5 Changes between 0.9.8k and 0.9.8l [xx XXX xxxx]
7 *) Disable renegotiation completely - this fixes a severe security
8 problem at the cost of breaking all renegotiation. Renegotiation
9 can be re-enabled by setting
10 OPENSSL_ENABLE_UNSAFE_LEGACY_SESSION_RENEGOTATION at
11 compile-time. This is really not recommended.
14 *) Fixes to stateless session resumption handling. Use initial_ctx when
15 issuing and attempting to decrypt tickets in case it has changed during
16 servername handling. Use a non-zero length session ID when attempting
17 stateless session resumption: this makes it possible to determine if
18 a resumption has occurred immediately after receiving server hello
19 (several places in OpenSSL subtly assume this) instead of later in
23 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
24 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
25 fixes for a few places where the return code is not checked
27 [Julia Lawall <julia@diku.dk>]
29 *) Add --strict-warnings option to Configure script to include devteam
30 warnings in other configurations.
33 *) Add support for --libdir option and LIBDIR variable in makefiles. This
34 makes it possible to install openssl libraries in locations which
35 have names other than "lib", for example "/usr/lib64" which some
37 [Steve Henson, based on patch from Jeremy Utley]
39 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
40 X690 8.9.12 and can produce some misleading textual output of OIDs.
41 [Steve Henson, reported by Dan Kaminsky]
43 *) Delete MD2 from algorithm tables. This follows the recommendation in
44 several standards that it is not used in new applications due to
45 several cryptographic weaknesses. For binary compatibility reasons
46 the MD2 API is still compiled in by default.
49 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
53 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
54 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
56 [Guenter <lists@gknw.net>]
58 *) Fix the server certificate chain building code to use X509_verify_cert(),
59 it used to have an ad-hoc builder which was unable to cope with anything
60 other than a simple chain.
61 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
63 *) Don't check self signed certificate signatures in X509_verify_cert()
64 by default (a flag can override this): it just wastes time without
65 adding any security. As a useful side effect self signed root CAs
66 with non-FIPS digests are now usable in FIPS mode.
69 *) In dtls1_process_out_of_seq_message() the check if the current message
70 is already buffered was missing. For every new message was memory
71 allocated, allowing an attacker to perform an denial of service attack
72 with sending out of seq handshake messages until there is no memory
73 left. Additionally every future messege was buffered, even if the
74 sequence number made no sense and would be part of another handshake.
75 So only messages with sequence numbers less than 10 in advance will be
76 buffered. (CVE-2009-1378)
77 [Robin Seggelmann, discovered by Daniel Mentz]
79 *) Records are buffered if they arrive with a future epoch to be
80 processed after finishing the corresponding handshake. There is
81 currently no limitation to this buffer allowing an attacker to perform
82 a DOS attack with sending records with future epochs until there is no
83 memory left. This patch adds the pqueue_size() function to detemine
84 the size of a buffer and limits the record buffer to 100 entries.
86 [Robin Seggelmann, discovered by Daniel Mentz]
88 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
89 parent structure is freed. (CVE-2009-1379)
92 *) Handle non-blocking I/O properly in SSL_shutdown() call.
93 [Darryl Miles <darryl-mailinglists@netbauds.net>]
96 [Ilya O. <vrghost@gmail.com>]
98 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
100 *) Don't set val to NULL when freeing up structures, it is freed up by
101 underlying code. If sizeof(void *) > sizeof(long) this can result in
102 zeroing past the valid field. (CVE-2009-0789)
103 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
105 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
106 checked correctly. This would allow some invalid signed attributes to
107 appear to verify correctly. (CVE-2009-0591)
108 [Ivan Nestlerode <inestlerode@us.ibm.com>]
110 *) Reject UniversalString and BMPString types with invalid lengths. This
111 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
112 a legal length. (CVE-2009-0590)
115 *) Set S/MIME signing as the default purpose rather than setting it
116 unconditionally. This allows applications to override it at the store
120 *) Permit restricted recursion of ASN1 strings. This is needed in practice
121 to handle some structures.
124 *) Improve efficiency of mem_gets: don't search whole buffer each time
126 [Jeremy Shapiro <jnshapir@us.ibm.com>]
128 *) New -hex option for openssl rand.
131 *) Print out UTF8String and NumericString when parsing ASN1.
134 *) Support NumericString type for name components.
137 *) Allow CC in the environment to override the automatically chosen
138 compiler. Note that nothing is done to ensure flags work with the
142 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
144 *) Properly check EVP_VerifyFinal() and similar return values
146 [Ben Laurie, Bodo Moeller, Google Security Team]
148 *) Enable TLS extensions by default.
151 *) Allow the CHIL engine to be loaded, whether the application is
152 multithreaded or not. (This does not release the developer from the
153 obligation to set up the dynamic locking callbacks.)
154 [Sander Temme <sander@temme.net>]
156 *) Use correct exit code if there is an error in dgst command.
157 [Steve Henson; problem pointed out by Roland Dirlewanger]
159 *) Tweak Configure so that you need to say "experimental-jpake" to enable
160 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
163 *) Add experimental JPAKE support, including demo authentication in
164 s_client and s_server.
167 *) Set the comparison function in v3_addr_canonize().
168 [Rob Austein <sra@hactrn.net>]
170 *) Add support for XMPP STARTTLS in s_client.
171 [Philip Paeps <philip@freebsd.org>]
173 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
174 to ensure that even with this option, only ciphersuites in the
175 server's preference list will be accepted. (Note that the option
176 applies only when resuming a session, so the earlier behavior was
177 just about the algorithm choice for symmetric cryptography.)
180 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
182 *) Fix NULL pointer dereference if a DTLS server recieved
183 ChangeCipherSpec as first record (CVE-2009-1386)
184 [Bodo Moeller, discovered by Alex Lam]
186 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
187 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
190 *) The fix in 0.9.8c that supposedly got rid of unsafe
191 double-checked locking was incomplete for RSA blinding,
192 addressing just one layer of what turns out to have been
193 doubly unsafe triple-checked locking.
195 So now fix this for real by retiring the MONT_HELPER macro
196 in crypto/rsa/rsa_eay.c.
198 [Bodo Moeller; problem pointed out by Marius Schilder]
200 *) Various precautionary measures:
202 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
204 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
205 (NB: This would require knowledge of the secret session ticket key
206 to exploit, in which case you'd be SOL either way.)
208 - Change bn_nist.c so that it will properly handle input BIGNUMs
209 outside the expected range.
211 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
214 [Neel Mehta, Bodo Moeller]
216 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
217 the load fails. Useful for distros.
218 [Ben Laurie and the FreeBSD team]
220 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
223 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
226 *) Expand ENGINE to support engine supplied SSL client certificate functions.
228 This work was sponsored by Logica.
231 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
232 keystores. Support for SSL/TLS client authentication too.
233 Not compiled unless enable-capieng specified to Configure.
235 This work was sponsored by Logica.
238 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
239 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
240 attribute creation routines such as certifcate requests and PKCS#12
244 Changes between 0.9.8g and 0.9.8h [28 May 2008]
246 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
247 handshake which could lead to a cilent crash as found using the
248 Codenomicon TLS test suite (CVE-2008-1672)
249 [Steve Henson, Mark Cox]
251 *) Fix double free in TLS server name extensions which could lead to
252 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
255 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
257 Clear the error queue to ensure that error entries left from
258 older function calls do not interfere with the correct operation.
259 [Lutz Jaenicke, Erik de Castro Lopo]
261 *) Remove root CA certificates of commercial CAs:
263 The OpenSSL project does not recommend any specific CA and does not
264 have any policy with respect to including or excluding any CA.
265 Therefore it does not make any sense to ship an arbitrary selection
266 of root CA certificates with the OpenSSL software.
269 *) RSA OAEP patches to fix two separate invalid memory reads.
270 The first one involves inputs when 'lzero' is greater than
271 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
272 before the beginning of from). The second one involves inputs where
273 the 'db' section contains nothing but zeroes (there is a one-byte
274 invalid read after the end of 'db').
275 [Ivan Nestlerode <inestlerode@us.ibm.com>]
277 *) Partial backport from 0.9.9-dev:
279 Introduce bn_mul_mont (dedicated Montgomery multiplication
280 procedure) as a candidate for BIGNUM assembler implementation.
281 While 0.9.9-dev uses assembler for various architectures, only
282 x86_64 is available by default here in the 0.9.8 branch, and
283 32-bit x86 is available through a compile-time setting.
285 To try the 32-bit x86 assembler implementation, use Configure
286 option "enable-montasm" (which exists only for this backport).
288 As "enable-montasm" for 32-bit x86 disclaims code stability
289 anyway, in this constellation we activate additional code
290 backported from 0.9.9-dev for further performance improvements,
291 namely BN_from_montgomery_word. (To enable this otherwise,
292 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
294 [Andy Polyakov (backport partially by Bodo Moeller)]
296 *) Add TLS session ticket callback. This allows an application to set
297 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
298 values. This is useful for key rollover for example where several key
299 sets may exist with different names.
302 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
303 This was broken until now in 0.9.8 releases, such that the only way
304 a registered ENGINE could be used (assuming it initialises
305 successfully on the host) was to explicitly set it as the default
306 for the relevant algorithms. This is in contradiction with 0.9.7
307 behaviour and the documentation. With this fix, when an ENGINE is
308 registered into a given algorithm's table of implementations, the
309 'uptodate' flag is reset so that auto-discovery will be used next
310 time a new context for that algorithm attempts to select an
312 [Ian Lister (tweaked by Geoff Thorpe)]
314 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
315 implemention in the following ways:
317 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
320 Lack of BER streaming support means one pass streaming processing is
321 only supported if data is detached: setting the streaming flag is
322 ignored for embedded content.
324 CMS support is disabled by default and must be explicitly enabled
325 with the enable-cms configuration option.
328 *) Update the GMP engine glue to do direct copies between BIGNUM and
329 mpz_t when openssl and GMP use the same limb size. Otherwise the
330 existing "conversion via a text string export" trick is still used.
331 [Paul Sheer <paulsheer@gmail.com>]
333 *) Zlib compression BIO. This is a filter BIO which compressed and
334 uncompresses any data passed through it.
337 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
338 RFC3394 compatible AES key wrapping.
341 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
342 sets string data without copying. X509_ALGOR_set0() and
343 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
344 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
345 from an X509_ATTRIBUTE structure optionally checking it occurs only
346 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
350 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
351 to get the expected BN_FLG_CONSTTIME behavior.
352 [Bodo Moeller (Google)]
356 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
357 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
358 - added some more tests to do_tests.pl
359 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
360 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
361 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
362 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
363 - various changes to netware.pl to enable gcc-cross builds on Win32
365 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
366 - various changes to fix missing prototype warnings
367 - fixed x86nasm.pl to create correct asm files for NASM COFF output
368 - added AES, WHIRLPOOL and CPUID assembler code to build files
369 - added missing AES assembler make rules to mk1mf.pl
370 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
371 [Guenter Knauf <eflash@gmx.net>]
373 *) Implement certificate status request TLS extension defined in RFC3546.
374 A client can set the appropriate parameters and receive the encoded
375 OCSP response via a callback. A server can query the supplied parameters
376 and set the encoded OCSP response in the callback. Add simplified examples
377 to s_client and s_server.
380 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
383 + Binary incompatibility of ssl_ctx_st structure
384 + DTLS interoperation with non-compliant servers
385 + Don't call get_session_cb() without proposed session
386 + Fix ia64 assembler code
387 [Andy Polyakov, Steve Henson]
389 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
391 *) DTLS Handshake overhaul. There were longstanding issues with
392 OpenSSL DTLS implementation, which were making it impossible for
393 RFC 4347 compliant client to communicate with OpenSSL server.
394 Unfortunately just fixing these incompatibilities would "cut off"
395 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
396 server keeps tolerating non RFC compliant syntax. The opposite is
397 not true, 0.9.8f client can not communicate with earlier server.
398 This update even addresses CVE-2007-4995.
401 *) Changes to avoid need for function casts in OpenSSL: some compilers
402 (gcc 4.2 and later) reject their use.
403 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
406 *) Add RFC4507 support to OpenSSL. This includes the corrections in
407 RFC4507bis. The encrypted ticket format is an encrypted encoded
408 SSL_SESSION structure, that way new session features are automatically
411 If a client application caches session in an SSL_SESSION structure
412 support is transparent because tickets are now stored in the encoded
415 The SSL_CTX structure automatically generates keys for ticket
416 protection in servers so again support should be possible
417 with no application modification.
419 If a client or server wishes to disable RFC4507 support then the option
420 SSL_OP_NO_TICKET can be set.
422 Add a TLS extension debugging callback to allow the contents of any client
423 or server extensions to be examined.
425 This work was sponsored by Google.
428 *) Add initial support for TLS extensions, specifically for the server_name
429 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
430 have new members for a host name. The SSL data structure has an
431 additional member SSL_CTX *initial_ctx so that new sessions can be
432 stored in that context to allow for session resumption, even after the
433 SSL has been switched to a new SSL_CTX in reaction to a client's
434 server_name extension.
436 New functions (subject to change):
439 SSL_get_servername_type()
442 New CTRL codes and macros (subject to change):
444 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
445 - SSL_CTX_set_tlsext_servername_callback()
446 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
447 - SSL_CTX_set_tlsext_servername_arg()
448 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
450 openssl s_client has a new '-servername ...' option.
452 openssl s_server has new options '-servername_host ...', '-cert2 ...',
453 '-key2 ...', '-servername_fatal' (subject to change). This allows
454 testing the HostName extension for a specific single host name ('-cert'
455 and '-key' remain fallbacks for handshakes without HostName
456 negotiation). If the unrecogninzed_name alert has to be sent, this by
457 default is a warning; it becomes fatal with the '-servername_fatal'
460 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
462 *) Add AES and SSE2 assembly language support to VC++ build.
465 *) Mitigate attack on final subtraction in Montgomery reduction.
468 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
469 (which previously caused an internal error).
472 *) Squeeze another 10% out of IGE mode when in != out.
475 *) AES IGE mode speedup.
476 [Dean Gaudet (Google)]
478 *) Add the Korean symmetric 128-bit cipher SEED (see
479 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
480 add SEED ciphersuites from RFC 4162:
482 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
483 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
484 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
485 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
487 To minimize changes between patchlevels in the OpenSSL 0.9.8
488 series, SEED remains excluded from compilation unless OpenSSL
489 is configured with 'enable-seed'.
492 *) Mitigate branch prediction attacks, which can be practical if a
493 single processor is shared, allowing a spy process to extract
494 information. For detailed background information, see
495 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
496 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
497 and Necessary Software Countermeasures"). The core of the change
498 are new versions BN_div_no_branch() and
499 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
500 respectively, which are slower, but avoid the security-relevant
501 conditional branches. These are automatically called by BN_div()
502 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
503 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
504 remove a conditional branch.
506 BN_FLG_CONSTTIME is the new name for the previous
507 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
508 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
509 in the exponent causes BN_mod_exp_mont() to use the alternative
510 implementation in BN_mod_exp_mont_consttime().) The old name
511 remains as a deprecated alias.
513 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
514 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
515 constant-time implementations for more than just exponentiation.
516 Here too the old name is kept as a deprecated alias.
518 BN_BLINDING_new() will now use BN_dup() for the modulus so that
519 the BN_BLINDING structure gets an independent copy of the
520 modulus. This means that the previous "BIGNUM *m" argument to
521 BN_BLINDING_new() and to BN_BLINDING_create_param() now
522 essentially becomes "const BIGNUM *m", although we can't actually
523 change this in the header file before 0.9.9. It allows
524 RSA_setup_blinding() to use BN_with_flags() on the modulus to
525 enable BN_FLG_CONSTTIME.
527 [Matthew D Wood (Intel Corp)]
529 *) In the SSL/TLS server implementation, be strict about session ID
530 context matching (which matters if an application uses a single
531 external cache for different purposes). Previously,
532 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
533 set. This did ensure strict client verification, but meant that,
534 with applications using a single external cache for quite
535 different requirements, clients could circumvent ciphersuite
536 restrictions for a given session ID context by starting a session
537 in a different context.
540 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
541 a ciphersuite string such as "DEFAULT:RSA" cannot enable
542 authentication-only ciphersuites.
545 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
546 not complete and could lead to a possible single byte overflow
547 (CVE-2007-5135) [Ben Laurie]
549 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
551 *) Since AES128 and AES256 (and similarly Camellia128 and
552 Camellia256) share a single mask bit in the logic of
553 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
554 kludge to work properly if AES128 is available and AES256 isn't
555 (or if Camellia128 is available and Camellia256 isn't).
558 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
559 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
560 When a point or a seed is encoded in a BIT STRING, we need to
561 prevent the removal of trailing zero bits to get the proper DER
562 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
563 of a NamedBitList, for which trailing 0 bits need to be removed.)
566 *) Have SSL/TLS server implementation tolerate "mismatched" record
567 protocol version while receiving ClientHello even if the
568 ClientHello is fragmented. (The server can't insist on the
569 particular protocol version it has chosen before the ServerHello
570 message has informed the client about his choice.)
573 *) Add RFC 3779 support.
574 [Rob Austein for ARIN, Ben Laurie]
576 *) Load error codes if they are not already present instead of using a
577 static variable. This allows them to be cleanly unloaded and reloaded.
578 Improve header file function name parsing.
581 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
582 or CAPABILITY handshake as required by RFCs.
585 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
587 *) Introduce limits to prevent malicious keys being able to
588 cause a denial of service. (CVE-2006-2940)
589 [Steve Henson, Bodo Moeller]
591 *) Fix ASN.1 parsing of certain invalid structures that can result
592 in a denial of service. (CVE-2006-2937) [Steve Henson]
594 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
595 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
597 *) Fix SSL client code which could crash if connecting to a
598 malicious SSLv2 server. (CVE-2006-4343)
599 [Tavis Ormandy and Will Drewry, Google Security Team]
601 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
602 match only those. Before that, "AES256-SHA" would be interpreted
603 as a pattern and match "AES128-SHA" too (since AES128-SHA got
604 the same strength classification in 0.9.7h) as we currently only
605 have a single AES bit in the ciphersuite description bitmap.
606 That change, however, also applied to ciphersuite strings such as
607 "RC4-MD5" that intentionally matched multiple ciphersuites --
608 namely, SSL 2.0 ciphersuites in addition to the more common ones
609 from SSL 3.0/TLS 1.0.
611 So we change the selection algorithm again: Naming an explicit
612 ciphersuite selects this one ciphersuite, and any other similar
613 ciphersuite (same bitmap) from *other* protocol versions.
614 Thus, "RC4-MD5" again will properly select both the SSL 2.0
615 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
617 Since SSL 2.0 does not have any ciphersuites for which the
618 128/256 bit distinction would be relevant, this works for now.
619 The proper fix will be to use different bits for AES128 and
620 AES256, which would have avoided the problems from the beginning;
621 however, bits are scarce, so we can only do this in a new release
622 (not just a patchlevel) when we can change the SSL_CIPHER
623 definition to split the single 'unsigned long mask' bitmap into
624 multiple values to extend the available space.
628 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
630 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
631 (CVE-2006-4339) [Ben Laurie and Google Security Team]
633 *) Add AES IGE and biIGE modes.
636 *) Change the Unix randomness entropy gathering to use poll() when
637 possible instead of select(), since the latter has some
638 undesirable limitations.
639 [Darryl Miles via Richard Levitte and Bodo Moeller]
641 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
642 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
643 cannot be implicitly activated as part of, e.g., the "AES" alias.
644 However, please upgrade to OpenSSL 0.9.9[-dev] for
645 non-experimental use of the ECC ciphersuites to get TLS extension
646 support, which is required for curve and point format negotiation
647 to avoid potential handshake problems.
650 *) Disable rogue ciphersuites:
652 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
653 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
654 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
656 The latter two were purportedly from
657 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
660 Also deactivate the remaining ciphersuites from
661 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
662 unofficial, and the ID has long expired.
665 *) Fix RSA blinding Heisenbug (problems sometimes occured on
666 dual-core machines) and other potential thread-safety issues.
669 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
670 versions), which is now available for royalty-free use
671 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
672 Also, add Camellia TLS ciphersuites from RFC 4132.
674 To minimize changes between patchlevels in the OpenSSL 0.9.8
675 series, Camellia remains excluded from compilation unless OpenSSL
676 is configured with 'enable-camellia'.
679 *) Disable the padding bug check when compression is in use. The padding
680 bug check assumes the first packet is of even length, this is not
681 necessarily true if compresssion is enabled and can result in false
682 positives causing handshake failure. The actual bug test is ancient
683 code so it is hoped that implementations will either have fixed it by
684 now or any which still have the bug do not support compression.
687 Changes between 0.9.8a and 0.9.8b [04 May 2006]
689 *) When applying a cipher rule check to see if string match is an explicit
690 cipher suite and only match that one cipher suite if it is.
693 *) Link in manifests for VC++ if needed.
694 [Austin Ziegler <halostatue@gmail.com>]
696 *) Update support for ECC-based TLS ciphersuites according to
697 draft-ietf-tls-ecc-12.txt with proposed changes (but without
698 TLS extensions, which are supported starting with the 0.9.9
699 branch, not in the OpenSSL 0.9.8 branch).
702 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
703 opaque EVP_CIPHER_CTX handling.
706 *) Fixes and enhancements to zlib compression code. We now only use
707 "zlib1.dll" and use the default __cdecl calling convention on Win32
708 to conform with the standards mentioned here:
709 http://www.zlib.net/DLL_FAQ.txt
710 Static zlib linking now works on Windows and the new --with-zlib-include
711 --with-zlib-lib options to Configure can be used to supply the location
712 of the headers and library. Gracefully handle case where zlib library
716 *) Several fixes and enhancements to the OID generation code. The old code
717 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
718 handle numbers larger than ULONG_MAX, truncated printing and had a
719 non standard OBJ_obj2txt() behaviour.
722 *) Add support for building of engines under engine/ as shared libraries
723 under VC++ build system.
726 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
727 Hopefully, we will not see any false combination of paths any more.
730 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
732 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
733 (part of SSL_OP_ALL). This option used to disable the
734 countermeasure against man-in-the-middle protocol-version
735 rollback in the SSL 2.0 server implementation, which is a bad
736 idea. (CVE-2005-2969)
738 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
739 for Information Security, National Institute of Advanced Industrial
740 Science and Technology [AIST], Japan)]
742 *) Add two function to clear and return the verify parameter flags.
745 *) Keep cipherlists sorted in the source instead of sorting them at
746 runtime, thus removing the need for a lock.
749 *) Avoid some small subgroup attacks in Diffie-Hellman.
750 [Nick Mathewson and Ben Laurie]
752 *) Add functions for well-known primes.
755 *) Extended Windows CE support.
756 [Satoshi Nakamura and Andy Polyakov]
758 *) Initialize SSL_METHOD structures at compile time instead of during
759 runtime, thus removing the need for a lock.
762 *) Make PKCS7_decrypt() work even if no certificate is supplied by
763 attempting to decrypt each encrypted key in turn. Add support to
767 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
769 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
772 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
775 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
776 key into the same file any more.
779 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
782 *) Add -utf8 command line and config file option to 'ca'.
783 [Stefan <stf@udoma.org]
785 *) Removed the macro des_crypt(), as it seems to conflict with some
786 libraries. Use DES_crypt().
789 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
790 involves renaming the source and generated shared-libs for
791 both. The engines will accept the corrected or legacy ids
792 ('ncipher' and '4758_cca' respectively) when binding. NB,
793 this only applies when building 'shared'.
794 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
796 *) Add attribute functions to EVP_PKEY structure. Modify
797 PKCS12_create() to recognize a CSP name attribute and
798 use it. Make -CSP option work again in pkcs12 utility.
801 *) Add new functionality to the bn blinding code:
802 - automatic re-creation of the BN_BLINDING parameters after
803 a fixed number of uses (currently 32)
804 - add new function for parameter creation
805 - introduce flags to control the update behaviour of the
806 BN_BLINDING parameters
807 - hide BN_BLINDING structure
808 Add a second BN_BLINDING slot to the RSA structure to improve
809 performance when a single RSA object is shared among several
813 *) Add support for DTLS.
814 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
816 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
817 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
820 *) Remove buggy and incompletet DH cert support from
821 ssl/ssl_rsa.c and ssl/s3_both.c
824 *) Use SHA-1 instead of MD5 as the default digest algorithm for
825 the apps/openssl applications.
828 *) Compile clean with "-Wall -Wmissing-prototypes
829 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
830 DEBUG_SAFESTACK must also be set.
833 *) Change ./Configure so that certain algorithms can be disabled by default.
834 The new counterpiece to "no-xxx" is "enable-xxx".
836 The patented RC5 and MDC2 algorithms will now be disabled unless
837 "enable-rc5" and "enable-mdc2", respectively, are specified.
839 (IDEA remains enabled despite being patented. This is because IDEA
840 is frequently required for interoperability, and there is no license
841 fee for non-commercial use. As before, "no-idea" can be used to
842 avoid this algorithm.)
846 *) Add processing of proxy certificates (see RFC 3820). This work was
847 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
848 EGEE (Enabling Grids for E-science in Europe).
851 *) RC4 performance overhaul on modern architectures/implementations, such
852 as Intel P4, IA-64 and AMD64.
855 *) New utility extract-section.pl. This can be used specify an alternative
856 section number in a pod file instead of having to treat each file as
857 a separate case in Makefile. This can be done by adding two lines to the
860 =for comment openssl_section:XXX
862 The blank line is mandatory.
866 *) New arguments -certform, -keyform and -pass for s_client and s_server
867 to allow alternative format key and certificate files and passphrase
871 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
872 update associated structures and add various utility functions.
874 Add new policy related verify parameters, include policy checking in
875 standard verify code. Enhance 'smime' application with extra parameters
876 to support policy checking and print out.
879 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
880 Nehemiah processors. These extensions support AES encryption in hardware
881 as well as RNG (though RNG support is currently disabled).
882 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
884 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
887 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
888 [Andy Polyakov and a number of other people]
890 *) Improved PowerPC platform support. Most notably BIGNUM assembler
891 implementation contributed by IBM.
892 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
894 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
895 exponent rather than 'unsigned long'. There is a corresponding change to
896 the new 'rsa_keygen' element of the RSA_METHOD structure.
897 [Jelte Jansen, Geoff Thorpe]
899 *) Functionality for creating the initial serial number file is now
900 moved from CA.pl to the 'ca' utility with a new option -create_serial.
902 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
903 number file to 1, which is bound to cause problems. To avoid
904 the problems while respecting compatibility between different 0.9.7
905 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
906 CA.pl for serial number initialization. With the new release 0.9.8,
907 we can fix the problem directly in the 'ca' utility.)
910 *) Reduced header interdepencies by declaring more opaque objects in
911 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
912 give fewer recursive includes, which could break lazy source code - so
913 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
914 developers should define this symbol when building and using openssl to
915 ensure they track the recommended behaviour, interfaces, [etc], but
916 backwards-compatible behaviour prevails when this isn't defined.
919 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
922 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
923 This will generate a random key of the appropriate length based on the
924 cipher context. The EVP_CIPHER can provide its own random key generation
925 routine to support keys of a specific form. This is used in the des and
926 3des routines to generate a key of the correct parity. Update S/MIME
927 code to use new functions and hence generate correct parity DES keys.
928 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
929 valid (weak or incorrect parity).
932 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
933 as looking them up. This is useful when the verified structure may contain
934 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
935 present unless the new PKCS7_NO_CRL flag is asserted.
938 *) Extend ASN1 oid configuration module. It now additionally accepts the
941 shortName = some long name, 1.2.3.4
944 *) Reimplemented the BN_CTX implementation. There is now no more static
945 limitation on the number of variables it can handle nor the depth of the
946 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
947 information can now expand as required, and rather than having a single
948 static array of bignums, BN_CTX now uses a linked-list of such arrays
949 allowing it to expand on demand whilst maintaining the usefulness of
953 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
954 to allow all RSA operations to function using a single BN_CTX.
957 *) Preliminary support for certificate policy evaluation and checking. This
958 is initially intended to pass the tests outlined in "Conformance Testing
959 of Relying Party Client Certificate Path Processing Logic" v1.07.
962 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
963 remained unused and not that useful. A variety of other little bignum
964 tweaks and fixes have also been made continuing on from the audit (see
968 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
969 associated ASN1, EVP and SSL functions and old ASN1 macros.
972 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
973 and this should never fail. So the return value from the use of
974 BN_set_word() (which can fail due to needless expansion) is now deprecated;
975 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
978 *) BN_CTX_get() should return zero-valued bignums, providing the same
979 initialised value as BN_new().
980 [Geoff Thorpe, suggested by Ulf Möller]
982 *) Support for inhibitAnyPolicy certificate extension.
985 *) An audit of the BIGNUM code is underway, for which debugging code is
986 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
987 is considered valid when processing BIGNUMs, and causes execution to
988 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
989 further steps are taken to deliberately pollute unused data in BIGNUM
990 structures to try and expose faulty code further on. For now, openssl will
991 (in its default mode of operation) continue to tolerate the inconsistent
992 forms that it has tolerated in the past, but authors and packagers should
993 consider trying openssl and their own applications when compiled with
994 these debugging symbols defined. It will help highlight potential bugs in
995 their own code, and will improve the test coverage for OpenSSL itself. At
996 some point, these tighter rules will become openssl's default to improve
997 maintainability, though the assert()s and other overheads will remain only
998 in debugging configurations. See bn.h for more details.
999 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1001 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1002 that can only be obtained through BN_CTX_new() (which implicitly
1003 initialises it). The presence of this function only made it possible
1004 to overwrite an existing structure (and cause memory leaks).
1007 *) Because of the callback-based approach for implementing LHASH as a
1008 template type, lh_insert() adds opaque objects to hash-tables and
1009 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1010 to clean up those corresponding objects before destroying the hash table
1011 (and losing the object pointers). So some over-zealous constifications in
1012 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1013 objects as "const" and the lh_doall[_arg] callback wrappers are not
1014 prototyped to have "const" restrictions on the object pointers they are
1015 given (and so aren't required to cast them away any more).
1018 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1019 (speed) prefers to use its own implementation. The two implementations
1020 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1021 its object type properly exposed (MS_TM) instead of casting to/from "char
1022 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1023 aren't necessarily the greatest nomenclatures - but this is what was used
1024 internally to the implementation so I've used that for now.
1027 *) Ensure that deprecated functions do not get compiled when
1028 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1029 the self-tests were still using deprecated key-generation functions so
1030 these have been updated also.
1033 *) Reorganise PKCS#7 code to separate the digest location functionality
1034 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1035 New function PKCS7_set_digest() to set the digest type for PKCS#7
1036 digestedData type. Add additional code to correctly generate the
1037 digestedData type and add support for this type in PKCS7 initialization
1041 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1042 structure of type "other".
1045 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1046 sure the loop does correctly stop and breaking ("division by zero")
1047 modulus operations are not performed. The (pre-generated) prime
1048 table crypto/bn/bn_prime.h was already correct, but it could not be
1049 re-generated on some platforms because of the "division by zero"
1050 situation in the script.
1051 [Ralf S. Engelschall]
1053 *) Update support for ECC-based TLS ciphersuites according to
1054 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1055 SHA-1 now is only used for "small" curves (where the
1056 representation of a field element takes up to 24 bytes); for
1057 larger curves, the field element resulting from ECDH is directly
1058 used as premaster secret.
1059 [Douglas Stebila (Sun Microsystems Laboratories)]
1061 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1062 curve secp160r1 to the tests.
1063 [Douglas Stebila (Sun Microsystems Laboratories)]
1065 *) Add the possibility to load symbols globally with DSO.
1066 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1068 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1069 control of the error stack.
1072 *) Add support for STORE in ENGINE.
1075 *) Add the STORE type. The intention is to provide a common interface
1076 to certificate and key stores, be they simple file-based stores, or
1077 HSM-type store, or LDAP stores, or...
1078 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1081 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1082 pass a list of arguments to any function as well as provide a way
1083 for a function to pass data back to the caller.
1086 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1087 works like BUF_strdup() but can be used to duplicate a portion of
1088 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1092 *) Add the function sk_find_ex() which works like sk_find(), but will
1093 return an index to an element even if an exact match couldn't be
1094 found. The index is guaranteed to point at the element where the
1095 searched-for key would be inserted to preserve sorting order.
1098 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1099 takes an extra flags argument for optional functionality. Currently,
1100 the following flags are defined:
1102 OBJ_BSEARCH_VALUE_ON_NOMATCH
1103 This one gets OBJ_bsearch_ex() to return a pointer to the first
1104 element where the comparing function returns a negative or zero
1107 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1108 This one gets OBJ_bsearch_ex() to return a pointer to the first
1109 element where the comparing function returns zero. This is useful
1110 if there are more than one element where the comparing function
1114 *) Make it possible to create self-signed certificates with 'openssl ca'
1115 in such a way that the self-signed certificate becomes part of the
1116 CA database and uses the same mechanisms for serial number generation
1117 as all other certificate signing. The new flag '-selfsign' enables
1118 this functionality. Adapt CA.sh and CA.pl.in.
1121 *) Add functionality to check the public key of a certificate request
1122 against a given private. This is useful to check that a certificate
1123 request can be signed by that key (self-signing).
1126 *) Make it possible to have multiple active certificates with the same
1127 subject in the CA index file. This is done only if the keyword
1128 'unique_subject' is set to 'no' in the main CA section (default
1129 if 'CA_default') of the configuration file. The value is saved
1130 with the database itself in a separate index attribute file,
1131 named like the index file with '.attr' appended to the name.
1134 *) Generate muti valued AVAs using '+' notation in config files for
1138 *) Support for nameConstraints certificate extension.
1141 *) Support for policyConstraints certificate extension.
1144 *) Support for policyMappings certificate extension.
1147 *) Make sure the default DSA_METHOD implementation only uses its
1148 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1149 and change its own handlers to be NULL so as to remove unnecessary
1150 indirection. This lets alternative implementations fallback to the
1151 default implementation more easily.
1154 *) Support for directoryName in GeneralName related extensions
1158 *) Make it possible to link applications using Makefile.shared.
1159 Make that possible even when linking against static libraries!
1162 *) Support for single pass processing for S/MIME signing. This now
1163 means that S/MIME signing can be done from a pipe, in addition
1164 cleartext signing (multipart/signed type) is effectively streaming
1165 and the signed data does not need to be all held in memory.
1167 This is done with a new flag PKCS7_STREAM. When this flag is set
1168 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1169 is done after the data is output (and digests calculated) in
1170 SMIME_write_PKCS7().
1173 *) Add full support for -rpath/-R, both in shared libraries and
1174 applications, at least on the platforms where it's known how
1178 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1179 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1180 will now compute a table of multiples of the generator that
1181 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1182 faster (notably in the case of a single point multiplication,
1183 scalar * generator).
1184 [Nils Larsch, Bodo Moeller]
1186 *) IPv6 support for certificate extensions. The various extensions
1187 which use the IP:a.b.c.d can now take IPv6 addresses using the
1188 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1192 *) Added an ENGINE that implements RSA by performing private key
1193 exponentiations with the GMP library. The conversions to and from
1194 GMP's mpz_t format aren't optimised nor are any montgomery forms
1195 cached, and on x86 it appears OpenSSL's own performance has caught up.
1196 However there are likely to be other architectures where GMP could
1197 provide a boost. This ENGINE is not built in by default, but it can be
1198 specified at Configure time and should be accompanied by the necessary
1199 linker additions, eg;
1200 ./config -DOPENSSL_USE_GMP -lgmp
1203 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1204 testing availability of engines with "-t" - the old behaviour is
1205 produced by increasing the feature's verbosity with "-tt".
1208 *) ECDSA routines: under certain error conditions uninitialized BN objects
1209 could be freed. Solution: make sure initialization is performed early
1210 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1214 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1215 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1216 software implementations. For DSA and DH, parameter generation can
1217 also be overriden by providing the appropriate method callbacks.
1220 *) Change the "progress" mechanism used in key-generation and
1221 primality testing to functions that take a new BN_GENCB pointer in
1222 place of callback/argument pairs. The new API functions have "_ex"
1223 postfixes and the older functions are reimplemented as wrappers for
1224 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1225 declarations of the old functions to help (graceful) attempts to
1226 migrate to the new functions. Also, the new key-generation API
1227 functions operate on a caller-supplied key-structure and return
1228 success/failure rather than returning a key or NULL - this is to
1229 help make "keygen" another member function of RSA_METHOD etc.
1231 Example for using the new callback interface:
1233 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1237 BN_GENCB_set(&my_cb, my_callback, my_arg);
1239 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1240 /* For the meaning of a, b in calls to my_callback(), see the
1241 * documentation of the function that calls the callback.
1242 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1243 * my_callback should return 1 if it wants BN_is_prime_ex()
1244 * to continue, or 0 to stop.
1249 *) Change the ZLIB compression method to be stateful, and make it
1250 available to TLS with the number defined in
1251 draft-ietf-tls-compression-04.txt.
1254 *) Add the ASN.1 structures and functions for CertificatePair, which
1255 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1257 CertificatePair ::= SEQUENCE {
1258 forward [0] Certificate OPTIONAL,
1259 reverse [1] Certificate OPTIONAL,
1260 -- at least one of the pair shall be present -- }
1262 Also implement the PEM functions to read and write certificate
1263 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1265 This needed to be defined, mostly for the sake of the LDAP
1266 attribute crossCertificatePair, but may prove useful elsewhere as
1270 *) Make it possible to inhibit symlinking of shared libraries in
1271 Makefile.shared, for Cygwin's sake.
1274 *) Extend the BIGNUM API by creating a function
1275 void BN_set_negative(BIGNUM *a, int neg);
1276 and a macro that behave like
1277 int BN_is_negative(const BIGNUM *a);
1279 to avoid the need to access 'a->neg' directly in applications.
1282 *) Implement fast modular reduction for pseudo-Mersenne primes
1283 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1284 EC_GROUP_new_curve_GFp() will now automatically use this
1286 [Nils Larsch <nla@trustcenter.de>]
1288 *) Add new lock type (CRYPTO_LOCK_BN).
1291 *) Change the ENGINE framework to automatically load engines
1292 dynamically from specific directories unless they could be
1293 found to already be built in or loaded. Move all the
1294 current engines except for the cryptodev one to a new
1296 The engines in engines/ are built as shared libraries if
1297 the "shared" options was given to ./Configure or ./config.
1298 Otherwise, they are inserted in libcrypto.a.
1299 /usr/local/ssl/engines is the default directory for dynamic
1300 engines, but that can be overriden at configure time through
1301 the usual use of --prefix and/or --openssldir, and at run
1302 time with the environment variable OPENSSL_ENGINES.
1303 [Geoff Thorpe and Richard Levitte]
1305 *) Add Makefile.shared, a helper makefile to build shared
1306 libraries. Addapt Makefile.org.
1309 *) Add version info to Win32 DLLs.
1310 [Peter 'Luna' Runestig" <peter@runestig.com>]
1312 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1313 can be added using this API to created arbitrary PKCS#12
1314 files while avoiding the low level API.
1316 New options to PKCS12_create(), key or cert can be NULL and
1317 will then be omitted from the output file. The encryption
1318 algorithm NIDs can be set to -1 for no encryption, the mac
1319 iteration count can be set to 0 to omit the mac.
1321 Enhance pkcs12 utility by making the -nokeys and -nocerts
1322 options work when creating a PKCS#12 file. New option -nomac
1323 to omit the mac, NONE can be set for an encryption algorithm.
1324 New code is modified to use the enhanced PKCS12_create()
1325 instead of the low level API.
1328 *) Extend ASN1 encoder to support indefinite length constructed
1329 encoding. This can output sequences tags and octet strings in
1330 this form. Modify pk7_asn1.c to support indefinite length
1331 encoding. This is experimental and needs additional code to
1332 be useful, such as an ASN1 bio and some enhanced streaming
1335 Extend template encode functionality so that tagging is passed
1336 down to the template encoder.
1339 *) Let 'openssl req' fail if an argument to '-newkey' is not
1340 recognized instead of using RSA as a default.
1343 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1344 As these are not official, they are not included in "ALL";
1345 the "ECCdraft" ciphersuite group alias can be used to select them.
1346 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1348 *) Add ECDH engine support.
1349 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1351 *) Add ECDH in new directory crypto/ecdh/.
1352 [Douglas Stebila (Sun Microsystems Laboratories)]
1354 *) Let BN_rand_range() abort with an error after 100 iterations
1355 without success (which indicates a broken PRNG).
1358 *) Change BN_mod_sqrt() so that it verifies that the input value
1359 is really the square of the return value. (Previously,
1360 BN_mod_sqrt would show GIGO behaviour.)
1363 *) Add named elliptic curves over binary fields from X9.62, SECG,
1364 and WAP/WTLS; add OIDs that were still missing.
1366 [Sheueling Chang Shantz and Douglas Stebila
1367 (Sun Microsystems Laboratories)]
1369 *) Extend the EC library for elliptic curves over binary fields
1370 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1373 EC_GF2m_simple_method
1377 EC_GROUP_new_curve_GF2m
1378 EC_GROUP_set_curve_GF2m
1379 EC_GROUP_get_curve_GF2m
1380 EC_POINT_set_affine_coordinates_GF2m
1381 EC_POINT_get_affine_coordinates_GF2m
1382 EC_POINT_set_compressed_coordinates_GF2m
1384 Point compression for binary fields is disabled by default for
1385 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1388 As binary polynomials are represented as BIGNUMs, various members
1389 of the EC_GROUP and EC_POINT data structures can be shared
1390 between the implementations for prime fields and binary fields;
1391 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1392 are essentially identical to their ..._GFp counterparts.
1393 (For simplicity, the '..._GFp' prefix has been dropped from
1394 various internal method names.)
1396 An internal 'field_div' method (similar to 'field_mul' and
1397 'field_sqr') has been added; this is used only for binary fields.
1399 [Sheueling Chang Shantz and Douglas Stebila
1400 (Sun Microsystems Laboratories)]
1402 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1403 through methods ('mul', 'precompute_mult').
1405 The generic implementations (now internally called 'ec_wNAF_mul'
1406 and 'ec_wNAF_precomputed_mult') remain the default if these
1407 methods are undefined.
1409 [Sheueling Chang Shantz and Douglas Stebila
1410 (Sun Microsystems Laboratories)]
1412 *) New function EC_GROUP_get_degree, which is defined through
1413 EC_METHOD. For curves over prime fields, this returns the bit
1414 length of the modulus.
1416 [Sheueling Chang Shantz and Douglas Stebila
1417 (Sun Microsystems Laboratories)]
1419 *) New functions EC_GROUP_dup, EC_POINT_dup.
1420 (These simply call ..._new and ..._copy).
1422 [Sheueling Chang Shantz and Douglas Stebila
1423 (Sun Microsystems Laboratories)]
1425 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1426 Polynomials are represented as BIGNUMs (where the sign bit is not
1427 used) in the following functions [macros]:
1430 BN_GF2m_sub [= BN_GF2m_add]
1431 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1432 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1433 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1435 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1436 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1437 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1438 BN_GF2m_cmp [= BN_ucmp]
1440 (Note that only the 'mod' functions are actually for fields GF(2^m).
1441 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1443 For some functions, an the irreducible polynomial defining a
1444 field can be given as an 'unsigned int[]' with strictly
1445 decreasing elements giving the indices of those bits that are set;
1446 i.e., p[] represents the polynomial
1447 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1449 p[0] > p[1] > ... > p[k] = 0.
1450 This applies to the following functions:
1455 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1456 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1458 BN_GF2m_mod_sqrt_arr
1459 BN_GF2m_mod_solve_quad_arr
1463 Conversion can be performed by the following functions:
1468 bntest.c has additional tests for binary polynomial arithmetic.
1470 Two implementations for BN_GF2m_mod_div() are available.
1471 The default algorithm simply uses BN_GF2m_mod_inv() and
1472 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1473 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1474 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1476 [Sheueling Chang Shantz and Douglas Stebila
1477 (Sun Microsystems Laboratories)]
1479 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1480 functionality is disabled at compile-time.
1481 [Douglas Stebila <douglas.stebila@sun.com>]
1483 *) Change default behaviour of 'openssl asn1parse' so that more
1484 information is visible when viewing, e.g., a certificate:
1486 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1487 mode the content of non-printable OCTET STRINGs is output in a
1488 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1489 avoid the appearance of a printable string.
1490 [Nils Larsch <nla@trustcenter.de>]
1492 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1494 EC_GROUP_set_asn1_flag()
1495 EC_GROUP_get_asn1_flag()
1496 EC_GROUP_set_point_conversion_form()
1497 EC_GROUP_get_point_conversion_form()
1498 These control ASN1 encoding details:
1499 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1500 has been set to OPENSSL_EC_NAMED_CURVE.
1501 - Points are encoded in uncompressed form by default; options for
1502 asn1_for are as for point2oct, namely
1503 POINT_CONVERSION_COMPRESSED
1504 POINT_CONVERSION_UNCOMPRESSED
1505 POINT_CONVERSION_HYBRID
1507 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1510 EC_GROUP_get0_seed()
1511 EC_GROUP_get_seed_len()
1512 This is used only for ASN1 purposes (so far).
1513 [Nils Larsch <nla@trustcenter.de>]
1515 *) Add 'field_type' member to EC_METHOD, which holds the NID
1516 of the appropriate field type OID. The new function
1517 EC_METHOD_get_field_type() returns this value.
1518 [Nils Larsch <nla@trustcenter.de>]
1523 EC_POINT_point2hex()
1524 EC_POINT_hex2point()
1525 providing useful interfaces to EC_POINT_point2oct() and
1526 EC_POINT_oct2point().
1527 [Nils Larsch <nla@trustcenter.de>]
1529 *) Change internals of the EC library so that the functions
1530 EC_GROUP_set_generator()
1531 EC_GROUP_get_generator()
1532 EC_GROUP_get_order()
1533 EC_GROUP_get_cofactor()
1534 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1535 to methods, which would lead to unnecessary code duplication when
1536 adding different types of curves.
1537 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1539 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1540 arithmetic, and such that modified wNAFs are generated
1541 (which avoid length expansion in many cases).
1544 *) Add a function EC_GROUP_check_discriminant() (defined via
1545 EC_METHOD) that verifies that the curve discriminant is non-zero.
1547 Add a function EC_GROUP_check() that makes some sanity tests
1548 on a EC_GROUP, its generator and order. This includes
1549 EC_GROUP_check_discriminant().
1550 [Nils Larsch <nla@trustcenter.de>]
1552 *) Add ECDSA in new directory crypto/ecdsa/.
1554 Add applications 'openssl ecparam' and 'openssl ecdsa'
1555 (these are based on 'openssl dsaparam' and 'openssl dsa').
1557 ECDSA support is also included in various other files across the
1558 library. Most notably,
1559 - 'openssl req' now has a '-newkey ecdsa:file' option;
1560 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1561 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1562 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1563 them suitable for ECDSA where domain parameters must be
1564 extracted before the specific public key;
1565 - ECDSA engine support has been added.
1566 [Nils Larsch <nla@trustcenter.de>]
1568 *) Include some named elliptic curves, and add OIDs from X9.62,
1569 SECG, and WAP/WTLS. Each curve can be obtained from the new
1571 EC_GROUP_new_by_curve_name(),
1572 and the list of available named curves can be obtained with
1573 EC_get_builtin_curves().
1574 Also add a 'curve_name' member to EC_GROUP objects, which can be
1576 EC_GROUP_set_curve_name()
1577 EC_GROUP_get_curve_name()
1578 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1580 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1581 was actually never needed) and in BN_mul(). The removal in BN_mul()
1582 required a small change in bn_mul_part_recursive() and the addition
1583 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1584 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1585 bn_sub_words() and bn_add_words() except they take arrays with
1589 Changes between 0.9.7m and 0.9.7n [xx XXX xxxx]
1591 *) In the SSL/TLS server implementation, be strict about session ID
1592 context matching (which matters if an application uses a single
1593 external cache for different purposes). Previously,
1594 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1595 set. This did ensure strict client verification, but meant that,
1596 with applications using a single external cache for quite
1597 different requirements, clients could circumvent ciphersuite
1598 restrictions for a given session ID context by starting a session
1599 in a different context.
1602 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1604 *) Cleanse PEM buffers before freeing them since they may contain
1606 [Benjamin Bennett <ben@psc.edu>]
1608 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1609 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1610 authentication-only ciphersuites.
1613 *) Since AES128 and AES256 share a single mask bit in the logic of
1614 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1615 kludge to work properly if AES128 is available and AES256 isn't.
1618 *) Expand security boundary to match 1.1.1 module.
1621 *) Remove redundant features: hash file source, editing of test vectors
1622 modify fipsld to use external fips_premain.c signature.
1625 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1626 run algorithm test programs.
1629 *) Make algorithm test programs more tolerant of whitespace.
1632 *) Have SSL/TLS server implementation tolerate "mismatched" record
1633 protocol version while receiving ClientHello even if the
1634 ClientHello is fragmented. (The server can't insist on the
1635 particular protocol version it has chosen before the ServerHello
1636 message has informed the client about his choice.)
1639 *) Load error codes if they are not already present instead of using a
1640 static variable. This allows them to be cleanly unloaded and reloaded.
1643 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1645 *) Introduce limits to prevent malicious keys being able to
1646 cause a denial of service. (CVE-2006-2940)
1647 [Steve Henson, Bodo Moeller]
1649 *) Fix ASN.1 parsing of certain invalid structures that can result
1650 in a denial of service. (CVE-2006-2937) [Steve Henson]
1652 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1653 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1655 *) Fix SSL client code which could crash if connecting to a
1656 malicious SSLv2 server. (CVE-2006-4343)
1657 [Tavis Ormandy and Will Drewry, Google Security Team]
1659 *) Change ciphersuite string processing so that an explicit
1660 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1661 will no longer include "AES128-SHA"), and any other similar
1662 ciphersuite (same bitmap) from *other* protocol versions (so that
1663 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1664 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1665 changes from 0.9.8b and 0.9.8d.
1668 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1670 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1671 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1673 *) Change the Unix randomness entropy gathering to use poll() when
1674 possible instead of select(), since the latter has some
1675 undesirable limitations.
1676 [Darryl Miles via Richard Levitte and Bodo Moeller]
1678 *) Disable rogue ciphersuites:
1680 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1681 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1682 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1684 The latter two were purportedly from
1685 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1688 Also deactive the remaining ciphersuites from
1689 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1690 unofficial, and the ID has long expired.
1693 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1694 dual-core machines) and other potential thread-safety issues.
1697 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1699 *) Adapt fipsld and the build system to link against the validated FIPS
1700 module in FIPS mode.
1703 *) Fixes for VC++ 2005 build under Windows.
1706 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1707 from a Windows bash shell such as MSYS. It is autodetected from the
1708 "config" script when run from a VC++ environment. Modify standard VC++
1709 build to use fipscanister.o from the GNU make build.
1712 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1714 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1715 The value now differs depending on if you build for FIPS or not.
1716 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1717 safely run with a non-FIPSed libcrypto, as it may crash because of
1718 the difference induced by this change.
1721 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1723 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1724 (part of SSL_OP_ALL). This option used to disable the
1725 countermeasure against man-in-the-middle protocol-version
1726 rollback in the SSL 2.0 server implementation, which is a bad
1727 idea. (CVE-2005-2969)
1729 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1730 for Information Security, National Institute of Advanced Industrial
1731 Science and Technology [AIST], Japan)]
1733 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1734 mainly for FIPS compliance and not fully integrated at this stage.
1737 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1738 the exponentiation using a fixed-length exponent. (Otherwise,
1739 the information leaked through timing could expose the secret key
1740 after many signatures; cf. Bleichenbacher's attack on DSA with
1744 *) Make a new fixed-window mod_exp implementation the default for
1745 RSA, DSA, and DH private-key operations so that the sequence of
1746 squares and multiplies and the memory access pattern are
1747 independent of the particular secret key. This will mitigate
1748 cache-timing and potential related attacks.
1750 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1751 and this is automatically used by BN_mod_exp_mont() if the new flag
1752 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1753 will use this BN flag for private exponents unless the flag
1754 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1755 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1757 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1759 *) Change the client implementation for SSLv23_method() and
1760 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1761 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1762 (Previously, the SSL 2.0 backwards compatible Client Hello
1763 message format would be used even with SSL_OP_NO_SSLv2.)
1766 *) Add support for smime-type MIME parameter in S/MIME messages which some
1770 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1771 a threadsafe manner. Modify rsa code to use new function and add calls
1772 to dsa and dh code (which had race conditions before).
1775 *) Include the fixed error library code in the C error file definitions
1776 instead of fixing them up at runtime. This keeps the error code
1777 structures constant.
1780 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1782 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1785 *) Fixes for newer kerberos headers. NB: the casts are needed because
1786 the 'length' field is signed on one version and unsigned on another
1787 with no (?) obvious way to tell the difference, without these VC++
1788 complains. Also the "definition" of FAR (blank) is no longer included
1789 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1790 some needed definitions.
1793 *) Undo Cygwin change.
1796 *) Added support for proxy certificates according to RFC 3820.
1797 Because they may be a security thread to unaware applications,
1798 they must be explicitely allowed in run-time. See
1799 docs/HOWTO/proxy_certificates.txt for further information.
1802 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1804 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1805 server and client random values. Previously
1806 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1807 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1809 This change has negligible security impact because:
1811 1. Server and client random values still have 24 bytes of pseudo random
1814 2. Server and client random values are sent in the clear in the initial
1817 3. The master secret is derived using the premaster secret (48 bytes in
1818 size for static RSA ciphersuites) as well as client server and random
1821 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1824 [Stephen Henson, reported by UK NISCC]
1826 *) Use Windows randomness collection on Cygwin.
1829 *) Fix hang in EGD/PRNGD query when communication socket is closed
1830 prematurely by EGD/PRNGD.
1831 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1833 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1836 *) Back-port of selected performance improvements from development
1837 branch, as well as improved support for PowerPC platforms.
1840 *) Add lots of checks for memory allocation failure, error codes to indicate
1841 failure and freeing up memory if a failure occurs.
1842 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1844 *) Add new -passin argument to dgst.
1847 *) Perform some character comparisons of different types in X509_NAME_cmp:
1848 this is needed for some certificates that reencode DNs into UTF8Strings
1849 (in violation of RFC3280) and can't or wont issue name rollover
1853 *) Make an explicit check during certificate validation to see that
1854 the CA setting in each certificate on the chain is correct. As a
1855 side effect always do the following basic checks on extensions,
1856 not just when there's an associated purpose to the check:
1858 - if there is an unhandled critical extension (unless the user
1859 has chosen to ignore this fault)
1860 - if the path length has been exceeded (if one is set at all)
1861 - that certain extensions fit the associated purpose (if one has
1865 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1867 *) Avoid a race condition when CRLs are checked in a multi threaded
1868 environment. This would happen due to the reordering of the revoked
1869 entries during signature checking and serial number lookup. Now the
1870 encoding is cached and the serial number sort performed under a lock.
1871 Add new STACK function sk_is_sorted().
1874 *) Add Delta CRL to the extension code.
1877 *) Various fixes to s3_pkt.c so alerts are sent properly.
1878 [David Holmes <d.holmes@f5.com>]
1880 *) Reduce the chances of duplicate issuer name and serial numbers (in
1881 violation of RFC3280) using the OpenSSL certificate creation utilities.
1882 This is done by creating a random 64 bit value for the initial serial
1883 number when a serial number file is created or when a self signed
1884 certificate is created using 'openssl req -x509'. The initial serial
1885 number file is created using 'openssl x509 -next_serial' in CA.pl
1886 rather than being initialized to 1.
1889 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1891 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1892 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1893 [Joe Orton, Steve Henson]
1895 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1897 [Joe Orton, Steve Henson]
1899 *) Make it possible to have multiple active certificates with the same
1900 subject in the CA index file. This is done only if the keyword
1901 'unique_subject' is set to 'no' in the main CA section (default
1902 if 'CA_default') of the configuration file. The value is saved
1903 with the database itself in a separate index attribute file,
1904 named like the index file with '.attr' appended to the name.
1907 *) X509 verify fixes. Disable broken certificate workarounds when
1908 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1909 keyUsage extension present. Don't accept CRLs with unhandled critical
1910 extensions: since verify currently doesn't process CRL extensions this
1911 rejects a CRL with *any* critical extensions. Add new verify error codes
1915 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1916 A clarification of RFC2560 will require the use of OCTET STRINGs and
1917 some implementations cannot handle the current raw format. Since OpenSSL
1918 copies and compares OCSP nonces as opaque blobs without any attempt at
1919 parsing them this should not create any compatibility issues.
1922 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1923 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1924 this HMAC (and other) operations are several times slower than OpenSSL
1928 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1929 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1931 *) Use the correct content when signing type "other".
1934 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1936 *) Fix various bugs revealed by running the NISCC test suite:
1938 Stop out of bounds reads in the ASN1 code when presented with
1939 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1941 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1943 If verify callback ignores invalid public key errors don't try to check
1944 certificate signature with the NULL public key.
1948 *) New -ignore_err option in ocsp application to stop the server
1949 exiting on the first error in a request.
1952 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1953 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1957 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1958 extra data after the compression methods not only for TLS 1.0
1959 but also for SSL 3.0 (as required by the specification).
1960 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1962 *) Change X509_certificate_type() to mark the key as exported/exportable
1963 when it's 512 *bits* long, not 512 bytes.
1966 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1967 blocks during encryption.
1970 *) Various fixes to base64 BIO and non blocking I/O. On write
1971 flushes were not handled properly if the BIO retried. On read
1972 data was not being buffered properly and had various logic bugs.
1973 This also affects blocking I/O when the data being decoded is a
1977 *) Various S/MIME bugfixes and compatibility changes:
1978 output correct application/pkcs7 MIME type if
1979 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1980 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1981 of files as .eml work). Correctly handle very long lines in MIME
1985 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1987 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1988 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1989 a protocol version number mismatch like a decryption error
1990 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1993 *) Turn on RSA blinding by default in the default implementation
1994 to avoid a timing attack. Applications that don't want it can call
1995 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1996 They would be ill-advised to do so in most cases.
1997 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1999 *) Change RSA blinding code so that it works when the PRNG is not
2000 seeded (in this case, the secret RSA exponent is abused as
2001 an unpredictable seed -- if it is not unpredictable, there
2002 is no point in blinding anyway). Make RSA blinding thread-safe
2003 by remembering the creator's thread ID in rsa->blinding and
2004 having all other threads use local one-time blinding factors
2005 (this requires more computation than sharing rsa->blinding, but
2006 avoids excessive locking; and if an RSA object is not shared
2007 between threads, blinding will still be very fast).
2010 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2011 ENGINE as defaults for all supported algorithms irrespective of
2012 the 'flags' parameter. 'flags' is now honoured, so applications
2013 should make sure they are passing it correctly.
2016 *) Target "mingw" now allows native Windows code to be generated in
2017 the Cygwin environment as well as with the MinGW compiler.
2020 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2022 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2023 via timing by performing a MAC computation even if incorrrect
2024 block cipher padding has been found. This is a countermeasure
2025 against active attacks where the attacker has to distinguish
2026 between bad padding and a MAC verification error. (CVE-2003-0078)
2028 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2029 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2030 Martin Vuagnoux (EPFL, Ilion)]
2032 *) Make the no-err option work as intended. The intention with no-err
2033 is not to have the whole error stack handling routines removed from
2034 libcrypto, it's only intended to remove all the function name and
2035 reason texts, thereby removing some of the footprint that may not
2036 be interesting if those errors aren't displayed anyway.
2038 NOTE: it's still possible for any application or module to have it's
2039 own set of error texts inserted. The routines are there, just not
2040 used by default when no-err is given.
2043 *) Add support for FreeBSD on IA64.
2044 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2046 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2047 Kerberos function mit_des_cbc_cksum(). Before this change,
2048 the value returned by DES_cbc_cksum() was like the one from
2049 mit_des_cbc_cksum(), except the bytes were swapped.
2050 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2052 *) Allow an application to disable the automatic SSL chain building.
2053 Before this a rather primitive chain build was always performed in
2054 ssl3_output_cert_chain(): an application had no way to send the
2055 correct chain if the automatic operation produced an incorrect result.
2057 Now the chain builder is disabled if either:
2059 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2061 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2063 The reasoning behind this is that an application would not want the
2064 auto chain building to take place if extra chain certificates are
2065 present and it might also want a means of sending no additional
2066 certificates (for example the chain has two certificates and the
2070 *) Add the possibility to build without the ENGINE framework.
2071 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2073 *) Under Win32 gmtime() can return NULL: check return value in
2074 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2077 *) DSA routines: under certain error conditions uninitialized BN objects
2078 could be freed. Solution: make sure initialization is performed early
2079 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2080 Nils Larsch <nla@trustcenter.de> via PR#459)
2083 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2084 checked on reconnect on the client side, therefore session resumption
2085 could still fail with a "ssl session id is different" error. This
2086 behaviour is masked when SSL_OP_ALL is used due to
2087 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2088 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2089 followup to PR #377.
2092 *) IA-32 assembler support enhancements: unified ELF targets, support
2093 for SCO/Caldera platforms, fix for Cygwin shared build.
2096 *) Add support for FreeBSD on sparc64. As a consequence, support for
2097 FreeBSD on non-x86 processors is separate from x86 processors on
2098 the config script, much like the NetBSD support.
2099 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2101 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2103 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2106 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2107 code (06) was taken as the first octet of the session ID and the last
2108 octet was ignored consequently. As a result SSLv2 client side session
2109 caching could not have worked due to the session ID mismatch between
2111 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2115 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2116 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2120 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2121 seems that in spite of existing for more than a year, many application
2122 author have done nothing to provide the necessary callbacks, which
2123 means that this particular engine will not work properly anywhere.
2124 This is a very unfortunate situation which forces us, in the name
2125 of usability, to give the hw_ncipher.c a static lock, which is part
2127 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2128 appear in 0.9.8 or later. We EXPECT application authors to have
2129 dealt properly with this when 0.9.8 is released (unless we actually
2130 make such changes in the libcrypto locking code that changes will
2131 have to be made anyway).
2134 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2135 octets have been read, EOF or an error occurs. Without this change
2136 some truncated ASN1 structures will not produce an error.
2139 *) Disable Heimdal support, since it hasn't been fully implemented.
2140 Still give the possibility to force the use of Heimdal, but with
2141 warnings and a request that patches get sent to openssl-dev.
2144 *) Add the VC-CE target, introduce the WINCE sysname, and add
2145 INSTALL.WCE and appropriate conditionals to make it build.
2146 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2148 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2149 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2150 edit numbers of the version.
2151 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2153 *) Introduce safe string copy and catenation functions
2154 (BUF_strlcpy() and BUF_strlcat()).
2155 [Ben Laurie (CHATS) and Richard Levitte]
2157 *) Avoid using fixed-size buffers for one-line DNs.
2158 [Ben Laurie (CHATS)]
2160 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2161 resizing buffers containing secrets, and use where appropriate.
2162 [Ben Laurie (CHATS)]
2164 *) Avoid using fixed size buffers for configuration file location.
2165 [Ben Laurie (CHATS)]
2167 *) Avoid filename truncation for various CA files.
2168 [Ben Laurie (CHATS)]
2170 *) Use sizeof in preference to magic numbers.
2171 [Ben Laurie (CHATS)]
2173 *) Avoid filename truncation in cert requests.
2174 [Ben Laurie (CHATS)]
2176 *) Add assertions to check for (supposedly impossible) buffer
2178 [Ben Laurie (CHATS)]
2180 *) Don't cache truncated DNS entries in the local cache (this could
2181 potentially lead to a spoofing attack).
2182 [Ben Laurie (CHATS)]
2184 *) Fix various buffers to be large enough for hex/decimal
2185 representations in a platform independent manner.
2186 [Ben Laurie (CHATS)]
2188 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2189 resizing buffers containing secrets, and use where appropriate.
2190 [Ben Laurie (CHATS)]
2192 *) Add BIO_indent() to avoid much slightly worrying code to do
2194 [Ben Laurie (CHATS)]
2196 *) Convert sprintf()/BIO_puts() to BIO_printf().
2197 [Ben Laurie (CHATS)]
2199 *) buffer_gets() could terminate with the buffer only half
2201 [Ben Laurie (CHATS)]
2203 *) Add assertions to prevent user-supplied crypto functions from
2204 overflowing internal buffers by having large block sizes, etc.
2205 [Ben Laurie (CHATS)]
2207 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2209 [Ben Laurie (CHATS)]
2211 *) Eliminate unused copy of key in RC4.
2212 [Ben Laurie (CHATS)]
2214 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2215 [Ben Laurie (CHATS)]
2217 *) Fix off-by-one error in EGD path.
2218 [Ben Laurie (CHATS)]
2220 *) If RANDFILE path is too long, ignore instead of truncating.
2221 [Ben Laurie (CHATS)]
2223 *) Eliminate unused and incorrectly sized X.509 structure
2225 [Ben Laurie (CHATS)]
2227 *) Eliminate unused and dangerous function knumber().
2228 [Ben Laurie (CHATS)]
2230 *) Eliminate unused and dangerous structure, KSSL_ERR.
2231 [Ben Laurie (CHATS)]
2233 *) Protect against overlong session ID context length in an encoded
2234 session object. Since these are local, this does not appear to be
2236 [Ben Laurie (CHATS)]
2238 *) Change from security patch (see 0.9.6e below) that did not affect
2239 the 0.9.6 release series:
2241 Remote buffer overflow in SSL3 protocol - an attacker could
2242 supply an oversized master key in Kerberos-enabled versions.
2244 [Ben Laurie (CHATS)]
2246 *) Change the SSL kerb5 codes to match RFC 2712.
2249 *) Make -nameopt work fully for req and add -reqopt switch.
2250 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2252 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2253 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2255 *) Make sure tests can be performed even if the corresponding algorithms
2256 have been removed entirely. This was also the last step to make
2257 OpenSSL compilable with DJGPP under all reasonable conditions.
2258 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2260 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2261 to allow version independent disabling of normally unselected ciphers,
2262 which may be activated as a side-effect of selecting a single cipher.
2264 (E.g., cipher list string "RSA" enables ciphersuites that are left
2265 out of "ALL" because they do not provide symmetric encryption.
2266 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2267 [Lutz Jaenicke, Bodo Moeller]
2269 *) Add appropriate support for separate platform-dependent build
2270 directories. The recommended way to make a platform-dependent
2271 build directory is the following (tested on Linux), maybe with
2274 # Place yourself outside of the OpenSSL source tree. In
2275 # this example, the environment variable OPENSSL_SOURCE
2276 # is assumed to contain the absolute OpenSSL source directory.
2277 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2278 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2279 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2280 mkdir -p `dirname $F`
2281 ln -s $OPENSSL_SOURCE/$F $F
2284 To be absolutely sure not to disturb the source tree, a "make clean"
2285 is a good thing. If it isn't successfull, don't worry about it,
2286 it probably means the source directory is very clean.
2289 *) Make sure any ENGINE control commands make local copies of string
2290 pointers passed to them whenever necessary. Otherwise it is possible
2291 the caller may have overwritten (or deallocated) the original string
2292 data when a later ENGINE operation tries to use the stored values.
2293 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2295 *) Improve diagnostics in file reading and command-line digests.
2296 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2298 *) Add AES modes CFB and OFB to the object database. Correct an
2299 error in AES-CFB decryption.
2302 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2303 allows existing EVP_CIPHER_CTX structures to be reused after
2304 calling EVP_*Final(). This behaviour is used by encryption
2305 BIOs and some applications. This has the side effect that
2306 applications must explicitly clean up cipher contexts with
2307 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2310 *) Check the values of dna and dnb in bn_mul_recursive before calling
2311 bn_mul_comba (a non zero value means the a or b arrays do not contain
2312 n2 elements) and fallback to bn_mul_normal if either is not zero.
2315 *) Fix escaping of non-ASCII characters when using the -subj option
2316 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2319 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2320 form for "surname", serialNumber has no short form.
2321 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2322 therefore remove "mail" short name for "internet 7".
2323 The OID for unique identifiers in X509 certificates is
2324 x500UniqueIdentifier, not uniqueIdentifier.
2325 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2328 *) Add an "init" command to the ENGINE config module and auto initialize
2329 ENGINEs. Without any "init" command the ENGINE will be initialized
2330 after all ctrl commands have been executed on it. If init=1 the
2331 ENGINE is initailized at that point (ctrls before that point are run
2332 on the uninitialized ENGINE and after on the initialized one). If
2333 init=0 then the ENGINE will not be iniatialized at all.
2336 *) Fix the 'app_verify_callback' interface so that the user-defined
2337 argument is actually passed to the callback: In the
2338 SSL_CTX_set_cert_verify_callback() prototype, the callback
2339 declaration has been changed from
2342 int (*cb)(X509_STORE_CTX *,void *);
2343 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2344 i=s->ctx->app_verify_callback(&ctx)
2345 has been changed into
2346 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2348 To update applications using SSL_CTX_set_cert_verify_callback(),
2349 a dummy argument can be added to their callback functions.
2350 [D. K. Smetters <smetters@parc.xerox.com>]
2352 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2353 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2355 *) Add and OPENSSL_LOAD_CONF define which will cause
2356 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2357 This allows older applications to transparently support certain
2358 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2359 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2360 load the config file and OPENSSL_add_all_algorithms_conf() which will
2361 always load it have also been added.
2364 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2365 Adjust NIDs and EVP layer.
2366 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2368 *) Config modules support in openssl utility.
2370 Most commands now load modules from the config file,
2371 though in a few (such as version) this isn't done
2372 because it couldn't be used for anything.
2374 In the case of ca and req the config file used is
2375 the same as the utility itself: that is the -config
2376 command line option can be used to specify an
2380 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2381 use "openssl_conf" if filename is NULL use default openssl config file.
2384 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2385 config section name. Add a new flag to tolerate a missing config file
2386 and move code to CONF_modules_load_file().
2389 *) Support for crypto accelerator cards from Accelerated Encryption
2390 Processing, www.aep.ie. (Use engine 'aep')
2391 The support was copied from 0.9.6c [engine] and adapted/corrected
2392 to work with the new engine framework.
2393 [AEP Inc. and Richard Levitte]
2395 *) Support for SureWare crypto accelerator cards from Baltimore
2396 Technologies. (Use engine 'sureware')
2397 The support was copied from 0.9.6c [engine] and adapted
2398 to work with the new engine framework.
2401 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2402 make the newer ENGINE framework commands for the CHIL engine work.
2403 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2405 *) Make it possible to produce shared libraries on ReliantUNIX.
2406 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2408 *) Add the configuration target debug-linux-ppro.
2409 Make 'openssl rsa' use the general key loading routines
2410 implemented in apps.c, and make those routines able to
2411 handle the key format FORMAT_NETSCAPE and the variant
2413 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2415 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2416 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2418 *) Add -keyform to rsautl, and document -engine.
2419 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2421 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2422 BIO_R_NO_SUCH_FILE error code rather than the generic
2423 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2426 *) Add new functions
2428 ERR_peek_last_error_line
2429 ERR_peek_last_error_line_data.
2430 These are similar to
2433 ERR_peek_error_line_data,
2434 but report on the latest error recorded rather than the first one
2435 still in the error queue.
2436 [Ben Laurie, Bodo Moeller]
2438 *) default_algorithms option in ENGINE config module. This allows things
2440 default_algorithms = ALL
2441 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2444 *) Prelminary ENGINE config module.
2447 *) New experimental application configuration code.
2450 *) Change the AES code to follow the same name structure as all other
2451 symmetric ciphers, and behave the same way. Move everything to
2452 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2453 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2455 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2456 [Ben Laurie and Theo de Raadt]
2458 *) Add option to output public keys in req command.
2459 [Massimiliano Pala madwolf@openca.org]
2461 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2462 (up to about 10% better than before for P-192 and P-224).
2465 *) New functions/macros
2467 SSL_CTX_set_msg_callback(ctx, cb)
2468 SSL_CTX_set_msg_callback_arg(ctx, arg)
2469 SSL_set_msg_callback(ssl, cb)
2470 SSL_set_msg_callback_arg(ssl, arg)
2472 to request calling a callback function
2474 void cb(int write_p, int version, int content_type,
2475 const void *buf, size_t len, SSL *ssl, void *arg)
2477 whenever a protocol message has been completely received
2478 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2479 protocol version according to which the SSL library interprets
2480 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2481 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2482 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2483 specification (change_cipher_spec(20), alert(21), handshake(22)).
2484 'buf' and 'len' point to the actual message, 'ssl' to the
2485 SSL object, and 'arg' is the application-defined value set by
2486 SSL[_CTX]_set_msg_callback_arg().
2488 'openssl s_client' and 'openssl s_server' have new '-msg' options
2489 to enable a callback that displays all protocol messages.
2492 *) Change the shared library support so shared libraries are built as
2493 soon as the corresponding static library is finished, and thereby get
2494 openssl and the test programs linked against the shared library.
2495 This still only happens when the keyword "shard" has been given to
2496 the configuration scripts.
2498 NOTE: shared library support is still an experimental thing, and
2499 backward binary compatibility is still not guaranteed.
2500 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2502 *) Add support for Subject Information Access extension.
2503 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2505 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2506 additional bytes when new memory had to be allocated, not just
2507 when reusing an existing buffer.
2510 *) New command line and configuration option 'utf8' for the req command.
2511 This allows field values to be specified as UTF8 strings.
2514 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2515 runs for the former and machine-readable output for the latter.
2518 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2519 of the e-mail address in the DN (i.e., it will go into a certificate
2520 extension only). The new configuration file option 'email_in_dn = no'
2521 has the same effect.
2522 [Massimiliano Pala madwolf@openca.org]
2524 *) Change all functions with names starting with des_ to be starting
2525 with DES_ instead. Add wrappers that are compatible with libdes,
2526 but are named _ossl_old_des_*. Finally, add macros that map the
2527 des_* symbols to the corresponding _ossl_old_des_* if libdes
2528 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2529 desired, the des_* symbols will be mapped to DES_*, with one
2532 Since we provide two compatibility mappings, the user needs to
2533 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2534 compatibility is desired. The default (i.e., when that macro
2535 isn't defined) is OpenSSL 0.9.6c compatibility.
2537 There are also macros that enable and disable the support of old
2538 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2539 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2540 are defined, the default will apply: to support the old des routines.
2542 In either case, one must include openssl/des.h to get the correct
2543 definitions. Do not try to just include openssl/des_old.h, that
2546 NOTE: This is a major break of an old API into a new one. Software
2547 authors are encouraged to switch to the DES_ style functions. Some
2548 time in the future, des_old.h and the libdes compatibility functions
2549 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2550 default), and then completely removed.
2553 *) Test for certificates which contain unsupported critical extensions.
2554 If such a certificate is found during a verify operation it is
2555 rejected by default: this behaviour can be overridden by either
2556 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2557 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2558 X509_supported_extension() has also been added which returns 1 if a
2559 particular extension is supported.
2562 *) Modify the behaviour of EVP cipher functions in similar way to digests
2563 to retain compatibility with existing code.
2566 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2567 compatibility with existing code. In particular the 'ctx' parameter does
2568 not have to be to be initialized before the call to EVP_DigestInit() and
2569 it is tidied up after a call to EVP_DigestFinal(). New function
2570 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2571 EVP_MD_CTX_copy() changed to not require the destination to be
2572 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2573 requires the destination to be valid.
2575 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2576 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2579 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2580 so that complete 'Handshake' protocol structures are kept in memory
2581 instead of overwriting 'msg_type' and 'length' with 'body' data.
2584 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2585 [Massimo Santin via Richard Levitte]
2587 *) Major restructuring to the underlying ENGINE code. This includes
2588 reduction of linker bloat, separation of pure "ENGINE" manipulation
2589 (initialisation, etc) from functionality dealing with implementations
2590 of specific crypto iterfaces. This change also introduces integrated
2591 support for symmetric ciphers and digest implementations - so ENGINEs
2592 can now accelerate these by providing EVP_CIPHER and EVP_MD
2593 implementations of their own. This is detailed in crypto/engine/README
2594 as it couldn't be adequately described here. However, there are a few
2595 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2596 were changed in the original introduction of ENGINE code have now
2597 reverted back - the hooking from this code to ENGINE is now a good
2598 deal more passive and at run-time, operations deal directly with
2599 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2600 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2601 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2602 they were not being used by the framework as there is no concept of a
2603 BIGNUM_METHOD and they could not be generalised to the new
2604 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2605 ENGINE_cpy() has been removed as it cannot be consistently defined in
2609 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2612 *) Change mkdef.pl to sort symbols that get the same entry number,
2613 and make sure the automatically generated functions ERR_load_*
2614 become part of libeay.num as well.
2617 *) New function SSL_renegotiate_pending(). This returns true once
2618 renegotiation has been requested (either SSL_renegotiate() call
2619 or HelloRequest/ClientHello receveived from the peer) and becomes
2620 false once a handshake has been completed.
2621 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2622 sends a HelloRequest, but does not ensure that a handshake takes
2623 place. SSL_renegotiate_pending() is useful for checking if the
2624 client has followed the request.)
2627 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2628 By default, clients may request session resumption even during
2629 renegotiation (if session ID contexts permit); with this option,
2630 session resumption is possible only in the first handshake.
2632 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2633 more bits available for options that should not be part of
2634 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2637 *) Add some demos for certificate and certificate request creation.
2640 *) Make maximum certificate chain size accepted from the peer application
2641 settable (SSL*_get/set_max_cert_list()), as proposed by
2642 "Douglas E. Engert" <deengert@anl.gov>.
2645 *) Add support for shared libraries for Unixware-7
2646 (Boyd Lynn Gerber <gerberb@zenez.com>).
2649 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2650 be done prior to destruction. Use this to unload error strings from
2651 ENGINEs that load their own error strings. NB: This adds two new API
2652 functions to "get" and "set" this destroy handler in an ENGINE.
2655 *) Alter all existing ENGINE implementations (except "openssl" and
2656 "openbsd") to dynamically instantiate their own error strings. This
2657 makes them more flexible to be built both as statically-linked ENGINEs
2658 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2659 Also, add stub code to each that makes building them as self-contained
2660 shared-libraries easier (see README.ENGINE).
2663 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2664 implementations into applications that are completely implemented in
2665 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2666 commands that can be used to configure what shared-library to load and
2667 to control aspects of the way it is handled. Also, made an update to
2668 the README.ENGINE file that brings its information up-to-date and
2669 provides some information and instructions on the "dynamic" ENGINE
2670 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2673 *) Make it possible to unload ranges of ERR strings with a new
2674 "ERR_unload_strings" function.
2677 *) Add a copy() function to EVP_MD.
2680 *) Make EVP_MD routines take a context pointer instead of just the
2681 md_data void pointer.
2684 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2685 that the digest can only process a single chunk of data
2686 (typically because it is provided by a piece of
2687 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2688 is only going to provide a single chunk of data, and hence the
2689 framework needn't accumulate the data for oneshot drivers.
2692 *) As with "ERR", make it possible to replace the underlying "ex_data"
2693 functions. This change also alters the storage and management of global
2694 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2695 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2696 index counters. The API functions that use this state have been changed
2697 to take a "class_index" rather than pointers to the class's local STACK
2698 and counter, and there is now an API function to dynamically create new
2699 classes. This centralisation allows us to (a) plug a lot of the
2700 thread-safety problems that existed, and (b) makes it possible to clean
2701 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2702 such data would previously have always leaked in application code and
2703 workarounds were in place to make the memory debugging turn a blind eye
2704 to it. Application code that doesn't use this new function will still
2705 leak as before, but their memory debugging output will announce it now
2706 rather than letting it slide.
2708 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2709 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2710 has a return value to indicate success or failure.
2713 *) Make it possible to replace the underlying "ERR" functions such that the
2714 global state (2 LHASH tables and 2 locks) is only used by the "default"
2715 implementation. This change also adds two functions to "get" and "set"
2716 the implementation prior to it being automatically set the first time
2717 any other ERR function takes place. Ie. an application can call "get",
2718 pass the return value to a module it has just loaded, and that module
2719 can call its own "set" function using that value. This means the
2720 module's "ERR" operations will use (and modify) the error state in the
2721 application and not in its own statically linked copy of OpenSSL code.
2724 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2725 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2726 the operation, and provides a more encapsulated way for external code
2727 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2728 to use these functions rather than manually incrementing the counts.
2730 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2733 *) Add EVP test program.
2736 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2739 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2740 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2741 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2742 These allow a CRL to be built without having to access X509_CRL fields
2743 directly. Modify 'ca' application to use new functions.
2746 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2747 bug workarounds. Rollback attack detection is a security feature.
2748 The problem will only arise on OpenSSL servers when TLSv1 is not
2749 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2750 Software authors not wanting to support TLSv1 will have special reasons
2751 for their choice and can explicitly enable this option.
2752 [Bodo Moeller, Lutz Jaenicke]
2754 *) Rationalise EVP so it can be extended: don't include a union of
2755 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2756 (similar to those existing for EVP_CIPHER_CTX).
2761 EVP_MD_CTX_init(&md); /* new function call */
2762 EVP_DigestInit(&md, EVP_sha1());
2763 EVP_DigestUpdate(&md, in, len);
2764 EVP_DigestFinal(&md, out, NULL);
2765 EVP_MD_CTX_cleanup(&md); /* new function call */
2769 *) Make DES key schedule conform to the usual scheme, as well as
2770 correcting its structure. This means that calls to DES functions
2771 now have to pass a pointer to a des_key_schedule instead of a
2772 plain des_key_schedule (which was actually always a pointer
2775 des_key_schedule ks;
2777 des_set_key_checked(..., &ks);
2778 des_ncbc_encrypt(..., &ks, ...);
2780 (Note that a later change renames 'des_...' into 'DES_...'.)
2783 *) Initial reduction of linker bloat: the use of some functions, such as
2784 PEM causes large amounts of unused functions to be linked in due to
2785 poor organisation. For example pem_all.c contains every PEM function
2786 which has a knock on effect of linking in large amounts of (unused)
2787 ASN1 code. Grouping together similar functions and splitting unrelated
2788 functions prevents this.
2791 *) Cleanup of EVP macros.
2794 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2795 correct _ecb suffix.
2798 *) Add initial OCSP responder support to ocsp application. The
2799 revocation information is handled using the text based index
2800 use by the ca application. The responder can either handle
2801 requests generated internally, supplied in files (for example
2802 via a CGI script) or using an internal minimal server.
2805 *) Add configuration choices to get zlib compression for TLS.
2808 *) Changes to Kerberos SSL for RFC 2712 compliance:
2809 1. Implemented real KerberosWrapper, instead of just using
2810 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2811 2. Implemented optional authenticator field of KerberosWrapper.
2813 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2814 and authenticator structs; see crypto/krb5/.
2816 Generalized Kerberos calls to support multiple Kerberos libraries.
2817 [Vern Staats <staatsvr@asc.hpc.mil>,
2818 Jeffrey Altman <jaltman@columbia.edu>
2819 via Richard Levitte]
2821 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2822 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2823 values for each of the key sizes rather than having just
2824 parameters (and 'speed' generating keys each time).
2827 *) Speed up EVP routines.
2830 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2831 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2832 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2833 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2835 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2836 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2837 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2840 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2842 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2845 *) Added the OS2-EMX target.
2846 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2848 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2849 to support NCONF routines in extension code. New function CONF_set_nconf()
2850 to allow functions which take an NCONF to also handle the old LHASH
2851 structure: this means that the old CONF compatible routines can be
2852 retained (in particular wrt extensions) without having to duplicate the
2853 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2856 *) Enhance the general user interface with mechanisms for inner control
2857 and with possibilities to have yes/no kind of prompts.
2860 *) Change all calls to low level digest routines in the library and
2861 applications to use EVP. Add missing calls to HMAC_cleanup() and
2862 don't assume HMAC_CTX can be copied using memcpy().
2863 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2865 *) Add the possibility to control engines through control names but with
2866 arbitrary arguments instead of just a string.
2867 Change the key loaders to take a UI_METHOD instead of a callback
2868 function pointer. NOTE: this breaks binary compatibility with earlier
2869 versions of OpenSSL [engine].
2870 Adapt the nCipher code for these new conditions and add a card insertion
2874 *) Enhance the general user interface with mechanisms to better support
2875 dialog box interfaces, application-defined prompts, the possibility
2876 to use defaults (for example default passwords from somewhere else)
2877 and interrupts/cancellations.
2880 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2881 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2884 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2885 tidy up some unnecessarily weird code in 'sk_new()').
2886 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2888 *) Change the key loading routines for ENGINEs to use the same kind
2889 callback (pem_password_cb) as all other routines that need this
2893 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2894 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2895 than this minimum value is recommended.
2898 *) New random seeder for OpenVMS, using the system process statistics
2899 that are easily reachable.
2902 *) Windows apparently can't transparently handle global
2903 variables defined in DLLs. Initialisations such as:
2905 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2907 wont compile. This is used by the any applications that need to
2908 declare their own ASN1 modules. This was fixed by adding the option
2909 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2910 needed for static libraries under Win32.
2913 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2914 setting of purpose and trust fields. New X509_STORE trust and
2915 purpose functions and tidy up setting in other SSL functions.
2918 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2919 structure. These are inherited by X509_STORE_CTX when it is
2920 initialised. This allows various defaults to be set in the
2921 X509_STORE structure (such as flags for CRL checking and custom
2922 purpose or trust settings) for functions which only use X509_STORE_CTX
2923 internally such as S/MIME.
2925 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2926 trust settings if they are not set in X509_STORE. This allows X509_STORE
2927 purposes and trust (in S/MIME for example) to override any set by default.
2929 Add command line options for CRL checking to smime, s_client and s_server
2933 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2934 are set then the CRL is looked up in the X509_STORE structure and
2935 its validity and signature checked, then if the certificate is found
2936 in the CRL the verify fails with a revoked error.
2938 Various new CRL related callbacks added to X509_STORE_CTX structure.
2940 Command line options added to 'verify' application to support this.
2942 This needs some additional work, such as being able to handle multiple
2943 CRLs with different times, extension based lookup (rather than just
2944 by subject name) and ultimately more complete V2 CRL extension
2948 *) Add a general user interface API (crypto/ui/). This is designed
2949 to replace things like des_read_password and friends (backward
2950 compatibility functions using this new API are provided).
2951 The purpose is to remove prompting functions from the DES code
2952 section as well as provide for prompting through dialog boxes in
2953 a window system and the like.
2956 *) Add "ex_data" support to ENGINE so implementations can add state at a
2957 per-structure level rather than having to store it globally.
2960 *) Make it possible for ENGINE structures to be copied when retrieved by
2961 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2962 This causes the "original" ENGINE structure to act like a template,
2963 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2964 operational state can be localised to each ENGINE structure, despite the
2965 fact they all share the same "methods". New ENGINE structures returned in
2966 this case have no functional references and the return value is the single
2967 structural reference. This matches the single structural reference returned
2968 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2972 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2973 needs to match any other type at all we need to manually clear the
2977 *) Changes to the "openssl engine" utility to include;
2978 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2979 about an ENGINE's available control commands.
2980 - executing control commands from command line arguments using the
2981 '-pre' and '-post' switches. '-post' is only used if '-t' is
2982 specified and the ENGINE is successfully initialised. The syntax for
2983 the individual commands are colon-separated, for example;
2984 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2987 *) New dynamic control command support for ENGINEs. ENGINEs can now
2988 declare their own commands (numbers), names (strings), descriptions,
2989 and input types for run-time discovery by calling applications. A
2990 subset of these commands are implicitly classed as "executable"
2991 depending on their input type, and only these can be invoked through
2992 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2993 can be based on user input, config files, etc). The distinction is
2994 that "executable" commands cannot return anything other than a boolean
2995 result and can only support numeric or string input, whereas some
2996 discoverable commands may only be for direct use through
2997 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2998 pointers, or other custom uses. The "executable" commands are to
2999 support parameterisations of ENGINE behaviour that can be
3000 unambiguously defined by ENGINEs and used consistently across any
3001 OpenSSL-based application. Commands have been added to all the
3002 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3003 control over shared-library paths without source code alterations.
3006 *) Changed all ENGINE implementations to dynamically allocate their
3007 ENGINEs rather than declaring them statically. Apart from this being
3008 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3009 this also allows the implementations to compile without using the
3010 internal engine_int.h header.
3013 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3014 'const' value. Any code that should be able to modify a RAND_METHOD
3015 should already have non-const pointers to it (ie. they should only
3016 modify their own ones).
3019 *) Made a variety of little tweaks to the ENGINE code.
3020 - "atalla" and "ubsec" string definitions were moved from header files
3021 to C code. "nuron" string definitions were placed in variables
3022 rather than hard-coded - allowing parameterisation of these values
3023 later on via ctrl() commands.
3024 - Removed unused "#if 0"'d code.
3025 - Fixed engine list iteration code so it uses ENGINE_free() to release
3026 structural references.
3027 - Constified the RAND_METHOD element of ENGINE structures.
3028 - Constified various get/set functions as appropriate and added
3029 missing functions (including a catch-all ENGINE_cpy that duplicates
3030 all ENGINE values onto a new ENGINE except reference counts/state).
3031 - Removed NULL parameter checks in get/set functions. Setting a method
3032 or function to NULL is a way of cancelling out a previously set
3033 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3034 and doesn't justify the extra error symbols and code.
3035 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3036 flags from engine_int.h to engine.h.
3037 - Changed prototypes for ENGINE handler functions (init(), finish(),
3038 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3041 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3042 to the algorithm using long division. The binary algorithm can be
3043 used only if the modulus is odd. On 32-bit systems, it is faster
3044 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3045 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3046 up to 450 bits. In 64-bit environments, the binary algorithm
3047 appears to be advantageous for much longer moduli; here we use it
3048 for moduli up to 2048 bits.
3051 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3052 could not support the combine flag in choice fields.
3055 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3056 extensions from a certificate request to the certificate.
3059 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3060 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3061 file: this allows the display of the certificate about to be
3062 signed to be customised, to allow certain fields to be included
3063 or excluded and extension details. The old system didn't display
3064 multicharacter strings properly, omitted fields not in the policy
3065 and couldn't display additional details such as extensions.
3068 *) Function EC_POINTs_mul for multiple scalar multiplication
3069 of an arbitrary number of elliptic curve points
3070 \sum scalars[i]*points[i],
3071 optionally including the generator defined for the EC_GROUP:
3072 scalar*generator + \sum scalars[i]*points[i].
3074 EC_POINT_mul is a simple wrapper function for the typical case
3075 that the point list has just one item (besides the optional
3079 *) First EC_METHODs for curves over GF(p):
3081 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3082 operations and provides various method functions that can also
3083 operate with faster implementations of modular arithmetic.
3085 EC_GFp_mont_method() reuses most functions that are part of
3086 EC_GFp_simple_method, but uses Montgomery arithmetic.
3088 [Bodo Moeller; point addition and point doubling
3089 implementation directly derived from source code provided by
3090 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3092 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3093 crypto/ec/ec_lib.c):
3095 Curves are EC_GROUP objects (with an optional group generator)
3096 based on EC_METHODs that are built into the library.
3098 Points are EC_POINT objects based on EC_GROUP objects.
3100 Most of the framework would be able to handle curves over arbitrary
3101 finite fields, but as there are no obvious types for fields other
3102 than GF(p), some functions are limited to that for now.
3105 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3106 that the file contains a complete HTTP response.
3109 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3110 change the def and num file printf format specifier from "%-40sXXX"
3111 to "%-39s XXX". The latter will always guarantee a space after the
3112 field while the former will cause them to run together if the field
3113 is 40 of more characters long.
3116 *) Constify the cipher and digest 'method' functions and structures
3117 and modify related functions to take constant EVP_MD and EVP_CIPHER
3121 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3122 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3125 *) Modify EVP_Digest*() routines so they now return values. Although the
3126 internal software routines can never fail additional hardware versions
3130 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3132 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3133 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3136 ERR_R_NESTED_ASN1_ERROR
3138 ERR_R_MISSING_ASN1_EOS
3139 were 4 .. 9, conflicting with
3140 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3142 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3143 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3145 Add new error code 'ERR_R_INTERNAL_ERROR'.
3148 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3152 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3153 sets the subject name for a new request or supersedes the
3154 subject name in a given request. Formats that can be parsed are
3155 'CN=Some Name, OU=myOU, C=IT'
3157 'CN=Some Name/OU=myOU/C=IT'.
3159 Add options '-batch' and '-verbose' to 'openssl req'.
3160 [Massimiliano Pala <madwolf@hackmasters.net>]
3162 *) Introduce the possibility to access global variables through
3163 functions on platform were that's the best way to handle exporting
3164 global variables in shared libraries. To enable this functionality,
3165 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3166 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3167 is normally done by Configure or something similar).
3169 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3170 in the source file (foo.c) like this:
3172 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3173 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3175 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3176 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3178 OPENSSL_DECLARE_GLOBAL(int,foo);
3179 #define foo OPENSSL_GLOBAL_REF(foo)
3180 OPENSSL_DECLARE_GLOBAL(double,bar);
3181 #define bar OPENSSL_GLOBAL_REF(bar)
3183 The #defines are very important, and therefore so is including the
3184 header file everywhere where the defined globals are used.
3186 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3187 of ASN.1 items, but that structure is a bit different.
3189 The largest change is in util/mkdef.pl which has been enhanced with
3190 better and easier to understand logic to choose which symbols should
3191 go into the Windows .def files as well as a number of fixes and code
3192 cleanup (among others, algorithm keywords are now sorted
3193 lexicographically to avoid constant rewrites).
3196 *) In BN_div() keep a copy of the sign of 'num' before writing the
3197 result to 'rm' because if rm==num the value will be overwritten
3198 and produce the wrong result if 'num' is negative: this caused
3199 problems with BN_mod() and BN_nnmod().
3202 *) Function OCSP_request_verify(). This checks the signature on an
3203 OCSP request and verifies the signer certificate. The signer
3204 certificate is just checked for a generic purpose and OCSP request
3208 *) Add OCSP_check_validity() function to check the validity of OCSP
3209 responses. OCSP responses are prepared in real time and may only
3210 be a few seconds old. Simply checking that the current time lies
3211 between thisUpdate and nextUpdate max reject otherwise valid responses
3212 caused by either OCSP responder or client clock inaccuracy. Instead
3213 we allow thisUpdate and nextUpdate to fall within a certain period of
3214 the current time. The age of the response can also optionally be
3215 checked. Two new options -validity_period and -status_age added to
3219 *) If signature or public key algorithm is unrecognized print out its
3220 OID rather that just UNKNOWN.
3223 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3224 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3225 ID to be generated from the issuer certificate alone which can then be
3226 passed to OCSP_id_issuer_cmp().
3229 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3230 ASN1 modules to export functions returning ASN1_ITEM pointers
3231 instead of the ASN1_ITEM structures themselves. This adds several
3232 new macros which allow the underlying ASN1 function/structure to
3233 be accessed transparently. As a result code should not use ASN1_ITEM
3234 references directly (such as &X509_it) but instead use the relevant
3235 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3236 use of the new ASN1 code on platforms where exporting structures
3237 is problematical (for example in shared libraries) but exporting
3238 functions returning pointers to structures is not.
3241 *) Add support for overriding the generation of SSL/TLS session IDs.
3242 These callbacks can be registered either in an SSL_CTX or per SSL.
3243 The purpose of this is to allow applications to control, if they wish,
3244 the arbitrary values chosen for use as session IDs, particularly as it
3245 can be useful for session caching in multiple-server environments. A
3246 command-line switch for testing this (and any client code that wishes
3247 to use such a feature) has been added to "s_server".
3248 [Geoff Thorpe, Lutz Jaenicke]
3250 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3251 of the form '#if defined(...) || defined(...) || ...' and
3252 '#if !defined(...) && !defined(...) && ...'. This also avoids
3253 the growing number of special cases it was previously handling.
3256 *) Make all configuration macros available for application by making
3257 sure they are available in opensslconf.h, by giving them names starting
3258 with "OPENSSL_" to avoid conflicts with other packages and by making
3259 sure e_os2.h will cover all platform-specific cases together with
3261 Additionally, it is now possible to define configuration/platform-
3262 specific names (called "system identities"). In the C code, these
3263 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3264 macro with the name beginning with "OPENSSL_SYS_", which is determined
3265 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3269 *) New option -set_serial to 'req' and 'x509' this allows the serial
3270 number to use to be specified on the command line. Previously self
3271 signed certificates were hard coded with serial number 0 and the
3272 CA options of 'x509' had to use a serial number in a file which was
3276 *) New options to 'ca' utility to support V2 CRL entry extensions.
3277 Currently CRL reason, invalidity date and hold instruction are
3278 supported. Add new CRL extensions to V3 code and some new objects.
3281 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3282 disable standard block padding (aka PKCS#5 padding) in the EVP
3283 API, which was previously mandatory. This means that the data is
3284 not padded in any way and so the total length much be a multiple
3285 of the block size, otherwise an error occurs.
3288 *) Initial (incomplete) OCSP SSL support.
3291 *) New function OCSP_parse_url(). This splits up a URL into its host,
3292 port and path components: primarily to parse OCSP URLs. New -url
3293 option to ocsp utility.
3296 *) New nonce behavior. The return value of OCSP_check_nonce() now
3297 reflects the various checks performed. Applications can decide
3298 whether to tolerate certain situations such as an absent nonce
3299 in a response when one was present in a request: the ocsp application
3300 just prints out a warning. New function OCSP_add1_basic_nonce()
3301 this is to allow responders to include a nonce in a response even if
3302 the request is nonce-less.
3305 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3306 skipped when using openssl x509 multiple times on a single input file,
3307 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3310 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3311 set string type: to handle setting ASN1_TIME structures. Fix ca
3312 utility to correctly initialize revocation date of CRLs.
3315 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3316 the clients preferred ciphersuites and rather use its own preferences.
3317 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3318 Internet Explorer by ensuring unchanged hash method during stepup.
3319 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3322 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3323 to aes and add a new 'exist' option to print out symbols that don't
3327 *) Additional options to ocsp utility to allow flags to be set and
3328 additional certificates supplied.
3331 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3332 OCSP client a number of certificate to only verify the response
3336 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3337 handle the new API. Currently only ECB, CBC modes supported. Add new
3340 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3341 Encryption Standard (AES) Ciphersuites for Transport Layer
3342 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3343 not enabled by default and were not part of the "ALL" ciphersuite
3344 alias because they were not yet official; they could be
3345 explicitly requested by specifying the "AESdraft" ciphersuite
3346 group alias. In the final release of OpenSSL 0.9.7, the group
3347 alias is called "AES" and is part of "ALL".)
3348 [Ben Laurie, Steve Henson, Bodo Moeller]
3350 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3351 request to response.
3354 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3355 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3356 extract information from a certificate request. OCSP_response_create()
3357 creates a response and optionally adds a basic response structure.
3358 OCSP_basic_add1_status() adds a complete single response to a basic
3359 response and returns the OCSP_SINGLERESP structure just added (to allow
3360 extensions to be included for example). OCSP_basic_add1_cert() adds a
3361 certificate to a basic response and OCSP_basic_sign() signs a basic
3362 response with various flags. New helper functions ASN1_TIME_check()
3363 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3364 (converts ASN1_TIME to GeneralizedTime).
3367 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3368 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3369 structure from a certificate. X509_pubkey_digest() digests the public_key
3370 contents: this is used in various key identifiers.
3373 *) Make sk_sort() tolerate a NULL argument.
3374 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3376 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3377 passed by the function are trusted implicitly. If any of them signed the
3378 response then it is assumed to be valid and is not verified.
3381 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3382 to data. This was previously part of the PKCS7 ASN1 code. This
3383 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3384 [Steve Henson, reported by Kenneth R. Robinette
3385 <support@securenetterm.com>]
3387 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3388 routines: without these tracing memory leaks is very painful.
3389 Fix leaks in PKCS12 and PKCS7 routines.
3392 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3393 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3394 effectively meant GeneralizedTime would never be used. Now it
3395 is initialised to -1 but X509_time_adj() now has to check the value
3396 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3397 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3398 [Steve Henson, reported by Kenneth R. Robinette
3399 <support@securenetterm.com>]
3401 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3402 result in a zero length in the ASN1_INTEGER structure which was
3403 not consistent with the structure when d2i_ASN1_INTEGER() was used
3404 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3405 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3406 where it did not print out a minus for negative ASN1_INTEGER.
3409 *) Add summary printout to ocsp utility. The various functions which
3410 convert status values to strings have been renamed to:
3411 OCSP_response_status_str(), OCSP_cert_status_str() and
3412 OCSP_crl_reason_str() and are no longer static. New options
3413 to verify nonce values and to disable verification. OCSP response
3414 printout format cleaned up.
3417 *) Add additional OCSP certificate checks. These are those specified
3418 in RFC2560. This consists of two separate checks: the CA of the
3419 certificate being checked must either be the OCSP signer certificate
3420 or the issuer of the OCSP signer certificate. In the latter case the
3421 OCSP signer certificate must contain the OCSP signing extended key
3422 usage. This check is performed by attempting to match the OCSP
3423 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3424 in the OCSP_CERTID structures of the response.
3427 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3428 and related routines. This uses the standard OpenSSL certificate
3429 verify routines to perform initial checks (just CA validity) and
3430 to obtain the certificate chain. Then additional checks will be
3431 performed on the chain. Currently the root CA is checked to see
3432 if it is explicitly trusted for OCSP signing. This is used to set
3433 a root CA as a global signing root: that is any certificate that
3434 chains to that CA is an acceptable OCSP signing certificate.
3437 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3438 extensions from a separate configuration file.
3439 As when reading extensions from the main configuration file,
3440 the '-extensions ...' option may be used for specifying the
3442 [Massimiliano Pala <madwolf@comune.modena.it>]
3444 *) New OCSP utility. Allows OCSP requests to be generated or
3445 read. The request can be sent to a responder and the output
3446 parsed, outputed or printed in text form. Not complete yet:
3447 still needs to check the OCSP response validity.
3450 *) New subcommands for 'openssl ca':
3451 'openssl ca -status <serial>' prints the status of the cert with
3452 the given serial number (according to the index file).
3453 'openssl ca -updatedb' updates the expiry status of certificates
3455 [Massimiliano Pala <madwolf@comune.modena.it>]
3457 *) New '-newreq-nodes' command option to CA.pl. This is like
3458 '-newreq', but calls 'openssl req' with the '-nodes' option
3459 so that the resulting key is not encrypted.
3460 [Damien Miller <djm@mindrot.org>]
3462 *) New configuration for the GNU Hurd.
3463 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3465 *) Initial code to implement OCSP basic response verify. This
3466 is currently incomplete. Currently just finds the signer's
3467 certificate and verifies the signature on the response.
3470 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3471 value of OPENSSLDIR. This is available via the new '-d' option
3472 to 'openssl version', and is also included in 'openssl version -a'.
3475 *) Allowing defining memory allocation callbacks that will be given
3476 file name and line number information in additional arguments
3477 (a const char* and an int). The basic functionality remains, as
3478 well as the original possibility to just replace malloc(),
3479 realloc() and free() by functions that do not know about these
3480 additional arguments. To register and find out the current
3481 settings for extended allocation functions, the following
3482 functions are provided:
3484 CRYPTO_set_mem_ex_functions
3485 CRYPTO_set_locked_mem_ex_functions
3486 CRYPTO_get_mem_ex_functions
3487 CRYPTO_get_locked_mem_ex_functions
3489 These work the same way as CRYPTO_set_mem_functions and friends.
3490 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3491 extended allocation function is enabled.
3492 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3493 a conventional allocation function is enabled.
3494 [Richard Levitte, Bodo Moeller]
3496 *) Finish off removing the remaining LHASH function pointer casts.
3497 There should no longer be any prototype-casting required when using
3498 the LHASH abstraction, and any casts that remain are "bugs". See
3499 the callback types and macros at the head of lhash.h for details
3500 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3503 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3504 If /dev/[u]random devices are not available or do not return enough
3505 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3507 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3508 /etc/entropy will be queried once each in this sequence, quering stops
3509 when enough entropy was collected without querying more sockets.
3512 *) Change the Unix RAND_poll() variant to be able to poll several
3513 random devices, as specified by DEVRANDOM, until a sufficient amount
3514 of data has been collected. We spend at most 10 ms on each file
3515 (select timeout) and read in non-blocking mode. DEVRANDOM now
3516 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3517 (previously it was just the string "/dev/urandom"), so on typical
3518 platforms the 10 ms delay will never occur.
3519 Also separate out the Unix variant to its own file, rand_unix.c.
3520 For VMS, there's a currently-empty rand_vms.c.
3523 *) Move OCSP client related routines to ocsp_cl.c. These
3524 provide utility functions which an application needing
3525 to issue a request to an OCSP responder and analyse the
3526 response will typically need: as opposed to those which an
3527 OCSP responder itself would need which will be added later.
3529 OCSP_request_sign() signs an OCSP request with an API similar
3530 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3531 response. OCSP_response_get1_basic() extracts basic response
3532 from response. OCSP_resp_find_status(): finds and extracts status
3533 information from an OCSP_CERTID structure (which will be created
3534 when the request structure is built). These are built from lower
3535 level functions which work on OCSP_SINGLERESP structures but
3536 wont normally be used unless the application wishes to examine
3537 extensions in the OCSP response for example.
3539 Replace nonce routines with a pair of functions.
3540 OCSP_request_add1_nonce() adds a nonce value and optionally
3541 generates a random value. OCSP_check_nonce() checks the
3542 validity of the nonce in an OCSP response.
3545 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3546 This doesn't copy the supplied OCSP_CERTID and avoids the
3547 need to free up the newly created id. Change return type
3548 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3549 This can then be used to add extensions to the request.
3550 Deleted OCSP_request_new(), since most of its functionality
3551 is now in OCSP_REQUEST_new() (and the case insensitive name
3552 clash) apart from the ability to set the request name which
3553 will be added elsewhere.
3556 *) Update OCSP API. Remove obsolete extensions argument from
3557 various functions. Extensions are now handled using the new
3558 OCSP extension code. New simple OCSP HTTP function which
3559 can be used to send requests and parse the response.
3562 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3563 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3564 uses the special reorder version of SET OF to sort the attributes
3565 and reorder them to match the encoded order. This resolves a long
3566 standing problem: a verify on a PKCS7 structure just after signing
3567 it used to fail because the attribute order did not match the
3568 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3569 it uses the received order. This is necessary to tolerate some broken
3570 software that does not order SET OF. This is handled by encoding
3571 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3572 to produce the required SET OF.
3575 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3576 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3577 files to get correct declarations of the ASN.1 item variables.
3580 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3581 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3582 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3583 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3584 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3585 ASN1_ITEM and no wrapper functions.
3588 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3589 replace the old function pointer based I/O routines. Change most of
3590 the *_d2i_bio() and *_d2i_fp() functions to use these.
3593 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3594 lines, recognice more "algorithms" that can be deselected, and make
3595 it complain about algorithm deselection that isn't recognised.
3598 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3599 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3600 to use new functions. Add NO_ASN1_OLD which can be set to remove
3601 some old style ASN1 functions: this can be used to determine if old
3602 code will still work when these eventually go away.
3605 *) New extension functions for OCSP structures, these follow the
3606 same conventions as certificates and CRLs.
3609 *) New function X509V3_add1_i2d(). This automatically encodes and
3610 adds an extension. Its behaviour can be customised with various
3611 flags to append, replace or delete. Various wrappers added for
3612 certifcates and CRLs.
3615 *) Fix to avoid calling the underlying ASN1 print routine when
3616 an extension cannot be parsed. Correct a typo in the
3617 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3620 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3621 entries for variables.
3624 *) Add functionality to apps/openssl.c for detecting locking
3625 problems: As the program is single-threaded, all we have
3626 to do is register a locking callback using an array for
3627 storing which locks are currently held by the program.
3630 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3631 SSL_get_ex_data_X509_STORE_idx(), which is used in
3632 ssl_verify_cert_chain() and thus can be called at any time
3633 during TLS/SSL handshakes so that thread-safety is essential.
3634 Unfortunately, the ex_data design is not at all suited
3635 for multi-threaded use, so it probably should be abolished.
3638 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3639 [Broadcom, tweaked and integrated by Geoff Thorpe]
3641 *) Move common extension printing code to new function
3642 X509V3_print_extensions(). Reorganise OCSP print routines and
3643 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3646 *) New function X509_signature_print() to remove duplication in some
3650 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3651 set (this was treated exactly the same as SET OF previously). This
3652 is used to reorder the STACK representing the structure to match the
3653 encoding. This will be used to get round a problem where a PKCS7
3654 structure which was signed could not be verified because the STACK
3655 order did not reflect the encoded order.
3658 *) Reimplement the OCSP ASN1 module using the new code.
3661 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3662 for its ASN1 operations. The old style function pointers still exist
3663 for now but they will eventually go away.
3666 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3667 completely replaces the old ASN1 functionality with a table driven
3668 encoder and decoder which interprets an ASN1_ITEM structure describing
3669 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3670 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3671 has also been converted to the new form.
3674 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3675 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3676 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3677 for negative moduli.
3680 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3681 of not touching the result's sign bit.
3684 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3688 *) Changed the LHASH code to use prototypes for callbacks, and created
3689 macros to declare and implement thin (optionally static) functions
3690 that provide type-safety and avoid function pointer casting for the
3691 type-specific callbacks.
3694 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3696 [Veers Staats <staatsvr@asc.hpc.mil>,
3697 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3699 *) Reformat the FAQ so the different questions and answers can be divided
3700 in sections depending on the subject.
3703 *) Have the zlib compression code load ZLIB.DLL dynamically under
3707 *) New function BN_mod_sqrt for computing square roots modulo a prime
3708 (using the probabilistic Tonelli-Shanks algorithm unless
3709 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3710 be handled deterministically).
3711 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3713 *) Make BN_mod_inverse faster by explicitly handling small quotients
3714 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3715 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3718 *) New function BN_kronecker.
3721 *) Fix BN_gcd so that it works on negative inputs; the result is
3722 positive unless both parameters are zero.
3723 Previously something reasonably close to an infinite loop was
3724 possible because numbers could be growing instead of shrinking
3725 in the implementation of Euclid's algorithm.
3728 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3729 sign of the number in question.
3731 Fix BN_is_word(a,w) to work correctly for w == 0.
3733 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3734 because its test if the absolute value of 'a' equals 'w'.
3735 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3736 it exists mostly for use in the implementations of BN_is_zero(),
3737 BN_is_one(), and BN_is_word().
3740 *) New function BN_swap.
3743 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3744 the exponentiation functions are more likely to produce reasonable
3745 results on negative inputs.
3748 *) Change BN_mod_mul so that the result is always non-negative.
3749 Previously, it could be negative if one of the factors was negative;
3750 I don't think anyone really wanted that behaviour.
3753 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3754 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3755 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3756 and add new functions:
3765 BN_mod_lshift1_quick
3769 These functions always generate non-negative results.
3771 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3772 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3774 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3775 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3776 be reduced modulo m.
3777 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3780 The following entry accidentily appeared in the CHANGES file
3781 distributed with OpenSSL 0.9.7. The modifications described in
3782 it do *not* apply to OpenSSL 0.9.7.
3784 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3785 was actually never needed) and in BN_mul(). The removal in BN_mul()
3786 required a small change in bn_mul_part_recursive() and the addition
3787 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3788 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3789 bn_sub_words() and bn_add_words() except they take arrays with
3794 *) In 'openssl passwd', verify passwords read from the terminal
3795 unless the '-salt' option is used (which usually means that
3796 verification would just waste user's time since the resulting
3797 hash is going to be compared with some given password hash)
3798 or the new '-noverify' option is used.
3800 This is an incompatible change, but it does not affect
3801 non-interactive use of 'openssl passwd' (passwords on the command
3802 line, '-stdin' option, '-in ...' option) and thus should not
3806 *) Remove all references to RSAref, since there's no more need for it.
3809 *) Make DSO load along a path given through an environment variable
3810 (SHLIB_PATH) with shl_load().
3813 *) Constify the ENGINE code as a result of BIGNUM constification.
3814 Also constify the RSA code and most things related to it. In a
3815 few places, most notable in the depth of the ASN.1 code, ugly
3816 casts back to non-const were required (to be solved at a later
3820 *) Make it so the openssl application has all engines loaded by default.
3823 *) Constify the BIGNUM routines a little more.
3826 *) Add the following functions:
3828 ENGINE_load_cswift()
3830 ENGINE_load_atalla()
3832 ENGINE_load_builtin_engines()
3834 That way, an application can itself choose if external engines that
3835 are built-in in OpenSSL shall ever be used or not. The benefit is
3836 that applications won't have to be linked with libdl or other dso
3837 libraries unless it's really needed.
3839 Changed 'openssl engine' to load all engines on demand.
3840 Changed the engine header files to avoid the duplication of some
3841 declarations (they differed!).
3844 *) 'openssl engine' can now list capabilities.
3847 *) Better error reporting in 'openssl engine'.
3850 *) Never call load_dh_param(NULL) in s_server.
3853 *) Add engine application. It can currently list engines by name and
3854 identity, and test if they are actually available.
3857 *) Improve RPM specification file by forcing symbolic linking and making
3858 sure the installed documentation is also owned by root.root.
3859 [Damien Miller <djm@mindrot.org>]
3861 *) Give the OpenSSL applications more possibilities to make use of
3862 keys (public as well as private) handled by engines.
3865 *) Add OCSP code that comes from CertCo.
3868 *) Add VMS support for the Rijndael code.
3871 *) Added untested support for Nuron crypto accelerator.
3874 *) Add support for external cryptographic devices. This code was
3875 previously distributed separately as the "engine" branch.
3876 [Geoff Thorpe, Richard Levitte]
3878 *) Rework the filename-translation in the DSO code. It is now possible to
3879 have far greater control over how a "name" is turned into a filename
3880 depending on the operating environment and any oddities about the
3881 different shared library filenames on each system.
3884 *) Support threads on FreeBSD-elf in Configure.
3887 *) Fix for SHA1 assembly problem with MASM: it produces
3888 warnings about corrupt line number information when assembling
3889 with debugging information. This is caused by the overlapping
3891 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3894 NCONF_get_number() has no error checking at all. As a replacement,
3895 NCONF_get_number_e() is defined (_e for "error checking") and is
3896 promoted strongly. The old NCONF_get_number is kept around for
3897 binary backward compatibility.
3898 Make it possible for methods to load from something other than a BIO,
3899 by providing a function pointer that is given a name instead of a BIO.
3900 For example, this could be used to load configuration data from an
3904 *) Fix for non blocking accept BIOs. Added new I/O special reason
3905 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3906 with non blocking I/O was not possible because no retry code was
3907 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3911 *) Added the beginnings of Rijndael support.
3914 *) Fix for bug in DirectoryString mask setting. Add support for
3915 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3916 to allow certificate printing to more controllable, additional
3917 'certopt' option to 'x509' to allow new printing options to be
3921 *) Clean old EAY MD5 hack from e_os.h.
3924 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3926 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3927 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3928 [Joe Orton, Steve Henson]
3930 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3932 *) Fix additional bug revealed by the NISCC test suite:
3934 Stop bug triggering large recursion when presented with
3935 certain ASN.1 tags (CVE-2003-0851)
3938 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3940 *) Fix various bugs revealed by running the NISCC test suite:
3942 Stop out of bounds reads in the ASN1 code when presented with
3943 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3945 If verify callback ignores invalid public key errors don't try to check
3946 certificate signature with the NULL public key.
3950 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3951 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3955 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3956 extra data after the compression methods not only for TLS 1.0
3957 but also for SSL 3.0 (as required by the specification).
3958 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3960 *) Change X509_certificate_type() to mark the key as exported/exportable
3961 when it's 512 *bits* long, not 512 bytes.
3964 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3966 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3967 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3968 a protocol version number mismatch like a decryption error
3969 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3972 *) Turn on RSA blinding by default in the default implementation
3973 to avoid a timing attack. Applications that don't want it can call
3974 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3975 They would be ill-advised to do so in most cases.
3976 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3978 *) Change RSA blinding code so that it works when the PRNG is not
3979 seeded (in this case, the secret RSA exponent is abused as
3980 an unpredictable seed -- if it is not unpredictable, there
3981 is no point in blinding anyway). Make RSA blinding thread-safe
3982 by remembering the creator's thread ID in rsa->blinding and
3983 having all other threads use local one-time blinding factors
3984 (this requires more computation than sharing rsa->blinding, but
3985 avoids excessive locking; and if an RSA object is not shared
3986 between threads, blinding will still be very fast).
3989 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3991 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3992 via timing by performing a MAC computation even if incorrrect
3993 block cipher padding has been found. This is a countermeasure
3994 against active attacks where the attacker has to distinguish
3995 between bad padding and a MAC verification error. (CVE-2003-0078)
3997 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3998 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3999 Martin Vuagnoux (EPFL, Ilion)]
4001 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4003 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4004 memory from it's contents. This is done with a counter that will
4005 place alternating values in each byte. This can be used to solve
4006 two issues: 1) the removal of calls to memset() by highly optimizing
4007 compilers, and 2) cleansing with other values than 0, since those can
4008 be read through on certain media, for example a swap space on disk.
4011 *) Bugfix: client side session caching did not work with external caching,
4012 because the session->cipher setting was not restored when reloading
4013 from the external cache. This problem was masked, when
4014 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4015 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4018 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4019 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4020 [Zeev Lieber <zeev-l@yahoo.com>]
4022 *) Undo an undocumented change introduced in 0.9.6e which caused
4023 repeated calls to OpenSSL_add_all_ciphers() and
4024 OpenSSL_add_all_digests() to be ignored, even after calling
4028 *) Change the default configuration reader to deal with last line not
4029 being properly terminated.
4032 *) Change X509_NAME_cmp() so it applies the special rules on handling
4033 DN values that are of type PrintableString, as well as RDNs of type
4034 emailAddress where the value has the type ia5String.
4035 [stefank@valicert.com via Richard Levitte]
4037 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4038 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4039 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4040 the bitwise-OR of the two for use by the majority of applications
4041 wanting this behaviour, and update the docs. The documented
4042 behaviour and actual behaviour were inconsistent and had been
4043 changing anyway, so this is more a bug-fix than a behavioural
4045 [Geoff Thorpe, diagnosed by Nadav Har'El]
4047 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4048 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4051 *) Fix initialization code race conditions in
4052 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4053 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4054 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4055 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4056 ssl2_get_cipher_by_char(),
4057 ssl3_get_cipher_by_char().
4058 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4060 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4061 the cached sessions are flushed, as the remove_cb() might use ex_data
4062 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4063 (see [openssl.org #212]).
4064 [Geoff Thorpe, Lutz Jaenicke]
4066 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4067 length, instead of the encoding length to d2i_ASN1_OBJECT.
4070 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4072 *) [In 0.9.6g-engine release:]
4073 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4074 [Lynn Gazis <lgazis@rainbow.com>]
4076 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4078 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4079 and get fix the header length calculation.
4080 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4081 Alon Kantor <alonk@checkpoint.com> (and others),
4084 *) Use proper error handling instead of 'assertions' in buffer
4085 overflow checks added in 0.9.6e. This prevents DoS (the
4086 assertions could call abort()).
4087 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4089 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4091 *) Add various sanity checks to asn1_get_length() to reject
4092 the ASN1 length bytes if they exceed sizeof(long), will appear
4093 negative or the content length exceeds the length of the
4095 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4097 *) Fix cipher selection routines: ciphers without encryption had no flags
4098 for the cipher strength set and where therefore not handled correctly
4099 by the selection routines (PR #130).
4102 *) Fix EVP_dsa_sha macro.
4106 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4107 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4108 that was added in OpenSSL 0.9.6d.
4110 As the countermeasure turned out to be incompatible with some
4111 broken SSL implementations, the new option is part of SSL_OP_ALL.
4112 SSL_OP_ALL is usually employed when compatibility with weird SSL
4113 implementations is desired (e.g. '-bugs' option to 's_client' and
4114 's_server'), so the new option is automatically set in many
4118 *) Changes in security patch:
4120 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4121 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4122 Air Force Materiel Command, USAF, under agreement number
4125 *) Add various sanity checks to asn1_get_length() to reject
4126 the ASN1 length bytes if they exceed sizeof(long), will appear
4127 negative or the content length exceeds the length of the
4128 supplied buffer. (CVE-2002-0659)
4129 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4131 *) Assertions for various potential buffer overflows, not known to
4133 [Ben Laurie (CHATS)]
4135 *) Various temporary buffers to hold ASCII versions of integers were
4136 too small for 64 bit platforms. (CVE-2002-0655)
4137 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4139 *) Remote buffer overflow in SSL3 protocol - an attacker could
4140 supply an oversized session ID to a client. (CVE-2002-0656)
4141 [Ben Laurie (CHATS)]
4143 *) Remote buffer overflow in SSL2 protocol - an attacker could
4144 supply an oversized client master key. (CVE-2002-0656)
4145 [Ben Laurie (CHATS)]
4147 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4149 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4150 encoded as NULL) with id-dsa-with-sha1.
4151 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4153 *) Check various X509_...() return values in apps/req.c.
4154 [Nils Larsch <nla@trustcenter.de>]
4156 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4157 an end-of-file condition would erronously be flagged, when the CRLF
4158 was just at the end of a processed block. The bug was discovered when
4159 processing data through a buffering memory BIO handing the data to a
4160 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4161 <ptsekov@syntrex.com> and Nedelcho Stanev.
4164 *) Implement a countermeasure against a vulnerability recently found
4165 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4166 before application data chunks to avoid the use of known IVs
4167 with data potentially chosen by the attacker.
4170 *) Fix length checks in ssl3_get_client_hello().
4173 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4174 to prevent ssl3_read_internal() from incorrectly assuming that
4175 ssl3_read_bytes() found application data while handshake
4176 processing was enabled when in fact s->s3->in_read_app_data was
4177 merely automatically cleared during the initial handshake.
4178 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4180 *) Fix object definitions for Private and Enterprise: they were not
4181 recognized in their shortname (=lowercase) representation. Extend
4182 obj_dat.pl to issue an error when using undefined keywords instead
4183 of silently ignoring the problem (Svenning Sorensen
4184 <sss@sss.dnsalias.net>).
4187 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4188 generators, i.e. generators other than 2 and 5. (Previously, the
4189 code did not properly initialise the 'add' and 'rem' values to
4190 BN_generate_prime().)
4192 In the new general case, we do not insist that 'generator' is
4193 actually a primitive root: This requirement is rather pointless;
4194 a generator of the order-q subgroup is just as good, if not
4198 *) Map new X509 verification errors to alerts. Discovered and submitted by
4199 Tom Wu <tom@arcot.com>.
4202 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4203 returning non-zero before the data has been completely received
4204 when using non-blocking I/O.
4205 [Bodo Moeller; problem pointed out by John Hughes]
4207 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4208 [Ben Laurie, Lutz Jaenicke]
4210 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4211 Yoram Zahavi <YoramZ@gilian.com>).
4214 *) Add information about CygWin 1.3 and on, and preserve proper
4215 configuration for the versions before that.
4216 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4218 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4219 check whether we deal with a copy of a session and do not delete from
4220 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4221 <izhar@checkpoint.com>.
4224 *) Do not store session data into the internal session cache, if it
4225 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4226 flag is set). Proposed by Aslam <aslam@funk.com>.
4229 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4233 *) [In 0.9.6d-engine release:]
4234 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4235 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4237 *) Add the configuration target linux-s390x.
4238 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4240 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4241 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4242 variable as an indication that a ClientHello message has been
4243 received. As the flag value will be lost between multiple
4244 invocations of ssl3_accept when using non-blocking I/O, the
4245 function may not be aware that a handshake has actually taken
4246 place, thus preventing a new session from being added to the
4249 To avoid this problem, we now set s->new_session to 2 instead of
4250 using a local variable.
4251 [Lutz Jaenicke, Bodo Moeller]
4253 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4254 if the SSL_R_LENGTH_MISMATCH error is detected.
4255 [Geoff Thorpe, Bodo Moeller]
4257 *) New 'shared_ldflag' column in Configure platform table.
4260 *) Fix EVP_CIPHER_mode macro.
4261 ["Dan S. Camper" <dan@bti.net>]
4263 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4264 type, we must throw them away by setting rr->length to 0.
4265 [D P Chang <dpc@qualys.com>]
4267 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4269 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4270 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4271 worked incorrectly for those cases where range = 10..._2 and
4272 3*range is two bits longer than range.)
4275 *) Only add signing time to PKCS7 structures if it is not already
4279 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4280 OBJ_ld_ce should be OBJ_id_ce.
4281 Also some ip-pda OIDs in crypto/objects/objects.txt were
4282 incorrect (cf. RFC 3039).
4283 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4285 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4286 returns early because it has nothing to do.
4287 [Andy Schneider <andy.schneider@bjss.co.uk>]
4289 *) [In 0.9.6c-engine release:]
4290 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4291 [Andy Schneider <andy.schneider@bjss.co.uk>]
4293 *) [In 0.9.6c-engine release:]
4294 Add support for Cryptographic Appliance's keyserver technology.
4295 (Use engine 'keyclient')
4296 [Cryptographic Appliances and Geoff Thorpe]
4298 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4299 is called via tools/c89.sh because arguments have to be
4300 rearranged (all '-L' options must appear before the first object
4302 [Richard Shapiro <rshapiro@abinitio.com>]
4304 *) [In 0.9.6c-engine release:]
4305 Add support for Broadcom crypto accelerator cards, backported
4307 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4309 *) [In 0.9.6c-engine release:]
4310 Add support for SureWare crypto accelerator cards from
4311 Baltimore Technologies. (Use engine 'sureware')
4312 [Baltimore Technologies and Mark Cox]
4314 *) [In 0.9.6c-engine release:]
4315 Add support for crypto accelerator cards from Accelerated
4316 Encryption Processing, www.aep.ie. (Use engine 'aep')
4317 [AEP Inc. and Mark Cox]
4319 *) Add a configuration entry for gcc on UnixWare.
4320 [Gary Benson <gbenson@redhat.com>]
4322 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4323 messages are stored in a single piece (fixed-length part and
4324 variable-length part combined) and fix various bugs found on the way.
4327 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4328 instead. BIO_gethostbyname() does not know what timeouts are
4329 appropriate, so entries would stay in cache even when they have
4331 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4333 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4334 faced with a pathologically small ClientHello fragment that does
4335 not contain client_version: Instead of aborting with an error,
4336 simply choose the highest available protocol version (i.e.,
4337 TLS 1.0 unless it is disabled). In practice, ClientHello
4338 messages are never sent like this, but this change gives us
4339 strictly correct behaviour at least for TLS.
4342 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4343 never resets s->method to s->ctx->method when called from within
4344 one of the SSL handshake functions.
4345 [Bodo Moeller; problem pointed out by Niko Baric]
4347 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4348 (sent using the client's version number) if client_version is
4349 smaller than the protocol version in use. Also change
4350 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4351 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4352 the client will at least see that alert.
4355 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4359 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4360 client receives HelloRequest while in a handshake.
4361 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4363 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4364 should end in 'break', not 'goto end' which circuments various
4365 cleanups done in state SSL_ST_OK. But session related stuff
4366 must be disabled for SSL_ST_OK in the case that we just sent a
4369 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4370 before just sending a HelloRequest.
4371 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4373 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4374 reveal whether illegal block cipher padding was found or a MAC
4375 verification error occured. (Neither SSLerr() codes nor alerts
4376 are directly visible to potential attackers, but the information
4377 may leak via logfiles.)
4379 Similar changes are not required for the SSL 2.0 implementation
4380 because the number of padding bytes is sent in clear for SSL 2.0,
4381 and the extra bytes are just ignored. However ssl/s2_pkt.c
4382 failed to verify that the purported number of padding bytes is in
4386 *) Add OpenUNIX-8 support including shared libraries
4387 (Boyd Lynn Gerber <gerberb@zenez.com>).
4390 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4391 'wristwatch attack' using huge encoding parameters (cf.
4392 James H. Manger's CRYPTO 2001 paper). Note that the
4393 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4394 encoding parameters and hence was not vulnerable.
4397 *) BN_sqr() bug fix.
4398 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4400 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4401 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4402 followed by modular reduction.
4403 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4405 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4406 equivalent based on BN_pseudo_rand() instead of BN_rand().
4409 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4410 This function was broken, as the check for a new client hello message
4411 to handle SGC did not allow these large messages.
4412 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4415 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4418 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4419 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4422 *) Rework the configuration and shared library support for Tru64 Unix.
4423 The configuration part makes use of modern compiler features and
4424 still retains old compiler behavior for those that run older versions
4425 of the OS. The shared library support part includes a variant that
4426 uses the RPATH feature, and is available through the special
4427 configuration target "alpha-cc-rpath", which will never be selected
4429 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4431 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4432 with the same message size as in ssl3_get_certificate_request().
4433 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4434 messages might inadvertently be reject as too long.
4435 [Petr Lampa <lampa@fee.vutbr.cz>]
4437 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4440 *) Modified SSL library such that the verify_callback that has been set
4441 specificly for an SSL object with SSL_set_verify() is actually being
4442 used. Before the change, a verify_callback set with this function was
4443 ignored and the verify_callback() set in the SSL_CTX at the time of
4444 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4445 to allow the necessary settings.
4448 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4449 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4450 done automatically (in contradiction to the requirements of the C
4451 standard). This made problems when used from OpenSSH.
4454 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4455 dh->length and always used
4457 BN_rand_range(priv_key, dh->p).
4459 BN_rand_range() is not necessary for Diffie-Hellman, and this
4460 specific range makes Diffie-Hellman unnecessarily inefficient if
4461 dh->length (recommended exponent length) is much smaller than the
4462 length of dh->p. We could use BN_rand_range() if the order of
4463 the subgroup was stored in the DH structure, but we only have
4468 BN_rand(priv_key, l, ...)
4470 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4476 RSA_eay_public_encrypt
4477 RSA_eay_private_decrypt
4478 RSA_eay_private_encrypt (signing)
4479 RSA_eay_public_decrypt (signature verification)
4481 (default implementations for RSA_public_encrypt,
4482 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4483 always reject numbers >= n.
4486 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4487 to synchronize access to 'locking_thread'. This is necessary on
4488 systems where access to 'locking_thread' (an 'unsigned long'
4489 variable) is not atomic.
4492 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4493 *before* setting the 'crypto_lock_rand' flag. The previous code had
4494 a race condition if 0 is a valid thread ID.
4495 [Travis Vitek <vitek@roguewave.com>]
4497 *) Add support for shared libraries under Irix.
4498 [Albert Chin-A-Young <china@thewrittenword.com>]
4500 *) Add configuration option to build on Linux on both big-endian and
4502 [Ralf Baechle <ralf@uni-koblenz.de>]
4504 *) Add the possibility to create shared libraries on HP-UX.
4507 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4509 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4510 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4511 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4512 PRNG state recovery was possible based on the output of
4513 one PRNG request appropriately sized to gain knowledge on
4514 'md' followed by enough consecutive 1-byte PRNG requests
4515 to traverse all of 'state'.
4517 1. When updating 'md_local' (the current thread's copy of 'md')
4518 during PRNG output generation, hash all of the previous
4519 'md_local' value, not just the half used for PRNG output.
4521 2. Make the number of bytes from 'state' included into the hash
4522 independent from the number of PRNG bytes requested.
4524 The first measure alone would be sufficient to avoid
4525 Markku-Juhani's attack. (Actually it had never occurred
4526 to me that the half of 'md_local' used for chaining was the
4527 half from which PRNG output bytes were taken -- I had always
4528 assumed that the secret half would be used.) The second
4529 measure makes sure that additional data from 'state' is never
4530 mixed into 'md_local' in small portions; this heuristically
4531 further strengthens the PRNG.
4534 *) Fix crypto/bn/asm/mips3.s.
4537 *) When only the key is given to "enc", the IV is undefined. Print out
4538 an error message in this case.
4541 *) Handle special case when X509_NAME is empty in X509 printing routines.
4544 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4545 positive and less than q.
4548 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4549 used: it isn't thread safe and the add_lock_callback should handle
4551 [Paul Rose <Paul.Rose@bridge.com>]
4553 *) Verify that incoming data obeys the block size in
4554 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4558 [Ulf Möller, Bodo Möller]
4560 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4561 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4562 when fixing the server behaviour for backwards-compatible 'client
4563 hello' messages. (Note that the attack is impractical against
4564 SSL 3.0 and TLS 1.0 anyway because length and version checking
4565 means that the probability of guessing a valid ciphertext is
4566 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4569 Before 0.9.5, the countermeasure (hide the error by generating a
4570 random 'decryption result') did not work properly because
4571 ERR_clear_error() was missing, meaning that SSL_get_error() would
4572 detect the supposedly ignored error.
4574 Both problems are now fixed.
4577 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4578 (previously it was 1024).
4581 *) Fix for compatibility mode trust settings: ignore trust settings
4582 unless some valid trust or reject settings are present.
4585 *) Fix for blowfish EVP: its a variable length cipher.
4588 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4589 parameters in DSA public key structures and return an error in the
4590 DSA routines if parameters are absent.
4593 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4594 in the current directory if neither $RANDFILE nor $HOME was set.
4595 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4596 caused some confusion to Windows users who haven't defined $HOME.
4597 Thus RAND_file_name() is changed again: e_os.h can define a
4598 DEFAULT_HOME, which will be used if $HOME is not set.
4599 For Windows, we use "C:"; on other platforms, we still require
4600 environment variables.
4602 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4603 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4604 having multiple threads call RAND_poll() concurrently.
4607 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4608 combination of a flag and a thread ID variable.
4609 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4610 flag), *other* threads can enter ssleay_add_bytes without obeying
4611 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4612 that they do not hold after the first thread unsets add_do_not_lock).
4615 *) Change bctest again: '-x' expressions are not available in all
4619 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4621 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4622 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4624 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4625 the default extension for executables, if any. Also, make the perl
4626 scripts that use symlink() to test if it really exists and use "cp"
4627 if it doesn't. All this made OpenSSL compilable and installable in
4631 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4632 If SEQUENCE is length is indefinite just set c->slen to the total
4633 amount of data available.
4634 [Steve Henson, reported by shige@FreeBSD.org]
4635 [This change does not apply to 0.9.7.]
4637 *) Change bctest to avoid here-documents inside command substitution
4638 (workaround for FreeBSD /bin/sh bug).
4639 For compatibility with Ultrix, avoid shell functions (introduced
4640 in the bctest version that searches along $PATH).
4643 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4644 with des_encrypt() defined on some operating systems, like Solaris
4648 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4649 On the Importance of Eliminating Errors in Cryptographic
4650 Computations, J. Cryptology 14 (2001) 2, 101-119,
4651 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4654 *) MIPS assembler BIGNUM division bug fix.
4657 *) Disabled incorrect Alpha assembler code.
4660 *) Fix PKCS#7 decode routines so they correctly update the length
4661 after reading an EOC for the EXPLICIT tag.
4663 [This change does not apply to 0.9.7.]
4665 *) Fix bug in PKCS#12 key generation routines. This was triggered
4666 if a 3DES key was generated with a 0 initial byte. Include
4667 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4668 (but broken) behaviour.
4671 *) Enhance bctest to search for a working bc along $PATH and print
4673 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4675 *) Fix memory leaks in err.c: free err_data string if necessary;
4676 don't write to the wrong index in ERR_set_error_data.
4679 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4683 *) Replace rdtsc with _emit statements for VC++ version 5.
4684 [Jeremy Cooper <jeremy@baymoo.org>]
4686 *) Make it possible to reuse SSLv2 sessions.
4689 *) In copy_email() check for >= 0 as a return value for
4690 X509_NAME_get_index_by_NID() since 0 is a valid index.
4691 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4693 *) Avoid coredump with unsupported or invalid public keys by checking if
4694 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4695 PKCS7_verify() fails with non detached data.
4698 *) Don't use getenv in library functions when run as setuid/setgid.
4699 New function OPENSSL_issetugid().
4702 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4703 due to incorrect handling of multi-threading:
4705 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4707 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4709 3. Count how many times MemCheck_off() has been called so that
4710 nested use can be treated correctly. This also avoids
4711 inband-signalling in the previous code (which relied on the
4712 assumption that thread ID 0 is impossible).
4715 *) Add "-rand" option also to s_client and s_server.
4718 *) Fix CPU detection on Irix 6.x.
4719 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4720 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4722 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4725 [This change does not apply to 0.9.7.]
4727 *) Use the cached encoding of an X509_NAME structure rather than
4728 copying it. This is apparently the reason for the libsafe "errors"
4729 but the code is actually correct.
4732 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4733 Bleichenbacher's DSA attack.
4734 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4735 to be set and top=0 forces the highest bit to be set; top=-1 is new
4736 and leaves the highest bit random.
4737 [Ulf Moeller, Bodo Moeller]
4739 *) In the NCONF_...-based implementations for CONF_... queries
4740 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4741 a temporary CONF structure with the data component set to NULL
4742 (which gives segmentation faults in lh_retrieve).
4743 Instead, use NULL for the CONF pointer in CONF_get_string and
4744 CONF_get_number (which may use environment variables) and directly
4745 return NULL from CONF_get_section.
4748 *) Fix potential buffer overrun for EBCDIC.
4751 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4752 keyUsage if basicConstraints absent for a CA.
4755 *) Make SMIME_write_PKCS7() write mail header values with a format that
4756 is more generally accepted (no spaces before the semicolon), since
4757 some programs can't parse those values properly otherwise. Also make
4758 sure BIO's that break lines after each write do not create invalid
4762 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4763 macros previously used would not encode an empty SEQUENCE OF
4764 and break the signature.
4766 [This change does not apply to 0.9.7.]
4768 *) Zero the premaster secret after deriving the master secret in
4772 *) Add some EVP_add_digest_alias registrations (as found in
4773 OpenSSL_add_all_digests()) to SSL_library_init()
4774 aka OpenSSL_add_ssl_algorithms(). This provides improved
4775 compatibility with peers using X.509 certificates
4776 with unconventional AlgorithmIdentifier OIDs.
4779 *) Fix for Irix with NO_ASM.
4780 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4782 *) ./config script fixes.
4783 [Ulf Moeller, Richard Levitte]
4785 *) Fix 'openssl passwd -1'.
4788 *) Change PKCS12_key_gen_asc() so it can cope with non null
4789 terminated strings whose length is passed in the passlen
4790 parameter, for example from PEM callbacks. This was done
4791 by adding an extra length parameter to asc2uni().
4792 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4794 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4795 call failed, free the DSA structure.
4798 *) Fix to uni2asc() to cope with zero length Unicode strings.
4799 These are present in some PKCS#12 files.
4802 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4803 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4804 when writing a 32767 byte record.
4805 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4807 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4808 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4810 (RSA objects have a reference count access to which is protected
4811 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4812 so they are meant to be shared between threads.)
4813 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4814 "Reddie, Steven" <Steven.Reddie@ca.com>]
4816 *) Fix a deadlock in CRYPTO_mem_leaks().
4819 *) Use better test patterns in bntest.
4822 *) rand_win.c fix for Borland C.
4825 *) BN_rshift bugfix for n == 0.
4828 *) Add a 'bctest' script that checks for some known 'bc' bugs
4829 so that 'make test' does not abort just because 'bc' is broken.
4832 *) Store verify_result within SSL_SESSION also for client side to
4833 avoid potential security hole. (Re-used sessions on the client side
4834 always resulted in verify_result==X509_V_OK, not using the original
4835 result of the server certificate verification.)
4838 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4839 SSL3_RT_APPLICATION_DATA, return 0.
4840 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4844 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4845 releases, have been re-implemented by renaming the previous
4846 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4847 and ssl3_read_internal, respectively, and adding 'peek' parameters
4848 to them. The new ssl[23]_{read,peek} functions are calls to
4849 ssl[23]_read_internal with the 'peek' flag set appropriately.
4850 A 'peek' parameter has also been added to ssl3_read_bytes, which
4851 does the actual work for ssl3_read_internal.
4854 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4855 the method-specific "init()" handler. Also clean up ex_data after
4856 calling the method-specific "finish()" handler. Previously, this was
4857 happening the other way round.
4860 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4861 The previous value, 12, was not always sufficient for BN_mod_exp().
4864 *) Make sure that shared libraries get the internal name engine with
4865 the full version number and not just 0. This should mark the
4866 shared libraries as not backward compatible. Of course, this should
4867 be changed again when we can guarantee backward binary compatibility.
4870 *) Fix typo in get_cert_by_subject() in by_dir.c
4871 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4873 *) Rework the system to generate shared libraries:
4875 - Make note of the expected extension for the shared libraries and
4876 if there is a need for symbolic links from for example libcrypto.so.0
4877 to libcrypto.so.0.9.7. There is extended info in Configure for
4880 - Make as few rebuilds of the shared libraries as possible.
4882 - Still avoid linking the OpenSSL programs with the shared libraries.
4884 - When installing, install the shared libraries separately from the
4888 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4890 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4891 and not in SSL_clear because the latter is also used by the
4892 accept/connect functions; previously, the settings made by
4893 SSL_set_read_ahead would be lost during the handshake.
4894 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4896 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4897 Previously, it would create entries for disableed algorithms no
4901 *) Added several new manual pages for SSL_* function.
4904 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4906 *) In ssl23_get_client_hello, generate an error message when faced
4907 with an initial SSL 3.0/TLS record that is too small to contain the
4908 first two bytes of the ClientHello message, i.e. client_version.
4909 (Note that this is a pathologic case that probably has never happened
4910 in real life.) The previous approach was to use the version number
4911 from the record header as a substitute; but our protocol choice
4912 should not depend on that one because it is not authenticated
4913 by the Finished messages.
4916 *) More robust randomness gathering functions for Windows.
4917 [Jeffrey Altman <jaltman@columbia.edu>]
4919 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4920 not set then we don't setup the error code for issuer check errors
4921 to avoid possibly overwriting other errors which the callback does
4922 handle. If an application does set the flag then we assume it knows
4923 what it is doing and can handle the new informational codes
4927 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4928 a general "ANY" type, as such it should be able to decode anything
4929 including tagged types. However it didn't check the class so it would
4930 wrongly interpret tagged types in the same way as their universal
4931 counterpart and unknown types were just rejected. Changed so that the
4932 tagged and unknown types are handled in the same way as a SEQUENCE:
4933 that is the encoding is stored intact. There is also a new type
4934 "V_ASN1_OTHER" which is used when the class is not universal, in this
4935 case we have no idea what the actual type is so we just lump them all
4939 *) On VMS, stdout may very well lead to a file that is written to
4940 in a record-oriented fashion. That means that every write() will
4941 write a separate record, which will be read separately by the
4942 programs trying to read from it. This can be very confusing.
4944 The solution is to put a BIO filter in the way that will buffer
4945 text until a linefeed is reached, and then write everything a
4946 line at a time, so every record written will be an actual line,
4947 not chunks of lines and not (usually doesn't happen, but I've
4948 seen it once) several lines in one record. BIO_f_linebuffer() is
4951 Currently, it's a VMS-only method, because that's where it has
4952 been tested well enough.
4955 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4956 it can return incorrect results.
4957 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4958 but it was in 0.9.6-beta[12].)
4961 *) Disable the check for content being present when verifying detached
4962 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4963 include zero length content when signing messages.
4966 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4967 BIO_ctrl (for BIO pairs).
4970 *) Add DSO method for VMS.
4973 *) Bug fix: Montgomery multiplication could produce results with the
4977 *) Add RPM specification openssl.spec and modify it to build three
4978 packages. The default package contains applications, application
4979 documentation and run-time libraries. The devel package contains
4980 include files, static libraries and function documentation. The
4981 doc package contains the contents of the doc directory. The original
4982 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4985 *) Add a large number of documentation files for many SSL routines.
4986 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4988 *) Add a configuration entry for Sony News 4.
4989 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4991 *) Don't set the two most significant bits to one when generating a
4992 random number < q in the DSA library.
4995 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4996 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4997 the underlying transport is blocking) if a handshake took place.
4998 (The default behaviour is needed by applications such as s_client
4999 and s_server that use select() to determine when to use SSL_read;
5000 but for applications that know in advance when to expect data, it
5001 just makes things more complicated.)
5004 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5008 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5009 work better on such systems.
5010 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5012 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5013 Update PKCS12_parse() so it copies the friendlyName and the
5014 keyid to the certificates aux info.
5017 *) Fix bug in PKCS7_verify() which caused an infinite loop
5018 if there was more than one signature.
5019 [Sven Uszpelkat <su@celocom.de>]
5021 *) Major change in util/mkdef.pl to include extra information
5022 about each symbol, as well as presentig variables as well
5023 as functions. This change means that there's n more need
5024 to rebuild the .num files when some algorithms are excluded.
5027 *) Allow the verify time to be set by an application,
5028 rather than always using the current time.
5031 *) Phase 2 verify code reorganisation. The certificate
5032 verify code now looks up an issuer certificate by a
5033 number of criteria: subject name, authority key id
5034 and key usage. It also verifies self signed certificates
5035 by the same criteria. The main comparison function is
5036 X509_check_issued() which performs these checks.
5038 Lot of changes were necessary in order to support this
5039 without completely rewriting the lookup code.
5041 Authority and subject key identifier are now cached.
5043 The LHASH 'certs' is X509_STORE has now been replaced
5044 by a STACK_OF(X509_OBJECT). This is mainly because an
5045 LHASH can't store or retrieve multiple objects with
5046 the same hash value.
5048 As a result various functions (which were all internal
5049 use only) have changed to handle the new X509_STORE
5050 structure. This will break anything that messed round
5051 with X509_STORE internally.
5053 The functions X509_STORE_add_cert() now checks for an
5054 exact match, rather than just subject name.
5056 The X509_STORE API doesn't directly support the retrieval
5057 of multiple certificates matching a given criteria, however
5058 this can be worked round by performing a lookup first
5059 (which will fill the cache with candidate certificates)
5060 and then examining the cache for matches. This is probably
5061 the best we can do without throwing out X509_LOOKUP
5062 entirely (maybe later...).
5064 The X509_VERIFY_CTX structure has been enhanced considerably.
5066 All certificate lookup operations now go via a get_issuer()
5067 callback. Although this currently uses an X509_STORE it
5068 can be replaced by custom lookups. This is a simple way
5069 to bypass the X509_STORE hackery necessary to make this
5070 work and makes it possible to use more efficient techniques
5071 in future. A very simple version which uses a simple
5072 STACK for its trusted certificate store is also provided
5073 using X509_STORE_CTX_trusted_stack().
5075 The verify_cb() and verify() callbacks now have equivalents
5076 in the X509_STORE_CTX structure.
5078 X509_STORE_CTX also has a 'flags' field which can be used
5079 to customise the verify behaviour.
5082 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5083 excludes S/MIME capabilities.
5086 *) When a certificate request is read in keep a copy of the
5087 original encoding of the signed data and use it when outputing
5088 again. Signatures then use the original encoding rather than
5089 a decoded, encoded version which may cause problems if the
5090 request is improperly encoded.
5093 *) For consistency with other BIO_puts implementations, call
5094 buffer_write(b, ...) directly in buffer_puts instead of calling
5097 In BIO_puts, increment b->num_write as in BIO_write.
5098 [Peter.Sylvester@EdelWeb.fr]
5100 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5101 BN_zero, we may not return a BIGNUM with an array consisting of
5105 *) Avoid calling abort() from within the library when problems are
5106 detected, except if preprocessor symbols have been defined
5107 (such as REF_CHECK, BN_DEBUG etc.).
5110 *) New openssl application 'rsautl'. This utility can be
5111 used for low level RSA operations. DER public key
5112 BIO/fp routines also added.
5115 *) New Configure entry and patches for compiling on QNX 4.
5116 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5118 *) A demo state-machine implementation was sponsored by
5119 Nuron (http://www.nuron.com/) and is now available in
5120 demos/state_machine.
5123 *) New options added to the 'dgst' utility for signature
5124 generation and verification.
5127 *) Unrecognized PKCS#7 content types are now handled via a
5128 catch all ASN1_TYPE structure. This allows unsupported
5129 types to be stored as a "blob" and an application can
5130 encode and decode it manually.
5133 *) Fix various signed/unsigned issues to make a_strex.c
5135 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5137 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5138 length if passed a buffer. ASN1_INTEGER_to_BN failed
5139 if passed a NULL BN and its argument was negative.
5140 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5142 *) Modification to PKCS#7 encoding routines to output definite
5143 length encoding. Since currently the whole structures are in
5144 memory there's not real point in using indefinite length
5145 constructed encoding. However if OpenSSL is compiled with
5146 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5149 *) Added BIO_vprintf() and BIO_vsnprintf().
5152 *) Added more prefixes to parse for in the the strings written
5153 through a logging bio, to cover all the levels that are available
5154 through syslog. The prefixes are now:
5156 PANIC, EMERG, EMR => LOG_EMERG
5157 ALERT, ALR => LOG_ALERT
5158 CRIT, CRI => LOG_CRIT
5159 ERROR, ERR => LOG_ERR
5160 WARNING, WARN, WAR => LOG_WARNING
5161 NOTICE, NOTE, NOT => LOG_NOTICE
5162 INFO, INF => LOG_INFO
5163 DEBUG, DBG => LOG_DEBUG
5165 and as before, if none of those prefixes are present at the
5166 beginning of the string, LOG_ERR is chosen.
5168 On Win32, the LOG_* levels are mapped according to this:
5170 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5171 LOG_WARNING => EVENTLOG_WARNING_TYPE
5172 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5176 *) Made it possible to reconfigure with just the configuration
5177 argument "reconf" or "reconfigure". The command line arguments
5178 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5179 and are retrieved from there when reconfiguring.
5183 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5185 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5188 *) The obj_dat.pl script was messing up the sorting of object
5189 names. The reason was that it compared the quoted version
5190 of strings as a result "OCSP" > "OCSP Signing" because
5191 " > SPACE. Changed script to store unquoted versions of
5192 names and add quotes on output. It was also omitting some
5193 names from the lookup table if they were given a default
5194 value (that is if SN is missing it is given the same
5195 value as LN and vice versa), these are now added on the
5196 grounds that if an object has a name we should be able to
5197 look it up. Finally added warning output when duplicate
5198 short or long names are found.
5201 *) Changes needed for Tandem NSK.
5202 [Scott Uroff <scott@xypro.com>]
5204 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5205 RSA_padding_check_SSLv23(), special padding was never detected
5206 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5207 version rollback attacks was not effective.
5209 In s23_clnt.c, don't use special rollback-attack detection padding
5210 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5211 client; similarly, in s23_srvr.c, don't do the rollback check if
5212 SSL 2.0 is the only protocol enabled in the server.
5215 *) Make it possible to get hexdumps of unprintable data with 'openssl
5216 asn1parse'. By implication, the functions ASN1_parse_dump() and
5217 BIO_dump_indent() are added.
5220 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5221 these print out strings and name structures based on various
5222 flags including RFC2253 support and proper handling of
5223 multibyte characters. Added options to the 'x509' utility
5224 to allow the various flags to be set.
5227 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5228 Also change the functions X509_cmp_current_time() and
5229 X509_gmtime_adj() work with an ASN1_TIME structure,
5230 this will enable certificates using GeneralizedTime in validity
5231 dates to be checked.
5234 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5235 negative public key encodings) on by default,
5236 NO_NEG_PUBKEY_BUG can be set to disable it.
5239 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5240 content octets. An i2c_ASN1_OBJECT is unnecessary because
5241 the encoding can be trivially obtained from the structure.
5244 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5245 not read locks (CRYPTO_r_[un]lock).
5248 *) A first attempt at creating official support for shared
5249 libraries through configuration. I've kept it so the
5250 default is static libraries only, and the OpenSSL programs
5251 are always statically linked for now, but there are
5252 preparations for dynamic linking in place.
5253 This has been tested on Linux and Tru64.
5256 *) Randomness polling function for Win9x, as described in:
5257 Peter Gutmann, Software Generation of Practically Strong
5261 *) Fix so PRNG is seeded in req if using an already existing
5265 *) New options to smime application. -inform and -outform
5266 allow alternative formats for the S/MIME message including
5267 PEM and DER. The -content option allows the content to be
5268 specified separately. This should allow things like Netscape
5269 form signing output easier to verify.
5272 *) Fix the ASN1 encoding of tags using the 'long form'.
5275 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5276 STRING types. These convert content octets to and from the
5277 underlying type. The actual tag and length octets are
5278 already assumed to have been read in and checked. These
5279 are needed because all other string types have virtually
5280 identical handling apart from the tag. By having versions
5281 of the ASN1 functions that just operate on content octets
5282 IMPLICIT tagging can be handled properly. It also allows
5283 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5284 and ASN1_INTEGER are identical apart from the tag.
5287 *) Change the handling of OID objects as follows:
5289 - New object identifiers are inserted in objects.txt, following
5290 the syntax given in objects.README.
5291 - objects.pl is used to process obj_mac.num and create a new
5293 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5296 This is currently kind of a hack, and the perl code in objects.pl
5297 isn't very elegant, but it works as I intended. The simplest way
5298 to check that it worked correctly is to look in obj_dat.h and
5299 check the array nid_objs and make sure the objects haven't moved
5300 around (this is important!). Additions are OK, as well as
5301 consistent name changes.
5304 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5307 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5308 The given file adds to whatever has already been seeded into the
5309 random pool through the RANDFILE configuration file option or
5310 environment variable, or the default random state file.
5313 *) mkstack.pl now sorts each macro group into lexical order.
5314 Previously the output order depended on the order the files
5315 appeared in the directory, resulting in needless rewriting
5319 *) Patches to make OpenSSL compile under Win32 again. Mostly
5320 work arounds for the VC++ problem that it treats func() as
5321 func(void). Also stripped out the parts of mkdef.pl that
5322 added extra typesafe functions: these no longer exist.
5325 *) Reorganisation of the stack code. The macros are now all
5326 collected in safestack.h . Each macro is defined in terms of
5327 a "stack macro" of the form SKM_<name>(type, a, b). The
5328 DEBUG_SAFESTACK is now handled in terms of function casts,
5329 this has the advantage of retaining type safety without the
5330 use of additional functions. If DEBUG_SAFESTACK is not defined
5331 then the non typesafe macros are used instead. Also modified the
5332 mkstack.pl script to handle the new form. Needs testing to see
5333 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5334 the default if no major problems. Similar behaviour for ASN1_SET_OF
5335 and PKCS12_STACK_OF.
5338 *) When some versions of IIS use the 'NET' form of private key the
5339 key derivation algorithm is different. Normally MD5(password) is
5340 used as a 128 bit RC4 key. In the modified case
5341 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5342 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5343 as the old Netscape_RSA functions except they have an additional
5344 'sgckey' parameter which uses the modified algorithm. Also added
5345 an -sgckey command line option to the rsa utility. Thanks to
5346 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5347 algorithm to openssl-dev.
5350 *) The evp_local.h macros were using 'c.##kname' which resulted in
5351 invalid expansion on some systems (SCO 5.0.5 for example).
5352 Corrected to 'c.kname'.
5353 [Phillip Porch <root@theporch.com>]
5355 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5356 a STACK of email addresses from a certificate or request, these look
5357 in the subject name and the subject alternative name extensions and
5358 omit any duplicate addresses.
5361 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5362 This makes DSA verification about 2 % faster.
5365 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5366 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5367 plus overhead for 1024 bit moduli).
5368 This makes exponentiations about 0.5 % faster for 1024 bit
5369 exponents (as measured by "openssl speed rsa2048").
5372 *) Rename memory handling macros to avoid conflicts with other
5374 Malloc => OPENSSL_malloc
5375 Malloc_locked => OPENSSL_malloc_locked
5376 Realloc => OPENSSL_realloc
5377 Free => OPENSSL_free
5380 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5381 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5384 *) CygWin32 support.
5385 [John Jarvie <jjarvie@newsguy.com>]
5387 *) The type-safe stack code has been rejigged. It is now only compiled
5388 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5389 by default all type-specific stack functions are "#define"d back to
5390 standard stack functions. This results in more streamlined output
5391 but retains the type-safety checking possibilities of the original
5395 *) The STACK code has been cleaned up, and certain type declarations
5396 that didn't make a lot of sense have been brought in line. This has
5397 also involved a cleanup of sorts in safestack.h to more correctly
5398 map type-safe stack functions onto their plain stack counterparts.
5399 This work has also resulted in a variety of "const"ifications of
5400 lots of the code, especially "_cmp" operations which should normally
5401 be prototyped with "const" parameters anyway.
5404 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5405 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5406 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5407 where all of 'md' is used each time the PRNG is used, but 'state'
5408 is used only indexed by a cyclic counter. As entropy may not be
5409 well distributed from the beginning, 'md' is important as a
5410 chaining variable. However, the output function chains only half
5411 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5412 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5413 in all of 'state' being rewritten, with the new values depending
5414 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5417 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5418 the handshake is continued after ssl_verify_cert_chain();
5419 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5420 can lead to 'unexplainable' connection aborts later.
5421 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5423 *) Major EVP API cipher revision.
5424 Add hooks for extra EVP features. This allows various cipher
5425 parameters to be set in the EVP interface. Support added for variable
5426 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5427 setting of RC2 and RC5 parameters.
5429 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5432 Remove lots of duplicated code from the EVP library. For example *every*
5433 cipher init() function handles the 'iv' in the same way according to the
5434 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5435 for CFB and OFB modes they zero ctx->num.
5437 New functionality allows removal of S/MIME code RC2 hack.
5439 Most of the routines have the same form and so can be declared in terms
5442 By shifting this to the top level EVP_CipherInit() it can be removed from
5443 all individual ciphers. If the cipher wants to handle IVs or keys
5444 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5447 Change lots of functions like EVP_EncryptUpdate() to now return a
5448 value: although software versions of the algorithms cannot fail
5449 any installed hardware versions can.
5452 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5453 this option is set, tolerate broken clients that send the negotiated
5454 protocol version number instead of the requested protocol version
5458 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5459 i.e. non-zero for export ciphersuites, zero otherwise.
5460 Previous versions had this flag inverted, inconsistent with
5461 rsa_tmp_cb (..._TMP_RSA_CB).
5462 [Bodo Moeller; problem reported by Amit Chopra]
5464 *) Add missing DSA library text string. Work around for some IIS
5465 key files with invalid SEQUENCE encoding.
5468 *) Add a document (doc/standards.txt) that list all kinds of standards
5469 and so on that are implemented in OpenSSL.
5472 *) Enhance c_rehash script. Old version would mishandle certificates
5473 with the same subject name hash and wouldn't handle CRLs at all.
5474 Added -fingerprint option to crl utility, to support new c_rehash
5478 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5481 *) Fix for SSL server purpose checking. Server checking was
5482 rejecting certificates which had extended key usage present
5483 but no ssl client purpose.
5484 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5486 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5487 is a little unclear about how a blank password is handled.
5488 Since the password in encoded as a BMPString with terminating
5489 double NULL a zero length password would end up as just the
5490 double NULL. However no password at all is different and is
5491 handled differently in the PKCS#12 key generation code. NS
5492 treats a blank password as zero length. MSIE treats it as no
5493 password on export: but it will try both on import. We now do
5494 the same: PKCS12_parse() tries zero length and no password if
5495 the password is set to "" or NULL (NULL is now a valid password:
5496 it wasn't before) as does the pkcs12 application.
5499 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5500 perror when PEM_read_bio_X509_REQ fails, the error message must
5501 be obtained from the error queue.
5504 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5505 it in ERR_remove_state if appropriate, and change ERR_get_state
5506 accordingly to avoid race conditions (this is necessary because
5507 thread_hash is no longer constant once set).
5510 *) Bugfix for linux-elf makefile.one.
5513 *) RSA_get_default_method() will now cause a default
5514 RSA_METHOD to be chosen if one doesn't exist already.
5515 Previously this was only set during a call to RSA_new()
5516 or RSA_new_method(NULL) meaning it was possible for
5517 RSA_get_default_method() to return NULL.
5520 *) Added native name translation to the existing DSO code
5521 that will convert (if the flag to do so is set) filenames
5522 that are sufficiently small and have no path information
5523 into a canonical native form. Eg. "blah" converted to
5524 "libblah.so" or "blah.dll" etc.
5527 *) New function ERR_error_string_n(e, buf, len) which is like
5528 ERR_error_string(e, buf), but writes at most 'len' bytes
5529 including the 0 terminator. For ERR_error_string_n, 'buf'
5531 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5533 *) CONF library reworked to become more general. A new CONF
5534 configuration file reader "class" is implemented as well as a
5535 new functions (NCONF_*, for "New CONF") to handle it. The now
5536 old CONF_* functions are still there, but are reimplemented to
5537 work in terms of the new functions. Also, a set of functions
5538 to handle the internal storage of the configuration data is
5539 provided to make it easier to write new configuration file
5540 reader "classes" (I can definitely see something reading a
5541 configuration file in XML format, for example), called _CONF_*,
5542 or "the configuration storage API"...
5544 The new configuration file reading functions are:
5546 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5547 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5549 NCONF_default, NCONF_WIN32
5551 NCONF_dump_fp, NCONF_dump_bio
5553 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5554 NCONF_new creates a new CONF object. This works in the same way
5555 as other interfaces in OpenSSL, like the BIO interface.
5556 NCONF_dump_* dump the internal storage of the configuration file,
5557 which is useful for debugging. All other functions take the same
5558 arguments as the old CONF_* functions wth the exception of the
5559 first that must be a `CONF *' instead of a `LHASH *'.
5561 To make it easer to use the new classes with the old CONF_* functions,
5562 the function CONF_set_default_method is provided.
5565 *) Add '-tls1' option to 'openssl ciphers', which was already
5566 mentioned in the documentation but had not been implemented.
5567 (This option is not yet really useful because even the additional
5568 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5571 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5572 OpenSSL-based applications) load shared libraries and bind to
5573 them in a portable way.
5574 [Geoff Thorpe, with contributions from Richard Levitte]
5576 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5578 *) Make sure _lrotl and _lrotr are only used with MSVC.
5580 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5581 (the default implementation of RAND_status).
5583 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5584 to '-clrext' (= clear extensions), as intended and documented.
5585 [Bodo Moeller; inconsistency pointed out by Michael Attili
5588 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5589 was larger than the MD block size.
5590 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5592 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5593 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5594 using the passed key: if the passed key was a private key the result
5595 of X509_print(), for example, would be to print out all the private key
5599 *) des_quad_cksum() byte order bug fix.
5600 [Ulf Möller, using the problem description in krb4-0.9.7, where
5601 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5603 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5605 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5607 *) For easily testing in shell scripts whether some command
5608 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5609 returns with exit code 0 iff no command of the given name is available.
5610 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5611 the output goes to stdout and nothing is printed to stderr.
5612 Additional arguments are always ignored.
5614 Since for each cipher there is a command of the same name,
5615 the 'no-cipher' compilation switches can be tested this way.
5617 ('openssl no-XXX' is not able to detect pseudo-commands such
5618 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5621 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5624 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5625 is set; it will be thrown away anyway because each handshake creates
5627 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5628 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5629 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5630 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5633 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5634 'Q' and 'R' lose their special meanings (quit/renegotiate).
5635 This is part of what -quiet does; unlike -quiet, -ign_eof
5636 does not suppress any output.
5639 *) Add compatibility options to the purpose and trust code. The
5640 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5641 accepts a certificate or CA, this was the previous behaviour,
5642 with all the associated security issues.
5644 X509_TRUST_COMPAT is the old trust behaviour: only and
5645 automatically trust self signed roots in certificate store. A
5646 new trust setting X509_TRUST_DEFAULT is used to specify that
5647 a purpose has no associated trust setting and it should instead
5648 use the value in the default purpose.
5651 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5652 and fix a memory leak.
5655 *) In util/mkerr.pl (which implements 'make errors'), preserve
5656 reason strings from the previous version of the .c file, as
5657 the default to have only downcase letters (and digits) in
5658 automatically generated reasons codes is not always appropriate.
5661 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5662 using strerror. Previously, ERR_reason_error_string() returned
5663 library names as reason strings for SYSerr; but SYSerr is a special
5664 case where small numbers are errno values, not library numbers.
5667 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5668 converts DSA parameters into DH parameters. (When creating parameters,
5669 DSA_generate_parameters is used.)
5672 *) Include 'length' (recommended exponent length) in C code generated
5673 by 'openssl dhparam -C'.
5676 *) The second argument to set_label in perlasm was already being used
5677 so couldn't be used as a "file scope" flag. Moved to third argument
5681 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5682 instead of RAND_bytes for encryption IVs and salts.
5685 *) Include RAND_status() into RAND_METHOD instead of implementing
5686 it only for md_rand.c Otherwise replacing the PRNG by calling
5687 RAND_set_rand_method would be impossible.
5690 *) Don't let DSA_generate_key() enter an infinite loop if the random
5691 number generation fails.
5694 *) New 'rand' application for creating pseudo-random output.
5697 *) Added configuration support for Linux/IA64
5698 [Rolf Haberrecker <rolf@suse.de>]
5700 *) Assembler module support for Mingw32.
5703 *) Shared library support for HPUX (in shlib/).
5704 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5706 *) Shared library support for Solaris gcc.
5707 [Lutz Behnke <behnke@trustcenter.de>]
5709 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5711 *) PKCS7_encrypt() was adding text MIME headers twice because they
5712 were added manually and by SMIME_crlf_copy().
5715 *) In bntest.c don't call BN_rand with zero bits argument.
5716 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5718 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5719 case was implemented. This caused BN_div_recp() to fail occasionally.
5722 *) Add an optional second argument to the set_label() in the perl
5723 assembly language builder. If this argument exists and is set
5724 to 1 it signals that the assembler should use a symbol whose
5725 scope is the entire file, not just the current function. This
5726 is needed with MASM which uses the format label:: for this scope.
5727 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5729 *) Change the ASN1 types so they are typedefs by default. Before
5730 almost all types were #define'd to ASN1_STRING which was causing
5731 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5735 *) Change names of new functions to the new get1/get0 naming
5736 convention: After 'get1', the caller owns a reference count
5737 and has to call ..._free; 'get0' returns a pointer to some
5738 data structure without incrementing reference counters.
5739 (Some of the existing 'get' functions increment a reference
5740 counter, some don't.)
5741 Similarly, 'set1' and 'add1' functions increase reference
5742 counters or duplicate objects.
5745 *) Allow for the possibility of temp RSA key generation failure:
5746 the code used to assume it always worked and crashed on failure.
5749 *) Fix potential buffer overrun problem in BIO_printf().
5750 [Ulf Möller, using public domain code by Patrick Powell; problem
5751 pointed out by David Sacerdote <das33@cornell.edu>]
5753 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5754 RAND_egd() and RAND_status(). In the command line application,
5755 the EGD socket can be specified like a seed file using RANDFILE
5759 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5760 Some CAs (e.g. Verisign) distribute certificates in this form.
5763 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5764 list to exclude them. This means that no special compilation option
5765 is needed to use anonymous DH: it just needs to be included in the
5769 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5770 EVP_MD_type. The old functionality is available in a new macro called
5771 EVP_MD_md(). Change code that uses it and update docs.
5774 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5775 where the 'void *' argument is replaced by a function pointer argument.
5776 Previously 'void *' was abused to point to functions, which works on
5777 many platforms, but is not correct. As these functions are usually
5778 called by macros defined in OpenSSL header files, most source code
5779 should work without changes.
5782 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5783 sections with information on -D... compiler switches used for
5784 compiling the library so that applications can see them. To enable
5785 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5786 must be defined. E.g.,
5787 #define OPENSSL_ALGORITHM_DEFINES
5788 #include <openssl/opensslconf.h>
5789 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5790 [Richard Levitte, Ulf and Bodo Möller]
5792 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5796 *) Change the 'other' type in certificate aux info to a STACK_OF
5797 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5798 the required ASN1 format: arbitrary types determined by an OID.
5801 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5802 argument to 'req'. This is not because the function is newer or
5803 better than others it just uses the work 'NEW' in the certificate
5804 request header lines. Some software needs this.
5807 *) Reorganise password command line arguments: now passwords can be
5808 obtained from various sources. Delete the PEM_cb function and make
5809 it the default behaviour: i.e. if the callback is NULL and the
5810 usrdata argument is not NULL interpret it as a null terminated pass
5811 phrase. If usrdata and the callback are NULL then the pass phrase
5812 is prompted for as usual.
5815 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5816 the support is automatically enabled. The resulting binaries will
5817 autodetect the card and use it if present.
5818 [Ben Laurie and Compaq Inc.]
5820 *) Work around for Netscape hang bug. This sends certificate request
5821 and server done in one record. Since this is perfectly legal in the
5822 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5823 the bugs/SSLv3 entry for more info.
5826 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5829 *) Add -rand argument to smime and pkcs12 applications and read/write
5833 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5836 *) Add command line password options to the remaining applications.
5839 *) Bug fix for BN_div_recp() for numerators with an even number of
5843 *) More tests in bntest.c, and changed test_bn output.
5846 *) ./config recognizes MacOS X now.
5849 *) Bug fix for BN_div() when the first words of num and divsor are
5850 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5853 *) Add support for various broken PKCS#8 formats, and command line
5854 options to produce them.
5857 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5858 get temporary BIGNUMs from a BN_CTX.
5861 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5865 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5866 include a #define from the old name to the new. The original intent
5867 was that statically linked binaries could for example just call
5868 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5869 link with digests. This never worked becayse SSLeay_add_all_digests()
5870 and SSLeay_add_all_ciphers() were in the same source file so calling
5871 one would link with the other. They are now in separate source files.
5874 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5877 *) Use a less unusual form of the Miller-Rabin primality test (it used
5878 a binary algorithm for exponentiation integrated into the Miller-Rabin
5879 loop, our standard modexp algorithms are faster).
5882 *) Support for the EBCDIC character set completed.
5883 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5885 *) Source code cleanups: use const where appropriate, eliminate casts,
5886 use void * instead of char * in lhash.
5889 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5890 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5891 this the server could overwrite ephemeral keys that the client
5895 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5896 using 50 iterations of the Rabin-Miller test.
5898 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5899 iterations of the Rabin-Miller test as required by the appendix
5900 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5901 As BN_is_prime_fasttest includes trial division, DSA parameter
5902 generation becomes much faster.
5904 This implies a change for the callback functions in DSA_is_prime
5905 and DSA_generate_parameters: The callback function is called once
5906 for each positive witness in the Rabin-Miller test, not just
5907 occasionally in the inner loop; and the parameters to the
5908 callback function now provide an iteration count for the outer
5909 loop rather than for the current invocation of the inner loop.
5910 DSA_generate_parameters additionally can call the callback
5911 function with an 'iteration count' of -1, meaning that a
5912 candidate has passed the trial division test (when q is generated
5913 from an application-provided seed, trial division is skipped).
5916 *) New function BN_is_prime_fasttest that optionally does trial
5917 division before starting the Rabin-Miller test and has
5918 an additional BN_CTX * argument (whereas BN_is_prime always
5919 has to allocate at least one BN_CTX).
5920 'callback(1, -1, cb_arg)' is called when a number has passed the
5921 trial division stage.
5924 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5928 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5931 *) New function BN_pseudo_rand().
5934 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5935 bignum version of BN_from_montgomery() with the working code from
5936 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5940 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5941 made it impossible to use the same SSL_SESSION data structure in
5942 SSL2 clients in multiple threads.
5945 *) The return value of RAND_load_file() no longer counts bytes obtained
5946 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5947 to seed the PRNG (previously an explicit byte count was required).
5948 [Ulf Möller, Bodo Möller]
5950 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5951 used (char *) instead of (void *) and had casts all over the place.
5954 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5957 *) Retain source code compatibility for BN_prime_checks macro:
5958 BN_is_prime(..., BN_prime_checks, ...) now uses
5959 BN_prime_checks_for_size to determine the appropriate number of
5960 Rabin-Miller iterations.
5963 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5964 DH_CHECK_P_NOT_SAFE_PRIME.
5965 (Check if this is true? OpenPGP calls them "strong".)
5968 *) Merge the functionality of "dh" and "gendh" programs into a new program
5969 "dhparam". The old programs are retained for now but will handle DH keys
5970 (instead of parameters) in future.
5973 *) Make the ciphers, s_server and s_client programs check the return values
5974 when a new cipher list is set.
5977 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5978 ciphers. Before when the 56bit ciphers were enabled the sorting was
5981 The syntax for the cipher sorting has been extended to support sorting by
5982 cipher-strength (using the strength_bits hard coded in the tables).
5983 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5985 Fix a bug in the cipher-command parser: when supplying a cipher command
5986 string with an "undefined" symbol (neither command nor alphanumeric
5987 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5988 an error is flagged.
5990 Due to the strength-sorting extension, the code of the
5991 ssl_create_cipher_list() function was completely rearranged. I hope that
5992 the readability was also increased :-)
5993 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5995 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5996 for the first serial number and places 2 in the serial number file. This
5997 avoids problems when the root CA is created with serial number zero and
5998 the first user certificate has the same issuer name and serial number
6002 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6003 the new code. Add documentation for this stuff.
6006 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6007 X509_*() to X509at_*() on the grounds that they don't handle X509
6008 structures and behave in an analagous way to the X509v3 functions:
6009 they shouldn't be called directly but wrapper functions should be used
6012 So we also now have some wrapper functions that call the X509at functions
6013 when passed certificate requests. (TO DO: similar things can be done with
6014 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6015 things. Some of these need some d2i or i2d and print functionality
6016 because they handle more complex structures.)
6019 *) Add missing #ifndefs that caused missing symbols when building libssl
6020 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6021 NO_RSA in ssl/s2*.c.
6022 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6024 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6025 has a return value which indicates the quality of the random data
6026 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6027 error queue. New function RAND_pseudo_bytes() generates output that is
6028 guaranteed to be unique but not unpredictable. RAND_add is like
6029 RAND_seed, but takes an extra argument for an entropy estimate
6030 (RAND_seed always assumes full entropy).
6033 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6034 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6035 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6036 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6037 false-positive rate of at most 2^-80 for random input.
6040 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6043 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6044 in the 0.9.5 release), this returns the chain
6045 from an X509_CTX structure with a dup of the stack and all
6046 the X509 reference counts upped: so the stack will exist
6047 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6050 Also make SSL_SESSION_print() print out the verify return
6054 *) Add manpage for the pkcs12 command. Also change the default
6055 behaviour so MAC iteration counts are used unless the new
6056 -nomaciter option is used. This improves file security and
6057 only older versions of MSIE (4.0 for example) need it.
6060 *) Honor the no-xxx Configure options when creating .DEF files.
6063 *) Add PKCS#10 attributes to field table: challengePassword,
6064 unstructuredName and unstructuredAddress. These are taken from
6065 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6066 international characters are used.
6068 More changes to X509_ATTRIBUTE code: allow the setting of types
6069 based on strings. Remove the 'loc' parameter when adding
6070 attributes because these will be a SET OF encoding which is sorted
6074 *) Initial changes to the 'req' utility to allow request generation
6075 automation. This will allow an application to just generate a template
6076 file containing all the field values and have req construct the
6079 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6080 used all over the place including certificate requests and PKCS#7
6081 structures. They are currently handled manually where necessary with
6082 some primitive wrappers for PKCS#7. The new functions behave in a
6083 manner analogous to the X509 extension functions: they allow
6084 attributes to be looked up by NID and added.
6086 Later something similar to the X509V3 code would be desirable to
6087 automatically handle the encoding, decoding and printing of the
6088 more complex types. The string types like challengePassword can
6089 be handled by the string table functions.
6091 Also modified the multi byte string table handling. Now there is
6092 a 'global mask' which masks out certain types. The table itself
6093 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6094 is useful when for example there is only one permissible type
6095 (as in countryName) and using the mask might result in no valid
6099 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6100 SSL_get_peer_finished to allow applications to obtain the latest
6101 Finished messages sent to the peer or expected from the peer,
6102 respectively. (SSL_get_peer_finished is usually the Finished message
6103 actually received from the peer, otherwise the protocol will be aborted.)
6105 As the Finished message are message digests of the complete handshake
6106 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6107 be used for external authentication procedures when the authentication
6108 provided by SSL/TLS is not desired or is not enough.
6111 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6112 the host supports BWX extension and if Compaq C is present on the
6113 $PATH. Just exploiting of the BWX extension results in 20-30%
6114 performance kick for some algorithms, e.g. DES and RC4 to mention
6115 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6119 *) Add support for MS "fast SGC". This is arguably a violation of the
6120 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6121 weak crypto and after checking the certificate is SGC a second one
6122 with strong crypto. MS SGC stops the first handshake after receiving
6123 the server certificate message and sends a second client hello. Since
6124 a server will typically do all the time consuming operations before
6125 expecting any further messages from the client (server key exchange
6126 is the most expensive) there is little difference between the two.
6128 To get OpenSSL to support MS SGC we have to permit a second client
6129 hello message after we have sent server done. In addition we have to
6130 reset the MAC if we do get this second client hello.
6133 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6134 if a DER encoded private key is RSA or DSA traditional format. Changed
6135 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6136 format DER encoded private key. Newer code should use PKCS#8 format which
6137 has the key type encoded in the ASN1 structure. Added DER private key
6138 support to pkcs8 application.
6141 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6142 ciphersuites has been selected (as required by the SSL 3/TLS 1
6143 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6144 is set, we interpret this as a request to violate the specification
6145 (the worst that can happen is a handshake failure, and 'correct'
6146 behaviour would result in a handshake failure anyway).
6149 *) In SSL_CTX_add_session, take into account that there might be multiple
6150 SSL_SESSION structures with the same session ID (e.g. when two threads
6151 concurrently obtain them from an external cache).
6152 The internal cache can handle only one SSL_SESSION with a given ID,
6153 so if there's a conflict, we now throw out the old one to achieve
6157 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6158 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6159 some routines that use cipher OIDs: some ciphers do not have OIDs
6160 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6164 *) Simplify the trust setting structure and code. Now we just have
6165 two sequences of OIDs for trusted and rejected settings. These will
6166 typically have values the same as the extended key usage extension
6167 and any application specific purposes.
6169 The trust checking code now has a default behaviour: it will just
6170 check for an object with the same NID as the passed id. Functions can
6171 be provided to override either the default behaviour or the behaviour
6172 for a given id. SSL client, server and email already have functions
6173 in place for compatibility: they check the NID and also return "trusted"
6174 if the certificate is self signed.
6177 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6178 traditional format into an EVP_PKEY structure.
6181 *) Add a password callback function PEM_cb() which either prompts for
6182 a password if usr_data is NULL or otherwise assumes it is a null
6183 terminated password. Allow passwords to be passed on command line
6184 environment or config files in a few more utilities.
6187 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6188 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6189 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6190 Update documentation.
6193 *) Support for ASN1 "NULL" type. This could be handled before by using
6194 ASN1_TYPE but there wasn't any function that would try to read a NULL
6195 and produce an error if it couldn't. For compatibility we also have
6196 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6197 don't allocate anything because they don't need to.
6200 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6202 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6204 *) Rebuild of the memory allocation routines used by OpenSSL code and
6205 possibly others as well. The purpose is to make an interface that
6206 provide hooks so anyone can build a separate set of allocation and
6207 deallocation routines to be used by OpenSSL, for example memory
6208 pool implementations, or something else, which was previously hard
6209 since Malloc(), Realloc() and Free() were defined as macros having
6210 the values malloc, realloc and free, respectively (except for Win32
6211 compilations). The same is provided for memory debugging code.
6212 OpenSSL already comes with functionality to find memory leaks, but
6213 this gives people a chance to debug other memory problems.
6215 With these changes, a new set of functions and macros have appeared:
6217 CRYPTO_set_mem_debug_functions() [F]
6218 CRYPTO_get_mem_debug_functions() [F]
6219 CRYPTO_dbg_set_options() [F]
6220 CRYPTO_dbg_get_options() [F]
6221 CRYPTO_malloc_debug_init() [M]
6223 The memory debug functions are NULL by default, unless the library
6224 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6225 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6226 gives the standard debugging functions that come with OpenSSL) or
6227 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6228 provided by the library user) must be used. When the standard
6229 debugging functions are used, CRYPTO_dbg_set_options can be used to
6230 request additional information:
6231 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6232 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6234 Also, things like CRYPTO_set_mem_functions will always give the
6235 expected result (the new set of functions is used for allocation
6236 and deallocation) at all times, regardless of platform and compiler
6239 To finish it up, some functions that were never use in any other
6240 way than through macros have a new API and new semantic:
6243 CRYPTO_dbg_realloc()
6246 All macros of value have retained their old syntax.
6247 [Richard Levitte and Bodo Moeller]
6249 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6250 ordering of SMIMECapabilities wasn't in "strength order" and there
6251 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6255 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6256 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6257 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6259 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6260 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6261 functionality to handle multipart/signed properly) and a utility
6262 called 'smime' to call all this stuff. This is based on code I
6263 originally wrote for Celo who have kindly allowed it to be
6264 included in OpenSSL.
6267 *) Add variants des_set_key_checked and des_set_key_unchecked of
6268 des_set_key (aka des_key_sched). Global variable des_check_key
6269 decides which of these is called by des_set_key; this way
6270 des_check_key behaves as it always did, but applications and
6271 the library itself, which was buggy for des_check_key == 1,
6272 have a cleaner way to pick the version they need.
6275 *) New function PKCS12_newpass() which changes the password of a
6279 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6280 dynamic mix. In both cases the ids can be used as an index into the
6281 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6282 functions so they accept a list of the field values and the
6283 application doesn't need to directly manipulate the X509_TRUST
6287 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6291 *) Modify the way the V3 extension code looks up extensions. This now
6292 works in a similar way to the object code: we have some "standard"
6293 extensions in a static table which is searched with OBJ_bsearch()
6294 and the application can add dynamic ones if needed. The file
6295 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6296 updated whenever a new extension is added to the core code and kept
6297 in ext_nid order. There is a simple program 'tabtest.c' which checks
6298 this. New extensions are not added too often so this file can readily
6299 be maintained manually.
6301 There are two big advantages in doing things this way. The extensions
6302 can be looked up immediately and no longer need to be "added" using
6303 X509V3_add_standard_extensions(): this function now does nothing.
6304 [Side note: I get *lots* of email saying the extension code doesn't
6305 work because people forget to call this function]
6306 Also no dynamic allocation is done unless new extensions are added:
6307 so if we don't add custom extensions there is no need to call
6308 X509V3_EXT_cleanup().
6311 *) Modify enc utility's salting as follows: make salting the default. Add a
6312 magic header, so unsalted files fail gracefully instead of just decrypting
6313 to garbage. This is because not salting is a big security hole, so people
6314 should be discouraged from doing it.
6317 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6318 digest to be passed on the command line but it only used this
6319 parameter when signing a certificate. Modified so all relevant
6320 operations are affected by the digest parameter including the
6321 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6322 DSA key was used because it didn't fix the digest.
6325 *) Initial certificate chain verify code. Currently tests the untrusted
6326 certificates for consistency with the verify purpose (which is set
6327 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6329 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6330 this is because it will reject chains with invalid extensions whereas
6331 every previous version of OpenSSL and SSLeay made no checks at all.
6333 Trust code: checks the root CA for the relevant trust settings. Trust
6334 settings have an initial value consistent with the verify purpose: e.g.
6335 if the verify purpose is for SSL client use it expects the CA to be
6336 trusted for SSL client use. However the default value can be changed to
6337 permit custom trust settings: one example of this would be to only trust
6338 certificates from a specific "secure" set of CAs.
6340 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6341 which should be used for version portability: especially since the
6342 verify structure is likely to change more often now.
6344 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6345 to set them. If not set then assume SSL clients will verify SSL servers
6348 Two new options to the verify program: -untrusted allows a set of
6349 untrusted certificates to be passed in and -purpose which sets the
6350 intended purpose of the certificate. If a purpose is set then the
6351 new chain verify code is used to check extension consistency.
6354 *) Support for the authority information access extension.
6357 *) Modify RSA and DSA PEM read routines to transparently handle
6358 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6359 public keys in a format compatible with certificate
6360 SubjectPublicKeyInfo structures. Unfortunately there were already
6361 functions called *_PublicKey_* which used various odd formats so
6362 these are retained for compatibility: however the DSA variants were
6363 never in a public release so they have been deleted. Changed dsa/rsa
6364 utilities to handle the new format: note no releases ever handled public
6365 keys so we should be OK.
6367 The primary motivation for this change is to avoid the same fiasco
6368 that dogs private keys: there are several incompatible private key
6369 formats some of which are standard and some OpenSSL specific and
6370 require various evil hacks to allow partial transparent handling and
6371 even then it doesn't work with DER formats. Given the option anything
6372 other than PKCS#8 should be dumped: but the other formats have to
6373 stay in the name of compatibility.
6375 With public keys and the benefit of hindsight one standard format
6376 is used which works with EVP_PKEY, RSA or DSA structures: though
6377 it clearly returns an error if you try to read the wrong kind of key.
6379 Added a -pubkey option to the 'x509' utility to output the public key.
6380 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6381 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6382 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6383 that do the same as the EVP_PKEY_assign_*() except they up the
6384 reference count of the added key (they don't "swallow" the
6388 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6389 CRLs would fail if the file contained no certificates or no CRLs:
6390 added a new function to read in both types and return the number
6391 read: this means that if none are read it will be an error. The
6392 DER versions of the certificate and CRL reader would always fail
6393 because it isn't possible to mix certificates and CRLs in DER format
6394 without choking one or the other routine. Changed this to just read
6395 a certificate: this is the best we can do. Also modified the code
6396 in apps/verify.c to take notice of return codes: it was previously
6397 attempting to read in certificates from NULL pointers and ignoring
6398 any errors: this is one reason why the cert and CRL reader seemed
6399 to work. It doesn't check return codes from the default certificate
6400 routines: these may well fail if the certificates aren't installed.
6403 *) Code to support otherName option in GeneralName.
6406 *) First update to verify code. Change the verify utility
6407 so it warns if it is passed a self signed certificate:
6408 for consistency with the normal behaviour. X509_verify
6409 has been modified to it will now verify a self signed
6410 certificate if *exactly* the same certificate appears
6411 in the store: it was previously impossible to trust a
6412 single self signed certificate. This means that:
6413 openssl verify ss.pem
6414 now gives a warning about a self signed certificate but
6415 openssl verify -CAfile ss.pem ss.pem
6419 *) For servers, store verify_result in SSL_SESSION data structure
6420 (and add it to external session representation).
6421 This is needed when client certificate verifications fails,
6422 but an application-provided verification callback (set by
6423 SSL_CTX_set_cert_verify_callback) allows accepting the session
6424 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6425 but returns 1): When the session is reused, we have to set
6426 ssl->verify_result to the appropriate error code to avoid
6428 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6430 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6431 case in PKCS7_dataInit() where the signed PKCS7 structure
6432 didn't contain any existing data because it was being created.
6433 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6435 *) Add a salt to the key derivation routines in enc.c. This
6436 forms the first 8 bytes of the encrypted file. Also add a
6437 -S option to allow a salt to be input on the command line.
6440 *) New function X509_cmp(). Oddly enough there wasn't a function
6441 to compare two certificates. We do this by working out the SHA1
6442 hash and comparing that. X509_cmp() will be needed by the trust
6446 *) SSL_get1_session() is like SSL_get_session(), but increments
6447 the reference count in the SSL_SESSION returned.
6448 [Geoff Thorpe <geoff@eu.c2.net>]
6450 *) Fix for 'req': it was adding a null to request attributes.
6451 Also change the X509_LOOKUP and X509_INFO code to handle
6452 certificate auxiliary information.
6455 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6459 *) Add the possibility to add extra information to the memory leak
6460 detecting output, to form tracebacks, showing from where each
6461 allocation was originated: CRYPTO_push_info("constant string") adds
6462 the string plus current file name and line number to a per-thread
6463 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6464 is like calling CYRPTO_pop_info() until the stack is empty.
6465 Also updated memory leak detection code to be multi-thread-safe.
6468 *) Add options -text and -noout to pkcs7 utility and delete the
6469 encryption options which never did anything. Update docs.
6472 *) Add options to some of the utilities to allow the pass phrase
6473 to be included on either the command line (not recommended on
6474 OSes like Unix) or read from the environment. Update the
6475 manpages and fix a few bugs.
6478 *) Add a few manpages for some of the openssl commands.
6481 *) Fix the -revoke option in ca. It was freeing up memory twice,
6482 leaking and not finding already revoked certificates.
6485 *) Extensive changes to support certificate auxiliary information.
6486 This involves the use of X509_CERT_AUX structure and X509_AUX
6487 functions. An X509_AUX function such as PEM_read_X509_AUX()
6488 can still read in a certificate file in the usual way but it
6489 will also read in any additional "auxiliary information". By
6490 doing things this way a fair degree of compatibility can be
6491 retained: existing certificates can have this information added
6492 using the new 'x509' options.
6494 Current auxiliary information includes an "alias" and some trust
6495 settings. The trust settings will ultimately be used in enhanced
6496 certificate chain verification routines: currently a certificate
6497 can only be trusted if it is self signed and then it is trusted
6501 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6502 The problem was that one of the replacement routines had not been working
6503 since SSLeay releases. For now the offending routine has been replaced
6504 with non-optimised assembler. Even so, this now gives around 95%
6505 performance improvement for 1024 bit RSA signs.
6508 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6509 handling. Most clients have the effective key size in bits equal to
6510 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6511 A few however don't do this and instead use the size of the decrypted key
6512 to determine the RC2 key length and the AlgorithmIdentifier to determine
6513 the effective key length. In this case the effective key length can still
6514 be 40 bits but the key length can be 168 bits for example. This is fixed
6515 by manually forcing an RC2 key into the EVP_PKEY structure because the
6516 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6517 the key length and effective key length are equal.
6520 *) Add a bunch of functions that should simplify the creation of
6521 X509_NAME structures. Now you should be able to do:
6522 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6523 and have it automatically work out the correct field type and fill in
6524 the structures. The more adventurous can try:
6525 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6526 and it will (hopefully) work out the correct multibyte encoding.
6529 *) Change the 'req' utility to use the new field handling and multibyte
6530 copy routines. Before the DN field creation was handled in an ad hoc
6531 way in req, ca, and x509 which was rather broken and didn't support
6532 BMPStrings or UTF8Strings. Since some software doesn't implement
6533 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6534 using the dirstring_type option. See the new comment in the default
6535 openssl.cnf for more info.
6538 *) Make crypto/rand/md_rand.c more robust:
6539 - Assure unique random numbers after fork().
6540 - Make sure that concurrent threads access the global counter and
6541 md serializably so that we never lose entropy in them
6542 or use exactly the same state in multiple threads.
6543 Access to the large state is not always serializable because
6544 the additional locking could be a performance killer, and
6545 md should be large enough anyway.
6548 *) New file apps/app_rand.c with commonly needed functionality
6549 for handling the random seed file.
6551 Use the random seed file in some applications that previously did not:
6553 dsaparam -genkey (which also ignored its '-rand' option),
6556 x509 (when signing).
6557 Except on systems with /dev/urandom, it is crucial to have a random
6558 seed file at least for key creation, DSA signing, and for DH exchanges;
6559 for RSA signatures we could do without one.
6561 gendh and gendsa (unlike genrsa) used to read only the first byte
6562 of each file listed in the '-rand' option. The function as previously
6563 found in genrsa is now in app_rand.c and is used by all programs
6564 that support '-rand'.
6567 *) In RAND_write_file, use mode 0600 for creating files;
6568 don't just chmod when it may be too late.
6571 *) Report an error from X509_STORE_load_locations
6572 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6575 *) New function ASN1_mbstring_copy() this copies a string in either
6576 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6577 into an ASN1_STRING type. A mask of permissible types is passed
6578 and it chooses the "minimal" type to use or an error if not type
6582 *) Add function equivalents to the various macros in asn1.h. The old
6583 macros are retained with an M_ prefix. Code inside the library can
6584 use the M_ macros. External code (including the openssl utility)
6585 should *NOT* in order to be "shared library friendly".
6588 *) Add various functions that can check a certificate's extensions
6589 to see if it usable for various purposes such as SSL client,
6590 server or S/MIME and CAs of these types. This is currently
6591 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6592 verification. Also added a -purpose flag to x509 utility to
6593 print out all the purposes.
6596 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6600 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6601 for, obtain and decode and extension and obtain its critical flag.
6602 This allows all the necessary extension code to be handled in a
6603 single function call.
6606 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6607 platforms. See crypto/rc4/rc4_enc.c for further details.
6610 *) New -noout option to asn1parse. This causes no output to be produced
6611 its main use is when combined with -strparse and -out to extract data
6612 from a file (which may not be in ASN.1 format).
6615 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6616 when producing the local key id.
6617 [Richard Levitte <levitte@stacken.kth.se>]
6619 *) New option -dhparam in s_server. This allows a DH parameter file to be
6620 stated explicitly. If it is not stated then it tries the first server
6621 certificate file. The previous behaviour hard coded the filename
6625 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6626 a public key to be input or output. For example:
6627 openssl rsa -in key.pem -pubout -out pubkey.pem
6628 Also added necessary DSA public key functions to handle this.
6631 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6632 in the message. This was handled by allowing
6633 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6634 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6636 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6637 to the end of the strings whereas this didn't. This would cause problems
6638 if strings read with d2i_ASN1_bytes() were later modified.
6639 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6641 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6642 data and it contains EOF it will end up returning an error. This is
6643 caused by input 46 bytes long. The cause is due to the way base64
6644 BIOs find the start of base64 encoded data. They do this by trying a
6645 trial decode on each line until they find one that works. When they
6646 do a flag is set and it starts again knowing it can pass all the
6647 data directly through the decoder. Unfortunately it doesn't reset
6648 the context it uses. This means that if EOF is reached an attempt
6649 is made to pass two EOFs through the context and this causes the
6650 resulting error. This can also cause other problems as well. As is
6651 usual with these problems it takes *ages* to find and the fix is
6652 trivial: move one line.
6653 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6655 *) Ugly workaround to get s_client and s_server working under Windows. The
6656 old code wouldn't work because it needed to select() on sockets and the
6657 tty (for keypresses and to see if data could be written). Win32 only
6658 supports select() on sockets so we select() with a 1s timeout on the
6659 sockets and then see if any characters are waiting to be read, if none
6660 are present then we retry, we also assume we can always write data to
6661 the tty. This isn't nice because the code then blocks until we've
6662 received a complete line of data and it is effectively polling the
6663 keyboard at 1s intervals: however it's quite a bit better than not
6664 working at all :-) A dedicated Windows application might handle this
6665 with an event loop for example.
6668 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6669 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6670 will be called when RSA_sign() and RSA_verify() are used. This is useful
6671 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6672 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6673 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6674 This necessitated the support of an extra signature type NID_md5_sha1
6675 for SSL signatures and modifications to the SSL library to use it instead
6676 of calling RSA_public_decrypt() and RSA_private_encrypt().
6679 *) Add new -verify -CAfile and -CApath options to the crl program, these
6680 will lookup a CRL issuers certificate and verify the signature in a
6681 similar way to the verify program. Tidy up the crl program so it
6682 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6683 less strict. It will now permit CRL extensions even if it is not
6684 a V2 CRL: this will allow it to tolerate some broken CRLs.
6687 *) Initialize all non-automatic variables each time one of the openssl
6688 sub-programs is started (this is necessary as they may be started
6689 multiple times from the "OpenSSL>" prompt).
6690 [Lennart Bang, Bodo Moeller]
6692 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6693 removing all other RSA functionality (this is what NO_RSA does). This
6694 is so (for example) those in the US can disable those operations covered
6695 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6699 *) Non-copying interface to BIO pairs.
6700 (still largely untested)
6703 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6704 ASCII string. This was handled independently in various places before.
6707 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6708 UTF8 strings a character at a time.
6711 *) Use client_version from client hello to select the protocol
6712 (s23_srvr.c) and for RSA client key exchange verification
6713 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6716 *) Add various utility functions to handle SPKACs, these were previously
6717 handled by poking round in the structure internals. Added new function
6718 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6719 print, verify and generate SPKACs. Based on an original idea from
6720 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6723 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6726 *) Allow the config file extension section to be overwritten on the
6727 command line. Based on an original idea from Massimiliano Pala
6728 <madwolf@comune.modena.it>. The new option is called -extensions
6729 and can be applied to ca, req and x509. Also -reqexts to override
6730 the request extensions in req and -crlexts to override the crl extensions
6734 *) Add new feature to the SPKAC handling in ca. Now you can include
6735 the same field multiple times by preceding it by "XXXX." for example:
6738 this is the same syntax as used in the req config file.
6741 *) Allow certificate extensions to be added to certificate requests. These
6742 are specified in a 'req_extensions' option of the req section of the
6743 config file. They can be printed out with the -text option to req but
6744 are otherwise ignored at present.
6747 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6748 data read consists of only the final block it would not decrypted because
6749 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6750 A misplaced 'break' also meant the decrypted final block might not be
6751 copied until the next read.
6754 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6755 a few extra parameters to the DH structure: these will be useful if
6756 for example we want the value of 'q' or implement X9.42 DH.
6759 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6760 provides hooks that allow the default DSA functions or functions on a
6761 "per key" basis to be replaced. This allows hardware acceleration and
6762 hardware key storage to be handled without major modification to the
6763 library. Also added low level modexp hooks and CRYPTO_EX structure and
6764 associated functions.
6767 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6768 as "read only": it can't be written to and the buffer it points to will
6769 not be freed. Reading from a read only BIO is much more efficient than
6770 a normal memory BIO. This was added because there are several times when
6771 an area of memory needs to be read from a BIO. The previous method was
6772 to create a memory BIO and write the data to it, this results in two
6773 copies of the data and an O(n^2) reading algorithm. There is a new
6774 function BIO_new_mem_buf() which creates a read only memory BIO from
6775 an area of memory. Also modified the PKCS#7 routines to use read only
6779 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6780 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6781 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6782 but a retry condition occured while trying to read the rest.
6785 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6786 NID_pkcs7_encrypted by default: this was wrong since this should almost
6787 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6788 the encrypted data type: this is a more sensible place to put it and it
6789 allows the PKCS#12 code to be tidied up that duplicated this
6793 *) Changed obj_dat.pl script so it takes its input and output files on
6794 the command line. This should avoid shell escape redirection problems
6798 *) Initial support for certificate extension requests, these are included
6799 in things like Xenroll certificate requests. Included functions to allow
6800 extensions to be obtained and added.
6803 *) -crlf option to s_client and s_server for sending newlines as
6804 CRLF (as required by many protocols).
6807 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6809 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6810 [Ralf S. Engelschall]
6812 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6813 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6815 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6819 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6820 DH parameters/keys (q is lost during that conversion, but the resulting
6821 DH parameters contain its length).
6823 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6824 much faster than DH_generate_parameters (which creates parameters
6825 where p = 2*q + 1), and also the smaller q makes DH computations
6826 much more efficient (160-bit exponentiation instead of 1024-bit
6827 exponentiation); so this provides a convenient way to support DHE
6828 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6829 utter importance to use
6830 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6832 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6833 when such DH parameters are used, because otherwise small subgroup
6834 attacks may become possible!
6837 *) Avoid memory leak in i2d_DHparams.
6840 *) Allow the -k option to be used more than once in the enc program:
6841 this allows the same encrypted message to be read by multiple recipients.
6844 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6845 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6846 it will always use the numerical form of the OID, even if it has a short
6850 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6851 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6852 otherwise bn_mod_exp was called. In the case of hardware keys for example
6853 no private key components need be present and it might store extra data
6854 in the RSA structure, which cannot be accessed from bn_mod_exp.
6855 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6856 private key operations.
6859 *) Added support for SPARC Linux.
6862 *) pem_password_cb function type incompatibly changed from
6863 typedef int pem_password_cb(char *buf, int size, int rwflag);
6865 ....(char *buf, int size, int rwflag, void *userdata);
6866 so that applications can pass data to their callbacks:
6867 The PEM[_ASN1]_{read,write}... functions and macros now take an
6868 additional void * argument, which is just handed through whenever
6869 the password callback is called.
6870 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6872 New function SSL_CTX_set_default_passwd_cb_userdata.
6874 Compatibility note: As many C implementations push function arguments
6875 onto the stack in reverse order, the new library version is likely to
6876 interoperate with programs that have been compiled with the old
6877 pem_password_cb definition (PEM_whatever takes some data that
6878 happens to be on the stack as its last argument, and the callback
6879 just ignores this garbage); but there is no guarantee whatsoever that
6882 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6883 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6884 problems not only on Windows, but also on some Unix platforms.
6885 To avoid problematic command lines, these definitions are now in an
6886 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6887 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6890 *) MIPS III/IV assembler module is reimplemented.
6893 *) More DES library cleanups: remove references to srand/rand and
6894 delete an unused file.
6897 *) Add support for the the free Netwide assembler (NASM) under Win32,
6898 since not many people have MASM (ml) and it can be hard to obtain.
6899 This is currently experimental but it seems to work OK and pass all
6900 the tests. Check out INSTALL.W32 for info.
6903 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6904 without temporary keys kept an extra copy of the server key,
6905 and connections with temporary keys did not free everything in case
6909 *) New function RSA_check_key and new openssl rsa option -check
6910 for verifying the consistency of RSA keys.
6911 [Ulf Moeller, Bodo Moeller]
6913 *) Various changes to make Win32 compile work:
6914 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6915 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6916 comparison" warnings.
6917 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6920 *) Add a debugging option to PKCS#5 v2 key generation function: when
6921 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6922 derived keys are printed to stderr.
6925 *) Copy the flags in ASN1_STRING_dup().
6926 [Roman E. Pavlov <pre@mo.msk.ru>]
6928 *) The x509 application mishandled signing requests containing DSA
6929 keys when the signing key was also DSA and the parameters didn't match.
6931 It was supposed to omit the parameters when they matched the signing key:
6932 the verifying software was then supposed to automatically use the CA's
6933 parameters if they were absent from the end user certificate.
6935 Omitting parameters is no longer recommended. The test was also
6936 the wrong way round! This was probably due to unusual behaviour in
6937 EVP_cmp_parameters() which returns 1 if the parameters match.
6938 This meant that parameters were omitted when they *didn't* match and
6939 the certificate was useless. Certificates signed with 'ca' didn't have
6941 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6943 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6944 The interface is as follows:
6945 Applications can use
6946 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6947 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6948 "off" is now the default.
6949 The library internally uses
6950 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6951 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6952 to disable memory-checking temporarily.
6954 Some inconsistent states that previously were possible (and were
6955 even the default) are now avoided.
6957 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6958 with each memory chunk allocated; this is occasionally more helpful
6959 than just having a counter.
6961 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6963 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6967 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6968 which largely parallels "options", but is for changing API behaviour,
6969 whereas "options" are about protocol behaviour.
6970 Initial "mode" flags are:
6972 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6973 a single record has been written.
6974 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6975 retries use the same buffer location.
6976 (But all of the contents must be
6980 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6983 *) Fix problems with no-hmac etc.
6984 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6986 *) New functions RSA_get_default_method(), RSA_set_method() and
6987 RSA_get_method(). These allows replacement of RSA_METHODs without having
6988 to mess around with the internals of an RSA structure.
6991 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6992 Also really enable memory leak checks in openssl.c and in some
6994 [Chad C. Mulligan, Bodo Moeller]
6996 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6997 up the length of negative integers. This has now been simplified to just
6998 store the length when it is first determined and use it later, rather
6999 than trying to keep track of where data is copied and updating it to
7001 [Steve Henson, reported by Brien Wheeler
7002 <bwheeler@authentica-security.com>]
7004 *) Add a new function PKCS7_signatureVerify. This allows the verification
7005 of a PKCS#7 signature but with the signing certificate passed to the
7006 function itself. This contrasts with PKCS7_dataVerify which assumes the
7007 certificate is present in the PKCS#7 structure. This isn't always the
7008 case: certificates can be omitted from a PKCS#7 structure and be
7009 distributed by "out of band" means (such as a certificate database).
7012 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7013 function prototypes in pem.h, also change util/mkdef.pl to add the
7014 necessary function names.
7017 *) mk1mf.pl (used by Windows builds) did not properly read the
7018 options set by Configure in the top level Makefile, and Configure
7019 was not even able to write more than one option correctly.
7020 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7023 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7024 file to be loaded from a BIO or FILE pointer. The BIO version will
7025 for example allow memory BIOs to contain config info.
7028 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7029 Whoever hopes to achieve shared-library compatibility across versions
7030 must use this, not the compile-time macro.
7031 (Exercise 0.9.4: Which is the minimum library version required by
7033 Note: All this applies only to multi-threaded programs, others don't
7037 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7038 through a BIO pair triggered the default case, i.e.
7039 SSLerr(...,SSL_R_UNKNOWN_STATE).
7042 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7043 can use the SSL library even if none of the specific BIOs is
7047 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7048 for the encoded length.
7049 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7051 *) Add initial documentation of the X509V3 functions.
7054 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7055 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7056 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7057 secure PKCS#8 private key format with a high iteration count.
7060 *) Fix determination of Perl interpreter: A perl or perl5
7061 _directory_ in $PATH was also accepted as the interpreter.
7062 [Ralf S. Engelschall]
7064 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7065 wrong with it but it was very old and did things like calling
7066 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7070 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7071 to use the new extension code.
7074 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7075 with macros. This should make it easier to change their form, add extra
7076 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7080 *) Add to configuration table a new entry that can specify an alternative
7081 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7082 according to Mark Crispin <MRC@Panda.COM>.
7086 *) DES CBC did not update the IV. Weird.
7089 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7090 Changing the behaviour of the former might break existing programs --
7091 where IV updating is needed, des_ncbc_encrypt can be used.
7094 *) When bntest is run from "make test" it drives bc to check its
7095 calculations, as well as internally checking them. If an internal check
7096 fails, it needs to cause bc to give a non-zero result or make test carries
7097 on without noticing the failure. Fixed.
7100 *) DES library cleanups.
7103 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7104 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7105 ciphers. NOTE: although the key derivation function has been verified
7106 against some published test vectors it has not been extensively tested
7107 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7111 *) Instead of "mkdir -p", which is not fully portable, use new
7112 Perl script "util/mkdir-p.pl".
7115 *) Rewrite the way password based encryption (PBE) is handled. It used to
7116 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7117 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7118 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7119 the 'parameter' field of the AlgorithmIdentifier is passed to the
7120 underlying key generation function so it must do its own ASN1 parsing.
7121 This has also changed the EVP_PBE_CipherInit() function which now has a
7122 'parameter' argument instead of literal salt and iteration count values
7123 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7126 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7127 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7128 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7129 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7130 value was just used as a "magic string" and not used directly its
7131 value doesn't matter.
7134 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7138 *) "linux-sparc64" configuration (ultrapenguin).
7139 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7140 "linux-sparc" configuration.
7141 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7143 *) config now generates no-xxx options for missing ciphers.
7146 *) Support the EBCDIC character set (work in progress).
7147 File ebcdic.c not yet included because it has a different license.
7148 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7150 *) Support BS2000/OSD-POSIX.
7151 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7153 *) Make callbacks for key generation use void * instead of char *.
7156 *) Make S/MIME samples compile (not yet tested).
7159 *) Additional typesafe stacks.
7162 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7166 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7168 *) New configuration variant "sco5-gcc".
7170 *) Updated some demos.
7171 [Sean O Riordain, Wade Scholine]
7173 *) Add missing BIO_free at exit of pkcs12 application.
7176 *) Fix memory leak in conf.c.
7179 *) Updates for Win32 to assembler version of MD5.
7182 *) Set #! path to perl in apps/der_chop to where we found it
7183 instead of using a fixed path.
7186 *) SHA library changes for irix64-mips4-cc.
7189 *) Improvements for VMS support.
7193 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7195 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7196 This also avoids the problems with SC4.2 and unpatched SC5.
7197 [Andy Polyakov <appro@fy.chalmers.se>]
7199 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7200 These are required because of the typesafe stack would otherwise break
7201 existing code. If old code used a structure member which used to be STACK
7202 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7203 sk_num or sk_value it would produce an error because the num, data members
7204 are not present in STACK_OF. Now it just produces a warning. sk_set
7205 replaces the old method of assigning a value to sk_value
7206 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7207 that does this will no longer work (and should use sk_set instead) but
7208 this could be regarded as a "questionable" behaviour anyway.
7211 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7212 correctly handle encrypted S/MIME data.
7215 *) Change type of various DES function arguments from des_cblock
7216 (which means, in function argument declarations, pointer to char)
7217 to des_cblock * (meaning pointer to array with 8 char elements),
7218 which allows the compiler to do more typechecking; it was like
7219 that back in SSLeay, but with lots of ugly casts.
7221 Introduce new type const_des_cblock.
7224 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7225 problems: find RecipientInfo structure that matches recipient certificate
7226 and initialise the ASN1 structures properly based on passed cipher.
7229 *) Belatedly make the BN tests actually check the results.
7232 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7233 to and from BNs: it was completely broken. New compilation option
7234 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7235 key elements as negative integers.
7238 *) Reorganize and speed up MD5.
7239 [Andy Polyakov <appro@fy.chalmers.se>]
7242 [Richard Levitte <richard@levitte.org>]
7244 *) New option -out to asn1parse to allow the parsed structure to be
7245 output to a file. This is most useful when combined with the -strparse
7246 option to examine the output of things like OCTET STRINGS.
7249 *) Make SSL library a little more fool-proof by not requiring any longer
7250 that SSL_set_{accept,connect}_state be called before
7251 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7252 in many applications because usually everything *appeared* to work as
7253 intended anyway -- now it really works as intended).
7256 *) Move openssl.cnf out of lib/.
7259 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7260 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7261 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7262 [Ralf S. Engelschall]
7264 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7265 handle PKCS#7 enveloped data properly.
7266 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7268 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7269 copying pointers. The cert_st handling is changed by this in
7270 various ways (and thus what used to be known as ctx->default_cert
7271 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7272 any longer when s->cert does not give us what we need).
7273 ssl_cert_instantiate becomes obsolete by this change.
7274 As soon as we've got the new code right (possibly it already is?),
7275 we have solved a couple of bugs of the earlier code where s->cert
7276 was used as if it could not have been shared with other SSL structures.
7278 Note that using the SSL API in certain dirty ways now will result
7279 in different behaviour than observed with earlier library versions:
7280 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7281 does not influence s as it used to.
7283 In order to clean up things more thoroughly, inside SSL_SESSION
7284 we don't use CERT any longer, but a new structure SESS_CERT
7285 that holds per-session data (if available); currently, this is
7286 the peer's certificate chain and, for clients, the server's certificate
7287 and temporary key. CERT holds only those values that can have
7288 meaningful defaults in an SSL_CTX.
7291 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7292 from the internal representation. Various PKCS#7 fixes: remove some
7293 evil casts and set the enc_dig_alg field properly based on the signing
7297 *) Allow PKCS#12 password to be set from the command line or the
7298 environment. Let 'ca' get its config file name from the environment
7299 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7303 *) Allow certificate policies extension to use an IA5STRING for the
7304 organization field. This is contrary to the PKIX definition but
7305 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7309 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7310 without disallowing inline assembler and the like for non-pedantic builds.
7313 *) Support Borland C++ builder.
7314 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7319 *) SHA-1 cleanups and performance enhancements.
7320 [Andy Polyakov <appro@fy.chalmers.se>]
7322 *) Sparc v8plus assembler for the bignum library.
7323 [Andy Polyakov <appro@fy.chalmers.se>]
7325 *) Accept any -xxx and +xxx compiler options in Configure.
7328 *) Update HPUX configuration.
7331 *) Add missing sk_<type>_unshift() function to safestack.h
7332 [Ralf S. Engelschall]
7334 *) New function SSL_CTX_use_certificate_chain_file that sets the
7335 "extra_cert"s in addition to the certificate. (This makes sense
7336 only for "PEM" format files, as chains as a whole are not
7340 *) Support verify_depth from the SSL API.
7341 x509_vfy.c had what can be considered an off-by-one-error:
7342 Its depth (which was not part of the external interface)
7343 was actually counting the number of certificates in a chain;
7344 now it really counts the depth.
7347 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7348 instead of X509err, which often resulted in confusing error
7349 messages since the error codes are not globally unique
7350 (e.g. an alleged error in ssl3_accept when a certificate
7351 didn't match the private key).
7353 *) New function SSL_CTX_set_session_id_context that allows to set a default
7354 value (so that you don't need SSL_set_session_id_context for each
7355 connection using the SSL_CTX).
7358 *) OAEP decoding bug fix.
7361 *) Support INSTALL_PREFIX for package builders, as proposed by
7365 *) New Configure options "threads" and "no-threads". For systems
7366 where the proper compiler options are known (currently Solaris
7367 and Linux), "threads" is the default.
7370 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7373 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7374 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7375 such as /usr/local/bin.
7378 *) "make linux-shared" to build shared libraries.
7379 [Niels Poppe <niels@netbox.org>]
7381 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7384 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7385 extension adding in x509 utility.
7388 *) Remove NOPROTO sections and error code comments.
7391 *) Partial rewrite of the DEF file generator to now parse the ANSI
7395 *) New Configure options --prefix=DIR and --openssldir=DIR.
7398 *) Complete rewrite of the error code script(s). It is all now handled
7399 by one script at the top level which handles error code gathering,
7400 header rewriting and C source file generation. It should be much better
7401 than the old method: it now uses a modified version of Ulf's parser to
7402 read the ANSI prototypes in all header files (thus the old K&R definitions
7403 aren't needed for error creation any more) and do a better job of
7404 translating function codes into names. The old 'ASN1 error code imbedded
7405 in a comment' is no longer necessary and it doesn't use .err files which
7406 have now been deleted. Also the error code call doesn't have to appear all
7407 on one line (which resulted in some large lines...).
7410 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7413 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7414 0 (which usually indicates a closed connection), but continue reading.
7417 *) Fix some race conditions.
7420 *) Add support for CRL distribution points extension. Add Certificate
7421 Policies and CRL distribution points documentation.
7424 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7427 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7428 8 of keying material. Merlin has also confirmed interop with this fix
7429 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7430 [Merlin Hughes <merlin@baltimore.ie>]
7432 *) Fix lots of warnings.
7433 [Richard Levitte <levitte@stacken.kth.se>]
7435 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7436 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7437 [Richard Levitte <levitte@stacken.kth.se>]
7439 *) Fix problems with sizeof(long) == 8.
7440 [Andy Polyakov <appro@fy.chalmers.se>]
7442 *) Change functions to ANSI C.
7445 *) Fix typos in error codes.
7446 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7448 *) Remove defunct assembler files from Configure.
7451 *) SPARC v8 assembler BIGNUM implementation.
7452 [Andy Polyakov <appro@fy.chalmers.se>]
7454 *) Support for Certificate Policies extension: both print and set.
7455 Various additions to support the r2i method this uses.
7458 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7459 return a const string when you are expecting an allocated buffer.
7462 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7463 types DirectoryString and DisplayText.
7466 *) Add code to allow r2i extensions to access the configuration database,
7467 add an LHASH database driver and add several ctx helper functions.
7470 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7471 fail when they extended the size of a BIGNUM.
7474 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7475 support typesafe stack.
7478 *) Fix typo in SSL_[gs]et_options().
7479 [Nils Frostberg <nils@medcom.se>]
7481 *) Delete various functions and files that belonged to the (now obsolete)
7482 old X509V3 handling code.
7485 *) New Configure option "rsaref".
7488 *) Don't auto-generate pem.h.
7491 *) Introduce type-safe ASN.1 SETs.
7494 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7495 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7497 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7498 that links with OpenSSL (well at least cause lots of warnings), but fear
7499 not: the conversion is trivial, and it eliminates loads of evil casts. A
7500 few STACKed things have been converted already. Feel free to convert more.
7501 In the fullness of time, I'll do away with the STACK type altogether.
7504 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7505 specified in <certfile> by updating the entry in the index.txt file.
7506 This way one no longer has to edit the index.txt file manually for
7507 revoking a certificate. The -revoke option does the gory details now.
7508 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7510 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7511 `-text' option at all and this way the `-noout -text' combination was
7512 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7513 [Ralf S. Engelschall]
7515 *) Make sure a corresponding plain text error message exists for the
7516 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7517 verify callback function determined that a certificate was revoked.
7518 [Ralf S. Engelschall]
7520 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7521 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7522 all available cipers including rc5, which was forgotten until now.
7523 In order to let the testing shell script know which algorithms
7524 are available, a new (up to now undocumented) command
7525 "openssl list-cipher-commands" is used.
7528 *) Bugfix: s_client occasionally would sleep in select() when
7529 it should have checked SSL_pending() first.
7532 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7533 the raw DSA values prior to ASN.1 encoding.
7536 *) Tweaks to Configure
7537 [Niels Poppe <niels@netbox.org>]
7539 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7543 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7546 *) New config option to avoid instructions that are illegal on the 80386.
7547 The default code is faster, but requires at least a 486.
7550 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7551 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7552 same as SSL2_VERSION anyway.
7555 *) New "-showcerts" option for s_client.
7558 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7559 application. Various cleanups and fixes.
7562 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7563 modify error routines to work internally. Add error codes and PBE init
7564 to library startup routines.
7567 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7568 packing functions to asn1 and evp. Changed function names and error
7569 codes along the way.
7572 *) PKCS12 integration: and so it begins... First of several patches to
7573 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7574 objects to objects.h
7577 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7578 and display support for Thawte strong extranet extension.
7581 *) Add LinuxPPC support.
7582 [Jeff Dubrule <igor@pobox.org>]
7584 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7585 bn_div_words in alpha.s.
7586 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7588 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7589 OAEP isn't supported when OpenSSL is built with RSAref.
7590 [Ulf Moeller <ulf@fitug.de>]
7592 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7593 so they no longer are missing under -DNOPROTO.
7594 [Soren S. Jorvang <soren@t.dk>]
7597 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7599 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7600 doesn't work when the session is reused. Coming soon!
7603 *) Fix a security hole, that allows sessions to be reused in the wrong
7604 context thus bypassing client cert protection! All software that uses
7605 client certs and session caches in multiple contexts NEEDS PATCHING to
7606 allow session reuse! A fuller solution is in the works.
7607 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7609 *) Some more source tree cleanups (removed obsolete files
7610 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7611 permission on "config" script to be executable) and a fix for the INSTALL
7613 [Ulf Moeller <ulf@fitug.de>]
7615 *) Remove some legacy and erroneous uses of malloc, free instead of
7617 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7619 *) Make rsa_oaep_test return non-zero on error.
7620 [Ulf Moeller <ulf@fitug.de>]
7622 *) Add support for native Solaris shared libraries. Configure
7623 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7624 if someone would make that last step automatic.
7625 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7627 *) ctx_size was not built with the right compiler during "make links". Fixed.
7630 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7631 except NULL ciphers". This means the default cipher list will no longer
7632 enable NULL ciphers. They need to be specifically enabled e.g. with
7633 the string "DEFAULT:eNULL".
7636 *) Fix to RSA private encryption routines: if p < q then it would
7637 occasionally produce an invalid result. This will only happen with
7638 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7641 *) Be less restrictive and allow also `perl util/perlpath.pl
7642 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7643 because this way one can also use an interpreter named `perl5' (which is
7644 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7645 installed as `perl').
7646 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7648 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7649 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7651 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7652 advapi32.lib to Win32 build and change the pem test comparision
7653 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7654 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7655 and crypto/des/ede_cbcm_enc.c.
7658 *) DES quad checksum was broken on big-endian architectures. Fixed.
7661 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7662 Win32 test batch file so it (might) work again. The Win32 test batch file
7663 is horrible: I feel ill....
7666 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7667 in e_os.h. Audit of header files to check ANSI and non ANSI
7668 sections: 10 functions were absent from non ANSI section and not exported
7669 from Windows DLLs. Fixed up libeay.num for new functions.
7672 *) Make `openssl version' output lines consistent.
7673 [Ralf S. Engelschall]
7675 *) Fix Win32 symbol export lists for BIO functions: Added
7676 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7677 to ms/libeay{16,32}.def.
7678 [Ralf S. Engelschall]
7680 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7681 fine under Unix and passes some trivial tests I've now added. But the
7682 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7683 added to make sure no one expects that this stuff really works in the
7684 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7685 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7687 [Ralf S. Engelschall]
7689 *) Fix the generation of two part addresses in perl.
7690 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7692 *) Add config entry for Linux on MIPS.
7693 [John Tobey <jtobey@channel1.com>]
7695 *) Make links whenever Configure is run, unless we are on Windoze.
7698 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7699 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7703 *) Add a useful kludge to allow package maintainers to specify compiler and
7704 other platforms details on the command line without having to patch the
7705 Configure script everytime: One now can use ``perl Configure
7706 <id>:<details>'', i.e. platform ids are allowed to have details appended
7707 to them (seperated by colons). This is treated as there would be a static
7708 pre-configured entry in Configure's %table under key <id> with value
7709 <details> and ``perl Configure <id>'' is called. So, when you want to
7710 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7711 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7712 now, which overrides the FreeBSD-elf entry on-the-fly.
7713 [Ralf S. Engelschall]
7715 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7718 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7719 on the `perl Configure ...' command line. This way one can compile
7720 OpenSSL libraries with Position Independent Code (PIC) which is needed
7721 for linking it into DSOs.
7722 [Ralf S. Engelschall]
7724 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7728 *) Cleaned up the LICENSE document: The official contact for any license
7729 questions now is the OpenSSL core team under openssl-core@openssl.org.
7730 And add a paragraph about the dual-license situation to make sure people
7731 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7732 to the OpenSSL toolkit.
7733 [Ralf S. Engelschall]
7735 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7736 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7737 Additonally cleaned up the `make links' target: Remove unnecessary
7738 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7739 to speed processing and no longer clutter the display with confusing
7740 stuff. Instead only the actually done links are displayed.
7741 [Ralf S. Engelschall]
7743 *) Permit null encryption ciphersuites, used for authentication only. It used
7744 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7745 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7749 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7750 signed attributes when verifying signatures (this would break them),
7751 the detached data encoding was wrong and public keys obtained using
7752 X509_get_pubkey() weren't freed.
7755 *) Add text documentation for the BUFFER functions. Also added a work around
7756 to a Win95 console bug. This was triggered by the password read stuff: the
7757 last character typed gets carried over to the next fread(). If you were
7758 generating a new cert request using 'req' for example then the last
7759 character of the passphrase would be CR which would then enter the first
7763 *) Added the new `Includes OpenSSL Cryptography Software' button as
7764 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7765 button and can be used by applications based on OpenSSL to show the
7766 relationship to the OpenSSL project.
7767 [Ralf S. Engelschall]
7769 *) Remove confusing variables in function signatures in files
7770 ssl/ssl_lib.c and ssl/ssl.h.
7771 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7773 *) Don't install bss_file.c under PREFIX/include/
7774 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7776 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7777 functions that return function pointers and has support for NT specific
7778 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7779 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7780 unsigned to signed types: this was killing the Win32 compile.
7783 *) Add new certificate file to stack functions,
7784 SSL_add_dir_cert_subjects_to_stack() and
7785 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7786 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7787 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7788 This means that Apache-SSL and similar packages don't have to mess around
7789 to add as many CAs as they want to the preferred list.
7792 *) Experiment with doxygen documentation. Currently only partially applied to
7794 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7795 openssl.doxy as the configuration file.
7798 *) Get rid of remaining C++-style comments which strict C compilers hate.
7799 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7801 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7802 compiled in by default: it has problems with large keys.
7805 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7806 DH private keys and/or callback functions which directly correspond to
7807 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7808 is needed for applications which have to configure certificates on a
7809 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7811 For the RSA certificate situation is makes no difference, but
7812 for the DSA certificate situation this fixes the "no shared cipher"
7813 problem where the OpenSSL cipher selection procedure failed because the
7814 temporary keys were not overtaken from the context and the API provided
7815 no way to reconfigure them.
7816 The new functions now let applications reconfigure the stuff and they
7817 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7818 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7819 non-public-API function ssl_cert_instantiate() is used as a helper
7820 function and also to reduce code redundancy inside ssl_rsa.c.
7821 [Ralf S. Engelschall]
7823 *) Move s_server -dcert and -dkey options out of the undocumented feature
7824 area because they are useful for the DSA situation and should be
7825 recognized by the users.
7826 [Ralf S. Engelschall]
7828 *) Fix the cipher decision scheme for export ciphers: the export bits are
7829 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7830 SSL_EXP_MASK. So, the original variable has to be used instead of the
7831 already masked variable.
7832 [Richard Levitte <levitte@stacken.kth.se>]
7834 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7835 [Richard Levitte <levitte@stacken.kth.se>]
7837 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7838 from `int' to `unsigned int' because it's a length and initialized by
7839 EVP_DigestFinal() which expects an `unsigned int *'.
7840 [Richard Levitte <levitte@stacken.kth.se>]
7842 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7843 script. Instead use the usual Shell->Perl transition trick.
7844 [Ralf S. Engelschall]
7846 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7847 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7848 -noout -modulus' as it's already the case for `openssl rsa -noout
7849 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7850 currently the public key is printed (a decision which was already done by
7851 `openssl dsa -modulus' in the past) which serves a similar purpose.
7852 Additionally the NO_RSA no longer completely removes the whole -modulus
7853 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7855 [Ralf S. Engelschall]
7857 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7858 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7859 [Arne Ansper <arne@ats.cyber.ee>]
7861 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7862 to be added. Now both 'req' and 'ca' can use new objects defined in the
7866 *) Add cool BIO that does syslog (or event log on NT).
7867 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7869 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7870 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7871 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7872 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7875 *) Add preliminary config info for new extension code.
7878 *) Make RSA_NO_PADDING really use no padding.
7879 [Ulf Moeller <ulf@fitug.de>]
7881 *) Generate errors when private/public key check is done.
7884 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7885 for some CRL extensions and new objects added.
7888 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7889 key usage extension and fuller support for authority key id.
7892 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7893 padding method for RSA, which is recommended for new applications in PKCS
7894 #1 v2.0 (RFC 2437, October 1998).
7895 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7896 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7897 against Bleichbacher's attack on RSA.
7898 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7901 *) Updates to the new SSL compression code
7902 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7904 *) Fix so that the version number in the master secret, when passed
7905 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7906 (because the server will not accept higher), that the version number
7907 is 0x03,0x01, not 0x03,0x00
7908 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7910 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7911 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7912 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7915 *) Support for RAW extensions where an arbitrary extension can be
7916 created by including its DER encoding. See apps/openssl.cnf for
7920 *) Make sure latest Perl versions don't interpret some generated C array
7921 code as Perl array code in the crypto/err/err_genc.pl script.
7922 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7924 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7925 not many people have the assembler. Various Win32 compilation fixes and
7926 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7930 *) Modify configure script 'Configure' to automatically create crypto/date.h
7931 file under Win32 and also build pem.h from pem.org. New script
7932 util/mkfiles.pl to create the MINFO file on environments that can't do a
7933 'make files': perl util/mkfiles.pl >MINFO should work.
7936 *) Major rework of DES function declarations, in the pursuit of correctness
7937 and purity. As a result, many evil casts evaporated, and some weirdness,
7938 too. You may find this causes warnings in your code. Zapping your evil
7939 casts will probably fix them. Mostly.
7942 *) Fix for a typo in asn1.h. Bug fix to object creation script
7943 obj_dat.pl. It considered a zero in an object definition to mean
7944 "end of object": none of the objects in objects.h have any zeros
7945 so it wasn't spotted.
7946 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7948 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7949 Masking (CBCM). In the absence of test vectors, the best I have been able
7950 to do is check that the decrypt undoes the encrypt, so far. Send me test
7951 vectors if you have them.
7954 *) Correct calculation of key length for export ciphers (too much space was
7955 allocated for null ciphers). This has not been tested!
7958 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7959 message is now correct (it understands "crypto" and "ssl" on its
7960 command line). There is also now an "update" option. This will update
7961 the util/ssleay.num and util/libeay.num files with any new functions.
7963 perl util/mkdef.pl crypto ssl update
7964 it will update them.
7967 *) Overhauled the Perl interface (perl/*):
7968 - ported BN stuff to OpenSSL's different BN library
7969 - made the perl/ source tree CVS-aware
7970 - renamed the package from SSLeay to OpenSSL (the files still contain
7971 their history because I've copied them in the repository)
7972 - removed obsolete files (the test scripts will be replaced
7973 by better Test::Harness variants in the future)
7974 [Ralf S. Engelschall]
7976 *) First cut for a very conservative source tree cleanup:
7977 1. merge various obsolete readme texts into doc/ssleay.txt
7978 where we collect the old documents and readme texts.
7979 2. remove the first part of files where I'm already sure that we no
7980 longer need them because of three reasons: either they are just temporary
7981 files which were left by Eric or they are preserved original files where
7982 I've verified that the diff is also available in the CVS via "cvs diff
7983 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7984 the crypto/md/ stuff).
7985 [Ralf S. Engelschall]
7987 *) More extension code. Incomplete support for subject and issuer alt
7988 name, issuer and authority key id. Change the i2v function parameters
7989 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7990 what that's for :-) Fix to ASN1 macro which messed up
7991 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7994 *) Preliminary support for ENUMERATED type. This is largely copied from the
7998 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7999 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8001 *) Make sure `make rehash' target really finds the `openssl' program.
8002 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8004 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8005 like to hear about it if this slows down other processors.
8008 *) Add CygWin32 platform information to Configure script.
8009 [Alan Batie <batie@aahz.jf.intel.com>]
8011 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8012 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8014 *) New program nseq to manipulate netscape certificate sequences
8017 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8021 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8022 but the BN code had some problems that would cause failures when
8023 doing certificate verification and some other functions.
8024 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8026 *) Add ASN1 and PEM code to support netscape certificate sequences.
8029 *) Add ASN1 and PEM code to support netscape certificate sequences.
8032 *) Add several PKIX and private extended key usage OIDs.
8035 *) Modify the 'ca' program to handle the new extension code. Modify
8036 openssl.cnf for new extension format, add comments.
8039 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8040 and add a sample to openssl.cnf so req -x509 now adds appropriate
8044 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8045 error code, add initial support to X509_print() and x509 application.
8048 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8049 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8050 stuff is currently isolated and isn't even compiled yet.
8053 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8054 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8055 Removed the versions check from X509 routines when loading extensions:
8056 this allows certain broken certificates that don't set the version
8057 properly to be processed.
8060 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8061 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8062 can still be regenerated with "make depend".
8065 *) Spelling mistake in C version of CAST-128.
8066 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8068 *) Changes to the error generation code. The perl script err-code.pl
8069 now reads in the old error codes and retains the old numbers, only
8070 adding new ones if necessary. It also only changes the .err files if new
8071 codes are added. The makefiles have been modified to only insert errors
8072 when needed (to avoid needlessly modifying header files). This is done
8073 by only inserting errors if the .err file is newer than the auto generated
8074 C file. To rebuild all the error codes from scratch (the old behaviour)
8075 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8076 or delete all the .err files.
8079 *) CAST-128 was incorrectly implemented for short keys. The C version has
8080 been fixed, but is untested. The assembler versions are also fixed, but
8081 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8082 to regenerate it if needed.
8083 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8084 Hagino <itojun@kame.net>]
8086 *) File was opened incorrectly in randfile.c.
8087 [Ulf Möller <ulf@fitug.de>]
8089 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8090 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8091 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8092 al: it's just almost always a UTCTime. Note this patch adds new error
8093 codes so do a "make errors" if there are problems.
8096 *) Correct Linux 1 recognition in config.
8097 [Ulf Möller <ulf@fitug.de>]
8099 *) Remove pointless MD5 hash when using DSA keys in ca.
8100 [Anonymous <nobody@replay.com>]
8102 *) Generate an error if given an empty string as a cert directory. Also
8103 generate an error if handed NULL (previously returned 0 to indicate an
8104 error, but didn't set one).
8105 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8107 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8110 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8111 parameters. This was causing a warning which killed off the Win32 compile.
8114 *) Remove C++ style comments from crypto/bn/bn_local.h.
8115 [Neil Costigan <neil.costigan@celocom.com>]
8117 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8118 based on a text string, looking up short and long names and finally
8119 "dot" format. The "dot" format stuff didn't work. Added new function
8120 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8121 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8122 OID is not part of the table.
8125 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8126 X509_LOOKUP_by_alias().
8129 *) Sort openssl functions by name.
8132 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8133 encryption from sample DSA keys (in case anyone is interested the password
8137 *) Make _all_ *_free functions accept a NULL pointer.
8138 [Frans Heymans <fheymans@isaserver.be>]
8140 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8142 [Anonymous <nobody@replay.com>]
8144 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8145 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8147 *) Don't blow it for numeric -newkey arguments to apps/req.
8148 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8150 *) Temp key "for export" tests were wrong in s3_srvr.c.
8151 [Anonymous <nobody@replay.com>]
8153 *) Add prototype for temp key callback functions
8154 SSL_CTX_set_tmp_{rsa,dh}_callback().
8157 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8158 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8161 *) X509_name_add_entry() freed the wrong thing after an error.
8162 [Arne Ansper <arne@ats.cyber.ee>]
8164 *) rsa_eay.c would attempt to free a NULL context.
8165 [Arne Ansper <arne@ats.cyber.ee>]
8167 *) BIO_s_socket() had a broken should_retry() on Windoze.
8168 [Arne Ansper <arne@ats.cyber.ee>]
8170 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8171 [Arne Ansper <arne@ats.cyber.ee>]
8173 *) Make sure the already existing X509_STORE->depth variable is initialized
8174 in X509_STORE_new(), but document the fact that this variable is still
8175 unused in the certificate verification process.
8176 [Ralf S. Engelschall]
8178 *) Fix the various library and apps files to free up pkeys obtained from
8179 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8182 *) Fix reference counting in X509_PUBKEY_get(). This makes
8183 demos/maurice/example2.c work, amongst others, probably.
8184 [Steve Henson and Ben Laurie]
8186 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8187 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8188 are no longer created. This way we have a single and consistent command
8189 line interface `openssl <command>', similar to `cvs <command>'.
8190 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8192 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8193 BIT STRING wrapper always have zero unused bits.
8196 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8199 *) Make the top-level INSTALL documentation easier to understand.
8202 *) Makefiles updated to exit if an error occurs in a sub-directory
8203 make (including if user presses ^C) [Paul Sutton]
8205 *) Make Montgomery context stuff explicit in RSA data structure.
8208 *) Fix build order of pem and err to allow for generated pem.h.
8211 *) Fix renumbering bug in X509_NAME_delete_entry().
8214 *) Enhanced the err-ins.pl script so it makes the error library number
8215 global and can add a library name. This is needed for external ASN1 and
8216 other error libraries.
8219 *) Fixed sk_insert which never worked properly.
8222 *) Fix ASN1 macros so they can handle indefinite length construted
8223 EXPLICIT tags. Some non standard certificates use these: they can now
8227 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8228 into a single doc/ssleay.txt bundle. This way the information is still
8229 preserved but no longer messes up this directory. Now it's new room for
8230 the new set of documenation files.
8231 [Ralf S. Engelschall]
8233 *) SETs were incorrectly DER encoded. This was a major pain, because they
8234 shared code with SEQUENCEs, which aren't coded the same. This means that
8235 almost everything to do with SETs or SEQUENCEs has either changed name or
8236 number of arguments.
8237 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8239 *) Fix test data to work with the above.
8242 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8243 was already fixed by Eric for 0.9.1 it seems.
8244 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8246 *) Autodetect FreeBSD3.
8249 *) Fix various bugs in Configure. This affects the following platforms:
8253 unixware-2.0-pentium
8257 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8258 before they are needed.
8261 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8265 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8267 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8268 changed SSLeay to OpenSSL in version strings.
8269 [Ralf S. Engelschall]
8271 *) Some fixups to the top-level documents.
8274 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8275 because the symlink to include/ was missing.
8276 [Ralf S. Engelschall]
8278 *) Incorporated the popular no-RSA/DSA-only patches
8279 which allow to compile a RSA-free SSLeay.
8280 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8282 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8283 when "ssleay" is still not found.
8284 [Ralf S. Engelschall]
8286 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8287 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8289 *) Updated the README file.
8290 [Ralf S. Engelschall]
8292 *) Added various .cvsignore files in the CVS repository subdirs
8293 to make a "cvs update" really silent.
8294 [Ralf S. Engelschall]
8296 *) Recompiled the error-definition header files and added
8297 missing symbols to the Win32 linker tables.
8298 [Ralf S. Engelschall]
8300 *) Cleaned up the top-level documents;
8301 o new files: CHANGES and LICENSE
8302 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8303 o merged COPYRIGHT into LICENSE
8304 o removed obsolete TODO file
8305 o renamed MICROSOFT to INSTALL.W32
8306 [Ralf S. Engelschall]
8308 *) Removed dummy files from the 0.9.1b source tree:
8309 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8310 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8311 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8312 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8313 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8314 [Ralf S. Engelschall]
8316 *) Added various platform portability fixes.
8319 *) The Genesis of the OpenSSL rpject:
8320 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8321 Young and Tim J. Hudson created while they were working for C2Net until
8323 [The OpenSSL Project]
8326 Changes between 0.9.0b and 0.9.1b [not released]
8328 *) Updated a few CA certificates under certs/
8331 *) Changed some BIGNUM api stuff.
8334 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8335 DGUX x86, Linux Alpha, etc.
8338 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8339 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8343 *) Add -strparse option to asn1pars program which parses nested
8345 [Dr Stephen Henson <shenson@bigfoot.com>]
8347 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8350 *) DSA fix for "ca" program.
8353 *) Added "-genkey" option to "dsaparam" program.
8356 *) Added RIPE MD160 (rmd160) message digest.
8359 *) Added -a (all) option to "ssleay version" command.
8362 *) Added PLATFORM define which is the id given to Configure.
8365 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8368 *) Extended the ASN.1 parser routines.
8371 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8374 *) Added a BN_CTX to the BN library.
8377 *) Fixed the weak key values in DES library
8380 *) Changed API in EVP library for cipher aliases.
8383 *) Added support for RC2/64bit cipher.
8386 *) Converted the lhash library to the crypto/mem.c functions.
8389 *) Added more recognized ASN.1 object ids.
8392 *) Added more RSA padding checks for SSL/TLS.
8395 *) Added BIO proxy/filter functionality.
8398 *) Added extra_certs to SSL_CTX which can be used
8399 send extra CA certificates to the client in the CA cert chain sending
8400 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8403 *) Now Fortezza is denied in the authentication phase because
8404 this is key exchange mechanism is not supported by SSLeay at all.
8407 *) Additional PKCS1 checks.
8410 *) Support the string "TLSv1" for all TLS v1 ciphers.
8413 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8414 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8417 *) Fixed a few memory leaks.
8420 *) Fixed various code and comment typos.
8423 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8424 bytes sent in the client random.
8425 [Edward Bishop <ebishop@spyglass.com>]