjkim *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI 1006264331Sjkim supporting platforms or when small records were transferred. 1007264331Sjkim [Andy Polyakov, Steve Henson] 1008246772Sjkim 1009246772Sjkim Changes between 1.0.1c and 1.0.1d [5 Feb 2013] 1010246772Sjkim 1011246772Sjkim *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. 1012246772Sjkim 1013246772Sjkim This addresses the flaw in CBC record processing discovered by 1014246772Sjkim Nadhem Alfardan and Kenny Paterson. Details of this attack can be found 1015246772Sjkim at: http://www.isg.rhul.ac.uk/tls/ 1016246772Sjkim 1017246772Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1018246772Sjkim Security Group at Royal Holloway, University of London 1019246772Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and 1020291721Sjkim Emilia K��sper for the initial patch. 1021246772Sjkim (CVE-2013-0169) 1022291721Sjkim [Emilia K��sper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] 1023246772Sjkim 1024246772Sjkim *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode 1025246772Sjkim ciphersuites which can be exploited in a denial of service attack. 1026246772Sjkim Thanks go to and to Adam Langley <agl@chromium.org> for discovering 1027246772Sjkim and detecting this bug and to Wolfgang Ettlinger 1028246772Sjkim <wolfgang.ettlinger@gmail.com> for independently discovering this issue. 1029246772Sjkim (CVE-2012-2686) 1030246772Sjkim [Adam Langley] 1031246772Sjkim 1032246772Sjkim *) Return an error when checking OCSP signatures when key is NULL. 1033246772Sjkim This fixes a DoS attack. (CVE-2013-0166) 1034246772Sjkim [Steve Henson] 1035246772Sjkim 1036246772Sjkim *) Make openssl verify return errors. 1037246772Sjkim [Chris Palmer <palmer@google.com> and Ben Laurie] 1038246772Sjkim 1039246772Sjkim *) Call OCSP Stapling callback after ciphersuite has been chosen, so 1040246772Sjkim the right response is stapled. Also change SSL_get_certificate() 1041246772Sjkim so it returns the certificate actually sent. 1042246772Sjkim See http://rt.openssl.org/Ticket/Display.html?id=2836. 1043246772Sjkim [Rob Stradling <rob.stradling@comodo.com>] 1044246772Sjkim 1045246772Sjkim *) Fix possible deadlock when decoding public keys. 1046246772Sjkim [Steve Henson] 1047246772Sjkim 1048246772Sjkim *) Don't use TLS 1.0 record version number in initial client hello 1049246772Sjkim if renegotiating. 1050246772Sjkim [Steve Henson] 1051246772Sjkim 1052238405Sjkim Changes between 1.0.1b and 1.0.1c [10 May 2012] 1053237657Sjkim 1054238405Sjkim *) Sanity check record length before skipping explicit IV in TLS 1055238405Sjkim 1.2, 1.1 and DTLS to fix DoS attack. 1056237657Sjkim 1057237657Sjkim Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic 1058237657Sjkim fuzzing as a service testing platform. 1059237657Sjkim (CVE-2012-2333) 1060237657Sjkim [Steve Henson] 1061237657Sjkim 1062237657Sjkim *) Initialise tkeylen properly when encrypting CMS messages. 1063237657Sjkim Thanks to Solar Designer of Openwall for reporting this issue. 1064237657Sjkim [Steve Henson] 1065237657Sjkim 1066238405Sjkim *) In FIPS mode don't try to use composite ciphers as they are not 1067238405Sjkim approved. 1068238405Sjkim [Steve Henson] 1069237657Sjkim 1070238405Sjkim Changes between 1.0.1a and 1.0.1b [26 Apr 2012] 1071237657Sjkim 1072238405Sjkim *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and 1073238405Sjkim 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately 1074238405Sjkim mean any application compiled against OpenSSL 1.0.0 headers setting 1075238405Sjkim SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng 1076238405Sjkim TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to 1077238405Sjkim 0x10000000L Any application which was previously compiled against 1078238405Sjkim OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 1079238405Sjkim will need to be recompiled as a result. Letting be results in 1080238405Sjkim inability to disable specifically TLS 1.1 and in client context, 1081238405Sjkim in unlike event, limit maximum offered version to TLS 1.0 [see below]. 1082238405Sjkim [Steve Henson] 1083237657Sjkim 1084238405Sjkim *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not 1085238405Sjkim disable just protocol X, but all protocols above X *if* there are 1086238405Sjkim protocols *below* X still enabled. In more practical terms it means 1087238405Sjkim that if application wants to disable TLS1.0 in favor of TLS1.1 and 1088238405Sjkim above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass 1089238405Sjkim SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to 1090238405Sjkim client side. 1091238405Sjkim [Andy Polyakov] 1092238405Sjkim 1093238405Sjkim Changes between 1.0.1 and 1.0.1a [19 Apr 2012] 1094238405Sjkim 1095237657Sjkim *) Check for potentially exploitable overflows in asn1_d2i_read_bio 1096237657Sjkim BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer 1097237657Sjkim in CRYPTO_realloc_clean. 1098237657Sjkim 1099237657Sjkim Thanks to Tavis Ormandy, Google Security Team, for discovering this 1100237657Sjkim issue and to Adam Langley <agl@chromium.org> for fixing it. 1101237657Sjkim (CVE-2012-2110) 1102237657Sjkim [Adam Langley (Google), Tavis Ormandy, Google Security Team] 1103237657Sjkim 1104238405Sjkim *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. 1105238405Sjkim [Adam Langley] 1106237657Sjkim 1107238405Sjkim *) Workarounds for some broken servers that "hang" if a client hello 1108238405Sjkim record length exceeds 255 bytes. 1109238405Sjkim 1110238405Sjkim 1. Do not use record version number > TLS 1.0 in initial client 1111238405Sjkim hello: some (but not all) hanging servers will now work. 1112238405Sjkim 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate 1113238405Sjkim the number of ciphers sent in the client hello. This should be 1114238405Sjkim set to an even number, such as 50, for example by passing: 1115238405Sjkim -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. 1116238405Sjkim Most broken servers should now work. 1117238405Sjkim 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable 1118238405Sjkim TLS 1.2 client support entirely. 1119238405Sjkim [Steve Henson] 1120238405Sjkim 1121238405Sjkim *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. 1122238405Sjkim [Andy Polyakov] 1123238405Sjkim 1124238405Sjkim Changes between 1.0.0h and 1.0.1 [14 Mar 2012] 1125238405Sjkim 1126238405Sjkim *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET 1127238405Sjkim STRING form instead of a DigestInfo. 1128238405Sjkim [Steve Henson] 1129238405Sjkim 1130238405Sjkim *) The format used for MDC2 RSA signatures is inconsistent between EVP 1131238405Sjkim and the RSA_sign/RSA_verify functions. This was made more apparent when 1132238405Sjkim OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular 1133238405Sjkim those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect 1134238405Sjkim the correct format in RSA_verify so both forms transparently work. 1135238405Sjkim [Steve Henson] 1136238405Sjkim 1137238405Sjkim *) Some servers which support TLS 1.0 can choke if we initially indicate 1138238405Sjkim support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA 1139238405Sjkim encrypted premaster secret. As a workaround use the maximum pemitted 1140238405Sjkim client version in client hello, this should keep such servers happy 1141238405Sjkim and still work with previous versions of OpenSSL. 1142238405Sjkim [Steve Henson] 1143238405Sjkim 1144238405Sjkim *) Add support for TLS/DTLS heartbeats. 1145238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 1146238405Sjkim 1147238405Sjkim *) Add support for SCTP. 1148238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 1149238405Sjkim 1150238405Sjkim *) Improved PRNG seeding for VOS. 1151238405Sjkim [Paul Green <Paul.Green@stratus.com>] 1152238405Sjkim 1153238405Sjkim *) Extensive assembler packs updates, most notably: 1154238405Sjkim 1155238405Sjkim - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; 1156238405Sjkim - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); 1157238405Sjkim - x86_64: bit-sliced AES implementation; 1158238405Sjkim - ARM: NEON support, contemporary platforms optimizations; 1159238405Sjkim - s390x: z196 support; 1160238405Sjkim - *: GHASH and GF(2^m) multiplication implementations; 1161238405Sjkim 1162238405Sjkim [Andy Polyakov] 1163238405Sjkim 1164238405Sjkim *) Make TLS-SRP code conformant with RFC 5054 API cleanup 1165238405Sjkim (removal of unnecessary code) 1166238405Sjkim [Peter Sylvester <peter.sylvester@edelweb.fr>] 1167238405Sjkim 1168238405Sjkim *) Add TLS key material exporter from RFC 5705. 1169238405Sjkim [Eric Rescorla] 1170238405Sjkim 1171238405Sjkim *) Add DTLS-SRTP negotiation from RFC 5764. 1172238405Sjkim [Eric Rescorla] 1173238405Sjkim 1174238405Sjkim *) Add Next Protocol Negotiation, 1175238405Sjkim http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be 1176238405Sjkim disabled with a no-npn flag to config or Configure. Code donated 1177238405Sjkim by Google. 1178238405Sjkim [Adam Langley <agl@google.com> and Ben Laurie] 1179238405Sjkim 1180238405Sjkim *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, 1181238405Sjkim NIST-P256, NIST-P521, with constant-time single point multiplication on 1182238405Sjkim typical inputs. Compiler support for the nonstandard type __uint128_t is 1183238405Sjkim required to use this (present in gcc 4.4 and later, for 64-bit builds). 1184238405Sjkim Code made available under Apache License version 2.0. 1185238405Sjkim 1186238405Sjkim Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command 1187238405Sjkim line to include this in your build of OpenSSL, and run "make depend" (or 1188238405Sjkim "make update"). This enables the following EC_METHODs: 1189238405Sjkim 1190238405Sjkim EC_GFp_nistp224_method() 1191238405Sjkim EC_GFp_nistp256_method() 1192238405Sjkim EC_GFp_nistp521_method() 1193238405Sjkim 1194238405Sjkim EC_GROUP_new_by_curve_name() will automatically use these (while 1195238405Sjkim EC_GROUP_new_curve_GFp() currently prefers the more flexible 1196238405Sjkim implementations). 1197291721Sjkim [Emilia K��sper, Adam Langley, Bodo Moeller (Google)] 1198238405Sjkim 1199238405Sjkim *) Use type ossl_ssize_t instad of ssize_t which isn't available on 1200238405Sjkim all platforms. Move ssize_t definition from e_os.h to the public 1201238405Sjkim header file e_os2.h as it now appears in public header file cms.h 1202238405Sjkim [Steve Henson] 1203238405Sjkim 1204238405Sjkim *) New -sigopt option to the ca, req and x509 utilities. Additional 1205238405Sjkim signature parameters can be passed using this option and in 1206238405Sjkim particular PSS. 1207238405Sjkim [Steve Henson] 1208238405Sjkim 1209238405Sjkim *) Add RSA PSS signing function. This will generate and set the 1210238405Sjkim appropriate AlgorithmIdentifiers for PSS based on those in the 1211238405Sjkim corresponding EVP_MD_CTX structure. No application support yet. 1212238405Sjkim [Steve Henson] 1213238405Sjkim 1214238405Sjkim *) Support for companion algorithm specific ASN1 signing routines. 1215238405Sjkim New function ASN1_item_sign_ctx() signs a pre-initialised 1216238405Sjkim EVP_MD_CTX structure and sets AlgorithmIdentifiers based on 1217238405Sjkim the appropriate parameters. 1218238405Sjkim [Steve Henson] 1219238405Sjkim 1220238405Sjkim *) Add new algorithm specific ASN1 verification initialisation function 1221238405Sjkim to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 1222238405Sjkim handling will be the same no matter what EVP_PKEY_METHOD is used. 1223238405Sjkim Add a PSS handler to support verification of PSS signatures: checked 1224238405Sjkim against a number of sample certificates. 1225238405Sjkim [Steve Henson] 1226238405Sjkim 1227238405Sjkim *) Add signature printing for PSS. Add PSS OIDs. 1228238405Sjkim [Steve Henson, Martin Kaiser <lists@kaiser.cx>] 1229238405Sjkim 1230238405Sjkim *) Add algorithm specific signature printing. An individual ASN1 method 1231238405Sjkim can now print out signatures instead of the standard hex dump. 1232238405Sjkim 1233238405Sjkim More complex signatures (e.g. PSS) can print out more meaningful 1234238405Sjkim information. Include DSA version that prints out the signature 1235238405Sjkim parameters r, s. 1236238405Sjkim [Steve Henson] 1237238405Sjkim 1238238405Sjkim *) Password based recipient info support for CMS library: implementing 1239238405Sjkim RFC3211. 1240238405Sjkim [Steve Henson] 1241238405Sjkim 1242238405Sjkim *) Split password based encryption into PBES2 and PBKDF2 functions. This 1243238405Sjkim neatly separates the code into cipher and PBE sections and is required 1244238405Sjkim for some algorithms that split PBES2 into separate pieces (such as 1245238405Sjkim password based CMS). 1246238405Sjkim [Steve Henson] 1247238405Sjkim 1248238405Sjkim *) Session-handling fixes: 1249238405Sjkim - Fix handling of connections that are resuming with a session ID, 1250238405Sjkim but also support Session Tickets. 1251238405Sjkim - Fix a bug that suppressed issuing of a new ticket if the client 1252238405Sjkim presented a ticket with an expired session. 1253238405Sjkim - Try to set the ticket lifetime hint to something reasonable. 1254238405Sjkim - Make tickets shorter by excluding irrelevant information. 1255238405Sjkim - On the client side, don't ignore renewed tickets. 1256238405Sjkim [Adam Langley, Bodo Moeller (Google)] 1257238405Sjkim 1258238405Sjkim *) Fix PSK session representation. 1259238405Sjkim [Bodo Moeller] 1260238405Sjkim 1261238405Sjkim *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. 1262238405Sjkim 1263238405Sjkim This work was sponsored by Intel. 1264238405Sjkim [Andy Polyakov] 1265238405Sjkim 1266238405Sjkim *) Add GCM support to TLS library. Some custom code is needed to split 1267238405Sjkim the IV between the fixed (from PRF) and explicit (from TLS record) 1268238405Sjkim portions. This adds all GCM ciphersuites supported by RFC5288 and 1269238405Sjkim RFC5289. Generalise some AES* cipherstrings to inlclude GCM and 1270238405Sjkim add a special AESGCM string for GCM only. 1271238405Sjkim [Steve Henson] 1272238405Sjkim 1273238405Sjkim *) Expand range of ctrls for AES GCM. Permit setting invocation 1274238405Sjkim field on decrypt and retrieval of invocation field only on encrypt. 1275238405Sjkim [Steve Henson] 1276238405Sjkim 1277238405Sjkim *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. 1278238405Sjkim As required by RFC5289 these ciphersuites cannot be used if for 1279238405Sjkim versions of TLS earlier than 1.2. 1280238405Sjkim [Steve Henson] 1281238405Sjkim 1282238405Sjkim *) For FIPS capable OpenSSL interpret a NULL default public key method 1283238405Sjkim as unset and return the appopriate default but do *not* set the default. 1284238405Sjkim This means we can return the appopriate method in applications that 1285238405Sjkim swicth between FIPS and non-FIPS modes. 1286238405Sjkim [Steve Henson] 1287238405Sjkim 1288238405Sjkim *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an 1289238405Sjkim ENGINE is used then we cannot handle that in the FIPS module so we 1290238405Sjkim keep original code iff non-FIPS operations are allowed. 1291238405Sjkim [Steve Henson] 1292238405Sjkim 1293238405Sjkim *) Add -attime option to openssl utilities. 1294238405Sjkim [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] 1295238405Sjkim 1296238405Sjkim *) Redirect DSA and DH operations to FIPS module in FIPS mode. 1297238405Sjkim [Steve Henson] 1298238405Sjkim 1299238405Sjkim *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use 1300238405Sjkim FIPS EC methods unconditionally for now. 1301238405Sjkim [Steve Henson] 1302238405Sjkim 1303238405Sjkim *) New build option no-ec2m to disable characteristic 2 code. 1304238405Sjkim [Steve Henson] 1305238405Sjkim 1306238405Sjkim *) Backport libcrypto audit of return value checking from 1.1.0-dev; not 1307238405Sjkim all cases can be covered as some introduce binary incompatibilities. 1308238405Sjkim [Steve Henson] 1309238405Sjkim 1310238405Sjkim *) Redirect RSA operations to FIPS module including keygen, 1311238405Sjkim encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. 1312238405Sjkim [Steve Henson] 1313238405Sjkim 1314238405Sjkim *) Add similar low level API blocking to ciphers. 1315238405Sjkim [Steve Henson] 1316238405Sjkim 1317238405Sjkim *) Low level digest APIs are not approved in FIPS mode: any attempt 1318238405Sjkim to use these will cause a fatal error. Applications that *really* want 1319238405Sjkim to use them can use the private_* version instead. 1320238405Sjkim [Steve Henson] 1321238405Sjkim 1322238405Sjkim *) Redirect cipher operations to FIPS module for FIPS builds. 1323238405Sjkim [Steve Henson] 1324238405Sjkim 1325238405Sjkim *) Redirect digest operations to FIPS module for FIPS builds. 1326238405Sjkim [Steve Henson] 1327238405Sjkim 1328238405Sjkim *) Update build system to add "fips" flag which will link in fipscanister.o 1329238405Sjkim for static and shared library builds embedding a signature if needed. 1330238405Sjkim [Steve Henson] 1331238405Sjkim 1332238405Sjkim *) Output TLS supported curves in preference order instead of numerical 1333238405Sjkim order. This is currently hardcoded for the highest order curves first. 1334238405Sjkim This should be configurable so applications can judge speed vs strength. 1335238405Sjkim [Steve Henson] 1336238405Sjkim 1337238405Sjkim *) Add TLS v1.2 server support for client authentication. 1338238405Sjkim [Steve Henson] 1339238405Sjkim 1340238405Sjkim *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers 1341238405Sjkim and enable MD5. 1342238405Sjkim [Steve Henson] 1343238405Sjkim 1344238405Sjkim *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying 1345238405Sjkim FIPS modules versions. 1346238405Sjkim [Steve Henson] 1347238405Sjkim 1348238405Sjkim *) Add TLS v1.2 client side support for client authentication. Keep cache 1349238405Sjkim of handshake records longer as we don't know the hash algorithm to use 1350238405Sjkim until after the certificate request message is received. 1351238405Sjkim [Steve Henson] 1352238405Sjkim 1353238405Sjkim *) Initial TLS v1.2 client support. Add a default signature algorithms 1354238405Sjkim extension including all the algorithms we support. Parse new signature 1355238405Sjkim format in client key exchange. Relax some ECC signing restrictions for 1356238405Sjkim TLS v1.2 as indicated in RFC5246. 1357238405Sjkim [Steve Henson] 1358238405Sjkim 1359238405Sjkim *) Add server support for TLS v1.2 signature algorithms extension. Switch 1360238405Sjkim to new signature format when needed using client digest preference. 1361238405Sjkim All server ciphersuites should now work correctly in TLS v1.2. No client 1362238405Sjkim support yet and no support for client certificates. 1363238405Sjkim [Steve Henson] 1364238405Sjkim 1365238405Sjkim *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch 1366238405Sjkim to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based 1367238405Sjkim ciphersuites. At present only RSA key exchange ciphersuites work with 1368238405Sjkim TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete 1369238405Sjkim SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods 1370238405Sjkim and version checking. 1371238405Sjkim [Steve Henson] 1372238405Sjkim 1373238405Sjkim *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled 1374238405Sjkim with this defined it will not be affected by any changes to ssl internal 1375238405Sjkim structures. Add several utility functions to allow openssl application 1376238405Sjkim to work with OPENSSL_NO_SSL_INTERN defined. 1377238405Sjkim [Steve Henson] 1378238405Sjkim 1379238405Sjkim *) Add SRP support. 1380238405Sjkim [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie] 1381238405Sjkim 1382238405Sjkim *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. 1383238405Sjkim [Steve Henson] 1384238405Sjkim 1385238405Sjkim *) Permit abbreviated handshakes when renegotiating using the function 1386238405Sjkim SSL_renegotiate_abbreviated(). 1387238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>] 1388238405Sjkim 1389238405Sjkim *) Add call to ENGINE_register_all_complete() to 1390238405Sjkim ENGINE_load_builtin_engines(), so some implementations get used 1391238405Sjkim automatically instead of needing explicit application support. 1392238405Sjkim [Steve Henson] 1393238405Sjkim 1394238405Sjkim *) Add support for TLS key exporter as described in RFC5705. 1395238405Sjkim [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] 1396238405Sjkim 1397238405Sjkim *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only 1398238405Sjkim a few changes are required: 1399238405Sjkim 1400238405Sjkim Add SSL_OP_NO_TLSv1_1 flag. 1401238405Sjkim Add TLSv1_1 methods. 1402238405Sjkim Update version checking logic to handle version 1.1. 1403238405Sjkim Add explicit IV handling (ported from DTLS code). 1404238405Sjkim Add command line options to s_client/s_server. 1405238405Sjkim [Steve Henson] 1406238405Sjkim 1407238405Sjkim Changes between 1.0.0g and 1.0.0h [12 Mar 2012] 1408238405Sjkim 1409237657Sjkim *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness 1410237657Sjkim in CMS and PKCS7 code. When RSA decryption fails use a random key for 1411237657Sjkim content decryption and always return the same error. Note: this attack 1412237657Sjkim needs on average 2^20 messages so it only affects automated senders. The 1413237657Sjkim old behaviour can be reenabled in the CMS code by setting the 1414237657Sjkim CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where 1415237657Sjkim an MMA defence is not necessary. 1416237657Sjkim Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering 1417237657Sjkim this issue. (CVE-2012-0884) 1418237657Sjkim [Steve Henson] 1419237657Sjkim 1420237657Sjkim *) Fix CVE-2011-4619: make sure we really are receiving a 1421237657Sjkim client hello before rejecting multiple SGC restarts. Thanks to 1422237657Sjkim Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. 1423237657Sjkim [Steve Henson] 1424237657Sjkim 1425238405Sjkim Changes between 1.0.0f and 1.0.0g [18 Jan 2012] 1426238405Sjkim 1427238405Sjkim *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. 1428238405Sjkim Thanks to Antonio Martin, Enterprise Secure Access Research and 1429238405Sjkim Development, Cisco Systems, Inc. for discovering this bug and 1430238405Sjkim preparing a fix. (CVE-2012-0050) 1431238405Sjkim [Antonio Martin] 1432238405Sjkim 1433238405Sjkim Changes between 1.0.0e and 1.0.0f [4 Jan 2012] 1434238405Sjkim 1435238405Sjkim *) Nadhem Alfardan and Kenny Paterson have discovered an extension 1436238405Sjkim of the Vaudenay padding oracle attack on CBC mode encryption 1437238405Sjkim which enables an efficient plaintext recovery attack against 1438238405Sjkim the OpenSSL implementation of DTLS. Their attack exploits timing 1439238405Sjkim differences arising during decryption processing. A research 1440238405Sjkim paper describing this attack can be found at: 1441238405Sjkim http://www.isg.rhul.ac.uk/~kp/dtls.pdf 1442238405Sjkim Thanks go to Nadhem Alfardan and Kenny Paterson of the Information 1443238405Sjkim Security Group at Royal Holloway, University of London 1444238405Sjkim (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann 1445238405Sjkim <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> 1446238405Sjkim for preparing the fix. (CVE-2011-4108) 1447238405Sjkim [Robin Seggelmann, Michael Tuexen] 1448238405Sjkim 1449238405Sjkim *) Clear bytes used for block padding of SSL 3.0 records. 1450238405Sjkim (CVE-2011-4576) 1451238405Sjkim [Adam Langley (Google)] 1452238405Sjkim 1453238405Sjkim *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George 1454238405Sjkim Kadianakis <desnacked@gmail.com> for discovering this issue and 1455238405Sjkim Adam Langley for preparing the fix. (CVE-2011-4619) 1456238405Sjkim [Adam Langley (Google)] 1457238405Sjkim 1458238405Sjkim *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) 1459238405Sjkim [Andrey Kulikov <amdeich@gmail.com>] 1460238405Sjkim 1461238405Sjkim *) Prevent malformed RFC3779 data triggering an assertion failure. 1462238405Sjkim Thanks to Andrew Chi, BBN Technologies, for discovering the flaw 1463238405Sjkim and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) 1464238405Sjkim [Rob Austein <sra@hactrn.net>] 1465238405Sjkim 1466238405Sjkim *) Improved PRNG seeding for VOS. 1467238405Sjkim [Paul Green <Paul.Green@stratus.com>] 1468238405Sjkim 1469238405Sjkim *) Fix ssl_ciph.c set-up race. 1470238405Sjkim [Adam Langley (Google)] 1471238405Sjkim 1472238405Sjkim *) Fix spurious failures in ecdsatest.c. 1473291721Sjkim [Emilia K��sper (Google)] 1474238405Sjkim 1475238405Sjkim *) Fix the BIO_f_buffer() implementation (which was mixing different 1476238405Sjkim interpretations of the '..._len' fields). 1477238405Sjkim [Adam Langley (Google)] 1478238405Sjkim 1479238405Sjkim *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than 1480238405Sjkim BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent 1481238405Sjkim threads won't reuse the same blinding coefficients. 1482238405Sjkim 1483238405Sjkim This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING 1484238405Sjkim lock to call BN_BLINDING_invert_ex, and avoids one use of 1485238405Sjkim BN_BLINDING_update for each BN_BLINDING structure (previously, 1486238405Sjkim the last update always remained unused). 1487291721Sjkim [Emilia K��sper (Google)] 1488238405Sjkim 1489238405Sjkim *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. 1490238405Sjkim [Bob Buckholz (Google)] 1491238405Sjkim 1492238405Sjkim Changes between 1.0.0d and 1.0.0e [6 Sep 2011] 1493238405Sjkim 1494238405Sjkim *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted 1495238405Sjkim by initialising X509_STORE_CTX properly. (CVE-2011-3207) 1496238405Sjkim [Kaspar Brand <ossl@velox.ch>] 1497238405Sjkim 1498238405Sjkim *) Fix SSL memory handling for (EC)DH ciphersuites, in particular 1499238405Sjkim for multi-threaded use of ECDH. (CVE-2011-3210) 1500238405Sjkim [Adam Langley (Google)] 1501238405Sjkim 1502238405Sjkim *) Fix x509_name_ex_d2i memory leak on bad inputs. 1503238405Sjkim [Bodo Moeller] 1504238405Sjkim 1505238405Sjkim *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check 1506238405Sjkim signature public key algorithm by using OID xref utilities instead. 1507238405Sjkim Before this you could only use some ECC ciphersuites with SHA1 only. 1508238405Sjkim [Steve Henson] 1509238405Sjkim 1510238405Sjkim *) Add protection against ECDSA timing attacks as mentioned in the paper 1511238405Sjkim by Billy Bob Brumley and Nicola Tuveri, see: 1512238405Sjkim 1513238405Sjkim http://eprint.iacr.org/2011/232.pdf 1514238405Sjkim 1515238405Sjkim [Billy Bob Brumley and Nicola Tuveri] 1516238405Sjkim 1517238405Sjkim Changes between 1.0.0c and 1.0.0d [8 Feb 2011] 1518238405Sjkim 1519238405Sjkim *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 1520238405Sjkim [Neel Mehta, Adam Langley, Bodo Moeller (Google)] 1521238405Sjkim 1522238405Sjkim *) Fix bug in string printing code: if *any* escaping is enabled we must 1523238405Sjkim escape the escape character (backslash) or the resulting string is 1524238405Sjkim ambiguous. 1525238405Sjkim [Steve Henson] 1526238405Sjkim 1527238405Sjkim Changes between 1.0.0b and 1.0.0c [2 Dec 2010] 1528238405Sjkim 1529238405Sjkim *) Disable code workaround for ancient and obsolete Netscape browsers 1530238405Sjkim and servers: an attacker can use it in a ciphersuite downgrade attack. 1531238405Sjkim Thanks to Martin Rex for discovering this bug. CVE-2010-4180 1532238405Sjkim [Steve Henson] 1533238405Sjkim 1534238405Sjkim *) Fixed J-PAKE implementation error, originally discovered by 1535238405Sjkim Sebastien Martini, further info and confirmation from Stefan 1536238405Sjkim Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 1537238405Sjkim [Ben Laurie] 1538238405Sjkim 1539238405Sjkim Changes between 1.0.0a and 1.0.0b [16 Nov 2010] 1540238405Sjkim 1541238405Sjkim *) Fix extension code to avoid race conditions which can result in a buffer 1542238405Sjkim overrun vulnerability: resumed sessions must not be modified as they can 1543238405Sjkim be shared by multiple threads. CVE-2010-3864 1544238405Sjkim [Steve Henson] 1545238405Sjkim 1546238405Sjkim *) Fix WIN32 build system to correctly link an ENGINE directory into 1547238405Sjkim a DLL. 1548238405Sjkim [Steve Henson] 1549238405Sjkim 1550238405Sjkim Changes between 1.0.0 and 1.0.0a [01 Jun 2010] 1551238405Sjkim 1552238405Sjkim *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover 1553238405Sjkim (CVE-2010-1633) 1554238405Sjkim [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] 1555238405Sjkim 1556238405Sjkim Changes between 0.9.8n and 1.0.0 [29 Mar 2010] 1557238405Sjkim 1558238405Sjkim *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher 1559238405Sjkim context. The operation can be customised via the ctrl mechanism in 1560238405Sjkim case ENGINEs want to include additional functionality. 1561238405Sjkim [Steve Henson] 1562238405Sjkim 1563238405Sjkim *) Tolerate yet another broken PKCS#8 key format: private key value negative. 1564238405Sjkim [Steve Henson] 1565238405Sjkim 1566238405Sjkim *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to 1567238405Sjkim output hashes compatible with older versions of OpenSSL. 1568238405Sjkim [Willy Weisz <weisz@vcpc.univie.ac.at>] 1569238405Sjkim 1570238405Sjkim *) Fix compression algorithm handling: if resuming a session use the 1571238405Sjkim compression algorithm of the resumed session instead of determining 1572238405Sjkim it from client hello again. Don't allow server to change algorithm. 1573238405Sjkim [Steve Henson] 1574238405Sjkim 1575238405Sjkim *) Add load_crls() function to apps tidying load_certs() too. Add option 1576238405Sjkim to verify utility to allow additional CRLs to be included. 1577238405Sjkim [Steve Henson] 1578238405Sjkim 1579238405Sjkim *) Update OCSP request code to permit adding custom headers to the request: 1580238405Sjkim some responders need this. 1581238405Sjkim [Steve Henson] 1582238405Sjkim 1583238405Sjkim *) The function EVP_PKEY_sign() returns <=0 on error: check return code 1584238405Sjkim correctly. 1585238405Sjkim [Julia Lawall <julia@diku.dk>] 1586238405Sjkim 1587238405Sjkim *) Update verify callback code in apps/s_cb.c and apps/verify.c, it 1588238405Sjkim needlessly dereferenced structures, used obsolete functions and 1589238405Sjkim didn't handle all updated verify codes correctly. 1590238405Sjkim [Steve Henson] 1591238405Sjkim 1592238405Sjkim *) Disable MD2 in the default configuration. 1593238405Sjkim [Steve Henson] 1594238405Sjkim 1595238405Sjkim *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to 1596238405Sjkim indicate the initial BIO being pushed or popped. This makes it possible 1597238405Sjkim to determine whether the BIO is the one explicitly called or as a result 1598238405Sjkim of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so 1599238405Sjkim it handles reference counts correctly and doesn't zero out the I/O bio 1600238405Sjkim when it is not being explicitly popped. WARNING: applications which 1601238405Sjkim included workarounds for the old buggy behaviour will need to be modified 1602238405Sjkim or they could free up already freed BIOs. 1603238405Sjkim [Steve Henson] 1604238405Sjkim 1605238405Sjkim *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni 1606238405Sjkim renaming to all platforms (within the 0.9.8 branch, this was 1607238405Sjkim done conditionally on Netware platforms to avoid a name clash). 1608238405Sjkim [Guenter <lists@gknw.net>] 1609238405Sjkim 1610238405Sjkim *) Add ECDHE and PSK support to DTLS. 1611238405Sjkim [Michael Tuexen <tuexen@fh-muenster.de>] 1612238405Sjkim 1613238405Sjkim *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't 1614238405Sjkim be used on C++. 1615238405Sjkim [Steve Henson] 1616238405Sjkim 1617238405Sjkim *) Add "missing" function EVP_MD_flags() (without this the only way to 1618238405Sjkim retrieve a digest flags is by accessing the structure directly. Update 1619238405Sjkim EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest 1620238405Sjkim or cipher is registered as in the "from" argument. Print out all 1621238405Sjkim registered digests in the dgst usage message instead of manually 1622238405Sjkim attempting to work them out. 1623238405Sjkim [Steve Henson] 1624238405Sjkim 1625238405Sjkim *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: 1626238405Sjkim this allows the use of compression and extensions. Change default cipher 1627238405Sjkim string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 1628238405Sjkim by default unless an application cipher string requests it. 1629238405Sjkim [Steve Henson] 1630238405Sjkim 1631238405Sjkim *) Alter match criteria in PKCS12_parse(). It used to try to use local 1632238405Sjkim key ids to find matching certificates and keys but some PKCS#12 files 1633238405Sjkim don't follow the (somewhat unwritten) rules and this strategy fails. 1634238405Sjkim Now just gather all certificates together and the first private key 1635238405Sjkim then look for the first certificate that matches the key. 1636238405Sjkim [Steve Henson] 1637238405Sjkim 1638238405Sjkim *) Support use of registered digest and cipher names for dgst and cipher 1639238405Sjkim commands instead of having to add each one as a special case. So now 1640238405Sjkim you can do: 1641238405Sjkim 1642238405Sjkim openssl sha256 foo 1643238405Sjkim 1644238405Sjkim as well as: 1645238405Sjkim 1646238405Sjkim openssl dgst -sha256 foo 1647238405Sjkim 1648238405Sjkim and this works for ENGINE based algorithms too. 1649238405Sjkim 1650238405Sjkim [Steve Henson] 1651238405Sjkim 1652238405Sjkim *) Update Gost ENGINE to support parameter files. 1653238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1654238405Sjkim 1655238405Sjkim *) Support GeneralizedTime in ca utility. 1656238405Sjkim [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] 1657238405Sjkim 1658238405Sjkim *) Enhance the hash format used for certificate directory links. The new 1659238405Sjkim form uses the canonical encoding (meaning equivalent names will work 1660238405Sjkim even if they aren't identical) and uses SHA1 instead of MD5. This form 1661238405Sjkim is incompatible with the older format and as a result c_rehash should 1662238405Sjkim be used to rebuild symbolic links. 1663238405Sjkim [Steve Henson] 1664238405Sjkim 1665238405Sjkim *) Make PKCS#8 the default write format for private keys, replacing the 1666238405Sjkim traditional format. This form is standardised, more secure and doesn't 1667238405Sjkim include an implicit MD5 dependency. 1668238405Sjkim [Steve Henson] 1669238405Sjkim 1670238405Sjkim *) Add a $gcc_devteam_warn option to Configure. The idea is that any code 1671238405Sjkim committed to OpenSSL should pass this lot as a minimum. 1672238405Sjkim [Steve Henson] 1673238405Sjkim 1674238405Sjkim *) Add session ticket override functionality for use by EAP-FAST. 1675238405Sjkim [Jouni Malinen <j@w1.fi>] 1676238405Sjkim 1677238405Sjkim *) Modify HMAC functions to return a value. Since these can be implemented 1678238405Sjkim in an ENGINE errors can occur. 1679238405Sjkim [Steve Henson] 1680238405Sjkim 1681238405Sjkim *) Type-checked OBJ_bsearch_ex. 1682238405Sjkim [Ben Laurie] 1683238405Sjkim 1684238405Sjkim *) Type-checked OBJ_bsearch. Also some constification necessitated 1685238405Sjkim by type-checking. Still to come: TXT_DB, bsearch(?), 1686238405Sjkim OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, 1687238405Sjkim CONF_VALUE. 1688238405Sjkim [Ben Laurie] 1689238405Sjkim 1690238405Sjkim *) New function OPENSSL_gmtime_adj() to add a specific number of days and 1691238405Sjkim seconds to a tm structure directly, instead of going through OS 1692238405Sjkim specific date routines. This avoids any issues with OS routines such 1693238405Sjkim as the year 2038 bug. New *_adj() functions for ASN1 time structures 1694238405Sjkim and X509_time_adj_ex() to cover the extended range. The existing 1695238405Sjkim X509_time_adj() is still usable and will no longer have any date issues. 1696238405Sjkim [Steve Henson] 1697238405Sjkim 1698238405Sjkim *) Delta CRL support. New use deltas option which will attempt to locate 1699238405Sjkim and search any appropriate delta CRLs available. 1700238405Sjkim 1701238405Sjkim This work was sponsored by Google. 1702238405Sjkim [Steve Henson] 1703238405Sjkim 1704238405Sjkim *) Support for CRLs partitioned by reason code. Reorganise CRL processing 1705238405Sjkim code and add additional score elements. Validate alternate CRL paths 1706238405Sjkim as part of the CRL checking and indicate a new error "CRL path validation 1707238405Sjkim error" in this case. Applications wanting additional details can use 1708238405Sjkim the verify callback and check the new "parent" field. If this is not 1709238405Sjkim NULL CRL path validation is taking place. Existing applications wont 1710238405Sjkim see this because it requires extended CRL support which is off by 1711238405Sjkim default. 1712238405Sjkim 1713238405Sjkim This work was sponsored by Google. 1714238405Sjkim [Steve Henson] 1715238405Sjkim 1716238405Sjkim *) Support for freshest CRL extension. 1717238405Sjkim 1718238405Sjkim This work was sponsored by Google. 1719238405Sjkim [Steve Henson] 1720238405Sjkim 1721238405Sjkim *) Initial indirect CRL support. Currently only supported in the CRLs 1722238405Sjkim passed directly and not via lookup. Process certificate issuer 1723238405Sjkim CRL entry extension and lookup CRL entries by bother issuer name 1724238405Sjkim and serial number. Check and process CRL issuer entry in IDP extension. 1725238405Sjkim 1726238405Sjkim This work was sponsored by Google. 1727238405Sjkim [Steve Henson] 1728238405Sjkim 1729238405Sjkim *) Add support for distinct certificate and CRL paths. The CRL issuer 1730238405Sjkim certificate is validated separately in this case. Only enabled if 1731238405Sjkim an extended CRL support flag is set: this flag will enable additional 1732238405Sjkim CRL functionality in future. 1733238405Sjkim 1734238405Sjkim This work was sponsored by Google. 1735238405Sjkim [Steve Henson] 1736238405Sjkim 1737238405Sjkim *) Add support for policy mappings extension. 1738238405Sjkim 1739238405Sjkim This work was sponsored by Google. 1740238405Sjkim [Steve Henson] 1741238405Sjkim 1742238405Sjkim *) Fixes to pathlength constraint, self issued certificate handling, 1743238405Sjkim policy processing to align with RFC3280 and PKITS tests. 1744238405Sjkim 1745238405Sjkim This work was sponsored by Google. 1746238405Sjkim [Steve Henson] 1747238405Sjkim 1748238405Sjkim *) Support for name constraints certificate extension. DN, email, DNS 1749238405Sjkim and URI types are currently supported. 1750238405Sjkim 1751238405Sjkim This work was sponsored by Google. 1752238405Sjkim [Steve Henson] 1753238405Sjkim 1754238405Sjkim *) To cater for systems that provide a pointer-based thread ID rather 1755238405Sjkim than numeric, deprecate the current numeric thread ID mechanism and 1756238405Sjkim replace it with a structure and associated callback type. This 1757238405Sjkim mechanism allows a numeric "hash" to be extracted from a thread ID in 1758238405Sjkim either case, and on platforms where pointers are larger than 'long', 1759238405Sjkim mixing is done to help ensure the numeric 'hash' is usable even if it 1760238405Sjkim can't be guaranteed unique. The default mechanism is to use "&errno" 1761238405Sjkim as a pointer-based thread ID to distinguish between threads. 1762238405Sjkim 1763238405Sjkim Applications that want to provide their own thread IDs should now use 1764238405Sjkim CRYPTO_THREADID_set_callback() to register a callback that will call 1765238405Sjkim either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). 1766238405Sjkim 1767238405Sjkim Note that ERR_remove_state() is now deprecated, because it is tied 1768238405Sjkim to the assumption that thread IDs are numeric. ERR_remove_state(0) 1769238405Sjkim to free the current thread's error state should be replaced by 1770238405Sjkim ERR_remove_thread_state(NULL). 1771238405Sjkim 1772238405Sjkim (This new approach replaces the functions CRYPTO_set_idptr_callback(), 1773238405Sjkim CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in 1774238405Sjkim OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an 1775238405Sjkim application was previously providing a numeric thread callback that 1776238405Sjkim was inappropriate for distinguishing threads, then uniqueness might 1777238405Sjkim have been obtained with &errno that happened immediately in the 1778238405Sjkim intermediate development versions of OpenSSL; this is no longer the 1779238405Sjkim case, the numeric thread callback will now override the automatic use 1780238405Sjkim of &errno.) 1781238405Sjkim [Geoff Thorpe, with help from Bodo Moeller] 1782238405Sjkim 1783238405Sjkim *) Initial support for different CRL issuing certificates. This covers a 1784238405Sjkim simple case where the self issued certificates in the chain exist and 1785238405Sjkim the real CRL issuer is higher in the existing chain. 1786238405Sjkim 1787238405Sjkim This work was sponsored by Google. 1788238405Sjkim [Steve Henson] 1789238405Sjkim 1790238405Sjkim *) Removed effectively defunct crypto/store from the build. 1791238405Sjkim [Ben Laurie] 1792238405Sjkim 1793238405Sjkim *) Revamp of STACK to provide stronger type-checking. Still to come: 1794238405Sjkim TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, 1795238405Sjkim ASN1_STRING, CONF_VALUE. 1796238405Sjkim [Ben Laurie] 1797238405Sjkim 1798238405Sjkim *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer 1799238405Sjkim RAM on SSL connections. This option can save about 34k per idle SSL. 1800238405Sjkim [Nick Mathewson] 1801238405Sjkim 1802238405Sjkim *) Revamp of LHASH to provide stronger type-checking. Still to come: 1803238405Sjkim STACK, TXT_DB, bsearch, qsort. 1804238405Sjkim [Ben Laurie] 1805238405Sjkim 1806238405Sjkim *) Initial support for Cryptographic Message Syntax (aka CMS) based 1807238405Sjkim on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, 1808238405Sjkim support for data, signedData, compressedData, digestedData and 1809238405Sjkim encryptedData, envelopedData types included. Scripts to check against 1810238405Sjkim RFC4134 examples draft and interop and consistency checks of many 1811238405Sjkim content types and variants. 1812238405Sjkim [Steve Henson] 1813238405Sjkim 1814238405Sjkim *) Add options to enc utility to support use of zlib compression BIO. 1815238405Sjkim [Steve Henson] 1816238405Sjkim 1817238405Sjkim *) Extend mk1mf to support importing of options and assembly language 1818238405Sjkim files from Configure script, currently only included in VC-WIN32. 1819238405Sjkim The assembly language rules can now optionally generate the source 1820238405Sjkim files from the associated perl scripts. 1821238405Sjkim [Steve Henson] 1822238405Sjkim 1823238405Sjkim *) Implement remaining functionality needed to support GOST ciphersuites. 1824238405Sjkim Interop testing has been performed using CryptoPro implementations. 1825238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1826238405Sjkim 1827238405Sjkim *) s390x assembler pack. 1828238405Sjkim [Andy Polyakov] 1829238405Sjkim 1830238405Sjkim *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU 1831238405Sjkim "family." 1832238405Sjkim [Andy Polyakov] 1833238405Sjkim 1834238405Sjkim *) Implement Opaque PRF Input TLS extension as specified in 1835238405Sjkim draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an 1836238405Sjkim official specification yet and no extension type assignment by 1837238405Sjkim IANA exists, this extension (for now) will have to be explicitly 1838238405Sjkim enabled when building OpenSSL by providing the extension number 1839238405Sjkim to use. For example, specify an option 1840238405Sjkim 1841238405Sjkim -DTLSEXT_TYPE_opaque_prf_input=0x9527 1842238405Sjkim 1843238405Sjkim to the "config" or "Configure" script to enable the extension, 1844238405Sjkim assuming extension number 0x9527 (which is a completely arbitrary 1845238405Sjkim and unofficial assignment based on the MD5 hash of the Internet 1846238405Sjkim Draft). Note that by doing so, you potentially lose 1847238405Sjkim interoperability with other TLS implementations since these might 1848238405Sjkim be using the same extension number for other purposes. 1849238405Sjkim 1850238405Sjkim SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the 1851238405Sjkim opaque PRF input value to use in the handshake. This will create 1852238405Sjkim an interal copy of the length-'len' string at 'src', and will 1853238405Sjkim return non-zero for success. 1854238405Sjkim 1855238405Sjkim To get more control and flexibility, provide a callback function 1856238405Sjkim by using 1857238405Sjkim 1858238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) 1859238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) 1860238405Sjkim 1861238405Sjkim where 1862238405Sjkim 1863238405Sjkim int (*cb)(SSL *, void *peerinput, size_t len, void *arg); 1864238405Sjkim void *arg; 1865238405Sjkim 1866238405Sjkim Callback function 'cb' will be called in handshakes, and is 1867238405Sjkim expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. 1868238405Sjkim Argument 'arg' is for application purposes (the value as given to 1869238405Sjkim SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly 1870238405Sjkim be provided to the callback function). The callback function 1871238405Sjkim has to return non-zero to report success: usually 1 to use opaque 1872238405Sjkim PRF input just if possible, or 2 to enforce use of the opaque PRF 1873238405Sjkim input. In the latter case, the library will abort the handshake 1874238405Sjkim if opaque PRF input is not successfully negotiated. 1875238405Sjkim 1876238405Sjkim Arguments 'peerinput' and 'len' given to the callback function 1877238405Sjkim will always be NULL and 0 in the case of a client. A server will 1878238405Sjkim see the client's opaque PRF input through these variables if 1879238405Sjkim available (NULL and 0 otherwise). Note that if the server 1880238405Sjkim provides an opaque PRF input, the length must be the same as the 1881238405Sjkim length of the client's opaque PRF input. 1882238405Sjkim 1883238405Sjkim Note that the callback function will only be called when creating 1884238405Sjkim a new session (session resumption can resume whatever was 1885238405Sjkim previously negotiated), and will not be called in SSL 2.0 1886238405Sjkim handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or 1887238405Sjkim SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended 1888238405Sjkim for applications that need to enforce opaque PRF input. 1889238405Sjkim 1890238405Sjkim [Bodo Moeller] 1891238405Sjkim 1892238405Sjkim *) Update ssl code to support digests other than SHA1+MD5 for handshake 1893238405Sjkim MAC. 1894238405Sjkim 1895238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>] 1896238405Sjkim 1897238405Sjkim *) Add RFC4507 support to OpenSSL. This includes the corrections in 1898238405Sjkim RFC4507bis. The encrypted ticket format is an encrypted encoded 1899238405Sjkim SSL_SESSION structure, that way new session features are automatically 1900238405Sjkim supported. 1901238405Sjkim 1902238405Sjkim If a client application caches session in an SSL_SESSION structure 1903238405Sjkim support is transparent because tickets are now stored in the encoded 1904238405Sjkim SSL_SESSION. 1905238405Sjkim 1906238405Sjkim The SSL_CTX structure automatically generates keys for ticket 1907238405Sjkim protection in servers so again support should be possible 1908238405Sjkim with no application modification. 1909238405Sjkim 1910238405Sjkim If a client or server wishes to disable RFC4507 support then the option 1911238405Sjkim SSL_OP_NO_TICKET can be set. 1912238405Sjkim 1913238405Sjkim Add a TLS extension debugging callback to allow the contents of any client 1914238405Sjkim or server extensions to be examined. 1915238405Sjkim 1916238405Sjkim This work was sponsored by Google. 1917238405Sjkim [Steve Henson] 1918238405Sjkim 1919238405Sjkim *) Final changes to avoid use of pointer pointer casts in OpenSSL. 1920238405Sjkim OpenSSL should now compile cleanly on gcc 4.2 1921238405Sjkim [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] 1922238405Sjkim 1923238405Sjkim *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC 1924238405Sjkim support including streaming MAC support: this is required for GOST 1925238405Sjkim ciphersuite support. 1926238405Sjkim [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] 1927238405Sjkim 1928238405Sjkim *) Add option -stream to use PKCS#7 streaming in smime utility. New 1929238405Sjkim function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() 1930238405Sjkim to output in BER and PEM format. 1931238405Sjkim [Steve Henson] 1932238405Sjkim 1933238405Sjkim *) Experimental support for use of HMAC via EVP_PKEY interface. This 1934238405Sjkim allows HMAC to be handled via the EVP_DigestSign*() interface. The 1935238405Sjkim EVP_PKEY "key" in this case is the HMAC key, potentially allowing 1936238405Sjkim ENGINE support for HMAC keys which are unextractable. New -mac and 1937238405Sjkim -macopt options to dgst utility. 1938238405Sjkim [Steve Henson] 1939238405Sjkim 1940238405Sjkim *) New option -sigopt to dgst utility. Update dgst to use 1941238405Sjkim EVP_Digest{Sign,Verify}*. These two changes make it possible to use 1942238405Sjkim alternative signing paramaters such as X9.31 or PSS in the dgst 1943238405Sjkim utility. 1944238405Sjkim [Steve Henson] 1945238405Sjkim 1946238405Sjkim *) Change ssl_cipher_apply_rule(), the internal function that does 1947238405Sjkim the work each time a ciphersuite string requests enabling 1948238405Sjkim ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or 1949238405Sjkim removing ("!foo+bar") a class of ciphersuites: Now it maintains 1950238405Sjkim the order of disabled ciphersuites such that those ciphersuites 1951238405Sjkim that most recently went from enabled to disabled not only stay 1952238405Sjkim in order with respect to each other, but also have higher priority 1953238405Sjkim than other disabled ciphersuites the next time ciphersuites are 1954238405Sjkim enabled again. 1955238405Sjkim 1956238405Sjkim This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable 1957238405Sjkim the same ciphersuites as with "HIGH" alone, but in a specific 1958238405Sjkim order where the PSK ciphersuites come first (since they are the 1959238405Sjkim most recently disabled ciphersuites when "HIGH" is parsed). 1960238405Sjkim 1961238405Sjkim Also, change ssl_create_cipher_list() (using this new 1962238405Sjkim funcionality) such that between otherwise identical 1963238405Sjkim cihpersuites, ephemeral ECDH is preferred over ephemeral DH in 1964238405Sjkim the default order. 1965238405Sjkim [Bodo Moeller] 1966238405Sjkim 1967238405Sjkim *) Change ssl_create_cipher_list() so that it automatically 1968238405Sjkim arranges the ciphersuites in reasonable order before starting 1969238405Sjkim to process the rule string. Thus, the definition for "DEFAULT" 1970238405Sjkim (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but 1971238405Sjkim remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". 1972238405Sjkim This makes it much easier to arrive at a reasonable default order 1973238405Sjkim in applications for which anonymous ciphers are OK (meaning 1974238405Sjkim that you can't actually use DEFAULT). 1975238405Sjkim [Bodo Moeller; suggested by Victor Duchovni] 1976238405Sjkim 1977238405Sjkim *) Split the SSL/TLS algorithm mask (as used for ciphersuite string 1978238405Sjkim processing) into multiple integers instead of setting 1979238405Sjkim "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", 1980238405Sjkim "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. 1981238405Sjkim (These masks as well as the individual bit definitions are hidden 1982238405Sjkim away into the non-exported interface ssl/ssl_locl.h, so this 1983238405Sjkim change to the definition of the SSL_CIPHER structure shouldn't 1984238405Sjkim affect applications.) This give us more bits for each of these 1985238405Sjkim categories, so there is no longer a need to coagulate AES128 and 1986238405Sjkim AES256 into a single algorithm bit, and to coagulate Camellia128 1987238405Sjkim and Camellia256 into a single algorithm bit, which has led to all 1988238405Sjkim kinds of kludges. 1989238405Sjkim 1990238405Sjkim Thus, among other things, the kludge introduced in 0.9.7m and 1991238405Sjkim 0.9.8e for masking out AES256 independently of AES128 or masking 1992238405Sjkim out Camellia256 independently of AES256 is not needed here in 0.9.9. 1993238405Sjkim 1994238405Sjkim With the change, we also introduce new ciphersuite aliases that 1995238405Sjkim so far were missing: "AES128", "AES256", "CAMELLIA128", and 1996238405Sjkim "CAMELLIA256". 1997238405Sjkim [Bodo Moeller] 1998238405Sjkim 1999238405Sjkim *) Add support for dsa-with-SHA224 and dsa-with-SHA256. 2000238405Sjkim Use the leftmost N bytes of the signature input if the input is 2001238405Sjkim larger than the prime q (with N being the size in bytes of q). 2002238405Sjkim [Nils Larsch] 2003238405Sjkim 2004238405Sjkim *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses 2005238405Sjkim it yet and it is largely untested. 2006238405Sjkim [Steve Henson] 2007238405Sjkim 2008238405Sjkim *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. 2009238405Sjkim [Nils Larsch] 2010238405Sjkim 2011238405Sjkim *) Initial incomplete changes to avoid need for function casts in OpenSSL 2012238405Sjkim some compilers (gcc 4.2 and later) reject their use. Safestack is 2013238405Sjkim reimplemented. Update ASN1 to avoid use of legacy functions. 2014238405Sjkim [Steve Henson] 2015238405Sjkim 2016238405Sjkim *) Win32/64 targets are linked with Winsock2. 2017238405Sjkim [Andy Polyakov] 2018238405Sjkim 2019238405Sjkim *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected 2020238405Sjkim to external functions. This can be used to increase CRL handling 2021238405Sjkim efficiency especially when CRLs are very large by (for example) storing 2022238405Sjkim the CRL revoked certificates in a database. 2023238405Sjkim [Steve Henson] 2024238405Sjkim 2025238405Sjkim *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so 2026238405Sjkim new CRLs added to a directory can be used. New command line option 2027238405Sjkim -verify_return_error to s_client and s_server. This causes real errors 2028238405Sjkim to be returned by the verify callback instead of carrying on no matter 2029238405Sjkim what. This reflects the way a "real world" verify callback would behave. 2030238405Sjkim [Steve Henson] 2031238405Sjkim 2032238405Sjkim *) GOST engine, supporting several GOST algorithms and public key formats. 2033238405Sjkim Kindly donated by Cryptocom. 2034238405Sjkim [Cryptocom] 2035238405Sjkim 2036238405Sjkim *) Partial support for Issuing Distribution Point CRL extension. CRLs 2037238405Sjkim partitioned by DP are handled but no indirect CRL or reason partitioning 2038238405Sjkim (yet). Complete overhaul of CRL handling: now the most suitable CRL is 2039238405Sjkim selected via a scoring technique which handles IDP and AKID in CRLs. 2040238405Sjkim [Steve Henson] 2041238405Sjkim 2042238405Sjkim *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which 2043238405Sjkim will ultimately be used for all verify operations: this will remove the 2044238405Sjkim X509_STORE dependency on certificate verification and allow alternative 2045238405Sjkim lookup methods. X509_STORE based implementations of these two callbacks. 2046238405Sjkim [Steve Henson] 2047238405Sjkim 2048238405Sjkim *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. 2049238405Sjkim Modify get_crl() to find a valid (unexpired) CRL if possible. 2050238405Sjkim [Steve Henson] 2051238405Sjkim 2052238405Sjkim *) New function X509_CRL_match() to check if two CRLs are identical. Normally 2053238405Sjkim this would be called X509_CRL_cmp() but that name is already used by 2054238405Sjkim a function that just compares CRL issuer names. Cache several CRL 2055238405Sjkim extensions in X509_CRL structure and cache CRLDP in X509. 2056238405Sjkim [Steve Henson] 2057238405Sjkim 2058238405Sjkim *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) 2059238405Sjkim this maps equivalent X509_NAME structures into a consistent structure. 2060238405Sjkim Name comparison can then be performed rapidly using memcmp(). 2061238405Sjkim [Steve Henson] 2062238405Sjkim 2063238405Sjkim *) Non-blocking OCSP request processing. Add -timeout option to ocsp 2064238405Sjkim utility. 2065238405Sjkim [Steve Henson] 2066238405Sjkim 2067238405Sjkim *) Allow digests to supply their own micalg string for S/MIME type using 2068238405Sjkim the ctrl EVP_MD_CTRL_MICALG. 2069238405Sjkim [Steve Henson] 2070238405Sjkim 2071238405Sjkim *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the 2072238405Sjkim EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN 2073238405Sjkim ctrl. It can then customise the structure before and/or after signing 2074238405Sjkim if necessary. 2075238405Sjkim [Steve Henson] 2076238405Sjkim 2077238405Sjkim *) New function OBJ_add_sigid() to allow application defined signature OIDs 2078238405Sjkim to be added to OpenSSLs internal tables. New function OBJ_sigid_free() 2079238405Sjkim to free up any added signature OIDs. 2080238405Sjkim [Steve Henson] 2081238405Sjkim 2082238405Sjkim *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), 2083238405Sjkim EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal 2084238405Sjkim digest and cipher tables. New options added to openssl utility: 2085238405Sjkim list-message-digest-algorithms and list-cipher-algorithms. 2086238405Sjkim [Steve Henson] 2087238405Sjkim 2088238405Sjkim *) Change the array representation of binary polynomials: the list 2089238405Sjkim of degrees of non-zero coefficients is now terminated with -1. 2090238405Sjkim Previously it was terminated with 0, which was also part of the 2091238405Sjkim value; thus, the array representation was not applicable to 2092238405Sjkim polynomials where t^0 has coefficient zero. This change makes 2093238405Sjkim the array representation useful in a more general context. 2094238405Sjkim [Douglas Stebila] 2095238405Sjkim 2096238405Sjkim *) Various modifications and fixes to SSL/TLS cipher string 2097238405Sjkim handling. For ECC, the code now distinguishes between fixed ECDH 2098238405Sjkim with RSA certificates on the one hand and with ECDSA certificates 2099238405Sjkim on the other hand, since these are separate ciphersuites. The 2100238405Sjkim unused code for Fortezza ciphersuites has been removed. 2101238405Sjkim 2102238405Sjkim For consistency with EDH, ephemeral ECDH is now called "EECDH" 2103238405Sjkim (not "ECDHE"). For consistency with the code for DH 2104238405Sjkim certificates, use of ECDH certificates is now considered ECDH 2105238405Sjkim authentication, not RSA or ECDSA authentication (the latter is 2106238405Sjkim merely the CA's signing algorithm and not actively used in the 2107238405Sjkim protocol). 2108238405Sjkim 2109238405Sjkim The temporary ciphersuite alias "ECCdraft" is no longer 2110238405Sjkim available, and ECC ciphersuites are no longer excluded from "ALL" 2111238405Sjkim and "DEFAULT". The following aliases now exist for RFC 4492 2112238405Sjkim ciphersuites, most of these by analogy with the DH case: 2113238405Sjkim 2114238405Sjkim kECDHr - ECDH cert, signed with RSA 2115238405Sjkim kECDHe - ECDH cert, signed with ECDSA 2116238405Sjkim kECDH - ECDH cert (signed with either RSA or ECDSA) 2117238405Sjkim kEECDH - ephemeral ECDH 2118238405Sjkim ECDH - ECDH cert or ephemeral ECDH 2119238405Sjkim 2120238405Sjkim aECDH - ECDH cert 2121238405Sjkim aECDSA - ECDSA cert 2122238405Sjkim ECDSA - ECDSA cert 2123238405Sjkim 2124238405Sjkim AECDH - anonymous ECDH 2125238405Sjkim EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") 2126238405Sjkim 2127238405Sjkim [Bodo Moeller] 2128238405Sjkim 2129238405Sjkim *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. 2130238405Sjkim Use correct micalg parameters depending on digest(s) in signed message. 2131238405Sjkim [Steve Henson] 2132238405Sjkim 2133238405Sjkim *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process 2134238405Sjkim an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. 2135238405Sjkim [Steve Henson] 2136238405Sjkim 2137238405Sjkim *) Initial engine support for EVP_PKEY_METHOD. New functions to permit 2138238405Sjkim an engine to register a method. Add ENGINE lookups for methods and 2139238405Sjkim functional reference processing. 2140238405Sjkim [Steve Henson] 2141238405Sjkim 2142238405Sjkim *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of 2143238405Sjkim EVP_{Sign,Verify}* which allow an application to customise the signature 2144238405Sjkim process. 2145238405Sjkim [Steve Henson] 2146238405Sjkim 2147238405Sjkim *) New -resign option to smime utility. This adds one or more signers 2148238405Sjkim to an existing PKCS#7 signedData structure. Also -md option to use an 2149238405Sjkim alternative message digest algorithm for signing. 2150238405Sjkim [Steve Henson] 2151238405Sjkim 2152238405Sjkim *) Tidy up PKCS#7 routines and add new functions to make it easier to 2153238405Sjkim create PKCS7 structures containing multiple signers. Update smime 2154238405Sjkim application to support multiple signers. 2155238405Sjkim [Steve Henson] 2156238405Sjkim 2157238405Sjkim *) New -macalg option to pkcs12 utility to allow setting of an alternative 2158238405Sjkim digest MAC. 2159238405Sjkim [Steve Henson] 2160238405Sjkim 2161238405Sjkim *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. 2162238405Sjkim Reorganize PBE internals to lookup from a static table using NIDs, 2163238405Sjkim add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: 2164238405Sjkim EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative 2165238405Sjkim PRF which will be automatically used with PBES2. 2166238405Sjkim [Steve Henson] 2167238405Sjkim 2168238405Sjkim *) Replace the algorithm specific calls to generate keys in "req" with the 2169238405Sjkim new API. 2170238405Sjkim [Steve Henson] 2171238405Sjkim 2172238405Sjkim *) Update PKCS#7 enveloped data routines to use new API. This is now 2173238405Sjkim supported by any public key method supporting the encrypt operation. A 2174238405Sjkim ctrl is added to allow the public key algorithm to examine or modify 2175238405Sjkim the PKCS#7 RecipientInfo structure if it needs to: for RSA this is 2176238405Sjkim a no op. 2177238405Sjkim [Steve Henson] 2178238405Sjkim 2179238405Sjkim *) Add a ctrl to asn1 method to allow a public key algorithm to express 2180238405Sjkim a default digest type to use. In most cases this will be SHA1 but some 2181238405Sjkim algorithms (such as GOST) need to specify an alternative digest. The 2182238405Sjkim return value indicates how strong the prefernce is 1 means optional and 2183238405Sjkim 2 is mandatory (that is it is the only supported type). Modify 2184238405Sjkim ASN1_item_sign() to accept a NULL digest argument to indicate it should 2185238405Sjkim use the default md. Update openssl utilities to use the default digest 2186238405Sjkim type for signing if it is not explicitly indicated. 2187238405Sjkim [Steve Henson] 2188238405Sjkim 2189238405Sjkim *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New 2190238405Sjkim EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant 2191238405Sjkim signing method from the key type. This effectively removes the link 2192238405Sjkim between digests and public key types. 2193238405Sjkim [Steve Henson] 2194238405Sjkim 2195238405Sjkim *) Add an OID cross reference table and utility functions. Its purpose is to 2196238405Sjkim translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, 2197238405Sjkim rsaEncryption. This will allow some of the algorithm specific hackery 2198238405Sjkim needed to use the correct OID to be removed. 2199238405Sjkim [Steve Henson] 2200238405Sjkim 2201238405Sjkim *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO 2202238405Sjkim structures for PKCS7_sign(). They are now set up by the relevant public 2203238405Sjkim key ASN1 method. 2204238405Sjkim [Steve Henson] 2205238405Sjkim 2206238405Sjkim *) Add provisional EC pkey method with support for ECDSA and ECDH. 2207238405Sjkim [Steve Henson] 2208238405Sjkim 2209238405Sjkim *) Add support for key derivation (agreement) in the API, DH method and 2210238405Sjkim pkeyutl. 2211238405Sjkim [Steve Henson] 2212238405Sjkim 2213238405Sjkim *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support 2214238405Sjkim public and private key formats. As a side effect these add additional 2215238405Sjkim command line functionality not previously available: DSA signatures can be 2216238405Sjkim generated and verified using pkeyutl and DH key support and generation in 2217238405Sjkim pkey, genpkey. 2218238405Sjkim [Steve Henson] 2219238405Sjkim 2220238405Sjkim *) BeOS support. 2221238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 2222238405Sjkim 2223238405Sjkim *) New make target "install_html_docs" installs HTML renditions of the 2224238405Sjkim manual pages. 2225238405Sjkim [Oliver Tappe <zooey@hirschkaefer.de>] 2226238405Sjkim 2227238405Sjkim *) New utility "genpkey" this is analagous to "genrsa" etc except it can 2228238405Sjkim generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to 2229238405Sjkim support key and parameter generation and add initial key generation 2230238405Sjkim functionality for RSA. 2231238405Sjkim [Steve Henson] 2232238405Sjkim 2233238405Sjkim *) Add functions for main EVP_PKEY_method operations. The undocumented 2234238405Sjkim functions EVP_PKEY_{encrypt,decrypt} have been renamed to 2235238405Sjkim EVP_PKEY_{encrypt,decrypt}_old. 2236238405Sjkim [Steve Henson] 2237238405Sjkim 2238238405Sjkim *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public 2239238405Sjkim key API, doesn't do much yet. 2240238405Sjkim [Steve Henson] 2241238405Sjkim 2242238405Sjkim *) New function EVP_PKEY_asn1_get0_info() to retrieve information about 2243238405Sjkim public key algorithms. New option to openssl utility: 2244238405Sjkim "list-public-key-algorithms" to print out info. 2245238405Sjkim [Steve Henson] 2246238405Sjkim 2247238405Sjkim *) Implement the Supported Elliptic Curves Extension for 2248238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 2249238405Sjkim [Douglas Stebila] 2250238405Sjkim 2251238405Sjkim *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or 2252238405Sjkim EVP_CIPHER structures to avoid later problems in EVP_cleanup(). 2253238405Sjkim [Steve Henson] 2254238405Sjkim 2255238405Sjkim *) New utilities pkey and pkeyparam. These are similar to algorithm specific 2256238405Sjkim utilities such as rsa, dsa, dsaparam etc except they process any key 2257238405Sjkim type. 2258238405Sjkim [Steve Henson] 2259238405Sjkim 2260238405Sjkim *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New 2261238405Sjkim functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), 2262238405Sjkim EVP_PKEY_print_param() to print public key data from an EVP_PKEY 2263238405Sjkim structure. 2264238405Sjkim [Steve Henson] 2265238405Sjkim 2266238405Sjkim *) Initial support for pluggable public key ASN1. 2267238405Sjkim De-spaghettify the public key ASN1 handling. Move public and private 2268238405Sjkim key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate 2269238405Sjkim algorithm specific handling to a single module within the relevant 2270238405Sjkim algorithm directory. Add functions to allow (near) opaque processing 2271238405Sjkim of public and private key structures. 2272238405Sjkim [Steve Henson] 2273238405Sjkim 2274238405Sjkim *) Implement the Supported Point Formats Extension for 2275238405Sjkim ECC ciphersuites from draft-ietf-tls-ecc-12.txt. 2276238405Sjkim [Douglas Stebila] 2277238405Sjkim 2278238405Sjkim *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members 2279238405Sjkim for the psk identity [hint] and the psk callback functions to the 2280238405Sjkim SSL_SESSION, SSL and SSL_CTX structure. 2281238405Sjkim 2282238405Sjkim New ciphersuites: 2283238405Sjkim PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, 2284238405Sjkim PSK-AES256-CBC-SHA 2285238405Sjkim 2286238405Sjkim New functions: 2287238405Sjkim SSL_CTX_use_psk_identity_hint 2288238405Sjkim SSL_get_psk_identity_hint 2289238405Sjkim SSL_get_psk_identity 2290238405Sjkim SSL_use_psk_identity_hint 2291238405Sjkim 2292238405Sjkim [Mika Kousa and Pasi Eronen of Nokia Corporation] 2293238405Sjkim 2294238405Sjkim *) Add RFC 3161 compliant time stamp request creation, response generation 2295238405Sjkim and response verification functionality. 2296291721Sjkim [Zolt��n Gl��zik <zglozik@opentsa.org>, The OpenTSA Project] 2297238405Sjkim 2298238405Sjkim *) Add initial support for TLS extensions, specifically for the server_name 2299238405Sjkim extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2300238405Sjkim have new members for a host name. The SSL data structure has an 2301238405Sjkim additional member SSL_CTX *initial_ctx so that new sessions can be 2302238405Sjkim stored in that context to allow for session resumption, even after the 2303238405Sjkim SSL has been switched to a new SSL_CTX in reaction to a client's 2304238405Sjkim server_name extension. 2305238405Sjkim 2306238405Sjkim New functions (subject to change): 2307238405Sjkim 2308238405Sjkim SSL_get_servername() 2309238405Sjkim SSL_get_servername_type() 2310238405Sjkim SSL_set_SSL_CTX() 2311238405Sjkim 2312238405Sjkim New CTRL codes and macros (subject to change): 2313238405Sjkim 2314238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2315238405Sjkim - SSL_CTX_set_tlsext_servername_callback() 2316238405Sjkim SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2317238405Sjkim - SSL_CTX_set_tlsext_servername_arg() 2318238405Sjkim SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2319238405Sjkim 2320238405Sjkim openssl s_client has a new '-servername ...' option. 2321238405Sjkim 2322238405Sjkim openssl s_server has new options '-servername_host ...', '-cert2 ...', 2323238405Sjkim '-key2 ...', '-servername_fatal' (subject to change). This allows 2324238405Sjkim testing the HostName extension for a specific single host name ('-cert' 2325238405Sjkim and '-key' remain fallbacks for handshakes without HostName 2326238405Sjkim negotiation). If the unrecogninzed_name alert has to be sent, this by 2327238405Sjkim default is a warning; it becomes fatal with the '-servername_fatal' 2328238405Sjkim option. 2329238405Sjkim 2330238405Sjkim [Peter Sylvester, Remy Allais, Christophe Renou] 2331238405Sjkim 2332238405Sjkim *) Whirlpool hash implementation is added. 2333238405Sjkim [Andy Polyakov] 2334238405Sjkim 2335238405Sjkim *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to 2336238405Sjkim bn(64,32). Because of instruction set limitations it doesn't have 2337238405Sjkim any negative impact on performance. This was done mostly in order 2338238405Sjkim to make it possible to share assembler modules, such as bn_mul_mont 2339238405Sjkim implementations, between 32- and 64-bit builds without hassle. 2340238405Sjkim [Andy Polyakov] 2341238405Sjkim 2342238405Sjkim *) Move code previously exiled into file crypto/ec/ec2_smpt.c 2343238405Sjkim to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP 2344238405Sjkim macro. 2345238405Sjkim [Bodo Moeller] 2346238405Sjkim 2347238405Sjkim *) New candidate for BIGNUM assembler implementation, bn_mul_mont, 2348238405Sjkim dedicated Montgomery multiplication procedure, is introduced. 2349238405Sjkim BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher 2350238405Sjkim "64-bit" performance on certain 32-bit targets. 2351238405Sjkim [Andy Polyakov] 2352238405Sjkim 2353238405Sjkim *) New option SSL_OP_NO_COMP to disable use of compression selectively 2354238405Sjkim in SSL structures. New SSL ctrl to set maximum send fragment size. 2355238405Sjkim Save memory by seeting the I/O buffer sizes dynamically instead of 2356238405Sjkim using the maximum available value. 2357238405Sjkim [Steve Henson] 2358238405Sjkim 2359238405Sjkim *) New option -V for 'openssl ciphers'. This prints the ciphersuite code 2360238405Sjkim in addition to the text details. 2361238405Sjkim [Bodo Moeller] 2362238405Sjkim 2363238405Sjkim *) Very, very preliminary EXPERIMENTAL support for printing of general 2364238405Sjkim ASN1 structures. This currently produces rather ugly output and doesn't 2365238405Sjkim handle several customised structures at all. 2366238405Sjkim [Steve Henson] 2367238405Sjkim 2368238405Sjkim *) Integrated support for PVK file format and some related formats such 2369238405Sjkim as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support 2370238405Sjkim these in the 'rsa' and 'dsa' utilities. 2371238405Sjkim [Steve Henson] 2372238405Sjkim 2373238405Sjkim *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. 2374238405Sjkim [Steve Henson] 2375238405Sjkim 2376238405Sjkim *) Remove the ancient ASN1_METHOD code. This was only ever used in one 2377238405Sjkim place for the (very old) "NETSCAPE" format certificates which are now 2378238405Sjkim handled using new ASN1 code equivalents. 2379238405Sjkim [Steve Henson] 2380238405Sjkim 2381238405Sjkim *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD 2382238405Sjkim pointer and make the SSL_METHOD parameter in SSL_CTX_new, 2383238405Sjkim SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. 2384238405Sjkim [Nils Larsch] 2385238405Sjkim 2386238405Sjkim *) Modify CRL distribution points extension code to print out previously 2387238405Sjkim unsupported fields. Enhance extension setting code to allow setting of 2388238405Sjkim all fields. 2389238405Sjkim [Steve Henson] 2390238405Sjkim 2391238405Sjkim *) Add print and set support for Issuing Distribution Point CRL extension. 2392238405Sjkim [Steve Henson] 2393238405Sjkim 2394238405Sjkim *) Change 'Configure' script to enable Camellia by default. 2395238405Sjkim [NTT] 2396238405Sjkim 2397206046Ssimon Changes between 0.9.8m and 0.9.8n [24 Mar 2010] 2398206046Ssimon 2399206046Ssimon *) When rejecting SSL/TLS records due to an incorrect version number, never 2400206046Ssimon update s->server with a new major version number. As of 2401206046Ssimon - OpenSSL 0.9.8m if 'short' is a 16-bit type, 2402206046Ssimon - OpenSSL 0.9.8f if 'short' is longer than 16 bits, 2403206046Ssimon the previous behavior could result in a read attempt at NULL when 2404206046Ssimon receiving specific incorrect SSL/TLS records once record payload 2405206046Ssimon protection is active. (CVE-2010-0740) 2406206046Ssimon [Bodo Moeller, Adam Langley <agl@chromium.org>] 2407206046Ssimon 2408206046Ssimon *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL 2409206046Ssimon could be crashed if the relevant tables were not present (e.g. chrooted). 2410206046Ssimon [Tomas Hoger <thoger@redhat.com>] 2411206046Ssimon 2412205128Ssimon Changes between 0.9.8l and 0.9.8m [25 Feb 2010] 2413205128Ssimon 2414205128Ssimon *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) 2415205128Ssimon [Martin Olsson, Neel Mehta] 2416205128Ssimon 2417205128Ssimon *) Fix X509_STORE locking: Every 'objs' access requires a lock (to 2418205128Ssimon accommodate for stack sorting, always a write lock!). 2419205128Ssimon [Bodo Moeller] 2420205128Ssimon 2421205128Ssimon *) On some versions of WIN32 Heap32Next is very slow. This can cause 2422205128Ssimon excessive delays in the RAND_poll(): over a minute. As a workaround 2423205128Ssimon include a time check in the inner Heap32Next loop too. 2424205128Ssimon [Steve Henson] 2425205128Ssimon 2426205128Ssimon *) The code that handled flushing of data in SSL/TLS originally used the 2427205128Ssimon BIO_CTRL_INFO ctrl to see if any data was pending first. This caused 2428205128Ssimon the problem outlined in PR#1949. The fix suggested there however can 2429205128Ssimon trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions 2430205128Ssimon of Apache). So instead simplify the code to flush unconditionally. 2431205128Ssimon This should be fine since flushing with no data to flush is a no op. 2432205128Ssimon [Steve Henson] 2433205128Ssimon 2434205128Ssimon *) Handle TLS versions 2.0 and later properly and correctly use the 2435205128Ssimon highest version of TLS/SSL supported. Although TLS >= 2.0 is some way 2436205128Ssimon off ancient servers have a habit of sticking around for a while... 2437205128Ssimon [Steve Henson] 2438205128Ssimon 2439205128Ssimon *) Modify compression code so it frees up structures without using the 2440205128Ssimon ex_data callbacks. This works around a problem where some applications 2441205128Ssimon call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when 2442205128Ssimon restarting) then use compression (e.g. SSL with compression) later. 2443205128Ssimon This results in significant per-connection memory leaks and 2444205128Ssimon has caused some security issues including CVE-2008-1678 and 2445205128Ssimon CVE-2009-4355. 2446205128Ssimon [Steve Henson] 2447205128Ssimon 2448205128Ssimon *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't 2449205128Ssimon change when encrypting or decrypting. 2450205128Ssimon [Bodo Moeller] 2451205128Ssimon 2452205128Ssimon *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to 2453205128Ssimon connect and renegotiate with servers which do not support RI. 2454205128Ssimon Until RI is more widely deployed this option is enabled by default. 2455205128Ssimon [Steve Henson] 2456205128Ssimon 2457205128Ssimon *) Add "missing" ssl ctrls to clear options and mode. 2458205128Ssimon [Steve Henson] 2459205128Ssimon 2460205128Ssimon *) If client attempts to renegotiate and doesn't support RI respond with 2461205128Ssimon a no_renegotiation alert as required by RFC5746. Some renegotiating 2462205128Ssimon TLS clients will continue a connection gracefully when they receive 2463205128Ssimon the alert. Unfortunately OpenSSL mishandled this alert and would hang 2464205128Ssimon waiting for a server hello which it will never receive. Now we treat a 2465205128Ssimon received no_renegotiation alert as a fatal error. This is because 2466205128Ssimon applications requesting a renegotiation might well expect it to succeed 2467205128Ssimon and would have no code in place to handle the server denying it so the 2468205128Ssimon only safe thing to do is to terminate the connection. 2469205128Ssimon [Steve Henson] 2470205128Ssimon 2471205128Ssimon *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if 2472205128Ssimon peer supports secure renegotiation and 0 otherwise. Print out peer 2473205128Ssimon renegotiation support in s_client/s_server. 2474205128Ssimon [Steve Henson] 2475205128Ssimon 2476205128Ssimon *) Replace the highly broken and deprecated SPKAC certification method with 2477205128Ssimon the updated NID creation version. This should correctly handle UTF8. 2478205128Ssimon [Steve Henson] 2479205128Ssimon 2480205128Ssimon *) Implement RFC5746. Re-enable renegotiation but require the extension 2481205128Ssimon as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION 2482205128Ssimon turns out to be a bad idea. It has been replaced by 2483205128Ssimon SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with 2484205128Ssimon SSL_CTX_set_options(). This is really not recommended unless you 2485205128Ssimon know what you are doing. 2486205128Ssimon [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] 2487205128Ssimon 2488205128Ssimon *) Fixes to stateless session resumption handling. Use initial_ctx when 2489205128Ssimon issuing and attempting to decrypt tickets in case it has changed during 2490205128Ssimon servername handling. Use a non-zero length session ID when attempting 2491205128Ssimon stateless session resumption: this makes it possible to determine if 2492205128Ssimon a resumption has occurred immediately after receiving server hello 2493205128Ssimon (several places in OpenSSL subtly assume this) instead of later in 2494205128Ssimon the handshake. 2495205128Ssimon [Steve Henson] 2496205128Ssimon 2497205128Ssimon *) The functions ENGINE_ctrl(), OPENSSL_isservice(), 2498205128Ssimon CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error 2499205128Ssimon fixes for a few places where the return code is not checked 2500205128Ssimon correctly. 2501205128Ssimon [Julia Lawall <julia@diku.dk>] 2502205128Ssimon 2503205128Ssimon *) Add --strict-warnings option to Configure script to include devteam 2504205128Ssimon warnings in other configurations. 2505205128Ssimon [Steve Henson] 2506205128Ssimon 2507205128Ssimon *) Add support for --libdir option and LIBDIR variable in makefiles. This 2508205128Ssimon makes it possible to install openssl libraries in locations which 2509205128Ssimon have names other than "lib", for example "/usr/lib64" which some 2510205128Ssimon systems need. 2511205128Ssimon [Steve Henson, based on patch from Jeremy Utley] 2512205128Ssimon 2513205128Ssimon *) Don't allow the use of leading 0x80 in OIDs. This is a violation of 2514205128Ssimon X690 8.9.12 and can produce some misleading textual output of OIDs. 2515205128Ssimon [Steve Henson, reported by Dan Kaminsky] 2516205128Ssimon 2517205128Ssimon *) Delete MD2 from algorithm tables. This follows the recommendation in 2518205128Ssimon several standards that it is not used in new applications due to 2519205128Ssimon several cryptographic weaknesses. For binary compatibility reasons 2520205128Ssimon the MD2 API is still compiled in by default. 2521205128Ssimon [Steve Henson] 2522205128Ssimon 2523205128Ssimon *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved 2524205128Ssimon and restored. 2525205128Ssimon [Steve Henson] 2526205128Ssimon 2527205128Ssimon *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and 2528205128Ssimon OPENSSL_asc2uni conditionally on Netware platforms to avoid a name 2529205128Ssimon clash. 2530205128Ssimon [Guenter <lists@gknw.net>] 2531205128Ssimon 2532205128Ssimon *) Fix the server certificate chain building code to use X509_verify_cert(), 2533205128Ssimon it used to have an ad-hoc builder which was unable to cope with anything 2534205128Ssimon other than a simple chain. 2535205128Ssimon [David Woodhouse <dwmw2@infradead.org>, Steve Henson] 2536205128Ssimon 2537205128Ssimon *) Don't check self signed certificate signatures in X509_verify_cert() 2538205128Ssimon by default (a flag can override this): it just wastes time without 2539205128Ssimon adding any security. As a useful side effect self signed root CAs 2540205128Ssimon with non-FIPS digests are now usable in FIPS mode. 2541205128Ssimon [Steve Henson] 2542205128Ssimon 2543205128Ssimon *) In dtls1_process_out_of_seq_message() the check if the current message 2544205128Ssimon is already buffered was missing. For every new message was memory 2545205128Ssimon allocated, allowing an attacker to perform an denial of service attack 2546205128Ssimon with sending out of seq handshake messages until there is no memory 2547205128Ssimon left. Additionally every future messege was buffered, even if the 2548205128Ssimon sequence number made no sense and would be part of another handshake. 2549205128Ssimon So only messages with sequence numbers less than 10 in advance will be 2550205128Ssimon buffered. (CVE-2009-1378) 2551205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2552205128Ssimon 2553205128Ssimon *) Records are buffered if they arrive with a future epoch to be 2554205128Ssimon processed after finishing the corresponding handshake. There is 2555205128Ssimon currently no limitation to this buffer allowing an attacker to perform 2556205128Ssimon a DOS attack with sending records with future epochs until there is no 2557205128Ssimon memory left. This patch adds the pqueue_size() function to detemine 2558205128Ssimon the size of a buffer and limits the record buffer to 100 entries. 2559205128Ssimon (CVE-2009-1377) 2560205128Ssimon [Robin Seggelmann, discovered by Daniel Mentz] 2561205128Ssimon 2562205128Ssimon *) Keep a copy of frag->msg_header.frag_len so it can be used after the 2563205128Ssimon parent structure is freed. (CVE-2009-1379) 2564205128Ssimon [Daniel Mentz] 2565205128Ssimon 2566205128Ssimon *) Handle non-blocking I/O properly in SSL_shutdown() call. 2567205128Ssimon [Darryl Miles <darryl-mailinglists@netbauds.net>] 2568205128Ssimon 2569205128Ssimon *) Add 2.5.4.* OIDs 2570205128Ssimon [Ilya O. <vrghost@gmail.com>] 2571205128Ssimon 2572205128Ssimon Changes between 0.9.8k and 0.9.8l [5 Nov 2009] 2573205128Ssimon 2574205128Ssimon *) Disable renegotiation completely - this fixes a severe security 2575205128Ssimon problem (CVE-2009-3555) at the cost of breaking all 2576205128Ssimon renegotiation. Renegotiation can be re-enabled by setting 2577205128Ssimon SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at 2578205128Ssimon run-time. This is really not recommended unless you know what 2579205128Ssimon you're doing. 2580205128Ssimon [Ben Laurie] 2581205128Ssimon 2582194206Ssimon Changes between 0.9.8j and 0.9.8k [25 Mar 2009] 2583167612Ssimon 2584194206Ssimon *) Don't set val to NULL when freeing up structures, it is freed up by 2585194206Ssimon underlying code. If sizeof(void *) > sizeof(long) this can result in 2586194206Ssimon zeroing past the valid field. (CVE-2009-0789) 2587194206Ssimon [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] 2588194206Ssimon 2589194206Ssimon *) Fix bug where return value of CMS_SignerInfo_verify_content() was not 2590194206Ssimon checked correctly. This would allow some invalid signed attributes to 2591194206Ssimon appear to verify correctly. (CVE-2009-0591) 2592194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2593194206Ssimon 2594194206Ssimon *) Reject UniversalString and BMPString types with invalid lengths. This 2595194206Ssimon prevents a crash in ASN1_STRING_print_ex() which assumes the strings have 2596194206Ssimon a legal length. (CVE-2009-0590) 2597194206Ssimon [Steve Henson] 2598194206Ssimon 2599194206Ssimon *) Set S/MIME signing as the default purpose rather than setting it 2600194206Ssimon unconditionally. This allows applications to override it at the store 2601194206Ssimon level. 2602194206Ssimon [Steve Henson] 2603194206Ssimon 2604194206Ssimon *) Permit restricted recursion of ASN1 strings. This is needed in practice 2605194206Ssimon to handle some structures. 2606194206Ssimon [Steve Henson] 2607194206Ssimon 2608194206Ssimon *) Improve efficiency of mem_gets: don't search whole buffer each time 2609194206Ssimon for a '\n' 2610194206Ssimon [Jeremy Shapiro <jnshapir@us.ibm.com>] 2611194206Ssimon 2612194206Ssimon *) New -hex option for openssl rand. 2613194206Ssimon [Matthieu Herrb] 2614194206Ssimon 2615194206Ssimon *) Print out UTF8String and NumericString when parsing ASN1. 2616194206Ssimon [Steve Henson] 2617194206Ssimon 2618194206Ssimon *) Support NumericString type for name components. 2619194206Ssimon [Steve Henson] 2620194206Ssimon 2621194206Ssimon *) Allow CC in the environment to override the automatically chosen 2622194206Ssimon compiler. Note that nothing is done to ensure flags work with the 2623194206Ssimon chosen compiler. 2624194206Ssimon [Ben Laurie] 2625194206Ssimon 2626194206Ssimon Changes between 0.9.8i and 0.9.8j [07 Jan 2009] 2627194206Ssimon 2628194206Ssimon *) Properly check EVP_VerifyFinal() and similar return values 2629194206Ssimon (CVE-2008-5077). 2630194206Ssimon [Ben Laurie, Bodo Moeller, Google Security Team] 2631194206Ssimon 2632194206Ssimon *) Enable TLS extensions by default. 2633194206Ssimon [Ben Laurie] 2634194206Ssimon 2635194206Ssimon *) Allow the CHIL engine to be loaded, whether the application is 2636194206Ssimon multithreaded or not. (This does not release the developer from the 2637194206Ssimon obligation to set up the dynamic locking callbacks.) 2638194206Ssimon [Sander Temme <sander@temme.net>] 2639194206Ssimon 2640194206Ssimon *) Use correct exit code if there is an error in dgst command. 2641194206Ssimon [Steve Henson; problem pointed out by Roland Dirlewanger] 2642194206Ssimon 2643194206Ssimon *) Tweak Configure so that you need to say "experimental-jpake" to enable 2644194206Ssimon JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. 2645194206Ssimon [Bodo Moeller] 2646194206Ssimon 2647194206Ssimon *) Add experimental JPAKE support, including demo authentication in 2648194206Ssimon s_client and s_server. 2649194206Ssimon [Ben Laurie] 2650194206Ssimon 2651194206Ssimon *) Set the comparison function in v3_addr_canonize(). 2652194206Ssimon [Rob Austein <sra@hactrn.net>] 2653194206Ssimon 2654194206Ssimon *) Add support for XMPP STARTTLS in s_client. 2655194206Ssimon [Philip Paeps <philip@freebsd.org>] 2656194206Ssimon 2657194206Ssimon *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior 2658194206Ssimon to ensure that even with this option, only ciphersuites in the 2659194206Ssimon server's preference list will be accepted. (Note that the option 2660194206Ssimon applies only when resuming a session, so the earlier behavior was 2661194206Ssimon just about the algorithm choice for symmetric cryptography.) 2662194206Ssimon [Bodo Moeller] 2663194206Ssimon 2664194206Ssimon Changes between 0.9.8h and 0.9.8i [15 Sep 2008] 2665194206Ssimon 2666205128Ssimon *) Fix NULL pointer dereference if a DTLS server received 2667205128Ssimon ChangeCipherSpec as first record (CVE-2009-1386). 2668205128Ssimon [PR #1679] 2669205128Ssimon 2670194206Ssimon *) Fix a state transitition in s3_srvr.c and d1_srvr.c 2671194206Ssimon (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). 2672194206Ssimon [Nagendra Modadugu] 2673194206Ssimon 2674194206Ssimon *) The fix in 0.9.8c that supposedly got rid of unsafe 2675194206Ssimon double-checked locking was incomplete for RSA blinding, 2676194206Ssimon addressing just one layer of what turns out to have been 2677194206Ssimon doubly unsafe triple-checked locking. 2678194206Ssimon 2679194206Ssimon So now fix this for real by retiring the MONT_HELPER macro 2680194206Ssimon in crypto/rsa/rsa_eay.c. 2681194206Ssimon 2682194206Ssimon [Bodo Moeller; problem pointed out by Marius Schilder] 2683194206Ssimon 2684194206Ssimon *) Various precautionary measures: 2685194206Ssimon 2686194206Ssimon - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). 2687194206Ssimon 2688194206Ssimon - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). 2689194206Ssimon (NB: This would require knowledge of the secret session ticket key 2690194206Ssimon to exploit, in which case you'd be SOL either way.) 2691194206Ssimon 2692194206Ssimon - Change bn_nist.c so that it will properly handle input BIGNUMs 2693194206Ssimon outside the expected range. 2694194206Ssimon 2695194206Ssimon - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG 2696194206Ssimon builds. 2697194206Ssimon 2698194206Ssimon [Neel Mehta, Bodo Moeller] 2699194206Ssimon 2700194206Ssimon *) Allow engines to be "soft loaded" - i.e. optionally don't die if 2701194206Ssimon the load fails. Useful for distros. 2702194206Ssimon [Ben Laurie and the FreeBSD team] 2703194206Ssimon 2704194206Ssimon *) Add support for Local Machine Keyset attribute in PKCS#12 files. 2705194206Ssimon [Steve Henson] 2706194206Ssimon 2707194206Ssimon *) Fix BN_GF2m_mod_arr() top-bit cleanup code. 2708194206Ssimon [Huang Ying] 2709194206Ssimon 2710194206Ssimon *) Expand ENGINE to support engine supplied SSL client certificate functions. 2711194206Ssimon 2712194206Ssimon This work was sponsored by Logica. 2713194206Ssimon [Steve Henson] 2714194206Ssimon 2715194206Ssimon *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows 2716194206Ssimon keystores. Support for SSL/TLS client authentication too. 2717194206Ssimon Not compiled unless enable-capieng specified to Configure. 2718194206Ssimon 2719194206Ssimon This work was sponsored by Logica. 2720194206Ssimon [Steve Henson] 2721194206Ssimon 2722194206Ssimon *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using 2723194206Ssimon ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain 2724194206Ssimon attribute creation routines such as certifcate requests and PKCS#12 2725194206Ssimon files. 2726194206Ssimon [Steve Henson] 2727194206Ssimon 2728194206Ssimon Changes between 0.9.8g and 0.9.8h [28 May 2008] 2729194206Ssimon 2730194206Ssimon *) Fix flaw if 'Server Key exchange message' is omitted from a TLS 2731194206Ssimon handshake which could lead to a cilent crash as found using the 2732194206Ssimon Codenomicon TLS test suite (CVE-2008-1672) 2733194206Ssimon [Steve Henson, Mark Cox] 2734194206Ssimon 2735194206Ssimon *) Fix double free in TLS server name extensions which could lead to 2736194206Ssimon a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) 2737194206Ssimon [Joe Orton] 2738194206Ssimon 2739194206Ssimon *) Clear error queue in SSL_CTX_use_certificate_chain_file() 2740194206Ssimon 2741194206Ssimon Clear the error queue to ensure that error entries left from 2742194206Ssimon older function calls do not interfere with the correct operation. 2743194206Ssimon [Lutz Jaenicke, Erik de Castro Lopo] 2744194206Ssimon 2745194206Ssimon *) Remove root CA certificates of commercial CAs: 2746194206Ssimon 2747194206Ssimon The OpenSSL project does not recommend any specific CA and does not 2748194206Ssimon have any policy with respect to including or excluding any CA. 2749194206Ssimon Therefore it does not make any sense to ship an arbitrary selection 2750194206Ssimon of root CA certificates with the OpenSSL software. 2751194206Ssimon [Lutz Jaenicke] 2752194206Ssimon 2753194206Ssimon *) RSA OAEP patches to fix two separate invalid memory reads. 2754194206Ssimon The first one involves inputs when 'lzero' is greater than 2755194206Ssimon 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes 2756194206Ssimon before the beginning of from). The second one involves inputs where 2757194206Ssimon the 'db' section contains nothing but zeroes (there is a one-byte 2758194206Ssimon invalid read after the end of 'db'). 2759194206Ssimon [Ivan Nestlerode <inestlerode@us.ibm.com>] 2760194206Ssimon 2761194206Ssimon *) Partial backport from 0.9.9-dev: 2762194206Ssimon 2763194206Ssimon Introduce bn_mul_mont (dedicated Montgomery multiplication 2764194206Ssimon procedure) as a candidate for BIGNUM assembler implementation. 2765194206Ssimon While 0.9.9-dev uses assembler for various architectures, only 2766194206Ssimon x86_64 is available by default here in the 0.9.8 branch, and 2767194206Ssimon 32-bit x86 is available through a compile-time setting. 2768194206Ssimon 2769194206Ssimon To try the 32-bit x86 assembler implementation, use Configure 2770194206Ssimon option "enable-montasm" (which exists only for this backport). 2771194206Ssimon 2772194206Ssimon As "enable-montasm" for 32-bit x86 disclaims code stability 2773194206Ssimon anyway, in this constellation we activate additional code 2774194206Ssimon backported from 0.9.9-dev for further performance improvements, 2775194206Ssimon namely BN_from_montgomery_word. (To enable this otherwise, 2776194206Ssimon e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) 2777194206Ssimon 2778194206Ssimon [Andy Polyakov (backport partially by Bodo Moeller)] 2779194206Ssimon 2780194206Ssimon *) Add TLS session ticket callback. This allows an application to set 2781194206Ssimon TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed 2782194206Ssimon values. This is useful for key rollover for example where several key 2783194206Ssimon sets may exist with different names. 2784194206Ssimon [Steve Henson] 2785194206Ssimon 2786194206Ssimon *) Reverse ENGINE-internal logic for caching default ENGINE handles. 2787194206Ssimon This was broken until now in 0.9.8 releases, such that the only way 2788194206Ssimon a registered ENGINE could be used (assuming it initialises 2789194206Ssimon successfully on the host) was to explicitly set it as the default 2790194206Ssimon for the relevant algorithms. This is in contradiction with 0.9.7 2791194206Ssimon behaviour and the documentation. With this fix, when an ENGINE is 2792194206Ssimon registered into a given algorithm's table of implementations, the 2793194206Ssimon 'uptodate' flag is reset so that auto-discovery will be used next 2794194206Ssimon time a new context for that algorithm attempts to select an 2795194206Ssimon implementation. 2796194206Ssimon [Ian Lister (tweaked by Geoff Thorpe)] 2797194206Ssimon 2798194206Ssimon *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 2799194206Ssimon implemention in the following ways: 2800194206Ssimon 2801194206Ssimon Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be 2802194206Ssimon hard coded. 2803194206Ssimon 2804194206Ssimon Lack of BER streaming support means one pass streaming processing is 2805194206Ssimon only supported if data is detached: setting the streaming flag is 2806194206Ssimon ignored for embedded content. 2807194206Ssimon 2808194206Ssimon CMS support is disabled by default and must be explicitly enabled 2809194206Ssimon with the enable-cms configuration option. 2810194206Ssimon [Steve Henson] 2811194206Ssimon 2812194206Ssimon *) Update the GMP engine glue to do direct copies between BIGNUM and 2813194206Ssimon mpz_t when openssl and GMP use the same limb size. Otherwise the 2814194206Ssimon existing "conversion via a text string export" trick is still used. 2815194206Ssimon [Paul Sheer <paulsheer@gmail.com>] 2816194206Ssimon 2817194206Ssimon *) Zlib compression BIO. This is a filter BIO which compressed and 2818194206Ssimon uncompresses any data passed through it. 2819194206Ssimon [Steve Henson] 2820194206Ssimon 2821194206Ssimon *) Add AES_wrap_key() and AES_unwrap_key() functions to implement 2822194206Ssimon RFC3394 compatible AES key wrapping. 2823194206Ssimon [Steve Henson] 2824194206Ssimon 2825194206Ssimon *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): 2826194206Ssimon sets string data without copying. X509_ALGOR_set0() and 2827194206Ssimon X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) 2828194206Ssimon data. Attribute function X509at_get0_data_by_OBJ(): retrieves data 2829194206Ssimon from an X509_ATTRIBUTE structure optionally checking it occurs only 2830194206Ssimon once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied 2831194206Ssimon data. 2832194206Ssimon [Steve Henson] 2833194206Ssimon 2834194206Ssimon *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() 2835194206Ssimon to get the expected BN_FLG_CONSTTIME behavior. 2836194206Ssimon [Bodo Moeller (Google)] 2837194206Ssimon 2838194206Ssimon *) Netware support: 2839194206Ssimon 2840194206Ssimon - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets 2841194206Ssimon - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) 2842194206Ssimon - added some more tests to do_tests.pl 2843194206Ssimon - fixed RunningProcess usage so that it works with newer LIBC NDKs too 2844194206Ssimon - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency 2845194206Ssimon - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, 2846194206Ssimon netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc 2847194206Ssimon - various changes to netware.pl to enable gcc-cross builds on Win32 2848194206Ssimon platform 2849194206Ssimon - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) 2850194206Ssimon - various changes to fix missing prototype warnings 2851194206Ssimon - fixed x86nasm.pl to create correct asm files for NASM COFF output 2852194206Ssimon - added AES, WHIRLPOOL and CPUID assembler code to build files 2853194206Ssimon - added missing AES assembler make rules to mk1mf.pl 2854194206Ssimon - fixed order of includes in apps/ocsp.c so that e_os.h settings apply 2855194206Ssimon [Guenter Knauf <eflash@gmx.net>] 2856194206Ssimon 2857194206Ssimon *) Implement certificate status request TLS extension defined in RFC3546. 2858194206Ssimon A client can set the appropriate parameters and receive the encoded 2859194206Ssimon OCSP response via a callback. A server can query the supplied parameters 2860194206Ssimon and set the encoded OCSP response in the callback. Add simplified examples 2861194206Ssimon to s_client and s_server. 2862194206Ssimon [Steve Henson] 2863194206Ssimon 2864194206Ssimon Changes between 0.9.8f and 0.9.8g [19 Oct 2007] 2865194206Ssimon 2866194206Ssimon *) Fix various bugs: 2867194206Ssimon + Binary incompatibility of ssl_ctx_st structure 2868194206Ssimon + DTLS interoperation with non-compliant servers 2869194206Ssimon + Don't call get_session_cb() without proposed session 2870194206Ssimon + Fix ia64 assembler code 2871194206Ssimon [Andy Polyakov, Steve Henson] 2872194206Ssimon 2873194206Ssimon Changes between 0.9.8e and 0.9.8f [11 Oct 2007] 2874194206Ssimon 2875194206Ssimon *) DTLS Handshake overhaul. There were longstanding issues with 2876194206Ssimon OpenSSL DTLS implementation, which were making it impossible for 2877194206Ssimon RFC 4347 compliant client to communicate with OpenSSL server. 2878194206Ssimon Unfortunately just fixing these incompatibilities would "cut off" 2879194206Ssimon pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e 2880194206Ssimon server keeps tolerating non RFC compliant syntax. The opposite is 2881194206Ssimon not true, 0.9.8f client can not communicate with earlier server. 2882194206Ssimon This update even addresses CVE-2007-4995. 2883194206Ssimon [Andy Polyakov] 2884194206Ssimon 2885194206Ssimon *) Changes to avoid need for function casts in OpenSSL: some compilers 2886194206Ssimon (gcc 4.2 and later) reject their use. 2887194206Ssimon [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, 2888194206Ssimon Steve Henson] 2889194206Ssimon 2890194206Ssimon *) Add RFC4507 support to OpenSSL. This includes the corrections in 2891194206Ssimon RFC4507bis. The encrypted ticket format is an encrypted encoded 2892194206Ssimon SSL_SESSION structure, that way new session features are automatically 2893194206Ssimon supported. 2894194206Ssimon 2895194206Ssimon If a client application caches session in an SSL_SESSION structure 2896194206Ssimon support is transparent because tickets are now stored in the encoded 2897194206Ssimon SSL_SESSION. 2898194206Ssimon 2899194206Ssimon The SSL_CTX structure automatically generates keys for ticket 2900194206Ssimon protection in servers so again support should be possible 2901194206Ssimon with no application modification. 2902194206Ssimon 2903194206Ssimon If a client or server wishes to disable RFC4507 support then the option 2904194206Ssimon SSL_OP_NO_TICKET can be set. 2905194206Ssimon 2906194206Ssimon Add a TLS extension debugging callback to allow the contents of any client 2907194206Ssimon or server extensions to be examined. 2908194206Ssimon 2909194206Ssimon This work was sponsored by Google. 2910194206Ssimon [Steve Henson] 2911194206Ssimon 2912194206Ssimon *) Add initial support for TLS extensions, specifically for the server_name 2913194206Ssimon extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now 2914194206Ssimon have new members for a host name. The SSL data structure has an 2915194206Ssimon additional member SSL_CTX *initial_ctx so that new sessions can be 2916194206Ssimon stored in that context to allow for session resumption, even after the 2917194206Ssimon SSL has been switched to a new SSL_CTX in reaction to a client's 2918194206Ssimon server_name extension. 2919194206Ssimon 2920194206Ssimon New functions (subject to change): 2921194206Ssimon 2922194206Ssimon SSL_get_servername() 2923194206Ssimon SSL_get_servername_type() 2924194206Ssimon SSL_set_SSL_CTX() 2925194206Ssimon 2926194206Ssimon New CTRL codes and macros (subject to change): 2927194206Ssimon 2928194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_CB 2929194206Ssimon - SSL_CTX_set_tlsext_servername_callback() 2930194206Ssimon SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG 2931194206Ssimon - SSL_CTX_set_tlsext_servername_arg() 2932194206Ssimon SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() 2933194206Ssimon 2934194206Ssimon openssl s_client has a new '-servername ...' option. 2935194206Ssimon 2936194206Ssimon openssl s_server has new options '-servername_host ...', '-cert2 ...', 2937194206Ssimon '-key2 ...', '-servername_fatal' (subject to change). This allows 2938194206Ssimon testing the HostName extension for a specific single host name ('-cert' 2939194206Ssimon and '-key' remain fallbacks for handshakes without HostName 2940194206Ssimon negotiation). If the unrecogninzed_name alert has to be sent, this by 2941194206Ssimon default is a warning; it becomes fatal with the '-servername_fatal' 2942194206Ssimon option. 2943194206Ssimon 2944194206Ssimon [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] 2945194206Ssimon 2946194206Ssimon *) Add AES and SSE2 assembly language support to VC++ build. 2947194206Ssimon [Steve Henson] 2948194206Ssimon 2949194206Ssimon *) Mitigate attack on final subtraction in Montgomery reduction. 2950194206Ssimon [Andy Polyakov] 2951194206Ssimon 2952194206Ssimon *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 2953194206Ssimon (which previously caused an internal error). 2954194206Ssimon [Bodo Moeller] 2955194206Ssimon 2956194206Ssimon *) Squeeze another 10% out of IGE mode when in != out. 2957194206Ssimon [Ben Laurie] 2958194206Ssimon 2959194206Ssimon *) AES IGE mode speedup. 2960194206Ssimon [Dean Gaudet (Google)] 2961194206Ssimon 2962194206Ssimon *) Add the Korean symmetric 128-bit cipher SEED (see 2963194206Ssimon http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and 2964194206Ssimon add SEED ciphersuites from RFC 4162: 2965194206Ssimon 2966194206Ssimon TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" 2967194206Ssimon TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" 2968194206Ssimon TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" 2969194206Ssimon TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" 2970194206Ssimon 2971194206Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 2972194206Ssimon series, SEED remains excluded from compilation unless OpenSSL 2973194206Ssimon is configured with 'enable-seed'. 2974194206Ssimon [KISA, Bodo Moeller] 2975194206Ssimon 2976194206Ssimon *) Mitigate branch prediction attacks, which can be practical if a 2977194206Ssimon single processor is shared, allowing a spy process to extract 2978194206Ssimon information. For detailed background information, see 2979194206Ssimon http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, 2980194206Ssimon J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL 2981194206Ssimon and Necessary Software Countermeasures"). The core of the change 2982194206Ssimon are new versions BN_div_no_branch() and 2983194206Ssimon BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), 2984194206Ssimon respectively, which are slower, but avoid the security-relevant 2985194206Ssimon conditional branches. These are automatically called by BN_div() 2986194206Ssimon and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one 2987194206Ssimon of the input BIGNUMs. Also, BN_is_bit_set() has been changed to 2988194206Ssimon remove a conditional branch. 2989194206Ssimon 2990194206Ssimon BN_FLG_CONSTTIME is the new name for the previous 2991194206Ssimon BN_FLG_EXP_CONSTTIME flag, since it now affects more than just 2992194206Ssimon modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag 2993194206Ssimon in the exponent causes BN_mod_exp_mont() to use the alternative 2994194206Ssimon implementation in BN_mod_exp_mont_consttime().) The old name 2995194206Ssimon remains as a deprecated alias. 2996194206Ssimon 2997194206Ssimon Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general 2998194206Ssimon RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses 2999194206Ssimon constant-time implementations for more than just exponentiation. 3000194206Ssimon Here too the old name is kept as a deprecated alias. 3001194206Ssimon 3002194206Ssimon BN_BLINDING_new() will now use BN_dup() for the modulus so that 3003194206Ssimon the BN_BLINDING structure gets an independent copy of the 3004194206Ssimon modulus. This means that the previous "BIGNUM *m" argument to 3005194206Ssimon BN_BLINDING_new() and to BN_BLINDING_create_param() now 3006194206Ssimon essentially becomes "const BIGNUM *m", although we can't actually 3007194206Ssimon change this in the header file before 0.9.9. It allows 3008194206Ssimon RSA_setup_blinding() to use BN_with_flags() on the modulus to 3009194206Ssimon enable BN_FLG_CONSTTIME. 3010194206Ssimon 3011194206Ssimon [Matthew D Wood (Intel Corp)] 3012194206Ssimon 3013194206Ssimon *) In the SSL/TLS server implementation, be strict about session ID 3014194206Ssimon context matching (which matters if an application uses a single 3015194206Ssimon external cache for different purposes). Previously, 3016194206Ssimon out-of-context reuse was forbidden only if SSL_VERIFY_PEER was 3017194206Ssimon set. This did ensure strict client verification, but meant that, 3018194206Ssimon with applications using a single external cache for quite 3019194206Ssimon different requirements, clients could circumvent ciphersuite 3020194206Ssimon restrictions for a given session ID context by starting a session 3021194206Ssimon in a different context. 3022194206Ssimon [Bodo Moeller] 3023194206Ssimon 3024167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 3025167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 3026167612Ssimon authentication-only ciphersuites. 3027167612Ssimon [Bodo Moeller] 3028167612Ssimon 3029194206Ssimon *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was 3030194206Ssimon not complete and could lead to a possible single byte overflow 3031194206Ssimon (CVE-2007-5135) [Ben Laurie] 3032194206Ssimon 3033194206Ssimon Changes between 0.9.8d and 0.9.8e [23 Feb 2007] 3034194206Ssimon 3035167612Ssimon *) Since AES128 and AES256 (and similarly Camellia128 and 3036167612Ssimon Camellia256) share a single mask bit in the logic of 3037167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 3038167612Ssimon kludge to work properly if AES128 is available and AES256 isn't 3039167612Ssimon (or if Camellia128 is available and Camellia256 isn't). 3040167612Ssimon [Victor Duchovni] 3041167612Ssimon 3042167612Ssimon *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c 3043167612Ssimon (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): 3044167612Ssimon When a point or a seed is encoded in a BIT STRING, we need to 3045167612Ssimon prevent the removal of trailing zero bits to get the proper DER 3046167612Ssimon encoding. (By default, crypto/asn1/a_bitstr.c assumes the case 3047167612Ssimon of a NamedBitList, for which trailing 0 bits need to be removed.) 3048167612Ssimon [Bodo Moeller] 3049167612Ssimon 3050167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 3051167612Ssimon protocol version while receiving ClientHello even if the 3052167612Ssimon ClientHello is fragmented. (The server can't insist on the 3053167612Ssimon particular protocol version it has chosen before the ServerHello 3054167612Ssimon message has informed the client about his choice.) 3055167612Ssimon [Bodo Moeller] 3056167612Ssimon 3057167612Ssimon *) Add RFC 3779 support. 3058167612Ssimon [Rob Austein for ARIN, Ben Laurie] 3059167612Ssimon 3060167612Ssimon *) Load error codes if they are not already present instead of using a 3061167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 3062167612Ssimon Improve header file function name parsing. 3063167612Ssimon [Steve Henson] 3064167612Ssimon 3065167612Ssimon *) extend SMTP and IMAP protocol emulation in s_client to use EHLO 3066167612Ssimon or CAPABILITY handshake as required by RFCs. 3067167612Ssimon [Goetz Babin-Ebell] 3068167612Ssimon 3069162911Ssimon Changes between 0.9.8c and 0.9.8d [28 Sep 2006] 3070162911Ssimon 3071162911Ssimon *) Introduce limits to prevent malicious keys being able to 3072162911Ssimon cause a denial of service. (CVE-2006-2940) 3073162911Ssimon [Steve Henson, Bodo Moeller] 3074162911Ssimon 3075162911Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 3076162911Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 3077162911Ssimon 3078162911Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 3079162911Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 3080162911Ssimon 3081162911Ssimon *) Fix SSL client code which could crash if connecting to a 3082162911Ssimon malicious SSLv2 server. (CVE-2006-4343) 3083162911Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 3084162911Ssimon 3085162911Ssimon *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites 3086162911Ssimon match only those. Before that, "AES256-SHA" would be interpreted 3087162911Ssimon as a pattern and match "AES128-SHA" too (since AES128-SHA got 3088162911Ssimon the same strength classification in 0.9.7h) as we currently only 3089162911Ssimon have a single AES bit in the ciphersuite description bitmap. 3090162911Ssimon That change, however, also applied to ciphersuite strings such as 3091162911Ssimon "RC4-MD5" that intentionally matched multiple ciphersuites -- 3092162911Ssimon namely, SSL 2.0 ciphersuites in addition to the more common ones 3093162911Ssimon from SSL 3.0/TLS 1.0. 3094162911Ssimon 3095162911Ssimon So we change the selection algorithm again: Naming an explicit 3096162911Ssimon ciphersuite selects this one ciphersuite, and any other similar 3097162911Ssimon ciphersuite (same bitmap) from *other* protocol versions. 3098162911Ssimon Thus, "RC4-MD5" again will properly select both the SSL 2.0 3099162911Ssimon ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. 3100162911Ssimon 3101162911Ssimon Since SSL 2.0 does not have any ciphersuites for which the 3102162911Ssimon 128/256 bit distinction would be relevant, this works for now. 3103162911Ssimon The proper fix will be to use different bits for AES128 and 3104162911Ssimon AES256, which would have avoided the problems from the beginning; 3105162911Ssimon however, bits are scarce, so we can only do this in a new release 3106162911Ssimon (not just a patchlevel) when we can change the SSL_CIPHER 3107162911Ssimon definition to split the single 'unsigned long mask' bitmap into 3108162911Ssimon multiple values to extend the available space. 3109162911Ssimon 3110162911Ssimon [Bodo Moeller] 3111162911Ssimon 3112162911Ssimon Changes between 0.9.8b and 0.9.8c [05 Sep 2006] 3113162911Ssimon 3114162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 3115162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 3116162911Ssimon 3117162911Ssimon *) Add AES IGE and biIGE modes. 3118162911Ssimon [Ben Laurie] 3119162911Ssimon 3120162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 3121162911Ssimon possible instead of select(), since the latter has some 3122162911Ssimon undesirable limitations. 3123162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 3124162911Ssimon 3125162911Ssimon *) Disable "ECCdraft" ciphersuites more thoroughly. Now special 3126162911Ssimon treatment in ssl/ssl_ciph.s makes sure that these ciphersuites 3127162911Ssimon cannot be implicitly activated as part of, e.g., the "AES" alias. 3128162911Ssimon However, please upgrade to OpenSSL 0.9.9[-dev] for 3129162911Ssimon non-experimental use of the ECC ciphersuites to get TLS extension 3130162911Ssimon support, which is required for curve and point format negotiation 3131162911Ssimon to avoid potential handshake problems. 3132162911Ssimon [Bodo Moeller] 3133162911Ssimon 3134162911Ssimon *) Disable rogue ciphersuites: 3135162911Ssimon 3136162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 3137162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 3138162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 3139162911Ssimon 3140162911Ssimon The latter two were purportedly from 3141162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 3142162911Ssimon appear there. 3143162911Ssimon 3144167612Ssimon Also deactivate the remaining ciphersuites from 3145162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 3146162911Ssimon unofficial, and the ID has long expired. 3147162911Ssimon [Bodo Moeller] 3148162911Ssimon 3149162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 3150162911Ssimon dual-core machines) and other potential thread-safety issues. 3151162911Ssimon [Bodo Moeller] 3152162911Ssimon 3153162911Ssimon *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key 3154162911Ssimon versions), which is now available for royalty-free use 3155162911Ssimon (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). 3156162911Ssimon Also, add Camellia TLS ciphersuites from RFC 4132. 3157162911Ssimon 3158162911Ssimon To minimize changes between patchlevels in the OpenSSL 0.9.8 3159162911Ssimon series, Camellia remains excluded from compilation unless OpenSSL 3160162911Ssimon is configured with 'enable-camellia'. 3161162911Ssimon [NTT] 3162162911Ssimon 3163162911Ssimon *) Disable the padding bug check when compression is in use. The padding 3164162911Ssimon bug check assumes the first packet is of even length, this is not 3165162911Ssimon necessarily true if compresssion is enabled and can result in false 3166162911Ssimon positives causing handshake failure. The actual bug test is ancient 3167162911Ssimon code so it is hoped that implementations will either have fixed it by 3168162911Ssimon now or any which still have the bug do not support compression. 3169162911Ssimon [Steve Henson] 3170162911Ssimon 3171160814Ssimon Changes between 0.9.8a and 0.9.8b [04 May 2006] 3172160814Ssimon 3173160814Ssimon *) When applying a cipher rule check to see if string match is an explicit 3174160814Ssimon cipher suite and only match that one cipher suite if it is. 3175160814Ssimon [Steve Henson] 3176160814Ssimon 3177160814Ssimon *) Link in manifests for VC++ if needed. 3178160814Ssimon [Austin Ziegler <halostatue@gmail.com>] 3179160814Ssimon 3180160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 3181160814Ssimon draft-ietf-tls-ecc-12.txt with proposed changes (but without 3182160814Ssimon TLS extensions, which are supported starting with the 0.9.9 3183160814Ssimon branch, not in the OpenSSL 0.9.8 branch). 3184160814Ssimon [Douglas Stebila] 3185160814Ssimon 3186160814Ssimon *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support 3187160814Ssimon opaque EVP_CIPHER_CTX handling. 3188160814Ssimon [Steve Henson] 3189160814Ssimon 3190160814Ssimon *) Fixes and enhancements to zlib compression code. We now only use 3191160814Ssimon "zlib1.dll" and use the default __cdecl calling convention on Win32 3192160814Ssimon to conform with the standards mentioned here: 3193160814Ssimon http://www.zlib.net/DLL_FAQ.txt 3194160814Ssimon Static zlib linking now works on Windows and the new --with-zlib-include 3195160814Ssimon --with-zlib-lib options to Configure can be used to supply the location 3196160814Ssimon of the headers and library. Gracefully handle case where zlib library 3197160814Ssimon can't be loaded. 3198160814Ssimon [Steve Henson] 3199160814Ssimon 3200160814Ssimon *) Several fixes and enhancements to the OID generation code. The old code 3201160814Ssimon sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't 3202160814Ssimon handle numbers larger than ULONG_MAX, truncated printing and had a 3203160814Ssimon non standard OBJ_obj2txt() behaviour. 3204160814Ssimon [Steve Henson] 3205160814Ssimon 3206160814Ssimon *) Add support for building of engines under engine/ as shared libraries 3207160814Ssimon under VC++ build system. 3208160814Ssimon [Steve Henson] 3209160814Ssimon 3210160814Ssimon *) Corrected the numerous bugs in the Win32 path splitter in DSO. 3211160814Ssimon Hopefully, we will not see any false combination of paths any more. 3212160814Ssimon [Richard Levitte] 3213160814Ssimon 3214160814Ssimon Changes between 0.9.8 and 0.9.8a [11 Oct 2005] 3215160814Ssimon 3216160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 3217160814Ssimon (part of SSL_OP_ALL). This option used to disable the 3218160814Ssimon countermeasure against man-in-the-middle protocol-version 3219160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 3220160814Ssimon idea. (CVE-2005-2969) 3221160814Ssimon 3222160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 3223160814Ssimon for Information Security, National Institute of Advanced Industrial 3224160814Ssimon Science and Technology [AIST], Japan)] 3225160814Ssimon 3226160814Ssimon *) Add two function to clear and return the verify parameter flags. 3227160814Ssimon [Steve Henson] 3228160814Ssimon 3229160814Ssimon *) Keep cipherlists sorted in the source instead of sorting them at 3230160814Ssimon runtime, thus removing the need for a lock. 3231160814Ssimon [Nils Larsch] 3232160814Ssimon 3233160814Ssimon *) Avoid some small subgroup attacks in Diffie-Hellman. 3234160814Ssimon [Nick Mathewson and Ben Laurie] 3235160814Ssimon 3236160814Ssimon *) Add functions for well-known primes. 3237160814Ssimon [Nick Mathewson] 3238160814Ssimon 3239160814Ssimon *) Extended Windows CE support. 3240160814Ssimon [Satoshi Nakamura and Andy Polyakov] 3241160814Ssimon 3242160814Ssimon *) Initialize SSL_METHOD structures at compile time instead of during 3243160814Ssimon runtime, thus removing the need for a lock. 3244160814Ssimon [Steve Henson] 3245160814Ssimon 3246160814Ssimon *) Make PKCS7_decrypt() work even if no certificate is supplied by 3247160814Ssimon attempting to decrypt each encrypted key in turn. Add support to 3248160814Ssimon smime utility. 3249160814Ssimon [Steve Henson] 3250160814Ssimon 3251160814Ssimon Changes between 0.9.7h and 0.9.8 [05 Jul 2005] 3252160814Ssimon 3253162911Ssimon [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after 3254162911Ssimon OpenSSL 0.9.8.] 3255162911Ssimon 3256160814Ssimon *) Add libcrypto.pc and libssl.pc for those who feel they need them. 3257160814Ssimon [Richard Levitte] 3258160814Ssimon 3259160814Ssimon *) Change CA.sh and CA.pl so they don't bundle the CSR and the private 3260160814Ssimon key into the same file any more. 3261160814Ssimon [Richard Levitte] 3262160814Ssimon 3263160814Ssimon *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. 3264160814Ssimon [Andy Polyakov] 3265160814Ssimon 3266160814Ssimon *) Add -utf8 command line and config file option to 'ca'. 3267160814Ssimon [Stefan <stf@udoma.org] 3268160814Ssimon 3269160814Ssimon *) Removed the macro des_crypt(), as it seems to conflict with some 3270160814Ssimon libraries. Use DES_crypt(). 3271160814Ssimon [Richard Levitte] 3272160814Ssimon 3273160814Ssimon *) Correct naming of the 'chil' and '4758cca' ENGINEs. This 3274160814Ssimon involves renaming the source and generated shared-libs for 3275160814Ssimon both. The engines will accept the corrected or legacy ids 3276160814Ssimon ('ncipher' and '4758_cca' respectively) when binding. NB, 3277160814Ssimon this only applies when building 'shared'. 3278160814Ssimon [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] 3279160814Ssimon 3280160814Ssimon *) Add attribute functions to EVP_PKEY structure. Modify 3281160814Ssimon PKCS12_create() to recognize a CSP name attribute and 3282160814Ssimon use it. Make -CSP option work again in pkcs12 utility. 3283160814Ssimon [Steve Henson] 3284160814Ssimon 3285160814Ssimon *) Add new functionality to the bn blinding code: 3286160814Ssimon - automatic re-creation of the BN_BLINDING parameters after 3287160814Ssimon a fixed number of uses (currently 32) 3288160814Ssimon - add new function for parameter creation 3289160814Ssimon - introduce flags to control the update behaviour of the 3290160814Ssimon BN_BLINDING parameters 3291160814Ssimon - hide BN_BLINDING structure 3292160814Ssimon Add a second BN_BLINDING slot to the RSA structure to improve 3293160814Ssimon performance when a single RSA object is shared among several 3294160814Ssimon threads. 3295160814Ssimon [Nils Larsch] 3296160814Ssimon 3297160814Ssimon *) Add support for DTLS. 3298160814Ssimon [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] 3299160814Ssimon 3300160814Ssimon *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) 3301160814Ssimon to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() 3302160814Ssimon [Walter Goulet] 3303160814Ssimon 3304160814Ssimon *) Remove buggy and incompletet DH cert support from 3305160814Ssimon ssl/ssl_rsa.c and ssl/s3_both.c 3306160814Ssimon [Nils Larsch] 3307160814Ssimon 3308160814Ssimon *) Use SHA-1 instead of MD5 as the default digest algorithm for 3309160814Ssimon the apps/openssl applications. 3310160814Ssimon [Nils Larsch] 3311160814Ssimon 3312160814Ssimon *) Compile clean with "-Wall -Wmissing-prototypes 3313160814Ssimon -Wstrict-prototypes -Wmissing-declarations -Werror". Currently 3314160814Ssimon DEBUG_SAFESTACK must also be set. 3315160814Ssimon [Ben Laurie] 3316160814Ssimon 3317160814Ssimon *) Change ./Configure so that certain algorithms can be disabled by default. 3318160814Ssimon The new counterpiece to "no-xxx" is "enable-xxx". 3319160814Ssimon 3320160814Ssimon The patented RC5 and MDC2 algorithms will now be disabled unless 3321160814Ssimon "enable-rc5" and "enable-mdc2", respectively, are specified. 3322160814Ssimon 3323160814Ssimon (IDEA remains enabled despite being patented. This is because IDEA 3324160814Ssimon is frequently required for interoperability, and there is no license 3325160814Ssimon fee for non-commercial use. As before, "no-idea" can be used to 3326160814Ssimon avoid this algorithm.) 3327160814Ssimon 3328160814Ssimon [Bodo Moeller] 3329160814Ssimon 3330160814Ssimon *) Add processing of proxy certificates (see RFC 3820). This work was 3331160814Ssimon sponsored by KTH (The Royal Institute of Technology in Stockholm) and 3332160814Ssimon EGEE (Enabling Grids for E-science in Europe). 3333160814Ssimon [Richard Levitte] 3334160814Ssimon 3335160814Ssimon *) RC4 performance overhaul on modern architectures/implementations, such 3336160814Ssimon as Intel P4, IA-64 and AMD64. 3337160814Ssimon [Andy Polyakov] 3338160814Ssimon 3339160814Ssimon *) New utility extract-section.pl. This can be used specify an alternative 3340160814Ssimon section number in a pod file instead of having to treat each file as 3341160814Ssimon a separate case in Makefile. This can be done by adding two lines to the 3342160814Ssimon pod file: 3343160814Ssimon 3344160814Ssimon =for comment openssl_section:XXX 3345160814Ssimon 3346160814Ssimon The blank line is mandatory. 3347160814Ssimon 3348160814Ssimon [Steve Henson] 3349160814Ssimon 3350160814Ssimon *) New arguments -certform, -keyform and -pass for s_client and s_server 3351160814Ssimon to allow alternative format key and certificate files and passphrase 3352160814Ssimon sources. 3353160814Ssimon [Steve Henson] 3354160814Ssimon 3355160814Ssimon *) New structure X509_VERIFY_PARAM which combines current verify parameters, 3356160814Ssimon update associated structures and add various utility functions. 3357160814Ssimon 3358160814Ssimon Add new policy related verify parameters, include policy checking in 3359160814Ssimon standard verify code. Enhance 'smime' application with extra parameters 3360160814Ssimon to support policy checking and print out. 3361160814Ssimon [Steve Henson] 3362160814Ssimon 3363160814Ssimon *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 3364160814Ssimon Nehemiah processors. These extensions support AES encryption in hardware 3365160814Ssimon as well as RNG (though RNG support is currently disabled). 3366160814Ssimon [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] 3367160814Ssimon 3368160814Ssimon *) Deprecate BN_[get|set]_params() functions (they were ignored internally). 3369160814Ssimon [Geoff Thorpe] 3370160814Ssimon 3371160814Ssimon *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. 3372160814Ssimon [Andy Polyakov and a number of other people] 3373160814Ssimon 3374160814Ssimon *) Improved PowerPC platform support. Most notably BIGNUM assembler 3375160814Ssimon implementation contributed by IBM. 3376160814Ssimon [Suresh Chari, Peter Waltenberg, Andy Polyakov] 3377160814Ssimon 3378160814Ssimon *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public 3379160814Ssimon exponent rather than 'unsigned long'. There is a corresponding change to 3380160814Ssimon the new 'rsa_keygen' element of the RSA_METHOD structure. 3381160814Ssimon [Jelte Jansen, Geoff Thorpe] 3382160814Ssimon 3383160814Ssimon *) Functionality for creating the initial serial number file is now 3384160814Ssimon moved from CA.pl to the 'ca' utility with a new option -create_serial. 3385160814Ssimon 3386160814Ssimon (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial 3387160814Ssimon number file to 1, which is bound to cause problems. To avoid 3388160814Ssimon the problems while respecting compatibility between different 0.9.7 3389160814Ssimon patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in 3390160814Ssimon CA.pl for serial number initialization. With the new release 0.9.8, 3391160814Ssimon we can fix the problem directly in the 'ca' utility.) 3392160814Ssimon [Steve Henson] 3393160814Ssimon 3394160814Ssimon *) Reduced header interdepencies by declaring more opaque objects in 3395160814Ssimon ossl_typ.h. As a consequence, including some headers (eg. engine.h) will 3396160814Ssimon give fewer recursive includes, which could break lazy source code - so 3397160814Ssimon this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, 3398160814Ssimon developers should define this symbol when building and using openssl to 3399160814Ssimon ensure they track the recommended behaviour, interfaces, [etc], but 3400160814Ssimon backwards-compatible behaviour prevails when this isn't defined. 3401160814Ssimon [Geoff Thorpe] 3402160814Ssimon 3403160814Ssimon *) New function X509_POLICY_NODE_print() which prints out policy nodes. 3404160814Ssimon [Steve Henson] 3405160814Ssimon 3406160814Ssimon *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. 3407160814Ssimon This will generate a random key of the appropriate length based on the 3408160814Ssimon cipher context. The EVP_CIPHER can provide its own random key generation 3409160814Ssimon routine to support keys of a specific form. This is used in the des and 3410160814Ssimon 3des routines to generate a key of the correct parity. Update S/MIME 3411160814Ssimon code to use new functions and hence generate correct parity DES keys. 3412160814Ssimon Add EVP_CHECK_DES_KEY #define to return an error if the key is not 3413160814Ssimon valid (weak or incorrect parity). 3414160814Ssimon [Steve Henson] 3415160814Ssimon 3416160814Ssimon *) Add a local set of CRLs that can be used by X509_verify_cert() as well 3417160814Ssimon as looking them up. This is useful when the verified structure may contain 3418160814Ssimon CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs 3419160814Ssimon present unless the new PKCS7_NO_CRL flag is asserted. 3420160814Ssimon [Steve Henson] 3421160814Ssimon 3422160814Ssimon *) Extend ASN1 oid configuration module. It now additionally accepts the 3423160814Ssimon syntax: 3424160814Ssimon 3425160814Ssimon shortName = some long name, 1.2.3.4 3426160814Ssimon [Steve Henson] 3427160814Ssimon 3428160814Ssimon *) Reimplemented the BN_CTX implementation. There is now no more static 3429160814Ssimon limitation on the number of variables it can handle nor the depth of the 3430160814Ssimon "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack 3431160814Ssimon information can now expand as required, and rather than having a single 3432160814Ssimon static array of bignums, BN_CTX now uses a linked-list of such arrays 3433160814Ssimon allowing it to expand on demand whilst maintaining the usefulness of 3434160814Ssimon BN_CTX's "bundling". 3435160814Ssimon [Geoff Thorpe] 3436160814Ssimon 3437160814Ssimon *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD 3438160814Ssimon to allow all RSA operations to function using a single BN_CTX. 3439160814Ssimon [Geoff Thorpe] 3440160814Ssimon 3441160814Ssimon *) Preliminary support for certificate policy evaluation and checking. This 3442160814Ssimon is initially intended to pass the tests outlined in "Conformance Testing 3443160814Ssimon of Relying Party Client Certificate Path Processing Logic" v1.07. 3444160814Ssimon [Steve Henson] 3445160814Ssimon 3446160814Ssimon *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and 3447160814Ssimon remained unused and not that useful. A variety of other little bignum 3448160814Ssimon tweaks and fixes have also been made continuing on from the audit (see 3449160814Ssimon below). 3450160814Ssimon [Geoff Thorpe] 3451160814Ssimon 3452160814Ssimon *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with 3453160814Ssimon associated ASN1, EVP and SSL functions and old ASN1 macros. 3454160814Ssimon [Richard Levitte] 3455160814Ssimon 3456160814Ssimon *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, 3457160814Ssimon and this should never fail. So the return value from the use of 3458160814Ssimon BN_set_word() (which can fail due to needless expansion) is now deprecated; 3459160814Ssimon if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. 3460160814Ssimon [Geoff Thorpe] 3461160814Ssimon 3462160814Ssimon *) BN_CTX_get() should return zero-valued bignums, providing the same 3463160814Ssimon initialised value as BN_new(). 3464291721Sjkim [Geoff Thorpe, suggested by Ulf M��ller] 3465160814Ssimon 3466160814Ssimon *) Support for inhibitAnyPolicy certificate extension. 3467160814Ssimon [Steve Henson] 3468160814Ssimon 3469160814Ssimon *) An audit of the BIGNUM code is underway, for which debugging code is 3470160814Ssimon enabled when BN_DEBUG is defined. This makes stricter enforcements on what 3471160814Ssimon is considered valid when processing BIGNUMs, and causes execution to 3472160814Ssimon assert() when a problem is discovered. If BN_DEBUG_RAND is defined, 3473160814Ssimon further steps are taken to deliberately pollute unused data in BIGNUM 3474160814Ssimon structures to try and expose faulty code further on. For now, openssl will 3475160814Ssimon (in its default mode of operation) continue to tolerate the inconsistent 3476160814Ssimon forms that it has tolerated in the past, but authors and packagers should 3477160814Ssimon consider trying openssl and their own applications when compiled with 3478160814Ssimon these debugging symbols defined. It will help highlight potential bugs in 3479160814Ssimon their own code, and will improve the test coverage for OpenSSL itself. At 3480160814Ssimon some point, these tighter rules will become openssl's default to improve 3481160814Ssimon maintainability, though the assert()s and other overheads will remain only 3482160814Ssimon in debugging configurations. See bn.h for more details. 3483291721Sjkim [Geoff Thorpe, Nils Larsch, Ulf M��ller] 3484160814Ssimon 3485160814Ssimon *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure 3486160814Ssimon that can only be obtained through BN_CTX_new() (which implicitly 3487160814Ssimon initialises it). The presence of this function only made it possible 3488160814Ssimon to overwrite an existing structure (and cause memory leaks). 3489160814Ssimon [Geoff Thorpe] 3490160814Ssimon 3491160814Ssimon *) Because of the callback-based approach for implementing LHASH as a 3492160814Ssimon template type, lh_insert() adds opaque objects to hash-tables and 3493160814Ssimon lh_doall() or lh_doall_arg() are typically used with a destructor callback 3494160814Ssimon to clean up those corresponding objects before destroying the hash table 3495160814Ssimon (and losing the object pointers). So some over-zealous constifications in 3496160814Ssimon LHASH have been relaxed so that lh_insert() does not take (nor store) the 3497160814Ssimon objects as "const" and the lh_doall[_arg] callback wrappers are not 3498160814Ssimon prototyped to have "const" restrictions on the object pointers they are 3499160814Ssimon given (and so aren't required to cast them away any more). 3500160814Ssimon [Geoff Thorpe] 3501160814Ssimon 3502160814Ssimon *) The tmdiff.h API was so ugly and minimal that our own timing utility 3503160814Ssimon (speed) prefers to use its own implementation. The two implementations 3504160814Ssimon haven't been consolidated as yet (volunteers?) but the tmdiff API has had 3505160814Ssimon its object type properly exposed (MS_TM) instead of casting to/from "char 3506160814Ssimon *". This may still change yet if someone realises MS_TM and "ms_time_***" 3507160814Ssimon aren't necessarily the greatest nomenclatures - but this is what was used 3508160814Ssimon internally to the implementation so I've used that for now. 3509160814Ssimon [Geoff Thorpe] 3510160814Ssimon 3511160814Ssimon *) Ensure that deprecated functions do not get compiled when 3512160814Ssimon OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of 3513160814Ssimon the self-tests were still using deprecated key-generation functions so 3514160814Ssimon these have been updated also. 3515160814Ssimon [Geoff Thorpe] 3516160814Ssimon 3517160814Ssimon *) Reorganise PKCS#7 code to separate the digest location functionality 3518160814Ssimon into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). 3519160814Ssimon New function PKCS7_set_digest() to set the digest type for PKCS#7 3520160814Ssimon digestedData type. Add additional code to correctly generate the 3521160814Ssimon digestedData type and add support for this type in PKCS7 initialization 3522160814Ssimon functions. 3523160814Ssimon [Steve Henson] 3524160814Ssimon 3525160814Ssimon *) New function PKCS7_set0_type_other() this initializes a PKCS7 3526160814Ssimon structure of type "other". 3527160814Ssimon [Steve Henson] 3528160814Ssimon 3529160814Ssimon *) Fix prime generation loop in crypto/bn/bn_prime.pl by making 3530160814Ssimon sure the loop does correctly stop and breaking ("division by zero") 3531160814Ssimon modulus operations are not performed. The (pre-generated) prime 3532160814Ssimon table crypto/bn/bn_prime.h was already correct, but it could not be 3533160814Ssimon re-generated on some platforms because of the "division by zero" 3534160814Ssimon situation in the script. 3535160814Ssimon [Ralf S. Engelschall] 3536160814Ssimon 3537160814Ssimon *) Update support for ECC-based TLS ciphersuites according to 3538160814Ssimon draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with 3539160814Ssimon SHA-1 now is only used for "small" curves (where the 3540160814Ssimon representation of a field element takes up to 24 bytes); for 3541160814Ssimon larger curves, the field element resulting from ECDH is directly 3542160814Ssimon used as premaster secret. 3543160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3544160814Ssimon 3545160814Ssimon *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 3546160814Ssimon curve secp160r1 to the tests. 3547160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3548160814Ssimon 3549160814Ssimon *) Add the possibility to load symbols globally with DSO. 3550291721Sjkim [G��tz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] 3551160814Ssimon 3552160814Ssimon *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better 3553160814Ssimon control of the error stack. 3554160814Ssimon [Richard Levitte] 3555160814Ssimon 3556160814Ssimon *) Add support for STORE in ENGINE. 3557160814Ssimon [Richard Levitte] 3558160814Ssimon 3559160814Ssimon *) Add the STORE type. The intention is to provide a common interface 3560160814Ssimon to certificate and key stores, be they simple file-based stores, or 3561160814Ssimon HSM-type store, or LDAP stores, or... 3562160814Ssimon NOTE: The code is currently UNTESTED and isn't really used anywhere. 3563160814Ssimon [Richard Levitte] 3564160814Ssimon 3565160814Ssimon *) Add a generic structure called OPENSSL_ITEM. This can be used to 3566160814Ssimon pass a list of arguments to any function as well as provide a way 3567160814Ssimon for a function to pass data back to the caller. 3568160814Ssimon [Richard Levitte] 3569160814Ssimon 3570160814Ssimon *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() 3571160814Ssimon works like BUF_strdup() but can be used to duplicate a portion of 3572160814Ssimon a string. The copy gets NUL-terminated. BUF_memdup() duplicates 3573160814Ssimon a memory area. 3574160814Ssimon [Richard Levitte] 3575160814Ssimon 3576160814Ssimon *) Add the function sk_find_ex() which works like sk_find(), but will 3577160814Ssimon return an index to an element even if an exact match couldn't be 3578160814Ssimon found. The index is guaranteed to point at the element where the 3579160814Ssimon searched-for key would be inserted to preserve sorting order. 3580160814Ssimon [Richard Levitte] 3581160814Ssimon 3582160814Ssimon *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but 3583160814Ssimon takes an extra flags argument for optional functionality. Currently, 3584160814Ssimon the following flags are defined: 3585160814Ssimon 3586160814Ssimon OBJ_BSEARCH_VALUE_ON_NOMATCH 3587160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3588160814Ssimon element where the comparing function returns a negative or zero 3589160814Ssimon number. 3590160814Ssimon 3591160814Ssimon OBJ_BSEARCH_FIRST_VALUE_ON_MATCH 3592160814Ssimon This one gets OBJ_bsearch_ex() to return a pointer to the first 3593160814Ssimon element where the comparing function returns zero. This is useful 3594160814Ssimon if there are more than one element where the comparing function 3595160814Ssimon returns zero. 3596160814Ssimon [Richard Levitte] 3597160814Ssimon 3598160814Ssimon *) Make it possible to create self-signed certificates with 'openssl ca' 3599160814Ssimon in such a way that the self-signed certificate becomes part of the 3600160814Ssimon CA database and uses the same mechanisms for serial number generation 3601160814Ssimon as all other certificate signing. The new flag '-selfsign' enables 3602160814Ssimon this functionality. Adapt CA.sh and CA.pl.in. 3603160814Ssimon [Richard Levitte] 3604160814Ssimon 3605160814Ssimon *) Add functionality to check the public key of a certificate request 3606160814Ssimon against a given private. This is useful to check that a certificate 3607160814Ssimon request can be signed by that key (self-signing). 3608160814Ssimon [Richard Levitte] 3609160814Ssimon 3610160814Ssimon *) Make it possible to have multiple active certificates with the same 3611160814Ssimon subject in the CA index file. This is done only if the keyword 3612160814Ssimon 'unique_subject' is set to 'no' in the main CA section (default 3613160814Ssimon if 'CA_default') of the configuration file. The value is saved 3614160814Ssimon with the database itself in a separate index attribute file, 3615160814Ssimon named like the index file with '.attr' appended to the name. 3616160814Ssimon [Richard Levitte] 3617160814Ssimon 3618160814Ssimon *) Generate muti valued AVAs using '+' notation in config files for 3619160814Ssimon req and dirName. 3620160814Ssimon [Steve Henson] 3621160814Ssimon 3622160814Ssimon *) Support for nameConstraints certificate extension. 3623160814Ssimon [Steve Henson] 3624160814Ssimon 3625160814Ssimon *) Support for policyConstraints certificate extension. 3626160814Ssimon [Steve Henson] 3627160814Ssimon 3628160814Ssimon *) Support for policyMappings certificate extension. 3629160814Ssimon [Steve Henson] 3630160814Ssimon 3631160814Ssimon *) Make sure the default DSA_METHOD implementation only uses its 3632160814Ssimon dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, 3633160814Ssimon and change its own handlers to be NULL so as to remove unnecessary 3634160814Ssimon indirection. This lets alternative implementations fallback to the 3635160814Ssimon default implementation more easily. 3636160814Ssimon [Geoff Thorpe] 3637160814Ssimon 3638160814Ssimon *) Support for directoryName in GeneralName related extensions 3639160814Ssimon in config files. 3640160814Ssimon [Steve Henson] 3641160814Ssimon 3642160814Ssimon *) Make it possible to link applications using Makefile.shared. 3643160814Ssimon Make that possible even when linking against static libraries! 3644160814Ssimon [Richard Levitte] 3645160814Ssimon 3646160814Ssimon *) Support for single pass processing for S/MIME signing. This now 3647160814Ssimon means that S/MIME signing can be done from a pipe, in addition 3648160814Ssimon cleartext signing (multipart/signed type) is effectively streaming 3649160814Ssimon and the signed data does not need to be all held in memory. 3650160814Ssimon 3651160814Ssimon This is done with a new flag PKCS7_STREAM. When this flag is set 3652160814Ssimon PKCS7_sign() only initializes the PKCS7 structure and the actual signing 3653160814Ssimon is done after the data is output (and digests calculated) in 3654160814Ssimon SMIME_write_PKCS7(). 3655160814Ssimon [Steve Henson] 3656160814Ssimon 3657160814Ssimon *) Add full support for -rpath/-R, both in shared libraries and 3658160814Ssimon applications, at least on the platforms where it's known how 3659160814Ssimon to do it. 3660160814Ssimon [Richard Levitte] 3661160814Ssimon 3662160814Ssimon *) In crypto/ec/ec_mult.c, implement fast point multiplication with 3663160814Ssimon precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() 3664160814Ssimon will now compute a table of multiples of the generator that 3665160814Ssimon makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() 3666160814Ssimon faster (notably in the case of a single point multiplication, 3667160814Ssimon scalar * generator). 3668160814Ssimon [Nils Larsch, Bodo Moeller] 3669160814Ssimon 3670160814Ssimon *) IPv6 support for certificate extensions. The various extensions 3671160814Ssimon which use the IP:a.b.c.d can now take IPv6 addresses using the 3672160814Ssimon formats of RFC1884 2.2 . IPv6 addresses are now also displayed 3673160814Ssimon correctly. 3674160814Ssimon [Steve Henson] 3675160814Ssimon 3676160814Ssimon *) Added an ENGINE that implements RSA by performing private key 3677160814Ssimon exponentiations with the GMP library. The conversions to and from 3678160814Ssimon GMP's mpz_t format aren't optimised nor are any montgomery forms 3679160814Ssimon cached, and on x86 it appears OpenSSL's own performance has caught up. 3680160814Ssimon However there are likely to be other architectures where GMP could 3681160814Ssimon provide a boost. This ENGINE is not built in by default, but it can be 3682160814Ssimon specified at Configure time and should be accompanied by the necessary 3683160814Ssimon linker additions, eg; 3684160814Ssimon ./config -DOPENSSL_USE_GMP -lgmp 3685160814Ssimon [Geoff Thorpe] 3686160814Ssimon 3687160814Ssimon *) "openssl engine" will not display ENGINE/DSO load failure errors when 3688160814Ssimon testing availability of engines with "-t" - the old behaviour is 3689160814Ssimon produced by increasing the feature's verbosity with "-tt". 3690160814Ssimon [Geoff Thorpe] 3691160814Ssimon 3692160814Ssimon *) ECDSA routines: under certain error conditions uninitialized BN objects 3693160814Ssimon could be freed. Solution: make sure initialization is performed early 3694160814Ssimon enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> 3695160814Ssimon via PR#459) 3696160814Ssimon [Lutz Jaenicke] 3697160814Ssimon 3698160814Ssimon *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD 3699160814Ssimon and DH_METHOD (eg. by ENGINE implementations) to override the normal 3700160814Ssimon software implementations. For DSA and DH, parameter generation can 3701160814Ssimon also be overriden by providing the appropriate method callbacks. 3702160814Ssimon [Geoff Thorpe] 3703160814Ssimon 3704160814Ssimon *) Change the "progress" mechanism used in key-generation and 3705160814Ssimon primality testing to functions that take a new BN_GENCB pointer in 3706160814Ssimon place of callback/argument pairs. The new API functions have "_ex" 3707160814Ssimon postfixes and the older functions are reimplemented as wrappers for 3708160814Ssimon the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide 3709160814Ssimon declarations of the old functions to help (graceful) attempts to 3710160814Ssimon migrate to the new functions. Also, the new key-generation API 3711160814Ssimon functions operate on a caller-supplied key-structure and return 3712160814Ssimon success/failure rather than returning a key or NULL - this is to 3713160814Ssimon help make "keygen" another member function of RSA_METHOD etc. 3714160814Ssimon 3715160814Ssimon Example for using the new callback interface: 3716160814Ssimon 3717160814Ssimon int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; 3718160814Ssimon void *my_arg = ...; 3719160814Ssimon BN_GENCB my_cb; 3720160814Ssimon 3721160814Ssimon BN_GENCB_set(&my_cb, my_callback, my_arg); 3722160814Ssimon 3723160814Ssimon return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); 3724160814Ssimon /* For the meaning of a, b in calls to my_callback(), see the 3725160814Ssimon * documentation of the function that calls the callback. 3726160814Ssimon * cb will point to my_cb; my_arg can be retrieved as cb->arg. 3727160814Ssimon * my_callback should return 1 if it wants BN_is_prime_ex() 3728160814Ssimon * to continue, or 0 to stop. 3729160814Ssimon */ 3730160814Ssimon 3731160814Ssimon [Geoff Thorpe] 3732160814Ssimon 3733160814Ssimon *) Change the ZLIB compression method to be stateful, and make it 3734160814Ssimon available to TLS with the number defined in 3735160814Ssimon draft-ietf-tls-compression-04.txt. 3736160814Ssimon [Richard Levitte] 3737160814Ssimon 3738160814Ssimon *) Add the ASN.1 structures and functions for CertificatePair, which 3739160814Ssimon is defined as follows (according to X.509_4thEditionDraftV6.pdf): 3740160814Ssimon 3741160814Ssimon CertificatePair ::= SEQUENCE { 3742160814Ssimon forward [0] Certificate OPTIONAL, 3743160814Ssimon reverse [1] Certificate OPTIONAL, 3744160814Ssimon -- at least one of the pair shall be present -- } 3745160814Ssimon 3746160814Ssimon Also implement the PEM functions to read and write certificate 3747160814Ssimon pairs, and defined the PEM tag as "CERTIFICATE PAIR". 3748160814Ssimon 3749160814Ssimon This needed to be defined, mostly for the sake of the LDAP 3750160814Ssimon attribute crossCertificatePair, but may prove useful elsewhere as 3751160814Ssimon well. 3752160814Ssimon [Richard Levitte] 3753160814Ssimon 3754160814Ssimon *) Make it possible to inhibit symlinking of shared libraries in 3755160814Ssimon Makefile.shared, for Cygwin's sake. 3756160814Ssimon [Richard Levitte] 3757160814Ssimon 3758160814Ssimon *) Extend the BIGNUM API by creating a function 3759160814Ssimon void BN_set_negative(BIGNUM *a, int neg); 3760160814Ssimon and a macro that behave like 3761160814Ssimon int BN_is_negative(const BIGNUM *a); 3762160814Ssimon 3763160814Ssimon to avoid the need to access 'a->neg' directly in applications. 3764160814Ssimon [Nils Larsch] 3765160814Ssimon 3766160814Ssimon *) Implement fast modular reduction for pseudo-Mersenne primes 3767160814Ssimon used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). 3768160814Ssimon EC_GROUP_new_curve_GFp() will now automatically use this 3769160814Ssimon if applicable. 3770160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3771160814Ssimon 3772160814Ssimon *) Add new lock type (CRYPTO_LOCK_BN). 3773160814Ssimon [Bodo Moeller] 3774160814Ssimon 3775160814Ssimon *) Change the ENGINE framework to automatically load engines 3776160814Ssimon dynamically from specific directories unless they could be 3777160814Ssimon found to already be built in or loaded. Move all the 3778160814Ssimon current engines except for the cryptodev one to a new 3779160814Ssimon directory engines/. 3780160814Ssimon The engines in engines/ are built as shared libraries if 3781160814Ssimon the "shared" options was given to ./Configure or ./config. 3782160814Ssimon Otherwise, they are inserted in libcrypto.a. 3783160814Ssimon /usr/local/ssl/engines is the default directory for dynamic 3784160814Ssimon engines, but that can be overriden at configure time through 3785160814Ssimon the usual use of --prefix and/or --openssldir, and at run 3786160814Ssimon time with the environment variable OPENSSL_ENGINES. 3787160814Ssimon [Geoff Thorpe and Richard Levitte] 3788160814Ssimon 3789160814Ssimon *) Add Makefile.shared, a helper makefile to build shared 3790160814Ssimon libraries. Addapt Makefile.org. 3791160814Ssimon [Richard Levitte] 3792160814Ssimon 3793160814Ssimon *) Add version info to Win32 DLLs. 3794160814Ssimon [Peter 'Luna' Runestig" <peter@runestig.com>] 3795160814Ssimon 3796160814Ssimon *) Add new 'medium level' PKCS#12 API. Certificates and keys 3797160814Ssimon can be added using this API to created arbitrary PKCS#12 3798160814Ssimon files while avoiding the low level API. 3799160814Ssimon 3800160814Ssimon New options to PKCS12_create(), key or cert can be NULL and 3801160814Ssimon will then be omitted from the output file. The encryption 3802160814Ssimon algorithm NIDs can be set to -1 for no encryption, the mac 3803160814Ssimon iteration count can be set to 0 to omit the mac. 3804160814Ssimon 3805160814Ssimon Enhance pkcs12 utility by making the -nokeys and -nocerts 3806160814Ssimon options work when creating a PKCS#12 file. New option -nomac 3807160814Ssimon to omit the mac, NONE can be set for an encryption algorithm. 3808160814Ssimon New code is modified to use the enhanced PKCS12_create() 3809160814Ssimon instead of the low level API. 3810160814Ssimon [Steve Henson] 3811160814Ssimon 3812160814Ssimon *) Extend ASN1 encoder to support indefinite length constructed 3813160814Ssimon encoding. This can output sequences tags and octet strings in 3814160814Ssimon this form. Modify pk7_asn1.c to support indefinite length 3815160814Ssimon encoding. This is experimental and needs additional code to 3816160814Ssimon be useful, such as an ASN1 bio and some enhanced streaming 3817160814Ssimon PKCS#7 code. 3818160814Ssimon 3819160814Ssimon Extend template encode functionality so that tagging is passed 3820160814Ssimon down to the template encoder. 3821160814Ssimon [Steve Henson] 3822160814Ssimon 3823160814Ssimon *) Let 'openssl req' fail if an argument to '-newkey' is not 3824160814Ssimon recognized instead of using RSA as a default. 3825160814Ssimon [Bodo Moeller] 3826160814Ssimon 3827160814Ssimon *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. 3828160814Ssimon As these are not official, they are not included in "ALL"; 3829160814Ssimon the "ECCdraft" ciphersuite group alias can be used to select them. 3830160814Ssimon [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] 3831160814Ssimon 3832160814Ssimon *) Add ECDH engine support. 3833160814Ssimon [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] 3834160814Ssimon 3835160814Ssimon *) Add ECDH in new directory crypto/ecdh/. 3836160814Ssimon [Douglas Stebila (Sun Microsystems Laboratories)] 3837160814Ssimon 3838160814Ssimon *) Let BN_rand_range() abort with an error after 100 iterations 3839160814Ssimon without success (which indicates a broken PRNG). 3840160814Ssimon [Bodo Moeller] 3841160814Ssimon 3842160814Ssimon *) Change BN_mod_sqrt() so that it verifies that the input value 3843160814Ssimon is really the square of the return value. (Previously, 3844160814Ssimon BN_mod_sqrt would show GIGO behaviour.) 3845160814Ssimon [Bodo Moeller] 3846160814Ssimon 3847160814Ssimon *) Add named elliptic curves over binary fields from X9.62, SECG, 3848160814Ssimon and WAP/WTLS; add OIDs that were still missing. 3849160814Ssimon 3850160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3851160814Ssimon (Sun Microsystems Laboratories)] 3852160814Ssimon 3853160814Ssimon *) Extend the EC library for elliptic curves over binary fields 3854160814Ssimon (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). 3855160814Ssimon New EC_METHOD: 3856160814Ssimon 3857160814Ssimon EC_GF2m_simple_method 3858160814Ssimon 3859160814Ssimon New API functions: 3860160814Ssimon 3861160814Ssimon EC_GROUP_new_curve_GF2m 3862160814Ssimon EC_GROUP_set_curve_GF2m 3863160814Ssimon EC_GROUP_get_curve_GF2m 3864160814Ssimon EC_POINT_set_affine_coordinates_GF2m 3865160814Ssimon EC_POINT_get_affine_coordinates_GF2m 3866160814Ssimon EC_POINT_set_compressed_coordinates_GF2m 3867160814Ssimon 3868160814Ssimon Point compression for binary fields is disabled by default for 3869160814Ssimon patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to 3870160814Ssimon enable it). 3871160814Ssimon 3872160814Ssimon As binary polynomials are represented as BIGNUMs, various members 3873160814Ssimon of the EC_GROUP and EC_POINT data structures can be shared 3874160814Ssimon between the implementations for prime fields and binary fields; 3875160814Ssimon the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) 3876160814Ssimon are essentially identical to their ..._GFp counterparts. 3877160814Ssimon (For simplicity, the '..._GFp' prefix has been dropped from 3878160814Ssimon various internal method names.) 3879160814Ssimon 3880160814Ssimon An internal 'field_div' method (similar to 'field_mul' and 3881160814Ssimon 'field_sqr') has been added; this is used only for binary fields. 3882160814Ssimon 3883160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3884160814Ssimon (Sun Microsystems Laboratories)] 3885160814Ssimon 3886160814Ssimon *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() 3887160814Ssimon through methods ('mul', 'precompute_mult'). 3888160814Ssimon 3889160814Ssimon The generic implementations (now internally called 'ec_wNAF_mul' 3890160814Ssimon and 'ec_wNAF_precomputed_mult') remain the default if these 3891160814Ssimon methods are undefined. 3892160814Ssimon 3893160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3894160814Ssimon (Sun Microsystems Laboratories)] 3895160814Ssimon 3896160814Ssimon *) New function EC_GROUP_get_degree, which is defined through 3897160814Ssimon EC_METHOD. For curves over prime fields, this returns the bit 3898160814Ssimon length of the modulus. 3899160814Ssimon 3900160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3901160814Ssimon (Sun Microsystems Laboratories)] 3902160814Ssimon 3903160814Ssimon *) New functions EC_GROUP_dup, EC_POINT_dup. 3904160814Ssimon (These simply call ..._new and ..._copy). 3905160814Ssimon 3906160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3907160814Ssimon (Sun Microsystems Laboratories)] 3908160814Ssimon 3909160814Ssimon *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. 3910160814Ssimon Polynomials are represented as BIGNUMs (where the sign bit is not 3911160814Ssimon used) in the following functions [macros]: 3912160814Ssimon 3913160814Ssimon BN_GF2m_add 3914160814Ssimon BN_GF2m_sub [= BN_GF2m_add] 3915160814Ssimon BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] 3916160814Ssimon BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] 3917160814Ssimon BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] 3918160814Ssimon BN_GF2m_mod_inv 3919160814Ssimon BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] 3920160814Ssimon BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] 3921160814Ssimon BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] 3922160814Ssimon BN_GF2m_cmp [= BN_ucmp] 3923160814Ssimon 3924160814Ssimon (Note that only the 'mod' functions are actually for fields GF(2^m). 3925160814Ssimon BN_GF2m_add() is misnomer, but this is for the sake of consistency.) 3926160814Ssimon 3927160814Ssimon For some functions, an the irreducible polynomial defining a 3928160814Ssimon field can be given as an 'unsigned int[]' with strictly 3929160814Ssimon decreasing elements giving the indices of those bits that are set; 3930160814Ssimon i.e., p[] represents the polynomial 3931160814Ssimon f(t) = t^p[0] + t^p[1] + ... + t^p[k] 3932160814Ssimon where 3933160814Ssimon p[0] > p[1] > ... > p[k] = 0. 3934160814Ssimon This applies to the following functions: 3935160814Ssimon 3936160814Ssimon BN_GF2m_mod_arr 3937160814Ssimon BN_GF2m_mod_mul_arr 3938160814Ssimon BN_GF2m_mod_sqr_arr 3939160814Ssimon BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] 3940160814Ssimon BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] 3941160814Ssimon BN_GF2m_mod_exp_arr 3942160814Ssimon BN_GF2m_mod_sqrt_arr 3943160814Ssimon BN_GF2m_mod_solve_quad_arr 3944160814Ssimon BN_GF2m_poly2arr 3945160814Ssimon BN_GF2m_arr2poly 3946160814Ssimon 3947160814Ssimon Conversion can be performed by the following functions: 3948160814Ssimon 3949160814Ssimon BN_GF2m_poly2arr 3950160814Ssimon BN_GF2m_arr2poly 3951160814Ssimon 3952160814Ssimon bntest.c has additional tests for binary polynomial arithmetic. 3953160814Ssimon 3954160814Ssimon Two implementations for BN_GF2m_mod_div() are available. 3955160814Ssimon The default algorithm simply uses BN_GF2m_mod_inv() and 3956160814Ssimon BN_GF2m_mod_mul(). The alternative algorithm is compiled in only 3957160814Ssimon if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the 3958160814Ssimon copyright notice in crypto/bn/bn_gf2m.c before enabling it). 3959160814Ssimon 3960160814Ssimon [Sheueling Chang Shantz and Douglas Stebila 3961160814Ssimon (Sun Microsystems Laboratories)] 3962160814Ssimon 3963160814Ssimon *) Add new error code 'ERR_R_DISABLED' that can be used when some 3964160814Ssimon functionality is disabled at compile-time. 3965160814Ssimon [Douglas Stebila <douglas.stebila@sun.com>] 3966160814Ssimon 3967160814Ssimon *) Change default behaviour of 'openssl asn1parse' so that more 3968160814Ssimon information is visible when viewing, e.g., a certificate: 3969160814Ssimon 3970160814Ssimon Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' 3971160814Ssimon mode the content of non-printable OCTET STRINGs is output in a 3972160814Ssimon style similar to INTEGERs, but with '[HEX DUMP]' prepended to 3973160814Ssimon avoid the appearance of a printable string. 3974160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3975160814Ssimon 3976160814Ssimon *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access 3977160814Ssimon functions 3978160814Ssimon EC_GROUP_set_asn1_flag() 3979160814Ssimon EC_GROUP_get_asn1_flag() 3980160814Ssimon EC_GROUP_set_point_conversion_form() 3981160814Ssimon EC_GROUP_get_point_conversion_form() 3982160814Ssimon These control ASN1 encoding details: 3983160814Ssimon - Curves (i.e., groups) are encoded explicitly unless asn1_flag 3984160814Ssimon has been set to OPENSSL_EC_NAMED_CURVE. 3985160814Ssimon - Points are encoded in uncompressed form by default; options for 3986160814Ssimon asn1_for are as for point2oct, namely 3987160814Ssimon POINT_CONVERSION_COMPRESSED 3988160814Ssimon POINT_CONVERSION_UNCOMPRESSED 3989160814Ssimon POINT_CONVERSION_HYBRID 3990160814Ssimon 3991160814Ssimon Also add 'seed' and 'seed_len' members to EC_GROUP with access 3992160814Ssimon functions 3993160814Ssimon EC_GROUP_set_seed() 3994160814Ssimon EC_GROUP_get0_seed() 3995160814Ssimon EC_GROUP_get_seed_len() 3996160814Ssimon This is used only for ASN1 purposes (so far). 3997160814Ssimon [Nils Larsch <nla@trustcenter.de>] 3998160814Ssimon 3999160814Ssimon *) Add 'field_type' member to EC_METHOD, which holds the NID 4000160814Ssimon of the appropriate field type OID. The new function 4001160814Ssimon EC_METHOD_get_field_type() returns this value. 4002160814Ssimon [Nils Larsch <nla@trustcenter.de>] 4003160814Ssimon 4004160814Ssimon *) Add functions 4005160814Ssimon EC_POINT_point2bn() 4006160814Ssimon EC_POINT_bn2point() 4007160814Ssimon EC_POINT_point2hex() 4008160814Ssimon EC_POINT_hex2point() 4009160814Ssimon providing useful interfaces to EC_POINT_point2oct() and 4010160814Ssimon EC_POINT_oct2point(). 4011160814Ssimon [Nils Larsch <nla@trustcenter.de>] 4012160814Ssimon 4013160814Ssimon *) Change internals of the EC library so that the functions 4014160814Ssimon EC_GROUP_set_generator() 4015160814Ssimon EC_GROUP_get_generator() 4016160814Ssimon EC_GROUP_get_order() 4017160814Ssimon EC_GROUP_get_cofactor() 4018160814Ssimon are implemented directly in crypto/ec/ec_lib.c and not dispatched 4019160814Ssimon to methods, which would lead to unnecessary code duplication when 4020160814Ssimon adding different types of curves. 4021160814Ssimon [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] 4022160814Ssimon 4023160814Ssimon *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM 4024160814Ssimon arithmetic, and such that modified wNAFs are generated 4025160814Ssimon (which avoid length expansion in many cases). 4026160814Ssimon [Bodo Moeller] 4027160814Ssimon 4028160814Ssimon *) Add a function EC_GROUP_check_discriminant() (defined via 4029160814Ssimon EC_METHOD) that verifies that the curve discriminant is non-zero. 4030160814Ssimon 4031160814Ssimon Add a function EC_GROUP_check() that makes some sanity tests 4032160814Ssimon on a EC_GROUP, its generator and order. This includes 4033160814Ssimon EC_GROUP_check_discriminant(). 4034160814Ssimon [Nils Larsch <nla@trustcenter.de>] 4035160814Ssimon 4036160814Ssimon *) Add ECDSA in new directory crypto/ecdsa/. 4037160814Ssimon 4038160814Ssimon Add applications 'openssl ecparam' and 'openssl ecdsa' 4039160814Ssimon (these are based on 'openssl dsaparam' and 'openssl dsa'). 4040160814Ssimon 4041160814Ssimon ECDSA support is also included in various other files across the 4042160814Ssimon library. Most notably, 4043160814Ssimon - 'openssl req' now has a '-newkey ecdsa:file' option; 4044160814Ssimon - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; 4045160814Ssimon - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and 4046160814Ssimon d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make 4047160814Ssimon them suitable for ECDSA where domain parameters must be 4048160814Ssimon extracted before the specific public key; 4049160814Ssimon - ECDSA engine support has been added. 4050160814Ssimon [Nils Larsch <nla@trustcenter.de>] 4051160814Ssimon 4052160814Ssimon *) Include some named elliptic curves, and add OIDs from X9.62, 4053160814Ssimon SECG, and WAP/WTLS. Each curve can be obtained from the new 4054160814Ssimon function 4055160814Ssimon EC_GROUP_new_by_curve_name(), 4056160814Ssimon and the list of available named curves can be obtained with 4057160814Ssimon EC_get_builtin_curves(). 4058160814Ssimon Also add a 'curve_name' member to EC_GROUP objects, which can be 4059160814Ssimon accessed via 4060160814Ssimon EC_GROUP_set_curve_name() 4061160814Ssimon EC_GROUP_get_curve_name() 4062160814Ssimon [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] 4063160814Ssimon 4064160814Ssimon *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 4065160814Ssimon was actually never needed) and in BN_mul(). The removal in BN_mul() 4066160814Ssimon required a small change in bn_mul_part_recursive() and the addition 4067160814Ssimon of the functions bn_cmp_part_words(), bn_sub_part_words() and 4068160814Ssimon bn_add_part_words(), which do the same thing as bn_cmp_words(), 4069160814Ssimon bn_sub_words() and bn_add_words() except they take arrays with 4070160814Ssimon differing sizes. 4071160814Ssimon [Richard Levitte] 4072160814Ssimon 4073194206Ssimon Changes between 0.9.7l and 0.9.7m [23 Feb 2007] 4074194206Ssimon 4075167612Ssimon *) Cleanse PEM buffers before freeing them since they may contain 4076167612Ssimon sensitive data. 4077167612Ssimon [Benjamin Bennett <ben@psc.edu>] 4078167612Ssimon 4079167612Ssimon *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that 4080167612Ssimon a ciphersuite string such as "DEFAULT:RSA" cannot enable 4081167612Ssimon authentication-only ciphersuites. 4082167612Ssimon [Bodo Moeller] 4083167612Ssimon 4084167612Ssimon *) Since AES128 and AES256 share a single mask bit in the logic of 4085167612Ssimon ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a 4086167612Ssimon kludge to work properly if AES128 is available and AES256 isn't. 4087167612Ssimon [Victor Duchovni] 4088167612Ssimon 4089194206Ssimon *) Expand security boundary to match 1.1.1 module. 4090194206Ssimon [Steve Henson] 4091194206Ssimon 4092194206Ssimon *) Remove redundant features: hash file source, editing of test vectors 4093194206Ssimon modify fipsld to use external fips_premain.c signature. 4094194206Ssimon [Steve Henson] 4095194206Ssimon 4096194206Ssimon *) New perl script mkfipsscr.pl to create shell scripts or batch files to 4097194206Ssimon run algorithm test programs. 4098194206Ssimon [Steve Henson] 4099194206Ssimon 4100194206Ssimon *) Make algorithm test programs more tolerant of whitespace. 4101194206Ssimon [Steve Henson] 4102194206Ssimon 4103167612Ssimon *) Have SSL/TLS server implementation tolerate "mismatched" record 4104167612Ssimon protocol version while receiving ClientHello even if the 4105167612Ssimon ClientHello is fragmented. (The server can't insist on the 4106167612Ssimon particular protocol version it has chosen before the ServerHello 4107167612Ssimon message has informed the client about his choice.) 4108167612Ssimon [Bodo Moeller] 4109167612Ssimon 4110167612Ssimon *) Load error codes if they are not already present instead of using a 4111167612Ssimon static variable. This allows them to be cleanly unloaded and reloaded. 4112167612Ssimon [Steve Henson] 4113167612Ssimon 4114167612Ssimon Changes between 0.9.7k and 0.9.7l [28 Sep 2006] 4115167612Ssimon 4116167612Ssimon *) Introduce limits to prevent malicious keys being able to 4117167612Ssimon cause a denial of service. (CVE-2006-2940) 4118167612Ssimon [Steve Henson, Bodo Moeller] 4119167612Ssimon 4120167612Ssimon *) Fix ASN.1 parsing of certain invalid structures that can result 4121167612Ssimon in a denial of service. (CVE-2006-2937) [Steve Henson] 4122167612Ssimon 4123167612Ssimon *) Fix buffer overflow in SSL_get_shared_ciphers() function. 4124167612Ssimon (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] 4125167612Ssimon 4126167612Ssimon *) Fix SSL client code which could crash if connecting to a 4127167612Ssimon malicious SSLv2 server. (CVE-2006-4343) 4128167612Ssimon [Tavis Ormandy and Will Drewry, Google Security Team] 4129167612Ssimon 4130162911Ssimon *) Change ciphersuite string processing so that an explicit 4131162911Ssimon ciphersuite selects this one ciphersuite (so that "AES256-SHA" 4132162911Ssimon will no longer include "AES128-SHA"), and any other similar 4133162911Ssimon ciphersuite (same bitmap) from *other* protocol versions (so that 4134162911Ssimon "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the 4135162911Ssimon SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining 4136162911Ssimon changes from 0.9.8b and 0.9.8d. 4137162911Ssimon [Bodo Moeller] 4138162911Ssimon 4139162911Ssimon Changes between 0.9.7j and 0.9.7k [05 Sep 2006] 4140162911Ssimon 4141162911Ssimon *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher 4142162911Ssimon (CVE-2006-4339) [Ben Laurie and Google Security Team] 4143162911Ssimon 4144162911Ssimon *) Change the Unix randomness entropy gathering to use poll() when 4145162911Ssimon possible instead of select(), since the latter has some 4146162911Ssimon undesirable limitations. 4147162911Ssimon [Darryl Miles via Richard Levitte and Bodo Moeller] 4148162911Ssimon 4149162911Ssimon *) Disable rogue ciphersuites: 4150162911Ssimon 4151162911Ssimon - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") 4152162911Ssimon - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") 4153162911Ssimon - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") 4154162911Ssimon 4155162911Ssimon The latter two were purportedly from 4156162911Ssimon draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really 4157162911Ssimon appear there. 4158162911Ssimon 4159162911Ssimon Also deactive the remaining ciphersuites from 4160162911Ssimon draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as 4161162911Ssimon unofficial, and the ID has long expired. 4162162911Ssimon [Bodo Moeller] 4163162911Ssimon 4164162911Ssimon *) Fix RSA blinding Heisenbug (problems sometimes occured on 4165162911Ssimon dual-core machines) and other potential thread-safety issues. 4166162911Ssimon [Bodo Moeller] 4167162911Ssimon 4168162911Ssimon Changes between 0.9.7i and 0.9.7j [04 May 2006] 4169162911Ssimon 4170162911Ssimon *) Adapt fipsld and the build system to link against the validated FIPS 4171162911Ssimon module in FIPS mode. 4172162911Ssimon [Steve Henson] 4173162911Ssimon 4174162911Ssimon *) Fixes for VC++ 2005 build under Windows. 4175162911Ssimon [Steve Henson] 4176162911Ssimon 4177162911Ssimon *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make 4178162911Ssimon from a Windows bash shell such as MSYS. It is autodetected from the 4179162911Ssimon "config" script when run from a VC++ environment. Modify standard VC++ 4180162911Ssimon build to use fipscanister.o from the GNU make build. 4181162911Ssimon [Steve Henson] 4182162911Ssimon 4183160814Ssimon Changes between 0.9.7h and 0.9.7i [14 Oct 2005] 4184160814Ssimon 4185160814Ssimon *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. 4186160814Ssimon The value now differs depending on if you build for FIPS or not. 4187160814Ssimon BEWARE! A program linked with a shared FIPSed libcrypto can't be 4188160814Ssimon safely run with a non-FIPSed libcrypto, as it may crash because of 4189160814Ssimon the difference induced by this change. 4190160814Ssimon [Andy Polyakov] 4191160814Ssimon 4192160814Ssimon Changes between 0.9.7g and 0.9.7h [11 Oct 2005] 4193160814Ssimon 4194160814Ssimon *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING 4195160814Ssimon (part of SSL_OP_ALL). This option used to disable the 4196160814Ssimon countermeasure against man-in-the-middle protocol-version 4197160814Ssimon rollback in the SSL 2.0 server implementation, which is a bad 4198160814Ssimon idea. (CVE-2005-2969) 4199160814Ssimon 4200160814Ssimon [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center 4201160814Ssimon for Information Security, National Institute of Advanced Industrial 4202160814Ssimon Science and Technology [AIST], Japan)] 4203160814Ssimon 4204160814Ssimon *) Minimal support for X9.31 signatures and PSS padding modes. This is 4205160814Ssimon mainly for FIPS compliance and not fully integrated at this stage. 4206160814Ssimon [Steve Henson] 4207160814Ssimon 4208160814Ssimon *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform 4209160814Ssimon the exponentiation using a fixed-length exponent. (Otherwise, 4210160814Ssimon the information leaked through timing could expose the secret key 4211160814Ssimon after many signatures; cf. Bleichenbacher's attack on DSA with 4212160814Ssimon biased k.) 4213160814Ssimon [Bodo Moeller] 4214160814Ssimon 4215160814Ssimon *) Make a new fixed-window mod_exp implementation the default for 4216160814Ssimon RSA, DSA, and DH private-key operations so that the sequence of 4217160814Ssimon squares and multiplies and the memory access pattern are 4218160814Ssimon independent of the particular secret key. This will mitigate 4219160814Ssimon cache-timing and potential related attacks. 4220160814Ssimon 4221160814Ssimon BN_mod_exp_mont_consttime() is the new exponentiation implementation, 4222160814Ssimon and this is automatically used by BN_mod_exp_mont() if the new flag 4223160814Ssimon BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH 4224160814Ssimon will use this BN flag for private exponents unless the flag 4225160814Ssimon RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or 4226160814Ssimon DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. 4227160814Ssimon 4228160814Ssimon [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] 4229160814Ssimon 4230160814Ssimon *) Change the client implementation for SSLv23_method() and 4231160814Ssimon SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 4232160814Ssimon Client Hello message format if the SSL_OP_NO_SSLv2 option is set. 4233160814Ssimon (Previously, the SSL 2.0 backwards compatible Client Hello 4234160814Ssimon message format would be used even with SSL_OP_NO_SSLv2.) 4235160814Ssimon [Bodo Moeller] 4236160814Ssimon 4237160814Ssimon *) Add support for smime-type MIME parameter in S/MIME messages which some 4238160814Ssimon clients need. 4239160814Ssimon [Steve Henson] 4240160814Ssimon 4241160814Ssimon *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in 4242160814Ssimon a threadsafe manner. Modify rsa code to use new function and add calls 4243160814Ssimon to dsa and dh code (which had race conditions before). 4244160814Ssimon [Steve Henson] 4245160814Ssimon 4246160814Ssimon *) Include the fixed error library code in the C error file definitions 4247160814Ssimon instead of fixing them up at runtime. This keeps the error code 4248160814Ssimon structures constant. 4249160814Ssimon [Steve Henson] 4250160814Ssimon 4251160814Ssimon Changes between 0.9.7f and 0.9.7g [11 Apr 2005] 4252160814Ssimon 4253160814Ssimon [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after 4254160814Ssimon OpenSSL 0.9.8.] 4255160814Ssimon 4256160814Ssimon *) Fixes for newer kerberos headers. NB: the casts are needed because 4257160814Ssimon the 'length' field is signed on one version and unsigned on another 4258160814Ssimon with no (?) obvious way to tell the difference, without these VC++ 4259160814Ssimon complains. Also the "definition" of FAR (blank) is no longer included 4260160814Ssimon nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up 4261160814Ssimon some needed definitions. 4262160814Ssimon [Steve Henson] 4263160814Ssimon 4264160814Ssimon *) Undo Cygwin change. 4265291721Sjkim [Ulf M��ller] 4266160814Ssimon 4267160814Ssimon *) Added support for proxy certificates according to RFC 3820. 4268160814Ssimon Because they may be a security thread to unaware applications, 4269160814Ssimon they must be explicitely allowed in run-time. See 4270160814Ssimon docs/HOWTO/proxy_certificates.txt for further information. 4271160814Ssimon [Richard Levitte] 4272160814Ssimon 4273160814Ssimon Changes between 0.9.7e and 0.9.7f [22 Mar 2005] 4274160814Ssimon 4275160814Ssimon *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating 4276160814Ssimon server and client random values. Previously 4277160814Ssimon (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in 4278160814Ssimon less random data when sizeof(time_t) > 4 (some 64 bit platforms). 4279160814Ssimon 4280160814Ssimon This change has negligible security impact because: 4281160814Ssimon 4282160814Ssimon 1. Server and client random values still have 24 bytes of pseudo random 4283160814Ssimon data. 4284160814Ssimon 4285160814Ssimon 2. Server and client random values are sent in the clear in the initial 4286160814Ssimon handshake. 4287160814Ssimon 4288160814Ssimon 3. The master secret is derived using the premaster secret (48 bytes in 4289160814Ssimon size for static RSA ciphersuites) as well as client server and random 4290160814Ssimon values. 4291160814Ssimon 4292160814Ssimon The OpenSSL team would like to thank the UK NISCC for bringing this issue 4293160814Ssimon to our attention. 4294160814Ssimon 4295160814Ssimon [Stephen Henson, reported by UK NISCC] 4296160814Ssimon 4297160814Ssimon *) Use Windows randomness collection on Cygwin. 4298291721Sjkim [Ulf M��ller] 4299160814Ssimon 4300160814Ssimon *) Fix hang in EGD/PRNGD query when communication socket is closed 4301160814Ssimon prematurely by EGD/PRNGD. 4302291721Sjkim [Darren Tucker <dtucker@zip.com.au> via Lutz J��nicke, resolves #1014] 4303160814Ssimon 4304160814Ssimon *) Prompt for pass phrases when appropriate for PKCS12 input format. 4305160814Ssimon [Steve Henson] 4306160814Ssimon 4307160814Ssimon *) Back-port of selected performance improvements from development 4308160814Ssimon branch, as well as improved support for PowerPC platforms. 4309160814Ssimon [Andy Polyakov] 4310160814Ssimon 4311160814Ssimon *) Add lots of checks for memory allocation failure, error codes to indicate 4312160814Ssimon failure and freeing up memory if a failure occurs. 4313160814Ssimon [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] 4314160814Ssimon 4315160814Ssimon *) Add new -passin argument to dgst. 4316160814Ssimon [Steve Henson] 4317160814Ssimon 4318160814Ssimon *) Perform some character comparisons of different types in X509_NAME_cmp: 4319160814Ssimon this is needed for some certificates that reencode DNs into UTF8Strings 4320160814Ssimon (in violation of RFC3280) and can't or wont issue name rollover 4321160814Ssimon certificates. 4322160814Ssimon [Steve Henson] 4323160814Ssimon 4324160814Ssimon *) Make an explicit check during certificate validation to see that 4325160814Ssimon the CA setting in each certificate on the chain is correct. As a 4326160814Ssimon side effect always do the following basic checks on extensions, 4327160814Ssimon not just when there's an associated purpose to the check: 4328160814Ssimon 4329160814Ssimon - if there is an unhandled critical extension (unless the user 4330160814Ssimon has chosen to ignore this fault) 4331160814Ssimon - if the path length has been exceeded (if one is set at all) 4332160814Ssimon - that certain extensions fit the associated purpose (if one has 4333160814Ssimon been given) 4334160814Ssimon [Richard Levitte] 4335160814Ssimon 4336142425Snectar Changes between 0.9.7d and 0.9.7e [25 Oct 2004] 4337142425Snectar 4338142425Snectar *) Avoid a race condition when CRLs are checked in a multi threaded 4339142425Snectar environment. This would happen due to the reordering of the revoked 4340142425Snectar entries during signature checking and serial number lookup. Now the 4341142425Snectar encoding is cached and the serial number sort performed under a lock. 4342142425Snectar Add new STACK function sk_is_sorted(). 4343142425Snectar [Steve Henson] 4344142425Snectar 4345142425Snectar *) Add Delta CRL to the extension code. 4346142425Snectar [Steve Henson] 4347142425Snectar 4348142425Snectar *) Various fixes to s3_pkt.c so alerts are sent properly. 4349142425Snectar [David Holmes <d.holmes@f5.com>] 4350142425Snectar 4351142425Snectar *) Reduce the chances of duplicate issuer name and serial numbers (in 4352142425Snectar violation of RFC3280) using the OpenSSL certificate creation utilities. 4353142425Snectar This is done by creating a random 64 bit value for the initial serial 4354142425Snectar number when a serial number file is created or when a self signed 4355142425Snectar certificate is created using 'openssl req -x509'. The initial serial 4356142425Snectar number file is created using 'openssl x509 -next_serial' in CA.pl 4357142425Snectar rather than being initialized to 1. 4358142425Snectar [Steve Henson] 4359142425Snectar 4360127128Snectar Changes between 0.9.7c and 0.9.7d [17 Mar 2004] 4361127128Snectar 4362127128Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 4363160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 4364127128Snectar [Joe Orton, Steve Henson] 4365127128Snectar 4366127128Snectar *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites 4367160814Ssimon (CVE-2004-0112) 4368127128Snectar [Joe Orton, Steve Henson] 4369127128Snectar 4370127128Snectar *) Make it possible to have multiple active certificates with the same 4371127128Snectar subject in the CA index file. This is done only if the keyword 4372127128Snectar 'unique_subject' is set to 'no' in the main CA section (default 4373127128Snectar if 'CA_default') of the configuration file. The value is saved 4374127128Snectar with the database itself in a separate index attribute file, 4375127128Snectar named like the index file with '.attr' appended to the name. 4376127128Snectar [Richard Levitte] 4377127128Snectar 4378127128Snectar *) X509 verify fixes. Disable broken certificate workarounds when 4379127128Snectar X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if 4380127128Snectar keyUsage extension present. Don't accept CRLs with unhandled critical 4381127128Snectar extensions: since verify currently doesn't process CRL extensions this 4382127128Snectar rejects a CRL with *any* critical extensions. Add new verify error codes 4383127128Snectar for these cases. 4384127128Snectar [Steve Henson] 4385127128Snectar 4386127128Snectar *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. 4387127128Snectar A clarification of RFC2560 will require the use of OCTET STRINGs and 4388127128Snectar some implementations cannot handle the current raw format. Since OpenSSL 4389127128Snectar copies and compares OCSP nonces as opaque blobs without any attempt at 4390127128Snectar parsing them this should not create any compatibility issues. 4391127128Snectar [Steve Henson] 4392127128Snectar 4393127128Snectar *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when 4394127128Snectar calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without 4395127128Snectar this HMAC (and other) operations are several times slower than OpenSSL 4396127128Snectar < 0.9.7. 4397127128Snectar [Steve Henson] 4398127128Snectar 4399127128Snectar *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). 4400127128Snectar [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4401127128Snectar 4402127128Snectar *) Use the correct content when signing type "other". 4403127128Snectar [Steve Henson] 4404127128Snectar 4405120631Snectar Changes between 0.9.7b and 0.9.7c [30 Sep 2003] 4406120631Snectar 4407120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 4408120631Snectar 4409120631Snectar Stop out of bounds reads in the ASN1 code when presented with 4410160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 4411120631Snectar 4412160814Ssimon Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). 4413120631Snectar 4414120631Snectar If verify callback ignores invalid public key errors don't try to check 4415120631Snectar certificate signature with the NULL public key. 4416120631Snectar 4417120631Snectar [Steve Henson] 4418120631Snectar 4419120631Snectar *) New -ignore_err option in ocsp application to stop the server 4420120631Snectar exiting on the first error in a request. 4421120631Snectar [Steve Henson] 4422120631Snectar 4423120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 4424120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 4425120631Snectar specifications. 4426120631Snectar [Steve Henson] 4427120631Snectar 4428120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 4429120631Snectar extra data after the compression methods not only for TLS 1.0 4430120631Snectar but also for SSL 3.0 (as required by the specification). 4431120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 4432120631Snectar 4433120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 4434120631Snectar when it's 512 *bits* long, not 512 bytes. 4435120631Snectar [Richard Levitte] 4436120631Snectar 4437120631Snectar *) Change AES_cbc_encrypt() so it outputs exact multiple of 4438120631Snectar blocks during encryption. 4439120631Snectar [Richard Levitte] 4440120631Snectar 4441120631Snectar *) Various fixes to base64 BIO and non blocking I/O. On write 4442120631Snectar flushes were not handled properly if the BIO retried. On read 4443120631Snectar data was not being buffered properly and had various logic bugs. 4444120631Snectar This also affects blocking I/O when the data being decoded is a 4445120631Snectar certain size. 4446120631Snectar [Steve Henson] 4447120631Snectar 4448120631Snectar *) Various S/MIME bugfixes and compatibility changes: 4449120631Snectar output correct application/pkcs7 MIME type if 4450120631Snectar PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. 4451120631Snectar Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening 4452120631Snectar of files as .eml work). Correctly handle very long lines in MIME 4453120631Snectar parser. 4454120631Snectar [Steve Henson] 4455120631Snectar 4456120631Snectar Changes between 0.9.7a and 0.9.7b [10 Apr 2003] 4457120631Snectar 4458120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 4459120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 4460120631Snectar a protocol version number mismatch like a decryption error 4461120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 4462120631Snectar [Bodo Moeller] 4463120631Snectar 4464120631Snectar *) Turn on RSA blinding by default in the default implementation 4465120631Snectar to avoid a timing attack. Applications that don't want it can call 4466120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 4467120631Snectar They would be ill-advised to do so in most cases. 4468120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 4469120631Snectar 4470120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 4471120631Snectar seeded (in this case, the secret RSA exponent is abused as 4472120631Snectar an unpredictable seed -- if it is not unpredictable, there 4473120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 4474120631Snectar by remembering the creator's thread ID in rsa->blinding and 4475120631Snectar having all other threads use local one-time blinding factors 4476120631Snectar (this requires more computation than sharing rsa->blinding, but 4477120631Snectar avoids excessive locking; and if an RSA object is not shared 4478120631Snectar between threads, blinding will still be very fast). 4479120631Snectar [Bodo Moeller] 4480120631Snectar 4481120631Snectar *) Fixed a typo bug that would cause ENGINE_set_default() to set an 4482120631Snectar ENGINE as defaults for all supported algorithms irrespective of 4483120631Snectar the 'flags' parameter. 'flags' is now honoured, so applications 4484120631Snectar should make sure they are passing it correctly. 4485120631Snectar [Geoff Thorpe] 4486120631Snectar 4487120631Snectar *) Target "mingw" now allows native Windows code to be generated in 4488120631Snectar the Cygwin environment as well as with the MinGW compiler. 4489120631Snectar [Ulf Moeller] 4490120631Snectar 4491111147Snectar Changes between 0.9.7 and 0.9.7a [19 Feb 2003] 4492111147Snectar 4493111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 4494111147Snectar via timing by performing a MAC computation even if incorrrect 4495111147Snectar block cipher padding has been found. This is a countermeasure 4496111147Snectar against active attacks where the attacker has to distinguish 4497160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 4498111147Snectar 4499111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 4500111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 4501111147Snectar Martin Vuagnoux (EPFL, Ilion)] 4502111147Snectar 4503111147Snectar *) Make the no-err option work as intended. The intention with no-err 4504111147Snectar is not to have the whole error stack handling routines removed from 4505111147Snectar libcrypto, it's only intended to remove all the function name and 4506111147Snectar reason texts, thereby removing some of the footprint that may not 4507111147Snectar be interesting if those errors aren't displayed anyway. 4508111147Snectar 4509111147Snectar NOTE: it's still possible for any application or module to have it's 4510111147Snectar own set of error texts inserted. The routines are there, just not 4511111147Snectar used by default when no-err is given. 4512111147Snectar [Richard Levitte] 4513111147Snectar 4514111147Snectar *) Add support for FreeBSD on IA64. 4515111147Snectar [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] 4516111147Snectar 4517111147Snectar *) Adjust DES_cbc_cksum() so it returns the same value as the MIT 4518111147Snectar Kerberos function mit_des_cbc_cksum(). Before this change, 4519111147Snectar the value returned by DES_cbc_cksum() was like the one from 4520111147Snectar mit_des_cbc_cksum(), except the bytes were swapped. 4521111147Snectar [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] 4522111147Snectar 4523111147Snectar *) Allow an application to disable the automatic SSL chain building. 4524111147Snectar Before this a rather primitive chain build was always performed in 4525111147Snectar ssl3_output_cert_chain(): an application had no way to send the 4526111147Snectar correct chain if the automatic operation produced an incorrect result. 4527111147Snectar 4528111147Snectar Now the chain builder is disabled if either: 4529111147Snectar 4530111147Snectar 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). 4531111147Snectar 4532111147Snectar 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. 4533111147Snectar 4534111147Snectar The reasoning behind this is that an application would not want the 4535111147Snectar auto chain building to take place if extra chain certificates are 4536111147Snectar present and it might also want a means of sending no additional 4537111147Snectar certificates (for example the chain has two certificates and the 4538111147Snectar root is omitted). 4539111147Snectar [Steve Henson] 4540111147Snectar 4541111147Snectar *) Add the possibility to build without the ENGINE framework. 4542111147Snectar [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4543111147Snectar 4544111147Snectar *) Under Win32 gmtime() can return NULL: check return value in 4545111147Snectar OPENSSL_gmtime(). Add error code for case where gmtime() fails. 4546111147Snectar [Steve Henson] 4547111147Snectar 4548111147Snectar *) DSA routines: under certain error conditions uninitialized BN objects 4549111147Snectar could be freed. Solution: make sure initialization is performed early 4550111147Snectar enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, 4551111147Snectar Nils Larsch <nla@trustcenter.de> via PR#459) 4552111147Snectar [Lutz Jaenicke] 4553111147Snectar 4554111147Snectar *) Another fix for SSLv2 session ID handling: the session ID was incorrectly 4555111147Snectar checked on reconnect on the client side, therefore session resumption 4556111147Snectar could still fail with a "ssl session id is different" error. This 4557111147Snectar behaviour is masked when SSL_OP_ALL is used due to 4558111147Snectar SSL_OP_MICROSOFT_SESS_ID_BUG being set. 4559111147Snectar Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4560111147Snectar followup to PR #377. 4561111147Snectar [Lutz Jaenicke] 4562111147Snectar 4563111147Snectar *) IA-32 assembler support enhancements: unified ELF targets, support 4564111147Snectar for SCO/Caldera platforms, fix for Cygwin shared build. 4565111147Snectar [Andy Polyakov] 4566111147Snectar 4567111147Snectar *) Add support for FreeBSD on sparc64. As a consequence, support for 4568111147Snectar FreeBSD on non-x86 processors is separate from x86 processors on 4569111147Snectar the config script, much like the NetBSD support. 4570111147Snectar [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] 4571111147Snectar 4572109998Smarkm Changes between 0.9.6h and 0.9.7 [31 Dec 2002] 4573109998Smarkm 4574120631Snectar [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after 4575120631Snectar OpenSSL 0.9.7.] 4576120631Snectar 4577109998Smarkm *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED 4578109998Smarkm code (06) was taken as the first octet of the session ID and the last 4579109998Smarkm octet was ignored consequently. As a result SSLv2 client side session 4580109998Smarkm caching could not have worked due to the session ID mismatch between 4581109998Smarkm client and server. 4582109998Smarkm Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as 4583109998Smarkm PR #377. 4584109998Smarkm [Lutz Jaenicke] 4585109998Smarkm 4586109998Smarkm *) Change the declaration of needed Kerberos libraries to use EX_LIBS 4587109998Smarkm instead of the special (and badly supported) LIBKRB5. LIBKRB5 is 4588109998Smarkm removed entirely. 4589109998Smarkm [Richard Levitte] 4590109998Smarkm 4591109998Smarkm *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it 4592109998Smarkm seems that in spite of existing for more than a year, many application 4593109998Smarkm author have done nothing to provide the necessary callbacks, which 4594109998Smarkm means that this particular engine will not work properly anywhere. 4595109998Smarkm This is a very unfortunate situation which forces us, in the name 4596109998Smarkm of usability, to give the hw_ncipher.c a static lock, which is part 4597109998Smarkm of libcrypto. 4598109998Smarkm NOTE: This is for the 0.9.7 series ONLY. This hack will never 4599109998Smarkm appear in 0.9.8 or later. We EXPECT application authors to have 4600109998Smarkm dealt properly with this when 0.9.8 is released (unless we actually 4601109998Smarkm make such changes in the libcrypto locking code that changes will 4602109998Smarkm have to be made anyway). 4603109998Smarkm [Richard Levitte] 4604109998Smarkm 4605109998Smarkm *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content 4606109998Smarkm octets have been read, EOF or an error occurs. Without this change 4607109998Smarkm some truncated ASN1 structures will not produce an error. 4608109998Smarkm [Steve Henson] 4609109998Smarkm 4610109998Smarkm *) Disable Heimdal support, since it hasn't been fully implemented. 4611109998Smarkm Still give the possibility to force the use of Heimdal, but with 4612109998Smarkm warnings and a request that patches get sent to openssl-dev. 4613109998Smarkm [Richard Levitte] 4614109998Smarkm 4615109998Smarkm *) Add the VC-CE target, introduce the WINCE sysname, and add 4616109998Smarkm INSTALL.WCE and appropriate conditionals to make it build. 4617109998Smarkm [Steven Reddie <smr@essemer.com.au> via Richard Levitte] 4618109998Smarkm 4619109998Smarkm *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and 4620109998Smarkm cygssl-x.y.z.dll, where x, y and z are the major, minor and 4621109998Smarkm edit numbers of the version. 4622109998Smarkm [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 4623109998Smarkm 4624109998Smarkm *) Introduce safe string copy and catenation functions 4625109998Smarkm (BUF_strlcpy() and BUF_strlcat()). 4626109998Smarkm [Ben Laurie (CHATS) and Richard Levitte] 4627109998Smarkm 4628109998Smarkm *) Avoid using fixed-size buffers for one-line DNs. 4629109998Smarkm [Ben Laurie (CHATS)] 4630109998Smarkm 4631109998Smarkm *) Add BUF_MEM_grow_clean() to avoid information leakage when 4632109998Smarkm resizing buffers containing secrets, and use where appropriate. 4633109998Smarkm [Ben Laurie (CHATS)] 4634109998Smarkm 4635109998Smarkm *) Avoid using fixed size buffers for configuration file location. 4636109998Smarkm [Ben Laurie (CHATS)] 4637109998Smarkm 4638109998Smarkm *) Avoid filename truncation for various CA files. 4639109998Smarkm [Ben Laurie (CHATS)] 4640109998Smarkm 4641109998Smarkm *) Use sizeof in preference to magic numbers. 4642109998Smarkm [Ben Laurie (CHATS)] 4643109998Smarkm 4644109998Smarkm *) Avoid filename truncation in cert requests. 4645109998Smarkm [Ben Laurie (CHATS)] 4646109998Smarkm 4647109998Smarkm *) Add assertions to check for (supposedly impossible) buffer 4648109998Smarkm overflows. 4649109998Smarkm [Ben Laurie (CHATS)] 4650109998Smarkm 4651109998Smarkm *) Don't cache truncated DNS entries in the local cache (this could 4652109998Smarkm potentially lead to a spoofing attack). 4653109998Smarkm [Ben Laurie (CHATS)] 4654109998Smarkm 4655109998Smarkm *) Fix various buffers to be large enough for hex/decimal 4656109998Smarkm representations in a platform independent manner. 4657109998Smarkm [Ben Laurie (CHATS)] 4658109998Smarkm 4659109998Smarkm *) Add CRYPTO_realloc_clean() to avoid information leakage when 4660109998Smarkm resizing buffers containing secrets, and use where appropriate. 4661109998Smarkm [Ben Laurie (CHATS)] 4662109998Smarkm 4663109998Smarkm *) Add BIO_indent() to avoid much slightly worrying code to do 4664109998Smarkm indents. 4665109998Smarkm [Ben Laurie (CHATS)] 4666109998Smarkm 4667109998Smarkm *) Convert sprintf()/BIO_puts() to BIO_printf(). 4668109998Smarkm [Ben Laurie (CHATS)] 4669109998Smarkm 4670109998Smarkm *) buffer_gets() could terminate with the buffer only half 4671109998Smarkm full. Fixed. 4672109998Smarkm [Ben Laurie (CHATS)] 4673109998Smarkm 4674109998Smarkm *) Add assertions to prevent user-supplied crypto functions from 4675109998Smarkm overflowing internal buffers by having large block sizes, etc. 4676109998Smarkm [Ben Laurie (CHATS)] 4677109998Smarkm 4678109998Smarkm *) New OPENSSL_assert() macro (similar to assert(), but enabled 4679109998Smarkm unconditionally). 4680109998Smarkm [Ben Laurie (CHATS)] 4681109998Smarkm 4682109998Smarkm *) Eliminate unused copy of key in RC4. 4683109998Smarkm [Ben Laurie (CHATS)] 4684109998Smarkm 4685109998Smarkm *) Eliminate unused and incorrectly sized buffers for IV in pem.h. 4686109998Smarkm [Ben Laurie (CHATS)] 4687109998Smarkm 4688109998Smarkm *) Fix off-by-one error in EGD path. 4689109998Smarkm [Ben Laurie (CHATS)] 4690109998Smarkm 4691109998Smarkm *) If RANDFILE path is too long, ignore instead of truncating. 4692109998Smarkm [Ben Laurie (CHATS)] 4693109998Smarkm 4694109998Smarkm *) Eliminate unused and incorrectly sized X.509 structure 4695109998Smarkm CBCParameter. 4696109998Smarkm [Ben Laurie (CHATS)] 4697109998Smarkm 4698109998Smarkm *) Eliminate unused and dangerous function knumber(). 4699109998Smarkm [Ben Laurie (CHATS)] 4700109998Smarkm 4701109998Smarkm *) Eliminate unused and dangerous structure, KSSL_ERR. 4702109998Smarkm [Ben Laurie (CHATS)] 4703109998Smarkm 4704109998Smarkm *) Protect against overlong session ID context length in an encoded 4705109998Smarkm session object. Since these are local, this does not appear to be 4706109998Smarkm exploitable. 4707109998Smarkm [Ben Laurie (CHATS)] 4708109998Smarkm 4709109998Smarkm *) Change from security patch (see 0.9.6e below) that did not affect 4710109998Smarkm the 0.9.6 release series: 4711109998Smarkm 4712109998Smarkm Remote buffer overflow in SSL3 protocol - an attacker could 4713109998Smarkm supply an oversized master key in Kerberos-enabled versions. 4714160814Ssimon (CVE-2002-0657) 4715109998Smarkm [Ben Laurie (CHATS)] 4716109998Smarkm 4717109998Smarkm *) Change the SSL kerb5 codes to match RFC 2712. 4718109998Smarkm [Richard Levitte] 4719109998Smarkm 4720109998Smarkm *) Make -nameopt work fully for req and add -reqopt switch. 4721109998Smarkm [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] 4722109998Smarkm 4723109998Smarkm *) The "block size" for block ciphers in CFB and OFB mode should be 1. 4724109998Smarkm [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] 4725109998Smarkm 4726109998Smarkm *) Make sure tests can be performed even if the corresponding algorithms 4727109998Smarkm have been removed entirely. This was also the last step to make 4728109998Smarkm OpenSSL compilable with DJGPP under all reasonable conditions. 4729109998Smarkm [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] 4730109998Smarkm 4731109998Smarkm *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT 4732109998Smarkm to allow version independent disabling of normally unselected ciphers, 4733109998Smarkm which may be activated as a side-effect of selecting a single cipher. 4734109998Smarkm 4735109998Smarkm (E.g., cipher list string "RSA" enables ciphersuites that are left 4736109998Smarkm out of "ALL" because they do not provide symmetric encryption. 4737109998Smarkm "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) 4738109998Smarkm [Lutz Jaenicke, Bodo Moeller] 4739109998Smarkm 4740109998Smarkm *) Add appropriate support for separate platform-dependent build 4741109998Smarkm directories. The recommended way to make a platform-dependent 4742109998Smarkm build directory is the following (tested on Linux), maybe with 4743109998Smarkm some local tweaks: 4744109998Smarkm 4745109998Smarkm # Place yourself outside of the OpenSSL source tree. In 4746109998Smarkm # this example, the environment variable OPENSSL_SOURCE 4747109998Smarkm # is assumed to contain the absolute OpenSSL source directory. 4748109998Smarkm mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" 4749109998Smarkm cd objtree/"`uname -s`-`uname -r`-`uname -m`" 4750111147Snectar (cd $OPENSSL_SOURCE; find . -type f) | while read F; do 4751109998Smarkm mkdir -p `dirname $F` 4752109998Smarkm ln -s $OPENSSL_SOURCE/$F $F 4753109998Smarkm done 4754109998Smarkm 4755109998Smarkm To be absolutely sure not to disturb the source tree, a "make clean" 4756109998Smarkm is a good thing. If it isn't successfull, don't worry about it, 4757109998Smarkm it probably means the source directory is very clean. 4758109998Smarkm [Richard Levitte] 4759109998Smarkm 4760109998Smarkm *) Make sure any ENGINE control commands make local copies of string 4761109998Smarkm pointers passed to them whenever necessary. Otherwise it is possible 4762109998Smarkm the caller may have overwritten (or deallocated) the original string 4763109998Smarkm data when a later ENGINE operation tries to use the stored values. 4764291721Sjkim [G��tz Babin-Ebell <babinebell@trustcenter.de>] 4765109998Smarkm 4766109998Smarkm *) Improve diagnostics in file reading and command-line digests. 4767109998Smarkm [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] 4768109998Smarkm 4769109998Smarkm *) Add AES modes CFB and OFB to the object database. Correct an 4770109998Smarkm error in AES-CFB decryption. 4771109998Smarkm [Richard Levitte] 4772109998Smarkm 4773109998Smarkm *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this 4774109998Smarkm allows existing EVP_CIPHER_CTX structures to be reused after 4775109998Smarkm calling EVP_*Final(). This behaviour is used by encryption 4776109998Smarkm BIOs and some applications. This has the side effect that 4777109998Smarkm applications must explicitly clean up cipher contexts with 4778109998Smarkm EVP_CIPHER_CTX_cleanup() or they will leak memory. 4779109998Smarkm [Steve Henson] 4780109998Smarkm 4781109998Smarkm *) Check the values of dna and dnb in bn_mul_recursive before calling 4782109998Smarkm bn_mul_comba (a non zero value means the a or b arrays do not contain 4783109998Smarkm n2 elements) and fallback to bn_mul_normal if either is not zero. 4784109998Smarkm [Steve Henson] 4785109998Smarkm 4786109998Smarkm *) Fix escaping of non-ASCII characters when using the -subj option 4787109998Smarkm of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) 4788109998Smarkm [Lutz Jaenicke] 4789109998Smarkm 4790109998Smarkm *) Make object definitions compliant to LDAP (RFC2256): SN is the short 4791109998Smarkm form for "surname", serialNumber has no short form. 4792109998Smarkm Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; 4793109998Smarkm therefore remove "mail" short name for "internet 7". 4794109998Smarkm The OID for unique identifiers in X509 certificates is 4795109998Smarkm x500UniqueIdentifier, not uniqueIdentifier. 4796109998Smarkm Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) 4797109998Smarkm [Lutz Jaenicke] 4798109998Smarkm 4799109998Smarkm *) Add an "init" command to the ENGINE config module and auto initialize 4800109998Smarkm ENGINEs. Without any "init" command the ENGINE will be initialized 4801109998Smarkm after all ctrl commands have been executed on it. If init=1 the 4802109998Smarkm ENGINE is initailized at that point (ctrls before that point are run 4803109998Smarkm on the uninitialized ENGINE and after on the initialized one). If 4804109998Smarkm init=0 then the ENGINE will not be iniatialized at all. 4805109998Smarkm [Steve Henson] 4806109998Smarkm 4807109998Smarkm *) Fix the 'app_verify_callback' interface so that the user-defined 4808109998Smarkm argument is actually passed to the callback: In the 4809109998Smarkm SSL_CTX_set_cert_verify_callback() prototype, the callback 4810109998Smarkm declaration has been changed from 4811109998Smarkm int (*cb)() 4812109998Smarkm into 4813109998Smarkm int (*cb)(X509_STORE_CTX *,void *); 4814109998Smarkm in ssl_verify_cert_chain (ssl/ssl_cert.c), the call 4815109998Smarkm i=s->ctx->app_verify_callback(&ctx) 4816109998Smarkm has been changed into 4817109998Smarkm i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). 4818109998Smarkm 4819109998Smarkm To update applications using SSL_CTX_set_cert_verify_callback(), 4820109998Smarkm a dummy argument can be added to their callback functions. 4821109998Smarkm [D. K. Smetters <smetters@parc.xerox.com>] 4822109998Smarkm 4823109998Smarkm *) Added the '4758cca' ENGINE to support IBM 4758 cards. 4824109998Smarkm [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] 4825109998Smarkm 4826109998Smarkm *) Add and OPENSSL_LOAD_CONF define which will cause 4827109998Smarkm OpenSSL_add_all_algorithms() to load the openssl.cnf config file. 4828109998Smarkm This allows older applications to transparently support certain 4829109998Smarkm OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. 4830109998Smarkm Two new functions OPENSSL_add_all_algorithms_noconf() which will never 4831109998Smarkm load the config file and OPENSSL_add_all_algorithms_conf() which will 4832109998Smarkm always load it have also been added. 4833109998Smarkm [Steve Henson] 4834109998Smarkm 4835109998Smarkm *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. 4836109998Smarkm Adjust NIDs and EVP layer. 4837109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4838109998Smarkm 4839109998Smarkm *) Config modules support in openssl utility. 4840109998Smarkm 4841109998Smarkm Most commands now load modules from the config file, 4842109998Smarkm though in a few (such as version) this isn't done 4843109998Smarkm because it couldn't be used for anything. 4844109998Smarkm 4845109998Smarkm In the case of ca and req the config file used is 4846109998Smarkm the same as the utility itself: that is the -config 4847109998Smarkm command line option can be used to specify an 4848109998Smarkm alternative file. 4849109998Smarkm [Steve Henson] 4850109998Smarkm 4851109998Smarkm *) Move default behaviour from OPENSSL_config(). If appname is NULL 4852109998Smarkm use "openssl_conf" if filename is NULL use default openssl config file. 4853109998Smarkm [Steve Henson] 4854109998Smarkm 4855109998Smarkm *) Add an argument to OPENSSL_config() to allow the use of an alternative 4856109998Smarkm config section name. Add a new flag to tolerate a missing config file 4857109998Smarkm and move code to CONF_modules_load_file(). 4858109998Smarkm [Steve Henson] 4859109998Smarkm 4860109998Smarkm *) Support for crypto accelerator cards from Accelerated Encryption 4861109998Smarkm Processing, www.aep.ie. (Use engine 'aep') 4862109998Smarkm The support was copied from 0.9.6c [engine] and adapted/corrected 4863109998Smarkm to work with the new engine framework. 4864109998Smarkm [AEP Inc. and Richard Levitte] 4865109998Smarkm 4866109998Smarkm *) Support for SureWare crypto accelerator cards from Baltimore 4867109998Smarkm Technologies. (Use engine 'sureware') 4868109998Smarkm The support was copied from 0.9.6c [engine] and adapted 4869109998Smarkm to work with the new engine framework. 4870109998Smarkm [Richard Levitte] 4871109998Smarkm 4872109998Smarkm *) Have the CHIL engine fork-safe (as defined by nCipher) and actually 4873109998Smarkm make the newer ENGINE framework commands for the CHIL engine work. 4874109998Smarkm [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] 4875109998Smarkm 4876109998Smarkm *) Make it possible to produce shared libraries on ReliantUNIX. 4877109998Smarkm [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] 4878109998Smarkm 4879109998Smarkm *) Add the configuration target debug-linux-ppro. 4880109998Smarkm Make 'openssl rsa' use the general key loading routines 4881109998Smarkm implemented in apps.c, and make those routines able to 4882109998Smarkm handle the key format FORMAT_NETSCAPE and the variant 4883109998Smarkm FORMAT_IISSGC. 4884109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4885109998Smarkm 4886109998Smarkm *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 4887109998Smarkm [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 4888109998Smarkm 4889109998Smarkm *) Add -keyform to rsautl, and document -engine. 4890109998Smarkm [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] 4891109998Smarkm 4892109998Smarkm *) Change BIO_new_file (crypto/bio/bss_file.c) to use new 4893109998Smarkm BIO_R_NO_SUCH_FILE error code rather than the generic 4894109998Smarkm ERR_R_SYS_LIB error code if fopen() fails with ENOENT. 4895109998Smarkm [Ben Laurie] 4896109998Smarkm 4897109998Smarkm *) Add new functions 4898109998Smarkm ERR_peek_last_error 4899109998Smarkm ERR_peek_last_error_line 4900109998Smarkm ERR_peek_last_error_line_data. 4901109998Smarkm These are similar to 4902109998Smarkm ERR_peek_error 4903109998Smarkm ERR_peek_error_line 4904109998Smarkm ERR_peek_error_line_data, 4905109998Smarkm but report on the latest error recorded rather than the first one 4906109998Smarkm still in the error queue. 4907109998Smarkm [Ben Laurie, Bodo Moeller] 4908109998Smarkm 4909109998Smarkm *) default_algorithms option in ENGINE config module. This allows things 4910109998Smarkm like: 4911109998Smarkm default_algorithms = ALL 4912109998Smarkm default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS 4913109998Smarkm [Steve Henson] 4914109998Smarkm 4915109998Smarkm *) Prelminary ENGINE config module. 4916109998Smarkm [Steve Henson] 4917109998Smarkm 4918109998Smarkm *) New experimental application configuration code. 4919109998Smarkm [Steve Henson] 4920109998Smarkm 4921109998Smarkm *) Change the AES code to follow the same name structure as all other 4922109998Smarkm symmetric ciphers, and behave the same way. Move everything to 4923109998Smarkm the directory crypto/aes, thereby obsoleting crypto/rijndael. 4924109998Smarkm [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] 4925109998Smarkm 4926109998Smarkm *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. 4927109998Smarkm [Ben Laurie and Theo de Raadt] 4928109998Smarkm 4929109998Smarkm *) Add option to output public keys in req command. 4930109998Smarkm [Massimiliano Pala madwolf@openca.org] 4931109998Smarkm 4932109998Smarkm *) Use wNAFs in EC_POINTs_mul() for improved efficiency 4933109998Smarkm (up to about 10% better than before for P-192 and P-224). 4934109998Smarkm [Bodo Moeller] 4935109998Smarkm 4936109998Smarkm *) New functions/macros 4937109998Smarkm 4938109998Smarkm SSL_CTX_set_msg_callback(ctx, cb) 4939109998Smarkm SSL_CTX_set_msg_callback_arg(ctx, arg) 4940109998Smarkm SSL_set_msg_callback(ssl, cb) 4941109998Smarkm SSL_set_msg_callback_arg(ssl, arg) 4942109998Smarkm 4943109998Smarkm to request calling a callback function 4944109998Smarkm 4945109998Smarkm void cb(int write_p, int version, int content_type, 4946109998Smarkm const void *buf, size_t len, SSL *ssl, void *arg) 4947109998Smarkm 4948109998Smarkm whenever a protocol message has been completely received 4949109998Smarkm (write_p == 0) or sent (write_p == 1). Here 'version' is the 4950109998Smarkm protocol version according to which the SSL library interprets 4951109998Smarkm the current protocol message (SSL2_VERSION, SSL3_VERSION, or 4952109998Smarkm TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or 4953109998Smarkm the content type as defined in the SSL 3.0/TLS 1.0 protocol 4954109998Smarkm specification (change_cipher_spec(20), alert(21), handshake(22)). 4955109998Smarkm 'buf' and 'len' point to the actual message, 'ssl' to the 4956109998Smarkm SSL object, and 'arg' is the application-defined value set by 4957109998Smarkm SSL[_CTX]_set_msg_callback_arg(). 4958109998Smarkm 4959109998Smarkm 'openssl s_client' and 'openssl s_server' have new '-msg' options 4960109998Smarkm to enable a callback that displays all protocol messages. 4961109998Smarkm [Bodo Moeller] 4962109998Smarkm 4963109998Smarkm *) Change the shared library support so shared libraries are built as 4964109998Smarkm soon as the corresponding static library is finished, and thereby get 4965109998Smarkm openssl and the test programs linked against the shared library. 4966109998Smarkm This still only happens when the keyword "shard" has been given to 4967109998Smarkm the configuration scripts. 4968109998Smarkm 4969109998Smarkm NOTE: shared library support is still an experimental thing, and 4970109998Smarkm backward binary compatibility is still not guaranteed. 4971109998Smarkm ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] 4972109998Smarkm 4973109998Smarkm *) Add support for Subject Information Access extension. 4974109998Smarkm [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] 4975109998Smarkm 4976109998Smarkm *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero 4977109998Smarkm additional bytes when new memory had to be allocated, not just 4978109998Smarkm when reusing an existing buffer. 4979109998Smarkm [Bodo Moeller] 4980109998Smarkm 4981109998Smarkm *) New command line and configuration option 'utf8' for the req command. 4982109998Smarkm This allows field values to be specified as UTF8 strings. 4983109998Smarkm [Steve Henson] 4984109998Smarkm 4985109998Smarkm *) Add -multi and -mr options to "openssl speed" - giving multiple parallel 4986109998Smarkm runs for the former and machine-readable output for the latter. 4987109998Smarkm [Ben Laurie] 4988109998Smarkm 4989109998Smarkm *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion 4990109998Smarkm of the e-mail address in the DN (i.e., it will go into a certificate 4991109998Smarkm extension only). The new configuration file option 'email_in_dn = no' 4992109998Smarkm has the same effect. 4993109998Smarkm [Massimiliano Pala madwolf@openca.org] 4994109998Smarkm 4995109998Smarkm *) Change all functions with names starting with des_ to be starting 4996109998Smarkm with DES_ instead. Add wrappers that are compatible with libdes, 4997109998Smarkm but are named _ossl_old_des_*. Finally, add macros that map the 4998109998Smarkm des_* symbols to the corresponding _ossl_old_des_* if libdes 4999109998Smarkm compatibility is desired. If OpenSSL 0.9.6c compatibility is 5000109998Smarkm desired, the des_* symbols will be mapped to DES_*, with one 5001109998Smarkm exception. 5002109998Smarkm 5003109998Smarkm Since we provide two compatibility mappings, the user needs to 5004109998Smarkm define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes 5005109998Smarkm compatibility is desired. The default (i.e., when that macro 5006109998Smarkm isn't defined) is OpenSSL 0.9.6c compatibility. 5007109998Smarkm 5008109998Smarkm There are also macros that enable and disable the support of old 5009109998Smarkm des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT 5010109998Smarkm and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those 5011109998Smarkm are defined, the default will apply: to support the old des routines. 5012109998Smarkm 5013109998Smarkm In either case, one must include openssl/des.h to get the correct 5014109998Smarkm definitions. Do not try to just include openssl/des_old.h, that 5015109998Smarkm won't work. 5016109998Smarkm 5017109998Smarkm NOTE: This is a major break of an old API into a new one. Software 5018109998Smarkm authors are encouraged to switch to the DES_ style functions. Some 5019109998Smarkm time in the future, des_old.h and the libdes compatibility functions 5020109998Smarkm will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the 5021109998Smarkm default), and then completely removed. 5022109998Smarkm [Richard Levitte] 5023109998Smarkm 5024109998Smarkm *) Test for certificates which contain unsupported critical extensions. 5025109998Smarkm If such a certificate is found during a verify operation it is 5026109998Smarkm rejected by default: this behaviour can be overridden by either 5027109998Smarkm handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or 5028109998Smarkm by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function 5029109998Smarkm X509_supported_extension() has also been added which returns 1 if a 5030109998Smarkm particular extension is supported. 5031109998Smarkm [Steve Henson] 5032109998Smarkm 5033109998Smarkm *) Modify the behaviour of EVP cipher functions in similar way to digests 5034109998Smarkm to retain compatibility with existing code. 5035109998Smarkm [Steve Henson] 5036109998Smarkm 5037109998Smarkm *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain 5038109998Smarkm compatibility with existing code. In particular the 'ctx' parameter does 5039109998Smarkm not have to be to be initialized before the call to EVP_DigestInit() and 5040109998Smarkm it is tidied up after a call to EVP_DigestFinal(). New function 5041109998Smarkm EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function 5042109998Smarkm EVP_MD_CTX_copy() changed to not require the destination to be 5043109998Smarkm initialized valid and new function EVP_MD_CTX_copy_ex() added which 5044109998Smarkm requires the destination to be valid. 5045109998Smarkm 5046109998Smarkm Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), 5047109998Smarkm EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). 5048109998Smarkm [Steve Henson] 5049109998Smarkm 5050109998Smarkm *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it 5051109998Smarkm so that complete 'Handshake' protocol structures are kept in memory 5052109998Smarkm instead of overwriting 'msg_type' and 'length' with 'body' data. 5053109998Smarkm [Bodo Moeller] 5054109998Smarkm 5055109998Smarkm *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. 5056109998Smarkm [Massimo Santin via Richard Levitte] 5057109998Smarkm 5058109998Smarkm *) Major restructuring to the underlying ENGINE code. This includes 5059109998Smarkm reduction of linker bloat, separation of pure "ENGINE" manipulation 5060109998Smarkm (initialisation, etc) from functionality dealing with implementations 5061109998Smarkm of specific crypto iterfaces. This change also introduces integrated 5062109998Smarkm support for symmetric ciphers and digest implementations - so ENGINEs 5063109998Smarkm can now accelerate these by providing EVP_CIPHER and EVP_MD 5064109998Smarkm implementations of their own. This is detailed in crypto/engine/README 5065109998Smarkm as it couldn't be adequately described here. However, there are a few 5066109998Smarkm API changes worth noting - some RSA, DSA, DH, and RAND functions that 5067109998Smarkm were changed in the original introduction of ENGINE code have now 5068109998Smarkm reverted back - the hooking from this code to ENGINE is now a good 5069109998Smarkm deal more passive and at run-time, operations deal directly with 5070109998Smarkm RSA_METHODs, DSA_METHODs (etc) as they did before, rather than 5071109998Smarkm dereferencing through an ENGINE pointer any more. Also, the ENGINE 5072109998Smarkm functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - 5073109998Smarkm they were not being used by the framework as there is no concept of a 5074109998Smarkm BIGNUM_METHOD and they could not be generalised to the new 5075109998Smarkm 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, 5076109998Smarkm ENGINE_cpy() has been removed as it cannot be consistently defined in 5077109998Smarkm the new code. 5078109998Smarkm [Geoff Thorpe] 5079109998Smarkm 5080109998Smarkm *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. 5081109998Smarkm [Steve Henson] 5082109998Smarkm 5083109998Smarkm *) Change mkdef.pl to sort symbols that get the same entry number, 5084109998Smarkm and make sure the automatically generated functions ERR_load_* 5085109998Smarkm become part of libeay.num as well. 5086109998Smarkm [Richard Levitte] 5087109998Smarkm 5088109998Smarkm *) New function SSL_renegotiate_pending(). This returns true once 5089109998Smarkm renegotiation has been requested (either SSL_renegotiate() call 5090109998Smarkm or HelloRequest/ClientHello receveived from the peer) and becomes 5091109998Smarkm false once a handshake has been completed. 5092109998Smarkm (For servers, SSL_renegotiate() followed by SSL_do_handshake() 5093109998Smarkm sends a HelloRequest, but does not ensure that a handshake takes 5094109998Smarkm place. SSL_renegotiate_pending() is useful for checking if the 5095109998Smarkm client has followed the request.) 5096109998Smarkm [Bodo Moeller] 5097109998Smarkm 5098109998Smarkm *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. 5099109998Smarkm By default, clients may request session resumption even during 5100109998Smarkm renegotiation (if session ID contexts permit); with this option, 5101109998Smarkm session resumption is possible only in the first handshake. 5102109998Smarkm 5103109998Smarkm SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes 5104109998Smarkm more bits available for options that should not be part of 5105109998Smarkm SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). 5106109998Smarkm [Bodo Moeller] 5107109998Smarkm 5108109998Smarkm *) Add some demos for certificate and certificate request creation. 5109109998Smarkm [Steve Henson] 5110109998Smarkm 5111109998Smarkm *) Make maximum certificate chain size accepted from the peer application 5112109998Smarkm settable (SSL*_get/set_max_cert_list()), as proposed by 5113109998Smarkm "Douglas E. Engert" <deengert@anl.gov>. 5114109998Smarkm [Lutz Jaenicke] 5115109998Smarkm 5116109998Smarkm *) Add support for shared libraries for Unixware-7 5117109998Smarkm (Boyd Lynn Gerber <gerberb@zenez.com>). 5118109998Smarkm [Lutz Jaenicke] 5119109998Smarkm 5120109998Smarkm *) Add a "destroy" handler to ENGINEs that allows structural cleanup to 5121109998Smarkm be done prior to destruction. Use this to unload error strings from 5122109998Smarkm ENGINEs that load their own error strings. NB: This adds two new API 5123109998Smarkm functions to "get" and "set" this destroy handler in an ENGINE. 5124109998Smarkm [Geoff Thorpe] 5125109998Smarkm 5126109998Smarkm *) Alter all existing ENGINE implementations (except "openssl" and 5127109998Smarkm "openbsd") to dynamically instantiate their own error strings. This 5128109998Smarkm makes them more flexible to be built both as statically-linked ENGINEs 5129109998Smarkm and self-contained shared-libraries loadable via the "dynamic" ENGINE. 5130109998Smarkm Also, add stub code to each that makes building them as self-contained 5131109998Smarkm shared-libraries easier (see README.ENGINE). 5132109998Smarkm [Geoff Thorpe] 5133109998Smarkm 5134109998Smarkm *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE 5135109998Smarkm implementations into applications that are completely implemented in 5136109998Smarkm self-contained shared-libraries. The "dynamic" ENGINE exposes control 5137109998Smarkm commands that can be used to configure what shared-library to load and 5138109998Smarkm to control aspects of the way it is handled. Also, made an update to 5139109998Smarkm the README.ENGINE file that brings its information up-to-date and 5140109998Smarkm provides some information and instructions on the "dynamic" ENGINE 5141109998Smarkm (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). 5142109998Smarkm [Geoff Thorpe] 5143109998Smarkm 5144109998Smarkm *) Make it possible to unload ranges of ERR strings with a new 5145109998Smarkm "ERR_unload_strings" function. 5146109998Smarkm [Geoff Thorpe] 5147109998Smarkm 5148109998Smarkm *) Add a copy() function to EVP_MD. 5149109998Smarkm [Ben Laurie] 5150109998Smarkm 5151109998Smarkm *) Make EVP_MD routines take a context pointer instead of just the 5152109998Smarkm md_data void pointer. 5153109998Smarkm [Ben Laurie] 5154109998Smarkm 5155109998Smarkm *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates 5156109998Smarkm that the digest can only process a single chunk of data 5157109998Smarkm (typically because it is provided by a piece of 5158109998Smarkm hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application 5159109998Smarkm is only going to provide a single chunk of data, and hence the 5160109998Smarkm framework needn't accumulate the data for oneshot drivers. 5161109998Smarkm [Ben Laurie] 5162109998Smarkm 5163109998Smarkm *) As with "ERR", make it possible to replace the underlying "ex_data" 5164109998Smarkm functions. This change also alters the storage and management of global 5165109998Smarkm ex_data state - it's now all inside ex_data.c and all "class" code (eg. 5166109998Smarkm RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class 5167109998Smarkm index counters. The API functions that use this state have been changed 5168109998Smarkm to take a "class_index" rather than pointers to the class's local STACK 5169109998Smarkm and counter, and there is now an API function to dynamically create new 5170109998Smarkm classes. This centralisation allows us to (a) plug a lot of the 5171109998Smarkm thread-safety problems that existed, and (b) makes it possible to clean 5172109998Smarkm up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) 5173109998Smarkm such data would previously have always leaked in application code and 5174109998Smarkm workarounds were in place to make the memory debugging turn a blind eye 5175109998Smarkm to it. Application code that doesn't use this new function will still 5176109998Smarkm leak as before, but their memory debugging output will announce it now 5177109998Smarkm rather than letting it slide. 5178109998Smarkm 5179109998Smarkm Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change 5180109998Smarkm induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now 5181109998Smarkm has a return value to indicate success or failure. 5182109998Smarkm [Geoff Thorpe] 5183109998Smarkm 5184109998Smarkm *) Make it possible to replace the underlying "ERR" functions such that the 5185109998Smarkm global state (2 LHASH tables and 2 locks) is only used by the "default" 5186109998Smarkm implementation. This change also adds two functions to "get" and "set" 5187109998Smarkm the implementation prior to it being automatically set the first time 5188109998Smarkm any other ERR function takes place. Ie. an application can call "get", 5189109998Smarkm pass the return value to a module it has just loaded, and that module 5190109998Smarkm can call its own "set" function using that value. This means the 5191109998Smarkm module's "ERR" operations will use (and modify) the error state in the 5192109998Smarkm application and not in its own statically linked copy of OpenSSL code. 5193109998Smarkm [Geoff Thorpe] 5194109998Smarkm 5195109998Smarkm *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment 5196109998Smarkm reference counts. This performs normal REF_PRINT/REF_CHECK macros on 5197109998Smarkm the operation, and provides a more encapsulated way for external code 5198109998Smarkm (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code 5199109998Smarkm to use these functions rather than manually incrementing the counts. 5200109998Smarkm 5201109998Smarkm Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". 5202109998Smarkm [Geoff Thorpe] 5203109998Smarkm 5204109998Smarkm *) Add EVP test program. 5205109998Smarkm [Ben Laurie] 5206109998Smarkm 5207109998Smarkm *) Add symmetric cipher support to ENGINE. Expect the API to change! 5208109998Smarkm [Ben Laurie] 5209109998Smarkm 5210109998Smarkm *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() 5211109998Smarkm X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), 5212109998Smarkm X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). 5213109998Smarkm These allow a CRL to be built without having to access X509_CRL fields 5214109998Smarkm directly. Modify 'ca' application to use new functions. 5215109998Smarkm [Steve Henson] 5216109998Smarkm 5217109998Smarkm *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended 5218109998Smarkm bug workarounds. Rollback attack detection is a security feature. 5219109998Smarkm The problem will only arise on OpenSSL servers when TLSv1 is not 5220109998Smarkm available (sslv3_server_method() or SSL_OP_NO_TLSv1). 5221109998Smarkm Software authors not wanting to support TLSv1 will have special reasons 5222109998Smarkm for their choice and can explicitly enable this option. 5223109998Smarkm [Bodo Moeller, Lutz Jaenicke] 5224109998Smarkm 5225109998Smarkm *) Rationalise EVP so it can be extended: don't include a union of 5226109998Smarkm cipher/digest structures, add init/cleanup functions for EVP_MD_CTX 5227109998Smarkm (similar to those existing for EVP_CIPHER_CTX). 5228109998Smarkm Usage example: 5229109998Smarkm 5230109998Smarkm EVP_MD_CTX md; 5231109998Smarkm 5232109998Smarkm EVP_MD_CTX_init(&md); /* new function call */ 5233109998Smarkm EVP_DigestInit(&md, EVP_sha1()); 5234109998Smarkm EVP_DigestUpdate(&md, in, len); 5235109998Smarkm EVP_DigestFinal(&md, out, NULL); 5236109998Smarkm EVP_MD_CTX_cleanup(&md); /* new function call */ 5237109998Smarkm 5238109998Smarkm [Ben Laurie] 5239109998Smarkm 5240109998Smarkm *) Make DES key schedule conform to the usual scheme, as well as 5241109998Smarkm correcting its structure. This means that calls to DES functions 5242109998Smarkm now have to pass a pointer to a des_key_schedule instead of a 5243109998Smarkm plain des_key_schedule (which was actually always a pointer 5244109998Smarkm anyway): E.g., 5245109998Smarkm 5246109998Smarkm des_key_schedule ks; 5247109998Smarkm 5248109998Smarkm des_set_key_checked(..., &ks); 5249109998Smarkm des_ncbc_encrypt(..., &ks, ...); 5250109998Smarkm 5251109998Smarkm (Note that a later change renames 'des_...' into 'DES_...'.) 5252109998Smarkm [Ben Laurie] 5253109998Smarkm 5254109998Smarkm *) Initial reduction of linker bloat: the use of some functions, such as 5255109998Smarkm PEM causes large amounts of unused functions to be linked in due to 5256109998Smarkm poor organisation. For example pem_all.c contains every PEM function 5257109998Smarkm which has a knock on effect of linking in large amounts of (unused) 5258109998Smarkm ASN1 code. Grouping together similar functions and splitting unrelated 5259109998Smarkm functions prevents this. 5260109998Smarkm [Steve Henson] 5261109998Smarkm 5262109998Smarkm *) Cleanup of EVP macros. 5263109998Smarkm [Ben Laurie] 5264109998Smarkm 5265109998Smarkm *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the 5266109998Smarkm correct _ecb suffix. 5267109998Smarkm [Ben Laurie] 5268109998Smarkm 5269109998Smarkm *) Add initial OCSP responder support to ocsp application. The 5270109998Smarkm revocation information is handled using the text based index 5271109998Smarkm use by the ca application. The responder can either handle 5272109998Smarkm requests generated internally, supplied in files (for example 5273109998Smarkm via a CGI script) or using an internal minimal server. 5274109998Smarkm [Steve Henson] 5275109998Smarkm 5276109998Smarkm *) Add configuration choices to get zlib compression for TLS. 5277109998Smarkm [Richard Levitte] 5278109998Smarkm 5279109998Smarkm *) Changes to Kerberos SSL for RFC 2712 compliance: 5280109998Smarkm 1. Implemented real KerberosWrapper, instead of just using 5281109998Smarkm KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] 5282109998Smarkm 2. Implemented optional authenticator field of KerberosWrapper. 5283109998Smarkm 5284109998Smarkm Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, 5285109998Smarkm and authenticator structs; see crypto/krb5/. 5286109998Smarkm 5287109998Smarkm Generalized Kerberos calls to support multiple Kerberos libraries. 5288109998Smarkm [Vern Staats <staatsvr@asc.hpc.mil>, 5289109998Smarkm Jeffrey Altman <jaltman@columbia.edu> 5290109998Smarkm via Richard Levitte] 5291109998Smarkm 5292109998Smarkm *) Cause 'openssl speed' to use fully hard-coded DSA keys as it 5293109998Smarkm already does with RSA. testdsa.h now has 'priv_key/pub_key' 5294109998Smarkm values for each of the key sizes rather than having just 5295109998Smarkm parameters (and 'speed' generating keys each time). 5296109998Smarkm [Geoff Thorpe] 5297109998Smarkm 5298109998Smarkm *) Speed up EVP routines. 5299109998Smarkm Before: 5300109998Smarkmencrypt 5301109998Smarkmtype 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 5302109998Smarkmdes-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k 5303109998Smarkmdes-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k 5304109998Smarkmdes-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k 5305109998Smarkmdecrypt 5306109998Smarkmdes-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k 5307109998Smarkmdes-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k 5308109998Smarkmdes-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k 5309109998Smarkm After: 5310109998Smarkmencrypt 5311109998Smarkmdes-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k 5312109998Smarkmdecrypt 5313109998Smarkmdes-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k 5314109998Smarkm [Ben Laurie] 5315109998Smarkm 5316109998Smarkm *) Added the OS2-EMX target. 5317109998Smarkm ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] 5318109998Smarkm 5319109998Smarkm *) Rewrite apps to use NCONF routines instead of the old CONF. New functions 5320109998Smarkm to support NCONF routines in extension code. New function CONF_set_nconf() 5321109998Smarkm to allow functions which take an NCONF to also handle the old LHASH 5322109998Smarkm structure: this means that the old CONF compatible routines can be 5323109998Smarkm retained (in particular wrt extensions) without having to duplicate the 5324109998Smarkm code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. 5325109998Smarkm [Steve Henson] 5326109998Smarkm 5327109998Smarkm *) Enhance the general user interface with mechanisms for inner control 5328109998Smarkm and with possibilities to have yes/no kind of prompts. 5329109998Smarkm [Richard Levitte] 5330109998Smarkm 5331109998Smarkm *) Change all calls to low level digest routines in the library and 5332109998Smarkm applications to use EVP. Add missing calls to HMAC_cleanup() and 5333109998Smarkm don't assume HMAC_CTX can be copied using memcpy(). 5334109998Smarkm [Verdon Walker <VWalker@novell.com>, Steve Henson] 5335109998Smarkm 5336109998Smarkm *) Add the possibility to control engines through control names but with 5337109998Smarkm arbitrary arguments instead of just a string. 5338109998Smarkm Change the key loaders to take a UI_METHOD instead of a callback 5339109998Smarkm function pointer. NOTE: this breaks binary compatibility with earlier 5340109998Smarkm versions of OpenSSL [engine]. 5341109998Smarkm Adapt the nCipher code for these new conditions and add a card insertion 5342109998Smarkm callback. 5343109998Smarkm [Richard Levitte] 5344109998Smarkm 5345109998Smarkm *) Enhance the general user interface with mechanisms to better support 5346109998Smarkm dialog box interfaces, application-defined prompts, the possibility 5347109998Smarkm to use defaults (for example default passwords from somewhere else) 5348109998Smarkm and interrupts/cancellations. 5349109998Smarkm [Richard Levitte] 5350109998Smarkm 5351109998Smarkm *) Tidy up PKCS#12 attribute handling. Add support for the CSP name 5352109998Smarkm attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. 5353109998Smarkm [Steve Henson] 5354109998Smarkm 5355109998Smarkm *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also 5356109998Smarkm tidy up some unnecessarily weird code in 'sk_new()'). 5357109998Smarkm [Geoff, reported by Diego Tartara <dtartara@novamens.com>] 5358109998Smarkm 5359109998Smarkm *) Change the key loading routines for ENGINEs to use the same kind 5360109998Smarkm callback (pem_password_cb) as all other routines that need this 5361109998Smarkm kind of callback. 5362109998Smarkm [Richard Levitte] 5363109998Smarkm 5364109998Smarkm *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with 5365109998Smarkm 256 bit (=32 byte) keys. Of course seeding with more entropy bytes 5366109998Smarkm than this minimum value is recommended. 5367109998Smarkm [Lutz Jaenicke] 5368109998Smarkm 5369109998Smarkm *) New random seeder for OpenVMS, using the system process statistics 5370109998Smarkm that are easily reachable. 5371109998Smarkm [Richard Levitte] 5372109998Smarkm 5373109998Smarkm *) Windows apparently can't transparently handle global 5374109998Smarkm variables defined in DLLs. Initialisations such as: 5375109998Smarkm 5376109998Smarkm const ASN1_ITEM *it = &ASN1_INTEGER_it; 5377109998Smarkm 5378109998Smarkm wont compile. This is used by the any applications that need to 5379109998Smarkm declare their own ASN1 modules. This was fixed by adding the option 5380109998Smarkm EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly 5381109998Smarkm needed for static libraries under Win32. 5382109998Smarkm [Steve Henson] 5383109998Smarkm 5384109998Smarkm *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle 5385109998Smarkm setting of purpose and trust fields. New X509_STORE trust and 5386109998Smarkm purpose functions and tidy up setting in other SSL functions. 5387109998Smarkm [Steve Henson] 5388109998Smarkm 5389109998Smarkm *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE 5390109998Smarkm structure. These are inherited by X509_STORE_CTX when it is 5391109998Smarkm initialised. This allows various defaults to be set in the 5392109998Smarkm X509_STORE structure (such as flags for CRL checking and custom 5393109998Smarkm purpose or trust settings) for functions which only use X509_STORE_CTX 5394109998Smarkm internally such as S/MIME. 5395109998Smarkm 5396109998Smarkm Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and 5397109998Smarkm trust settings if they are not set in X509_STORE. This allows X509_STORE 5398109998Smarkm purposes and trust (in S/MIME for example) to override any set by default. 5399109998Smarkm 5400109998Smarkm Add command line options for CRL checking to smime, s_client and s_server 5401109998Smarkm applications. 5402109998Smarkm [Steve Henson] 5403109998Smarkm 5404109998Smarkm *) Initial CRL based revocation checking. If the CRL checking flag(s) 5405109998Smarkm are set then the CRL is looked up in the X509_STORE structure and 5406109998Smarkm its validity and signature checked, then if the certificate is found 5407109998Smarkm in the CRL the verify fails with a revoked error. 5408109998Smarkm 5409109998Smarkm Various new CRL related callbacks added to X509_STORE_CTX structure. 5410109998Smarkm 5411109998Smarkm Command line options added to 'verify' application to support this. 5412109998Smarkm 5413109998Smarkm This needs some additional work, such as being able to handle multiple 5414109998Smarkm CRLs with different times, extension based lookup (rather than just 5415109998Smarkm by subject name) and ultimately more complete V2 CRL extension 5416109998Smarkm handling. 5417109998Smarkm [Steve Henson] 5418109998Smarkm 5419109998Smarkm *) Add a general user interface API (crypto/ui/). This is designed 5420109998Smarkm to replace things like des_read_password and friends (backward 5421109998Smarkm compatibility functions using this new API are provided). 5422109998Smarkm The purpose is to remove prompting functions from the DES code 5423109998Smarkm section as well as provide for prompting through dialog boxes in 5424109998Smarkm a window system and the like. 5425109998Smarkm [Richard Levitte] 5426109998Smarkm 5427109998Smarkm *) Add "ex_data" support to ENGINE so implementations can add state at a 5428109998Smarkm per-structure level rather than having to store it globally. 5429109998Smarkm [Geoff] 5430109998Smarkm 5431109998Smarkm *) Make it possible for ENGINE structures to be copied when retrieved by 5432109998Smarkm ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. 5433109998Smarkm This causes the "original" ENGINE structure to act like a template, 5434109998Smarkm analogous to the RSA vs. RSA_METHOD type of separation. Because of this 5435109998Smarkm operational state can be localised to each ENGINE structure, despite the 5436109998Smarkm fact they all share the same "methods". New ENGINE structures returned in 5437109998Smarkm this case have no functional references and the return value is the single 5438109998Smarkm structural reference. This matches the single structural reference returned 5439109998Smarkm by ENGINE_by_id() normally, when it is incremented on the pre-existing 5440109998Smarkm ENGINE structure. 5441109998Smarkm [Geoff] 5442109998Smarkm 5443109998Smarkm *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this 5444109998Smarkm needs to match any other type at all we need to manually clear the 5445109998Smarkm tag cache. 5446109998Smarkm [Steve Henson] 5447109998Smarkm 5448109998Smarkm *) Changes to the "openssl engine" utility to include; 5449109998Smarkm - verbosity levels ('-v', '-vv', and '-vvv') that provide information 5450109998Smarkm about an ENGINE's available control commands. 5451109998Smarkm - executing control commands from command line arguments using the 5452109998Smarkm '-pre' and '-post' switches. '-post' is only used if '-t' is 5453109998Smarkm specified and the ENGINE is successfully initialised. The syntax for 5454109998Smarkm the individual commands are colon-separated, for example; 5455109998Smarkm openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so 5456109998Smarkm [Geoff] 5457109998Smarkm 5458109998Smarkm *) New dynamic control command support for ENGINEs. ENGINEs can now 5459109998Smarkm declare their own commands (numbers), names (strings), descriptions, 5460109998Smarkm and input types for run-time discovery by calling applications. A 5461109998Smarkm subset of these commands are implicitly classed as "executable" 5462109998Smarkm depending on their input type, and only these can be invoked through 5463109998Smarkm the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this 5464109998Smarkm can be based on user input, config files, etc). The distinction is 5465109998Smarkm that "executable" commands cannot return anything other than a boolean 5466109998Smarkm result and can only support numeric or string input, whereas some 5467109998Smarkm discoverable commands may only be for direct use through 5468109998Smarkm ENGINE_ctrl(), eg. supporting the exchange of binary data, function 5469109998Smarkm pointers, or other custom uses. The "executable" commands are to 5470109998Smarkm support parameterisations of ENGINE behaviour that can be 5471109998Smarkm unambiguously defined by ENGINEs and used consistently across any 5472109998Smarkm OpenSSL-based application. Commands have been added to all the 5473109998Smarkm existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow 5474109998Smarkm control over shared-library paths without source code alterations. 5475109998Smarkm [Geoff] 5476109998Smarkm 5477109998Smarkm *) Changed all ENGINE implementations to dynamically allocate their 5478109998Smarkm ENGINEs rather than declaring them statically. Apart from this being 5479109998Smarkm necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, 5480109998Smarkm this also allows the implementations to compile without using the 5481109998Smarkm internal engine_int.h header. 5482109998Smarkm [Geoff] 5483109998Smarkm 5484109998Smarkm *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a 5485109998Smarkm 'const' value. Any code that should be able to modify a RAND_METHOD 5486109998Smarkm should already have non-const pointers to it (ie. they should only 5487109998Smarkm modify their own ones). 5488109998Smarkm [Geoff] 5489109998Smarkm 5490109998Smarkm *) Made a variety of little tweaks to the ENGINE code. 5491109998Smarkm - "atalla" and "ubsec" string definitions were moved from header files 5492109998Smarkm to C code. "nuron" string definitions were placed in variables 5493109998Smarkm rather than hard-coded - allowing parameterisation of these values 5494109998Smarkm later on via ctrl() commands. 5495109998Smarkm - Removed unused "#if 0"'d code. 5496109998Smarkm - Fixed engine list iteration code so it uses ENGINE_free() to release 5497109998Smarkm structural references. 5498109998Smarkm - Constified the RAND_METHOD element of ENGINE structures. 5499109998Smarkm - Constified various get/set functions as appropriate and added 5500109998Smarkm missing functions (including a catch-all ENGINE_cpy that duplicates 5501109998Smarkm all ENGINE values onto a new ENGINE except reference counts/state). 5502109998Smarkm - Removed NULL parameter checks in get/set functions. Setting a method 5503109998Smarkm or function to NULL is a way of cancelling out a previously set 5504109998Smarkm value. Passing a NULL ENGINE parameter is just plain stupid anyway 5505109998Smarkm and doesn't justify the extra error symbols and code. 5506109998Smarkm - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for 5507109998Smarkm flags from engine_int.h to engine.h. 5508109998Smarkm - Changed prototypes for ENGINE handler functions (init(), finish(), 5509109998Smarkm ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. 5510109998Smarkm [Geoff] 5511109998Smarkm 5512109998Smarkm *) Implement binary inversion algorithm for BN_mod_inverse in addition 5513109998Smarkm to the algorithm using long division. The binary algorithm can be 5514109998Smarkm used only if the modulus is odd. On 32-bit systems, it is faster 5515109998Smarkm only for relatively small moduli (roughly 20-30% for 128-bit moduli, 5516109998Smarkm roughly 5-15% for 256-bit moduli), so we use it only for moduli 5517109998Smarkm up to 450 bits. In 64-bit environments, the binary algorithm 5518109998Smarkm appears to be advantageous for much longer moduli; here we use it 5519109998Smarkm for moduli up to 2048 bits. 5520109998Smarkm [Bodo Moeller] 5521109998Smarkm 5522109998Smarkm *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code 5523109998Smarkm could not support the combine flag in choice fields. 5524109998Smarkm [Steve Henson] 5525109998Smarkm 5526109998Smarkm *) Add a 'copy_extensions' option to the 'ca' utility. This copies 5527109998Smarkm extensions from a certificate request to the certificate. 5528109998Smarkm [Steve Henson] 5529109998Smarkm 5530109998Smarkm *) Allow multiple 'certopt' and 'nameopt' options to be separated 5531109998Smarkm by commas. Add 'namopt' and 'certopt' options to the 'ca' config 5532109998Smarkm file: this allows the display of the certificate about to be 5533109998Smarkm signed to be customised, to allow certain fields to be included 5534109998Smarkm or excluded and extension details. The old system didn't display 5535109998Smarkm multicharacter strings properly, omitted fields not in the policy 5536109998Smarkm and couldn't display additional details such as extensions. 5537109998Smarkm [Steve Henson] 5538109998Smarkm 5539109998Smarkm *) Function EC_POINTs_mul for multiple scalar multiplication 5540109998Smarkm of an arbitrary number of elliptic curve points 5541109998Smarkm \sum scalars[i]*points[i], 5542109998Smarkm optionally including the generator defined for the EC_GROUP: 5543109998Smarkm scalar*generator + \sum scalars[i]*points[i]. 5544109998Smarkm 5545109998Smarkm EC_POINT_mul is a simple wrapper function for the typical case 5546109998Smarkm that the point list has just one item (besides the optional 5547109998Smarkm generator). 5548109998Smarkm [Bodo Moeller] 5549109998Smarkm 5550109998Smarkm *) First EC_METHODs for curves over GF(p): 5551109998Smarkm 5552109998Smarkm EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr 5553109998Smarkm operations and provides various method functions that can also 5554109998Smarkm operate with faster implementations of modular arithmetic. 5555109998Smarkm 5556109998Smarkm EC_GFp_mont_method() reuses most functions that are part of 5557109998Smarkm EC_GFp_simple_method, but uses Montgomery arithmetic. 5558109998Smarkm 5559109998Smarkm [Bodo Moeller; point addition and point doubling 5560109998Smarkm implementation directly derived from source code provided by 5561109998Smarkm Lenka Fibikova <fibikova@exp-math.uni-essen.de>] 5562109998Smarkm 5563109998Smarkm *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, 5564109998Smarkm crypto/ec/ec_lib.c): 5565109998Smarkm 5566109998Smarkm Curves are EC_GROUP objects (with an optional group generator) 5567109998Smarkm based on EC_METHODs that are built into the library. 5568109998Smarkm 5569109998Smarkm Points are EC_POINT objects based on EC_GROUP objects. 5570109998Smarkm 5571109998Smarkm Most of the framework would be able to handle curves over arbitrary 5572109998Smarkm finite fields, but as there are no obvious types for fields other 5573109998Smarkm than GF(p), some functions are limited to that for now. 5574109998Smarkm [Bodo Moeller] 5575109998Smarkm 5576109998Smarkm *) Add the -HTTP option to s_server. It is similar to -WWW, but requires 5577109998Smarkm that the file contains a complete HTTP response. 5578109998Smarkm [Richard Levitte] 5579109998Smarkm 5580109998Smarkm *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl 5581109998Smarkm change the def and num file printf format specifier from "%-40sXXX" 5582109998Smarkm to "%-39s XXX". The latter will always guarantee a space after the 5583109998Smarkm field while the former will cause them to run together if the field 5584109998Smarkm is 40 of more characters long. 5585109998Smarkm [Steve Henson] 5586109998Smarkm 5587109998Smarkm *) Constify the cipher and digest 'method' functions and structures 5588109998Smarkm and modify related functions to take constant EVP_MD and EVP_CIPHER 5589109998Smarkm pointers. 5590109998Smarkm [Steve Henson] 5591109998Smarkm 5592109998Smarkm *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them 5593109998Smarkm in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. 5594109998Smarkm [Bodo Moeller] 5595109998Smarkm 5596109998Smarkm *) Modify EVP_Digest*() routines so they now return values. Although the 5597109998Smarkm internal software routines can never fail additional hardware versions 5598109998Smarkm might. 5599109998Smarkm [Steve Henson] 5600109998Smarkm 5601109998Smarkm *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: 5602109998Smarkm 5603109998Smarkm Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 5604109998Smarkm (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. 5605109998Smarkm 5606109998Smarkm ASN1 error codes 5607109998Smarkm ERR_R_NESTED_ASN1_ERROR 5608109998Smarkm ... 5609109998Smarkm ERR_R_MISSING_ASN1_EOS 5610109998Smarkm were 4 .. 9, conflicting with 5611109998Smarkm ERR_LIB_RSA (= ERR_R_RSA_LIB) 5612109998Smarkm ... 5613109998Smarkm ERR_LIB_PEM (= ERR_R_PEM_LIB). 5614109998Smarkm They are now 58 .. 63 (i.e., just below ERR_R_FATAL). 5615109998Smarkm 5616109998Smarkm Add new error code 'ERR_R_INTERNAL_ERROR'. 5617109998Smarkm [Bodo Moeller] 5618109998Smarkm 5619109998Smarkm *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock 5620109998Smarkm suffices. 5621109998Smarkm [Bodo Moeller] 5622109998Smarkm 5623109998Smarkm *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This 5624109998Smarkm sets the subject name for a new request or supersedes the 5625109998Smarkm subject name in a given request. Formats that can be parsed are 5626109998Smarkm 'CN=Some Name, OU=myOU, C=IT' 5627109998Smarkm and 5628109998Smarkm 'CN=Some Name/OU=myOU/C=IT'. 5629109998Smarkm 5630109998Smarkm Add options '-batch' and '-verbose' to 'openssl req'. 5631109998Smarkm [Massimiliano Pala <madwolf@hackmasters.net>] 5632109998Smarkm 5633109998Smarkm *) Introduce the possibility to access global variables through 5634109998Smarkm functions on platform were that's the best way to handle exporting 5635109998Smarkm global variables in shared libraries. To enable this functionality, 5636109998Smarkm one must configure with "EXPORT_VAR_AS_FN" or defined the C macro 5637109998Smarkm "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter 5638109998Smarkm is normally done by Configure or something similar). 5639109998Smarkm 5640109998Smarkm To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL 5641109998Smarkm in the source file (foo.c) like this: 5642109998Smarkm 5643109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; 5644109998Smarkm OPENSSL_IMPLEMENT_GLOBAL(double,bar); 5645109998Smarkm 5646109998Smarkm To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL 5647109998Smarkm and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: 5648109998Smarkm 5649109998Smarkm OPENSSL_DECLARE_GLOBAL(int,foo); 5650109998Smarkm #define foo OPENSSL_GLOBAL_REF(foo) 5651109998Smarkm OPENSSL_DECLARE_GLOBAL(double,bar); 5652109998Smarkm #define bar OPENSSL_GLOBAL_REF(bar) 5653109998Smarkm 5654109998Smarkm The #defines are very important, and therefore so is including the 5655109998Smarkm header file everywhere where the defined globals are used. 5656109998Smarkm 5657109998Smarkm The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition 5658109998Smarkm of ASN.1 items, but that structure is a bit different. 5659109998Smarkm 5660109998Smarkm The largest change is in util/mkdef.pl which has been enhanced with 5661109998Smarkm better and easier to understand logic to choose which symbols should 5662109998Smarkm go into the Windows .def files as well as a number of fixes and code 5663109998Smarkm cleanup (among others, algorithm keywords are now sorted 5664109998Smarkm lexicographically to avoid constant rewrites). 5665109998Smarkm [Richard Levitte] 5666109998Smarkm 5667109998Smarkm *) In BN_div() keep a copy of the sign of 'num' before writing the 5668109998Smarkm result to 'rm' because if rm==num the value will be overwritten 5669109998Smarkm and produce the wrong result if 'num' is negative: this caused 5670109998Smarkm problems with BN_mod() and BN_nnmod(). 5671109998Smarkm [Steve Henson] 5672109998Smarkm 5673109998Smarkm *) Function OCSP_request_verify(). This checks the signature on an 5674109998Smarkm OCSP request and verifies the signer certificate. The signer 5675109998Smarkm certificate is just checked for a generic purpose and OCSP request 5676109998Smarkm trust settings. 5677109998Smarkm [Steve Henson] 5678109998Smarkm 5679109998Smarkm *) Add OCSP_check_validity() function to check the validity of OCSP 5680109998Smarkm responses. OCSP responses are prepared in real time and may only 5681109998Smarkm be a few seconds old. Simply checking that the current time lies 5682109998Smarkm between thisUpdate and nextUpdate max reject otherwise valid responses 5683109998Smarkm caused by either OCSP responder or client clock inaccuracy. Instead 5684109998Smarkm we allow thisUpdate and nextUpdate to fall within a certain period of 5685109998Smarkm the current time. The age of the response can also optionally be 5686109998Smarkm checked. Two new options -validity_period and -status_age added to 5687109998Smarkm ocsp utility. 5688109998Smarkm [Steve Henson] 5689109998Smarkm 5690109998Smarkm *) If signature or public key algorithm is unrecognized print out its 5691109998Smarkm OID rather that just UNKNOWN. 5692109998Smarkm [Steve Henson] 5693109998Smarkm 5694109998Smarkm *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and 5695109998Smarkm OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate 5696109998Smarkm ID to be generated from the issuer certificate alone which can then be 5697109998Smarkm passed to OCSP_id_issuer_cmp(). 5698109998Smarkm [Steve Henson] 5699109998Smarkm 5700109998Smarkm *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new 5701109998Smarkm ASN1 modules to export functions returning ASN1_ITEM pointers 5702109998Smarkm instead of the ASN1_ITEM structures themselves. This adds several 5703109998Smarkm new macros which allow the underlying ASN1 function/structure to 5704109998Smarkm be accessed transparently. As a result code should not use ASN1_ITEM 5705109998Smarkm references directly (such as &X509_it) but instead use the relevant 5706109998Smarkm macros (such as ASN1_ITEM_rptr(X509)). This option is to allow 5707109998Smarkm use of the new ASN1 code on platforms where exporting structures 5708109998Smarkm is problematical (for example in shared libraries) but exporting 5709109998Smarkm functions returning pointers to structures is not. 5710109998Smarkm [Steve Henson] 5711109998Smarkm 5712109998Smarkm *) Add support for overriding the generation of SSL/TLS session IDs. 5713109998Smarkm These callbacks can be registered either in an SSL_CTX or per SSL. 5714109998Smarkm The purpose of this is to allow applications to control, if they wish, 5715109998Smarkm the arbitrary values chosen for use as session IDs, particularly as it 5716109998Smarkm can be useful for session caching in multiple-server environments. A 5717109998Smarkm command-line switch for testing this (and any client code that wishes 5718109998Smarkm to use such a feature) has been added to "s_server". 5719109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 5720109998Smarkm 5721109998Smarkm *) Modify mkdef.pl to recognise and parse preprocessor conditionals 5722109998Smarkm of the form '#if defined(...) || defined(...) || ...' and 5723109998Smarkm '#if !defined(...) && !defined(...) && ...'. This also avoids 5724109998Smarkm the growing number of special cases it was previously handling. 5725109998Smarkm [Richard Levitte] 5726109998Smarkm 5727109998Smarkm *) Make all configuration macros available for application by making 5728109998Smarkm sure they are available in opensslconf.h, by giving them names starting 5729109998Smarkm with "OPENSSL_" to avoid conflicts with other packages and by making 5730109998Smarkm sure e_os2.h will cover all platform-specific cases together with 5731109998Smarkm opensslconf.h. 5732109998Smarkm Additionally, it is now possible to define configuration/platform- 5733109998Smarkm specific names (called "system identities"). In the C code, these 5734109998Smarkm are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another 5735109998Smarkm macro with the name beginning with "OPENSSL_SYS_", which is determined 5736109998Smarkm from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on 5737109998Smarkm what is available. 5738109998Smarkm [Richard Levitte] 5739109998Smarkm 5740109998Smarkm *) New option -set_serial to 'req' and 'x509' this allows the serial 5741109998Smarkm number to use to be specified on the command line. Previously self 5742109998Smarkm signed certificates were hard coded with serial number 0 and the 5743109998Smarkm CA options of 'x509' had to use a serial number in a file which was 5744109998Smarkm auto incremented. 5745109998Smarkm [Steve Henson] 5746109998Smarkm 5747109998Smarkm *) New options to 'ca' utility to support V2 CRL entry extensions. 5748109998Smarkm Currently CRL reason, invalidity date and hold instruction are 5749109998Smarkm supported. Add new CRL extensions to V3 code and some new objects. 5750109998Smarkm [Steve Henson] 5751109998Smarkm 5752109998Smarkm *) New function EVP_CIPHER_CTX_set_padding() this is used to 5753109998Smarkm disable standard block padding (aka PKCS#5 padding) in the EVP 5754109998Smarkm API, which was previously mandatory. This means that the data is 5755109998Smarkm not padded in any way and so the total length much be a multiple 5756109998Smarkm of the block size, otherwise an error occurs. 5757109998Smarkm [Steve Henson] 5758109998Smarkm 5759109998Smarkm *) Initial (incomplete) OCSP SSL support. 5760109998Smarkm [Steve Henson] 5761109998Smarkm 5762109998Smarkm *) New function OCSP_parse_url(). This splits up a URL into its host, 5763109998Smarkm port and path components: primarily to parse OCSP URLs. New -url 5764109998Smarkm option to ocsp utility. 5765109998Smarkm [Steve Henson] 5766109998Smarkm 5767109998Smarkm *) New nonce behavior. The return value of OCSP_check_nonce() now 5768109998Smarkm reflects the various checks performed. Applications can decide 5769109998Smarkm whether to tolerate certain situations such as an absent nonce 5770109998Smarkm in a response when one was present in a request: the ocsp application 5771109998Smarkm just prints out a warning. New function OCSP_add1_basic_nonce() 5772109998Smarkm this is to allow responders to include a nonce in a response even if 5773109998Smarkm the request is nonce-less. 5774109998Smarkm [Steve Henson] 5775109998Smarkm 5776109998Smarkm *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are 5777109998Smarkm skipped when using openssl x509 multiple times on a single input file, 5778109998Smarkm e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". 5779109998Smarkm [Bodo Moeller] 5780109998Smarkm 5781109998Smarkm *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() 5782109998Smarkm set string type: to handle setting ASN1_TIME structures. Fix ca 5783109998Smarkm utility to correctly initialize revocation date of CRLs. 5784109998Smarkm [Steve Henson] 5785109998Smarkm 5786109998Smarkm *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override 5787109998Smarkm the clients preferred ciphersuites and rather use its own preferences. 5788109998Smarkm Should help to work around M$ SGC (Server Gated Cryptography) bug in 5789109998Smarkm Internet Explorer by ensuring unchanged hash method during stepup. 5790109998Smarkm (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) 5791109998Smarkm [Lutz Jaenicke] 5792109998Smarkm 5793109998Smarkm *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael 5794109998Smarkm to aes and add a new 'exist' option to print out symbols that don't 5795109998Smarkm appear to exist. 5796109998Smarkm [Steve Henson] 5797109998Smarkm 5798109998Smarkm *) Additional options to ocsp utility to allow flags to be set and 5799109998Smarkm additional certificates supplied. 5800109998Smarkm [Steve Henson] 5801109998Smarkm 5802109998Smarkm *) Add the option -VAfile to 'openssl ocsp', so the user can give the 5803109998Smarkm OCSP client a number of certificate to only verify the response 5804109998Smarkm signature against. 5805109998Smarkm [Richard Levitte] 5806109998Smarkm 5807109998Smarkm *) Update Rijndael code to version 3.0 and change EVP AES ciphers to 5808109998Smarkm handle the new API. Currently only ECB, CBC modes supported. Add new 5809109998Smarkm AES OIDs. 5810109998Smarkm 5811109998Smarkm Add TLS AES ciphersuites as described in RFC3268, "Advanced 5812109998Smarkm Encryption Standard (AES) Ciphersuites for Transport Layer 5813109998Smarkm Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were 5814109998Smarkm not enabled by default and were not part of the "ALL" ciphersuite 5815109998Smarkm alias because they were not yet official; they could be 5816109998Smarkm explicitly requested by specifying the "AESdraft" ciphersuite 5817109998Smarkm group alias. In the final release of OpenSSL 0.9.7, the group 5818109998Smarkm alias is called "AES" and is part of "ALL".) 5819109998Smarkm [Ben Laurie, Steve Henson, Bodo Moeller] 5820109998Smarkm 5821109998Smarkm *) New function OCSP_copy_nonce() to copy nonce value (if present) from 5822109998Smarkm request to response. 5823109998Smarkm [Steve Henson] 5824109998Smarkm 5825109998Smarkm *) Functions for OCSP responders. OCSP_request_onereq_count(), 5826109998Smarkm OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() 5827109998Smarkm extract information from a certificate request. OCSP_response_create() 5828109998Smarkm creates a response and optionally adds a basic response structure. 5829109998Smarkm OCSP_basic_add1_status() adds a complete single response to a basic 5830109998Smarkm response and returns the OCSP_SINGLERESP structure just added (to allow 5831109998Smarkm extensions to be included for example). OCSP_basic_add1_cert() adds a 5832109998Smarkm certificate to a basic response and OCSP_basic_sign() signs a basic 5833109998Smarkm response with various flags. New helper functions ASN1_TIME_check() 5834109998Smarkm (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() 5835109998Smarkm (converts ASN1_TIME to GeneralizedTime). 5836109998Smarkm [Steve Henson] 5837109998Smarkm 5838109998Smarkm *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() 5839109998Smarkm in a single operation. X509_get0_pubkey_bitstr() extracts the public_key 5840109998Smarkm structure from a certificate. X509_pubkey_digest() digests the public_key 5841109998Smarkm contents: this is used in various key identifiers. 5842109998Smarkm [Steve Henson] 5843109998Smarkm 5844109998Smarkm *) Make sk_sort() tolerate a NULL argument. 5845109998Smarkm [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] 5846109998Smarkm 5847109998Smarkm *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates 5848109998Smarkm passed by the function are trusted implicitly. If any of them signed the 5849109998Smarkm response then it is assumed to be valid and is not verified. 5850109998Smarkm [Steve Henson] 5851109998Smarkm 5852109998Smarkm *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT 5853109998Smarkm to data. This was previously part of the PKCS7 ASN1 code. This 5854109998Smarkm was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. 5855109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5856109998Smarkm <support@securenetterm.com>] 5857109998Smarkm 5858109998Smarkm *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 5859109998Smarkm routines: without these tracing memory leaks is very painful. 5860109998Smarkm Fix leaks in PKCS12 and PKCS7 routines. 5861109998Smarkm [Steve Henson] 5862109998Smarkm 5863109998Smarkm *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). 5864109998Smarkm Previously it initialised the 'type' argument to V_ASN1_UTCTIME which 5865109998Smarkm effectively meant GeneralizedTime would never be used. Now it 5866109998Smarkm is initialised to -1 but X509_time_adj() now has to check the value 5867109998Smarkm and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or 5868109998Smarkm V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. 5869109998Smarkm [Steve Henson, reported by Kenneth R. Robinette 5870109998Smarkm <support@securenetterm.com>] 5871109998Smarkm 5872109998Smarkm *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously 5873109998Smarkm result in a zero length in the ASN1_INTEGER structure which was 5874109998Smarkm not consistent with the structure when d2i_ASN1_INTEGER() was used 5875109998Smarkm and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() 5876109998Smarkm to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() 5877109998Smarkm where it did not print out a minus for negative ASN1_INTEGER. 5878109998Smarkm [Steve Henson] 5879109998Smarkm 5880109998Smarkm *) Add summary printout to ocsp utility. The various functions which 5881109998Smarkm convert status values to strings have been renamed to: 5882109998Smarkm OCSP_response_status_str(), OCSP_cert_status_str() and 5883109998Smarkm OCSP_crl_reason_str() and are no longer static. New options 5884109998Smarkm to verify nonce values and to disable verification. OCSP response 5885109998Smarkm printout format cleaned up. 5886109998Smarkm [Steve Henson] 5887109998Smarkm 5888109998Smarkm *) Add additional OCSP certificate checks. These are those specified 5889109998Smarkm in RFC2560. This consists of two separate checks: the CA of the 5890109998Smarkm certificate being checked must either be the OCSP signer certificate 5891109998Smarkm or the issuer of the OCSP signer certificate. In the latter case the 5892109998Smarkm OCSP signer certificate must contain the OCSP signing extended key 5893109998Smarkm usage. This check is performed by attempting to match the OCSP 5894109998Smarkm signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash 5895109998Smarkm in the OCSP_CERTID structures of the response. 5896109998Smarkm [Steve Henson] 5897109998Smarkm 5898109998Smarkm *) Initial OCSP certificate verification added to OCSP_basic_verify() 5899109998Smarkm and related routines. This uses the standard OpenSSL certificate 5900109998Smarkm verify routines to perform initial checks (just CA validity) and 5901109998Smarkm to obtain the certificate chain. Then additional checks will be 5902109998Smarkm performed on the chain. Currently the root CA is checked to see 5903109998Smarkm if it is explicitly trusted for OCSP signing. This is used to set 5904109998Smarkm a root CA as a global signing root: that is any certificate that 5905109998Smarkm chains to that CA is an acceptable OCSP signing certificate. 5906109998Smarkm [Steve Henson] 5907109998Smarkm 5908109998Smarkm *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 5909109998Smarkm extensions from a separate configuration file. 5910109998Smarkm As when reading extensions from the main configuration file, 5911109998Smarkm the '-extensions ...' option may be used for specifying the 5912109998Smarkm section to use. 5913109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5914109998Smarkm 5915109998Smarkm *) New OCSP utility. Allows OCSP requests to be generated or 5916109998Smarkm read. The request can be sent to a responder and the output 5917109998Smarkm parsed, outputed or printed in text form. Not complete yet: 5918109998Smarkm still needs to check the OCSP response validity. 5919109998Smarkm [Steve Henson] 5920109998Smarkm 5921109998Smarkm *) New subcommands for 'openssl ca': 5922109998Smarkm 'openssl ca -status <serial>' prints the status of the cert with 5923109998Smarkm the given serial number (according to the index file). 5924109998Smarkm 'openssl ca -updatedb' updates the expiry status of certificates 5925109998Smarkm in the index file. 5926109998Smarkm [Massimiliano Pala <madwolf@comune.modena.it>] 5927109998Smarkm 5928109998Smarkm *) New '-newreq-nodes' command option to CA.pl. This is like 5929109998Smarkm '-newreq', but calls 'openssl req' with the '-nodes' option 5930109998Smarkm so that the resulting key is not encrypted. 5931109998Smarkm [Damien Miller <djm@mindrot.org>] 5932109998Smarkm 5933109998Smarkm *) New configuration for the GNU Hurd. 5934109998Smarkm [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] 5935109998Smarkm 5936109998Smarkm *) Initial code to implement OCSP basic response verify. This 5937109998Smarkm is currently incomplete. Currently just finds the signer's 5938109998Smarkm certificate and verifies the signature on the response. 5939109998Smarkm [Steve Henson] 5940109998Smarkm 5941109998Smarkm *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in 5942109998Smarkm value of OPENSSLDIR. This is available via the new '-d' option 5943109998Smarkm to 'openssl version', and is also included in 'openssl version -a'. 5944109998Smarkm [Bodo Moeller] 5945109998Smarkm 5946109998Smarkm *) Allowing defining memory allocation callbacks that will be given 5947109998Smarkm file name and line number information in additional arguments 5948109998Smarkm (a const char* and an int). The basic functionality remains, as 5949109998Smarkm well as the original possibility to just replace malloc(), 5950109998Smarkm realloc() and free() by functions that do not know about these 5951109998Smarkm additional arguments. To register and find out the current 5952109998Smarkm settings for extended allocation functions, the following 5953109998Smarkm functions are provided: 5954109998Smarkm 5955109998Smarkm CRYPTO_set_mem_ex_functions 5956109998Smarkm CRYPTO_set_locked_mem_ex_functions 5957109998Smarkm CRYPTO_get_mem_ex_functions 5958109998Smarkm CRYPTO_get_locked_mem_ex_functions 5959109998Smarkm 5960109998Smarkm These work the same way as CRYPTO_set_mem_functions and friends. 5961109998Smarkm CRYPTO_get_[locked_]mem_functions now writes 0 where such an 5962109998Smarkm extended allocation function is enabled. 5963109998Smarkm Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where 5964109998Smarkm a conventional allocation function is enabled. 5965109998Smarkm [Richard Levitte, Bodo Moeller] 5966109998Smarkm 5967109998Smarkm *) Finish off removing the remaining LHASH function pointer casts. 5968109998Smarkm There should no longer be any prototype-casting required when using 5969109998Smarkm the LHASH abstraction, and any casts that remain are "bugs". See 5970109998Smarkm the callback types and macros at the head of lhash.h for details 5971109998Smarkm (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). 5972109998Smarkm [Geoff Thorpe] 5973109998Smarkm 5974109998Smarkm *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. 5975109998Smarkm If /dev/[u]random devices are not available or do not return enough 5976109998Smarkm entropy, EGD style sockets (served by EGD or PRNGD) will automatically 5977109998Smarkm be queried. 5978109998Smarkm The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and 5979109998Smarkm /etc/entropy will be queried once each in this sequence, quering stops 5980109998Smarkm when enough entropy was collected without querying more sockets. 5981109998Smarkm [Lutz Jaenicke] 5982109998Smarkm 5983109998Smarkm *) Change the Unix RAND_poll() variant to be able to poll several 5984109998Smarkm random devices, as specified by DEVRANDOM, until a sufficient amount 5985109998Smarkm of data has been collected. We spend at most 10 ms on each file 5986109998Smarkm (select timeout) and read in non-blocking mode. DEVRANDOM now 5987109998Smarkm defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" 5988109998Smarkm (previously it was just the string "/dev/urandom"), so on typical 5989109998Smarkm platforms the 10 ms delay will never occur. 5990109998Smarkm Also separate out the Unix variant to its own file, rand_unix.c. 5991109998Smarkm For VMS, there's a currently-empty rand_vms.c. 5992109998Smarkm [Richard Levitte] 5993109998Smarkm 5994109998Smarkm *) Move OCSP client related routines to ocsp_cl.c. These 5995109998Smarkm provide utility functions which an application needing 5996109998Smarkm to issue a request to an OCSP responder and analyse the 5997109998Smarkm response will typically need: as opposed to those which an 5998109998Smarkm OCSP responder itself would need which will be added later. 5999109998Smarkm 6000109998Smarkm OCSP_request_sign() signs an OCSP request with an API similar 6001109998Smarkm to PKCS7_sign(). OCSP_response_status() returns status of OCSP 6002109998Smarkm response. OCSP_response_get1_basic() extracts basic response 6003109998Smarkm from response. OCSP_resp_find_status(): finds and extracts status 6004109998Smarkm information from an OCSP_CERTID structure (which will be created 6005109998Smarkm when the request structure is built). These are built from lower 6006109998Smarkm level functions which work on OCSP_SINGLERESP structures but 6007109998Smarkm wont normally be used unless the application wishes to examine 6008109998Smarkm extensions in the OCSP response for example. 6009109998Smarkm 6010109998Smarkm Replace nonce routines with a pair of functions. 6011109998Smarkm OCSP_request_add1_nonce() adds a nonce value and optionally 6012109998Smarkm generates a random value. OCSP_check_nonce() checks the 6013109998Smarkm validity of the nonce in an OCSP response. 6014109998Smarkm [Steve Henson] 6015109998Smarkm 6016109998Smarkm *) Change function OCSP_request_add() to OCSP_request_add0_id(). 6017109998Smarkm This doesn't copy the supplied OCSP_CERTID and avoids the 6018109998Smarkm need to free up the newly created id. Change return type 6019109998Smarkm to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. 6020109998Smarkm This can then be used to add extensions to the request. 6021109998Smarkm Deleted OCSP_request_new(), since most of its functionality 6022109998Smarkm is now in OCSP_REQUEST_new() (and the case insensitive name 6023109998Smarkm clash) apart from the ability to set the request name which 6024109998Smarkm will be added elsewhere. 6025109998Smarkm [Steve Henson] 6026109998Smarkm 6027109998Smarkm *) Update OCSP API. Remove obsolete extensions argument from 6028109998Smarkm various functions. Extensions are now handled using the new 6029109998Smarkm OCSP extension code. New simple OCSP HTTP function which 6030109998Smarkm can be used to send requests and parse the response. 6031109998Smarkm [Steve Henson] 6032109998Smarkm 6033109998Smarkm *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new 6034109998Smarkm ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN 6035109998Smarkm uses the special reorder version of SET OF to sort the attributes 6036109998Smarkm and reorder them to match the encoded order. This resolves a long 6037109998Smarkm standing problem: a verify on a PKCS7 structure just after signing 6038109998Smarkm it used to fail because the attribute order did not match the 6039109998Smarkm encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: 6040109998Smarkm it uses the received order. This is necessary to tolerate some broken 6041109998Smarkm software that does not order SET OF. This is handled by encoding 6042109998Smarkm as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) 6043109998Smarkm to produce the required SET OF. 6044109998Smarkm [Steve Henson] 6045109998Smarkm 6046109998Smarkm *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and 6047109998Smarkm OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header 6048109998Smarkm files to get correct declarations of the ASN.1 item variables. 6049109998Smarkm [Richard Levitte] 6050109998Smarkm 6051109998Smarkm *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many 6052109998Smarkm PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: 6053109998Smarkm asn1_check_tlen() would sometimes attempt to use 'ctx' when it was 6054109998Smarkm NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). 6055109998Smarkm New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant 6056109998Smarkm ASN1_ITEM and no wrapper functions. 6057109998Smarkm [Steve Henson] 6058109998Smarkm 6059109998Smarkm *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These 6060109998Smarkm replace the old function pointer based I/O routines. Change most of 6061109998Smarkm the *_d2i_bio() and *_d2i_fp() functions to use these. 6062109998Smarkm [Steve Henson] 6063109998Smarkm 6064109998Smarkm *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor 6065109998Smarkm lines, recognice more "algorithms" that can be deselected, and make 6066109998Smarkm it complain about algorithm deselection that isn't recognised. 6067109998Smarkm [Richard Levitte] 6068109998Smarkm 6069109998Smarkm *) New ASN1 functions to handle dup, sign, verify, digest, pack and 6070109998Smarkm unpack operations in terms of ASN1_ITEM. Modify existing wrappers 6071109998Smarkm to use new functions. Add NO_ASN1_OLD which can be set to remove 6072109998Smarkm some old style ASN1 functions: this can be used to determine if old 6073109998Smarkm code will still work when these eventually go away. 6074109998Smarkm [Steve Henson] 6075109998Smarkm 6076109998Smarkm *) New extension functions for OCSP structures, these follow the 6077109998Smarkm same conventions as certificates and CRLs. 6078109998Smarkm [Steve Henson] 6079109998Smarkm 6080109998Smarkm *) New function X509V3_add1_i2d(). This automatically encodes and 6081109998Smarkm adds an extension. Its behaviour can be customised with various 6082109998Smarkm flags to append, replace or delete. Various wrappers added for 6083109998Smarkm certifcates and CRLs. 6084109998Smarkm [Steve Henson] 6085109998Smarkm 6086109998Smarkm *) Fix to avoid calling the underlying ASN1 print routine when 6087109998Smarkm an extension cannot be parsed. Correct a typo in the 6088109998Smarkm OCSP_SERVICELOC extension. Tidy up print OCSP format. 6089109998Smarkm [Steve Henson] 6090109998Smarkm 6091109998Smarkm *) Make mkdef.pl parse some of the ASN1 macros and add apropriate 6092109998Smarkm entries for variables. 6093109998Smarkm [Steve Henson] 6094109998Smarkm 6095109998Smarkm *) Add functionality to apps/openssl.c for detecting locking 6096109998Smarkm problems: As the program is single-threaded, all we have 6097109998Smarkm to do is register a locking callback using an array for 6098109998Smarkm storing which locks are currently held by the program. 6099109998Smarkm [Bodo Moeller] 6100109998Smarkm 6101109998Smarkm *) Use a lock around the call to CRYPTO_get_ex_new_index() in 6102109998Smarkm SSL_get_ex_data_X509_STORE_idx(), which is used in 6103109998Smarkm ssl_verify_cert_chain() and thus can be called at any time 6104109998Smarkm during TLS/SSL handshakes so that thread-safety is essential. 6105109998Smarkm Unfortunately, the ex_data design is not at all suited 6106109998Smarkm for multi-threaded use, so it probably should be abolished. 6107109998Smarkm [Bodo Moeller] 6108109998Smarkm 6109109998Smarkm *) Added Broadcom "ubsec" ENGINE to OpenSSL. 6110109998Smarkm [Broadcom, tweaked and integrated by Geoff Thorpe] 6111109998Smarkm 6112109998Smarkm *) Move common extension printing code to new function 6113109998Smarkm X509V3_print_extensions(). Reorganise OCSP print routines and 6114109998Smarkm implement some needed OCSP ASN1 functions. Add OCSP extensions. 6115109998Smarkm [Steve Henson] 6116109998Smarkm 6117109998Smarkm *) New function X509_signature_print() to remove duplication in some 6118109998Smarkm print routines. 6119109998Smarkm [Steve Henson] 6120109998Smarkm 6121109998Smarkm *) Add a special meaning when SET OF and SEQUENCE OF flags are both 6122109998Smarkm set (this was treated exactly the same as SET OF previously). This 6123109998Smarkm is used to reorder the STACK representing the structure to match the 6124109998Smarkm encoding. This will be used to get round a problem where a PKCS7 6125109998Smarkm structure which was signed could not be verified because the STACK 6126109998Smarkm order did not reflect the encoded order. 6127109998Smarkm [Steve Henson] 6128109998Smarkm 6129109998Smarkm *) Reimplement the OCSP ASN1 module using the new code. 6130109998Smarkm [Steve Henson] 6131109998Smarkm 6132109998Smarkm *) Update the X509V3 code to permit the use of an ASN1_ITEM structure 6133109998Smarkm for its ASN1 operations. The old style function pointers still exist 6134109998Smarkm for now but they will eventually go away. 6135109998Smarkm [Steve Henson] 6136109998Smarkm 6137109998Smarkm *) Merge in replacement ASN1 code from the ASN1 branch. This almost 6138109998Smarkm completely replaces the old ASN1 functionality with a table driven 6139109998Smarkm encoder and decoder which interprets an ASN1_ITEM structure describing 6140109998Smarkm the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is 6141109998Smarkm largely maintained. Almost all of the old asn1_mac.h macro based ASN1 6142109998Smarkm has also been converted to the new form. 6143109998Smarkm [Steve Henson] 6144109998Smarkm 6145109998Smarkm *) Change BN_mod_exp_recp so that negative moduli are tolerated 6146109998Smarkm (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set 6147109998Smarkm so that BN_mod_exp_mont and BN_mod_exp_mont_word work 6148109998Smarkm for negative moduli. 6149109998Smarkm [Bodo Moeller] 6150109998Smarkm 6151109998Smarkm *) Fix BN_uadd and BN_usub: Always return non-negative results instead 6152109998Smarkm of not touching the result's sign bit. 6153109998Smarkm [Bodo Moeller] 6154109998Smarkm 6155109998Smarkm *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be 6156109998Smarkm set. 6157109998Smarkm [Bodo Moeller] 6158109998Smarkm 6159109998Smarkm *) Changed the LHASH code to use prototypes for callbacks, and created 6160109998Smarkm macros to declare and implement thin (optionally static) functions 6161109998Smarkm that provide type-safety and avoid function pointer casting for the 6162109998Smarkm type-specific callbacks. 6163109998Smarkm [Geoff Thorpe] 6164109998Smarkm 6165109998Smarkm *) Added Kerberos Cipher Suites to be used with TLS, as written in 6166109998Smarkm RFC 2712. 6167109998Smarkm [Veers Staats <staatsvr@asc.hpc.mil>, 6168109998Smarkm Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] 6169109998Smarkm 6170109998Smarkm *) Reformat the FAQ so the different questions and answers can be divided 6171109998Smarkm in sections depending on the subject. 6172109998Smarkm [Richard Levitte] 6173109998Smarkm 6174109998Smarkm *) Have the zlib compression code load ZLIB.DLL dynamically under 6175109998Smarkm Windows. 6176109998Smarkm [Richard Levitte] 6177109998Smarkm 6178109998Smarkm *) New function BN_mod_sqrt for computing square roots modulo a prime 6179109998Smarkm (using the probabilistic Tonelli-Shanks algorithm unless 6180109998Smarkm p == 3 (mod 4) or p == 5 (mod 8), which are cases that can 6181109998Smarkm be handled deterministically). 6182109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 6183109998Smarkm 6184109998Smarkm *) Make BN_mod_inverse faster by explicitly handling small quotients 6185109998Smarkm in the Euclid loop. (Speed gain about 20% for small moduli [256 or 6186109998Smarkm 512 bits], about 30% for larger ones [1024 or 2048 bits].) 6187109998Smarkm [Bodo Moeller] 6188109998Smarkm 6189109998Smarkm *) New function BN_kronecker. 6190109998Smarkm [Bodo Moeller] 6191109998Smarkm 6192109998Smarkm *) Fix BN_gcd so that it works on negative inputs; the result is 6193109998Smarkm positive unless both parameters are zero. 6194109998Smarkm Previously something reasonably close to an infinite loop was 6195109998Smarkm possible because numbers could be growing instead of shrinking 6196109998Smarkm in the implementation of Euclid's algorithm. 6197109998Smarkm [Bodo Moeller] 6198109998Smarkm 6199109998Smarkm *) Fix BN_is_word() and BN_is_one() macros to take into account the 6200109998Smarkm sign of the number in question. 6201109998Smarkm 6202109998Smarkm Fix BN_is_word(a,w) to work correctly for w == 0. 6203109998Smarkm 6204109998Smarkm The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) 6205109998Smarkm because its test if the absolute value of 'a' equals 'w'. 6206109998Smarkm Note that BN_abs_is_word does *not* handle w == 0 reliably; 6207109998Smarkm it exists mostly for use in the implementations of BN_is_zero(), 6208109998Smarkm BN_is_one(), and BN_is_word(). 6209109998Smarkm [Bodo Moeller] 6210109998Smarkm 6211109998Smarkm *) New function BN_swap. 6212109998Smarkm [Bodo Moeller] 6213109998Smarkm 6214109998Smarkm *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that 6215109998Smarkm the exponentiation functions are more likely to produce reasonable 6216109998Smarkm results on negative inputs. 6217109998Smarkm [Bodo Moeller] 6218109998Smarkm 6219109998Smarkm *) Change BN_mod_mul so that the result is always non-negative. 6220109998Smarkm Previously, it could be negative if one of the factors was negative; 6221109998Smarkm I don't think anyone really wanted that behaviour. 6222109998Smarkm [Bodo Moeller] 6223109998Smarkm 6224109998Smarkm *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c 6225109998Smarkm (except for exponentiation, which stays in crypto/bn/bn_exp.c, 6226109998Smarkm and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) 6227109998Smarkm and add new functions: 6228109998Smarkm 6229109998Smarkm BN_nnmod 6230109998Smarkm BN_mod_sqr 6231109998Smarkm BN_mod_add 6232109998Smarkm BN_mod_add_quick 6233109998Smarkm BN_mod_sub 6234109998Smarkm BN_mod_sub_quick 6235109998Smarkm BN_mod_lshift1 6236109998Smarkm BN_mod_lshift1_quick 6237109998Smarkm BN_mod_lshift 6238109998Smarkm BN_mod_lshift_quick 6239109998Smarkm 6240109998Smarkm These functions always generate non-negative results. 6241109998Smarkm 6242109998Smarkm BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r 6243109998Smarkm such that |m| < r < 0, BN_nnmod will output rem + |m| instead). 6244109998Smarkm 6245109998Smarkm BN_mod_XXX_quick(r, a, [b,] m) generates the same result as 6246109998Smarkm BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] 6247109998Smarkm be reduced modulo m. 6248109998Smarkm [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] 6249109998Smarkm 6250111147Snectar#if 0 6251111147Snectar The following entry accidentily appeared in the CHANGES file 6252111147Snectar distributed with OpenSSL 0.9.7. The modifications described in 6253111147Snectar it do *not* apply to OpenSSL 0.9.7. 6254111147Snectar 6255109998Smarkm *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there 6256109998Smarkm was actually never needed) and in BN_mul(). The removal in BN_mul() 6257109998Smarkm required a small change in bn_mul_part_recursive() and the addition 6258109998Smarkm of the functions bn_cmp_part_words(), bn_sub_part_words() and 6259109998Smarkm bn_add_part_words(), which do the same thing as bn_cmp_words(), 6260109998Smarkm bn_sub_words() and bn_add_words() except they take arrays with 6261109998Smarkm differing sizes. 6262109998Smarkm [Richard Levitte] 6263111147Snectar#endif 6264109998Smarkm 6265109998Smarkm *) In 'openssl passwd', verify passwords read from the terminal 6266109998Smarkm unless the '-salt' option is used (which usually means that 6267109998Smarkm verification would just waste user's time since the resulting 6268109998Smarkm hash is going to be compared with some given password hash) 6269109998Smarkm or the new '-noverify' option is used. 6270109998Smarkm 6271109998Smarkm This is an incompatible change, but it does not affect 6272109998Smarkm non-interactive use of 'openssl passwd' (passwords on the command 6273109998Smarkm line, '-stdin' option, '-in ...' option) and thus should not 6274109998Smarkm cause any problems. 6275109998Smarkm [Bodo Moeller] 6276109998Smarkm 6277109998Smarkm *) Remove all references to RSAref, since there's no more need for it. 6278109998Smarkm [Richard Levitte] 6279109998Smarkm 6280109998Smarkm *) Make DSO load along a path given through an environment variable 6281109998Smarkm (SHLIB_PATH) with shl_load(). 6282109998Smarkm [Richard Levitte] 6283109998Smarkm 6284109998Smarkm *) Constify the ENGINE code as a result of BIGNUM constification. 6285109998Smarkm Also constify the RSA code and most things related to it. In a 6286109998Smarkm few places, most notable in the depth of the ASN.1 code, ugly 6287109998Smarkm casts back to non-const were required (to be solved at a later 6288109998Smarkm time) 6289109998Smarkm [Richard Levitte] 6290109998Smarkm 6291109998Smarkm *) Make it so the openssl application has all engines loaded by default. 6292109998Smarkm [Richard Levitte] 6293109998Smarkm 6294109998Smarkm *) Constify the BIGNUM routines a little more. 6295109998Smarkm [Richard Levitte] 6296109998Smarkm 6297109998Smarkm *) Add the following functions: 6298109998Smarkm 6299109998Smarkm ENGINE_load_cswift() 6300109998Smarkm ENGINE_load_chil() 6301109998Smarkm ENGINE_load_atalla() 6302109998Smarkm ENGINE_load_nuron() 6303109998Smarkm ENGINE_load_builtin_engines() 6304109998Smarkm 6305109998Smarkm That way, an application can itself choose if external engines that 6306109998Smarkm are built-in in OpenSSL shall ever be used or not. The benefit is 6307109998Smarkm that applications won't have to be linked with libdl or other dso 6308109998Smarkm libraries unless it's really needed. 6309109998Smarkm 6310109998Smarkm Changed 'openssl engine' to load all engines on demand. 6311109998Smarkm Changed the engine header files to avoid the duplication of some 6312109998Smarkm declarations (they differed!). 6313109998Smarkm [Richard Levitte] 6314109998Smarkm 6315109998Smarkm *) 'openssl engine' can now list capabilities. 6316109998Smarkm [Richard Levitte] 6317109998Smarkm 6318109998Smarkm *) Better error reporting in 'openssl engine'. 6319109998Smarkm [Richard Levitte] 6320109998Smarkm 6321109998Smarkm *) Never call load_dh_param(NULL) in s_server. 6322109998Smarkm [Bodo Moeller] 6323109998Smarkm 6324109998Smarkm *) Add engine application. It can currently list engines by name and 6325109998Smarkm identity, and test if they are actually available. 6326109998Smarkm [Richard Levitte] 6327109998Smarkm 6328109998Smarkm *) Improve RPM specification file by forcing symbolic linking and making 6329109998Smarkm sure the installed documentation is also owned by root.root. 6330109998Smarkm [Damien Miller <djm@mindrot.org>] 6331109998Smarkm 6332109998Smarkm *) Give the OpenSSL applications more possibilities to make use of 6333109998Smarkm keys (public as well as private) handled by engines. 6334109998Smarkm [Richard Levitte] 6335109998Smarkm 6336109998Smarkm *) Add OCSP code that comes from CertCo. 6337109998Smarkm [Richard Levitte] 6338109998Smarkm 6339109998Smarkm *) Add VMS support for the Rijndael code. 6340109998Smarkm [Richard Levitte] 6341109998Smarkm 6342109998Smarkm *) Added untested support for Nuron crypto accelerator. 6343109998Smarkm [Ben Laurie] 6344109998Smarkm 6345109998Smarkm *) Add support for external cryptographic devices. This code was 6346109998Smarkm previously distributed separately as the "engine" branch. 6347109998Smarkm [Geoff Thorpe, Richard Levitte] 6348109998Smarkm 6349109998Smarkm *) Rework the filename-translation in the DSO code. It is now possible to 6350109998Smarkm have far greater control over how a "name" is turned into a filename 6351109998Smarkm depending on the operating environment and any oddities about the 6352109998Smarkm different shared library filenames on each system. 6353109998Smarkm [Geoff Thorpe] 6354109998Smarkm 6355109998Smarkm *) Support threads on FreeBSD-elf in Configure. 6356109998Smarkm [Richard Levitte] 6357109998Smarkm 6358109998Smarkm *) Fix for SHA1 assembly problem with MASM: it produces 6359109998Smarkm warnings about corrupt line number information when assembling 6360109998Smarkm with debugging information. This is caused by the overlapping 6361109998Smarkm of two sections. 6362109998Smarkm [Bernd Matthes <mainbug@celocom.de>, Steve Henson] 6363109998Smarkm 6364109998Smarkm *) NCONF changes. 6365109998Smarkm NCONF_get_number() has no error checking at all. As a replacement, 6366109998Smarkm NCONF_get_number_e() is defined (_e for "error checking") and is 6367109998Smarkm promoted strongly. The old NCONF_get_number is kept around for 6368109998Smarkm binary backward compatibility. 6369109998Smarkm Make it possible for methods to load from something other than a BIO, 6370109998Smarkm by providing a function pointer that is given a name instead of a BIO. 6371109998Smarkm For example, this could be used to load configuration data from an 6372109998Smarkm LDAP server. 6373109998Smarkm [Richard Levitte] 6374109998Smarkm 6375109998Smarkm *) Fix for non blocking accept BIOs. Added new I/O special reason 6376109998Smarkm BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs 6377109998Smarkm with non blocking I/O was not possible because no retry code was 6378109998Smarkm implemented. Also added new SSL code SSL_WANT_ACCEPT to cover 6379109998Smarkm this case. 6380109998Smarkm [Steve Henson] 6381109998Smarkm 6382109998Smarkm *) Added the beginnings of Rijndael support. 6383109998Smarkm [Ben Laurie] 6384109998Smarkm 6385109998Smarkm *) Fix for bug in DirectoryString mask setting. Add support for 6386109998Smarkm X509_NAME_print_ex() in 'req' and X509_print_ex() function 6387109998Smarkm to allow certificate printing to more controllable, additional 6388109998Smarkm 'certopt' option to 'x509' to allow new printing options to be 6389109998Smarkm set. 6390109998Smarkm [Steve Henson] 6391109998Smarkm 6392109998Smarkm *) Clean old EAY MD5 hack from e_os.h. 6393109998Smarkm [Richard Levitte] 6394109998Smarkm 6395142425Snectar Changes between 0.9.6l and 0.9.6m [17 Mar 2004] 6396142425Snectar 6397142425Snectar *) Fix null-pointer assignment in do_change_cipher_spec() revealed 6398160814Ssimon by using the Codenomicon TLS Test Tool (CVE-2004-0079) 6399142425Snectar [Joe Orton, Steve Henson] 6400142425Snectar 6401142425Snectar Changes between 0.9.6k and 0.9.6l [04 Nov 2003] 6402142425Snectar 6403142425Snectar *) Fix additional bug revealed by the NISCC test suite: 6404142425Snectar 6405142425Snectar Stop bug triggering large recursion when presented with 6406160814Ssimon certain ASN.1 tags (CVE-2003-0851) 6407142425Snectar [Steve Henson] 6408142425Snectar 6409120631Snectar Changes between 0.9.6j and 0.9.6k [30 Sep 2003] 6410120631Snectar 6411120631Snectar *) Fix various bugs revealed by running the NISCC test suite: 6412120631Snectar 6413120631Snectar Stop out of bounds reads in the ASN1 code when presented with 6414160814Ssimon invalid tags (CVE-2003-0543 and CVE-2003-0544). 6415120631Snectar 6416120631Snectar If verify callback ignores invalid public key errors don't try to check 6417120631Snectar certificate signature with the NULL public key. 6418120631Snectar 6419120631Snectar [Steve Henson] 6420120631Snectar 6421120631Snectar *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate 6422120631Snectar if the server requested one: as stated in TLS 1.0 and SSL 3.0 6423120631Snectar specifications. 6424120631Snectar [Steve Henson] 6425120631Snectar 6426120631Snectar *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional 6427120631Snectar extra data after the compression methods not only for TLS 1.0 6428120631Snectar but also for SSL 3.0 (as required by the specification). 6429120631Snectar [Bodo Moeller; problem pointed out by Matthias Loepfe] 6430120631Snectar 6431120631Snectar *) Change X509_certificate_type() to mark the key as exported/exportable 6432120631Snectar when it's 512 *bits* long, not 512 bytes. 6433120631Snectar [Richard Levitte] 6434120631Snectar 6435120631Snectar Changes between 0.9.6i and 0.9.6j [10 Apr 2003] 6436120631Snectar 6437120631Snectar *) Countermeasure against the Klima-Pokorny-Rosa extension of 6438120631Snectar Bleichbacher's attack on PKCS #1 v1.5 padding: treat 6439120631Snectar a protocol version number mismatch like a decryption error 6440120631Snectar in ssl3_get_client_key_exchange (ssl/s3_srvr.c). 6441120631Snectar [Bodo Moeller] 6442120631Snectar 6443120631Snectar *) Turn on RSA blinding by default in the default implementation 6444120631Snectar to avoid a timing attack. Applications that don't want it can call 6445120631Snectar RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. 6446120631Snectar They would be ill-advised to do so in most cases. 6447120631Snectar [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] 6448120631Snectar 6449120631Snectar *) Change RSA blinding code so that it works when the PRNG is not 6450120631Snectar seeded (in this case, the secret RSA exponent is abused as 6451120631Snectar an unpredictable seed -- if it is not unpredictable, there 6452120631Snectar is no point in blinding anyway). Make RSA blinding thread-safe 6453120631Snectar by remembering the creator's thread ID in rsa->blinding and 6454120631Snectar having all other threads use local one-time blinding factors 6455120631Snectar (this requires more computation than sharing rsa->blinding, but 6456120631Snectar avoids excessive locking; and if an RSA object is not shared 6457120631Snectar between threads, blinding will still be very fast). 6458120631Snectar [Bodo Moeller] 6459120631Snectar 6460111147Snectar Changes between 0.9.6h and 0.9.6i [19 Feb 2003] 6461111147Snectar 6462111147Snectar *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked 6463111147Snectar via timing by performing a MAC computation even if incorrrect 6464111147Snectar block cipher padding has been found. This is a countermeasure 6465111147Snectar against active attacks where the attacker has to distinguish 6466160814Ssimon between bad padding and a MAC verification error. (CVE-2003-0078) 6467111147Snectar 6468111147Snectar [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), 6469111147Snectar Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and 6470111147Snectar Martin Vuagnoux (EPFL, Ilion)] 6471111147Snectar 6472109998Smarkm Changes between 0.9.6g and 0.9.6h [5 Dec 2002] 6473109998Smarkm 6474109998Smarkm *) New function OPENSSL_cleanse(), which is used to cleanse a section of 6475109998Smarkm memory from it's contents. This is done with a counter that will 6476109998Smarkm place alternating values in each byte. This can be used to solve 6477109998Smarkm two issues: 1) the removal of calls to memset() by highly optimizing 6478109998Smarkm compilers, and 2) cleansing with other values than 0, since those can 6479109998Smarkm be read through on certain media, for example a swap space on disk. 6480109998Smarkm [Geoff Thorpe] 6481109998Smarkm 6482109998Smarkm *) Bugfix: client side session caching did not work with external caching, 6483109998Smarkm because the session->cipher setting was not restored when reloading 6484109998Smarkm from the external cache. This problem was masked, when 6485109998Smarkm SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. 6486109998Smarkm (Found by Steve Haslam <steve@araqnid.ddts.net>.) 6487109998Smarkm [Lutz Jaenicke] 6488109998Smarkm 6489109998Smarkm *) Fix client_certificate (ssl/s2_clnt.c): The permissible total 6490109998Smarkm length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. 6491109998Smarkm [Zeev Lieber <zeev-l@yahoo.com>] 6492109998Smarkm 6493109998Smarkm *) Undo an undocumented change introduced in 0.9.6e which caused 6494109998Smarkm repeated calls to OpenSSL_add_all_ciphers() and 6495109998Smarkm OpenSSL_add_all_digests() to be ignored, even after calling 6496109998Smarkm EVP_cleanup(). 6497109998Smarkm [Richard Levitte] 6498109998Smarkm 6499109998Smarkm *) Change the default configuration reader to deal with last line not 6500109998Smarkm being properly terminated. 6501109998Smarkm [Richard Levitte] 6502109998Smarkm 6503109998Smarkm *) Change X509_NAME_cmp() so it applies the special rules on handling 6504109998Smarkm DN values that are of type PrintableString, as well as RDNs of type 6505109998Smarkm emailAddress where the value has the type ia5String. 6506109998Smarkm [stefank@valicert.com via Richard Levitte] 6507109998Smarkm 6508109998Smarkm *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half 6509109998Smarkm the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently 6510109998Smarkm doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be 6511109998Smarkm the bitwise-OR of the two for use by the majority of applications 6512109998Smarkm wanting this behaviour, and update the docs. The documented 6513109998Smarkm behaviour and actual behaviour were inconsistent and had been 6514109998Smarkm changing anyway, so this is more a bug-fix than a behavioural 6515109998Smarkm change. 6516109998Smarkm [Geoff Thorpe, diagnosed by Nadav Har'El] 6517109998Smarkm 6518109998Smarkm *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c 6519109998Smarkm (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). 6520109998Smarkm [Bodo Moeller] 6521109998Smarkm 6522109998Smarkm *) Fix initialization code race conditions in 6523109998Smarkm SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), 6524109998Smarkm SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), 6525109998Smarkm SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), 6526109998Smarkm TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), 6527109998Smarkm ssl2_get_cipher_by_char(), 6528109998Smarkm ssl3_get_cipher_by_char(). 6529109998Smarkm [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] 6530109998Smarkm 6531109998Smarkm *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after 6532109998Smarkm the cached sessions are flushed, as the remove_cb() might use ex_data 6533109998Smarkm contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> 6534109998Smarkm (see [openssl.org #212]). 6535109998Smarkm [Geoff Thorpe, Lutz Jaenicke] 6536109998Smarkm 6537109998Smarkm *) Fix typo in OBJ_txt2obj which incorrectly passed the content 6538109998Smarkm length, instead of the encoding length to d2i_ASN1_OBJECT. 6539109998Smarkm [Steve Henson] 6540109998Smarkm 6541101618Snectar Changes between 0.9.6f and 0.9.6g [9 Aug 2002] 6542101618Snectar 6543101618Snectar *) [In 0.9.6g-engine release:] 6544101618Snectar Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). 6545101618Snectar [Lynn Gazis <lgazis@rainbow.com>] 6546101618Snectar 6547101613Snectar Changes between 0.9.6e and 0.9.6f [8 Aug 2002] 6548101613Snectar 6549101613Snectar *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX 6550101613Snectar and get fix the header length calculation. 6551101613Snectar [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, 6552101613Snectar Alon Kantor <alonk@checkpoint.com> (and others), 6553101613Snectar Steve Henson] 6554101613Snectar 6555101613Snectar *) Use proper error handling instead of 'assertions' in buffer 6556101613Snectar overflow checks added in 0.9.6e. This prevents DoS (the 6557101613Snectar assertions could call abort()). 6558101613Snectar [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] 6559101613Snectar 6560100936Snectar Changes between 0.9.6d and 0.9.6e [30 Jul 2002] 6561100936Snectar 6562109998Smarkm *) Add various sanity checks to asn1_get_length() to reject 6563109998Smarkm the ASN1 length bytes if they exceed sizeof(long), will appear 6564109998Smarkm negative or the content length exceeds the length of the 6565109998Smarkm supplied buffer. 6566109998Smarkm [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6567109998Smarkm 6568100936Snectar *) Fix cipher selection routines: ciphers without encryption had no flags 6569100936Snectar for the cipher strength set and where therefore not handled correctly 6570100936Snectar by the selection routines (PR #130). 6571100936Snectar [Lutz Jaenicke] 6572100936Snectar 6573100936Snectar *) Fix EVP_dsa_sha macro. 6574100936Snectar [Nils Larsch] 6575100936Snectar 6576100936Snectar *) New option 6577100936Snectar SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS 6578100936Snectar for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure 6579100936Snectar that was added in OpenSSL 0.9.6d. 6580100936Snectar 6581100936Snectar As the countermeasure turned out to be incompatible with some 6582100936Snectar broken SSL implementations, the new option is part of SSL_OP_ALL. 6583100936Snectar SSL_OP_ALL is usually employed when compatibility with weird SSL 6584100936Snectar implementations is desired (e.g. '-bugs' option to 's_client' and 6585100936Snectar 's_server'), so the new option is automatically set in many 6586100936Snectar applications. 6587100936Snectar [Bodo Moeller] 6588100936Snectar 6589100936Snectar *) Changes in security patch: 6590100936Snectar 6591100936Snectar Changes marked "(CHATS)" were sponsored by the Defense Advanced 6592100936Snectar Research Projects Agency (DARPA) and Air Force Research Laboratory, 6593100936Snectar Air Force Materiel Command, USAF, under agreement number 6594100936Snectar F30602-01-2-0537. 6595100936Snectar 6596100936Snectar *) Add various sanity checks to asn1_get_length() to reject 6597100936Snectar the ASN1 length bytes if they exceed sizeof(long), will appear 6598100936Snectar negative or the content length exceeds the length of the 6599160814Ssimon supplied buffer. (CVE-2002-0659) 6600100936Snectar [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] 6601100936Snectar 6602100936Snectar *) Assertions for various potential buffer overflows, not known to 6603100936Snectar happen in practice. 6604100936Snectar [Ben Laurie (CHATS)] 6605100936Snectar 6606100936Snectar *) Various temporary buffers to hold ASCII versions of integers were 6607160814Ssimon too small for 64 bit platforms. (CVE-2002-0655) 6608100936Snectar [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> 6609100936Snectar 6610100936Snectar *) Remote buffer overflow in SSL3 protocol - an attacker could 6611160814Ssimon supply an oversized session ID to a client. (CVE-2002-0656) 6612100936Snectar [Ben Laurie (CHATS)] 6613100936Snectar 6614100936Snectar *) Remote buffer overflow in SSL2 protocol - an attacker could 6615160814Ssimon supply an oversized client master key. (CVE-2002-0656) 6616100936Snectar [Ben Laurie (CHATS)] 6617100936Snectar 6618100928Snectar Changes between 0.9.6c and 0.9.6d [9 May 2002] 6619100928Snectar 6620100928Snectar *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not 6621100928Snectar encoded as NULL) with id-dsa-with-sha1. 6622100928Snectar [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] 6623100928Snectar 6624100928Snectar *) Check various X509_...() return values in apps/req.c. 6625100928Snectar [Nils Larsch <nla@trustcenter.de>] 6626100928Snectar 6627100928Snectar *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: 6628100928Snectar an end-of-file condition would erronously be flagged, when the CRLF 6629100928Snectar was just at the end of a processed block. The bug was discovered when 6630100928Snectar processing data through a buffering memory BIO handing the data to a 6631100928Snectar BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov 6632100928Snectar <ptsekov@syntrex.com> and Nedelcho Stanev. 6633100928Snectar [Lutz Jaenicke] 6634100928Snectar 6635100928Snectar *) Implement a countermeasure against a vulnerability recently found 6636100928Snectar in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment 6637100928Snectar before application data chunks to avoid the use of known IVs 6638100928Snectar with data potentially chosen by the attacker. 6639100928Snectar [Bodo Moeller] 6640100928Snectar 6641100928Snectar *) Fix length checks in ssl3_get_client_hello(). 6642100928Snectar [Bodo Moeller] 6643100928Snectar 6644100928Snectar *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently 6645100928Snectar to prevent ssl3_read_internal() from incorrectly assuming that 6646100928Snectar ssl3_read_bytes() found application data while handshake 6647100928Snectar processing was enabled when in fact s->s3->in_read_app_data was 6648100928Snectar merely automatically cleared during the initial handshake. 6649100928Snectar [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] 6650100928Snectar 6651100928Snectar *) Fix object definitions for Private and Enterprise: they were not 6652100928Snectar recognized in their shortname (=lowercase) representation. Extend 6653100928Snectar obj_dat.pl to issue an error when using undefined keywords instead 6654100928Snectar of silently ignoring the problem (Svenning Sorensen 6655100928Snectar <sss@sss.dnsalias.net>). 6656100928Snectar [Lutz Jaenicke] 6657100928Snectar 6658100928Snectar *) Fix DH_generate_parameters() so that it works for 'non-standard' 6659100928Snectar generators, i.e. generators other than 2 and 5. (Previously, the 6660100928Snectar code did not properly initialise the 'add' and 'rem' values to 6661100928Snectar BN_generate_prime().) 6662100928Snectar 6663100928Snectar In the new general case, we do not insist that 'generator' is 6664100928Snectar actually a primitive root: This requirement is rather pointless; 6665100928Snectar a generator of the order-q subgroup is just as good, if not 6666100928Snectar better. 6667100928Snectar [Bodo Moeller] 6668100928Snectar 6669100928Snectar *) Map new X509 verification errors to alerts. Discovered and submitted by 6670100928Snectar Tom Wu <tom@arcot.com>. 6671100928Snectar [Lutz Jaenicke] 6672100928Snectar 6673100928Snectar *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from 6674100928Snectar returning non-zero before the data has been completely received 6675100928Snectar when using non-blocking I/O. 6676100928Snectar [Bodo Moeller; problem pointed out by John Hughes] 6677100928Snectar 6678100928Snectar *) Some of the ciphers missed the strength entry (SSL_LOW etc). 6679100928Snectar [Ben Laurie, Lutz Jaenicke] 6680100928Snectar 6681100928Snectar *) Fix bug in SSL_clear(): bad sessions were not removed (found by 6682100928Snectar Yoram Zahavi <YoramZ@gilian.com>). 6683100928Snectar [Lutz Jaenicke] 6684100928Snectar 6685100928Snectar *) Add information about CygWin 1.3 and on, and preserve proper 6686100928Snectar configuration for the versions before that. 6687100928Snectar [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] 6688100928Snectar 6689100928Snectar *) Make removal from session cache (SSL_CTX_remove_session()) more robust: 6690100928Snectar check whether we deal with a copy of a session and do not delete from 6691100928Snectar the cache in this case. Problem reported by "Izhar Shoshani Levi" 6692100928Snectar <izhar@checkpoint.com>. 6693100928Snectar [Lutz Jaenicke] 6694100928Snectar 6695100928Snectar *) Do not store session data into the internal session cache, if it 6696100928Snectar is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP 6697100928Snectar flag is set). Proposed by Aslam <aslam@funk.com>. 6698100928Snectar [Lutz Jaenicke] 6699100928Snectar 6700100928Snectar *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested 6701100928Snectar value is 0. 6702100928Snectar [Richard Levitte] 6703100928Snectar 6704109998Smarkm *) [In 0.9.6d-engine release:] 6705109998Smarkm Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). 6706100928Snectar [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] 6707100928Snectar 6708100928Snectar *) Add the configuration target linux-s390x. 6709100928Snectar [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] 6710100928Snectar 6711100928Snectar *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of 6712100928Snectar ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag 6713100928Snectar variable as an indication that a ClientHello message has been 6714100928Snectar received. As the flag value will be lost between multiple 6715100928Snectar invocations of ssl3_accept when using non-blocking I/O, the 6716100928Snectar function may not be aware that a handshake has actually taken 6717100928Snectar place, thus preventing a new session from being added to the 6718100928Snectar session cache. 6719100928Snectar 6720100928Snectar To avoid this problem, we now set s->new_session to 2 instead of 6721100928Snectar using a local variable. 6722100928Snectar [Lutz Jaenicke, Bodo Moeller] 6723100928Snectar 6724100928Snectar *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) 6725100928Snectar if the SSL_R_LENGTH_MISMATCH error is detected. 6726100928Snectar [Geoff Thorpe, Bodo Moeller] 6727100928Snectar 6728100928Snectar *) New 'shared_ldflag' column in Configure platform table. 6729100928Snectar [Richard Levitte] 6730100928Snectar 6731100928Snectar *) Fix EVP_CIPHER_mode macro. 6732100928Snectar ["Dan S. Camper" <dan@bti.net>] 6733100928Snectar 6734100928Snectar *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown 6735100928Snectar type, we must throw them away by setting rr->length to 0. 6736100928Snectar [D P Chang <dpc@qualys.com>] 6737100928Snectar 673889837Skris Changes between 0.9.6b and 0.9.6c [21 dec 2001] 673989837Skris 674089837Skris *) Fix BN_rand_range bug pointed out by Dominikus Scherkl 674189837Skris <Dominikus.Scherkl@biodata.com>. (The previous implementation 674289837Skris worked incorrectly for those cases where range = 10..._2 and 674389837Skris 3*range is two bits longer than range.) 674489837Skris [Bodo Moeller] 674589837Skris 674689837Skris *) Only add signing time to PKCS7 structures if it is not already 674789837Skris present. 674889837Skris [Steve Henson] 674989837Skris 675089837Skris *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", 675189837Skris OBJ_ld_ce should be OBJ_id_ce. 675289837Skris Also some ip-pda OIDs in crypto/objects/objects.txt were 675389837Skris incorrect (cf. RFC 3039). 675489837Skris [Matt Cooper, Frederic Giudicelli, Bodo Moeller] 675589837Skris 675689837Skris *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() 675789837Skris returns early because it has nothing to do. 675889837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 675989837Skris 676089837Skris *) [In 0.9.6c-engine release:] 676189837Skris Fix mutex callback return values in crypto/engine/hw_ncipher.c. 676289837Skris [Andy Schneider <andy.schneider@bjss.co.uk>] 676389837Skris 676489837Skris *) [In 0.9.6c-engine release:] 676589837Skris Add support for Cryptographic Appliance's keyserver technology. 676689837Skris (Use engine 'keyclient') 676789837Skris [Cryptographic Appliances and Geoff Thorpe] 676889837Skris 676989837Skris *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' 677089837Skris is called via tools/c89.sh because arguments have to be 677189837Skris rearranged (all '-L' options must appear before the first object 677289837Skris modules). 677389837Skris [Richard Shapiro <rshapiro@abinitio.com>] 677489837Skris 677589837Skris *) [In 0.9.6c-engine release:] 677689837Skris Add support for Broadcom crypto accelerator cards, backported 677789837Skris from 0.9.7. 677889837Skris [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] 677989837Skris 678089837Skris *) [In 0.9.6c-engine release:] 678189837Skris Add support for SureWare crypto accelerator cards from 678289837Skris Baltimore Technologies. (Use engine 'sureware') 678389837Skris [Baltimore Technologies and Mark Cox] 678489837Skris 678589837Skris *) [In 0.9.6c-engine release:] 678689837Skris Add support for crypto accelerator cards from Accelerated 678789837Skris Encryption Processing, www.aep.ie. (Use engine 'aep') 678889837Skris [AEP Inc. and Mark Cox] 678989837Skris 679089837Skris *) Add a configuration entry for gcc on UnixWare. 679189837Skris [Gary Benson <gbenson@redhat.com>] 679289837Skris 679389837Skris *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake 679489837Skris messages are stored in a single piece (fixed-length part and 679589837Skris variable-length part combined) and fix various bugs found on the way. 679689837Skris [Bodo Moeller] 679789837Skris 679889837Skris *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() 679989837Skris instead. BIO_gethostbyname() does not know what timeouts are 680089837Skris appropriate, so entries would stay in cache even when they have 680189837Skris become invalid. 680289837Skris [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> 680389837Skris 680489837Skris *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when 680589837Skris faced with a pathologically small ClientHello fragment that does 680689837Skris not contain client_version: Instead of aborting with an error, 680789837Skris simply choose the highest available protocol version (i.e., 680889837Skris TLS 1.0 unless it is disabled). In practice, ClientHello 680989837Skris messages are never sent like this, but this change gives us 681089837Skris strictly correct behaviour at least for TLS. 681189837Skris [Bodo Moeller] 681289837Skris 681389837Skris *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() 681489837Skris never resets s->method to s->ctx->method when called from within 681589837Skris one of the SSL handshake functions. 681689837Skris [Bodo Moeller; problem pointed out by Niko Baric] 681789837Skris 681889837Skris *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert 681989837Skris (sent using the client's version number) if client_version is 682089837Skris smaller than the protocol version in use. Also change 682189837Skris ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if 682289837Skris the client demanded SSL 3.0 but only TLS 1.0 is enabled; then 682389837Skris the client will at least see that alert. 682489837Skris [Bodo Moeller] 682589837Skris 682689837Skris *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation 682789837Skris correctly. 682889837Skris [Bodo Moeller] 682989837Skris 683089837Skris *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a 683189837Skris client receives HelloRequest while in a handshake. 683289837Skris [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] 683389837Skris 683489837Skris *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C 683589837Skris should end in 'break', not 'goto end' which circuments various 683689837Skris cleanups done in state SSL_ST_OK. But session related stuff 683789837Skris must be disabled for SSL_ST_OK in the case that we just sent a 683889837Skris HelloRequest. 683989837Skris 684089837Skris Also avoid some overhead by not calling ssl_init_wbio_buffer() 684189837Skris before just sending a HelloRequest. 684289837Skris [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] 684389837Skris 684489837Skris *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't 684589837Skris reveal whether illegal block cipher padding was found or a MAC 684689837Skris verification error occured. (Neither SSLerr() codes nor alerts 684789837Skris are directly visible to potential attackers, but the information 684889837Skris may leak via logfiles.) 684989837Skris 685089837Skris Similar changes are not required for the SSL 2.0 implementation 685189837Skris because the number of padding bytes is sent in clear for SSL 2.0, 685289837Skris and the extra bytes are just ignored. However ssl/s2_pkt.c 685389837Skris failed to verify that the purported number of padding bytes is in 685489837Skris the legal range. 685589837Skris [Bodo Moeller] 685689837Skris 685789837Skris *) Add OpenUNIX-8 support including shared libraries 685889837Skris (Boyd Lynn Gerber <gerberb@zenez.com>). 685989837Skris [Lutz Jaenicke] 686089837Skris 686189837Skris *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid 686289837Skris 'wristwatch attack' using huge encoding parameters (cf. 686389837Skris James H. Manger's CRYPTO 2001 paper). Note that the 686489837Skris RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use 686589837Skris encoding parameters and hence was not vulnerable. 686689837Skris [Bodo Moeller] 686789837Skris 686889837Skris *) BN_sqr() bug fix. 6869291721Sjkim [Ulf M��ller, reported by Jim Ellis <jim.ellis@cavium.com>] 687089837Skris 687189837Skris *) Rabin-Miller test analyses assume uniformly distributed witnesses, 687289837Skris so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() 687389837Skris followed by modular reduction. 687489837Skris [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] 687589837Skris 687689837Skris *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() 687789837Skris equivalent based on BN_pseudo_rand() instead of BN_rand(). 687889837Skris [Bodo Moeller] 687989837Skris 688089837Skris *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). 688189837Skris This function was broken, as the check for a new client hello message 688289837Skris to handle SGC did not allow these large messages. 688389837Skris (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) 688489837Skris [Lutz Jaenicke] 688589837Skris 688689837Skris *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). 688789837Skris [Lutz Jaenicke] 688889837Skris 688989837Skris *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() 689089837Skris for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). 689189837Skris [Lutz Jaenicke] 689289837Skris 689389837Skris *) Rework the configuration and shared library support for Tru64 Unix. 689489837Skris The configuration part makes use of modern compiler features and 689589837Skris still retains old compiler behavior for those that run older versions 689689837Skris of the OS. The shared library support part includes a variant that 689789837Skris uses the RPATH feature, and is available through the special 689889837Skris configuration target "alpha-cc-rpath", which will never be selected 689989837Skris automatically. 690089837Skris [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] 690189837Skris 690289837Skris *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() 690389837Skris with the same message size as in ssl3_get_certificate_request(). 690489837Skris Otherwise, if no ServerKeyExchange message occurs, CertificateRequest 690589837Skris messages might inadvertently be reject as too long. 690689837Skris [Petr Lampa <lampa@fee.vutbr.cz>] 690789837Skris 690889837Skris *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). 690989837Skris [Andy Polyakov] 691089837Skris 691189837Skris *) Modified SSL library such that the verify_callback that has been set 691289837Skris specificly for an SSL object with SSL_set_verify() is actually being 691389837Skris used. Before the change, a verify_callback set with this function was 691489837Skris ignored and the verify_callback() set in the SSL_CTX at the time of 691589837Skris the call was used. New function X509_STORE_CTX_set_verify_cb() introduced 691689837Skris to allow the necessary settings. 691789837Skris [Lutz Jaenicke] 691889837Skris 691989837Skris *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c 692089837Skris explicitly to NULL, as at least on Solaris 8 this seems not always to be 692189837Skris done automatically (in contradiction to the requirements of the C 692289837Skris standard). This made problems when used from OpenSSH. 692389837Skris [Lutz Jaenicke] 692489837Skris 692589837Skris *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored 692689837Skris dh->length and always used 692789837Skris 692889837Skris BN_rand_range(priv_key, dh->p). 692989837Skris 693089837Skris BN_rand_range() is not necessary for Diffie-Hellman, and this 693189837Skris specific range makes Diffie-Hellman unnecessarily inefficient if 693289837Skris dh->length (recommended exponent length) is much smaller than the 693389837Skris length of dh->p. We could use BN_rand_range() if the order of 693489837Skris the subgroup was stored in the DH structure, but we only have 693589837Skris dh->length. 693689837Skris 693789837Skris So switch back to 693889837Skris 693989837Skris BN_rand(priv_key, l, ...) 694089837Skris 694189837Skris where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 694289837Skris otherwise. 694389837Skris [Bodo Moeller] 694489837Skris 694589837Skris *) In 694689837Skris 694789837Skris RSA_eay_public_encrypt 694889837Skris RSA_eay_private_decrypt 694989837Skris RSA_eay_private_encrypt (signing) 695089837Skris RSA_eay_public_decrypt (signature verification) 695189837Skris 695289837Skris (default implementations for RSA_public_encrypt, 695389837Skris RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), 695489837Skris always reject numbers >= n. 695589837Skris [Bodo Moeller] 695689837Skris 695789837Skris *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 695889837Skris to synchronize access to 'locking_thread'. This is necessary on 695989837Skris systems where access to 'locking_thread' (an 'unsigned long' 696089837Skris variable) is not atomic. 696189837Skris [Bodo Moeller] 696289837Skris 696389837Skris *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID 696489837Skris *before* setting the 'crypto_lock_rand' flag. The previous code had 696589837Skris a race condition if 0 is a valid thread ID. 696689837Skris [Travis Vitek <vitek@roguewave.com>] 696789837Skris 696889837Skris *) Add support for shared libraries under Irix. 696989837Skris [Albert Chin-A-Young <china@thewrittenword.com>] 697089837Skris 697189837Skris *) Add configuration option to build on Linux on both big-endian and 697289837Skris little-endian MIPS. 697389837Skris [Ralf Baechle <ralf@uni-koblenz.de>] 697489837Skris 697589837Skris *) Add the possibility to create shared libraries on HP-UX. 697689837Skris [Richard Levitte] 697789837Skris 697879998Skris Changes between 0.9.6a and 0.9.6b [9 Jul 2001] 697979998Skris 698079998Skris *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) 698179998Skris to avoid a SSLeay/OpenSSL PRNG weakness pointed out by 698279998Skris Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: 698379998Skris PRNG state recovery was possible based on the output of 698479998Skris one PRNG request appropriately sized to gain knowledge on 698579998Skris 'md' followed by enough consecutive 1-byte PRNG requests 698679998Skris to traverse all of 'state'. 698779998Skris 698879998Skris 1. When updating 'md_local' (the current thread's copy of 'md') 698979998Skris during PRNG output generation, hash all of the previous 699079998Skris 'md_local' value, not just the half used for PRNG output. 699179998Skris 699279998Skris 2. Make the number of bytes from 'state' included into the hash 699379998Skris independent from the number of PRNG bytes requested. 699479998Skris 699579998Skris The first measure alone would be sufficient to avoid 699679998Skris Markku-Juhani's attack. (Actually it had never occurred 699779998Skris to me that the half of 'md_local' used for chaining was the 699879998Skris half from which PRNG output bytes were taken -- I had always 699979998Skris assumed that the secret half would be used.) The second 700079998Skris measure makes sure that additional data from 'state' is never 700179998Skris mixed into 'md_local' in small portions; this heuristically 700279998Skris further strengthens the PRNG. 700379998Skris [Bodo Moeller] 700479998Skris 700579998Skris *) Fix crypto/bn/asm/mips3.s. 700679998Skris [Andy Polyakov] 700779998Skris 700879998Skris *) When only the key is given to "enc", the IV is undefined. Print out 700979998Skris an error message in this case. 701079998Skris [Lutz Jaenicke] 701179998Skris 701279998Skris *) Handle special case when X509_NAME is empty in X509 printing routines. 701379998Skris [Steve Henson] 701479998Skris 701579998Skris *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are 701679998Skris positive and less than q. 701779998Skris [Bodo Moeller] 701879998Skris 701979998Skris *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is 702079998Skris used: it isn't thread safe and the add_lock_callback should handle 702179998Skris that itself. 702279998Skris [Paul Rose <Paul.Rose@bridge.com>] 702379998Skris 702479998Skris *) Verify that incoming data obeys the block size in 702579998Skris ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). 702679998Skris [Bodo Moeller] 702779998Skris 702879998Skris *) Fix OAEP check. 7029291721Sjkim [Ulf M��ller, Bodo M��ller] 703079998Skris 703179998Skris *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 703289837Skris RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 703379998Skris when fixing the server behaviour for backwards-compatible 'client 703479998Skris hello' messages. (Note that the attack is impractical against 703579998Skris SSL 3.0 and TLS 1.0 anyway because length and version checking 703679998Skris means that the probability of guessing a valid ciphertext is 703779998Skris around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 703879998Skris paper.) 703979998Skris 704079998Skris Before 0.9.5, the countermeasure (hide the error by generating a 704179998Skris random 'decryption result') did not work properly because 704279998Skris ERR_clear_error() was missing, meaning that SSL_get_error() would 704379998Skris detect the supposedly ignored error. 704479998Skris 704579998Skris Both problems are now fixed. 704679998Skris [Bodo Moeller] 704779998Skris 704879998Skris *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 704979998Skris (previously it was 1024). 705079998Skris [Bodo Moeller] 705179998Skris 705279998Skris *) Fix for compatibility mode trust settings: ignore trust settings 705379998Skris unless some valid trust or reject settings are present. 705479998Skris [Steve Henson] 705579998Skris 705679998Skris *) Fix for blowfish EVP: its a variable length cipher. 705779998Skris [Steve Henson] 705879998Skris 705979998Skris *) Fix various bugs related to DSA S/MIME verification. Handle missing 706079998Skris parameters in DSA public key structures and return an error in the 706179998Skris DSA routines if parameters are absent. 706279998Skris [Steve Henson] 706379998Skris 706479998Skris *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" 706579998Skris in the current directory if neither $RANDFILE nor $HOME was set. 706679998Skris RAND_file_name() in 0.9.6a returned NULL in this case. This has 706779998Skris caused some confusion to Windows users who haven't defined $HOME. 706879998Skris Thus RAND_file_name() is changed again: e_os.h can define a 706979998Skris DEFAULT_HOME, which will be used if $HOME is not set. 707079998Skris For Windows, we use "C:"; on other platforms, we still require 707179998Skris environment variables. 707279998Skris 707379998Skris *) Move 'if (!initialized) RAND_poll()' into regions protected by 707479998Skris CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids 707579998Skris having multiple threads call RAND_poll() concurrently. 707679998Skris [Bodo Moeller] 707779998Skris 707879998Skris *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a 707979998Skris combination of a flag and a thread ID variable. 708079998Skris Otherwise while one thread is in ssleay_rand_bytes (which sets the 708179998Skris flag), *other* threads can enter ssleay_add_bytes without obeying 708289837Skris the CRYPTO_LOCK_RAND lock (and may even illegally release the lock 708379998Skris that they do not hold after the first thread unsets add_do_not_lock). 708479998Skris [Bodo Moeller] 708579998Skris 708679998Skris *) Change bctest again: '-x' expressions are not available in all 708779998Skris versions of 'test'. 708879998Skris [Bodo Moeller] 708979998Skris 709076866Skris Changes between 0.9.6 and 0.9.6a [5 Apr 2001] 709172613Skris 709276866Skris *) Fix a couple of memory leaks in PKCS7_dataDecode() 709376866Skris [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] 709476866Skris 709576866Skris *) Change Configure and Makefiles to provide EXE_EXT, which will contain 709676866Skris the default extension for executables, if any. Also, make the perl 709776866Skris scripts that use symlink() to test if it really exists and use "cp" 709876866Skris if it doesn't. All this made OpenSSL compilable and installable in 709976866Skris CygWin. 710076866Skris [Richard Levitte] 710176866Skris 710276866Skris *) Fix for asn1_GetSequence() for indefinite length constructed data. 710376866Skris If SEQUENCE is length is indefinite just set c->slen to the total 710476866Skris amount of data available. 710576866Skris [Steve Henson, reported by shige@FreeBSD.org] 710676866Skris [This change does not apply to 0.9.7.] 710776866Skris 710876866Skris *) Change bctest to avoid here-documents inside command substitution 710976866Skris (workaround for FreeBSD /bin/sh bug). 711076866Skris For compatibility with Ultrix, avoid shell functions (introduced 711176866Skris in the bctest version that searches along $PATH). 711276866Skris [Bodo Moeller] 711376866Skris 711476866Skris *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes 711576866Skris with des_encrypt() defined on some operating systems, like Solaris 711676866Skris and UnixWare. 711776866Skris [Richard Levitte] 711876866Skris 711976866Skris *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: 712076866Skris On the Importance of Eliminating Errors in Cryptographic 712176866Skris Computations, J. Cryptology 14 (2001) 2, 101-119, 712276866Skris http://theory.stanford.edu/~dabo/papers/faults.ps.gz). 712376866Skris [Ulf Moeller] 712476866Skris 712576866Skris *) MIPS assembler BIGNUM division bug fix. 712676866Skris [Andy Polyakov] 712776866Skris 712876866Skris *) Disabled incorrect Alpha assembler code. 712976866Skris [Richard Levitte] 713076866Skris 713176866Skris *) Fix PKCS#7 decode routines so they correctly update the length 713276866Skris after reading an EOC for the EXPLICIT tag. 713376866Skris [Steve Henson] 713476866Skris [This change does not apply to 0.9.7.] 713576866Skris 713676866Skris *) Fix bug in PKCS#12 key generation routines. This was triggered 713776866Skris if a 3DES key was generated with a 0 initial byte. Include 713876866Skris PKCS12_BROKEN_KEYGEN compilation option to retain the old 713976866Skris (but broken) behaviour. 714076866Skris [Steve Henson] 714176866Skris 714276866Skris *) Enhance bctest to search for a working bc along $PATH and print 714376866Skris it when found. 714476866Skris [Tim Rice <tim@multitalents.net> via Richard Levitte] 714576866Skris 714676866Skris *) Fix memory leaks in err.c: free err_data string if necessary; 714776866Skris don't write to the wrong index in ERR_set_error_data. 714876866Skris [Bodo Moeller] 714976866Skris 715076866Skris *) Implement ssl23_peek (analogous to ssl23_read), which previously 715176866Skris did not exist. 715276866Skris [Bodo Moeller] 715376866Skris 715476866Skris *) Replace rdtsc with _emit statements for VC++ version 5. 715576866Skris [Jeremy Cooper <jeremy@baymoo.org>] 715676866Skris 715776866Skris *) Make it possible to reuse SSLv2 sessions. 715876866Skris [Richard Levitte] 715976866Skris 716076866Skris *) In copy_email() check for >= 0 as a return value for 716176866Skris X509_NAME_get_index_by_NID() since 0 is a valid index. 716276866Skris [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] 716376866Skris 716476866Skris *) Avoid coredump with unsupported or invalid public keys by checking if 716576866Skris X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when 716676866Skris PKCS7_verify() fails with non detached data. 716776866Skris [Steve Henson] 716876866Skris 716976866Skris *) Don't use getenv in library functions when run as setuid/setgid. 717076866Skris New function OPENSSL_issetugid(). 717176866Skris [Ulf Moeller] 717276866Skris 717376866Skris *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) 717476866Skris due to incorrect handling of multi-threading: 717576866Skris 717676866Skris 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). 717776866Skris 717876866Skris 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). 717976866Skris 718076866Skris 3. Count how many times MemCheck_off() has been called so that 718176866Skris nested use can be treated correctly. This also avoids 718276866Skris inband-signalling in the previous code (which relied on the 718376866Skris assumption that thread ID 0 is impossible). 718476866Skris [Bodo Moeller] 718576866Skris 718676866Skris *) Add "-rand" option also to s_client and s_server. 718776866Skris [Lutz Jaenicke] 718876866Skris 718976866Skris *) Fix CPU detection on Irix 6.x. 719076866Skris [Kurt Hockenbury <khockenb@stevens-tech.edu> and 719176866Skris "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 719276866Skris 719376866Skris *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME 719476866Skris was empty. 719576866Skris [Steve Henson] 719676866Skris [This change does not apply to 0.9.7.] 719776866Skris 719876866Skris *) Use the cached encoding of an X509_NAME structure rather than 719976866Skris copying it. This is apparently the reason for the libsafe "errors" 720076866Skris but the code is actually correct. 720176866Skris [Steve Henson] 720276866Skris 720372613Skris *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent 720472613Skris Bleichenbacher's DSA attack. 720576866Skris Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits 720676866Skris to be set and top=0 forces the highest bit to be set; top=-1 is new 720776866Skris and leaves the highest bit random. 720876866Skris [Ulf Moeller, Bodo Moeller] 720972613Skris 721072613Skris *) In the NCONF_...-based implementations for CONF_... queries 721172613Skris (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using 721272613Skris a temporary CONF structure with the data component set to NULL 721372613Skris (which gives segmentation faults in lh_retrieve). 721472613Skris Instead, use NULL for the CONF pointer in CONF_get_string and 721572613Skris CONF_get_number (which may use environment variables) and directly 721672613Skris return NULL from CONF_get_section. 721772613Skris [Bodo Moeller] 721872613Skris 721972613Skris *) Fix potential buffer overrun for EBCDIC. 722072613Skris [Ulf Moeller] 722172613Skris 722272613Skris *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign 722372613Skris keyUsage if basicConstraints absent for a CA. 722472613Skris [Steve Henson] 722572613Skris 722672613Skris *) Make SMIME_write_PKCS7() write mail header values with a format that 722772613Skris is more generally accepted (no spaces before the semicolon), since 722872613Skris some programs can't parse those values properly otherwise. Also make 722972613Skris sure BIO's that break lines after each write do not create invalid 723072613Skris headers. 723172613Skris [Richard Levitte] 723272613Skris 723372613Skris *) Make the CRL encoding routines work with empty SEQUENCE OF. The 723472613Skris macros previously used would not encode an empty SEQUENCE OF 723572613Skris and break the signature. 723672613Skris [Steve Henson] 723776866Skris [This change does not apply to 0.9.7.] 723872613Skris 723972613Skris *) Zero the premaster secret after deriving the master secret in 724072613Skris DH ciphersuites. 724172613Skris [Steve Henson] 724272613Skris 724372613Skris *) Add some EVP_add_digest_alias registrations (as found in 724472613Skris OpenSSL_add_all_digests()) to SSL_library_init() 724572613Skris aka OpenSSL_add_ssl_algorithms(). This provides improved 724672613Skris compatibility with peers using X.509 certificates 724772613Skris with unconventional AlgorithmIdentifier OIDs. 724872613Skris [Bodo Moeller] 724972613Skris 725072613Skris *) Fix for Irix with NO_ASM. 725172613Skris ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] 725272613Skris 725372613Skris *) ./config script fixes. 725472613Skris [Ulf Moeller, Richard Levitte] 725572613Skris 725672613Skris *) Fix 'openssl passwd -1'. 725772613Skris [Bodo Moeller] 725872613Skris 725972613Skris *) Change PKCS12_key_gen_asc() so it can cope with non null 726072613Skris terminated strings whose length is passed in the passlen 726172613Skris parameter, for example from PEM callbacks. This was done 726272613Skris by adding an extra length parameter to asc2uni(). 726372613Skris [Steve Henson, reported by <oddissey@samsung.co.kr>] 726472613Skris 726572613Skris *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn 726672613Skris call failed, free the DSA structure. 726772613Skris [Bodo Moeller] 726872613Skris 726972613Skris *) Fix to uni2asc() to cope with zero length Unicode strings. 727072613Skris These are present in some PKCS#12 files. 727172613Skris [Steve Henson] 727272613Skris 727372613Skris *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). 727472613Skris Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits 727572613Skris when writing a 32767 byte record. 727672613Skris [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] 727772613Skris 727872613Skris *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), 727972613Skris obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. 728072613Skris 728172613Skris (RSA objects have a reference count access to which is protected 728272613Skris by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], 728372613Skris so they are meant to be shared between threads.) 728472613Skris [Bodo Moeller, Geoff Thorpe; original patch submitted by 728572613Skris "Reddie, Steven" <Steven.Reddie@ca.com>] 728672613Skris 728772613Skris *) Fix a deadlock in CRYPTO_mem_leaks(). 728872613Skris [Bodo Moeller] 728972613Skris 729076866Skris *) Use better test patterns in bntest. 7291291721Sjkim [Ulf M��ller] 729276866Skris 729372613Skris *) rand_win.c fix for Borland C. 7294291721Sjkim [Ulf M��ller] 729572613Skris 729672613Skris *) BN_rshift bugfix for n == 0. 729772613Skris [Bodo Moeller] 729872613Skris 729976866Skris *) Add a 'bctest' script that checks for some known 'bc' bugs 730076866Skris so that 'make test' does not abort just because 'bc' is broken. 730176866Skris [Bodo Moeller] 730276866Skris 730372613Skris *) Store verify_result within SSL_SESSION also for client side to 730472613Skris avoid potential security hole. (Re-used sessions on the client side 730572613Skris always resulted in verify_result==X509_V_OK, not using the original 730672613Skris result of the server certificate verification.) 730772613Skris [Lutz Jaenicke] 730872613Skris 730972613Skris *) Fix ssl3_pending: If the record in s->s3->rrec is not of type 731072613Skris SSL3_RT_APPLICATION_DATA, return 0. 731172613Skris Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. 731272613Skris [Bodo Moeller] 731372613Skris 731472613Skris *) Fix SSL_peek: 731572613Skris Both ssl2_peek and ssl3_peek, which were totally broken in earlier 731672613Skris releases, have been re-implemented by renaming the previous 731772613Skris implementations of ssl2_read and ssl3_read to ssl2_read_internal 731872613Skris and ssl3_read_internal, respectively, and adding 'peek' parameters 731972613Skris to them. The new ssl[23]_{read,peek} functions are calls to 732072613Skris ssl[23]_read_internal with the 'peek' flag set appropriately. 732172613Skris A 'peek' parameter has also been added to ssl3_read_bytes, which 732272613Skris does the actual work for ssl3_read_internal. 732372613Skris [Bodo Moeller] 732472613Skris 732576866Skris *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling 732676866Skris the method-specific "init()" handler. Also clean up ex_data after 732776866Skris calling the method-specific "finish()" handler. Previously, this was 732876866Skris happening the other way round. 732976866Skris [Geoff Thorpe] 733076866Skris 733172613Skris *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. 733272613Skris The previous value, 12, was not always sufficient for BN_mod_exp(). 733372613Skris [Bodo Moeller] 733472613Skris 733576866Skris *) Make sure that shared libraries get the internal name engine with 733676866Skris the full version number and not just 0. This should mark the 733776866Skris shared libraries as not backward compatible. Of course, this should 733876866Skris be changed again when we can guarantee backward binary compatibility. 733976866Skris [Richard Levitte] 734076866Skris 734172613Skris *) Fix typo in get_cert_by_subject() in by_dir.c 734272613Skris [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] 734372613Skris 734476866Skris *) Rework the system to generate shared libraries: 734576866Skris 734676866Skris - Make note of the expected extension for the shared libraries and 734776866Skris if there is a need for symbolic links from for example libcrypto.so.0 734876866Skris to libcrypto.so.0.9.7. There is extended info in Configure for 734976866Skris that. 735076866Skris 735176866Skris - Make as few rebuilds of the shared libraries as possible. 735276866Skris 735376866Skris - Still avoid linking the OpenSSL programs with the shared libraries. 735476866Skris 735576866Skris - When installing, install the shared libraries separately from the 735676866Skris static ones. 735776866Skris [Richard Levitte] 735876866Skris 735972613Skris *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. 736072613Skris 736172613Skris Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new 736272613Skris and not in SSL_clear because the latter is also used by the 736372613Skris accept/connect functions; previously, the settings made by 736472613Skris SSL_set_read_ahead would be lost during the handshake. 736572613Skris [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] 736672613Skris 736772613Skris *) Correct util/mkdef.pl to be selective about disabled algorithms. 736872613Skris Previously, it would create entries for disableed algorithms no 736972613Skris matter what. 737072613Skris [Richard Levitte] 737172613Skris 737272613Skris *) Added several new manual pages for SSL_* function. 737372613Skris [Lutz Jaenicke] 737472613Skris 737568651Skris Changes between 0.9.5a and 0.9.6 [24 Sep 2000] 737668651Skris 737768651Skris *) In ssl23_get_client_hello, generate an error message when faced 737868651Skris with an initial SSL 3.0/TLS record that is too small to contain the 737968651Skris first two bytes of the ClientHello message, i.e. client_version. 738068651Skris (Note that this is a pathologic case that probably has never happened 738168651Skris in real life.) The previous approach was to use the version number 738268651Skris from the record header as a substitute; but our protocol choice 738368651Skris should not depend on that one because it is not authenticated 738468651Skris by the Finished messages. 738568651Skris [Bodo Moeller] 738668651Skris 738768651Skris *) More robust randomness gathering functions for Windows. 738868651Skris [Jeffrey Altman <jaltman@columbia.edu>] 738968651Skris 739068651Skris *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is 739168651Skris not set then we don't setup the error code for issuer check errors 739268651Skris to avoid possibly overwriting other errors which the callback does 739368651Skris handle. If an application does set the flag then we assume it knows 739468651Skris what it is doing and can handle the new informational codes 739568651Skris appropriately. 739668651Skris [Steve Henson] 739768651Skris 739868651Skris *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for 739968651Skris a general "ANY" type, as such it should be able to decode anything 740068651Skris including tagged types. However it didn't check the class so it would 740168651Skris wrongly interpret tagged types in the same way as their universal 740268651Skris counterpart and unknown types were just rejected. Changed so that the 740368651Skris tagged and unknown types are handled in the same way as a SEQUENCE: 740468651Skris that is the encoding is stored intact. There is also a new type 740568651Skris "V_ASN1_OTHER" which is used when the class is not universal, in this 740668651Skris case we have no idea what the actual type is so we just lump them all 740768651Skris together. 740868651Skris [Steve Henson] 740968651Skris 741068651Skris *) On VMS, stdout may very well lead to a file that is written to 741168651Skris in a record-oriented fashion. That means that every write() will 741268651Skris write a separate record, which will be read separately by the 741368651Skris programs trying to read from it. This can be very confusing. 741468651Skris 741568651Skris The solution is to put a BIO filter in the way that will buffer 741668651Skris text until a linefeed is reached, and then write everything a 741768651Skris line at a time, so every record written will be an actual line, 741868651Skris not chunks of lines and not (usually doesn't happen, but I've 741968651Skris seen it once) several lines in one record. BIO_f_linebuffer() is 742068651Skris the answer. 742168651Skris 742268651Skris Currently, it's a VMS-only method, because that's where it has 742368651Skris been tested well enough. 742468651Skris [Richard Levitte] 742568651Skris 742668651Skris *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, 742768651Skris it can return incorrect results. 742868651Skris (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, 742968651Skris but it was in 0.9.6-beta[12].) 743068651Skris [Bodo Moeller] 743168651Skris 743268651Skris *) Disable the check for content being present when verifying detached 743368651Skris signatures in pk7_smime.c. Some versions of Netscape (wrongly) 743468651Skris include zero length content when signing messages. 743568651Skris [Steve Henson] 743668651Skris 743768651Skris *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR 743868651Skris BIO_ctrl (for BIO pairs). 7439291721Sjkim [Bodo M��ller] 744068651Skris 744168651Skris *) Add DSO method for VMS. 744268651Skris [Richard Levitte] 744368651Skris 744468651Skris *) Bug fix: Montgomery multiplication could produce results with the 744568651Skris wrong sign. 7446291721Sjkim [Ulf M��ller] 744768651Skris 744868651Skris *) Add RPM specification openssl.spec and modify it to build three 744968651Skris packages. The default package contains applications, application 745068651Skris documentation and run-time libraries. The devel package contains 745168651Skris include files, static libraries and function documentation. The 745268651Skris doc package contains the contents of the doc directory. The original 745368651Skris openssl.spec was provided by Damien Miller <djm@mindrot.org>. 745468651Skris [Richard Levitte] 745568651Skris 745668651Skris *) Add a large number of documentation files for many SSL routines. 745768651Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 745868651Skris 745968651Skris *) Add a configuration entry for Sony News 4. 746068651Skris [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] 746168651Skris 746268651Skris *) Don't set the two most significant bits to one when generating a 746368651Skris random number < q in the DSA library. 7464291721Sjkim [Ulf M��ller] 746568651Skris 746668651Skris *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default 746768651Skris behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if 746868651Skris the underlying transport is blocking) if a handshake took place. 746968651Skris (The default behaviour is needed by applications such as s_client 747068651Skris and s_server that use select() to determine when to use SSL_read; 747168651Skris but for applications that know in advance when to expect data, it 747268651Skris just makes things more complicated.) 747368651Skris [Bodo Moeller] 747468651Skris 747568651Skris *) Add RAND_egd_bytes(), which gives control over the number of bytes read 747668651Skris from EGD. 747768651Skris [Ben Laurie] 747868651Skris 747968651Skris *) Add a few more EBCDIC conditionals that make `req' and `x509' 748068651Skris work better on such systems. 748168651Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 748268651Skris 748368651Skris *) Add two demo programs for PKCS12_parse() and PKCS12_create(). 748468651Skris Update PKCS12_parse() so it copies the friendlyName and the 748568651Skris keyid to the certificates aux info. 748668651Skris [Steve Henson] 748768651Skris 748868651Skris *) Fix bug in PKCS7_verify() which caused an infinite loop 748968651Skris if there was more than one signature. 749068651Skris [Sven Uszpelkat <su@celocom.de>] 749168651Skris 749268651Skris *) Major change in util/mkdef.pl to include extra information 749368651Skris about each symbol, as well as presentig variables as well 749468651Skris as functions. This change means that there's n more need 749568651Skris to rebuild the .num files when some algorithms are excluded. 749668651Skris [Richard Levitte] 749768651Skris 749868651Skris *) Allow the verify time to be set by an application, 749968651Skris rather than always using the current time. 750068651Skris [Steve Henson] 750168651Skris 750268651Skris *) Phase 2 verify code reorganisation. The certificate 750368651Skris verify code now looks up an issuer certificate by a 750468651Skris number of criteria: subject name, authority key id 750568651Skris and key usage. It also verifies self signed certificates 750668651Skris by the same criteria. The main comparison function is 750768651Skris X509_check_issued() which performs these checks. 750868651Skris 750968651Skris Lot of changes were necessary in order to support this 751068651Skris without completely rewriting the lookup code. 751168651Skris 751268651Skris Authority and subject key identifier are now cached. 751368651Skris 751468651Skris The LHASH 'certs' is X509_STORE has now been replaced 751568651Skris by a STACK_OF(X509_OBJECT). This is mainly because an 751668651Skris LHASH can't store or retrieve multiple objects with 751768651Skris the same hash value. 751868651Skris 751968651Skris As a result various functions (which were all internal 752068651Skris use only) have changed to handle the new X509_STORE 752168651Skris structure. This will break anything that messed round 752268651Skris with X509_STORE internally. 752368651Skris 752468651Skris The functions X509_STORE_add_cert() now checks for an 752568651Skris exact match, rather than just subject name. 752668651Skris 752768651Skris The X509_STORE API doesn't directly support the retrieval 752868651Skris of multiple certificates matching a given criteria, however 752968651Skris this can be worked round by performing a lookup first 753068651Skris (which will fill the cache with candidate certificates) 753168651Skris and then examining the cache for matches. This is probably 753268651Skris the best we can do without throwing out X509_LOOKUP 753368651Skris entirely (maybe later...). 753468651Skris 753568651Skris The X509_VERIFY_CTX structure has been enhanced considerably. 753668651Skris 753768651Skris All certificate lookup operations now go via a get_issuer() 753868651Skris callback. Although this currently uses an X509_STORE it 753968651Skris can be replaced by custom lookups. This is a simple way 754068651Skris to bypass the X509_STORE hackery necessary to make this 754168651Skris work and makes it possible to use more efficient techniques 754268651Skris in future. A very simple version which uses a simple 754368651Skris STACK for its trusted certificate store is also provided 754468651Skris using X509_STORE_CTX_trusted_stack(). 754568651Skris 754668651Skris The verify_cb() and verify() callbacks now have equivalents 754768651Skris in the X509_STORE_CTX structure. 754868651Skris 754968651Skris X509_STORE_CTX also has a 'flags' field which can be used 755068651Skris to customise the verify behaviour. 755168651Skris [Steve Henson] 755268651Skris 755368651Skris *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which 755468651Skris excludes S/MIME capabilities. 755568651Skris [Steve Henson] 755668651Skris 755768651Skris *) When a certificate request is read in keep a copy of the 755868651Skris original encoding of the signed data and use it when outputing 755968651Skris again. Signatures then use the original encoding rather than 756068651Skris a decoded, encoded version which may cause problems if the 756168651Skris request is improperly encoded. 756268651Skris [Steve Henson] 756368651Skris 756468651Skris *) For consistency with other BIO_puts implementations, call 756568651Skris buffer_write(b, ...) directly in buffer_puts instead of calling 756668651Skris BIO_write(b, ...). 756768651Skris 756868651Skris In BIO_puts, increment b->num_write as in BIO_write. 756968651Skris [Peter.Sylvester@EdelWeb.fr] 757068651Skris 757168651Skris *) Fix BN_mul_word for the case where the word is 0. (We have to use 757268651Skris BN_zero, we may not return a BIGNUM with an array consisting of 757368651Skris words set to zero.) 757468651Skris [Bodo Moeller] 757568651Skris 757668651Skris *) Avoid calling abort() from within the library when problems are 757768651Skris detected, except if preprocessor symbols have been defined 757868651Skris (such as REF_CHECK, BN_DEBUG etc.). 757968651Skris [Bodo Moeller] 758068651Skris 758168651Skris *) New openssl application 'rsautl'. This utility can be 758268651Skris used for low level RSA operations. DER public key 758368651Skris BIO/fp routines also added. 758468651Skris [Steve Henson] 758568651Skris 758668651Skris *) New Configure entry and patches for compiling on QNX 4. 758768651Skris [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] 758868651Skris 758968651Skris *) A demo state-machine implementation was sponsored by 759068651Skris Nuron (http://www.nuron.com/) and is now available in 759168651Skris demos/state_machine. 759268651Skris [Ben Laurie] 759368651Skris 759468651Skris *) New options added to the 'dgst' utility for signature 759568651Skris generation and verification. 759668651Skris [Steve Henson] 759768651Skris 759868651Skris *) Unrecognized PKCS#7 content types are now handled via a 759968651Skris catch all ASN1_TYPE structure. This allows unsupported 760068651Skris types to be stored as a "blob" and an application can 760168651Skris encode and decode it manually. 760268651Skris [Steve Henson] 760368651Skris 760468651Skris *) Fix various signed/unsigned issues to make a_strex.c 760568651Skris compile under VC++. 760668651Skris [Oscar Jacobsson <oscar.jacobsson@celocom.com>] 760768651Skris 760868651Skris *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct 760968651Skris length if passed a buffer. ASN1_INTEGER_to_BN failed 761068651Skris if passed a NULL BN and its argument was negative. 761168651Skris [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] 761268651Skris 761368651Skris *) Modification to PKCS#7 encoding routines to output definite 761468651Skris length encoding. Since currently the whole structures are in 761568651Skris memory there's not real point in using indefinite length 761668651Skris constructed encoding. However if OpenSSL is compiled with 761768651Skris the flag PKCS7_INDEFINITE_ENCODING the old form is used. 761868651Skris [Steve Henson] 761968651Skris 762068651Skris *) Added BIO_vprintf() and BIO_vsnprintf(). 762168651Skris [Richard Levitte] 762268651Skris 762368651Skris *) Added more prefixes to parse for in the the strings written 762468651Skris through a logging bio, to cover all the levels that are available 762568651Skris through syslog. The prefixes are now: 762668651Skris 762768651Skris PANIC, EMERG, EMR => LOG_EMERG 762868651Skris ALERT, ALR => LOG_ALERT 762968651Skris CRIT, CRI => LOG_CRIT 763068651Skris ERROR, ERR => LOG_ERR 763168651Skris WARNING, WARN, WAR => LOG_WARNING 763268651Skris NOTICE, NOTE, NOT => LOG_NOTICE 763368651Skris INFO, INF => LOG_INFO 763468651Skris DEBUG, DBG => LOG_DEBUG 763568651Skris 763668651Skris and as before, if none of those prefixes are present at the 763768651Skris beginning of the string, LOG_ERR is chosen. 763868651Skris 763968651Skris On Win32, the LOG_* levels are mapped according to this: 764068651Skris 764168651Skris LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE 764268651Skris LOG_WARNING => EVENTLOG_WARNING_TYPE 764368651Skris LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE 764468651Skris 764568651Skris [Richard Levitte] 764668651Skris 764768651Skris *) Made it possible to reconfigure with just the configuration 764868651Skris argument "reconf" or "reconfigure". The command line arguments 764968651Skris are stored in Makefile.ssl in the variable CONFIGURE_ARGS, 765068651Skris and are retrieved from there when reconfiguring. 765168651Skris [Richard Levitte] 765268651Skris 765368651Skris *) MD4 implemented. 765468651Skris [Assar Westerlund <assar@sics.se>, Richard Levitte] 765568651Skris 765668651Skris *) Add the arguments -CAfile and -CApath to the pkcs12 utility. 765768651Skris [Richard Levitte] 765868651Skris 765968651Skris *) The obj_dat.pl script was messing up the sorting of object 766068651Skris names. The reason was that it compared the quoted version 766168651Skris of strings as a result "OCSP" > "OCSP Signing" because 766268651Skris " > SPACE. Changed script to store unquoted versions of 766368651Skris names and add quotes on output. It was also omitting some 766468651Skris names from the lookup table if they were given a default 766568651Skris value (that is if SN is missing it is given the same 766668651Skris value as LN and vice versa), these are now added on the 766768651Skris grounds that if an object has a name we should be able to 766868651Skris look it up. Finally added warning output when duplicate 766968651Skris short or long names are found. 767068651Skris [Steve Henson] 767168651Skris 767268651Skris *) Changes needed for Tandem NSK. 767368651Skris [Scott Uroff <scott@xypro.com>] 767468651Skris 767568651Skris *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in 767668651Skris RSA_padding_check_SSLv23(), special padding was never detected 767768651Skris and thus the SSL 3.0/TLS 1.0 countermeasure against protocol 767868651Skris version rollback attacks was not effective. 767968651Skris 768068651Skris In s23_clnt.c, don't use special rollback-attack detection padding 768168651Skris (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the 768268651Skris client; similarly, in s23_srvr.c, don't do the rollback check if 768368651Skris SSL 2.0 is the only protocol enabled in the server. 768468651Skris [Bodo Moeller] 768568651Skris 768668651Skris *) Make it possible to get hexdumps of unprintable data with 'openssl 768768651Skris asn1parse'. By implication, the functions ASN1_parse_dump() and 768868651Skris BIO_dump_indent() are added. 768968651Skris [Richard Levitte] 769068651Skris 769168651Skris *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() 769268651Skris these print out strings and name structures based on various 769368651Skris flags including RFC2253 support and proper handling of 769468651Skris multibyte characters. Added options to the 'x509' utility 769568651Skris to allow the various flags to be set. 769668651Skris [Steve Henson] 769768651Skris 769868651Skris *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. 769968651Skris Also change the functions X509_cmp_current_time() and 770068651Skris X509_gmtime_adj() work with an ASN1_TIME structure, 770168651Skris this will enable certificates using GeneralizedTime in validity 770268651Skris dates to be checked. 770368651Skris [Steve Henson] 770468651Skris 770568651Skris *) Make the NEG_PUBKEY_BUG code (which tolerates invalid 770668651Skris negative public key encodings) on by default, 770768651Skris NO_NEG_PUBKEY_BUG can be set to disable it. 770868651Skris [Steve Henson] 770968651Skris 771068651Skris *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT 771168651Skris content octets. An i2c_ASN1_OBJECT is unnecessary because 771268651Skris the encoding can be trivially obtained from the structure. 771368651Skris [Steve Henson] 771468651Skris 771568651Skris *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), 771668651Skris not read locks (CRYPTO_r_[un]lock). 771768651Skris [Bodo Moeller] 771868651Skris 771968651Skris *) A first attempt at creating official support for shared 772068651Skris libraries through configuration. I've kept it so the 772168651Skris default is static libraries only, and the OpenSSL programs 772268651Skris are always statically linked for now, but there are 772368651Skris preparations for dynamic linking in place. 772489837Skris This has been tested on Linux and Tru64. 772568651Skris [Richard Levitte] 772668651Skris 772768651Skris *) Randomness polling function for Win9x, as described in: 772868651Skris Peter Gutmann, Software Generation of Practically Strong 772968651Skris Random Numbers. 7730291721Sjkim [Ulf M��ller] 773168651Skris 773268651Skris *) Fix so PRNG is seeded in req if using an already existing 773368651Skris DSA key. 773468651Skris [Steve Henson] 773568651Skris 773668651Skris *) New options to smime application. -inform and -outform 773768651Skris allow alternative formats for the S/MIME message including 773868651Skris PEM and DER. The -content option allows the content to be 773968651Skris specified separately. This should allow things like Netscape 774068651Skris form signing output easier to verify. 774168651Skris [Steve Henson] 774268651Skris 774368651Skris *) Fix the ASN1 encoding of tags using the 'long form'. 774468651Skris [Steve Henson] 774568651Skris 774668651Skris *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT 774768651Skris STRING types. These convert content octets to and from the 774868651Skris underlying type. The actual tag and length octets are 774968651Skris already assumed to have been read in and checked. These 775068651Skris are needed because all other string types have virtually 775168651Skris identical handling apart from the tag. By having versions 775268651Skris of the ASN1 functions that just operate on content octets 775368651Skris IMPLICIT tagging can be handled properly. It also allows 775468651Skris the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED 775568651Skris and ASN1_INTEGER are identical apart from the tag. 775668651Skris [Steve Henson] 775768651Skris 775868651Skris *) Change the handling of OID objects as follows: 775968651Skris 776068651Skris - New object identifiers are inserted in objects.txt, following 776168651Skris the syntax given in objects.README. 776268651Skris - objects.pl is used to process obj_mac.num and create a new 776368651Skris obj_mac.h. 776468651Skris - obj_dat.pl is used to create a new obj_dat.h, using the data in 776568651Skris obj_mac.h. 776668651Skris 776768651Skris This is currently kind of a hack, and the perl code in objects.pl 776868651Skris isn't very elegant, but it works as I intended. The simplest way 776968651Skris to check that it worked correctly is to look in obj_dat.h and 777068651Skris check the array nid_objs and make sure the objects haven't moved 777168651Skris around (this is important!). Additions are OK, as well as 777268651Skris consistent name changes. 777368651Skris [Richard Levitte] 777468651Skris 777568651Skris *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). 777668651Skris [Bodo Moeller] 777768651Skris 777868651Skris *) Addition of the command line parameter '-rand file' to 'openssl req'. 777968651Skris The given file adds to whatever has already been seeded into the 778068651Skris random pool through the RANDFILE configuration file option or 778168651Skris environment variable, or the default random state file. 778268651Skris [Richard Levitte] 778368651Skris 778468651Skris *) mkstack.pl now sorts each macro group into lexical order. 778568651Skris Previously the output order depended on the order the files 778668651Skris appeared in the directory, resulting in needless rewriting 778768651Skris of safestack.h . 778868651Skris [Steve Henson] 778968651Skris 779068651Skris *) Patches to make OpenSSL compile under Win32 again. Mostly 779168651Skris work arounds for the VC++ problem that it treats func() as 779268651Skris func(void). Also stripped out the parts of mkdef.pl that 779368651Skris added extra typesafe functions: these no longer exist. 779468651Skris [Steve Henson] 779568651Skris 779668651Skris *) Reorganisation of the stack code. The macros are now all 779768651Skris collected in safestack.h . Each macro is defined in terms of 779868651Skris a "stack macro" of the form SKM_<name>(type, a, b). The 779968651Skris DEBUG_SAFESTACK is now handled in terms of function casts, 780068651Skris this has the advantage of retaining type safety without the 780168651Skris use of additional functions. If DEBUG_SAFESTACK is not defined 780268651Skris then the non typesafe macros are used instead. Also modified the 780368651Skris mkstack.pl script to handle the new form. Needs testing to see 780468651Skris if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK 780568651Skris the default if no major problems. Similar behaviour for ASN1_SET_OF 780668651Skris and PKCS12_STACK_OF. 780768651Skris [Steve Henson] 780868651Skris 780968651Skris *) When some versions of IIS use the 'NET' form of private key the 781068651Skris key derivation algorithm is different. Normally MD5(password) is 781168651Skris used as a 128 bit RC4 key. In the modified case 781268651Skris MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some 781368651Skris new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same 781468651Skris as the old Netscape_RSA functions except they have an additional 781568651Skris 'sgckey' parameter which uses the modified algorithm. Also added 781668651Skris an -sgckey command line option to the rsa utility. Thanks to 781768651Skris Adrian Peck <bertie@ncipher.com> for posting details of the modified 781868651Skris algorithm to openssl-dev. 781968651Skris [Steve Henson] 782068651Skris 782168651Skris *) The evp_local.h macros were using 'c.##kname' which resulted in 782268651Skris invalid expansion on some systems (SCO 5.0.5 for example). 782368651Skris Corrected to 'c.kname'. 782468651Skris [Phillip Porch <root@theporch.com>] 782568651Skris 782668651Skris *) New X509_get1_email() and X509_REQ_get1_email() functions that return 782768651Skris a STACK of email addresses from a certificate or request, these look 782868651Skris in the subject name and the subject alternative name extensions and 782968651Skris omit any duplicate addresses. 783068651Skris [Steve Henson] 783168651Skris 783268651Skris *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. 783368651Skris This makes DSA verification about 2 % faster. 783468651Skris [Bodo Moeller] 783568651Skris 783668651Skris *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 783768651Skris (meaning that now 2^5 values will be precomputed, which is only 4 KB 783868651Skris plus overhead for 1024 bit moduli). 783968651Skris This makes exponentiations about 0.5 % faster for 1024 bit 784068651Skris exponents (as measured by "openssl speed rsa2048"). 784168651Skris [Bodo Moeller] 784268651Skris 784368651Skris *) Rename memory handling macros to avoid conflicts with other 784468651Skris software: 784568651Skris Malloc => OPENSSL_malloc 784668651Skris Malloc_locked => OPENSSL_malloc_locked 784768651Skris Realloc => OPENSSL_realloc 784868651Skris Free => OPENSSL_free 784968651Skris [Richard Levitte] 785068651Skris 785168651Skris *) New function BN_mod_exp_mont_word for small bases (roughly 15% 785268651Skris faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). 785368651Skris [Bodo Moeller] 785468651Skris 785568651Skris *) CygWin32 support. 785668651Skris [John Jarvie <jjarvie@newsguy.com>] 785768651Skris 785868651Skris *) The type-safe stack code has been rejigged. It is now only compiled 785968651Skris in when OpenSSL is configured with the DEBUG_SAFESTACK option and 786068651Skris by default all type-specific stack functions are "#define"d back to 786168651Skris standard stack functions. This results in more streamlined output 786268651Skris but retains the type-safety checking possibilities of the original 786368651Skris approach. 786468651Skris [Geoff Thorpe] 786568651Skris 786668651Skris *) The STACK code has been cleaned up, and certain type declarations 786768651Skris that didn't make a lot of sense have been brought in line. This has 786868651Skris also involved a cleanup of sorts in safestack.h to more correctly 786968651Skris map type-safe stack functions onto their plain stack counterparts. 787068651Skris This work has also resulted in a variety of "const"ifications of 787168651Skris lots of the code, especially "_cmp" operations which should normally 787268651Skris be prototyped with "const" parameters anyway. 787368651Skris [Geoff Thorpe] 787468651Skris 787568651Skris *) When generating bytes for the first time in md_rand.c, 'stir the pool' 787668651Skris by seeding with STATE_SIZE dummy bytes (with zero entropy count). 787768651Skris (The PRNG state consists of two parts, the large pool 'state' and 'md', 787868651Skris where all of 'md' is used each time the PRNG is used, but 'state' 787968651Skris is used only indexed by a cyclic counter. As entropy may not be 788068651Skris well distributed from the beginning, 'md' is important as a 788168651Skris chaining variable. However, the output function chains only half 788268651Skris of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains 788368651Skris all of 'md', and seeding with STATE_SIZE dummy bytes will result 788468651Skris in all of 'state' being rewritten, with the new values depending 788568651Skris on virtually all of 'md'. This overcomes the 80 bit limitation.) 788668651Skris [Bodo Moeller] 788768651Skris 788868651Skris *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when 788968651Skris the handshake is continued after ssl_verify_cert_chain(); 789068651Skris otherwise, if SSL_VERIFY_NONE is set, remaining error codes 789168651Skris can lead to 'unexplainable' connection aborts later. 789268651Skris [Bodo Moeller; problem tracked down by Lutz Jaenicke] 789368651Skris 789468651Skris *) Major EVP API cipher revision. 789568651Skris Add hooks for extra EVP features. This allows various cipher 789668651Skris parameters to be set in the EVP interface. Support added for variable 789768651Skris key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and 789868651Skris setting of RC2 and RC5 parameters. 789968651Skris 790068651Skris Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length 790168651Skris ciphers. 790268651Skris 790368651Skris Remove lots of duplicated code from the EVP library. For example *every* 790468651Skris cipher init() function handles the 'iv' in the same way according to the 790568651Skris cipher mode. They also all do nothing if the 'key' parameter is NULL and 790668651Skris for CFB and OFB modes they zero ctx->num. 790768651Skris 790868651Skris New functionality allows removal of S/MIME code RC2 hack. 790968651Skris 791068651Skris Most of the routines have the same form and so can be declared in terms 791168651Skris of macros. 791268651Skris 791368651Skris By shifting this to the top level EVP_CipherInit() it can be removed from 791468651Skris all individual ciphers. If the cipher wants to handle IVs or keys 791568651Skris differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT 791668651Skris flags. 791768651Skris 791868651Skris Change lots of functions like EVP_EncryptUpdate() to now return a 791968651Skris value: although software versions of the algorithms cannot fail 792068651Skris any installed hardware versions can. 792168651Skris [Steve Henson] 792268651Skris 792368651Skris *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if 792468651Skris this option is set, tolerate broken clients that send the negotiated 792568651Skris protocol version number instead of the requested protocol version 792668651Skris number. 792768651Skris [Bodo Moeller] 792868651Skris 792968651Skris *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; 793068651Skris i.e. non-zero for export ciphersuites, zero otherwise. 793168651Skris Previous versions had this flag inverted, inconsistent with 793268651Skris rsa_tmp_cb (..._TMP_RSA_CB). 793368651Skris [Bodo Moeller; problem reported by Amit Chopra] 793468651Skris 793568651Skris *) Add missing DSA library text string. Work around for some IIS 793668651Skris key files with invalid SEQUENCE encoding. 793768651Skris [Steve Henson] 793868651Skris 793968651Skris *) Add a document (doc/standards.txt) that list all kinds of standards 794068651Skris and so on that are implemented in OpenSSL. 794168651Skris [Richard Levitte] 794268651Skris 794368651Skris *) Enhance c_rehash script. Old version would mishandle certificates 794468651Skris with the same subject name hash and wouldn't handle CRLs at all. 794568651Skris Added -fingerprint option to crl utility, to support new c_rehash 794668651Skris features. 794768651Skris [Steve Henson] 794868651Skris 794968651Skris *) Eliminate non-ANSI declarations in crypto.h and stack.h. 7950291721Sjkim [Ulf M��ller] 795168651Skris 795268651Skris *) Fix for SSL server purpose checking. Server checking was 795368651Skris rejecting certificates which had extended key usage present 795468651Skris but no ssl client purpose. 795568651Skris [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] 795668651Skris 795768651Skris *) Make PKCS#12 code work with no password. The PKCS#12 spec 795868651Skris is a little unclear about how a blank password is handled. 795968651Skris Since the password in encoded as a BMPString with terminating 796068651Skris double NULL a zero length password would end up as just the 796168651Skris double NULL. However no password at all is different and is 796268651Skris handled differently in the PKCS#12 key generation code. NS 796368651Skris treats a blank password as zero length. MSIE treats it as no 796468651Skris password on export: but it will try both on import. We now do 796568651Skris the same: PKCS12_parse() tries zero length and no password if 796668651Skris the password is set to "" or NULL (NULL is now a valid password: 796768651Skris it wasn't before) as does the pkcs12 application. 796868651Skris [Steve Henson] 796968651Skris 797068651Skris *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use 797168651Skris perror when PEM_read_bio_X509_REQ fails, the error message must 797268651Skris be obtained from the error queue. 797368651Skris [Bodo Moeller] 797468651Skris 797568651Skris *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing 797668651Skris it in ERR_remove_state if appropriate, and change ERR_get_state 797768651Skris accordingly to avoid race conditions (this is necessary because 797868651Skris thread_hash is no longer constant once set). 797968651Skris [Bodo Moeller] 798068651Skris 798168651Skris *) Bugfix for linux-elf makefile.one. 7982291721Sjkim [Ulf M��ller] 798368651Skris 798468651Skris *) RSA_get_default_method() will now cause a default 798568651Skris RSA_METHOD to be chosen if one doesn't exist already. 798668651Skris Previously this was only set during a call to RSA_new() 798768651Skris or RSA_new_method(NULL) meaning it was possible for 798868651Skris RSA_get_default_method() to return NULL. 798968651Skris [Geoff Thorpe] 799068651Skris 799168651Skris *) Added native name translation to the existing DSO code 799268651Skris that will convert (if the flag to do so is set) filenames 799368651Skris that are sufficiently small and have no path information 799468651Skris into a canonical native form. Eg. "blah" converted to 799568651Skris "libblah.so" or "blah.dll" etc. 799668651Skris [Geoff Thorpe] 799768651Skris 799868651Skris *) New function ERR_error_string_n(e, buf, len) which is like 799968651Skris ERR_error_string(e, buf), but writes at most 'len' bytes 800068651Skris including the 0 terminator. For ERR_error_string_n, 'buf' 800168651Skris may not be NULL. 800268651Skris [Damien Miller <djm@mindrot.org>, Bodo Moeller] 800368651Skris 800468651Skris *) CONF library reworked to become more general. A new CONF 800568651Skris configuration file reader "class" is implemented as well as a 800668651Skris new functions (NCONF_*, for "New CONF") to handle it. The now 800768651Skris old CONF_* functions are still there, but are reimplemented to 800868651Skris work in terms of the new functions. Also, a set of functions 800968651Skris to handle the internal storage of the configuration data is 801068651Skris provided to make it easier to write new configuration file 801168651Skris reader "classes" (I can definitely see something reading a 801268651Skris configuration file in XML format, for example), called _CONF_*, 801368651Skris or "the configuration storage API"... 801468651Skris 801568651Skris The new configuration file reading functions are: 801668651Skris 801768651Skris NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, 801868651Skris NCONF_get_section, NCONF_get_string, NCONF_get_numbre 801968651Skris 802068651Skris NCONF_default, NCONF_WIN32 802168651Skris 802268651Skris NCONF_dump_fp, NCONF_dump_bio 802368651Skris 802468651Skris NCONF_default and NCONF_WIN32 are method (or "class") choosers, 802568651Skris NCONF_new creates a new CONF object. This works in the same way 802668651Skris as other interfaces in OpenSSL, like the BIO interface. 802768651Skris NCONF_dump_* dump the internal storage of the configuration file, 802868651Skris which is useful for debugging. All other functions take the same 802968651Skris arguments as the old CONF_* functions wth the exception of the 803068651Skris first that must be a `CONF *' instead of a `LHASH *'. 803168651Skris 803268651Skris To make it easer to use the new classes with the old CONF_* functions, 803368651Skris the function CONF_set_default_method is provided. 803468651Skris [Richard Levitte] 803568651Skris 803668651Skris *) Add '-tls1' option to 'openssl ciphers', which was already 803768651Skris mentioned in the documentation but had not been implemented. 803868651Skris (This option is not yet really useful because even the additional 803968651Skris experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) 804068651Skris [Bodo Moeller] 804168651Skris 804268651Skris *) Initial DSO code added into libcrypto for letting OpenSSL (and 804368651Skris OpenSSL-based applications) load shared libraries and bind to 804468651Skris them in a portable way. 804568651Skris [Geoff Thorpe, with contributions from Richard Levitte] 804668651Skris 804759191Skris Changes between 0.9.5 and 0.9.5a [1 Apr 2000] 804859191Skris 804959191Skris *) Make sure _lrotl and _lrotr are only used with MSVC. 805059191Skris 805159191Skris *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status 805259191Skris (the default implementation of RAND_status). 805359191Skris 805459191Skris *) Rename openssl x509 option '-crlext', which was added in 0.9.5, 805559191Skris to '-clrext' (= clear extensions), as intended and documented. 805659191Skris [Bodo Moeller; inconsistency pointed out by Michael Attili 805759191Skris <attili@amaxo.com>] 805859191Skris 805959191Skris *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length 806059191Skris was larger than the MD block size. 806159191Skris [Steve Henson, pointed out by Yost William <YostW@tce.com>] 806259191Skris 806359191Skris *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument 806459191Skris fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() 806559191Skris using the passed key: if the passed key was a private key the result 806659191Skris of X509_print(), for example, would be to print out all the private key 806759191Skris components. 806859191Skris [Steve Henson] 806959191Skris 807059191Skris *) des_quad_cksum() byte order bug fix. 8071291721Sjkim [Ulf M��ller, using the problem description in krb4-0.9.7, where 807259191Skris the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] 807359191Skris 807459191Skris *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly 807559191Skris discouraged. 807659191Skris [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] 807759191Skris 807859191Skris *) For easily testing in shell scripts whether some command 807959191Skris 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' 808059191Skris returns with exit code 0 iff no command of the given name is available. 808159191Skris 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, 808259191Skris the output goes to stdout and nothing is printed to stderr. 808359191Skris Additional arguments are always ignored. 808459191Skris 808559191Skris Since for each cipher there is a command of the same name, 808659191Skris the 'no-cipher' compilation switches can be tested this way. 808759191Skris 808859191Skris ('openssl no-XXX' is not able to detect pseudo-commands such 808959191Skris as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) 809059191Skris [Bodo Moeller] 809159191Skris 809259191Skris *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. 809359191Skris [Bodo Moeller] 809459191Skris 809559191Skris *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE 809659191Skris is set; it will be thrown away anyway because each handshake creates 809759191Skris its own key. 809859191Skris ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition 809959191Skris to parameters -- in previous versions (since OpenSSL 0.9.3) the 810059191Skris 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining 810159191Skris you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. 810259191Skris [Bodo Moeller] 810359191Skris 810459191Skris *) New s_client option -ign_eof: EOF at stdin is ignored, and 810559191Skris 'Q' and 'R' lose their special meanings (quit/renegotiate). 810659191Skris This is part of what -quiet does; unlike -quiet, -ign_eof 810759191Skris does not suppress any output. 810859191Skris [Richard Levitte] 810959191Skris 811059191Skris *) Add compatibility options to the purpose and trust code. The 811159191Skris purpose X509_PURPOSE_ANY is "any purpose" which automatically 811259191Skris accepts a certificate or CA, this was the previous behaviour, 811359191Skris with all the associated security issues. 811459191Skris 811559191Skris X509_TRUST_COMPAT is the old trust behaviour: only and 811659191Skris automatically trust self signed roots in certificate store. A 811759191Skris new trust setting X509_TRUST_DEFAULT is used to specify that 811859191Skris a purpose has no associated trust setting and it should instead 811959191Skris use the value in the default purpose. 812059191Skris [Steve Henson] 812159191Skris 812259191Skris *) Fix the PKCS#8 DSA private key code so it decodes keys again 812359191Skris and fix a memory leak. 812459191Skris [Steve Henson] 812559191Skris 812659191Skris *) In util/mkerr.pl (which implements 'make errors'), preserve 812759191Skris reason strings from the previous version of the .c file, as 812859191Skris the default to have only downcase letters (and digits) in 812959191Skris automatically generated reasons codes is not always appropriate. 813059191Skris [Bodo Moeller] 813159191Skris 813259191Skris *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table 813359191Skris using strerror. Previously, ERR_reason_error_string() returned 813459191Skris library names as reason strings for SYSerr; but SYSerr is a special 813559191Skris case where small numbers are errno values, not library numbers. 813659191Skris [Bodo Moeller] 813759191Skris 813859191Skris *) Add '-dsaparam' option to 'openssl dhparam' application. This 813959191Skris converts DSA parameters into DH parameters. (When creating parameters, 814059191Skris DSA_generate_parameters is used.) 814159191Skris [Bodo Moeller] 814259191Skris 814359191Skris *) Include 'length' (recommended exponent length) in C code generated 814459191Skris by 'openssl dhparam -C'. 814559191Skris [Bodo Moeller] 814659191Skris 814759191Skris *) The second argument to set_label in perlasm was already being used 814859191Skris so couldn't be used as a "file scope" flag. Moved to third argument 814959191Skris which was free. 815059191Skris [Steve Henson] 815159191Skris 815259191Skris *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes 815359191Skris instead of RAND_bytes for encryption IVs and salts. 815459191Skris [Bodo Moeller] 815559191Skris 815659191Skris *) Include RAND_status() into RAND_METHOD instead of implementing 815759191Skris it only for md_rand.c Otherwise replacing the PRNG by calling 815859191Skris RAND_set_rand_method would be impossible. 815959191Skris [Bodo Moeller] 816059191Skris 816159191Skris *) Don't let DSA_generate_key() enter an infinite loop if the random 816259191Skris number generation fails. 816359191Skris [Bodo Moeller] 816459191Skris 816559191Skris *) New 'rand' application for creating pseudo-random output. 816659191Skris [Bodo Moeller] 816759191Skris 816859191Skris *) Added configuration support for Linux/IA64 816959191Skris [Rolf Haberrecker <rolf@suse.de>] 817059191Skris 817159191Skris *) Assembler module support for Mingw32. 8172291721Sjkim [Ulf M��ller] 817359191Skris 817459191Skris *) Shared library support for HPUX (in shlib/). 817559191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] 817659191Skris 817759191Skris *) Shared library support for Solaris gcc. 817859191Skris [Lutz Behnke <behnke@trustcenter.de>] 817959191Skris 818059191Skris Changes between 0.9.4 and 0.9.5 [28 Feb 2000] 818159191Skris 818259191Skris *) PKCS7_encrypt() was adding text MIME headers twice because they 818359191Skris were added manually and by SMIME_crlf_copy(). 818459191Skris [Steve Henson] 818559191Skris 818659191Skris *) In bntest.c don't call BN_rand with zero bits argument. 818759191Skris [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] 818859191Skris 818959191Skris *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] 819059191Skris case was implemented. This caused BN_div_recp() to fail occasionally. 8191291721Sjkim [Ulf M��ller] 819259191Skris 819359191Skris *) Add an optional second argument to the set_label() in the perl 819459191Skris assembly language builder. If this argument exists and is set 819559191Skris to 1 it signals that the assembler should use a symbol whose 819659191Skris scope is the entire file, not just the current function. This 819759191Skris is needed with MASM which uses the format label:: for this scope. 819859191Skris [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] 819959191Skris 820059191Skris *) Change the ASN1 types so they are typedefs by default. Before 820159191Skris almost all types were #define'd to ASN1_STRING which was causing 820259191Skris STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) 820359191Skris for example. 820459191Skris [Steve Henson] 820559191Skris 820659191Skris *) Change names of new functions to the new get1/get0 naming 820759191Skris convention: After 'get1', the caller owns a reference count 820859191Skris and has to call ..._free; 'get0' returns a pointer to some 820959191Skris data structure without incrementing reference counters. 821059191Skris (Some of the existing 'get' functions increment a reference 821159191Skris counter, some don't.) 821259191Skris Similarly, 'set1' and 'add1' functions increase reference 821359191Skris counters or duplicate objects. 821459191Skris [Steve Henson] 821559191Skris 821659191Skris *) Allow for the possibility of temp RSA key generation failure: 821759191Skris the code used to assume it always worked and crashed on failure. 821859191Skris [Steve Henson] 821959191Skris 822059191Skris *) Fix potential buffer overrun problem in BIO_printf(). 8221291721Sjkim [Ulf M��ller, using public domain code by Patrick Powell; problem 822259191Skris pointed out by David Sacerdote <das33@cornell.edu>] 822359191Skris 822459191Skris *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions 822559191Skris RAND_egd() and RAND_status(). In the command line application, 822659191Skris the EGD socket can be specified like a seed file using RANDFILE 822759191Skris or -rand. 8228291721Sjkim [Ulf M��ller] 822959191Skris 823059191Skris *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. 823159191Skris Some CAs (e.g. Verisign) distribute certificates in this form. 823259191Skris [Steve Henson] 823359191Skris 823459191Skris *) Remove the SSL_ALLOW_ADH compile option and set the default cipher 823559191Skris list to exclude them. This means that no special compilation option 823659191Skris is needed to use anonymous DH: it just needs to be included in the 823759191Skris cipher list. 823859191Skris [Steve Henson] 823959191Skris 824059191Skris *) Change the EVP_MD_CTX_type macro so its meaning consistent with 824159191Skris EVP_MD_type. The old functionality is available in a new macro called 824259191Skris EVP_MD_md(). Change code that uses it and update docs. 824359191Skris [Steve Henson] 824459191Skris 824559191Skris *) ..._ctrl functions now have corresponding ..._callback_ctrl functions 824659191Skris where the 'void *' argument is replaced by a function pointer argument. 824759191Skris Previously 'void *' was abused to point to functions, which works on 824859191Skris many platforms, but is not correct. As these functions are usually 824959191Skris called by macros defined in OpenSSL header files, most source code 825059191Skris should work without changes. 825159191Skris [Richard Levitte] 825259191Skris 825359191Skris *) <openssl/opensslconf.h> (which is created by Configure) now contains 825459191Skris sections with information on -D... compiler switches used for 825559191Skris compiling the library so that applications can see them. To enable 825659191Skris one of these sections, a pre-processor symbol OPENSSL_..._DEFINES 825759191Skris must be defined. E.g., 825859191Skris #define OPENSSL_ALGORITHM_DEFINES 825959191Skris #include <openssl/opensslconf.h> 826059191Skris defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. 8261291721Sjkim [Richard Levitte, Ulf and Bodo M��ller] 826259191Skris 826359191Skris *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS 826459191Skris record layer. 826559191Skris [Bodo Moeller] 826659191Skris 826759191Skris *) Change the 'other' type in certificate aux info to a STACK_OF 826859191Skris X509_ALGOR. Although not an AlgorithmIdentifier as such it has 826959191Skris the required ASN1 format: arbitrary types determined by an OID. 827059191Skris [Steve Henson] 827159191Skris 827259191Skris *) Add some PEM_write_X509_REQ_NEW() functions and a command line 827359191Skris argument to 'req'. This is not because the function is newer or 827459191Skris better than others it just uses the work 'NEW' in the certificate 827559191Skris request header lines. Some software needs this. 827659191Skris [Steve Henson] 827759191Skris 827859191Skris *) Reorganise password command line arguments: now passwords can be 827959191Skris obtained from various sources. Delete the PEM_cb function and make 828059191Skris it the default behaviour: i.e. if the callback is NULL and the 828159191Skris usrdata argument is not NULL interpret it as a null terminated pass 828259191Skris phrase. If usrdata and the callback are NULL then the pass phrase 828359191Skris is prompted for as usual. 828459191Skris [Steve Henson] 828559191Skris 828659191Skris *) Add support for the Compaq Atalla crypto accelerator. If it is installed, 828759191Skris the support is automatically enabled. The resulting binaries will 828859191Skris autodetect the card and use it if present. 828959191Skris [Ben Laurie and Compaq Inc.] 829059191Skris 829159191Skris *) Work around for Netscape hang bug. This sends certificate request 829259191Skris and server done in one record. Since this is perfectly legal in the 829359191Skris SSL/TLS protocol it isn't a "bug" option and is on by default. See 829459191Skris the bugs/SSLv3 entry for more info. 829559191Skris [Steve Henson] 829659191Skris 829759191Skris *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. 829859191Skris [Andy Polyakov] 829959191Skris 830059191Skris *) Add -rand argument to smime and pkcs12 applications and read/write 830159191Skris of seed file. 830259191Skris [Steve Henson] 830359191Skris 830459191Skris *) New 'passwd' tool for crypt(3) and apr1 password hashes. 830559191Skris [Bodo Moeller] 830659191Skris 830759191Skris *) Add command line password options to the remaining applications. 830859191Skris [Steve Henson] 830959191Skris 831059191Skris *) Bug fix for BN_div_recp() for numerators with an even number of 831159191Skris bits. 8312291721Sjkim [Ulf M��ller] 831359191Skris 831459191Skris *) More tests in bntest.c, and changed test_bn output. 8315291721Sjkim [Ulf M��ller] 831659191Skris 831759191Skris *) ./config recognizes MacOS X now. 831859191Skris [Andy Polyakov] 831959191Skris 832059191Skris *) Bug fix for BN_div() when the first words of num and divsor are 832159191Skris equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). 8322291721Sjkim [Ulf M��ller] 832359191Skris 832459191Skris *) Add support for various broken PKCS#8 formats, and command line 832559191Skris options to produce them. 832659191Skris [Steve Henson] 832759191Skris 832859191Skris *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to 832959191Skris get temporary BIGNUMs from a BN_CTX. 8330291721Sjkim [Ulf M��ller] 833159191Skris 833259191Skris *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() 833359191Skris for p == 0. 8334291721Sjkim [Ulf M��ller] 833559191Skris 833659191Skris *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and 833759191Skris include a #define from the old name to the new. The original intent 833859191Skris was that statically linked binaries could for example just call 833959191Skris SSLeay_add_all_ciphers() to just add ciphers to the table and not 834059191Skris link with digests. This never worked becayse SSLeay_add_all_digests() 834159191Skris and SSLeay_add_all_ciphers() were in the same source file so calling 834259191Skris one would link with the other. They are now in separate source files. 834359191Skris [Steve Henson] 834459191Skris 834559191Skris *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. 834659191Skris [Steve Henson] 834759191Skris 834859191Skris *) Use a less unusual form of the Miller-Rabin primality test (it used 834959191Skris a binary algorithm for exponentiation integrated into the Miller-Rabin 835059191Skris loop, our standard modexp algorithms are faster). 835159191Skris [Bodo Moeller] 835259191Skris 835359191Skris *) Support for the EBCDIC character set completed. 835459191Skris [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] 835559191Skris 835659191Skris *) Source code cleanups: use const where appropriate, eliminate casts, 835759191Skris use void * instead of char * in lhash. 8358291721Sjkim [Ulf M��ller] 835959191Skris 836059191Skris *) Bugfix: ssl3_send_server_key_exchange was not restartable 836159191Skris (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of 836259191Skris this the server could overwrite ephemeral keys that the client 836359191Skris has already seen). 836459191Skris [Bodo Moeller] 836559191Skris 836659191Skris *) Turn DSA_is_prime into a macro that calls BN_is_prime, 836759191Skris using 50 iterations of the Rabin-Miller test. 836859191Skris 836959191Skris DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 837059191Skris iterations of the Rabin-Miller test as required by the appendix 837159191Skris to FIPS PUB 186[-1]) instead of DSA_is_prime. 837259191Skris As BN_is_prime_fasttest includes trial division, DSA parameter 837359191Skris generation becomes much faster. 837459191Skris 837559191Skris This implies a change for the callback functions in DSA_is_prime 837659191Skris and DSA_generate_parameters: The callback function is called once 837759191Skris for each positive witness in the Rabin-Miller test, not just 837859191Skris occasionally in the inner loop; and the parameters to the 837959191Skris callback function now provide an iteration count for the outer 838059191Skris loop rather than for the current invocation of the inner loop. 838159191Skris DSA_generate_parameters additionally can call the callback 838259191Skris function with an 'iteration count' of -1, meaning that a 838359191Skris candidate has passed the trial division test (when q is generated 838459191Skris from an application-provided seed, trial division is skipped). 838559191Skris [Bodo Moeller] 838659191Skris 838759191Skris *) New function BN_is_prime_fasttest that optionally does trial 838859191Skris division before starting the Rabin-Miller test and has 838959191Skris an additional BN_CTX * argument (whereas BN_is_prime always 839059191Skris has to allocate at least one BN_CTX). 839159191Skris 'callback(1, -1, cb_arg)' is called when a number has passed the 839259191Skris trial division stage. 839359191Skris [Bodo Moeller] 839459191Skris 839559191Skris *) Fix for bug in CRL encoding. The validity dates weren't being handled 839659191Skris as ASN1_TIME. 839759191Skris [Steve Henson] 839859191Skris 839959191Skris *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. 840059191Skris [Steve Henson] 840159191Skris 840259191Skris *) New function BN_pseudo_rand(). 8403291721Sjkim [Ulf M��ller] 840459191Skris 840559191Skris *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) 840659191Skris bignum version of BN_from_montgomery() with the working code from 840759191Skris SSLeay 0.9.0 (the word based version is faster anyway), and clean up 840859191Skris the comments. 8409291721Sjkim [Ulf M��ller] 841059191Skris 841159191Skris *) Avoid a race condition in s2_clnt.c (function get_server_hello) that 841259191Skris made it impossible to use the same SSL_SESSION data structure in 841359191Skris SSL2 clients in multiple threads. 841459191Skris [Bodo Moeller] 841559191Skris 841659191Skris *) The return value of RAND_load_file() no longer counts bytes obtained 841759191Skris by stat(). RAND_load_file(..., -1) is new and uses the complete file 841859191Skris to seed the PRNG (previously an explicit byte count was required). 8419291721Sjkim [Ulf M��ller, Bodo M��ller] 842059191Skris 842159191Skris *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes 842259191Skris used (char *) instead of (void *) and had casts all over the place. 842359191Skris [Steve Henson] 842459191Skris 842559191Skris *) Make BN_generate_prime() return NULL on error if ret!=NULL. 8426291721Sjkim [Ulf M��ller] 842759191Skris 842859191Skris *) Retain source code compatibility for BN_prime_checks macro: 842959191Skris BN_is_prime(..., BN_prime_checks, ...) now uses 843059191Skris BN_prime_checks_for_size to determine the appropriate number of 843159191Skris Rabin-Miller iterations. 8432291721Sjkim [Ulf M��ller] 843359191Skris 843459191Skris *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to 843559191Skris DH_CHECK_P_NOT_SAFE_PRIME. 843659191Skris (Check if this is true? OpenPGP calls them "strong".) 8437291721Sjkim [Ulf M��ller] 843859191Skris 843959191Skris *) Merge the functionality of "dh" and "gendh" programs into a new program 844059191Skris "dhparam". The old programs are retained for now but will handle DH keys 844159191Skris (instead of parameters) in future. 844259191Skris [Steve Henson] 844359191Skris 844459191Skris *) Make the ciphers, s_server and s_client programs check the return values 844559191Skris when a new cipher list is set. 844659191Skris [Steve Henson] 844759191Skris 844859191Skris *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit 844959191Skris ciphers. Before when the 56bit ciphers were enabled the sorting was 845059191Skris wrong. 845159191Skris 845259191Skris The syntax for the cipher sorting has been extended to support sorting by 845359191Skris cipher-strength (using the strength_bits hard coded in the tables). 845459191Skris The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). 845559191Skris 845659191Skris Fix a bug in the cipher-command parser: when supplying a cipher command 845759191Skris string with an "undefined" symbol (neither command nor alphanumeric 845859191Skris [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now 845959191Skris an error is flagged. 846059191Skris 846159191Skris Due to the strength-sorting extension, the code of the 846259191Skris ssl_create_cipher_list() function was completely rearranged. I hope that 846359191Skris the readability was also increased :-) 846459191Skris [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] 846559191Skris 846659191Skris *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 846759191Skris for the first serial number and places 2 in the serial number file. This 846859191Skris avoids problems when the root CA is created with serial number zero and 846959191Skris the first user certificate has the same issuer name and serial number 847059191Skris as the root CA. 847159191Skris [Steve Henson] 847259191Skris 847359191Skris *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses 847459191Skris the new code. Add documentation for this stuff. 847559191Skris [Steve Henson] 847659191Skris 847759191Skris *) Changes to X509_ATTRIBUTE utilities. These have been renamed from 847859191Skris X509_*() to X509at_*() on the grounds that they don't handle X509 847959191Skris structures and behave in an analagous way to the X509v3 functions: 848059191Skris they shouldn't be called directly but wrapper functions should be used 848159191Skris instead. 848259191Skris 848359191Skris So we also now have some wrapper functions that call the X509at functions 848459191Skris when passed certificate requests. (TO DO: similar things can be done with 848559191Skris PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other 848659191Skris things. Some of these need some d2i or i2d and print functionality 848759191Skris because they handle more complex structures.) 848859191Skris [Steve Henson] 848959191Skris 849059191Skris *) Add missing #ifndefs that caused missing symbols when building libssl 849159191Skris as a shared library without RSA. Use #ifndef NO_SSL2 instead of 849259191Skris NO_RSA in ssl/s2*.c. 8493291721Sjkim [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf M��ller] 849459191Skris 849559191Skris *) Precautions against using the PRNG uninitialized: RAND_bytes() now 849659191Skris has a return value which indicates the quality of the random data 849759191Skris (1 = ok, 0 = not seeded). Also an error is recorded on the thread's 849859191Skris error queue. New function RAND_pseudo_bytes() generates output that is 849959191Skris guaranteed to be unique but not unpredictable. RAND_add is like 850059191Skris RAND_seed, but takes an extra argument for an entropy estimate 850159191Skris (RAND_seed always assumes full entropy). 8502291721Sjkim [Ulf M��ller] 850359191Skris 850459191Skris *) Do more iterations of Rabin-Miller probable prime test (specifically, 850559191Skris 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes 850659191Skris instead of only 2 for all lengths; see BN_prime_checks_for_size definition 850759191Skris in crypto/bn/bn_prime.c for the complete table). This guarantees a 850859191Skris false-positive rate of at most 2^-80 for random input. 850959191Skris [Bodo Moeller] 851059191Skris 851159191Skris *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. 851259191Skris [Bodo Moeller] 851359191Skris 851459191Skris *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain 851559191Skris in the 0.9.5 release), this returns the chain 851659191Skris from an X509_CTX structure with a dup of the stack and all 851759191Skris the X509 reference counts upped: so the stack will exist 851859191Skris after X509_CTX_cleanup() has been called. Modify pkcs12.c 851959191Skris to use this. 852059191Skris 852159191Skris Also make SSL_SESSION_print() print out the verify return 852259191Skris code. 852359191Skris [Steve Henson] 852459191Skris 852559191Skris *) Add manpage for the pkcs12 command. Also change the default 852659191Skris behaviour so MAC iteration counts are used unless the new 852759191Skris -nomaciter option is used. This improves file security and 852859191Skris only older versions of MSIE (4.0 for example) need it. 852959191Skris [Steve Henson] 853059191Skris 853159191Skris *) Honor the no-xxx Configure options when creating .DEF files. 8532291721Sjkim [Ulf M��ller] 853359191Skris 853459191Skris *) Add PKCS#10 attributes to field table: challengePassword, 853559191Skris unstructuredName and unstructuredAddress. These are taken from 853659191Skris draft PKCS#9 v2.0 but are compatible with v1.2 provided no 853759191Skris international characters are used. 853859191Skris 853959191Skris More changes to X509_ATTRIBUTE code: allow the setting of types 854059191Skris based on strings. Remove the 'loc' parameter when adding 854159191Skris attributes because these will be a SET OF encoding which is sorted 854259191Skris in ASN1 order. 854359191Skris [Steve Henson] 854459191Skris 854559191Skris *) Initial changes to the 'req' utility to allow request generation 854659191Skris automation. This will allow an application to just generate a template 854759191Skris file containing all the field values and have req construct the 854859191Skris request. 854959191Skris 855059191Skris Initial support for X509_ATTRIBUTE handling. Stacks of these are 855159191Skris used all over the place including certificate requests and PKCS#7 855259191Skris structures. They are currently handled manually where necessary with 855359191Skris some primitive wrappers for PKCS#7. The new functions behave in a 855459191Skris manner analogous to the X509 extension functions: they allow 855559191Skris attributes to be looked up by NID and added. 855659191Skris 855759191Skris Later something similar to the X509V3 code would be desirable to 855859191Skris automatically handle the encoding, decoding and printing of the 855959191Skris more complex types. The string types like challengePassword can 856059191Skris be handled by the string table functions. 856159191Skris 856259191Skris Also modified the multi byte string table handling. Now there is 856359191Skris a 'global mask' which masks out certain types. The table itself 856459191Skris can use the flag STABLE_NO_MASK to ignore the mask setting: this 856559191Skris is useful when for example there is only one permissible type 856659191Skris (as in countryName) and using the mask might result in no valid 856759191Skris types at all. 856859191Skris [Steve Henson] 856959191Skris 857059191Skris *) Clean up 'Finished' handling, and add functions SSL_get_finished and 857159191Skris SSL_get_peer_finished to allow applications to obtain the latest 857259191Skris Finished messages sent to the peer or expected from the peer, 857359191Skris respectively. (SSL_get_peer_finished is usually the Finished message 857459191Skris actually received from the peer, otherwise the protocol will be aborted.) 857559191Skris 857659191Skris As the Finished message are message digests of the complete handshake 857759191Skris (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can 857859191Skris be used for external authentication procedures when the authentication 857959191Skris provided by SSL/TLS is not desired or is not enough. 858059191Skris [Bodo Moeller] 858159191Skris 858259191Skris *) Enhanced support for Alpha Linux is added. Now ./config checks if 858359191Skris the host supports BWX extension and if Compaq C is present on the 858459191Skris $PATH. Just exploiting of the BWX extension results in 20-30% 858559191Skris performance kick for some algorithms, e.g. DES and RC4 to mention 858659191Skris a couple. Compaq C in turn generates ~20% faster code for MD5 and 858759191Skris SHA1. 858859191Skris [Andy Polyakov] 858959191Skris 859059191Skris *) Add support for MS "fast SGC". This is arguably a violation of the 859159191Skris SSL3/TLS protocol. Netscape SGC does two handshakes: the first with 859259191Skris weak crypto and after checking the certificate is SGC a second one 859359191Skris with strong crypto. MS SGC stops the first handshake after receiving 859459191Skris the server certificate message and sends a second client hello. Since 859559191Skris a server will typically do all the time consuming operations before 859659191Skris expecting any further messages from the client (server key exchange 859759191Skris is the most expensive) there is little difference between the two. 859859191Skris 859959191Skris To get OpenSSL to support MS SGC we have to permit a second client 860059191Skris hello message after we have sent server done. In addition we have to 860159191Skris reset the MAC if we do get this second client hello. 860259191Skris [Steve Henson] 860359191Skris 860459191Skris *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide 860559191Skris if a DER encoded private key is RSA or DSA traditional format. Changed 860659191Skris d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" 860759191Skris format DER encoded private key. Newer code should use PKCS#8 format which 860859191Skris has the key type encoded in the ASN1 structure. Added DER private key 860959191Skris support to pkcs8 application. 861059191Skris [Steve Henson] 861159191Skris 861259191Skris *) SSL 3/TLS 1 servers now don't request certificates when an anonymous 861359191Skris ciphersuites has been selected (as required by the SSL 3/TLS 1 861459191Skris specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT 861559191Skris is set, we interpret this as a request to violate the specification 861659191Skris (the worst that can happen is a handshake failure, and 'correct' 861759191Skris behaviour would result in a handshake failure anyway). 861859191Skris [Bodo Moeller] 861959191Skris 862059191Skris *) In SSL_CTX_add_session, take into account that there might be multiple 862159191Skris SSL_SESSION structures with the same session ID (e.g. when two threads 862259191Skris concurrently obtain them from an external cache). 862359191Skris The internal cache can handle only one SSL_SESSION with a given ID, 862459191Skris so if there's a conflict, we now throw out the old one to achieve 862559191Skris consistency. 862659191Skris [Bodo Moeller] 862759191Skris 862859191Skris *) Add OIDs for idea and blowfish in CBC mode. This will allow both 862959191Skris to be used in PKCS#5 v2.0 and S/MIME. Also add checking to 863059191Skris some routines that use cipher OIDs: some ciphers do not have OIDs 863159191Skris defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for 863259191Skris example. 863359191Skris [Steve Henson] 863459191Skris 863559191Skris *) Simplify the trust setting structure and code. Now we just have 863659191Skris two sequences of OIDs for trusted and rejected settings. These will 863759191Skris typically have values the same as the extended key usage extension 863859191Skris and any application specific purposes. 863959191Skris 864059191Skris The trust checking code now has a default behaviour: it will just 864159191Skris check for an object with the same NID as the passed id. Functions can 864259191Skris be provided to override either the default behaviour or the behaviour 864359191Skris for a given id. SSL client, server and email already have functions 864459191Skris in place for compatibility: they check the NID and also return "trusted" 864559191Skris if the certificate is self signed. 864659191Skris [Steve Henson] 864759191Skris 864859191Skris *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the 864959191Skris traditional format into an EVP_PKEY structure. 865059191Skris [Steve Henson] 865159191Skris 865259191Skris *) Add a password callback function PEM_cb() which either prompts for 865359191Skris a password if usr_data is NULL or otherwise assumes it is a null 865459191Skris terminated password. Allow passwords to be passed on command line 865559191Skris environment or config files in a few more utilities. 865659191Skris [Steve Henson] 865759191Skris 865859191Skris *) Add a bunch of DER and PEM functions to handle PKCS#8 format private 865959191Skris keys. Add some short names for PKCS#8 PBE algorithms and allow them 866059191Skris to be specified on the command line for the pkcs8 and pkcs12 utilities. 866159191Skris Update documentation. 866259191Skris [Steve Henson] 866359191Skris 866459191Skris *) Support for ASN1 "NULL" type. This could be handled before by using 866559191Skris ASN1_TYPE but there wasn't any function that would try to read a NULL 866659191Skris and produce an error if it couldn't. For compatibility we also have 866759191Skris ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and 866859191Skris don't allocate anything because they don't need to. 866959191Skris [Steve Henson] 867059191Skris 867159191Skris *) Initial support for MacOS is now provided. Examine INSTALL.MacOS 867259191Skris for details. 867359191Skris [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] 867459191Skris 867559191Skris *) Rebuild of the memory allocation routines used by OpenSSL code and 867659191Skris possibly others as well. The purpose is to make an interface that 867759191Skris provide hooks so anyone can build a separate set of allocation and 867859191Skris deallocation routines to be used by OpenSSL, for example memory 867959191Skris pool implementations, or something else, which was previously hard 868059191Skris since Malloc(), Realloc() and Free() were defined as macros having 868159191Skris the values malloc, realloc and free, respectively (except for Win32 868259191Skris compilations). The same is provided for memory debugging code. 868359191Skris OpenSSL already comes with functionality to find memory leaks, but 868459191Skris this gives people a chance to debug other memory problems. 868559191Skris 868659191Skris With these changes, a new set of functions and macros have appeared: 868759191Skris 868868651Skris CRYPTO_set_mem_debug_functions() [F] 868968651Skris CRYPTO_get_mem_debug_functions() [F] 869068651Skris CRYPTO_dbg_set_options() [F] 869168651Skris CRYPTO_dbg_get_options() [F] 869268651Skris CRYPTO_malloc_debug_init() [M] 869359191Skris 869459191Skris The memory debug functions are NULL by default, unless the library 869559191Skris is compiled with CRYPTO_MDEBUG or friends is defined. If someone 869659191Skris wants to debug memory anyway, CRYPTO_malloc_debug_init() (which 869759191Skris gives the standard debugging functions that come with OpenSSL) or 869859191Skris CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions 869959191Skris provided by the library user) must be used. When the standard 870059191Skris debugging functions are used, CRYPTO_dbg_set_options can be used to 870159191Skris request additional information: 870259191Skris CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting 870359191Skris the CRYPTO_MDEBUG_xxx macro when compiling the library. 870459191Skris 870559191Skris Also, things like CRYPTO_set_mem_functions will always give the 870659191Skris expected result (the new set of functions is used for allocation 870759191Skris and deallocation) at all times, regardless of platform and compiler 870859191Skris options. 870959191Skris 871059191Skris To finish it up, some functions that were never use in any other 871159191Skris way than through macros have a new API and new semantic: 871259191Skris 871359191Skris CRYPTO_dbg_malloc() 871459191Skris CRYPTO_dbg_realloc() 871559191Skris CRYPTO_dbg_free() 871659191Skris 871759191Skris All macros of value have retained their old syntax. 871859191Skris [Richard Levitte and Bodo Moeller] 871959191Skris 872059191Skris *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the 872159191Skris ordering of SMIMECapabilities wasn't in "strength order" and there 872259191Skris was a missing NULL in the AlgorithmIdentifier for the SHA1 signature 872359191Skris algorithm. 872459191Skris [Steve Henson] 872559191Skris 872659191Skris *) Some ASN1 types with illegal zero length encoding (INTEGER, 872759191Skris ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. 872859191Skris [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] 872959191Skris 873059191Skris *) Merge in my S/MIME library for OpenSSL. This provides a simple 873159191Skris S/MIME API on top of the PKCS#7 code, a MIME parser (with enough 873259191Skris functionality to handle multipart/signed properly) and a utility 873359191Skris called 'smime' to call all this stuff. This is based on code I 873459191Skris originally wrote for Celo who have kindly allowed it to be 873559191Skris included in OpenSSL. 873659191Skris [Steve Henson] 873759191Skris 873859191Skris *) Add variants des_set_key_checked and des_set_key_unchecked of 873959191Skris des_set_key (aka des_key_sched). Global variable des_check_key 874059191Skris decides which of these is called by des_set_key; this way 874159191Skris des_check_key behaves as it always did, but applications and 874259191Skris the library itself, which was buggy for des_check_key == 1, 874359191Skris have a cleaner way to pick the version they need. 874459191Skris [Bodo Moeller] 874559191Skris 874659191Skris *) New function PKCS12_newpass() which changes the password of a 874759191Skris PKCS12 structure. 874859191Skris [Steve Henson] 874959191Skris 875059191Skris *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and 875159191Skris dynamic mix. In both cases the ids can be used as an index into the 875259191Skris table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() 875359191Skris functions so they accept a list of the field values and the 875459191Skris application doesn't need to directly manipulate the X509_TRUST 875559191Skris structure. 875659191Skris [Steve Henson] 875759191Skris 875859191Skris *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't 875959191Skris need initialising. 876059191Skris [Steve Henson] 876159191Skris 876259191Skris *) Modify the way the V3 extension code looks up extensions. This now 876359191Skris works in a similar way to the object code: we have some "standard" 876459191Skris extensions in a static table which is searched with OBJ_bsearch() 876559191Skris and the application can add dynamic ones if needed. The file 876659191Skris crypto/x509v3/ext_dat.h now has the info: this file needs to be 876759191Skris updated whenever a new extension is added to the core code and kept 876859191Skris in ext_nid order. There is a simple program 'tabtest.c' which checks 876959191Skris this. New extensions are not added too often so this file can readily 877059191Skris be maintained manually. 877159191Skris 877259191Skris There are two big advantages in doing things this way. The extensions 877359191Skris can be looked up immediately and no longer need to be "added" using 877459191Skris X509V3_add_standard_extensions(): this function now does nothing. 877559191Skris [Side note: I get *lots* of email saying the extension code doesn't 877659191Skris work because people forget to call this function] 877759191Skris Also no dynamic allocation is done unless new extensions are added: 877859191Skris so if we don't add custom extensions there is no need to call 877959191Skris X509V3_EXT_cleanup(). 878059191Skris [Steve Henson] 878159191Skris 878259191Skris *) Modify enc utility's salting as follows: make salting the default. Add a 878359191Skris magic header, so unsalted files fail gracefully instead of just decrypting 878459191Skris to garbage. This is because not salting is a big security hole, so people 878559191Skris should be discouraged from doing it. 878659191Skris [Ben Laurie] 878759191Skris 878859191Skris *) Fixes and enhancements to the 'x509' utility. It allowed a message 878959191Skris digest to be passed on the command line but it only used this 879059191Skris parameter when signing a certificate. Modified so all relevant 879159191Skris operations are affected by the digest parameter including the 879259191Skris -fingerprint and -x509toreq options. Also -x509toreq choked if a 879359191Skris DSA key was used because it didn't fix the digest. 879459191Skris [Steve Henson] 879559191Skris 879659191Skris *) Initial certificate chain verify code. Currently tests the untrusted 879759191Skris certificates for consistency with the verify purpose (which is set 879859191Skris when the X509_STORE_CTX structure is set up) and checks the pathlength. 879959191Skris 880059191Skris There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: 880159191Skris this is because it will reject chains with invalid extensions whereas 880259191Skris every previous version of OpenSSL and SSLeay made no checks at all. 880359191Skris 880459191Skris Trust code: checks the root CA for the relevant trust settings. Trust 880559191Skris settings have an initial value consistent with the verify purpose: e.g. 880659191Skris if the verify purpose is for SSL client use it expects the CA to be 880759191Skris trusted for SSL client use. However the default value can be changed to 880859191Skris permit custom trust settings: one example of this would be to only trust 880959191Skris certificates from a specific "secure" set of CAs. 881059191Skris 881159191Skris Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions 881259191Skris which should be used for version portability: especially since the 881359191Skris verify structure is likely to change more often now. 881459191Skris 881559191Skris SSL integration. Add purpose and trust to SSL_CTX and SSL and functions 881659191Skris to set them. If not set then assume SSL clients will verify SSL servers 881759191Skris and vice versa. 881859191Skris 881959191Skris Two new options to the verify program: -untrusted allows a set of 882059191Skris untrusted certificates to be passed in and -purpose which sets the 882159191Skris intended purpose of the certificate. If a purpose is set then the 882259191Skris new chain verify code is used to check extension consistency. 882359191Skris [Steve Henson] 882459191Skris 882559191Skris *) Support for the authority information access extension. 882659191Skris [Steve Henson] 882759191Skris 882859191Skris *) Modify RSA and DSA PEM read routines to transparently handle 882959191Skris PKCS#8 format private keys. New *_PUBKEY_* functions that handle 883059191Skris public keys in a format compatible with certificate 883159191Skris SubjectPublicKeyInfo structures. Unfortunately there were already 883259191Skris functions called *_PublicKey_* which used various odd formats so 883359191Skris these are retained for compatibility: however the DSA variants were 883459191Skris never in a public release so they have been deleted. Changed dsa/rsa 883559191Skris utilities to handle the new format: note no releases ever handled public 883659191Skris keys so we should be OK. 883759191Skris 883859191Skris The primary motivation for this change is to avoid the same fiasco 883959191Skris that dogs private keys: there are several incompatible private key 884059191Skris formats some of which are standard and some OpenSSL specific and 884159191Skris require various evil hacks to allow partial transparent handling and 884259191Skris even then it doesn't work with DER formats. Given the option anything 884359191Skris other than PKCS#8 should be dumped: but the other formats have to 884459191Skris stay in the name of compatibility. 884559191Skris 884659191Skris With public keys and the benefit of hindsight one standard format 884759191Skris is used which works with EVP_PKEY, RSA or DSA structures: though 884859191Skris it clearly returns an error if you try to read the wrong kind of key. 884959191Skris 885059191Skris Added a -pubkey option to the 'x509' utility to output the public key. 885159191Skris Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() 885259191Skris (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add 885359191Skris EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) 885459191Skris that do the same as the EVP_PKEY_assign_*() except they up the 885559191Skris reference count of the added key (they don't "swallow" the 885659191Skris supplied key). 885759191Skris [Steve Henson] 885859191Skris 885959191Skris *) Fixes to crypto/x509/by_file.c the code to read in certificates and 886059191Skris CRLs would fail if the file contained no certificates or no CRLs: 886159191Skris added a new function to read in both types and return the number 886259191Skris read: this means that if none are read it will be an error. The 886359191Skris DER versions of the certificate and CRL reader would always fail 886459191Skris because it isn't possible to mix certificates and CRLs in DER format 886559191Skris without choking one or the other routine. Changed this to just read 886659191Skris a certificate: this is the best we can do. Also modified the code 886759191Skris in apps/verify.c to take notice of return codes: it was previously 886859191Skris attempting to read in certificates from NULL pointers and ignoring 886959191Skris any errors: this is one reason why the cert and CRL reader seemed 887059191Skris to work. It doesn't check return codes from the default certificate 887159191Skris routines: these may well fail if the certificates aren't installed. 887259191Skris [Steve Henson] 887359191Skris 887459191Skris *) Code to support otherName option in GeneralName. 887559191Skris [Steve Henson] 887659191Skris 887759191Skris *) First update to verify code. Change the verify utility 887859191Skris so it warns if it is passed a self signed certificate: 887959191Skris for consistency with the normal behaviour. X509_verify 888059191Skris has been modified to it will now verify a self signed 888159191Skris certificate if *exactly* the same certificate appears 888259191Skris in the store: it was previously impossible to trust a 888359191Skris single self signed certificate. This means that: 888459191Skris openssl verify ss.pem 888559191Skris now gives a warning about a self signed certificate but 888659191Skris openssl verify -CAfile ss.pem ss.pem 888759191Skris is OK. 888859191Skris [Steve Henson] 888959191Skris 889059191Skris *) For servers, store verify_result in SSL_SESSION data structure 889159191Skris (and add it to external session representation). 889259191Skris This is needed when client certificate verifications fails, 889359191Skris but an application-provided verification callback (set by 889459191Skris SSL_CTX_set_cert_verify_callback) allows accepting the session 889559191Skris anyway (i.e. leaves x509_store_ctx->error != X509_V_OK 889659191Skris but returns 1): When the session is reused, we have to set 889759191Skris ssl->verify_result to the appropriate error code to avoid 889859191Skris security holes. 889959191Skris [Bodo Moeller, problem pointed out by Lutz Jaenicke] 890059191Skris 890159191Skris *) Fix a bug in the new PKCS#7 code: it didn't consider the 890259191Skris case in PKCS7_dataInit() where the signed PKCS7 structure 890359191Skris didn't contain any existing data because it was being created. 890459191Skris [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] 890559191Skris 890659191Skris *) Add a salt to the key derivation routines in enc.c. This 890759191Skris forms the first 8 bytes of the encrypted file. Also add a 890859191Skris -S option to allow a salt to be input on the command line. 890959191Skris [Steve Henson] 891059191Skris 891159191Skris *) New function X509_cmp(). Oddly enough there wasn't a function 891259191Skris to compare two certificates. We do this by working out the SHA1 891359191Skris hash and comparing that. X509_cmp() will be needed by the trust 891459191Skris code. 891559191Skris [Steve Henson] 891659191Skris 891759191Skris *) SSL_get1_session() is like SSL_get_session(), but increments 891859191Skris the reference count in the SSL_SESSION returned. 891959191Skris [Geoff Thorpe <geoff@eu.c2.net>] 892059191Skris 892159191Skris *) Fix for 'req': it was adding a null to request attributes. 892259191Skris Also change the X509_LOOKUP and X509_INFO code to handle 892359191Skris certificate auxiliary information. 892459191Skris [Steve Henson] 892559191Skris 892659191Skris *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document 892759191Skris the 'enc' command. 892859191Skris [Steve Henson] 892959191Skris 893059191Skris *) Add the possibility to add extra information to the memory leak 893159191Skris detecting output, to form tracebacks, showing from where each 893259191Skris allocation was originated: CRYPTO_push_info("constant string") adds 893359191Skris the string plus current file name and line number to a per-thread 893459191Skris stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() 893559191Skris is like calling CYRPTO_pop_info() until the stack is empty. 893659191Skris Also updated memory leak detection code to be multi-thread-safe. 893759191Skris [Richard Levitte] 893859191Skris 893959191Skris *) Add options -text and -noout to pkcs7 utility and delete the 894059191Skris encryption options which never did anything. Update docs. 894159191Skris [Steve Henson] 894259191Skris 894359191Skris *) Add options to some of the utilities to allow the pass phrase 894459191Skris to be included on either the command line (not recommended on 894559191Skris OSes like Unix) or read from the environment. Update the 894659191Skris manpages and fix a few bugs. 894759191Skris [Steve Henson] 894859191Skris 894959191Skris *) Add a few manpages for some of the openssl commands. 895059191Skris [Steve Henson] 895159191Skris 895259191Skris *) Fix the -revoke option in ca. It was freeing up memory twice, 895359191Skris leaking and not finding already revoked certificates. 895459191Skris [Steve Henson] 895559191Skris 895659191Skris *) Extensive changes to support certificate auxiliary information. 895759191Skris This involves the use of X509_CERT_AUX structure and X509_AUX 895859191Skris functions. An X509_AUX function such as PEM_read_X509_AUX() 895959191Skris can still read in a certificate file in the usual way but it 896059191Skris will also read in any additional "auxiliary information". By 896159191Skris doing things this way a fair degree of compatibility can be 896259191Skris retained: existing certificates can have this information added 896359191Skris using the new 'x509' options. 896459191Skris 896559191Skris Current auxiliary information includes an "alias" and some trust 896659191Skris settings. The trust settings will ultimately be used in enhanced 896759191Skris certificate chain verification routines: currently a certificate 896859191Skris can only be trusted if it is self signed and then it is trusted 896959191Skris for all purposes. 897059191Skris [Steve Henson] 897159191Skris 897259191Skris *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). 897359191Skris The problem was that one of the replacement routines had not been working 897459191Skris since SSLeay releases. For now the offending routine has been replaced 897559191Skris with non-optimised assembler. Even so, this now gives around 95% 897659191Skris performance improvement for 1024 bit RSA signs. 897759191Skris [Mark Cox] 897859191Skris 897959191Skris *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 898059191Skris handling. Most clients have the effective key size in bits equal to 898159191Skris the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. 898259191Skris A few however don't do this and instead use the size of the decrypted key 898359191Skris to determine the RC2 key length and the AlgorithmIdentifier to determine 898459191Skris the effective key length. In this case the effective key length can still 898559191Skris be 40 bits but the key length can be 168 bits for example. This is fixed 898659191Skris by manually forcing an RC2 key into the EVP_PKEY structure because the 898759191Skris EVP code can't currently handle unusual RC2 key sizes: it always assumes 898859191Skris the key length and effective key length are equal. 898959191Skris [Steve Henson] 899059191Skris 899159191Skris *) Add a bunch of functions that should simplify the creation of 899259191Skris X509_NAME structures. Now you should be able to do: 899359191Skris X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); 899459191Skris and have it automatically work out the correct field type and fill in 899559191Skris the structures. The more adventurous can try: 899659191Skris X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); 899759191Skris and it will (hopefully) work out the correct multibyte encoding. 899859191Skris [Steve Henson] 899959191Skris 900059191Skris *) Change the 'req' utility to use the new field handling and multibyte 900159191Skris copy routines. Before the DN field creation was handled in an ad hoc 900259191Skris way in req, ca, and x509 which was rather broken and didn't support 900359191Skris BMPStrings or UTF8Strings. Since some software doesn't implement 900459191Skris BMPStrings or UTF8Strings yet, they can be enabled using the config file 900559191Skris using the dirstring_type option. See the new comment in the default 900659191Skris openssl.cnf for more info. 900759191Skris [Steve Henson] 900859191Skris 900959191Skris *) Make crypto/rand/md_rand.c more robust: 901059191Skris - Assure unique random numbers after fork(). 901159191Skris - Make sure that concurrent threads access the global counter and 901259191Skris md serializably so that we never lose entropy in them 901359191Skris or use exactly the same state in multiple threads. 901459191Skris Access to the large state is not always serializable because 901559191Skris the additional locking could be a performance killer, and 901659191Skris md should be large enough anyway. 901759191Skris [Bodo Moeller] 901859191Skris 901959191Skris *) New file apps/app_rand.c with commonly needed functionality 902059191Skris for handling the random seed file. 902159191Skris 902259191Skris Use the random seed file in some applications that previously did not: 902359191Skris ca, 902459191Skris dsaparam -genkey (which also ignored its '-rand' option), 902559191Skris s_client, 902659191Skris s_server, 902759191Skris x509 (when signing). 902859191Skris Except on systems with /dev/urandom, it is crucial to have a random 902959191Skris seed file at least for key creation, DSA signing, and for DH exchanges; 903059191Skris for RSA signatures we could do without one. 903159191Skris 903259191Skris gendh and gendsa (unlike genrsa) used to read only the first byte 903359191Skris of each file listed in the '-rand' option. The function as previously 903459191Skris found in genrsa is now in app_rand.c and is used by all programs 903559191Skris that support '-rand'. 903659191Skris [Bodo Moeller] 903759191Skris 903859191Skris *) In RAND_write_file, use mode 0600 for creating files; 903959191Skris don't just chmod when it may be too late. 904059191Skris [Bodo Moeller] 904159191Skris 904259191Skris *) Report an error from X509_STORE_load_locations 904359191Skris when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. 904459191Skris [Bill Perry] 904559191Skris 904659191Skris *) New function ASN1_mbstring_copy() this copies a string in either 904759191Skris ASCII, Unicode, Universal (4 bytes per character) or UTF8 format 904859191Skris into an ASN1_STRING type. A mask of permissible types is passed 904959191Skris and it chooses the "minimal" type to use or an error if not type 905059191Skris is suitable. 905159191Skris [Steve Henson] 905259191Skris 905359191Skris *) Add function equivalents to the various macros in asn1.h. The old 905459191Skris macros are retained with an M_ prefix. Code inside the library can 905559191Skris use the M_ macros. External code (including the openssl utility) 905659191Skris should *NOT* in order to be "shared library friendly". 905759191Skris [Steve Henson] 905859191Skris 905959191Skris *) Add various functions that can check a certificate's extensions 906059191Skris to see if it usable for various purposes such as SSL client, 906159191Skris server or S/MIME and CAs of these types. This is currently 906259191Skris VERY EXPERIMENTAL but will ultimately be used for certificate chain 906359191Skris verification. Also added a -purpose flag to x509 utility to 906459191Skris print out all the purposes. 906559191Skris [Steve Henson] 906659191Skris 906759191Skris *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated 906859191Skris functions. 906959191Skris [Steve Henson] 907059191Skris 907159191Skris *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search 907259191Skris for, obtain and decode and extension and obtain its critical flag. 907359191Skris This allows all the necessary extension code to be handled in a 907459191Skris single function call. 907559191Skris [Steve Henson] 907659191Skris 907759191Skris *) RC4 tune-up featuring 30-40% performance improvement on most RISC 907859191Skris platforms. See crypto/rc4/rc4_enc.c for further details. 907959191Skris [Andy Polyakov] 908059191Skris 908159191Skris *) New -noout option to asn1parse. This causes no output to be produced 908259191Skris its main use is when combined with -strparse and -out to extract data 908359191Skris from a file (which may not be in ASN.1 format). 908459191Skris [Steve Henson] 908559191Skris 908659191Skris *) Fix for pkcs12 program. It was hashing an invalid certificate pointer 908759191Skris when producing the local key id. 908859191Skris [Richard Levitte <levitte@stacken.kth.se>] 908959191Skris 909059191Skris *) New option -dhparam in s_server. This allows a DH parameter file to be 909159191Skris stated explicitly. If it is not stated then it tries the first server 909259191Skris certificate file. The previous behaviour hard coded the filename 909359191Skris "server.pem". 909459191Skris [Steve Henson] 909559191Skris 909659191Skris *) Add -pubin and -pubout options to the rsa and dsa commands. These allow 909759191Skris a public key to be input or output. For example: 909859191Skris openssl rsa -in key.pem -pubout -out pubkey.pem 909959191Skris Also added necessary DSA public key functions to handle this. 910059191Skris [Steve Henson] 910159191Skris 910259191Skris *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained 910359191Skris in the message. This was handled by allowing 910459191Skris X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. 910559191Skris [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] 910659191Skris 910759191Skris *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null 910859191Skris to the end of the strings whereas this didn't. This would cause problems 910959191Skris if strings read with d2i_ASN1_bytes() were later modified. 911059191Skris [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] 911159191Skris 911259191Skris *) Fix for base64 decode bug. When a base64 bio reads only one line of 911359191Skris data and it contains EOF it will end up returning an error. This is 911459191Skris caused by input 46 bytes long. The cause is due to the way base64 911559191Skris BIOs find the start of base64 encoded data. They do this by trying a 911659191Skris trial decode on each line until they find one that works. When they 911759191Skris do a flag is set and it starts again knowing it can pass all the 911859191Skris data directly through the decoder. Unfortunately it doesn't reset 911959191Skris the context it uses. This means that if EOF is reached an attempt 912059191Skris is made to pass two EOFs through the context and this causes the 912159191Skris resulting error. This can also cause other problems as well. As is 912259191Skris usual with these problems it takes *ages* to find and the fix is 912359191Skris trivial: move one line. 912459191Skris [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] 912559191Skris 912659191Skris *) Ugly workaround to get s_client and s_server working under Windows. The 912759191Skris old code wouldn't work because it needed to select() on sockets and the 912859191Skris tty (for keypresses and to see if data could be written). Win32 only 912959191Skris supports select() on sockets so we select() with a 1s timeout on the 913059191Skris sockets and then see if any characters are waiting to be read, if none 913159191Skris are present then we retry, we also assume we can always write data to 913259191Skris the tty. This isn't nice because the code then blocks until we've 913359191Skris received a complete line of data and it is effectively polling the 913459191Skris keyboard at 1s intervals: however it's quite a bit better than not 913559191Skris working at all :-) A dedicated Windows application might handle this 913659191Skris with an event loop for example. 913759191Skris [Steve Henson] 913859191Skris 913959191Skris *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign 914059191Skris and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions 914159191Skris will be called when RSA_sign() and RSA_verify() are used. This is useful 914259191Skris if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. 914359191Skris For this to work properly RSA_public_decrypt() and RSA_private_encrypt() 914459191Skris should *not* be used: RSA_sign() and RSA_verify() must be used instead. 914559191Skris This necessitated the support of an extra signature type NID_md5_sha1 914659191Skris for SSL signatures and modifications to the SSL library to use it instead 914759191Skris of calling RSA_public_decrypt() and RSA_private_encrypt(). 914859191Skris [Steve Henson] 914959191Skris 915059191Skris *) Add new -verify -CAfile and -CApath options to the crl program, these 915159191Skris will lookup a CRL issuers certificate and verify the signature in a 915259191Skris similar way to the verify program. Tidy up the crl program so it 915359191Skris no longer accesses structures directly. Make the ASN1 CRL parsing a bit 915459191Skris less strict. It will now permit CRL extensions even if it is not 915559191Skris a V2 CRL: this will allow it to tolerate some broken CRLs. 915659191Skris [Steve Henson] 915759191Skris 915859191Skris *) Initialize all non-automatic variables each time one of the openssl 915959191Skris sub-programs is started (this is necessary as they may be started 916059191Skris multiple times from the "OpenSSL>" prompt). 916159191Skris [Lennart Bang, Bodo Moeller] 916259191Skris 916359191Skris *) Preliminary compilation option RSA_NULL which disables RSA crypto without 916459191Skris removing all other RSA functionality (this is what NO_RSA does). This 916559191Skris is so (for example) those in the US can disable those operations covered 916659191Skris by the RSA patent while allowing storage and parsing of RSA keys and RSA 916759191Skris key generation. 916859191Skris [Steve Henson] 916959191Skris 917059191Skris *) Non-copying interface to BIO pairs. 917159191Skris (still largely untested) 917259191Skris [Bodo Moeller] 917359191Skris 917459191Skris *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive 917559191Skris ASCII string. This was handled independently in various places before. 917659191Skris [Steve Henson] 917759191Skris 917859191Skris *) New functions UTF8_getc() and UTF8_putc() that parse and generate 917959191Skris UTF8 strings a character at a time. 918059191Skris [Steve Henson] 918159191Skris 918259191Skris *) Use client_version from client hello to select the protocol 918359191Skris (s23_srvr.c) and for RSA client key exchange verification 918459191Skris (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. 918559191Skris [Bodo Moeller] 918659191Skris 918759191Skris *) Add various utility functions to handle SPKACs, these were previously 918859191Skris handled by poking round in the structure internals. Added new function 918959191Skris NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to 919059191Skris print, verify and generate SPKACs. Based on an original idea from 919159191Skris Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. 919259191Skris [Steve Henson] 919359191Skris 919459191Skris *) RIPEMD160 is operational on all platforms and is back in 'make test'. 919559191Skris [Andy Polyakov] 919659191Skris 919759191Skris *) Allow the config file extension section to be overwritten on the 919859191Skris command line. Based on an original idea from Massimiliano Pala 919959191Skris <madwolf@comune.modena.it>. The new option is called -extensions 920059191Skris and can be applied to ca, req and x509. Also -reqexts to override 920159191Skris the request extensions in req and -crlexts to override the crl extensions 920259191Skris in ca. 920359191Skris [Steve Henson] 920459191Skris 920559191Skris *) Add new feature to the SPKAC handling in ca. Now you can include 920659191Skris the same field multiple times by preceding it by "XXXX." for example: 920759191Skris 1.OU="Unit name 1" 920859191Skris 2.OU="Unit name 2" 920959191Skris this is the same syntax as used in the req config file. 921059191Skris [Steve Henson] 921159191Skris 921259191Skris *) Allow certificate extensions to be added to certificate requests. These 921359191Skris are specified in a 'req_extensions' option of the req section of the 921459191Skris config file. They can be printed out with the -text option to req but 921559191Skris are otherwise ignored at present. 921659191Skris [Steve Henson] 921759191Skris 921859191Skris *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first 921959191Skris data read consists of only the final block it would not decrypted because 922059191Skris EVP_CipherUpdate() would correctly report zero bytes had been decrypted. 922159191Skris A misplaced 'break' also meant the decrypted final block might not be 922259191Skris copied until the next read. 922359191Skris [Steve Henson] 922459191Skris 922559191Skris *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added 922659191Skris a few extra parameters to the DH structure: these will be useful if 922759191Skris for example we want the value of 'q' or implement X9.42 DH. 922859191Skris [Steve Henson] 922959191Skris 923059191Skris *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and 923159191Skris provides hooks that allow the default DSA functions or functions on a 923259191Skris "per key" basis to be replaced. This allows hardware acceleration and 923359191Skris hardware key storage to be handled without major modification to the 923459191Skris library. Also added low level modexp hooks and CRYPTO_EX structure and 923559191Skris associated functions. 923659191Skris [Steve Henson] 923759191Skris 923859191Skris *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO 923959191Skris as "read only": it can't be written to and the buffer it points to will 924059191Skris not be freed. Reading from a read only BIO is much more efficient than 924159191Skris a normal memory BIO. This was added because there are several times when 924259191Skris an area of memory needs to be read from a BIO. The previous method was 924359191Skris to create a memory BIO and write the data to it, this results in two 924459191Skris copies of the data and an O(n^2) reading algorithm. There is a new 924559191Skris function BIO_new_mem_buf() which creates a read only memory BIO from 924659191Skris an area of memory. Also modified the PKCS#7 routines to use read only 924759191Skris memory BIOs. 924859191Skris [Steve Henson] 924959191Skris 925059191Skris *) Bugfix: ssl23_get_client_hello did not work properly when called in 925159191Skris state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of 925259191Skris a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, 925359191Skris but a retry condition occured while trying to read the rest. 925459191Skris [Bodo Moeller] 925559191Skris 925659191Skris *) The PKCS7_ENC_CONTENT_new() function was setting the content type as 925759191Skris NID_pkcs7_encrypted by default: this was wrong since this should almost 925859191Skris always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle 925959191Skris the encrypted data type: this is a more sensible place to put it and it 926059191Skris allows the PKCS#12 code to be tidied up that duplicated this 926159191Skris functionality. 926259191Skris [Steve Henson] 926359191Skris 926459191Skris *) Changed obj_dat.pl script so it takes its input and output files on 926559191Skris the command line. This should avoid shell escape redirection problems 926659191Skris under Win32. 926759191Skris [Steve Henson] 926859191Skris 926959191Skris *) Initial support for certificate extension requests, these are included 927059191Skris in things like Xenroll certificate requests. Included functions to allow 927159191Skris extensions to be obtained and added. 927259191Skris [Steve Henson] 927359191Skris 927459191Skris *) -crlf option to s_client and s_server for sending newlines as 927559191Skris CRLF (as required by many protocols). 927659191Skris [Bodo Moeller] 927759191Skris 927855714Skris Changes between 0.9.3a and 0.9.4 [09 Aug 1999] 927955714Skris 928055714Skris *) Install libRSAglue.a when OpenSSL is built with RSAref. 928155714Skris [Ralf S. Engelschall] 928255714Skris 928355714Skris *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. 928455714Skris [Andrija Antonijevic <TheAntony2@bigfoot.com>] 928555714Skris 928655714Skris *) Fix -startdate and -enddate (which was missing) arguments to 'ca' 928755714Skris program. 928855714Skris [Steve Henson] 928955714Skris 929055714Skris *) New function DSA_dup_DH, which duplicates DSA parameters/keys as 929155714Skris DH parameters/keys (q is lost during that conversion, but the resulting 929255714Skris DH parameters contain its length). 929355714Skris 929455714Skris For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is 929555714Skris much faster than DH_generate_parameters (which creates parameters 929655714Skris where p = 2*q + 1), and also the smaller q makes DH computations 929755714Skris much more efficient (160-bit exponentiation instead of 1024-bit 929855714Skris exponentiation); so this provides a convenient way to support DHE 929955714Skris ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of 930055714Skris utter importance to use 930155714Skris SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 930255714Skris or 930355714Skris SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); 930455714Skris when such DH parameters are used, because otherwise small subgroup 930555714Skris attacks may become possible! 930655714Skris [Bodo Moeller] 930755714Skris 930855714Skris *) Avoid memory leak in i2d_DHparams. 930955714Skris [Bodo Moeller] 931055714Skris 931155714Skris *) Allow the -k option to be used more than once in the enc program: 931255714Skris this allows the same encrypted message to be read by multiple recipients. 931355714Skris [Steve Henson] 931455714Skris 931555714Skris *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts 931655714Skris an ASN1_OBJECT to a text string. If the "no_name" parameter is set then 931755714Skris it will always use the numerical form of the OID, even if it has a short 931855714Skris or long name. 931955714Skris [Steve Henson] 932055714Skris 932155714Skris *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp 932255714Skris method only got called if p,q,dmp1,dmq1,iqmp components were present, 932355714Skris otherwise bn_mod_exp was called. In the case of hardware keys for example 932455714Skris no private key components need be present and it might store extra data 932559191Skris in the RSA structure, which cannot be accessed from bn_mod_exp. 932659191Skris By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for 932759191Skris private key operations. 932855714Skris [Steve Henson] 932955714Skris 933055714Skris *) Added support for SPARC Linux. 933155714Skris [Andy Polyakov] 933255714Skris 933355714Skris *) pem_password_cb function type incompatibly changed from 933455714Skris typedef int pem_password_cb(char *buf, int size, int rwflag); 933555714Skris to 933655714Skris ....(char *buf, int size, int rwflag, void *userdata); 933755714Skris so that applications can pass data to their callbacks: 933855714Skris The PEM[_ASN1]_{read,write}... functions and macros now take an 933955714Skris additional void * argument, which is just handed through whenever 934055714Skris the password callback is called. 934159191Skris [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] 934255714Skris 934355714Skris New function SSL_CTX_set_default_passwd_cb_userdata. 934455714Skris 934555714Skris Compatibility note: As many C implementations push function arguments 934655714Skris onto the stack in reverse order, the new library version is likely to 934755714Skris interoperate with programs that have been compiled with the old 934855714Skris pem_password_cb definition (PEM_whatever takes some data that 934955714Skris happens to be on the stack as its last argument, and the callback 935055714Skris just ignores this garbage); but there is no guarantee whatsoever that 935155714Skris this will work. 935255714Skris 935355714Skris *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... 935455714Skris (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused 935555714Skris problems not only on Windows, but also on some Unix platforms. 935655714Skris To avoid problematic command lines, these definitions are now in an 935755714Skris auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl 935855714Skris for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). 935955714Skris [Bodo Moeller] 936055714Skris 936155714Skris *) MIPS III/IV assembler module is reimplemented. 936255714Skris [Andy Polyakov] 936355714Skris 936455714Skris *) More DES library cleanups: remove references to srand/rand and 936555714Skris delete an unused file. 9366291721Sjkim [Ulf M��ller] 936755714Skris 936855714Skris *) Add support for the the free Netwide assembler (NASM) under Win32, 936955714Skris since not many people have MASM (ml) and it can be hard to obtain. 937055714Skris This is currently experimental but it seems to work OK and pass all 937155714Skris the tests. Check out INSTALL.W32 for info. 937255714Skris [Steve Henson] 937355714Skris 937455714Skris *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections 937555714Skris without temporary keys kept an extra copy of the server key, 937655714Skris and connections with temporary keys did not free everything in case 937755714Skris of an error. 937855714Skris [Bodo Moeller] 937955714Skris 938055714Skris *) New function RSA_check_key and new openssl rsa option -check 938155714Skris for verifying the consistency of RSA keys. 938255714Skris [Ulf Moeller, Bodo Moeller] 938355714Skris 938455714Skris *) Various changes to make Win32 compile work: 938555714Skris 1. Casts to avoid "loss of data" warnings in p5_crpt2.c 938655714Skris 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned 938755714Skris comparison" warnings. 938855714Skris 3. Add sk_<TYPE>_sort to DEF file generator and do make update. 938955714Skris [Steve Henson] 939055714Skris 939155714Skris *) Add a debugging option to PKCS#5 v2 key generation function: when 939255714Skris you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and 939355714Skris derived keys are printed to stderr. 939455714Skris [Steve Henson] 939555714Skris 939655714Skris *) Copy the flags in ASN1_STRING_dup(). 939755714Skris [Roman E. Pavlov <pre@mo.msk.ru>] 939855714Skris 939955714Skris *) The x509 application mishandled signing requests containing DSA 940055714Skris keys when the signing key was also DSA and the parameters didn't match. 940155714Skris 940255714Skris It was supposed to omit the parameters when they matched the signing key: 940355714Skris the verifying software was then supposed to automatically use the CA's 940455714Skris parameters if they were absent from the end user certificate. 940555714Skris 940655714Skris Omitting parameters is no longer recommended. The test was also 940755714Skris the wrong way round! This was probably due to unusual behaviour in 940855714Skris EVP_cmp_parameters() which returns 1 if the parameters match. 940955714Skris This meant that parameters were omitted when they *didn't* match and 941055714Skris the certificate was useless. Certificates signed with 'ca' didn't have 941155714Skris this bug. 941255714Skris [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] 941355714Skris 941455714Skris *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. 941555714Skris The interface is as follows: 941655714Skris Applications can use 941755714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), 941855714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); 941955714Skris "off" is now the default. 942055714Skris The library internally uses 942155714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), 942255714Skris CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() 942355714Skris to disable memory-checking temporarily. 942455714Skris 942555714Skris Some inconsistent states that previously were possible (and were 942655714Skris even the default) are now avoided. 942755714Skris 942855714Skris -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time 942955714Skris with each memory chunk allocated; this is occasionally more helpful 943055714Skris than just having a counter. 943155714Skris 943255714Skris -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. 943355714Skris 943455714Skris -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future 943555714Skris extensions. 943655714Skris [Bodo Moeller] 943755714Skris 943855714Skris *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), 943955714Skris which largely parallels "options", but is for changing API behaviour, 944055714Skris whereas "options" are about protocol behaviour. 944155714Skris Initial "mode" flags are: 944255714Skris 944355714Skris SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when 944455714Skris a single record has been written. 944555714Skris SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write 944655714Skris retries use the same buffer location. 944755714Skris (But all of the contents must be 944855714Skris copied!) 944955714Skris [Bodo Moeller] 945055714Skris 945179998Skris *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options 945255714Skris worked. 945355714Skris 945455714Skris *) Fix problems with no-hmac etc. 9455291721Sjkim [Ulf M��ller, pointed out by Brian Wellington <bwelling@tislabs.com>] 945655714Skris 945755714Skris *) New functions RSA_get_default_method(), RSA_set_method() and 945855714Skris RSA_get_method(). These allows replacement of RSA_METHODs without having 945955714Skris to mess around with the internals of an RSA structure. 946055714Skris [Steve Henson] 946155714Skris 946255714Skris *) Fix memory leaks in DSA_do_sign and DSA_is_prime. 946355714Skris Also really enable memory leak checks in openssl.c and in some 946455714Skris test programs. 946555714Skris [Chad C. Mulligan, Bodo Moeller] 946655714Skris 946755714Skris *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess 946855714Skris up the length of negative integers. This has now been simplified to just 946955714Skris store the length when it is first determined and use it later, rather 947055714Skris than trying to keep track of where data is copied and updating it to 947155714Skris point to the end. 947255714Skris [Steve Henson, reported by Brien Wheeler 947355714Skris <bwheeler@authentica-security.com>] 947455714Skris 947555714Skris *) Add a new function PKCS7_signatureVerify. This allows the verification 947655714Skris of a PKCS#7 signature but with the signing certificate passed to the 947755714Skris function itself. This contrasts with PKCS7_dataVerify which assumes the 947855714Skris certificate is present in the PKCS#7 structure. This isn't always the 947955714Skris case: certificates can be omitted from a PKCS#7 structure and be 948055714Skris distributed by "out of band" means (such as a certificate database). 948155714Skris [Steve Henson] 948255714Skris 948355714Skris *) Complete the PEM_* macros with DECLARE_PEM versions to replace the 948455714Skris function prototypes in pem.h, also change util/mkdef.pl to add the 948555714Skris necessary function names. 948655714Skris [Steve Henson] 948755714Skris 948855714Skris *) mk1mf.pl (used by Windows builds) did not properly read the 948955714Skris options set by Configure in the top level Makefile, and Configure 949055714Skris was not even able to write more than one option correctly. 949155714Skris Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. 949255714Skris [Bodo Moeller] 949355714Skris 949455714Skris *) New functions CONF_load_bio() and CONF_load_fp() to allow a config 949555714Skris file to be loaded from a BIO or FILE pointer. The BIO version will 949655714Skris for example allow memory BIOs to contain config info. 949755714Skris [Steve Henson] 949855714Skris 949955714Skris *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. 950055714Skris Whoever hopes to achieve shared-library compatibility across versions 950155714Skris must use this, not the compile-time macro. 950255714Skris (Exercise 0.9.4: Which is the minimum library version required by 950355714Skris such programs?) 950455714Skris Note: All this applies only to multi-threaded programs, others don't 950555714Skris need locks. 950655714Skris [Bodo Moeller] 950755714Skris 950855714Skris *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests 950955714Skris through a BIO pair triggered the default case, i.e. 951055714Skris SSLerr(...,SSL_R_UNKNOWN_STATE). 951155714Skris [Bodo Moeller] 951255714Skris 951355714Skris *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications 951455714Skris can use the SSL library even if none of the specific BIOs is 951555714Skris appropriate. 951655714Skris [Bodo Moeller] 951755714Skris 951855714Skris *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value 951955714Skris for the encoded length. 952055714Skris [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] 952155714Skris 952255714Skris *) Add initial documentation of the X509V3 functions. 952355714Skris [Steve Henson] 952455714Skris 952555714Skris *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 952655714Skris PEM_write_bio_PKCS8PrivateKey() that are equivalent to 952755714Skris PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more 952855714Skris secure PKCS#8 private key format with a high iteration count. 952955714Skris [Steve Henson] 953055714Skris 953155714Skris *) Fix determination of Perl interpreter: A perl or perl5 953255714Skris _directory_ in $PATH was also accepted as the interpreter. 953355714Skris [Ralf S. Engelschall] 953455714Skris 953555714Skris *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking 953655714Skris wrong with it but it was very old and did things like calling 953755714Skris PEM_ASN1_read() directly and used MD5 for the hash not to mention some 953855714Skris unusual formatting. 953955714Skris [Steve Henson] 954055714Skris 954155714Skris *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed 954255714Skris to use the new extension code. 954355714Skris [Steve Henson] 954455714Skris 954555714Skris *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c 954655714Skris with macros. This should make it easier to change their form, add extra 954755714Skris arguments etc. Fix a few PEM prototypes which didn't have cipher as a 954855714Skris constant. 954955714Skris [Steve Henson] 955055714Skris 955155714Skris *) Add to configuration table a new entry that can specify an alternative 955255714Skris name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, 955355714Skris according to Mark Crispin <MRC@Panda.COM>. 955455714Skris [Bodo Moeller] 955555714Skris 955655714Skris#if 0 955755714Skris *) DES CBC did not update the IV. Weird. 955855714Skris [Ben Laurie] 955955714Skris#else 956055714Skris des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. 956155714Skris Changing the behaviour of the former might break existing programs -- 956255714Skris where IV updating is needed, des_ncbc_encrypt can be used. 956355714Skris#endif 956455714Skris 956555714Skris *) When bntest is run from "make test" it drives bc to check its 956655714Skris calculations, as well as internally checking them. If an internal check 956755714Skris fails, it needs to cause bc to give a non-zero result or make test carries 956855714Skris on without noticing the failure. Fixed. 956955714Skris [Ben Laurie] 957055714Skris 957155714Skris *) DES library cleanups. 9572291721Sjkim [Ulf M��ller] 957355714Skris 957455714Skris *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be 957555714Skris used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit 957655714Skris ciphers. NOTE: although the key derivation function has been verified 957755714Skris against some published test vectors it has not been extensively tested 957855714Skris yet. Added a -v2 "cipher" option to pkcs8 application to allow the use 957955714Skris of v2.0. 958055714Skris [Steve Henson] 958155714Skris 958255714Skris *) Instead of "mkdir -p", which is not fully portable, use new 958355714Skris Perl script "util/mkdir-p.pl". 958455714Skris [Bodo Moeller] 958555714Skris 958655714Skris *) Rewrite the way password based encryption (PBE) is handled. It used to 958755714Skris assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter 958855714Skris structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms 958955714Skris but doesn't apply to PKCS#5 v2.0 where it can be something else. Now 959055714Skris the 'parameter' field of the AlgorithmIdentifier is passed to the 959155714Skris underlying key generation function so it must do its own ASN1 parsing. 959255714Skris This has also changed the EVP_PBE_CipherInit() function which now has a 959355714Skris 'parameter' argument instead of literal salt and iteration count values 959455714Skris and the function EVP_PBE_ALGOR_CipherInit() has been deleted. 959555714Skris [Steve Henson] 959655714Skris 959755714Skris *) Support for PKCS#5 v1.5 compatible password based encryption algorithms 959855714Skris and PKCS#8 functionality. New 'pkcs8' application linked to openssl. 959955714Skris Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE 960055714Skris KEY" because this clashed with PKCS#8 unencrypted string. Since this 960155714Skris value was just used as a "magic string" and not used directly its 960255714Skris value doesn't matter. 960355714Skris [Steve Henson] 960455714Skris 960555714Skris *) Introduce some semblance of const correctness to BN. Shame C doesn't 960655714Skris support mutable. 960755714Skris [Ben Laurie] 960855714Skris 960955714Skris *) "linux-sparc64" configuration (ultrapenguin). 961055714Skris [Ray Miller <ray.miller@oucs.ox.ac.uk>] 961155714Skris "linux-sparc" configuration. 961255714Skris [Christian Forster <fo@hawo.stw.uni-erlangen.de>] 961355714Skris 961455714Skris *) config now generates no-xxx options for missing ciphers. 9615291721Sjkim [Ulf M��ller] 961655714Skris 961755714Skris *) Support the EBCDIC character set (work in progress). 961855714Skris File ebcdic.c not yet included because it has a different license. 961955714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 962055714Skris 962155714Skris *) Support BS2000/OSD-POSIX. 962255714Skris [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] 962355714Skris 962455714Skris *) Make callbacks for key generation use void * instead of char *. 962555714Skris [Ben Laurie] 962655714Skris 962755714Skris *) Make S/MIME samples compile (not yet tested). 962855714Skris [Ben Laurie] 962955714Skris 963055714Skris *) Additional typesafe stacks. 963155714Skris [Ben Laurie] 963255714Skris 963355714Skris *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). 963455714Skris [Bodo Moeller] 963555714Skris 963655714Skris 963755714Skris Changes between 0.9.3 and 0.9.3a [29 May 1999] 963855714Skris 963955714Skris *) New configuration variant "sco5-gcc". 964055714Skris 964155714Skris *) Updated some demos. 964255714Skris [Sean O Riordain, Wade Scholine] 964355714Skris 964455714Skris *) Add missing BIO_free at exit of pkcs12 application. 964555714Skris [Wu Zhigang] 964655714Skris 964755714Skris *) Fix memory leak in conf.c. 964855714Skris [Steve Henson] 964955714Skris 965055714Skris *) Updates for Win32 to assembler version of MD5. 965155714Skris [Steve Henson] 965255714Skris 965355714Skris *) Set #! path to perl in apps/der_chop to where we found it 965455714Skris instead of using a fixed path. 965555714Skris [Bodo Moeller] 965655714Skris 965755714Skris *) SHA library changes for irix64-mips4-cc. 965855714Skris [Andy Polyakov] 965955714Skris 966055714Skris *) Improvements for VMS support. 966155714Skris [Richard Levitte] 966255714Skris 966355714Skris 966455714Skris Changes between 0.9.2b and 0.9.3 [24 May 1999] 966555714Skris 966655714Skris *) Bignum library bug fix. IRIX 6 passes "make test" now! 966755714Skris This also avoids the problems with SC4.2 and unpatched SC5. 966855714Skris [Andy Polyakov <appro@fy.chalmers.se>] 966955714Skris 967055714Skris *) New functions sk_num, sk_value and sk_set to replace the previous macros. 967155714Skris These are required because of the typesafe stack would otherwise break 967255714Skris existing code. If old code used a structure member which used to be STACK 967355714Skris and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with 967455714Skris sk_num or sk_value it would produce an error because the num, data members 967555714Skris are not present in STACK_OF. Now it just produces a warning. sk_set 967655714Skris replaces the old method of assigning a value to sk_value 967755714Skris (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code 967855714Skris that does this will no longer work (and should use sk_set instead) but 967955714Skris this could be regarded as a "questionable" behaviour anyway. 968055714Skris [Steve Henson] 968155714Skris 968255714Skris *) Fix most of the other PKCS#7 bugs. The "experimental" code can now 968355714Skris correctly handle encrypted S/MIME data. 968455714Skris [Steve Henson] 968555714Skris 968655714Skris *) Change type of various DES function arguments from des_cblock 968755714Skris (which means, in function argument declarations, pointer to char) 968855714Skris to des_cblock * (meaning pointer to array with 8 char elements), 968955714Skris which allows the compiler to do more typechecking; it was like 969055714Skris that back in SSLeay, but with lots of ugly casts. 969155714Skris 969255714Skris Introduce new type const_des_cblock. 969355714Skris [Bodo Moeller] 969455714Skris 969555714Skris *) Reorganise the PKCS#7 library and get rid of some of the more obvious 969655714Skris problems: find RecipientInfo structure that matches recipient certificate 969755714Skris and initialise the ASN1 structures properly based on passed cipher. 969855714Skris [Steve Henson] 969955714Skris 970055714Skris *) Belatedly make the BN tests actually check the results. 970155714Skris [Ben Laurie] 970255714Skris 970355714Skris *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion 970455714Skris to and from BNs: it was completely broken. New compilation option 970555714Skris NEG_PUBKEY_BUG to allow for some broken certificates that encode public 970655714Skris key elements as negative integers. 970755714Skris [Steve Henson] 970855714Skris 970955714Skris *) Reorganize and speed up MD5. 971055714Skris [Andy Polyakov <appro@fy.chalmers.se>] 971155714Skris 971255714Skris *) VMS support. 971355714Skris [Richard Levitte <richard@levitte.org>] 971455714Skris 971555714Skris *) New option -out to asn1parse to allow the parsed structure to be 971655714Skris output to a file. This is most useful when combined with the -strparse 971755714Skris option to examine the output of things like OCTET STRINGS. 971855714Skris [Steve Henson] 971955714Skris 972055714Skris *) Make SSL library a little more fool-proof by not requiring any longer 972155714Skris that SSL_set_{accept,connect}_state be called before 972255714Skris SSL_{accept,connect} may be used (SSL_set_..._state is omitted 972355714Skris in many applications because usually everything *appeared* to work as 972455714Skris intended anyway -- now it really works as intended). 972555714Skris [Bodo Moeller] 972655714Skris 972755714Skris *) Move openssl.cnf out of lib/. 9728291721Sjkim [Ulf M��ller] 972955714Skris 973055714Skris *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall 973155714Skris -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes 973255714Skris -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 973355714Skris [Ralf S. Engelschall] 973455714Skris 973555714Skris *) Various fixes to the EVP and PKCS#7 code. It may now be able to 973655714Skris handle PKCS#7 enveloped data properly. 973755714Skris [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] 973855714Skris 973955714Skris *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of 974055714Skris copying pointers. The cert_st handling is changed by this in 974155714Skris various ways (and thus what used to be known as ctx->default_cert 974255714Skris is now called ctx->cert, since we don't resort to s->ctx->[default_]cert 974355714Skris any longer when s->cert does not give us what we need). 974455714Skris ssl_cert_instantiate becomes obsolete by this change. 974555714Skris As soon as we've got the new code right (possibly it already is?), 974655714Skris we have solved a couple of bugs of the earlier code where s->cert 974755714Skris was used as if it could not have been shared with other SSL structures. 974855714Skris 974955714Skris Note that using the SSL API in certain dirty ways now will result 975055714Skris in different behaviour than observed with earlier library versions: 975155714Skris Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) 975255714Skris does not influence s as it used to. 975355714Skris 975455714Skris In order to clean up things more thoroughly, inside SSL_SESSION 975555714Skris we don't use CERT any longer, but a new structure SESS_CERT 975655714Skris that holds per-session data (if available); currently, this is 975755714Skris the peer's certificate chain and, for clients, the server's certificate 975855714Skris and temporary key. CERT holds only those values that can have 975955714Skris meaningful defaults in an SSL_CTX. 976055714Skris [Bodo Moeller] 976155714Skris 976255714Skris *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure 976355714Skris from the internal representation. Various PKCS#7 fixes: remove some 976455714Skris evil casts and set the enc_dig_alg field properly based on the signing 976555714Skris key type. 976655714Skris [Steve Henson] 976755714Skris 976855714Skris *) Allow PKCS#12 password to be set from the command line or the 976955714Skris environment. Let 'ca' get its config file name from the environment 977055714Skris variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' 977155714Skris and 'x509'). 977255714Skris [Steve Henson] 977355714Skris 977455714Skris *) Allow certificate policies extension to use an IA5STRING for the 977555714Skris organization field. This is contrary to the PKIX definition but 977655714Skris VeriSign uses it and IE5 only recognises this form. Document 'x509' 977755714Skris extension option. 977855714Skris [Steve Henson] 977955714Skris 978055714Skris *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, 978155714Skris without disallowing inline assembler and the like for non-pedantic builds. 978255714Skris [Ben Laurie] 978355714Skris 978455714Skris *) Support Borland C++ builder. 9785291721Sjkim [Janez Jere <jj@void.si>, modified by Ulf M��ller] 978655714Skris 978755714Skris *) Support Mingw32. 9788291721Sjkim [Ulf M��ller] 978955714Skris 979055714Skris *) SHA-1 cleanups and performance enhancements. 979155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 979255714Skris 979355714Skris *) Sparc v8plus assembler for the bignum library. 979455714Skris [Andy Polyakov <appro@fy.chalmers.se>] 979555714Skris 979655714Skris *) Accept any -xxx and +xxx compiler options in Configure. 9797291721Sjkim [Ulf M��ller] 979855714Skris 979955714Skris *) Update HPUX configuration. 980055714Skris [Anonymous] 980155714Skris 980255714Skris *) Add missing sk_<type>_unshift() function to safestack.h 980355714Skris [Ralf S. Engelschall] 980455714Skris 980555714Skris *) New function SSL_CTX_use_certificate_chain_file that sets the 980655714Skris "extra_cert"s in addition to the certificate. (This makes sense 980755714Skris only for "PEM" format files, as chains as a whole are not 980855714Skris DER-encoded.) 980955714Skris [Bodo Moeller] 981055714Skris 981155714Skris *) Support verify_depth from the SSL API. 981255714Skris x509_vfy.c had what can be considered an off-by-one-error: 981355714Skris Its depth (which was not part of the external interface) 981455714Skris was actually counting the number of certificates in a chain; 981555714Skris now it really counts the depth. 981655714Skris [Bodo Moeller] 981755714Skris 981855714Skris *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used 981955714Skris instead of X509err, which often resulted in confusing error 982055714Skris messages since the error codes are not globally unique 982155714Skris (e.g. an alleged error in ssl3_accept when a certificate 982255714Skris didn't match the private key). 982355714Skris 982455714Skris *) New function SSL_CTX_set_session_id_context that allows to set a default 982555714Skris value (so that you don't need SSL_set_session_id_context for each 982655714Skris connection using the SSL_CTX). 982755714Skris [Bodo Moeller] 982855714Skris 982955714Skris *) OAEP decoding bug fix. 9830291721Sjkim [Ulf M��ller] 983155714Skris 983255714Skris *) Support INSTALL_PREFIX for package builders, as proposed by 983355714Skris David Harris. 983455714Skris [Bodo Moeller] 983555714Skris 983655714Skris *) New Configure options "threads" and "no-threads". For systems 983755714Skris where the proper compiler options are known (currently Solaris 983855714Skris and Linux), "threads" is the default. 983955714Skris [Bodo Moeller] 984055714Skris 984155714Skris *) New script util/mklink.pl as a faster substitute for util/mklink.sh. 984255714Skris [Bodo Moeller] 984355714Skris 984455714Skris *) Install various scripts to $(OPENSSLDIR)/misc, not to 984555714Skris $(INSTALLTOP)/bin -- they shouldn't clutter directories 984655714Skris such as /usr/local/bin. 984755714Skris [Bodo Moeller] 984855714Skris 984955714Skris *) "make linux-shared" to build shared libraries. 985055714Skris [Niels Poppe <niels@netbox.org>] 985155714Skris 985255714Skris *) New Configure option no-<cipher> (rsa, idea, rc5, ...). 9853291721Sjkim [Ulf M��ller] 985455714Skris 985555714Skris *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for 985655714Skris extension adding in x509 utility. 985755714Skris [Steve Henson] 985855714Skris 985955714Skris *) Remove NOPROTO sections and error code comments. 9860291721Sjkim [Ulf M��ller] 986155714Skris 986255714Skris *) Partial rewrite of the DEF file generator to now parse the ANSI 986355714Skris prototypes. 986455714Skris [Steve Henson] 986555714Skris 986655714Skris *) New Configure options --prefix=DIR and --openssldir=DIR. 9867291721Sjkim [Ulf M��ller] 986855714Skris 986955714Skris *) Complete rewrite of the error code script(s). It is all now handled 987055714Skris by one script at the top level which handles error code gathering, 987155714Skris header rewriting and C source file generation. It should be much better 987255714Skris than the old method: it now uses a modified version of Ulf's parser to 987355714Skris read the ANSI prototypes in all header files (thus the old K&R definitions 987455714Skris aren't needed for error creation any more) and do a better job of 987555714Skris translating function codes into names. The old 'ASN1 error code imbedded 987655714Skris in a comment' is no longer necessary and it doesn't use .err files which 987755714Skris have now been deleted. Also the error code call doesn't have to appear all 987855714Skris on one line (which resulted in some large lines...). 987955714Skris [Steve Henson] 988055714Skris 988155714Skris *) Change #include filenames from <foo.h> to <openssl/foo.h>. 988255714Skris [Bodo Moeller] 988355714Skris 988455714Skris *) Change behaviour of ssl2_read when facing length-0 packets: Don't return 988555714Skris 0 (which usually indicates a closed connection), but continue reading. 988655714Skris [Bodo Moeller] 988755714Skris 988855714Skris *) Fix some race conditions. 988955714Skris [Bodo Moeller] 989055714Skris 989155714Skris *) Add support for CRL distribution points extension. Add Certificate 989255714Skris Policies and CRL distribution points documentation. 989355714Skris [Steve Henson] 989455714Skris 989555714Skris *) Move the autogenerated header file parts to crypto/opensslconf.h. 9896291721Sjkim [Ulf M��ller] 989755714Skris 989855714Skris *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of 989955714Skris 8 of keying material. Merlin has also confirmed interop with this fix 990055714Skris between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. 990155714Skris [Merlin Hughes <merlin@baltimore.ie>] 990255714Skris 990355714Skris *) Fix lots of warnings. 990455714Skris [Richard Levitte <levitte@stacken.kth.se>] 990555714Skris 990655714Skris *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if 990755714Skris the directory spec didn't end with a LIST_SEPARATOR_CHAR. 990855714Skris [Richard Levitte <levitte@stacken.kth.se>] 990955714Skris 991055714Skris *) Fix problems with sizeof(long) == 8. 991155714Skris [Andy Polyakov <appro@fy.chalmers.se>] 991255714Skris 991355714Skris *) Change functions to ANSI C. 9914291721Sjkim [Ulf M��ller] 991555714Skris 991655714Skris *) Fix typos in error codes. 9917291721Sjkim [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf M��ller] 991855714Skris 991955714Skris *) Remove defunct assembler files from Configure. 9920291721Sjkim [Ulf M��ller] 992155714Skris 992255714Skris *) SPARC v8 assembler BIGNUM implementation. 992355714Skris [Andy Polyakov <appro@fy.chalmers.se>] 992455714Skris 992555714Skris *) Support for Certificate Policies extension: both print and set. 992655714Skris Various additions to support the r2i method this uses. 992755714Skris [Steve Henson] 992855714Skris 992955714Skris *) A lot of constification, and fix a bug in X509_NAME_oneline() that could 993055714Skris return a const string when you are expecting an allocated buffer. 993155714Skris [Ben Laurie] 993255714Skris 993355714Skris *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE 993455714Skris types DirectoryString and DisplayText. 993555714Skris [Steve Henson] 993655714Skris 993755714Skris *) Add code to allow r2i extensions to access the configuration database, 993855714Skris add an LHASH database driver and add several ctx helper functions. 993955714Skris [Steve Henson] 994055714Skris 994155714Skris *) Fix an evil bug in bn_expand2() which caused various BN functions to 994255714Skris fail when they extended the size of a BIGNUM. 994355714Skris [Steve Henson] 994455714Skris 994555714Skris *) Various utility functions to handle SXNet extension. Modify mkdef.pl to 994655714Skris support typesafe stack. 994755714Skris [Steve Henson] 994855714Skris 994955714Skris *) Fix typo in SSL_[gs]et_options(). 995055714Skris [Nils Frostberg <nils@medcom.se>] 995155714Skris 995255714Skris *) Delete various functions and files that belonged to the (now obsolete) 995355714Skris old X509V3 handling code. 995455714Skris [Steve Henson] 995555714Skris 995655714Skris *) New Configure option "rsaref". 9957291721Sjkim [Ulf M��ller] 995855714Skris 995955714Skris *) Don't auto-generate pem.h. 996055714Skris [Bodo Moeller] 996155714Skris 996255714Skris *) Introduce type-safe ASN.1 SETs. 996355714Skris [Ben Laurie] 996455714Skris 996555714Skris *) Convert various additional casted stacks to type-safe STACK_OF() variants. 996655714Skris [Ben Laurie, Ralf S. Engelschall, Steve Henson] 996755714Skris 996855714Skris *) Introduce type-safe STACKs. This will almost certainly break lots of code 996955714Skris that links with OpenSSL (well at least cause lots of warnings), but fear 997055714Skris not: the conversion is trivial, and it eliminates loads of evil casts. A 997155714Skris few STACKed things have been converted already. Feel free to convert more. 997255714Skris In the fullness of time, I'll do away with the STACK type altogether. 997355714Skris [Ben Laurie] 997455714Skris 997555714Skris *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate 997655714Skris specified in <certfile> by updating the entry in the index.txt file. 997755714Skris This way one no longer has to edit the index.txt file manually for 997855714Skris revoking a certificate. The -revoke option does the gory details now. 997955714Skris [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] 998055714Skris 998155714Skris *) Fix `openssl crl -noout -text' combination where `-noout' killed the 998255714Skris `-text' option at all and this way the `-noout -text' combination was 998355714Skris inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. 998455714Skris [Ralf S. Engelschall] 998555714Skris 998655714Skris *) Make sure a corresponding plain text error message exists for the 998755714Skris X509_V_ERR_CERT_REVOKED/23 error number which can occur when a 998855714Skris verify callback function determined that a certificate was revoked. 998955714Skris [Ralf S. Engelschall] 999055714Skris 999155714Skris *) Bugfix: In test/testenc, don't test "openssl <cipher>" for 999255714Skris ciphers that were excluded, e.g. by -DNO_IDEA. Also, test 999355714Skris all available cipers including rc5, which was forgotten until now. 999455714Skris In order to let the testing shell script know which algorithms 999555714Skris are available, a new (up to now undocumented) command 999655714Skris "openssl list-cipher-commands" is used. 999755714Skris [Bodo Moeller] 999855714Skris 999955714Skris *) Bugfix: s_client occasionally would sleep in select() when 1000055714Skris it should have checked SSL_pending() first. 1000155714Skris [Bodo Moeller] 1000255714Skris 1000355714Skris *) New functions DSA_do_sign and DSA_do_verify to provide access to 1000455714Skris the raw DSA values prior to ASN.1 encoding. 10005291721Sjkim [Ulf M��ller] 1000655714Skris 1000755714Skris *) Tweaks to Configure 1000855714Skris [Niels Poppe <niels@netbox.org>] 1000955714Skris 1001055714Skris *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, 1001155714Skris yet... 1001255714Skris [Steve Henson] 1001355714Skris 1001455714Skris *) New variables $(RANLIB) and $(PERL) in the Makefiles. 10015291721Sjkim [Ulf M��ller] 1001655714Skris 1001755714Skris *) New config option to avoid instructions that are illegal on the 80386. 1001855714Skris The default code is faster, but requires at least a 486. 10019291721Sjkim [Ulf M��ller] 1002055714Skris 1002155714Skris *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and 1002255714Skris SSL2_SERVER_VERSION (not used at all) macros, which are now the 1002355714Skris same as SSL2_VERSION anyway. 1002455714Skris [Bodo Moeller] 1002555714Skris 1002655714Skris *) New "-showcerts" option for s_client. 1002755714Skris [Bodo Moeller] 1002855714Skris 1002955714Skris *) Still more PKCS#12 integration. Add pkcs12 application to openssl 1003055714Skris application. Various cleanups and fixes. 1003155714Skris [Steve Henson] 1003255714Skris 1003355714Skris *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and 1003455714Skris modify error routines to work internally. Add error codes and PBE init 1003555714Skris to library startup routines. 1003655714Skris [Steve Henson] 1003755714Skris 1003855714Skris *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and 1003955714Skris packing functions to asn1 and evp. Changed function names and error 1004055714Skris codes along the way. 1004155714Skris [Steve Henson] 1004255714Skris 1004355714Skris *) PKCS12 integration: and so it begins... First of several patches to 1004455714Skris slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 1004555714Skris objects to objects.h 1004655714Skris [Steve Henson] 1004755714Skris 1004855714Skris *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 1004955714Skris and display support for Thawte strong extranet extension. 1005055714Skris [Steve Henson] 1005155714Skris 1005255714Skris *) Add LinuxPPC support. 1005355714Skris [Jeff Dubrule <igor@pobox.org>] 1005455714Skris 1005555714Skris *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to 1005655714Skris bn_div_words in alpha.s. 1005755714Skris [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] 1005855714Skris 1005955714Skris *) Make sure the RSA OAEP test is skipped under -DRSAref because 1006055714Skris OAEP isn't supported when OpenSSL is built with RSAref. 1006155714Skris [Ulf Moeller <ulf@fitug.de>] 1006255714Skris 1006355714Skris *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 1006455714Skris so they no longer are missing under -DNOPROTO. 1006555714Skris [Soren S. Jorvang <soren@t.dk>] 1006655714Skris 1006755714Skris 1006855714Skris Changes between 0.9.1c and 0.9.2b [22 Mar 1999] 1006955714Skris 1007055714Skris *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still 1007155714Skris doesn't work when the session is reused. Coming soon! 1007255714Skris [Ben Laurie] 1007355714Skris 1007455714Skris *) Fix a security hole, that allows sessions to be reused in the wrong 1007555714Skris context thus bypassing client cert protection! All software that uses 1007655714Skris client certs and session caches in multiple contexts NEEDS PATCHING to 1007755714Skris allow session reuse! A fuller solution is in the works. 1007855714Skris [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] 1007955714Skris 1008055714Skris *) Some more source tree cleanups (removed obsolete files 1008155714Skris crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed 1008255714Skris permission on "config" script to be executable) and a fix for the INSTALL 1008355714Skris document. 1008455714Skris [Ulf Moeller <ulf@fitug.de>] 1008555714Skris 1008655714Skris *) Remove some legacy and erroneous uses of malloc, free instead of 1008755714Skris Malloc, Free. 1008855714Skris [Lennart Bang <lob@netstream.se>, with minor changes by Steve] 1008955714Skris 1009055714Skris *) Make rsa_oaep_test return non-zero on error. 1009155714Skris [Ulf Moeller <ulf@fitug.de>] 1009255714Skris 1009355714Skris *) Add support for native Solaris shared libraries. Configure 1009455714Skris solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice 1009555714Skris if someone would make that last step automatic. 1009655714Skris [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] 1009755714Skris 1009855714Skris *) ctx_size was not built with the right compiler during "make links". Fixed. 1009955714Skris [Ben Laurie] 1010055714Skris 1010155714Skris *) Change the meaning of 'ALL' in the cipher list. It now means "everything 1010255714Skris except NULL ciphers". This means the default cipher list will no longer 1010355714Skris enable NULL ciphers. They need to be specifically enabled e.g. with 1010455714Skris the string "DEFAULT:eNULL". 1010555714Skris [Steve Henson] 1010655714Skris 1010755714Skris *) Fix to RSA private encryption routines: if p < q then it would 1010855714Skris occasionally produce an invalid result. This will only happen with 1010955714Skris externally generated keys because OpenSSL (and SSLeay) ensure p > q. 1011055714Skris [Steve Henson] 1011155714Skris 1011255714Skris *) Be less restrictive and allow also `perl util/perlpath.pl 1011355714Skris /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', 1011455714Skris because this way one can also use an interpreter named `perl5' (which is 1011555714Skris usually the name of Perl 5.xxx on platforms where an Perl 4.x is still 1011655714Skris installed as `perl'). 1011755714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 1011855714Skris 1011955714Skris *) Let util/clean-depend.pl work also with older Perl 5.00x versions. 1012055714Skris [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 1012155714Skris 1012255714Skris *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add 1012355714Skris advapi32.lib to Win32 build and change the pem test comparision 1012455714Skris to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the 1012555714Skris suggestion). Fix misplaced ASNI prototypes and declarations in evp.h 1012655714Skris and crypto/des/ede_cbcm_enc.c. 1012755714Skris [Steve Henson] 1012855714Skris 1012955714Skris *) DES quad checksum was broken on big-endian architectures. Fixed. 1013055714Skris [Ben Laurie] 1013155714Skris 1013255714Skris *) Comment out two functions in bio.h that aren't implemented. Fix up the 1013355714Skris Win32 test batch file so it (might) work again. The Win32 test batch file 1013455714Skris is horrible: I feel ill.... 1013555714Skris [Steve Henson] 1013655714Skris 1013755714Skris *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected 1013855714Skris in e_os.h. Audit of header files to check ANSI and non ANSI 1013955714Skris sections: 10 functions were absent from non ANSI section and not exported 1014055714Skris from Windows DLLs. Fixed up libeay.num for new functions. 1014155714Skris [Steve Henson] 1014255714Skris 1014355714Skris *) Make `openssl version' output lines consistent. 1014455714Skris [Ralf S. Engelschall] 1014555714Skris 1014655714Skris *) Fix Win32 symbol export lists for BIO functions: Added 1014755714Skris BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data 1014855714Skris to ms/libeay{16,32}.def. 1014955714Skris [Ralf S. Engelschall] 1015055714Skris 1015155714Skris *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled 1015255714Skris fine under Unix and passes some trivial tests I've now added. But the 1015355714Skris whole stuff is horribly incomplete, so a README.1ST with a disclaimer was 1015455714Skris added to make sure no one expects that this stuff really works in the 1015555714Skris OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources 1015655714Skris up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and 1015755714Skris openssl_bio.xs. 1015855714Skris [Ralf S. Engelschall] 1015955714Skris 1016055714Skris *) Fix the generation of two part addresses in perl. 1016155714Skris [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] 1016255714Skris 1016355714Skris *) Add config entry for Linux on MIPS. 1016455714Skris [John Tobey <jtobey@channel1.com>] 1016555714Skris 1016655714Skris *) Make links whenever Configure is run, unless we are on Windoze. 1016755714Skris [Ben Laurie] 1016855714Skris 1016955714Skris *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. 1017055714Skris Currently only issuerAltName and AuthorityKeyIdentifier make any sense 1017155714Skris in CRLs. 1017255714Skris [Steve Henson] 1017355714Skris 1017455714Skris *) Add a useful kludge to allow package maintainers to specify compiler and 1017555714Skris other platforms details on the command line without having to patch the 1017655714Skris Configure script everytime: One now can use ``perl Configure 1017755714Skris <id>:<details>'', i.e. platform ids are allowed to have details appended 1017855714Skris to them (seperated by colons). This is treated as there would be a static 1017955714Skris pre-configured entry in Configure's %table under key <id> with value 1018055714Skris <details> and ``perl Configure <id>'' is called. So, when you want to 1018155714Skris perform a quick test-compile under FreeBSD 3.1 with pgcc and without 1018255714Skris assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' 1018355714Skris now, which overrides the FreeBSD-elf entry on-the-fly. 1018455714Skris [Ralf S. Engelschall] 1018555714Skris 1018655714Skris *) Disable new TLS1 ciphersuites by default: they aren't official yet. 1018755714Skris [Ben Laurie] 1018855714Skris 1018955714Skris *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified 1019055714Skris on the `perl Configure ...' command line. This way one can compile 1019155714Skris OpenSSL libraries with Position Independent Code (PIC) which is needed 1019255714Skris for linking it into DSOs. 1019355714Skris [Ralf S. Engelschall] 1019455714Skris 1019555714Skris *) Remarkably, export ciphers were totally broken and no-one had noticed! 1019655714Skris Fixed. 1019755714Skris [Ben Laurie] 1019855714Skris 1019955714Skris *) Cleaned up the LICENSE document: The official contact for any license 1020055714Skris questions now is the OpenSSL core team under openssl-core@openssl.org. 1020155714Skris And add a paragraph about the dual-license situation to make sure people 1020255714Skris recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply 1020355714Skris to the OpenSSL toolkit. 1020455714Skris [Ralf S. Engelschall] 1020555714Skris 1020655714Skris *) General source tree makefile cleanups: Made `making xxx in yyy...' 1020755714Skris display consistent in the source tree and replaced `/bin/rm' by `rm'. 1020855714Skris Additonally cleaned up the `make links' target: Remove unnecessary 1020955714Skris semicolons, subsequent redundant removes, inline point.sh into mklink.sh 1021055714Skris to speed processing and no longer clutter the display with confusing 1021155714Skris stuff. Instead only the actually done links are displayed. 1021255714Skris [Ralf S. Engelschall] 1021355714Skris 1021455714Skris *) Permit null encryption ciphersuites, used for authentication only. It used 1021555714Skris to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. 1021655714Skris It is now necessary to set SSL_FORBID_ENULL to prevent the use of null 1021755714Skris encryption. 1021855714Skris [Ben Laurie] 1021955714Skris 1022055714Skris *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder 1022155714Skris signed attributes when verifying signatures (this would break them), 1022255714Skris the detached data encoding was wrong and public keys obtained using 1022355714Skris X509_get_pubkey() weren't freed. 1022455714Skris [Steve Henson] 1022555714Skris 1022655714Skris *) Add text documentation for the BUFFER functions. Also added a work around 1022755714Skris to a Win95 console bug. This was triggered by the password read stuff: the 1022855714Skris last character typed gets carried over to the next fread(). If you were 1022955714Skris generating a new cert request using 'req' for example then the last 1023055714Skris character of the passphrase would be CR which would then enter the first 1023155714Skris field as blank. 1023255714Skris [Steve Henson] 1023355714Skris 1023455714Skris *) Added the new `Includes OpenSSL Cryptography Software' button as 1023555714Skris doc/openssl_button.{gif,html} which is similar in style to the old SSLeay 1023655714Skris button and can be used by applications based on OpenSSL to show the 1023755714Skris relationship to the OpenSSL project. 1023855714Skris [Ralf S. Engelschall] 1023955714Skris 1024055714Skris *) Remove confusing variables in function signatures in files 1024155714Skris ssl/ssl_lib.c and ssl/ssl.h. 1024255714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 1024355714Skris 1024455714Skris *) Don't install bss_file.c under PREFIX/include/ 1024555714Skris [Lennart Bong <lob@kulthea.stacken.kth.se>] 1024655714Skris 1024755714Skris *) Get the Win32 compile working again. Modify mkdef.pl so it can handle 1024855714Skris functions that return function pointers and has support for NT specific 1024955714Skris stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various 1025055714Skris #ifdef WIN32 and WINNTs sprinkled about the place and some changes from 1025155714Skris unsigned to signed types: this was killing the Win32 compile. 1025255714Skris [Steve Henson] 1025355714Skris 1025455714Skris *) Add new certificate file to stack functions, 1025555714Skris SSL_add_dir_cert_subjects_to_stack() and 1025655714Skris SSL_add_file_cert_subjects_to_stack(). These largely supplant 1025755714Skris SSL_load_client_CA_file(), and can be used to add multiple certs easily 1025855714Skris to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). 1025955714Skris This means that Apache-SSL and similar packages don't have to mess around 1026055714Skris to add as many CAs as they want to the preferred list. 1026155714Skris [Ben Laurie] 1026255714Skris 1026355714Skris *) Experiment with doxygen documentation. Currently only partially applied to 1026455714Skris ssl/ssl_lib.c. 1026555714Skris See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with 1026655714Skris openssl.doxy as the configuration file. 1026755714Skris [Ben Laurie] 1026855714Skris 1026955714Skris *) Get rid of remaining C++-style comments which strict C compilers hate. 1027055714Skris [Ralf S. Engelschall, pointed out by Carlos Amengual] 1027155714Skris 1027255714Skris *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not 1027355714Skris compiled in by default: it has problems with large keys. 1027455714Skris [Steve Henson] 1027555714Skris 1027655714Skris *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and 1027755714Skris DH private keys and/or callback functions which directly correspond to 1027855714Skris their SSL_CTX_xxx() counterparts but work on a per-connection basis. This 1027955714Skris is needed for applications which have to configure certificates on a 1028055714Skris per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis 1028155714Skris (e.g. s_server). 1028255714Skris For the RSA certificate situation is makes no difference, but 1028355714Skris for the DSA certificate situation this fixes the "no shared cipher" 1028455714Skris problem where the OpenSSL cipher selection procedure failed because the 1028555714Skris temporary keys were not overtaken from the context and the API provided 1028655714Skris no way to reconfigure them. 1028755714Skris The new functions now let applications reconfigure the stuff and they 1028855714Skris are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, 1028955714Skris SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new 1029055714Skris non-public-API function ssl_cert_instantiate() is used as a helper 1029155714Skris function and also to reduce code redundancy inside ssl_rsa.c. 1029255714Skris [Ralf S. Engelschall] 1029355714Skris 1029455714Skris *) Move s_server -dcert and -dkey options out of the undocumented feature 1029555714Skris area because they are useful for the DSA situation and should be 1029655714Skris recognized by the users. 1029755714Skris [Ralf S. Engelschall] 1029855714Skris 1029955714Skris *) Fix the cipher decision scheme for export ciphers: the export bits are 1030055714Skris *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within 1030155714Skris SSL_EXP_MASK. So, the original variable has to be used instead of the 1030255714Skris already masked variable. 1030355714Skris [Richard Levitte <levitte@stacken.kth.se>] 1030455714Skris 1030555714Skris *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c 1030655714Skris [Richard Levitte <levitte@stacken.kth.se>] 1030755714Skris 1030855714Skris *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() 1030955714Skris from `int' to `unsigned int' because it's a length and initialized by 1031055714Skris EVP_DigestFinal() which expects an `unsigned int *'. 1031155714Skris [Richard Levitte <levitte@stacken.kth.se>] 1031255714Skris 1031355714Skris *) Don't hard-code path to Perl interpreter on shebang line of Configure 1031455714Skris script. Instead use the usual Shell->Perl transition trick. 1031555714Skris [Ralf S. Engelschall] 1031655714Skris 1031755714Skris *) Make `openssl x509 -noout -modulus' functional also for DSA certificates 1031855714Skris (in addition to RSA certificates) to match the behaviour of `openssl dsa 1031955714Skris -noout -modulus' as it's already the case for `openssl rsa -noout 1032055714Skris -modulus'. For RSA the -modulus is the real "modulus" while for DSA 1032155714Skris currently the public key is printed (a decision which was already done by 1032255714Skris `openssl dsa -modulus' in the past) which serves a similar purpose. 1032355714Skris Additionally the NO_RSA no longer completely removes the whole -modulus 1032455714Skris option; it now only avoids using the RSA stuff. Same applies to NO_DSA 1032555714Skris now, too. 1032655714Skris [Ralf S. Engelschall] 1032755714Skris 1032855714Skris *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested 1032955714Skris BIO. See the source (crypto/evp/bio_ok.c) for more info. 1033055714Skris [Arne Ansper <arne@ats.cyber.ee>] 1033155714Skris 1033255714Skris *) Dump the old yucky req code that tried (and failed) to allow raw OIDs 1033355714Skris to be added. Now both 'req' and 'ca' can use new objects defined in the 1033455714Skris config file. 1033555714Skris [Steve Henson] 1033655714Skris 1033755714Skris *) Add cool BIO that does syslog (or event log on NT). 1033855714Skris [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] 1033955714Skris 1034055714Skris *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, 1034155714Skris TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and 1034255714Skris TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher 1034355714Skris Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. 1034455714Skris [Ben Laurie] 1034555714Skris 1034655714Skris *) Add preliminary config info for new extension code. 1034755714Skris [Steve Henson] 1034855714Skris 1034955714Skris *) Make RSA_NO_PADDING really use no padding. 1035055714Skris [Ulf Moeller <ulf@fitug.de>] 1035155714Skris 1035255714Skris *) Generate errors when private/public key check is done. 1035355714Skris [Ben Laurie] 1035455714Skris 1035555714Skris *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support 1035655714Skris for some CRL extensions and new objects added. 1035755714Skris [Steve Henson] 1035855714Skris 1035955714Skris *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private 1036055714Skris key usage extension and fuller support for authority key id. 1036155714Skris [Steve Henson] 1036255714Skris 1036355714Skris *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved 1036455714Skris padding method for RSA, which is recommended for new applications in PKCS 1036555714Skris #1 v2.0 (RFC 2437, October 1998). 1036655714Skris OAEP (Optimal Asymmetric Encryption Padding) has better theoretical 1036755714Skris foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure 1036855714Skris against Bleichbacher's attack on RSA. 1036955714Skris [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by 1037055714Skris Ben Laurie] 1037155714Skris 1037255714Skris *) Updates to the new SSL compression code 1037355714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1037455714Skris 1037555714Skris *) Fix so that the version number in the master secret, when passed 1037655714Skris via RSA, checks that if TLS was proposed, but we roll back to SSLv3 1037755714Skris (because the server will not accept higher), that the version number 1037855714Skris is 0x03,0x01, not 0x03,0x00 1037955714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1038055714Skris 1038155714Skris *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory 1038255714Skris leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes 1038355714Skris in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c 1038455714Skris [Steve Henson] 1038555714Skris 1038655714Skris *) Support for RAW extensions where an arbitrary extension can be 1038755714Skris created by including its DER encoding. See apps/openssl.cnf for 1038855714Skris an example. 1038955714Skris [Steve Henson] 1039055714Skris 1039155714Skris *) Make sure latest Perl versions don't interpret some generated C array 1039255714Skris code as Perl array code in the crypto/err/err_genc.pl script. 1039355714Skris [Lars Weber <3weber@informatik.uni-hamburg.de>] 1039455714Skris 1039555714Skris *) Modify ms/do_ms.bat to not generate assembly language makefiles since 1039655714Skris not many people have the assembler. Various Win32 compilation fixes and 1039755714Skris update to the INSTALL.W32 file with (hopefully) more accurate Win32 1039855714Skris build instructions. 1039955714Skris [Steve Henson] 1040055714Skris 1040155714Skris *) Modify configure script 'Configure' to automatically create crypto/date.h 1040255714Skris file under Win32 and also build pem.h from pem.org. New script 1040355714Skris util/mkfiles.pl to create the MINFO file on environments that can't do a 1040455714Skris 'make files': perl util/mkfiles.pl >MINFO should work. 1040555714Skris [Steve Henson] 1040655714Skris 1040755714Skris *) Major rework of DES function declarations, in the pursuit of correctness 1040855714Skris and purity. As a result, many evil casts evaporated, and some weirdness, 1040955714Skris too. You may find this causes warnings in your code. Zapping your evil 1041055714Skris casts will probably fix them. Mostly. 1041155714Skris [Ben Laurie] 1041255714Skris 1041355714Skris *) Fix for a typo in asn1.h. Bug fix to object creation script 1041455714Skris obj_dat.pl. It considered a zero in an object definition to mean 1041555714Skris "end of object": none of the objects in objects.h have any zeros 1041655714Skris so it wasn't spotted. 1041755714Skris [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] 1041855714Skris 1041955714Skris *) Add support for Triple DES Cipher Block Chaining with Output Feedback 1042055714Skris Masking (CBCM). In the absence of test vectors, the best I have been able 1042155714Skris to do is check that the decrypt undoes the encrypt, so far. Send me test 1042255714Skris vectors if you have them. 1042355714Skris [Ben Laurie] 1042455714Skris 1042555714Skris *) Correct calculation of key length for export ciphers (too much space was 1042655714Skris allocated for null ciphers). This has not been tested! 1042755714Skris [Ben Laurie] 1042855714Skris 1042955714Skris *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage 1043055714Skris message is now correct (it understands "crypto" and "ssl" on its 1043155714Skris command line). There is also now an "update" option. This will update 1043255714Skris the util/ssleay.num and util/libeay.num files with any new functions. 1043355714Skris If you do a: 1043455714Skris perl util/mkdef.pl crypto ssl update 1043555714Skris it will update them. 1043655714Skris [Steve Henson] 1043755714Skris 1043855714Skris *) Overhauled the Perl interface (perl/*): 1043955714Skris - ported BN stuff to OpenSSL's different BN library 1044055714Skris - made the perl/ source tree CVS-aware 1044155714Skris - renamed the package from SSLeay to OpenSSL (the files still contain 1044255714Skris their history because I've copied them in the repository) 1044355714Skris - removed obsolete files (the test scripts will be replaced 1044455714Skris by better Test::Harness variants in the future) 1044555714Skris [Ralf S. Engelschall] 1044655714Skris 1044755714Skris *) First cut for a very conservative source tree cleanup: 1044855714Skris 1. merge various obsolete readme texts into doc/ssleay.txt 1044955714Skris where we collect the old documents and readme texts. 1045055714Skris 2. remove the first part of files where I'm already sure that we no 1045155714Skris longer need them because of three reasons: either they are just temporary 1045255714Skris files which were left by Eric or they are preserved original files where 1045355714Skris I've verified that the diff is also available in the CVS via "cvs diff 1045455714Skris -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for 1045555714Skris the crypto/md/ stuff). 1045655714Skris [Ralf S. Engelschall] 1045755714Skris 1045855714Skris *) More extension code. Incomplete support for subject and issuer alt 1045955714Skris name, issuer and authority key id. Change the i2v function parameters 1046055714Skris and add an extra 'crl' parameter in the X509V3_CTX structure: guess 1046155714Skris what that's for :-) Fix to ASN1 macro which messed up 1046255714Skris IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. 1046355714Skris [Steve Henson] 1046455714Skris 1046555714Skris *) Preliminary support for ENUMERATED type. This is largely copied from the 1046655714Skris INTEGER code. 1046755714Skris [Steve Henson] 1046855714Skris 1046955714Skris *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. 1047055714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1047155714Skris 1047255714Skris *) Make sure `make rehash' target really finds the `openssl' program. 1047355714Skris [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] 1047455714Skris 1047555714Skris *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd 1047655714Skris like to hear about it if this slows down other processors. 1047755714Skris [Ben Laurie] 1047855714Skris 1047955714Skris *) Add CygWin32 platform information to Configure script. 1048055714Skris [Alan Batie <batie@aahz.jf.intel.com>] 1048155714Skris 1048255714Skris *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' 1048355714Skris [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] 1048455714Skris 1048555714Skris *) New program nseq to manipulate netscape certificate sequences 1048655714Skris [Steve Henson] 1048755714Skris 1048855714Skris *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a 1048955714Skris few typos. 1049055714Skris [Steve Henson] 1049155714Skris 1049255714Skris *) Fixes to BN code. Previously the default was to define BN_RECURSION 1049355714Skris but the BN code had some problems that would cause failures when 1049455714Skris doing certificate verification and some other functions. 1049555714Skris [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] 1049655714Skris 1049755714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 1049855714Skris [Steve Henson] 1049955714Skris 1050055714Skris *) Add ASN1 and PEM code to support netscape certificate sequences. 1050155714Skris [Steve Henson] 1050255714Skris 1050355714Skris *) Add several PKIX and private extended key usage OIDs. 1050455714Skris [Steve Henson] 1050555714Skris 1050655714Skris *) Modify the 'ca' program to handle the new extension code. Modify 1050755714Skris openssl.cnf for new extension format, add comments. 1050855714Skris [Steve Henson] 1050955714Skris 1051055714Skris *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' 1051155714Skris and add a sample to openssl.cnf so req -x509 now adds appropriate 1051255714Skris CA extensions. 1051355714Skris [Steve Henson] 1051455714Skris 1051555714Skris *) Continued X509 V3 changes. Add to other makefiles, integrate with the 1051655714Skris error code, add initial support to X509_print() and x509 application. 1051755714Skris [Steve Henson] 1051855714Skris 1051955714Skris *) Takes a deep breath and start addding X509 V3 extension support code. Add 1052055714Skris files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this 1052155714Skris stuff is currently isolated and isn't even compiled yet. 1052255714Skris [Steve Henson] 1052355714Skris 1052455714Skris *) Continuing patches for GeneralizedTime. Fix up certificate and CRL 1052555714Skris ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. 1052655714Skris Removed the versions check from X509 routines when loading extensions: 1052755714Skris this allows certain broken certificates that don't set the version 1052855714Skris properly to be processed. 1052955714Skris [Steve Henson] 1053055714Skris 1053155714Skris *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another 1053255714Skris Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which 1053355714Skris can still be regenerated with "make depend". 1053455714Skris [Ben Laurie] 1053555714Skris 1053655714Skris *) Spelling mistake in C version of CAST-128. 1053755714Skris [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] 1053855714Skris 1053955714Skris *) Changes to the error generation code. The perl script err-code.pl 1054055714Skris now reads in the old error codes and retains the old numbers, only 1054155714Skris adding new ones if necessary. It also only changes the .err files if new 1054255714Skris codes are added. The makefiles have been modified to only insert errors 1054355714Skris when needed (to avoid needlessly modifying header files). This is done 1054455714Skris by only inserting errors if the .err file is newer than the auto generated 1054555714Skris C file. To rebuild all the error codes from scratch (the old behaviour) 1054655714Skris either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl 1054755714Skris or delete all the .err files. 1054855714Skris [Steve Henson] 1054955714Skris 1055055714Skris *) CAST-128 was incorrectly implemented for short keys. The C version has 1055155714Skris been fixed, but is untested. The assembler versions are also fixed, but 1055255714Skris new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing 1055355714Skris to regenerate it if needed. 1055455714Skris [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun 1055555714Skris Hagino <itojun@kame.net>] 1055655714Skris 1055755714Skris *) File was opened incorrectly in randfile.c. 10558291721Sjkim [Ulf M��ller <ulf@fitug.de>] 1055955714Skris 1056055714Skris *) Beginning of support for GeneralizedTime. d2i, i2d, check and print 1056155714Skris functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or 1056255714Skris GeneralizedTime. ASN1_TIME is the proper type used in certificates et 1056355714Skris al: it's just almost always a UTCTime. Note this patch adds new error 1056455714Skris codes so do a "make errors" if there are problems. 1056555714Skris [Steve Henson] 1056655714Skris 1056755714Skris *) Correct Linux 1 recognition in config. 10568291721Sjkim [Ulf M��ller <ulf@fitug.de>] 1056955714Skris 1057055714Skris *) Remove pointless MD5 hash when using DSA keys in ca. 1057155714Skris [Anonymous <nobody@replay.com>] 1057255714Skris 1057355714Skris *) Generate an error if given an empty string as a cert directory. Also 1057455714Skris generate an error if handed NULL (previously returned 0 to indicate an 1057555714Skris error, but didn't set one). 1057655714Skris [Ben Laurie, reported by Anonymous <nobody@replay.com>] 1057755714Skris 1057855714Skris *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. 1057955714Skris [Ben Laurie] 1058055714Skris 1058155714Skris *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct 1058255714Skris parameters. This was causing a warning which killed off the Win32 compile. 1058355714Skris [Steve Henson] 1058455714Skris 1058555714Skris *) Remove C++ style comments from crypto/bn/bn_local.h. 1058655714Skris [Neil Costigan <neil.costigan@celocom.com>] 1058755714Skris 1058855714Skris *) The function OBJ_txt2nid was broken. It was supposed to return a nid 1058955714Skris based on a text string, looking up short and long names and finally 1059055714Skris "dot" format. The "dot" format stuff didn't work. Added new function 1059155714Skris OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 1059255714Skris OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the 1059355714Skris OID is not part of the table. 1059455714Skris [Steve Henson] 1059555714Skris 1059655714Skris *) Add prototypes to X509 lookup/verify methods, fixing a bug in 1059755714Skris X509_LOOKUP_by_alias(). 1059855714Skris [Ben Laurie] 1059955714Skris 1060055714Skris *) Sort openssl functions by name. 1060155714Skris [Ben Laurie] 1060255714Skris 1060355714Skris *) Get the gendsa program working (hopefully) and add it to app list. Remove 1060455714Skris encryption from sample DSA keys (in case anyone is interested the password 1060555714Skris was "1234"). 1060655714Skris [Steve Henson] 1060755714Skris 1060855714Skris *) Make _all_ *_free functions accept a NULL pointer. 1060955714Skris [Frans Heymans <fheymans@isaserver.be>] 1061055714Skris 1061155714Skris *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use 1061255714Skris NULL pointers. 1061355714Skris [Anonymous <nobody@replay.com>] 1061455714Skris 1061555714Skris *) s_server should send the CAfile as acceptable CAs, not its own cert. 1061655714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1061755714Skris 1061855714Skris *) Don't blow it for numeric -newkey arguments to apps/req. 1061955714Skris [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] 1062055714Skris 1062155714Skris *) Temp key "for export" tests were wrong in s3_srvr.c. 1062255714Skris [Anonymous <nobody@replay.com>] 1062355714Skris 1062455714Skris *) Add prototype for temp key callback functions 1062555714Skris SSL_CTX_set_tmp_{rsa,dh}_callback(). 1062655714Skris [Ben Laurie] 1062755714Skris 1062855714Skris *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and 1062955714Skris DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). 1063055714Skris [Steve Henson] 1063155714Skris 1063255714Skris *) X509_name_add_entry() freed the wrong thing after an error. 1063355714Skris [Arne Ansper <arne@ats.cyber.ee>] 1063455714Skris 1063555714Skris *) rsa_eay.c would attempt to free a NULL context. 1063655714Skris [Arne Ansper <arne@ats.cyber.ee>] 1063755714Skris 1063855714Skris *) BIO_s_socket() had a broken should_retry() on Windoze. 1063955714Skris [Arne Ansper <arne@ats.cyber.ee>] 1064055714Skris 1064155714Skris *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. 1064255714Skris [Arne Ansper <arne@ats.cyber.ee>] 1064355714Skris 1064455714Skris *) Make sure the already existing X509_STORE->depth variable is initialized 1064555714Skris in X509_STORE_new(), but document the fact that this variable is still 1064655714Skris unused in the certificate verification process. 1064755714Skris [Ralf S. Engelschall] 1064855714Skris 1064955714Skris *) Fix the various library and apps files to free up pkeys obtained from 1065055714Skris X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. 1065155714Skris [Steve Henson] 1065255714Skris 1065355714Skris *) Fix reference counting in X509_PUBKEY_get(). This makes 1065455714Skris demos/maurice/example2.c work, amongst others, probably. 1065555714Skris [Steve Henson and Ben Laurie] 1065655714Skris 1065755714Skris *) First cut of a cleanup for apps/. First the `ssleay' program is now named 1065855714Skris `openssl' and second, the shortcut symlinks for the `openssl <command>' 1065955714Skris are no longer created. This way we have a single and consistent command 1066055714Skris line interface `openssl <command>', similar to `cvs <command>'. 1066155714Skris [Ralf S. Engelschall, Paul Sutton and Ben Laurie] 1066255714Skris 1066355714Skris *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey 1066455714Skris BIT STRING wrapper always have zero unused bits. 1066555714Skris [Steve Henson] 1066655714Skris 1066755714Skris *) Add CA.pl, perl version of CA.sh, add extended key usage OID. 1066855714Skris [Steve Henson] 1066955714Skris 1067055714Skris *) Make the top-level INSTALL documentation easier to understand. 1067155714Skris [Paul Sutton] 1067255714Skris 1067355714Skris *) Makefiles updated to exit if an error occurs in a sub-directory 1067455714Skris make (including if user presses ^C) [Paul Sutton] 1067555714Skris 1067655714Skris *) Make Montgomery context stuff explicit in RSA data structure. 1067755714Skris [Ben Laurie] 1067855714Skris 1067955714Skris *) Fix build order of pem and err to allow for generated pem.h. 1068055714Skris [Ben Laurie] 1068155714Skris 1068255714Skris *) Fix renumbering bug in X509_NAME_delete_entry(). 1068355714Skris [Ben Laurie] 1068455714Skris 1068555714Skris *) Enhanced the err-ins.pl script so it makes the error library number 1068655714Skris global and can add a library name. This is needed for external ASN1 and 1068755714Skris other error libraries. 1068855714Skris [Steve Henson] 1068955714Skris 1069055714Skris *) Fixed sk_insert which never worked properly. 1069155714Skris [Steve Henson] 1069255714Skris 1069355714Skris *) Fix ASN1 macros so they can handle indefinite length construted 1069455714Skris EXPLICIT tags. Some non standard certificates use these: they can now 1069555714Skris be read in. 1069655714Skris [Steve Henson] 1069755714Skris 1069855714Skris *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) 1069955714Skris into a single doc/ssleay.txt bundle. This way the information is still 1070055714Skris preserved but no longer messes up this directory. Now it's new room for 1070155714Skris the new set of documenation files. 1070255714Skris [Ralf S. Engelschall] 1070355714Skris 1070455714Skris *) SETs were incorrectly DER encoded. This was a major pain, because they 1070555714Skris shared code with SEQUENCEs, which aren't coded the same. This means that 1070655714Skris almost everything to do with SETs or SEQUENCEs has either changed name or 1070755714Skris number of arguments. 1070855714Skris [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] 1070955714Skris 1071055714Skris *) Fix test data to work with the above. 1071155714Skris [Ben Laurie] 1071255714Skris 1071355714Skris *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but 1071455714Skris was already fixed by Eric for 0.9.1 it seems. 10715291721Sjkim [Ben Laurie - pointed out by Ulf M��ller <ulf@fitug.de>] 1071655714Skris 1071755714Skris *) Autodetect FreeBSD3. 1071855714Skris [Ben Laurie] 1071955714Skris 1072055714Skris *) Fix various bugs in Configure. This affects the following platforms: 1072155714Skris nextstep 1072255714Skris ncr-scde 1072355714Skris unixware-2.0 1072455714Skris unixware-2.0-pentium 1072555714Skris sco5-cc. 1072655714Skris [Ben Laurie] 1072755714Skris 1072855714Skris *) Eliminate generated files from CVS. Reorder tests to regenerate files 1072955714Skris before they are needed. 1073055714Skris [Ben Laurie] 1073155714Skris 1073255714Skris *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). 1073355714Skris [Ben Laurie] 1073455714Skris 1073555714Skris 1073655714Skris Changes between 0.9.1b and 0.9.1c [23-Dec-1998] 1073755714Skris 1073855714Skris *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 1073955714Skris changed SSLeay to OpenSSL in version strings. 1074055714Skris [Ralf S. Engelschall] 1074155714Skris 1074255714Skris *) Some fixups to the top-level documents. 1074355714Skris [Paul Sutton] 1074455714Skris 1074555714Skris *) Fixed the nasty bug where rsaref.h was not found under compile-time 1074655714Skris because the symlink to include/ was missing. 1074755714Skris [Ralf S. Engelschall] 1074855714Skris 1074955714Skris *) Incorporated the popular no-RSA/DSA-only patches 1075055714Skris which allow to compile a RSA-free SSLeay. 1075155714Skris [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] 1075255714Skris 1075355714Skris *) Fixed nasty rehash problem under `make -f Makefile.ssl links' 1075455714Skris when "ssleay" is still not found. 1075555714Skris [Ralf S. Engelschall] 1075655714Skris 1075755714Skris *) Added more platforms to Configure: Cray T3E, HPUX 11, 1075855714Skris [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] 1075955714Skris 1076055714Skris *) Updated the README file. 1076155714Skris [Ralf S. Engelschall] 1076255714Skris 1076355714Skris *) Added various .cvsignore files in the CVS repository subdirs 1076455714Skris to make a "cvs update" really silent. 1076555714Skris [Ralf S. Engelschall] 1076655714Skris 1076755714Skris *) Recompiled the error-definition header files and added 1076855714Skris missing symbols to the Win32 linker tables. 1076955714Skris [Ralf S. Engelschall] 1077055714Skris 1077155714Skris *) Cleaned up the top-level documents; 1077255714Skris o new files: CHANGES and LICENSE 1077355714Skris o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 1077455714Skris o merged COPYRIGHT into LICENSE 1077555714Skris o removed obsolete TODO file 1077655714Skris o renamed MICROSOFT to INSTALL.W32 1077755714Skris [Ralf S. Engelschall] 1077855714Skris 1077955714Skris *) Removed dummy files from the 0.9.1b source tree: 1078055714Skris crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi 1078155714Skris crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f 1078255714Skris crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f 1078355714Skris crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f 1078455714Skris util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f 1078555714Skris [Ralf S. Engelschall] 1078655714Skris 1078755714Skris *) Added various platform portability fixes. 1078855714Skris [Mark J. Cox] 1078955714Skris 1079055714Skris *) The Genesis of the OpenSSL rpject: 1079155714Skris We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. 1079255714Skris Young and Tim J. Hudson created while they were working for C2Net until 1079355714Skris summer 1998. 1079455714Skris [The OpenSSL Project] 1079555714Skris 1079655714Skris 1079755714Skris Changes between 0.9.0b and 0.9.1b [not released] 1079855714Skris 1079955714Skris *) Updated a few CA certificates under certs/ 1080055714Skris [Eric A. Young] 1080155714Skris 1080255714Skris *) Changed some BIGNUM api stuff. 1080355714Skris [Eric A. Young] 1080455714Skris 1080555714Skris *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 1080655714Skris DGUX x86, Linux Alpha, etc. 1080755714Skris [Eric A. Young] 1080855714Skris 1080955714Skris *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 1081055714Skris RLE (dummy implemented) and ZLIB (really implemented when ZLIB is 1081155714Skris available). 1081255714Skris [Eric A. Young] 1081355714Skris 1081455714Skris *) Add -strparse option to asn1pars program which parses nested 1081555714Skris binary structures 1081655714Skris [Dr Stephen Henson <shenson@bigfoot.com>] 1081755714Skris 1081855714Skris *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. 1081955714Skris [Eric A. Young] 1082055714Skris 1082155714Skris *) DSA fix for "ca" program. 1082255714Skris [Eric A. Young] 1082355714Skris 1082455714Skris *) Added "-genkey" option to "dsaparam" program. 1082555714Skris [Eric A. Young] 1082655714Skris 1082755714Skris *) Added RIPE MD160 (rmd160) message digest. 1082855714Skris [Eric A. Young] 1082955714Skris 1083055714Skris *) Added -a (all) option to "ssleay version" command. 1083155714Skris [Eric A. Young] 1083255714Skris 1083355714Skris *) Added PLATFORM define which is the id given to Configure. 1083455714Skris [Eric A. Young] 1083555714Skris 1083655714Skris *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. 1083755714Skris [Eric A. Young] 1083855714Skris 1083955714Skris *) Extended the ASN.1 parser routines. 1084055714Skris [Eric A. Young] 1084155714Skris 1084255714Skris *) Extended BIO routines to support REUSEADDR, seek, tell, etc. 1084355714Skris [Eric A. Young] 1084455714Skris 1084555714Skris *) Added a BN_CTX to the BN library. 1084655714Skris [Eric A. Young] 1084755714Skris 1084855714Skris *) Fixed the weak key values in DES library 1084955714Skris [Eric A. Young] 1085055714Skris 1085155714Skris *) Changed API in EVP library for cipher aliases. 1085255714Skris [Eric A. Young] 1085355714Skris 1085455714Skris *) Added support for RC2/64bit cipher. 1085555714Skris [Eric A. Young] 1085655714Skris 1085755714Skris *) Converted the lhash library to the crypto/mem.c functions. 1085855714Skris [Eric A. Young] 1085955714Skris 1086055714Skris *) Added more recognized ASN.1 object ids. 1086155714Skris [Eric A. Young] 1086255714Skris 1086355714Skris *) Added more RSA padding checks for SSL/TLS. 1086455714Skris [Eric A. Young] 1086555714Skris 1086655714Skris *) Added BIO proxy/filter functionality. 1086755714Skris [Eric A. Young] 1086855714Skris 1086955714Skris *) Added extra_certs to SSL_CTX which can be used 1087055714Skris send extra CA certificates to the client in the CA cert chain sending 1087155714Skris process. It can be configured with SSL_CTX_add_extra_chain_cert(). 1087255714Skris [Eric A. Young] 1087355714Skris 1087455714Skris *) Now Fortezza is denied in the authentication phase because 1087555714Skris this is key exchange mechanism is not supported by SSLeay at all. 1087655714Skris [Eric A. Young] 1087755714Skris 1087855714Skris *) Additional PKCS1 checks. 1087955714Skris [Eric A. Young] 1088055714Skris 1088155714Skris *) Support the string "TLSv1" for all TLS v1 ciphers. 1088255714Skris [Eric A. Young] 1088355714Skris 1088455714Skris *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the 1088555714Skris ex_data index of the SSL context in the X509_STORE_CTX ex_data. 1088655714Skris [Eric A. Young] 1088755714Skris 1088855714Skris *) Fixed a few memory leaks. 1088955714Skris [Eric A. Young] 1089055714Skris 1089155714Skris *) Fixed various code and comment typos. 1089255714Skris [Eric A. Young] 1089355714Skris 1089455714Skris *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 1089555714Skris bytes sent in the client random. 1089655714Skris [Edward Bishop <ebishop@spyglass.com>] 1089755714Skris 10898