5 Changes between 0.9.8l (?) and 0.9.8m (?) [xx XXX xxxx]
7 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8 connect (but not renegotiate) with servers which do not support RI.
9 Until RI is more widely deployed this option is enabled by default.
12 *) Add "missing" ssl ctrls to clear options and mode.
15 *) If client attempts to renegotiate and doesn't support RI respond with
16 a no_renegotiation alert as required by draft-ietf-tls-renegotiation.
17 Some renegotiating TLS clients will continue a connection gracefully
18 when they receive the alert. Unfortunately OpenSSL mishandled
19 this alert and would hang waiting for a server hello which it will never
20 receive. Now we treat a received no_renegotiation alert as a fatal
21 error. This is because applications requesting a renegotiation might well
22 expect it to succeed and would have no code in place to handle the server
23 denying it so the only safe thing to do is to terminate the connection.
26 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
27 peer supports secure renegotiation and 0 otherwise. Print out peer
28 renegotiation support in s_client/s_server.
31 *) Replace the highly broken and deprecated SPKAC certification method with
32 the updated NID creation version. This should correctly handle UTF8.
35 *) Implement draft-ietf-tls-renegotiation. Re-enable
36 renegotiation but require the extension as needed. Unfortunately,
37 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION turns out to be a
38 bad idea. It has been replaced by
39 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
40 SSL_CTX_set_options(). This is really not recommended unless you
41 know what you are doing.
42 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
44 *) Fixes to stateless session resumption handling. Use initial_ctx when
45 issuing and attempting to decrypt tickets in case it has changed during
46 servername handling. Use a non-zero length session ID when attempting
47 stateless session resumption: this makes it possible to determine if
48 a resumption has occurred immediately after receiving server hello
49 (several places in OpenSSL subtly assume this) instead of later in
53 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
54 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
55 fixes for a few places where the return code is not checked
57 [Julia Lawall <julia@diku.dk>]
59 *) Add --strict-warnings option to Configure script to include devteam
60 warnings in other configurations.
63 *) Add support for --libdir option and LIBDIR variable in makefiles. This
64 makes it possible to install openssl libraries in locations which
65 have names other than "lib", for example "/usr/lib64" which some
67 [Steve Henson, based on patch from Jeremy Utley]
69 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
70 X690 8.9.12 and can produce some misleading textual output of OIDs.
71 [Steve Henson, reported by Dan Kaminsky]
73 *) Delete MD2 from algorithm tables. This follows the recommendation in
74 several standards that it is not used in new applications due to
75 several cryptographic weaknesses. For binary compatibility reasons
76 the MD2 API is still compiled in by default.
79 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
83 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
84 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
86 [Guenter <lists@gknw.net>]
88 *) Fix the server certificate chain building code to use X509_verify_cert(),
89 it used to have an ad-hoc builder which was unable to cope with anything
90 other than a simple chain.
91 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
93 *) Don't check self signed certificate signatures in X509_verify_cert()
94 by default (a flag can override this): it just wastes time without
95 adding any security. As a useful side effect self signed root CAs
96 with non-FIPS digests are now usable in FIPS mode.
99 *) In dtls1_process_out_of_seq_message() the check if the current message
100 is already buffered was missing. For every new message was memory
101 allocated, allowing an attacker to perform an denial of service attack
102 with sending out of seq handshake messages until there is no memory
103 left. Additionally every future messege was buffered, even if the
104 sequence number made no sense and would be part of another handshake.
105 So only messages with sequence numbers less than 10 in advance will be
106 buffered. (CVE-2009-1378)
107 [Robin Seggelmann, discovered by Daniel Mentz]
109 *) Records are buffered if they arrive with a future epoch to be
110 processed after finishing the corresponding handshake. There is
111 currently no limitation to this buffer allowing an attacker to perform
112 a DOS attack with sending records with future epochs until there is no
113 memory left. This patch adds the pqueue_size() function to detemine
114 the size of a buffer and limits the record buffer to 100 entries.
116 [Robin Seggelmann, discovered by Daniel Mentz]
118 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
119 parent structure is freed. (CVE-2009-1379)
122 *) Handle non-blocking I/O properly in SSL_shutdown() call.
123 [Darryl Miles <darryl-mailinglists@netbauds.net>]
126 [Ilya O. <vrghost@gmail.com>]
128 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
130 *) Disable renegotiation completely - this fixes a severe security
131 problem (CVE-2009-3555) at the cost of breaking all
132 renegotiation. Renegotiation can be re-enabled by setting
133 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
134 run-time. This is really not recommended unless you know what
138 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
140 *) Don't set val to NULL when freeing up structures, it is freed up by
141 underlying code. If sizeof(void *) > sizeof(long) this can result in
142 zeroing past the valid field. (CVE-2009-0789)
143 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
145 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
146 checked correctly. This would allow some invalid signed attributes to
147 appear to verify correctly. (CVE-2009-0591)
148 [Ivan Nestlerode <inestlerode@us.ibm.com>]
150 *) Reject UniversalString and BMPString types with invalid lengths. This
151 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
152 a legal length. (CVE-2009-0590)
155 *) Set S/MIME signing as the default purpose rather than setting it
156 unconditionally. This allows applications to override it at the store
160 *) Permit restricted recursion of ASN1 strings. This is needed in practice
161 to handle some structures.
164 *) Improve efficiency of mem_gets: don't search whole buffer each time
166 [Jeremy Shapiro <jnshapir@us.ibm.com>]
168 *) New -hex option for openssl rand.
171 *) Print out UTF8String and NumericString when parsing ASN1.
174 *) Support NumericString type for name components.
177 *) Allow CC in the environment to override the automatically chosen
178 compiler. Note that nothing is done to ensure flags work with the
182 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
184 *) Properly check EVP_VerifyFinal() and similar return values
186 [Ben Laurie, Bodo Moeller, Google Security Team]
188 *) Enable TLS extensions by default.
191 *) Allow the CHIL engine to be loaded, whether the application is
192 multithreaded or not. (This does not release the developer from the
193 obligation to set up the dynamic locking callbacks.)
194 [Sander Temme <sander@temme.net>]
196 *) Use correct exit code if there is an error in dgst command.
197 [Steve Henson; problem pointed out by Roland Dirlewanger]
199 *) Tweak Configure so that you need to say "experimental-jpake" to enable
200 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
203 *) Add experimental JPAKE support, including demo authentication in
204 s_client and s_server.
207 *) Set the comparison function in v3_addr_canonize().
208 [Rob Austein <sra@hactrn.net>]
210 *) Add support for XMPP STARTTLS in s_client.
211 [Philip Paeps <philip@freebsd.org>]
213 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
214 to ensure that even with this option, only ciphersuites in the
215 server's preference list will be accepted. (Note that the option
216 applies only when resuming a session, so the earlier behavior was
217 just about the algorithm choice for symmetric cryptography.)
220 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
222 *) Fix NULL pointer dereference if a DTLS server received
223 ChangeCipherSpec as first record (CVE-2009-1386).
226 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
227 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
230 *) The fix in 0.9.8c that supposedly got rid of unsafe
231 double-checked locking was incomplete for RSA blinding,
232 addressing just one layer of what turns out to have been
233 doubly unsafe triple-checked locking.
235 So now fix this for real by retiring the MONT_HELPER macro
236 in crypto/rsa/rsa_eay.c.
238 [Bodo Moeller; problem pointed out by Marius Schilder]
240 *) Various precautionary measures:
242 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
244 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
245 (NB: This would require knowledge of the secret session ticket key
246 to exploit, in which case you'd be SOL either way.)
248 - Change bn_nist.c so that it will properly handle input BIGNUMs
249 outside the expected range.
251 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
254 [Neel Mehta, Bodo Moeller]
256 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
257 the load fails. Useful for distros.
258 [Ben Laurie and the FreeBSD team]
260 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
263 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
266 *) Expand ENGINE to support engine supplied SSL client certificate functions.
268 This work was sponsored by Logica.
271 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
272 keystores. Support for SSL/TLS client authentication too.
273 Not compiled unless enable-capieng specified to Configure.
275 This work was sponsored by Logica.
278 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
279 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
280 attribute creation routines such as certifcate requests and PKCS#12
284 Changes between 0.9.8g and 0.9.8h [28 May 2008]
286 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
287 handshake which could lead to a cilent crash as found using the
288 Codenomicon TLS test suite (CVE-2008-1672)
289 [Steve Henson, Mark Cox]
291 *) Fix double free in TLS server name extensions which could lead to
292 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
295 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
297 Clear the error queue to ensure that error entries left from
298 older function calls do not interfere with the correct operation.
299 [Lutz Jaenicke, Erik de Castro Lopo]
301 *) Remove root CA certificates of commercial CAs:
303 The OpenSSL project does not recommend any specific CA and does not
304 have any policy with respect to including or excluding any CA.
305 Therefore it does not make any sense to ship an arbitrary selection
306 of root CA certificates with the OpenSSL software.
309 *) RSA OAEP patches to fix two separate invalid memory reads.
310 The first one involves inputs when 'lzero' is greater than
311 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
312 before the beginning of from). The second one involves inputs where
313 the 'db' section contains nothing but zeroes (there is a one-byte
314 invalid read after the end of 'db').
315 [Ivan Nestlerode <inestlerode@us.ibm.com>]
317 *) Partial backport from 0.9.9-dev:
319 Introduce bn_mul_mont (dedicated Montgomery multiplication
320 procedure) as a candidate for BIGNUM assembler implementation.
321 While 0.9.9-dev uses assembler for various architectures, only
322 x86_64 is available by default here in the 0.9.8 branch, and
323 32-bit x86 is available through a compile-time setting.
325 To try the 32-bit x86 assembler implementation, use Configure
326 option "enable-montasm" (which exists only for this backport).
328 As "enable-montasm" for 32-bit x86 disclaims code stability
329 anyway, in this constellation we activate additional code
330 backported from 0.9.9-dev for further performance improvements,
331 namely BN_from_montgomery_word. (To enable this otherwise,
332 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
334 [Andy Polyakov (backport partially by Bodo Moeller)]
336 *) Add TLS session ticket callback. This allows an application to set
337 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
338 values. This is useful for key rollover for example where several key
339 sets may exist with different names.
342 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
343 This was broken until now in 0.9.8 releases, such that the only way
344 a registered ENGINE could be used (assuming it initialises
345 successfully on the host) was to explicitly set it as the default
346 for the relevant algorithms. This is in contradiction with 0.9.7
347 behaviour and the documentation. With this fix, when an ENGINE is
348 registered into a given algorithm's table of implementations, the
349 'uptodate' flag is reset so that auto-discovery will be used next
350 time a new context for that algorithm attempts to select an
352 [Ian Lister (tweaked by Geoff Thorpe)]
354 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
355 implemention in the following ways:
357 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
360 Lack of BER streaming support means one pass streaming processing is
361 only supported if data is detached: setting the streaming flag is
362 ignored for embedded content.
364 CMS support is disabled by default and must be explicitly enabled
365 with the enable-cms configuration option.
368 *) Update the GMP engine glue to do direct copies between BIGNUM and
369 mpz_t when openssl and GMP use the same limb size. Otherwise the
370 existing "conversion via a text string export" trick is still used.
371 [Paul Sheer <paulsheer@gmail.com>]
373 *) Zlib compression BIO. This is a filter BIO which compressed and
374 uncompresses any data passed through it.
377 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
378 RFC3394 compatible AES key wrapping.
381 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
382 sets string data without copying. X509_ALGOR_set0() and
383 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
384 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
385 from an X509_ATTRIBUTE structure optionally checking it occurs only
386 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
390 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
391 to get the expected BN_FLG_CONSTTIME behavior.
392 [Bodo Moeller (Google)]
396 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
397 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
398 - added some more tests to do_tests.pl
399 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
400 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
401 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
402 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
403 - various changes to netware.pl to enable gcc-cross builds on Win32
405 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
406 - various changes to fix missing prototype warnings
407 - fixed x86nasm.pl to create correct asm files for NASM COFF output
408 - added AES, WHIRLPOOL and CPUID assembler code to build files
409 - added missing AES assembler make rules to mk1mf.pl
410 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
411 [Guenter Knauf <eflash@gmx.net>]
413 *) Implement certificate status request TLS extension defined in RFC3546.
414 A client can set the appropriate parameters and receive the encoded
415 OCSP response via a callback. A server can query the supplied parameters
416 and set the encoded OCSP response in the callback. Add simplified examples
417 to s_client and s_server.
420 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
423 + Binary incompatibility of ssl_ctx_st structure
424 + DTLS interoperation with non-compliant servers
425 + Don't call get_session_cb() without proposed session
426 + Fix ia64 assembler code
427 [Andy Polyakov, Steve Henson]
429 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
431 *) DTLS Handshake overhaul. There were longstanding issues with
432 OpenSSL DTLS implementation, which were making it impossible for
433 RFC 4347 compliant client to communicate with OpenSSL server.
434 Unfortunately just fixing these incompatibilities would "cut off"
435 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
436 server keeps tolerating non RFC compliant syntax. The opposite is
437 not true, 0.9.8f client can not communicate with earlier server.
438 This update even addresses CVE-2007-4995.
441 *) Changes to avoid need for function casts in OpenSSL: some compilers
442 (gcc 4.2 and later) reject their use.
443 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
446 *) Add RFC4507 support to OpenSSL. This includes the corrections in
447 RFC4507bis. The encrypted ticket format is an encrypted encoded
448 SSL_SESSION structure, that way new session features are automatically
451 If a client application caches session in an SSL_SESSION structure
452 support is transparent because tickets are now stored in the encoded
455 The SSL_CTX structure automatically generates keys for ticket
456 protection in servers so again support should be possible
457 with no application modification.
459 If a client or server wishes to disable RFC4507 support then the option
460 SSL_OP_NO_TICKET can be set.
462 Add a TLS extension debugging callback to allow the contents of any client
463 or server extensions to be examined.
465 This work was sponsored by Google.
468 *) Add initial support for TLS extensions, specifically for the server_name
469 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
470 have new members for a host name. The SSL data structure has an
471 additional member SSL_CTX *initial_ctx so that new sessions can be
472 stored in that context to allow for session resumption, even after the
473 SSL has been switched to a new SSL_CTX in reaction to a client's
474 server_name extension.
476 New functions (subject to change):
479 SSL_get_servername_type()
482 New CTRL codes and macros (subject to change):
484 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
485 - SSL_CTX_set_tlsext_servername_callback()
486 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
487 - SSL_CTX_set_tlsext_servername_arg()
488 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
490 openssl s_client has a new '-servername ...' option.
492 openssl s_server has new options '-servername_host ...', '-cert2 ...',
493 '-key2 ...', '-servername_fatal' (subject to change). This allows
494 testing the HostName extension for a specific single host name ('-cert'
495 and '-key' remain fallbacks for handshakes without HostName
496 negotiation). If the unrecogninzed_name alert has to be sent, this by
497 default is a warning; it becomes fatal with the '-servername_fatal'
500 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
502 *) Add AES and SSE2 assembly language support to VC++ build.
505 *) Mitigate attack on final subtraction in Montgomery reduction.
508 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
509 (which previously caused an internal error).
512 *) Squeeze another 10% out of IGE mode when in != out.
515 *) AES IGE mode speedup.
516 [Dean Gaudet (Google)]
518 *) Add the Korean symmetric 128-bit cipher SEED (see
519 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
520 add SEED ciphersuites from RFC 4162:
522 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
523 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
524 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
525 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
527 To minimize changes between patchlevels in the OpenSSL 0.9.8
528 series, SEED remains excluded from compilation unless OpenSSL
529 is configured with 'enable-seed'.
532 *) Mitigate branch prediction attacks, which can be practical if a
533 single processor is shared, allowing a spy process to extract
534 information. For detailed background information, see
535 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
536 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
537 and Necessary Software Countermeasures"). The core of the change
538 are new versions BN_div_no_branch() and
539 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
540 respectively, which are slower, but avoid the security-relevant
541 conditional branches. These are automatically called by BN_div()
542 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
543 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
544 remove a conditional branch.
546 BN_FLG_CONSTTIME is the new name for the previous
547 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
548 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
549 in the exponent causes BN_mod_exp_mont() to use the alternative
550 implementation in BN_mod_exp_mont_consttime().) The old name
551 remains as a deprecated alias.
553 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
554 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
555 constant-time implementations for more than just exponentiation.
556 Here too the old name is kept as a deprecated alias.
558 BN_BLINDING_new() will now use BN_dup() for the modulus so that
559 the BN_BLINDING structure gets an independent copy of the
560 modulus. This means that the previous "BIGNUM *m" argument to
561 BN_BLINDING_new() and to BN_BLINDING_create_param() now
562 essentially becomes "const BIGNUM *m", although we can't actually
563 change this in the header file before 0.9.9. It allows
564 RSA_setup_blinding() to use BN_with_flags() on the modulus to
565 enable BN_FLG_CONSTTIME.
567 [Matthew D Wood (Intel Corp)]
569 *) In the SSL/TLS server implementation, be strict about session ID
570 context matching (which matters if an application uses a single
571 external cache for different purposes). Previously,
572 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
573 set. This did ensure strict client verification, but meant that,
574 with applications using a single external cache for quite
575 different requirements, clients could circumvent ciphersuite
576 restrictions for a given session ID context by starting a session
577 in a different context.
580 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
581 a ciphersuite string such as "DEFAULT:RSA" cannot enable
582 authentication-only ciphersuites.
585 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
586 not complete and could lead to a possible single byte overflow
587 (CVE-2007-5135) [Ben Laurie]
589 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
591 *) Since AES128 and AES256 (and similarly Camellia128 and
592 Camellia256) share a single mask bit in the logic of
593 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
594 kludge to work properly if AES128 is available and AES256 isn't
595 (or if Camellia128 is available and Camellia256 isn't).
598 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
599 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
600 When a point or a seed is encoded in a BIT STRING, we need to
601 prevent the removal of trailing zero bits to get the proper DER
602 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
603 of a NamedBitList, for which trailing 0 bits need to be removed.)
606 *) Have SSL/TLS server implementation tolerate "mismatched" record
607 protocol version while receiving ClientHello even if the
608 ClientHello is fragmented. (The server can't insist on the
609 particular protocol version it has chosen before the ServerHello
610 message has informed the client about his choice.)
613 *) Add RFC 3779 support.
614 [Rob Austein for ARIN, Ben Laurie]
616 *) Load error codes if they are not already present instead of using a
617 static variable. This allows them to be cleanly unloaded and reloaded.
618 Improve header file function name parsing.
621 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
622 or CAPABILITY handshake as required by RFCs.
625 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
627 *) Introduce limits to prevent malicious keys being able to
628 cause a denial of service. (CVE-2006-2940)
629 [Steve Henson, Bodo Moeller]
631 *) Fix ASN.1 parsing of certain invalid structures that can result
632 in a denial of service. (CVE-2006-2937) [Steve Henson]
634 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
635 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
637 *) Fix SSL client code which could crash if connecting to a
638 malicious SSLv2 server. (CVE-2006-4343)
639 [Tavis Ormandy and Will Drewry, Google Security Team]
641 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
642 match only those. Before that, "AES256-SHA" would be interpreted
643 as a pattern and match "AES128-SHA" too (since AES128-SHA got
644 the same strength classification in 0.9.7h) as we currently only
645 have a single AES bit in the ciphersuite description bitmap.
646 That change, however, also applied to ciphersuite strings such as
647 "RC4-MD5" that intentionally matched multiple ciphersuites --
648 namely, SSL 2.0 ciphersuites in addition to the more common ones
649 from SSL 3.0/TLS 1.0.
651 So we change the selection algorithm again: Naming an explicit
652 ciphersuite selects this one ciphersuite, and any other similar
653 ciphersuite (same bitmap) from *other* protocol versions.
654 Thus, "RC4-MD5" again will properly select both the SSL 2.0
655 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
657 Since SSL 2.0 does not have any ciphersuites for which the
658 128/256 bit distinction would be relevant, this works for now.
659 The proper fix will be to use different bits for AES128 and
660 AES256, which would have avoided the problems from the beginning;
661 however, bits are scarce, so we can only do this in a new release
662 (not just a patchlevel) when we can change the SSL_CIPHER
663 definition to split the single 'unsigned long mask' bitmap into
664 multiple values to extend the available space.
668 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
670 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
671 (CVE-2006-4339) [Ben Laurie and Google Security Team]
673 *) Add AES IGE and biIGE modes.
676 *) Change the Unix randomness entropy gathering to use poll() when
677 possible instead of select(), since the latter has some
678 undesirable limitations.
679 [Darryl Miles via Richard Levitte and Bodo Moeller]
681 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
682 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
683 cannot be implicitly activated as part of, e.g., the "AES" alias.
684 However, please upgrade to OpenSSL 0.9.9[-dev] for
685 non-experimental use of the ECC ciphersuites to get TLS extension
686 support, which is required for curve and point format negotiation
687 to avoid potential handshake problems.
690 *) Disable rogue ciphersuites:
692 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
693 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
694 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
696 The latter two were purportedly from
697 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
700 Also deactivate the remaining ciphersuites from
701 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
702 unofficial, and the ID has long expired.
705 *) Fix RSA blinding Heisenbug (problems sometimes occured on
706 dual-core machines) and other potential thread-safety issues.
709 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
710 versions), which is now available for royalty-free use
711 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
712 Also, add Camellia TLS ciphersuites from RFC 4132.
714 To minimize changes between patchlevels in the OpenSSL 0.9.8
715 series, Camellia remains excluded from compilation unless OpenSSL
716 is configured with 'enable-camellia'.
719 *) Disable the padding bug check when compression is in use. The padding
720 bug check assumes the first packet is of even length, this is not
721 necessarily true if compresssion is enabled and can result in false
722 positives causing handshake failure. The actual bug test is ancient
723 code so it is hoped that implementations will either have fixed it by
724 now or any which still have the bug do not support compression.
727 Changes between 0.9.8a and 0.9.8b [04 May 2006]
729 *) When applying a cipher rule check to see if string match is an explicit
730 cipher suite and only match that one cipher suite if it is.
733 *) Link in manifests for VC++ if needed.
734 [Austin Ziegler <halostatue@gmail.com>]
736 *) Update support for ECC-based TLS ciphersuites according to
737 draft-ietf-tls-ecc-12.txt with proposed changes (but without
738 TLS extensions, which are supported starting with the 0.9.9
739 branch, not in the OpenSSL 0.9.8 branch).
742 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
743 opaque EVP_CIPHER_CTX handling.
746 *) Fixes and enhancements to zlib compression code. We now only use
747 "zlib1.dll" and use the default __cdecl calling convention on Win32
748 to conform with the standards mentioned here:
749 http://www.zlib.net/DLL_FAQ.txt
750 Static zlib linking now works on Windows and the new --with-zlib-include
751 --with-zlib-lib options to Configure can be used to supply the location
752 of the headers and library. Gracefully handle case where zlib library
756 *) Several fixes and enhancements to the OID generation code. The old code
757 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
758 handle numbers larger than ULONG_MAX, truncated printing and had a
759 non standard OBJ_obj2txt() behaviour.
762 *) Add support for building of engines under engine/ as shared libraries
763 under VC++ build system.
766 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
767 Hopefully, we will not see any false combination of paths any more.
770 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
772 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
773 (part of SSL_OP_ALL). This option used to disable the
774 countermeasure against man-in-the-middle protocol-version
775 rollback in the SSL 2.0 server implementation, which is a bad
776 idea. (CVE-2005-2969)
778 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
779 for Information Security, National Institute of Advanced Industrial
780 Science and Technology [AIST], Japan)]
782 *) Add two function to clear and return the verify parameter flags.
785 *) Keep cipherlists sorted in the source instead of sorting them at
786 runtime, thus removing the need for a lock.
789 *) Avoid some small subgroup attacks in Diffie-Hellman.
790 [Nick Mathewson and Ben Laurie]
792 *) Add functions for well-known primes.
795 *) Extended Windows CE support.
796 [Satoshi Nakamura and Andy Polyakov]
798 *) Initialize SSL_METHOD structures at compile time instead of during
799 runtime, thus removing the need for a lock.
802 *) Make PKCS7_decrypt() work even if no certificate is supplied by
803 attempting to decrypt each encrypted key in turn. Add support to
807 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
809 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
812 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
815 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
816 key into the same file any more.
819 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
822 *) Add -utf8 command line and config file option to 'ca'.
823 [Stefan <stf@udoma.org]
825 *) Removed the macro des_crypt(), as it seems to conflict with some
826 libraries. Use DES_crypt().
829 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
830 involves renaming the source and generated shared-libs for
831 both. The engines will accept the corrected or legacy ids
832 ('ncipher' and '4758_cca' respectively) when binding. NB,
833 this only applies when building 'shared'.
834 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
836 *) Add attribute functions to EVP_PKEY structure. Modify
837 PKCS12_create() to recognize a CSP name attribute and
838 use it. Make -CSP option work again in pkcs12 utility.
841 *) Add new functionality to the bn blinding code:
842 - automatic re-creation of the BN_BLINDING parameters after
843 a fixed number of uses (currently 32)
844 - add new function for parameter creation
845 - introduce flags to control the update behaviour of the
846 BN_BLINDING parameters
847 - hide BN_BLINDING structure
848 Add a second BN_BLINDING slot to the RSA structure to improve
849 performance when a single RSA object is shared among several
853 *) Add support for DTLS.
854 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
856 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
857 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
860 *) Remove buggy and incompletet DH cert support from
861 ssl/ssl_rsa.c and ssl/s3_both.c
864 *) Use SHA-1 instead of MD5 as the default digest algorithm for
865 the apps/openssl applications.
868 *) Compile clean with "-Wall -Wmissing-prototypes
869 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
870 DEBUG_SAFESTACK must also be set.
873 *) Change ./Configure so that certain algorithms can be disabled by default.
874 The new counterpiece to "no-xxx" is "enable-xxx".
876 The patented RC5 and MDC2 algorithms will now be disabled unless
877 "enable-rc5" and "enable-mdc2", respectively, are specified.
879 (IDEA remains enabled despite being patented. This is because IDEA
880 is frequently required for interoperability, and there is no license
881 fee for non-commercial use. As before, "no-idea" can be used to
882 avoid this algorithm.)
886 *) Add processing of proxy certificates (see RFC 3820). This work was
887 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
888 EGEE (Enabling Grids for E-science in Europe).
891 *) RC4 performance overhaul on modern architectures/implementations, such
892 as Intel P4, IA-64 and AMD64.
895 *) New utility extract-section.pl. This can be used specify an alternative
896 section number in a pod file instead of having to treat each file as
897 a separate case in Makefile. This can be done by adding two lines to the
900 =for comment openssl_section:XXX
902 The blank line is mandatory.
906 *) New arguments -certform, -keyform and -pass for s_client and s_server
907 to allow alternative format key and certificate files and passphrase
911 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
912 update associated structures and add various utility functions.
914 Add new policy related verify parameters, include policy checking in
915 standard verify code. Enhance 'smime' application with extra parameters
916 to support policy checking and print out.
919 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
920 Nehemiah processors. These extensions support AES encryption in hardware
921 as well as RNG (though RNG support is currently disabled).
922 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
924 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
927 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
928 [Andy Polyakov and a number of other people]
930 *) Improved PowerPC platform support. Most notably BIGNUM assembler
931 implementation contributed by IBM.
932 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
934 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
935 exponent rather than 'unsigned long'. There is a corresponding change to
936 the new 'rsa_keygen' element of the RSA_METHOD structure.
937 [Jelte Jansen, Geoff Thorpe]
939 *) Functionality for creating the initial serial number file is now
940 moved from CA.pl to the 'ca' utility with a new option -create_serial.
942 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
943 number file to 1, which is bound to cause problems. To avoid
944 the problems while respecting compatibility between different 0.9.7
945 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
946 CA.pl for serial number initialization. With the new release 0.9.8,
947 we can fix the problem directly in the 'ca' utility.)
950 *) Reduced header interdepencies by declaring more opaque objects in
951 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
952 give fewer recursive includes, which could break lazy source code - so
953 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
954 developers should define this symbol when building and using openssl to
955 ensure they track the recommended behaviour, interfaces, [etc], but
956 backwards-compatible behaviour prevails when this isn't defined.
959 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
962 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
963 This will generate a random key of the appropriate length based on the
964 cipher context. The EVP_CIPHER can provide its own random key generation
965 routine to support keys of a specific form. This is used in the des and
966 3des routines to generate a key of the correct parity. Update S/MIME
967 code to use new functions and hence generate correct parity DES keys.
968 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
969 valid (weak or incorrect parity).
972 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
973 as looking them up. This is useful when the verified structure may contain
974 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
975 present unless the new PKCS7_NO_CRL flag is asserted.
978 *) Extend ASN1 oid configuration module. It now additionally accepts the
981 shortName = some long name, 1.2.3.4
984 *) Reimplemented the BN_CTX implementation. There is now no more static
985 limitation on the number of variables it can handle nor the depth of the
986 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
987 information can now expand as required, and rather than having a single
988 static array of bignums, BN_CTX now uses a linked-list of such arrays
989 allowing it to expand on demand whilst maintaining the usefulness of
993 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
994 to allow all RSA operations to function using a single BN_CTX.
997 *) Preliminary support for certificate policy evaluation and checking. This
998 is initially intended to pass the tests outlined in "Conformance Testing
999 of Relying Party Client Certificate Path Processing Logic" v1.07.
1002 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1003 remained unused and not that useful. A variety of other little bignum
1004 tweaks and fixes have also been made continuing on from the audit (see
1008 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1009 associated ASN1, EVP and SSL functions and old ASN1 macros.
1012 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1013 and this should never fail. So the return value from the use of
1014 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1015 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1018 *) BN_CTX_get() should return zero-valued bignums, providing the same
1019 initialised value as BN_new().
1020 [Geoff Thorpe, suggested by Ulf Möller]
1022 *) Support for inhibitAnyPolicy certificate extension.
1025 *) An audit of the BIGNUM code is underway, for which debugging code is
1026 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1027 is considered valid when processing BIGNUMs, and causes execution to
1028 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1029 further steps are taken to deliberately pollute unused data in BIGNUM
1030 structures to try and expose faulty code further on. For now, openssl will
1031 (in its default mode of operation) continue to tolerate the inconsistent
1032 forms that it has tolerated in the past, but authors and packagers should
1033 consider trying openssl and their own applications when compiled with
1034 these debugging symbols defined. It will help highlight potential bugs in
1035 their own code, and will improve the test coverage for OpenSSL itself. At
1036 some point, these tighter rules will become openssl's default to improve
1037 maintainability, though the assert()s and other overheads will remain only
1038 in debugging configurations. See bn.h for more details.
1039 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1041 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1042 that can only be obtained through BN_CTX_new() (which implicitly
1043 initialises it). The presence of this function only made it possible
1044 to overwrite an existing structure (and cause memory leaks).
1047 *) Because of the callback-based approach for implementing LHASH as a
1048 template type, lh_insert() adds opaque objects to hash-tables and
1049 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1050 to clean up those corresponding objects before destroying the hash table
1051 (and losing the object pointers). So some over-zealous constifications in
1052 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1053 objects as "const" and the lh_doall[_arg] callback wrappers are not
1054 prototyped to have "const" restrictions on the object pointers they are
1055 given (and so aren't required to cast them away any more).
1058 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1059 (speed) prefers to use its own implementation. The two implementations
1060 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1061 its object type properly exposed (MS_TM) instead of casting to/from "char
1062 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1063 aren't necessarily the greatest nomenclatures - but this is what was used
1064 internally to the implementation so I've used that for now.
1067 *) Ensure that deprecated functions do not get compiled when
1068 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1069 the self-tests were still using deprecated key-generation functions so
1070 these have been updated also.
1073 *) Reorganise PKCS#7 code to separate the digest location functionality
1074 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1075 New function PKCS7_set_digest() to set the digest type for PKCS#7
1076 digestedData type. Add additional code to correctly generate the
1077 digestedData type and add support for this type in PKCS7 initialization
1081 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1082 structure of type "other".
1085 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1086 sure the loop does correctly stop and breaking ("division by zero")
1087 modulus operations are not performed. The (pre-generated) prime
1088 table crypto/bn/bn_prime.h was already correct, but it could not be
1089 re-generated on some platforms because of the "division by zero"
1090 situation in the script.
1091 [Ralf S. Engelschall]
1093 *) Update support for ECC-based TLS ciphersuites according to
1094 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1095 SHA-1 now is only used for "small" curves (where the
1096 representation of a field element takes up to 24 bytes); for
1097 larger curves, the field element resulting from ECDH is directly
1098 used as premaster secret.
1099 [Douglas Stebila (Sun Microsystems Laboratories)]
1101 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1102 curve secp160r1 to the tests.
1103 [Douglas Stebila (Sun Microsystems Laboratories)]
1105 *) Add the possibility to load symbols globally with DSO.
1106 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1108 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1109 control of the error stack.
1112 *) Add support for STORE in ENGINE.
1115 *) Add the STORE type. The intention is to provide a common interface
1116 to certificate and key stores, be they simple file-based stores, or
1117 HSM-type store, or LDAP stores, or...
1118 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1121 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1122 pass a list of arguments to any function as well as provide a way
1123 for a function to pass data back to the caller.
1126 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1127 works like BUF_strdup() but can be used to duplicate a portion of
1128 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1132 *) Add the function sk_find_ex() which works like sk_find(), but will
1133 return an index to an element even if an exact match couldn't be
1134 found. The index is guaranteed to point at the element where the
1135 searched-for key would be inserted to preserve sorting order.
1138 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1139 takes an extra flags argument for optional functionality. Currently,
1140 the following flags are defined:
1142 OBJ_BSEARCH_VALUE_ON_NOMATCH
1143 This one gets OBJ_bsearch_ex() to return a pointer to the first
1144 element where the comparing function returns a negative or zero
1147 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1148 This one gets OBJ_bsearch_ex() to return a pointer to the first
1149 element where the comparing function returns zero. This is useful
1150 if there are more than one element where the comparing function
1154 *) Make it possible to create self-signed certificates with 'openssl ca'
1155 in such a way that the self-signed certificate becomes part of the
1156 CA database and uses the same mechanisms for serial number generation
1157 as all other certificate signing. The new flag '-selfsign' enables
1158 this functionality. Adapt CA.sh and CA.pl.in.
1161 *) Add functionality to check the public key of a certificate request
1162 against a given private. This is useful to check that a certificate
1163 request can be signed by that key (self-signing).
1166 *) Make it possible to have multiple active certificates with the same
1167 subject in the CA index file. This is done only if the keyword
1168 'unique_subject' is set to 'no' in the main CA section (default
1169 if 'CA_default') of the configuration file. The value is saved
1170 with the database itself in a separate index attribute file,
1171 named like the index file with '.attr' appended to the name.
1174 *) Generate muti valued AVAs using '+' notation in config files for
1178 *) Support for nameConstraints certificate extension.
1181 *) Support for policyConstraints certificate extension.
1184 *) Support for policyMappings certificate extension.
1187 *) Make sure the default DSA_METHOD implementation only uses its
1188 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1189 and change its own handlers to be NULL so as to remove unnecessary
1190 indirection. This lets alternative implementations fallback to the
1191 default implementation more easily.
1194 *) Support for directoryName in GeneralName related extensions
1198 *) Make it possible to link applications using Makefile.shared.
1199 Make that possible even when linking against static libraries!
1202 *) Support for single pass processing for S/MIME signing. This now
1203 means that S/MIME signing can be done from a pipe, in addition
1204 cleartext signing (multipart/signed type) is effectively streaming
1205 and the signed data does not need to be all held in memory.
1207 This is done with a new flag PKCS7_STREAM. When this flag is set
1208 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1209 is done after the data is output (and digests calculated) in
1210 SMIME_write_PKCS7().
1213 *) Add full support for -rpath/-R, both in shared libraries and
1214 applications, at least on the platforms where it's known how
1218 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1219 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1220 will now compute a table of multiples of the generator that
1221 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1222 faster (notably in the case of a single point multiplication,
1223 scalar * generator).
1224 [Nils Larsch, Bodo Moeller]
1226 *) IPv6 support for certificate extensions. The various extensions
1227 which use the IP:a.b.c.d can now take IPv6 addresses using the
1228 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1232 *) Added an ENGINE that implements RSA by performing private key
1233 exponentiations with the GMP library. The conversions to and from
1234 GMP's mpz_t format aren't optimised nor are any montgomery forms
1235 cached, and on x86 it appears OpenSSL's own performance has caught up.
1236 However there are likely to be other architectures where GMP could
1237 provide a boost. This ENGINE is not built in by default, but it can be
1238 specified at Configure time and should be accompanied by the necessary
1239 linker additions, eg;
1240 ./config -DOPENSSL_USE_GMP -lgmp
1243 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1244 testing availability of engines with "-t" - the old behaviour is
1245 produced by increasing the feature's verbosity with "-tt".
1248 *) ECDSA routines: under certain error conditions uninitialized BN objects
1249 could be freed. Solution: make sure initialization is performed early
1250 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1254 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1255 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1256 software implementations. For DSA and DH, parameter generation can
1257 also be overriden by providing the appropriate method callbacks.
1260 *) Change the "progress" mechanism used in key-generation and
1261 primality testing to functions that take a new BN_GENCB pointer in
1262 place of callback/argument pairs. The new API functions have "_ex"
1263 postfixes and the older functions are reimplemented as wrappers for
1264 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1265 declarations of the old functions to help (graceful) attempts to
1266 migrate to the new functions. Also, the new key-generation API
1267 functions operate on a caller-supplied key-structure and return
1268 success/failure rather than returning a key or NULL - this is to
1269 help make "keygen" another member function of RSA_METHOD etc.
1271 Example for using the new callback interface:
1273 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1277 BN_GENCB_set(&my_cb, my_callback, my_arg);
1279 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1280 /* For the meaning of a, b in calls to my_callback(), see the
1281 * documentation of the function that calls the callback.
1282 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1283 * my_callback should return 1 if it wants BN_is_prime_ex()
1284 * to continue, or 0 to stop.
1289 *) Change the ZLIB compression method to be stateful, and make it
1290 available to TLS with the number defined in
1291 draft-ietf-tls-compression-04.txt.
1294 *) Add the ASN.1 structures and functions for CertificatePair, which
1295 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1297 CertificatePair ::= SEQUENCE {
1298 forward [0] Certificate OPTIONAL,
1299 reverse [1] Certificate OPTIONAL,
1300 -- at least one of the pair shall be present -- }
1302 Also implement the PEM functions to read and write certificate
1303 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1305 This needed to be defined, mostly for the sake of the LDAP
1306 attribute crossCertificatePair, but may prove useful elsewhere as
1310 *) Make it possible to inhibit symlinking of shared libraries in
1311 Makefile.shared, for Cygwin's sake.
1314 *) Extend the BIGNUM API by creating a function
1315 void BN_set_negative(BIGNUM *a, int neg);
1316 and a macro that behave like
1317 int BN_is_negative(const BIGNUM *a);
1319 to avoid the need to access 'a->neg' directly in applications.
1322 *) Implement fast modular reduction for pseudo-Mersenne primes
1323 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1324 EC_GROUP_new_curve_GFp() will now automatically use this
1326 [Nils Larsch <nla@trustcenter.de>]
1328 *) Add new lock type (CRYPTO_LOCK_BN).
1331 *) Change the ENGINE framework to automatically load engines
1332 dynamically from specific directories unless they could be
1333 found to already be built in or loaded. Move all the
1334 current engines except for the cryptodev one to a new
1336 The engines in engines/ are built as shared libraries if
1337 the "shared" options was given to ./Configure or ./config.
1338 Otherwise, they are inserted in libcrypto.a.
1339 /usr/local/ssl/engines is the default directory for dynamic
1340 engines, but that can be overriden at configure time through
1341 the usual use of --prefix and/or --openssldir, and at run
1342 time with the environment variable OPENSSL_ENGINES.
1343 [Geoff Thorpe and Richard Levitte]
1345 *) Add Makefile.shared, a helper makefile to build shared
1346 libraries. Addapt Makefile.org.
1349 *) Add version info to Win32 DLLs.
1350 [Peter 'Luna' Runestig" <peter@runestig.com>]
1352 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1353 can be added using this API to created arbitrary PKCS#12
1354 files while avoiding the low level API.
1356 New options to PKCS12_create(), key or cert can be NULL and
1357 will then be omitted from the output file. The encryption
1358 algorithm NIDs can be set to -1 for no encryption, the mac
1359 iteration count can be set to 0 to omit the mac.
1361 Enhance pkcs12 utility by making the -nokeys and -nocerts
1362 options work when creating a PKCS#12 file. New option -nomac
1363 to omit the mac, NONE can be set for an encryption algorithm.
1364 New code is modified to use the enhanced PKCS12_create()
1365 instead of the low level API.
1368 *) Extend ASN1 encoder to support indefinite length constructed
1369 encoding. This can output sequences tags and octet strings in
1370 this form. Modify pk7_asn1.c to support indefinite length
1371 encoding. This is experimental and needs additional code to
1372 be useful, such as an ASN1 bio and some enhanced streaming
1375 Extend template encode functionality so that tagging is passed
1376 down to the template encoder.
1379 *) Let 'openssl req' fail if an argument to '-newkey' is not
1380 recognized instead of using RSA as a default.
1383 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1384 As these are not official, they are not included in "ALL";
1385 the "ECCdraft" ciphersuite group alias can be used to select them.
1386 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1388 *) Add ECDH engine support.
1389 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1391 *) Add ECDH in new directory crypto/ecdh/.
1392 [Douglas Stebila (Sun Microsystems Laboratories)]
1394 *) Let BN_rand_range() abort with an error after 100 iterations
1395 without success (which indicates a broken PRNG).
1398 *) Change BN_mod_sqrt() so that it verifies that the input value
1399 is really the square of the return value. (Previously,
1400 BN_mod_sqrt would show GIGO behaviour.)
1403 *) Add named elliptic curves over binary fields from X9.62, SECG,
1404 and WAP/WTLS; add OIDs that were still missing.
1406 [Sheueling Chang Shantz and Douglas Stebila
1407 (Sun Microsystems Laboratories)]
1409 *) Extend the EC library for elliptic curves over binary fields
1410 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1413 EC_GF2m_simple_method
1417 EC_GROUP_new_curve_GF2m
1418 EC_GROUP_set_curve_GF2m
1419 EC_GROUP_get_curve_GF2m
1420 EC_POINT_set_affine_coordinates_GF2m
1421 EC_POINT_get_affine_coordinates_GF2m
1422 EC_POINT_set_compressed_coordinates_GF2m
1424 Point compression for binary fields is disabled by default for
1425 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1428 As binary polynomials are represented as BIGNUMs, various members
1429 of the EC_GROUP and EC_POINT data structures can be shared
1430 between the implementations for prime fields and binary fields;
1431 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1432 are essentially identical to their ..._GFp counterparts.
1433 (For simplicity, the '..._GFp' prefix has been dropped from
1434 various internal method names.)
1436 An internal 'field_div' method (similar to 'field_mul' and
1437 'field_sqr') has been added; this is used only for binary fields.
1439 [Sheueling Chang Shantz and Douglas Stebila
1440 (Sun Microsystems Laboratories)]
1442 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1443 through methods ('mul', 'precompute_mult').
1445 The generic implementations (now internally called 'ec_wNAF_mul'
1446 and 'ec_wNAF_precomputed_mult') remain the default if these
1447 methods are undefined.
1449 [Sheueling Chang Shantz and Douglas Stebila
1450 (Sun Microsystems Laboratories)]
1452 *) New function EC_GROUP_get_degree, which is defined through
1453 EC_METHOD. For curves over prime fields, this returns the bit
1454 length of the modulus.
1456 [Sheueling Chang Shantz and Douglas Stebila
1457 (Sun Microsystems Laboratories)]
1459 *) New functions EC_GROUP_dup, EC_POINT_dup.
1460 (These simply call ..._new and ..._copy).
1462 [Sheueling Chang Shantz and Douglas Stebila
1463 (Sun Microsystems Laboratories)]
1465 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1466 Polynomials are represented as BIGNUMs (where the sign bit is not
1467 used) in the following functions [macros]:
1470 BN_GF2m_sub [= BN_GF2m_add]
1471 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1472 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1473 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1475 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1476 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1477 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1478 BN_GF2m_cmp [= BN_ucmp]
1480 (Note that only the 'mod' functions are actually for fields GF(2^m).
1481 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1483 For some functions, an the irreducible polynomial defining a
1484 field can be given as an 'unsigned int[]' with strictly
1485 decreasing elements giving the indices of those bits that are set;
1486 i.e., p[] represents the polynomial
1487 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1489 p[0] > p[1] > ... > p[k] = 0.
1490 This applies to the following functions:
1495 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1496 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1498 BN_GF2m_mod_sqrt_arr
1499 BN_GF2m_mod_solve_quad_arr
1503 Conversion can be performed by the following functions:
1508 bntest.c has additional tests for binary polynomial arithmetic.
1510 Two implementations for BN_GF2m_mod_div() are available.
1511 The default algorithm simply uses BN_GF2m_mod_inv() and
1512 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1513 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1514 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1516 [Sheueling Chang Shantz and Douglas Stebila
1517 (Sun Microsystems Laboratories)]
1519 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1520 functionality is disabled at compile-time.
1521 [Douglas Stebila <douglas.stebila@sun.com>]
1523 *) Change default behaviour of 'openssl asn1parse' so that more
1524 information is visible when viewing, e.g., a certificate:
1526 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1527 mode the content of non-printable OCTET STRINGs is output in a
1528 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1529 avoid the appearance of a printable string.
1530 [Nils Larsch <nla@trustcenter.de>]
1532 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1534 EC_GROUP_set_asn1_flag()
1535 EC_GROUP_get_asn1_flag()
1536 EC_GROUP_set_point_conversion_form()
1537 EC_GROUP_get_point_conversion_form()
1538 These control ASN1 encoding details:
1539 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1540 has been set to OPENSSL_EC_NAMED_CURVE.
1541 - Points are encoded in uncompressed form by default; options for
1542 asn1_for are as for point2oct, namely
1543 POINT_CONVERSION_COMPRESSED
1544 POINT_CONVERSION_UNCOMPRESSED
1545 POINT_CONVERSION_HYBRID
1547 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1550 EC_GROUP_get0_seed()
1551 EC_GROUP_get_seed_len()
1552 This is used only for ASN1 purposes (so far).
1553 [Nils Larsch <nla@trustcenter.de>]
1555 *) Add 'field_type' member to EC_METHOD, which holds the NID
1556 of the appropriate field type OID. The new function
1557 EC_METHOD_get_field_type() returns this value.
1558 [Nils Larsch <nla@trustcenter.de>]
1563 EC_POINT_point2hex()
1564 EC_POINT_hex2point()
1565 providing useful interfaces to EC_POINT_point2oct() and
1566 EC_POINT_oct2point().
1567 [Nils Larsch <nla@trustcenter.de>]
1569 *) Change internals of the EC library so that the functions
1570 EC_GROUP_set_generator()
1571 EC_GROUP_get_generator()
1572 EC_GROUP_get_order()
1573 EC_GROUP_get_cofactor()
1574 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1575 to methods, which would lead to unnecessary code duplication when
1576 adding different types of curves.
1577 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1579 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1580 arithmetic, and such that modified wNAFs are generated
1581 (which avoid length expansion in many cases).
1584 *) Add a function EC_GROUP_check_discriminant() (defined via
1585 EC_METHOD) that verifies that the curve discriminant is non-zero.
1587 Add a function EC_GROUP_check() that makes some sanity tests
1588 on a EC_GROUP, its generator and order. This includes
1589 EC_GROUP_check_discriminant().
1590 [Nils Larsch <nla@trustcenter.de>]
1592 *) Add ECDSA in new directory crypto/ecdsa/.
1594 Add applications 'openssl ecparam' and 'openssl ecdsa'
1595 (these are based on 'openssl dsaparam' and 'openssl dsa').
1597 ECDSA support is also included in various other files across the
1598 library. Most notably,
1599 - 'openssl req' now has a '-newkey ecdsa:file' option;
1600 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1601 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1602 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1603 them suitable for ECDSA where domain parameters must be
1604 extracted before the specific public key;
1605 - ECDSA engine support has been added.
1606 [Nils Larsch <nla@trustcenter.de>]
1608 *) Include some named elliptic curves, and add OIDs from X9.62,
1609 SECG, and WAP/WTLS. Each curve can be obtained from the new
1611 EC_GROUP_new_by_curve_name(),
1612 and the list of available named curves can be obtained with
1613 EC_get_builtin_curves().
1614 Also add a 'curve_name' member to EC_GROUP objects, which can be
1616 EC_GROUP_set_curve_name()
1617 EC_GROUP_get_curve_name()
1618 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1620 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1621 was actually never needed) and in BN_mul(). The removal in BN_mul()
1622 required a small change in bn_mul_part_recursive() and the addition
1623 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1624 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1625 bn_sub_words() and bn_add_words() except they take arrays with
1629 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1631 *) Cleanse PEM buffers before freeing them since they may contain
1633 [Benjamin Bennett <ben@psc.edu>]
1635 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1636 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1637 authentication-only ciphersuites.
1640 *) Since AES128 and AES256 share a single mask bit in the logic of
1641 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1642 kludge to work properly if AES128 is available and AES256 isn't.
1645 *) Expand security boundary to match 1.1.1 module.
1648 *) Remove redundant features: hash file source, editing of test vectors
1649 modify fipsld to use external fips_premain.c signature.
1652 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1653 run algorithm test programs.
1656 *) Make algorithm test programs more tolerant of whitespace.
1659 *) Have SSL/TLS server implementation tolerate "mismatched" record
1660 protocol version while receiving ClientHello even if the
1661 ClientHello is fragmented. (The server can't insist on the
1662 particular protocol version it has chosen before the ServerHello
1663 message has informed the client about his choice.)
1666 *) Load error codes if they are not already present instead of using a
1667 static variable. This allows them to be cleanly unloaded and reloaded.
1670 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1672 *) Introduce limits to prevent malicious keys being able to
1673 cause a denial of service. (CVE-2006-2940)
1674 [Steve Henson, Bodo Moeller]
1676 *) Fix ASN.1 parsing of certain invalid structures that can result
1677 in a denial of service. (CVE-2006-2937) [Steve Henson]
1679 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1680 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1682 *) Fix SSL client code which could crash if connecting to a
1683 malicious SSLv2 server. (CVE-2006-4343)
1684 [Tavis Ormandy and Will Drewry, Google Security Team]
1686 *) Change ciphersuite string processing so that an explicit
1687 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1688 will no longer include "AES128-SHA"), and any other similar
1689 ciphersuite (same bitmap) from *other* protocol versions (so that
1690 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1691 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1692 changes from 0.9.8b and 0.9.8d.
1695 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1697 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1698 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1700 *) Change the Unix randomness entropy gathering to use poll() when
1701 possible instead of select(), since the latter has some
1702 undesirable limitations.
1703 [Darryl Miles via Richard Levitte and Bodo Moeller]
1705 *) Disable rogue ciphersuites:
1707 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1708 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1709 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1711 The latter two were purportedly from
1712 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1715 Also deactive the remaining ciphersuites from
1716 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1717 unofficial, and the ID has long expired.
1720 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1721 dual-core machines) and other potential thread-safety issues.
1724 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1726 *) Adapt fipsld and the build system to link against the validated FIPS
1727 module in FIPS mode.
1730 *) Fixes for VC++ 2005 build under Windows.
1733 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1734 from a Windows bash shell such as MSYS. It is autodetected from the
1735 "config" script when run from a VC++ environment. Modify standard VC++
1736 build to use fipscanister.o from the GNU make build.
1739 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1741 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1742 The value now differs depending on if you build for FIPS or not.
1743 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1744 safely run with a non-FIPSed libcrypto, as it may crash because of
1745 the difference induced by this change.
1748 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1750 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1751 (part of SSL_OP_ALL). This option used to disable the
1752 countermeasure against man-in-the-middle protocol-version
1753 rollback in the SSL 2.0 server implementation, which is a bad
1754 idea. (CVE-2005-2969)
1756 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1757 for Information Security, National Institute of Advanced Industrial
1758 Science and Technology [AIST], Japan)]
1760 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1761 mainly for FIPS compliance and not fully integrated at this stage.
1764 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1765 the exponentiation using a fixed-length exponent. (Otherwise,
1766 the information leaked through timing could expose the secret key
1767 after many signatures; cf. Bleichenbacher's attack on DSA with
1771 *) Make a new fixed-window mod_exp implementation the default for
1772 RSA, DSA, and DH private-key operations so that the sequence of
1773 squares and multiplies and the memory access pattern are
1774 independent of the particular secret key. This will mitigate
1775 cache-timing and potential related attacks.
1777 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1778 and this is automatically used by BN_mod_exp_mont() if the new flag
1779 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1780 will use this BN flag for private exponents unless the flag
1781 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1782 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1784 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1786 *) Change the client implementation for SSLv23_method() and
1787 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1788 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1789 (Previously, the SSL 2.0 backwards compatible Client Hello
1790 message format would be used even with SSL_OP_NO_SSLv2.)
1793 *) Add support for smime-type MIME parameter in S/MIME messages which some
1797 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1798 a threadsafe manner. Modify rsa code to use new function and add calls
1799 to dsa and dh code (which had race conditions before).
1802 *) Include the fixed error library code in the C error file definitions
1803 instead of fixing them up at runtime. This keeps the error code
1804 structures constant.
1807 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1809 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1812 *) Fixes for newer kerberos headers. NB: the casts are needed because
1813 the 'length' field is signed on one version and unsigned on another
1814 with no (?) obvious way to tell the difference, without these VC++
1815 complains. Also the "definition" of FAR (blank) is no longer included
1816 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1817 some needed definitions.
1820 *) Undo Cygwin change.
1823 *) Added support for proxy certificates according to RFC 3820.
1824 Because they may be a security thread to unaware applications,
1825 they must be explicitely allowed in run-time. See
1826 docs/HOWTO/proxy_certificates.txt for further information.
1829 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1831 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1832 server and client random values. Previously
1833 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1834 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1836 This change has negligible security impact because:
1838 1. Server and client random values still have 24 bytes of pseudo random
1841 2. Server and client random values are sent in the clear in the initial
1844 3. The master secret is derived using the premaster secret (48 bytes in
1845 size for static RSA ciphersuites) as well as client server and random
1848 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1851 [Stephen Henson, reported by UK NISCC]
1853 *) Use Windows randomness collection on Cygwin.
1856 *) Fix hang in EGD/PRNGD query when communication socket is closed
1857 prematurely by EGD/PRNGD.
1858 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1860 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1863 *) Back-port of selected performance improvements from development
1864 branch, as well as improved support for PowerPC platforms.
1867 *) Add lots of checks for memory allocation failure, error codes to indicate
1868 failure and freeing up memory if a failure occurs.
1869 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1871 *) Add new -passin argument to dgst.
1874 *) Perform some character comparisons of different types in X509_NAME_cmp:
1875 this is needed for some certificates that reencode DNs into UTF8Strings
1876 (in violation of RFC3280) and can't or wont issue name rollover
1880 *) Make an explicit check during certificate validation to see that
1881 the CA setting in each certificate on the chain is correct. As a
1882 side effect always do the following basic checks on extensions,
1883 not just when there's an associated purpose to the check:
1885 - if there is an unhandled critical extension (unless the user
1886 has chosen to ignore this fault)
1887 - if the path length has been exceeded (if one is set at all)
1888 - that certain extensions fit the associated purpose (if one has
1892 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1894 *) Avoid a race condition when CRLs are checked in a multi threaded
1895 environment. This would happen due to the reordering of the revoked
1896 entries during signature checking and serial number lookup. Now the
1897 encoding is cached and the serial number sort performed under a lock.
1898 Add new STACK function sk_is_sorted().
1901 *) Add Delta CRL to the extension code.
1904 *) Various fixes to s3_pkt.c so alerts are sent properly.
1905 [David Holmes <d.holmes@f5.com>]
1907 *) Reduce the chances of duplicate issuer name and serial numbers (in
1908 violation of RFC3280) using the OpenSSL certificate creation utilities.
1909 This is done by creating a random 64 bit value for the initial serial
1910 number when a serial number file is created or when a self signed
1911 certificate is created using 'openssl req -x509'. The initial serial
1912 number file is created using 'openssl x509 -next_serial' in CA.pl
1913 rather than being initialized to 1.
1916 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1918 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1919 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1920 [Joe Orton, Steve Henson]
1922 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1924 [Joe Orton, Steve Henson]
1926 *) Make it possible to have multiple active certificates with the same
1927 subject in the CA index file. This is done only if the keyword
1928 'unique_subject' is set to 'no' in the main CA section (default
1929 if 'CA_default') of the configuration file. The value is saved
1930 with the database itself in a separate index attribute file,
1931 named like the index file with '.attr' appended to the name.
1934 *) X509 verify fixes. Disable broken certificate workarounds when
1935 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1936 keyUsage extension present. Don't accept CRLs with unhandled critical
1937 extensions: since verify currently doesn't process CRL extensions this
1938 rejects a CRL with *any* critical extensions. Add new verify error codes
1942 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1943 A clarification of RFC2560 will require the use of OCTET STRINGs and
1944 some implementations cannot handle the current raw format. Since OpenSSL
1945 copies and compares OCSP nonces as opaque blobs without any attempt at
1946 parsing them this should not create any compatibility issues.
1949 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1950 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1951 this HMAC (and other) operations are several times slower than OpenSSL
1955 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1956 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1958 *) Use the correct content when signing type "other".
1961 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1963 *) Fix various bugs revealed by running the NISCC test suite:
1965 Stop out of bounds reads in the ASN1 code when presented with
1966 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1968 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1970 If verify callback ignores invalid public key errors don't try to check
1971 certificate signature with the NULL public key.
1975 *) New -ignore_err option in ocsp application to stop the server
1976 exiting on the first error in a request.
1979 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1980 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1984 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1985 extra data after the compression methods not only for TLS 1.0
1986 but also for SSL 3.0 (as required by the specification).
1987 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1989 *) Change X509_certificate_type() to mark the key as exported/exportable
1990 when it's 512 *bits* long, not 512 bytes.
1993 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1994 blocks during encryption.
1997 *) Various fixes to base64 BIO and non blocking I/O. On write
1998 flushes were not handled properly if the BIO retried. On read
1999 data was not being buffered properly and had various logic bugs.
2000 This also affects blocking I/O when the data being decoded is a
2004 *) Various S/MIME bugfixes and compatibility changes:
2005 output correct application/pkcs7 MIME type if
2006 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2007 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2008 of files as .eml work). Correctly handle very long lines in MIME
2012 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2014 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2015 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2016 a protocol version number mismatch like a decryption error
2017 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2020 *) Turn on RSA blinding by default in the default implementation
2021 to avoid a timing attack. Applications that don't want it can call
2022 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2023 They would be ill-advised to do so in most cases.
2024 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2026 *) Change RSA blinding code so that it works when the PRNG is not
2027 seeded (in this case, the secret RSA exponent is abused as
2028 an unpredictable seed -- if it is not unpredictable, there
2029 is no point in blinding anyway). Make RSA blinding thread-safe
2030 by remembering the creator's thread ID in rsa->blinding and
2031 having all other threads use local one-time blinding factors
2032 (this requires more computation than sharing rsa->blinding, but
2033 avoids excessive locking; and if an RSA object is not shared
2034 between threads, blinding will still be very fast).
2037 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2038 ENGINE as defaults for all supported algorithms irrespective of
2039 the 'flags' parameter. 'flags' is now honoured, so applications
2040 should make sure they are passing it correctly.
2043 *) Target "mingw" now allows native Windows code to be generated in
2044 the Cygwin environment as well as with the MinGW compiler.
2047 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2049 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2050 via timing by performing a MAC computation even if incorrrect
2051 block cipher padding has been found. This is a countermeasure
2052 against active attacks where the attacker has to distinguish
2053 between bad padding and a MAC verification error. (CVE-2003-0078)
2055 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2056 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2057 Martin Vuagnoux (EPFL, Ilion)]
2059 *) Make the no-err option work as intended. The intention with no-err
2060 is not to have the whole error stack handling routines removed from
2061 libcrypto, it's only intended to remove all the function name and
2062 reason texts, thereby removing some of the footprint that may not
2063 be interesting if those errors aren't displayed anyway.
2065 NOTE: it's still possible for any application or module to have it's
2066 own set of error texts inserted. The routines are there, just not
2067 used by default when no-err is given.
2070 *) Add support for FreeBSD on IA64.
2071 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2073 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2074 Kerberos function mit_des_cbc_cksum(). Before this change,
2075 the value returned by DES_cbc_cksum() was like the one from
2076 mit_des_cbc_cksum(), except the bytes were swapped.
2077 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2079 *) Allow an application to disable the automatic SSL chain building.
2080 Before this a rather primitive chain build was always performed in
2081 ssl3_output_cert_chain(): an application had no way to send the
2082 correct chain if the automatic operation produced an incorrect result.
2084 Now the chain builder is disabled if either:
2086 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2088 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2090 The reasoning behind this is that an application would not want the
2091 auto chain building to take place if extra chain certificates are
2092 present and it might also want a means of sending no additional
2093 certificates (for example the chain has two certificates and the
2097 *) Add the possibility to build without the ENGINE framework.
2098 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2100 *) Under Win32 gmtime() can return NULL: check return value in
2101 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2104 *) DSA routines: under certain error conditions uninitialized BN objects
2105 could be freed. Solution: make sure initialization is performed early
2106 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2107 Nils Larsch <nla@trustcenter.de> via PR#459)
2110 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2111 checked on reconnect on the client side, therefore session resumption
2112 could still fail with a "ssl session id is different" error. This
2113 behaviour is masked when SSL_OP_ALL is used due to
2114 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2115 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2116 followup to PR #377.
2119 *) IA-32 assembler support enhancements: unified ELF targets, support
2120 for SCO/Caldera platforms, fix for Cygwin shared build.
2123 *) Add support for FreeBSD on sparc64. As a consequence, support for
2124 FreeBSD on non-x86 processors is separate from x86 processors on
2125 the config script, much like the NetBSD support.
2126 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2128 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2130 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2133 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2134 code (06) was taken as the first octet of the session ID and the last
2135 octet was ignored consequently. As a result SSLv2 client side session
2136 caching could not have worked due to the session ID mismatch between
2138 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2142 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2143 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2147 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2148 seems that in spite of existing for more than a year, many application
2149 author have done nothing to provide the necessary callbacks, which
2150 means that this particular engine will not work properly anywhere.
2151 This is a very unfortunate situation which forces us, in the name
2152 of usability, to give the hw_ncipher.c a static lock, which is part
2154 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2155 appear in 0.9.8 or later. We EXPECT application authors to have
2156 dealt properly with this when 0.9.8 is released (unless we actually
2157 make such changes in the libcrypto locking code that changes will
2158 have to be made anyway).
2161 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2162 octets have been read, EOF or an error occurs. Without this change
2163 some truncated ASN1 structures will not produce an error.
2166 *) Disable Heimdal support, since it hasn't been fully implemented.
2167 Still give the possibility to force the use of Heimdal, but with
2168 warnings and a request that patches get sent to openssl-dev.
2171 *) Add the VC-CE target, introduce the WINCE sysname, and add
2172 INSTALL.WCE and appropriate conditionals to make it build.
2173 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2175 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2176 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2177 edit numbers of the version.
2178 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2180 *) Introduce safe string copy and catenation functions
2181 (BUF_strlcpy() and BUF_strlcat()).
2182 [Ben Laurie (CHATS) and Richard Levitte]
2184 *) Avoid using fixed-size buffers for one-line DNs.
2185 [Ben Laurie (CHATS)]
2187 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2188 resizing buffers containing secrets, and use where appropriate.
2189 [Ben Laurie (CHATS)]
2191 *) Avoid using fixed size buffers for configuration file location.
2192 [Ben Laurie (CHATS)]
2194 *) Avoid filename truncation for various CA files.
2195 [Ben Laurie (CHATS)]
2197 *) Use sizeof in preference to magic numbers.
2198 [Ben Laurie (CHATS)]
2200 *) Avoid filename truncation in cert requests.
2201 [Ben Laurie (CHATS)]
2203 *) Add assertions to check for (supposedly impossible) buffer
2205 [Ben Laurie (CHATS)]
2207 *) Don't cache truncated DNS entries in the local cache (this could
2208 potentially lead to a spoofing attack).
2209 [Ben Laurie (CHATS)]
2211 *) Fix various buffers to be large enough for hex/decimal
2212 representations in a platform independent manner.
2213 [Ben Laurie (CHATS)]
2215 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2216 resizing buffers containing secrets, and use where appropriate.
2217 [Ben Laurie (CHATS)]
2219 *) Add BIO_indent() to avoid much slightly worrying code to do
2221 [Ben Laurie (CHATS)]
2223 *) Convert sprintf()/BIO_puts() to BIO_printf().
2224 [Ben Laurie (CHATS)]
2226 *) buffer_gets() could terminate with the buffer only half
2228 [Ben Laurie (CHATS)]
2230 *) Add assertions to prevent user-supplied crypto functions from
2231 overflowing internal buffers by having large block sizes, etc.
2232 [Ben Laurie (CHATS)]
2234 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2236 [Ben Laurie (CHATS)]
2238 *) Eliminate unused copy of key in RC4.
2239 [Ben Laurie (CHATS)]
2241 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2242 [Ben Laurie (CHATS)]
2244 *) Fix off-by-one error in EGD path.
2245 [Ben Laurie (CHATS)]
2247 *) If RANDFILE path is too long, ignore instead of truncating.
2248 [Ben Laurie (CHATS)]
2250 *) Eliminate unused and incorrectly sized X.509 structure
2252 [Ben Laurie (CHATS)]
2254 *) Eliminate unused and dangerous function knumber().
2255 [Ben Laurie (CHATS)]
2257 *) Eliminate unused and dangerous structure, KSSL_ERR.
2258 [Ben Laurie (CHATS)]
2260 *) Protect against overlong session ID context length in an encoded
2261 session object. Since these are local, this does not appear to be
2263 [Ben Laurie (CHATS)]
2265 *) Change from security patch (see 0.9.6e below) that did not affect
2266 the 0.9.6 release series:
2268 Remote buffer overflow in SSL3 protocol - an attacker could
2269 supply an oversized master key in Kerberos-enabled versions.
2271 [Ben Laurie (CHATS)]
2273 *) Change the SSL kerb5 codes to match RFC 2712.
2276 *) Make -nameopt work fully for req and add -reqopt switch.
2277 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2279 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2280 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2282 *) Make sure tests can be performed even if the corresponding algorithms
2283 have been removed entirely. This was also the last step to make
2284 OpenSSL compilable with DJGPP under all reasonable conditions.
2285 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2287 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2288 to allow version independent disabling of normally unselected ciphers,
2289 which may be activated as a side-effect of selecting a single cipher.
2291 (E.g., cipher list string "RSA" enables ciphersuites that are left
2292 out of "ALL" because they do not provide symmetric encryption.
2293 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2294 [Lutz Jaenicke, Bodo Moeller]
2296 *) Add appropriate support for separate platform-dependent build
2297 directories. The recommended way to make a platform-dependent
2298 build directory is the following (tested on Linux), maybe with
2301 # Place yourself outside of the OpenSSL source tree. In
2302 # this example, the environment variable OPENSSL_SOURCE
2303 # is assumed to contain the absolute OpenSSL source directory.
2304 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2305 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2306 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2307 mkdir -p `dirname $F`
2308 ln -s $OPENSSL_SOURCE/$F $F
2311 To be absolutely sure not to disturb the source tree, a "make clean"
2312 is a good thing. If it isn't successfull, don't worry about it,
2313 it probably means the source directory is very clean.
2316 *) Make sure any ENGINE control commands make local copies of string
2317 pointers passed to them whenever necessary. Otherwise it is possible
2318 the caller may have overwritten (or deallocated) the original string
2319 data when a later ENGINE operation tries to use the stored values.
2320 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2322 *) Improve diagnostics in file reading and command-line digests.
2323 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2325 *) Add AES modes CFB and OFB to the object database. Correct an
2326 error in AES-CFB decryption.
2329 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2330 allows existing EVP_CIPHER_CTX structures to be reused after
2331 calling EVP_*Final(). This behaviour is used by encryption
2332 BIOs and some applications. This has the side effect that
2333 applications must explicitly clean up cipher contexts with
2334 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2337 *) Check the values of dna and dnb in bn_mul_recursive before calling
2338 bn_mul_comba (a non zero value means the a or b arrays do not contain
2339 n2 elements) and fallback to bn_mul_normal if either is not zero.
2342 *) Fix escaping of non-ASCII characters when using the -subj option
2343 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2346 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2347 form for "surname", serialNumber has no short form.
2348 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2349 therefore remove "mail" short name for "internet 7".
2350 The OID for unique identifiers in X509 certificates is
2351 x500UniqueIdentifier, not uniqueIdentifier.
2352 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2355 *) Add an "init" command to the ENGINE config module and auto initialize
2356 ENGINEs. Without any "init" command the ENGINE will be initialized
2357 after all ctrl commands have been executed on it. If init=1 the
2358 ENGINE is initailized at that point (ctrls before that point are run
2359 on the uninitialized ENGINE and after on the initialized one). If
2360 init=0 then the ENGINE will not be iniatialized at all.
2363 *) Fix the 'app_verify_callback' interface so that the user-defined
2364 argument is actually passed to the callback: In the
2365 SSL_CTX_set_cert_verify_callback() prototype, the callback
2366 declaration has been changed from
2369 int (*cb)(X509_STORE_CTX *,void *);
2370 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2371 i=s->ctx->app_verify_callback(&ctx)
2372 has been changed into
2373 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2375 To update applications using SSL_CTX_set_cert_verify_callback(),
2376 a dummy argument can be added to their callback functions.
2377 [D. K. Smetters <smetters@parc.xerox.com>]
2379 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2380 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2382 *) Add and OPENSSL_LOAD_CONF define which will cause
2383 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2384 This allows older applications to transparently support certain
2385 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2386 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2387 load the config file and OPENSSL_add_all_algorithms_conf() which will
2388 always load it have also been added.
2391 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2392 Adjust NIDs and EVP layer.
2393 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2395 *) Config modules support in openssl utility.
2397 Most commands now load modules from the config file,
2398 though in a few (such as version) this isn't done
2399 because it couldn't be used for anything.
2401 In the case of ca and req the config file used is
2402 the same as the utility itself: that is the -config
2403 command line option can be used to specify an
2407 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2408 use "openssl_conf" if filename is NULL use default openssl config file.
2411 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2412 config section name. Add a new flag to tolerate a missing config file
2413 and move code to CONF_modules_load_file().
2416 *) Support for crypto accelerator cards from Accelerated Encryption
2417 Processing, www.aep.ie. (Use engine 'aep')
2418 The support was copied from 0.9.6c [engine] and adapted/corrected
2419 to work with the new engine framework.
2420 [AEP Inc. and Richard Levitte]
2422 *) Support for SureWare crypto accelerator cards from Baltimore
2423 Technologies. (Use engine 'sureware')
2424 The support was copied from 0.9.6c [engine] and adapted
2425 to work with the new engine framework.
2428 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2429 make the newer ENGINE framework commands for the CHIL engine work.
2430 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2432 *) Make it possible to produce shared libraries on ReliantUNIX.
2433 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2435 *) Add the configuration target debug-linux-ppro.
2436 Make 'openssl rsa' use the general key loading routines
2437 implemented in apps.c, and make those routines able to
2438 handle the key format FORMAT_NETSCAPE and the variant
2440 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2442 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2443 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2445 *) Add -keyform to rsautl, and document -engine.
2446 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2448 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2449 BIO_R_NO_SUCH_FILE error code rather than the generic
2450 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2453 *) Add new functions
2455 ERR_peek_last_error_line
2456 ERR_peek_last_error_line_data.
2457 These are similar to
2460 ERR_peek_error_line_data,
2461 but report on the latest error recorded rather than the first one
2462 still in the error queue.
2463 [Ben Laurie, Bodo Moeller]
2465 *) default_algorithms option in ENGINE config module. This allows things
2467 default_algorithms = ALL
2468 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2471 *) Prelminary ENGINE config module.
2474 *) New experimental application configuration code.
2477 *) Change the AES code to follow the same name structure as all other
2478 symmetric ciphers, and behave the same way. Move everything to
2479 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2480 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2482 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2483 [Ben Laurie and Theo de Raadt]
2485 *) Add option to output public keys in req command.
2486 [Massimiliano Pala madwolf@openca.org]
2488 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2489 (up to about 10% better than before for P-192 and P-224).
2492 *) New functions/macros
2494 SSL_CTX_set_msg_callback(ctx, cb)
2495 SSL_CTX_set_msg_callback_arg(ctx, arg)
2496 SSL_set_msg_callback(ssl, cb)
2497 SSL_set_msg_callback_arg(ssl, arg)
2499 to request calling a callback function
2501 void cb(int write_p, int version, int content_type,
2502 const void *buf, size_t len, SSL *ssl, void *arg)
2504 whenever a protocol message has been completely received
2505 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2506 protocol version according to which the SSL library interprets
2507 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2508 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2509 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2510 specification (change_cipher_spec(20), alert(21), handshake(22)).
2511 'buf' and 'len' point to the actual message, 'ssl' to the
2512 SSL object, and 'arg' is the application-defined value set by
2513 SSL[_CTX]_set_msg_callback_arg().
2515 'openssl s_client' and 'openssl s_server' have new '-msg' options
2516 to enable a callback that displays all protocol messages.
2519 *) Change the shared library support so shared libraries are built as
2520 soon as the corresponding static library is finished, and thereby get
2521 openssl and the test programs linked against the shared library.
2522 This still only happens when the keyword "shard" has been given to
2523 the configuration scripts.
2525 NOTE: shared library support is still an experimental thing, and
2526 backward binary compatibility is still not guaranteed.
2527 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2529 *) Add support for Subject Information Access extension.
2530 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2532 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2533 additional bytes when new memory had to be allocated, not just
2534 when reusing an existing buffer.
2537 *) New command line and configuration option 'utf8' for the req command.
2538 This allows field values to be specified as UTF8 strings.
2541 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2542 runs for the former and machine-readable output for the latter.
2545 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2546 of the e-mail address in the DN (i.e., it will go into a certificate
2547 extension only). The new configuration file option 'email_in_dn = no'
2548 has the same effect.
2549 [Massimiliano Pala madwolf@openca.org]
2551 *) Change all functions with names starting with des_ to be starting
2552 with DES_ instead. Add wrappers that are compatible with libdes,
2553 but are named _ossl_old_des_*. Finally, add macros that map the
2554 des_* symbols to the corresponding _ossl_old_des_* if libdes
2555 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2556 desired, the des_* symbols will be mapped to DES_*, with one
2559 Since we provide two compatibility mappings, the user needs to
2560 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2561 compatibility is desired. The default (i.e., when that macro
2562 isn't defined) is OpenSSL 0.9.6c compatibility.
2564 There are also macros that enable and disable the support of old
2565 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2566 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2567 are defined, the default will apply: to support the old des routines.
2569 In either case, one must include openssl/des.h to get the correct
2570 definitions. Do not try to just include openssl/des_old.h, that
2573 NOTE: This is a major break of an old API into a new one. Software
2574 authors are encouraged to switch to the DES_ style functions. Some
2575 time in the future, des_old.h and the libdes compatibility functions
2576 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2577 default), and then completely removed.
2580 *) Test for certificates which contain unsupported critical extensions.
2581 If such a certificate is found during a verify operation it is
2582 rejected by default: this behaviour can be overridden by either
2583 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2584 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2585 X509_supported_extension() has also been added which returns 1 if a
2586 particular extension is supported.
2589 *) Modify the behaviour of EVP cipher functions in similar way to digests
2590 to retain compatibility with existing code.
2593 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2594 compatibility with existing code. In particular the 'ctx' parameter does
2595 not have to be to be initialized before the call to EVP_DigestInit() and
2596 it is tidied up after a call to EVP_DigestFinal(). New function
2597 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2598 EVP_MD_CTX_copy() changed to not require the destination to be
2599 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2600 requires the destination to be valid.
2602 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2603 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2606 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2607 so that complete 'Handshake' protocol structures are kept in memory
2608 instead of overwriting 'msg_type' and 'length' with 'body' data.
2611 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2612 [Massimo Santin via Richard Levitte]
2614 *) Major restructuring to the underlying ENGINE code. This includes
2615 reduction of linker bloat, separation of pure "ENGINE" manipulation
2616 (initialisation, etc) from functionality dealing with implementations
2617 of specific crypto iterfaces. This change also introduces integrated
2618 support for symmetric ciphers and digest implementations - so ENGINEs
2619 can now accelerate these by providing EVP_CIPHER and EVP_MD
2620 implementations of their own. This is detailed in crypto/engine/README
2621 as it couldn't be adequately described here. However, there are a few
2622 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2623 were changed in the original introduction of ENGINE code have now
2624 reverted back - the hooking from this code to ENGINE is now a good
2625 deal more passive and at run-time, operations deal directly with
2626 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2627 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2628 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2629 they were not being used by the framework as there is no concept of a
2630 BIGNUM_METHOD and they could not be generalised to the new
2631 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2632 ENGINE_cpy() has been removed as it cannot be consistently defined in
2636 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2639 *) Change mkdef.pl to sort symbols that get the same entry number,
2640 and make sure the automatically generated functions ERR_load_*
2641 become part of libeay.num as well.
2644 *) New function SSL_renegotiate_pending(). This returns true once
2645 renegotiation has been requested (either SSL_renegotiate() call
2646 or HelloRequest/ClientHello receveived from the peer) and becomes
2647 false once a handshake has been completed.
2648 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2649 sends a HelloRequest, but does not ensure that a handshake takes
2650 place. SSL_renegotiate_pending() is useful for checking if the
2651 client has followed the request.)
2654 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2655 By default, clients may request session resumption even during
2656 renegotiation (if session ID contexts permit); with this option,
2657 session resumption is possible only in the first handshake.
2659 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2660 more bits available for options that should not be part of
2661 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2664 *) Add some demos for certificate and certificate request creation.
2667 *) Make maximum certificate chain size accepted from the peer application
2668 settable (SSL*_get/set_max_cert_list()), as proposed by
2669 "Douglas E. Engert" <deengert@anl.gov>.
2672 *) Add support for shared libraries for Unixware-7
2673 (Boyd Lynn Gerber <gerberb@zenez.com>).
2676 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2677 be done prior to destruction. Use this to unload error strings from
2678 ENGINEs that load their own error strings. NB: This adds two new API
2679 functions to "get" and "set" this destroy handler in an ENGINE.
2682 *) Alter all existing ENGINE implementations (except "openssl" and
2683 "openbsd") to dynamically instantiate their own error strings. This
2684 makes them more flexible to be built both as statically-linked ENGINEs
2685 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2686 Also, add stub code to each that makes building them as self-contained
2687 shared-libraries easier (see README.ENGINE).
2690 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2691 implementations into applications that are completely implemented in
2692 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2693 commands that can be used to configure what shared-library to load and
2694 to control aspects of the way it is handled. Also, made an update to
2695 the README.ENGINE file that brings its information up-to-date and
2696 provides some information and instructions on the "dynamic" ENGINE
2697 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2700 *) Make it possible to unload ranges of ERR strings with a new
2701 "ERR_unload_strings" function.
2704 *) Add a copy() function to EVP_MD.
2707 *) Make EVP_MD routines take a context pointer instead of just the
2708 md_data void pointer.
2711 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2712 that the digest can only process a single chunk of data
2713 (typically because it is provided by a piece of
2714 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2715 is only going to provide a single chunk of data, and hence the
2716 framework needn't accumulate the data for oneshot drivers.
2719 *) As with "ERR", make it possible to replace the underlying "ex_data"
2720 functions. This change also alters the storage and management of global
2721 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2722 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2723 index counters. The API functions that use this state have been changed
2724 to take a "class_index" rather than pointers to the class's local STACK
2725 and counter, and there is now an API function to dynamically create new
2726 classes. This centralisation allows us to (a) plug a lot of the
2727 thread-safety problems that existed, and (b) makes it possible to clean
2728 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2729 such data would previously have always leaked in application code and
2730 workarounds were in place to make the memory debugging turn a blind eye
2731 to it. Application code that doesn't use this new function will still
2732 leak as before, but their memory debugging output will announce it now
2733 rather than letting it slide.
2735 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2736 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2737 has a return value to indicate success or failure.
2740 *) Make it possible to replace the underlying "ERR" functions such that the
2741 global state (2 LHASH tables and 2 locks) is only used by the "default"
2742 implementation. This change also adds two functions to "get" and "set"
2743 the implementation prior to it being automatically set the first time
2744 any other ERR function takes place. Ie. an application can call "get",
2745 pass the return value to a module it has just loaded, and that module
2746 can call its own "set" function using that value. This means the
2747 module's "ERR" operations will use (and modify) the error state in the
2748 application and not in its own statically linked copy of OpenSSL code.
2751 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2752 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2753 the operation, and provides a more encapsulated way for external code
2754 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2755 to use these functions rather than manually incrementing the counts.
2757 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2760 *) Add EVP test program.
2763 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2766 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2767 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2768 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2769 These allow a CRL to be built without having to access X509_CRL fields
2770 directly. Modify 'ca' application to use new functions.
2773 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2774 bug workarounds. Rollback attack detection is a security feature.
2775 The problem will only arise on OpenSSL servers when TLSv1 is not
2776 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2777 Software authors not wanting to support TLSv1 will have special reasons
2778 for their choice and can explicitly enable this option.
2779 [Bodo Moeller, Lutz Jaenicke]
2781 *) Rationalise EVP so it can be extended: don't include a union of
2782 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2783 (similar to those existing for EVP_CIPHER_CTX).
2788 EVP_MD_CTX_init(&md); /* new function call */
2789 EVP_DigestInit(&md, EVP_sha1());
2790 EVP_DigestUpdate(&md, in, len);
2791 EVP_DigestFinal(&md, out, NULL);
2792 EVP_MD_CTX_cleanup(&md); /* new function call */
2796 *) Make DES key schedule conform to the usual scheme, as well as
2797 correcting its structure. This means that calls to DES functions
2798 now have to pass a pointer to a des_key_schedule instead of a
2799 plain des_key_schedule (which was actually always a pointer
2802 des_key_schedule ks;
2804 des_set_key_checked(..., &ks);
2805 des_ncbc_encrypt(..., &ks, ...);
2807 (Note that a later change renames 'des_...' into 'DES_...'.)
2810 *) Initial reduction of linker bloat: the use of some functions, such as
2811 PEM causes large amounts of unused functions to be linked in due to
2812 poor organisation. For example pem_all.c contains every PEM function
2813 which has a knock on effect of linking in large amounts of (unused)
2814 ASN1 code. Grouping together similar functions and splitting unrelated
2815 functions prevents this.
2818 *) Cleanup of EVP macros.
2821 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2822 correct _ecb suffix.
2825 *) Add initial OCSP responder support to ocsp application. The
2826 revocation information is handled using the text based index
2827 use by the ca application. The responder can either handle
2828 requests generated internally, supplied in files (for example
2829 via a CGI script) or using an internal minimal server.
2832 *) Add configuration choices to get zlib compression for TLS.
2835 *) Changes to Kerberos SSL for RFC 2712 compliance:
2836 1. Implemented real KerberosWrapper, instead of just using
2837 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2838 2. Implemented optional authenticator field of KerberosWrapper.
2840 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2841 and authenticator structs; see crypto/krb5/.
2843 Generalized Kerberos calls to support multiple Kerberos libraries.
2844 [Vern Staats <staatsvr@asc.hpc.mil>,
2845 Jeffrey Altman <jaltman@columbia.edu>
2846 via Richard Levitte]
2848 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2849 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2850 values for each of the key sizes rather than having just
2851 parameters (and 'speed' generating keys each time).
2854 *) Speed up EVP routines.
2857 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2858 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2859 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2860 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2862 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2863 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2864 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2867 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2869 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2872 *) Added the OS2-EMX target.
2873 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2875 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2876 to support NCONF routines in extension code. New function CONF_set_nconf()
2877 to allow functions which take an NCONF to also handle the old LHASH
2878 structure: this means that the old CONF compatible routines can be
2879 retained (in particular wrt extensions) without having to duplicate the
2880 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2883 *) Enhance the general user interface with mechanisms for inner control
2884 and with possibilities to have yes/no kind of prompts.
2887 *) Change all calls to low level digest routines in the library and
2888 applications to use EVP. Add missing calls to HMAC_cleanup() and
2889 don't assume HMAC_CTX can be copied using memcpy().
2890 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2892 *) Add the possibility to control engines through control names but with
2893 arbitrary arguments instead of just a string.
2894 Change the key loaders to take a UI_METHOD instead of a callback
2895 function pointer. NOTE: this breaks binary compatibility with earlier
2896 versions of OpenSSL [engine].
2897 Adapt the nCipher code for these new conditions and add a card insertion
2901 *) Enhance the general user interface with mechanisms to better support
2902 dialog box interfaces, application-defined prompts, the possibility
2903 to use defaults (for example default passwords from somewhere else)
2904 and interrupts/cancellations.
2907 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2908 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2911 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2912 tidy up some unnecessarily weird code in 'sk_new()').
2913 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2915 *) Change the key loading routines for ENGINEs to use the same kind
2916 callback (pem_password_cb) as all other routines that need this
2920 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2921 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2922 than this minimum value is recommended.
2925 *) New random seeder for OpenVMS, using the system process statistics
2926 that are easily reachable.
2929 *) Windows apparently can't transparently handle global
2930 variables defined in DLLs. Initialisations such as:
2932 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2934 wont compile. This is used by the any applications that need to
2935 declare their own ASN1 modules. This was fixed by adding the option
2936 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2937 needed for static libraries under Win32.
2940 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2941 setting of purpose and trust fields. New X509_STORE trust and
2942 purpose functions and tidy up setting in other SSL functions.
2945 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2946 structure. These are inherited by X509_STORE_CTX when it is
2947 initialised. This allows various defaults to be set in the
2948 X509_STORE structure (such as flags for CRL checking and custom
2949 purpose or trust settings) for functions which only use X509_STORE_CTX
2950 internally such as S/MIME.
2952 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2953 trust settings if they are not set in X509_STORE. This allows X509_STORE
2954 purposes and trust (in S/MIME for example) to override any set by default.
2956 Add command line options for CRL checking to smime, s_client and s_server
2960 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2961 are set then the CRL is looked up in the X509_STORE structure and
2962 its validity and signature checked, then if the certificate is found
2963 in the CRL the verify fails with a revoked error.
2965 Various new CRL related callbacks added to X509_STORE_CTX structure.
2967 Command line options added to 'verify' application to support this.
2969 This needs some additional work, such as being able to handle multiple
2970 CRLs with different times, extension based lookup (rather than just
2971 by subject name) and ultimately more complete V2 CRL extension
2975 *) Add a general user interface API (crypto/ui/). This is designed
2976 to replace things like des_read_password and friends (backward
2977 compatibility functions using this new API are provided).
2978 The purpose is to remove prompting functions from the DES code
2979 section as well as provide for prompting through dialog boxes in
2980 a window system and the like.
2983 *) Add "ex_data" support to ENGINE so implementations can add state at a
2984 per-structure level rather than having to store it globally.
2987 *) Make it possible for ENGINE structures to be copied when retrieved by
2988 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2989 This causes the "original" ENGINE structure to act like a template,
2990 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2991 operational state can be localised to each ENGINE structure, despite the
2992 fact they all share the same "methods". New ENGINE structures returned in
2993 this case have no functional references and the return value is the single
2994 structural reference. This matches the single structural reference returned
2995 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2999 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3000 needs to match any other type at all we need to manually clear the
3004 *) Changes to the "openssl engine" utility to include;
3005 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3006 about an ENGINE's available control commands.
3007 - executing control commands from command line arguments using the
3008 '-pre' and '-post' switches. '-post' is only used if '-t' is
3009 specified and the ENGINE is successfully initialised. The syntax for
3010 the individual commands are colon-separated, for example;
3011 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3014 *) New dynamic control command support for ENGINEs. ENGINEs can now
3015 declare their own commands (numbers), names (strings), descriptions,
3016 and input types for run-time discovery by calling applications. A
3017 subset of these commands are implicitly classed as "executable"
3018 depending on their input type, and only these can be invoked through
3019 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3020 can be based on user input, config files, etc). The distinction is
3021 that "executable" commands cannot return anything other than a boolean
3022 result and can only support numeric or string input, whereas some
3023 discoverable commands may only be for direct use through
3024 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3025 pointers, or other custom uses. The "executable" commands are to
3026 support parameterisations of ENGINE behaviour that can be
3027 unambiguously defined by ENGINEs and used consistently across any
3028 OpenSSL-based application. Commands have been added to all the
3029 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3030 control over shared-library paths without source code alterations.
3033 *) Changed all ENGINE implementations to dynamically allocate their
3034 ENGINEs rather than declaring them statically. Apart from this being
3035 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3036 this also allows the implementations to compile without using the
3037 internal engine_int.h header.
3040 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3041 'const' value. Any code that should be able to modify a RAND_METHOD
3042 should already have non-const pointers to it (ie. they should only
3043 modify their own ones).
3046 *) Made a variety of little tweaks to the ENGINE code.
3047 - "atalla" and "ubsec" string definitions were moved from header files
3048 to C code. "nuron" string definitions were placed in variables
3049 rather than hard-coded - allowing parameterisation of these values
3050 later on via ctrl() commands.
3051 - Removed unused "#if 0"'d code.
3052 - Fixed engine list iteration code so it uses ENGINE_free() to release
3053 structural references.
3054 - Constified the RAND_METHOD element of ENGINE structures.
3055 - Constified various get/set functions as appropriate and added
3056 missing functions (including a catch-all ENGINE_cpy that duplicates
3057 all ENGINE values onto a new ENGINE except reference counts/state).
3058 - Removed NULL parameter checks in get/set functions. Setting a method
3059 or function to NULL is a way of cancelling out a previously set
3060 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3061 and doesn't justify the extra error symbols and code.
3062 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3063 flags from engine_int.h to engine.h.
3064 - Changed prototypes for ENGINE handler functions (init(), finish(),
3065 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3068 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3069 to the algorithm using long division. The binary algorithm can be
3070 used only if the modulus is odd. On 32-bit systems, it is faster
3071 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3072 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3073 up to 450 bits. In 64-bit environments, the binary algorithm
3074 appears to be advantageous for much longer moduli; here we use it
3075 for moduli up to 2048 bits.
3078 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3079 could not support the combine flag in choice fields.
3082 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3083 extensions from a certificate request to the certificate.
3086 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3087 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3088 file: this allows the display of the certificate about to be
3089 signed to be customised, to allow certain fields to be included
3090 or excluded and extension details. The old system didn't display
3091 multicharacter strings properly, omitted fields not in the policy
3092 and couldn't display additional details such as extensions.
3095 *) Function EC_POINTs_mul for multiple scalar multiplication
3096 of an arbitrary number of elliptic curve points
3097 \sum scalars[i]*points[i],
3098 optionally including the generator defined for the EC_GROUP:
3099 scalar*generator + \sum scalars[i]*points[i].
3101 EC_POINT_mul is a simple wrapper function for the typical case
3102 that the point list has just one item (besides the optional
3106 *) First EC_METHODs for curves over GF(p):
3108 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3109 operations and provides various method functions that can also
3110 operate with faster implementations of modular arithmetic.
3112 EC_GFp_mont_method() reuses most functions that are part of
3113 EC_GFp_simple_method, but uses Montgomery arithmetic.
3115 [Bodo Moeller; point addition and point doubling
3116 implementation directly derived from source code provided by
3117 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3119 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3120 crypto/ec/ec_lib.c):
3122 Curves are EC_GROUP objects (with an optional group generator)
3123 based on EC_METHODs that are built into the library.
3125 Points are EC_POINT objects based on EC_GROUP objects.
3127 Most of the framework would be able to handle curves over arbitrary
3128 finite fields, but as there are no obvious types for fields other
3129 than GF(p), some functions are limited to that for now.
3132 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3133 that the file contains a complete HTTP response.
3136 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3137 change the def and num file printf format specifier from "%-40sXXX"
3138 to "%-39s XXX". The latter will always guarantee a space after the
3139 field while the former will cause them to run together if the field
3140 is 40 of more characters long.
3143 *) Constify the cipher and digest 'method' functions and structures
3144 and modify related functions to take constant EVP_MD and EVP_CIPHER
3148 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3149 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3152 *) Modify EVP_Digest*() routines so they now return values. Although the
3153 internal software routines can never fail additional hardware versions
3157 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3159 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3160 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3163 ERR_R_NESTED_ASN1_ERROR
3165 ERR_R_MISSING_ASN1_EOS
3166 were 4 .. 9, conflicting with
3167 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3169 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3170 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3172 Add new error code 'ERR_R_INTERNAL_ERROR'.
3175 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3179 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3180 sets the subject name for a new request or supersedes the
3181 subject name in a given request. Formats that can be parsed are
3182 'CN=Some Name, OU=myOU, C=IT'
3184 'CN=Some Name/OU=myOU/C=IT'.
3186 Add options '-batch' and '-verbose' to 'openssl req'.
3187 [Massimiliano Pala <madwolf@hackmasters.net>]
3189 *) Introduce the possibility to access global variables through
3190 functions on platform were that's the best way to handle exporting
3191 global variables in shared libraries. To enable this functionality,
3192 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3193 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3194 is normally done by Configure or something similar).
3196 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3197 in the source file (foo.c) like this:
3199 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3200 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3202 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3203 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3205 OPENSSL_DECLARE_GLOBAL(int,foo);
3206 #define foo OPENSSL_GLOBAL_REF(foo)
3207 OPENSSL_DECLARE_GLOBAL(double,bar);
3208 #define bar OPENSSL_GLOBAL_REF(bar)
3210 The #defines are very important, and therefore so is including the
3211 header file everywhere where the defined globals are used.
3213 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3214 of ASN.1 items, but that structure is a bit different.
3216 The largest change is in util/mkdef.pl which has been enhanced with
3217 better and easier to understand logic to choose which symbols should
3218 go into the Windows .def files as well as a number of fixes and code
3219 cleanup (among others, algorithm keywords are now sorted
3220 lexicographically to avoid constant rewrites).
3223 *) In BN_div() keep a copy of the sign of 'num' before writing the
3224 result to 'rm' because if rm==num the value will be overwritten
3225 and produce the wrong result if 'num' is negative: this caused
3226 problems with BN_mod() and BN_nnmod().
3229 *) Function OCSP_request_verify(). This checks the signature on an
3230 OCSP request and verifies the signer certificate. The signer
3231 certificate is just checked for a generic purpose and OCSP request
3235 *) Add OCSP_check_validity() function to check the validity of OCSP
3236 responses. OCSP responses are prepared in real time and may only
3237 be a few seconds old. Simply checking that the current time lies
3238 between thisUpdate and nextUpdate max reject otherwise valid responses
3239 caused by either OCSP responder or client clock inaccuracy. Instead
3240 we allow thisUpdate and nextUpdate to fall within a certain period of
3241 the current time. The age of the response can also optionally be
3242 checked. Two new options -validity_period and -status_age added to
3246 *) If signature or public key algorithm is unrecognized print out its
3247 OID rather that just UNKNOWN.
3250 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3251 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3252 ID to be generated from the issuer certificate alone which can then be
3253 passed to OCSP_id_issuer_cmp().
3256 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3257 ASN1 modules to export functions returning ASN1_ITEM pointers
3258 instead of the ASN1_ITEM structures themselves. This adds several
3259 new macros which allow the underlying ASN1 function/structure to
3260 be accessed transparently. As a result code should not use ASN1_ITEM
3261 references directly (such as &X509_it) but instead use the relevant
3262 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3263 use of the new ASN1 code on platforms where exporting structures
3264 is problematical (for example in shared libraries) but exporting
3265 functions returning pointers to structures is not.
3268 *) Add support for overriding the generation of SSL/TLS session IDs.
3269 These callbacks can be registered either in an SSL_CTX or per SSL.
3270 The purpose of this is to allow applications to control, if they wish,
3271 the arbitrary values chosen for use as session IDs, particularly as it
3272 can be useful for session caching in multiple-server environments. A
3273 command-line switch for testing this (and any client code that wishes
3274 to use such a feature) has been added to "s_server".
3275 [Geoff Thorpe, Lutz Jaenicke]
3277 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3278 of the form '#if defined(...) || defined(...) || ...' and
3279 '#if !defined(...) && !defined(...) && ...'. This also avoids
3280 the growing number of special cases it was previously handling.
3283 *) Make all configuration macros available for application by making
3284 sure they are available in opensslconf.h, by giving them names starting
3285 with "OPENSSL_" to avoid conflicts with other packages and by making
3286 sure e_os2.h will cover all platform-specific cases together with
3288 Additionally, it is now possible to define configuration/platform-
3289 specific names (called "system identities"). In the C code, these
3290 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3291 macro with the name beginning with "OPENSSL_SYS_", which is determined
3292 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3296 *) New option -set_serial to 'req' and 'x509' this allows the serial
3297 number to use to be specified on the command line. Previously self
3298 signed certificates were hard coded with serial number 0 and the
3299 CA options of 'x509' had to use a serial number in a file which was
3303 *) New options to 'ca' utility to support V2 CRL entry extensions.
3304 Currently CRL reason, invalidity date and hold instruction are
3305 supported. Add new CRL extensions to V3 code and some new objects.
3308 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3309 disable standard block padding (aka PKCS#5 padding) in the EVP
3310 API, which was previously mandatory. This means that the data is
3311 not padded in any way and so the total length much be a multiple
3312 of the block size, otherwise an error occurs.
3315 *) Initial (incomplete) OCSP SSL support.
3318 *) New function OCSP_parse_url(). This splits up a URL into its host,
3319 port and path components: primarily to parse OCSP URLs. New -url
3320 option to ocsp utility.
3323 *) New nonce behavior. The return value of OCSP_check_nonce() now
3324 reflects the various checks performed. Applications can decide
3325 whether to tolerate certain situations such as an absent nonce
3326 in a response when one was present in a request: the ocsp application
3327 just prints out a warning. New function OCSP_add1_basic_nonce()
3328 this is to allow responders to include a nonce in a response even if
3329 the request is nonce-less.
3332 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3333 skipped when using openssl x509 multiple times on a single input file,
3334 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3337 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3338 set string type: to handle setting ASN1_TIME structures. Fix ca
3339 utility to correctly initialize revocation date of CRLs.
3342 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3343 the clients preferred ciphersuites and rather use its own preferences.
3344 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3345 Internet Explorer by ensuring unchanged hash method during stepup.
3346 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3349 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3350 to aes and add a new 'exist' option to print out symbols that don't
3354 *) Additional options to ocsp utility to allow flags to be set and
3355 additional certificates supplied.
3358 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3359 OCSP client a number of certificate to only verify the response
3363 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3364 handle the new API. Currently only ECB, CBC modes supported. Add new
3367 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3368 Encryption Standard (AES) Ciphersuites for Transport Layer
3369 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3370 not enabled by default and were not part of the "ALL" ciphersuite
3371 alias because they were not yet official; they could be
3372 explicitly requested by specifying the "AESdraft" ciphersuite
3373 group alias. In the final release of OpenSSL 0.9.7, the group
3374 alias is called "AES" and is part of "ALL".)
3375 [Ben Laurie, Steve Henson, Bodo Moeller]
3377 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3378 request to response.
3381 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3382 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3383 extract information from a certificate request. OCSP_response_create()
3384 creates a response and optionally adds a basic response structure.
3385 OCSP_basic_add1_status() adds a complete single response to a basic
3386 response and returns the OCSP_SINGLERESP structure just added (to allow
3387 extensions to be included for example). OCSP_basic_add1_cert() adds a
3388 certificate to a basic response and OCSP_basic_sign() signs a basic
3389 response with various flags. New helper functions ASN1_TIME_check()
3390 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3391 (converts ASN1_TIME to GeneralizedTime).
3394 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3395 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3396 structure from a certificate. X509_pubkey_digest() digests the public_key
3397 contents: this is used in various key identifiers.
3400 *) Make sk_sort() tolerate a NULL argument.
3401 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3403 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3404 passed by the function are trusted implicitly. If any of them signed the
3405 response then it is assumed to be valid and is not verified.
3408 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3409 to data. This was previously part of the PKCS7 ASN1 code. This
3410 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3411 [Steve Henson, reported by Kenneth R. Robinette
3412 <support@securenetterm.com>]
3414 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3415 routines: without these tracing memory leaks is very painful.
3416 Fix leaks in PKCS12 and PKCS7 routines.
3419 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3420 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3421 effectively meant GeneralizedTime would never be used. Now it
3422 is initialised to -1 but X509_time_adj() now has to check the value
3423 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3424 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3425 [Steve Henson, reported by Kenneth R. Robinette
3426 <support@securenetterm.com>]
3428 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3429 result in a zero length in the ASN1_INTEGER structure which was
3430 not consistent with the structure when d2i_ASN1_INTEGER() was used
3431 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3432 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3433 where it did not print out a minus for negative ASN1_INTEGER.
3436 *) Add summary printout to ocsp utility. The various functions which
3437 convert status values to strings have been renamed to:
3438 OCSP_response_status_str(), OCSP_cert_status_str() and
3439 OCSP_crl_reason_str() and are no longer static. New options
3440 to verify nonce values and to disable verification. OCSP response
3441 printout format cleaned up.
3444 *) Add additional OCSP certificate checks. These are those specified
3445 in RFC2560. This consists of two separate checks: the CA of the
3446 certificate being checked must either be the OCSP signer certificate
3447 or the issuer of the OCSP signer certificate. In the latter case the
3448 OCSP signer certificate must contain the OCSP signing extended key
3449 usage. This check is performed by attempting to match the OCSP
3450 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3451 in the OCSP_CERTID structures of the response.
3454 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3455 and related routines. This uses the standard OpenSSL certificate
3456 verify routines to perform initial checks (just CA validity) and
3457 to obtain the certificate chain. Then additional checks will be
3458 performed on the chain. Currently the root CA is checked to see
3459 if it is explicitly trusted for OCSP signing. This is used to set
3460 a root CA as a global signing root: that is any certificate that
3461 chains to that CA is an acceptable OCSP signing certificate.
3464 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3465 extensions from a separate configuration file.
3466 As when reading extensions from the main configuration file,
3467 the '-extensions ...' option may be used for specifying the
3469 [Massimiliano Pala <madwolf@comune.modena.it>]
3471 *) New OCSP utility. Allows OCSP requests to be generated or
3472 read. The request can be sent to a responder and the output
3473 parsed, outputed or printed in text form. Not complete yet:
3474 still needs to check the OCSP response validity.
3477 *) New subcommands for 'openssl ca':
3478 'openssl ca -status <serial>' prints the status of the cert with
3479 the given serial number (according to the index file).
3480 'openssl ca -updatedb' updates the expiry status of certificates
3482 [Massimiliano Pala <madwolf@comune.modena.it>]
3484 *) New '-newreq-nodes' command option to CA.pl. This is like
3485 '-newreq', but calls 'openssl req' with the '-nodes' option
3486 so that the resulting key is not encrypted.
3487 [Damien Miller <djm@mindrot.org>]
3489 *) New configuration for the GNU Hurd.
3490 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3492 *) Initial code to implement OCSP basic response verify. This
3493 is currently incomplete. Currently just finds the signer's
3494 certificate and verifies the signature on the response.
3497 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3498 value of OPENSSLDIR. This is available via the new '-d' option
3499 to 'openssl version', and is also included in 'openssl version -a'.
3502 *) Allowing defining memory allocation callbacks that will be given
3503 file name and line number information in additional arguments
3504 (a const char* and an int). The basic functionality remains, as
3505 well as the original possibility to just replace malloc(),
3506 realloc() and free() by functions that do not know about these
3507 additional arguments. To register and find out the current
3508 settings for extended allocation functions, the following
3509 functions are provided:
3511 CRYPTO_set_mem_ex_functions
3512 CRYPTO_set_locked_mem_ex_functions
3513 CRYPTO_get_mem_ex_functions
3514 CRYPTO_get_locked_mem_ex_functions
3516 These work the same way as CRYPTO_set_mem_functions and friends.
3517 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3518 extended allocation function is enabled.
3519 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3520 a conventional allocation function is enabled.
3521 [Richard Levitte, Bodo Moeller]
3523 *) Finish off removing the remaining LHASH function pointer casts.
3524 There should no longer be any prototype-casting required when using
3525 the LHASH abstraction, and any casts that remain are "bugs". See
3526 the callback types and macros at the head of lhash.h for details
3527 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3530 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3531 If /dev/[u]random devices are not available or do not return enough
3532 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3534 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3535 /etc/entropy will be queried once each in this sequence, quering stops
3536 when enough entropy was collected without querying more sockets.
3539 *) Change the Unix RAND_poll() variant to be able to poll several
3540 random devices, as specified by DEVRANDOM, until a sufficient amount
3541 of data has been collected. We spend at most 10 ms on each file
3542 (select timeout) and read in non-blocking mode. DEVRANDOM now
3543 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3544 (previously it was just the string "/dev/urandom"), so on typical
3545 platforms the 10 ms delay will never occur.
3546 Also separate out the Unix variant to its own file, rand_unix.c.
3547 For VMS, there's a currently-empty rand_vms.c.
3550 *) Move OCSP client related routines to ocsp_cl.c. These
3551 provide utility functions which an application needing
3552 to issue a request to an OCSP responder and analyse the
3553 response will typically need: as opposed to those which an
3554 OCSP responder itself would need which will be added later.
3556 OCSP_request_sign() signs an OCSP request with an API similar
3557 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3558 response. OCSP_response_get1_basic() extracts basic response
3559 from response. OCSP_resp_find_status(): finds and extracts status
3560 information from an OCSP_CERTID structure (which will be created
3561 when the request structure is built). These are built from lower
3562 level functions which work on OCSP_SINGLERESP structures but
3563 wont normally be used unless the application wishes to examine
3564 extensions in the OCSP response for example.
3566 Replace nonce routines with a pair of functions.
3567 OCSP_request_add1_nonce() adds a nonce value and optionally
3568 generates a random value. OCSP_check_nonce() checks the
3569 validity of the nonce in an OCSP response.
3572 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3573 This doesn't copy the supplied OCSP_CERTID and avoids the
3574 need to free up the newly created id. Change return type
3575 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3576 This can then be used to add extensions to the request.
3577 Deleted OCSP_request_new(), since most of its functionality
3578 is now in OCSP_REQUEST_new() (and the case insensitive name
3579 clash) apart from the ability to set the request name which
3580 will be added elsewhere.
3583 *) Update OCSP API. Remove obsolete extensions argument from
3584 various functions. Extensions are now handled using the new
3585 OCSP extension code. New simple OCSP HTTP function which
3586 can be used to send requests and parse the response.
3589 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3590 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3591 uses the special reorder version of SET OF to sort the attributes
3592 and reorder them to match the encoded order. This resolves a long
3593 standing problem: a verify on a PKCS7 structure just after signing
3594 it used to fail because the attribute order did not match the
3595 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3596 it uses the received order. This is necessary to tolerate some broken
3597 software that does not order SET OF. This is handled by encoding
3598 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3599 to produce the required SET OF.
3602 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3603 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3604 files to get correct declarations of the ASN.1 item variables.
3607 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3608 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3609 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3610 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3611 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3612 ASN1_ITEM and no wrapper functions.
3615 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3616 replace the old function pointer based I/O routines. Change most of
3617 the *_d2i_bio() and *_d2i_fp() functions to use these.
3620 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3621 lines, recognice more "algorithms" that can be deselected, and make
3622 it complain about algorithm deselection that isn't recognised.
3625 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3626 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3627 to use new functions. Add NO_ASN1_OLD which can be set to remove
3628 some old style ASN1 functions: this can be used to determine if old
3629 code will still work when these eventually go away.
3632 *) New extension functions for OCSP structures, these follow the
3633 same conventions as certificates and CRLs.
3636 *) New function X509V3_add1_i2d(). This automatically encodes and
3637 adds an extension. Its behaviour can be customised with various
3638 flags to append, replace or delete. Various wrappers added for
3639 certifcates and CRLs.
3642 *) Fix to avoid calling the underlying ASN1 print routine when
3643 an extension cannot be parsed. Correct a typo in the
3644 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3647 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3648 entries for variables.
3651 *) Add functionality to apps/openssl.c for detecting locking
3652 problems: As the program is single-threaded, all we have
3653 to do is register a locking callback using an array for
3654 storing which locks are currently held by the program.
3657 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3658 SSL_get_ex_data_X509_STORE_idx(), which is used in
3659 ssl_verify_cert_chain() and thus can be called at any time
3660 during TLS/SSL handshakes so that thread-safety is essential.
3661 Unfortunately, the ex_data design is not at all suited
3662 for multi-threaded use, so it probably should be abolished.
3665 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3666 [Broadcom, tweaked and integrated by Geoff Thorpe]
3668 *) Move common extension printing code to new function
3669 X509V3_print_extensions(). Reorganise OCSP print routines and
3670 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3673 *) New function X509_signature_print() to remove duplication in some
3677 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3678 set (this was treated exactly the same as SET OF previously). This
3679 is used to reorder the STACK representing the structure to match the
3680 encoding. This will be used to get round a problem where a PKCS7
3681 structure which was signed could not be verified because the STACK
3682 order did not reflect the encoded order.
3685 *) Reimplement the OCSP ASN1 module using the new code.
3688 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3689 for its ASN1 operations. The old style function pointers still exist
3690 for now but they will eventually go away.
3693 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3694 completely replaces the old ASN1 functionality with a table driven
3695 encoder and decoder which interprets an ASN1_ITEM structure describing
3696 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3697 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3698 has also been converted to the new form.
3701 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3702 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3703 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3704 for negative moduli.
3707 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3708 of not touching the result's sign bit.
3711 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3715 *) Changed the LHASH code to use prototypes for callbacks, and created
3716 macros to declare and implement thin (optionally static) functions
3717 that provide type-safety and avoid function pointer casting for the
3718 type-specific callbacks.
3721 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3723 [Veers Staats <staatsvr@asc.hpc.mil>,
3724 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3726 *) Reformat the FAQ so the different questions and answers can be divided
3727 in sections depending on the subject.
3730 *) Have the zlib compression code load ZLIB.DLL dynamically under
3734 *) New function BN_mod_sqrt for computing square roots modulo a prime
3735 (using the probabilistic Tonelli-Shanks algorithm unless
3736 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3737 be handled deterministically).
3738 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3740 *) Make BN_mod_inverse faster by explicitly handling small quotients
3741 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3742 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3745 *) New function BN_kronecker.
3748 *) Fix BN_gcd so that it works on negative inputs; the result is
3749 positive unless both parameters are zero.
3750 Previously something reasonably close to an infinite loop was
3751 possible because numbers could be growing instead of shrinking
3752 in the implementation of Euclid's algorithm.
3755 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3756 sign of the number in question.
3758 Fix BN_is_word(a,w) to work correctly for w == 0.
3760 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3761 because its test if the absolute value of 'a' equals 'w'.
3762 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3763 it exists mostly for use in the implementations of BN_is_zero(),
3764 BN_is_one(), and BN_is_word().
3767 *) New function BN_swap.
3770 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3771 the exponentiation functions are more likely to produce reasonable
3772 results on negative inputs.
3775 *) Change BN_mod_mul so that the result is always non-negative.
3776 Previously, it could be negative if one of the factors was negative;
3777 I don't think anyone really wanted that behaviour.
3780 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3781 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3782 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3783 and add new functions:
3792 BN_mod_lshift1_quick
3796 These functions always generate non-negative results.
3798 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3799 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3801 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3802 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3803 be reduced modulo m.
3804 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3807 The following entry accidentily appeared in the CHANGES file
3808 distributed with OpenSSL 0.9.7. The modifications described in
3809 it do *not* apply to OpenSSL 0.9.7.
3811 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3812 was actually never needed) and in BN_mul(). The removal in BN_mul()
3813 required a small change in bn_mul_part_recursive() and the addition
3814 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3815 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3816 bn_sub_words() and bn_add_words() except they take arrays with
3821 *) In 'openssl passwd', verify passwords read from the terminal
3822 unless the '-salt' option is used (which usually means that
3823 verification would just waste user's time since the resulting
3824 hash is going to be compared with some given password hash)
3825 or the new '-noverify' option is used.
3827 This is an incompatible change, but it does not affect
3828 non-interactive use of 'openssl passwd' (passwords on the command
3829 line, '-stdin' option, '-in ...' option) and thus should not
3833 *) Remove all references to RSAref, since there's no more need for it.
3836 *) Make DSO load along a path given through an environment variable
3837 (SHLIB_PATH) with shl_load().
3840 *) Constify the ENGINE code as a result of BIGNUM constification.
3841 Also constify the RSA code and most things related to it. In a
3842 few places, most notable in the depth of the ASN.1 code, ugly
3843 casts back to non-const were required (to be solved at a later
3847 *) Make it so the openssl application has all engines loaded by default.
3850 *) Constify the BIGNUM routines a little more.
3853 *) Add the following functions:
3855 ENGINE_load_cswift()
3857 ENGINE_load_atalla()
3859 ENGINE_load_builtin_engines()
3861 That way, an application can itself choose if external engines that
3862 are built-in in OpenSSL shall ever be used or not. The benefit is
3863 that applications won't have to be linked with libdl or other dso
3864 libraries unless it's really needed.
3866 Changed 'openssl engine' to load all engines on demand.
3867 Changed the engine header files to avoid the duplication of some
3868 declarations (they differed!).
3871 *) 'openssl engine' can now list capabilities.
3874 *) Better error reporting in 'openssl engine'.
3877 *) Never call load_dh_param(NULL) in s_server.
3880 *) Add engine application. It can currently list engines by name and
3881 identity, and test if they are actually available.
3884 *) Improve RPM specification file by forcing symbolic linking and making
3885 sure the installed documentation is also owned by root.root.
3886 [Damien Miller <djm@mindrot.org>]
3888 *) Give the OpenSSL applications more possibilities to make use of
3889 keys (public as well as private) handled by engines.
3892 *) Add OCSP code that comes from CertCo.
3895 *) Add VMS support for the Rijndael code.
3898 *) Added untested support for Nuron crypto accelerator.
3901 *) Add support for external cryptographic devices. This code was
3902 previously distributed separately as the "engine" branch.
3903 [Geoff Thorpe, Richard Levitte]
3905 *) Rework the filename-translation in the DSO code. It is now possible to
3906 have far greater control over how a "name" is turned into a filename
3907 depending on the operating environment and any oddities about the
3908 different shared library filenames on each system.
3911 *) Support threads on FreeBSD-elf in Configure.
3914 *) Fix for SHA1 assembly problem with MASM: it produces
3915 warnings about corrupt line number information when assembling
3916 with debugging information. This is caused by the overlapping
3918 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3921 NCONF_get_number() has no error checking at all. As a replacement,
3922 NCONF_get_number_e() is defined (_e for "error checking") and is
3923 promoted strongly. The old NCONF_get_number is kept around for
3924 binary backward compatibility.
3925 Make it possible for methods to load from something other than a BIO,
3926 by providing a function pointer that is given a name instead of a BIO.
3927 For example, this could be used to load configuration data from an
3931 *) Fix for non blocking accept BIOs. Added new I/O special reason
3932 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3933 with non blocking I/O was not possible because no retry code was
3934 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3938 *) Added the beginnings of Rijndael support.
3941 *) Fix for bug in DirectoryString mask setting. Add support for
3942 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3943 to allow certificate printing to more controllable, additional
3944 'certopt' option to 'x509' to allow new printing options to be
3948 *) Clean old EAY MD5 hack from e_os.h.
3951 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3953 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3954 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3955 [Joe Orton, Steve Henson]
3957 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3959 *) Fix additional bug revealed by the NISCC test suite:
3961 Stop bug triggering large recursion when presented with
3962 certain ASN.1 tags (CVE-2003-0851)
3965 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3967 *) Fix various bugs revealed by running the NISCC test suite:
3969 Stop out of bounds reads in the ASN1 code when presented with
3970 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3972 If verify callback ignores invalid public key errors don't try to check
3973 certificate signature with the NULL public key.
3977 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3978 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3982 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3983 extra data after the compression methods not only for TLS 1.0
3984 but also for SSL 3.0 (as required by the specification).
3985 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3987 *) Change X509_certificate_type() to mark the key as exported/exportable
3988 when it's 512 *bits* long, not 512 bytes.
3991 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3993 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3994 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3995 a protocol version number mismatch like a decryption error
3996 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3999 *) Turn on RSA blinding by default in the default implementation
4000 to avoid a timing attack. Applications that don't want it can call
4001 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4002 They would be ill-advised to do so in most cases.
4003 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4005 *) Change RSA blinding code so that it works when the PRNG is not
4006 seeded (in this case, the secret RSA exponent is abused as
4007 an unpredictable seed -- if it is not unpredictable, there
4008 is no point in blinding anyway). Make RSA blinding thread-safe
4009 by remembering the creator's thread ID in rsa->blinding and
4010 having all other threads use local one-time blinding factors
4011 (this requires more computation than sharing rsa->blinding, but
4012 avoids excessive locking; and if an RSA object is not shared
4013 between threads, blinding will still be very fast).
4016 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4018 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4019 via timing by performing a MAC computation even if incorrrect
4020 block cipher padding has been found. This is a countermeasure
4021 against active attacks where the attacker has to distinguish
4022 between bad padding and a MAC verification error. (CVE-2003-0078)
4024 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4025 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4026 Martin Vuagnoux (EPFL, Ilion)]
4028 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4030 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4031 memory from it's contents. This is done with a counter that will
4032 place alternating values in each byte. This can be used to solve
4033 two issues: 1) the removal of calls to memset() by highly optimizing
4034 compilers, and 2) cleansing with other values than 0, since those can
4035 be read through on certain media, for example a swap space on disk.
4038 *) Bugfix: client side session caching did not work with external caching,
4039 because the session->cipher setting was not restored when reloading
4040 from the external cache. This problem was masked, when
4041 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4042 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4045 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4046 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4047 [Zeev Lieber <zeev-l@yahoo.com>]
4049 *) Undo an undocumented change introduced in 0.9.6e which caused
4050 repeated calls to OpenSSL_add_all_ciphers() and
4051 OpenSSL_add_all_digests() to be ignored, even after calling
4055 *) Change the default configuration reader to deal with last line not
4056 being properly terminated.
4059 *) Change X509_NAME_cmp() so it applies the special rules on handling
4060 DN values that are of type PrintableString, as well as RDNs of type
4061 emailAddress where the value has the type ia5String.
4062 [stefank@valicert.com via Richard Levitte]
4064 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4065 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4066 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4067 the bitwise-OR of the two for use by the majority of applications
4068 wanting this behaviour, and update the docs. The documented
4069 behaviour and actual behaviour were inconsistent and had been
4070 changing anyway, so this is more a bug-fix than a behavioural
4072 [Geoff Thorpe, diagnosed by Nadav Har'El]
4074 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4075 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4078 *) Fix initialization code race conditions in
4079 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4080 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4081 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4082 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4083 ssl2_get_cipher_by_char(),
4084 ssl3_get_cipher_by_char().
4085 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4087 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4088 the cached sessions are flushed, as the remove_cb() might use ex_data
4089 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4090 (see [openssl.org #212]).
4091 [Geoff Thorpe, Lutz Jaenicke]
4093 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4094 length, instead of the encoding length to d2i_ASN1_OBJECT.
4097 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4099 *) [In 0.9.6g-engine release:]
4100 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4101 [Lynn Gazis <lgazis@rainbow.com>]
4103 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4105 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4106 and get fix the header length calculation.
4107 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4108 Alon Kantor <alonk@checkpoint.com> (and others),
4111 *) Use proper error handling instead of 'assertions' in buffer
4112 overflow checks added in 0.9.6e. This prevents DoS (the
4113 assertions could call abort()).
4114 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4116 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4118 *) Add various sanity checks to asn1_get_length() to reject
4119 the ASN1 length bytes if they exceed sizeof(long), will appear
4120 negative or the content length exceeds the length of the
4122 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4124 *) Fix cipher selection routines: ciphers without encryption had no flags
4125 for the cipher strength set and where therefore not handled correctly
4126 by the selection routines (PR #130).
4129 *) Fix EVP_dsa_sha macro.
4133 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4134 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4135 that was added in OpenSSL 0.9.6d.
4137 As the countermeasure turned out to be incompatible with some
4138 broken SSL implementations, the new option is part of SSL_OP_ALL.
4139 SSL_OP_ALL is usually employed when compatibility with weird SSL
4140 implementations is desired (e.g. '-bugs' option to 's_client' and
4141 's_server'), so the new option is automatically set in many
4145 *) Changes in security patch:
4147 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4148 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4149 Air Force Materiel Command, USAF, under agreement number
4152 *) Add various sanity checks to asn1_get_length() to reject
4153 the ASN1 length bytes if they exceed sizeof(long), will appear
4154 negative or the content length exceeds the length of the
4155 supplied buffer. (CVE-2002-0659)
4156 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4158 *) Assertions for various potential buffer overflows, not known to
4160 [Ben Laurie (CHATS)]
4162 *) Various temporary buffers to hold ASCII versions of integers were
4163 too small for 64 bit platforms. (CVE-2002-0655)
4164 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4166 *) Remote buffer overflow in SSL3 protocol - an attacker could
4167 supply an oversized session ID to a client. (CVE-2002-0656)
4168 [Ben Laurie (CHATS)]
4170 *) Remote buffer overflow in SSL2 protocol - an attacker could
4171 supply an oversized client master key. (CVE-2002-0656)
4172 [Ben Laurie (CHATS)]
4174 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4176 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4177 encoded as NULL) with id-dsa-with-sha1.
4178 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4180 *) Check various X509_...() return values in apps/req.c.
4181 [Nils Larsch <nla@trustcenter.de>]
4183 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4184 an end-of-file condition would erronously be flagged, when the CRLF
4185 was just at the end of a processed block. The bug was discovered when
4186 processing data through a buffering memory BIO handing the data to a
4187 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4188 <ptsekov@syntrex.com> and Nedelcho Stanev.
4191 *) Implement a countermeasure against a vulnerability recently found
4192 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4193 before application data chunks to avoid the use of known IVs
4194 with data potentially chosen by the attacker.
4197 *) Fix length checks in ssl3_get_client_hello().
4200 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4201 to prevent ssl3_read_internal() from incorrectly assuming that
4202 ssl3_read_bytes() found application data while handshake
4203 processing was enabled when in fact s->s3->in_read_app_data was
4204 merely automatically cleared during the initial handshake.
4205 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4207 *) Fix object definitions for Private and Enterprise: they were not
4208 recognized in their shortname (=lowercase) representation. Extend
4209 obj_dat.pl to issue an error when using undefined keywords instead
4210 of silently ignoring the problem (Svenning Sorensen
4211 <sss@sss.dnsalias.net>).
4214 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4215 generators, i.e. generators other than 2 and 5. (Previously, the
4216 code did not properly initialise the 'add' and 'rem' values to
4217 BN_generate_prime().)
4219 In the new general case, we do not insist that 'generator' is
4220 actually a primitive root: This requirement is rather pointless;
4221 a generator of the order-q subgroup is just as good, if not
4225 *) Map new X509 verification errors to alerts. Discovered and submitted by
4226 Tom Wu <tom@arcot.com>.
4229 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4230 returning non-zero before the data has been completely received
4231 when using non-blocking I/O.
4232 [Bodo Moeller; problem pointed out by John Hughes]
4234 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4235 [Ben Laurie, Lutz Jaenicke]
4237 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4238 Yoram Zahavi <YoramZ@gilian.com>).
4241 *) Add information about CygWin 1.3 and on, and preserve proper
4242 configuration for the versions before that.
4243 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4245 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4246 check whether we deal with a copy of a session and do not delete from
4247 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4248 <izhar@checkpoint.com>.
4251 *) Do not store session data into the internal session cache, if it
4252 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4253 flag is set). Proposed by Aslam <aslam@funk.com>.
4256 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4260 *) [In 0.9.6d-engine release:]
4261 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4262 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4264 *) Add the configuration target linux-s390x.
4265 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4267 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4268 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4269 variable as an indication that a ClientHello message has been
4270 received. As the flag value will be lost between multiple
4271 invocations of ssl3_accept when using non-blocking I/O, the
4272 function may not be aware that a handshake has actually taken
4273 place, thus preventing a new session from being added to the
4276 To avoid this problem, we now set s->new_session to 2 instead of
4277 using a local variable.
4278 [Lutz Jaenicke, Bodo Moeller]
4280 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4281 if the SSL_R_LENGTH_MISMATCH error is detected.
4282 [Geoff Thorpe, Bodo Moeller]
4284 *) New 'shared_ldflag' column in Configure platform table.
4287 *) Fix EVP_CIPHER_mode macro.
4288 ["Dan S. Camper" <dan@bti.net>]
4290 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4291 type, we must throw them away by setting rr->length to 0.
4292 [D P Chang <dpc@qualys.com>]
4294 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4296 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4297 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4298 worked incorrectly for those cases where range = 10..._2 and
4299 3*range is two bits longer than range.)
4302 *) Only add signing time to PKCS7 structures if it is not already
4306 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4307 OBJ_ld_ce should be OBJ_id_ce.
4308 Also some ip-pda OIDs in crypto/objects/objects.txt were
4309 incorrect (cf. RFC 3039).
4310 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4312 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4313 returns early because it has nothing to do.
4314 [Andy Schneider <andy.schneider@bjss.co.uk>]
4316 *) [In 0.9.6c-engine release:]
4317 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4318 [Andy Schneider <andy.schneider@bjss.co.uk>]
4320 *) [In 0.9.6c-engine release:]
4321 Add support for Cryptographic Appliance's keyserver technology.
4322 (Use engine 'keyclient')
4323 [Cryptographic Appliances and Geoff Thorpe]
4325 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4326 is called via tools/c89.sh because arguments have to be
4327 rearranged (all '-L' options must appear before the first object
4329 [Richard Shapiro <rshapiro@abinitio.com>]
4331 *) [In 0.9.6c-engine release:]
4332 Add support for Broadcom crypto accelerator cards, backported
4334 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4336 *) [In 0.9.6c-engine release:]
4337 Add support for SureWare crypto accelerator cards from
4338 Baltimore Technologies. (Use engine 'sureware')
4339 [Baltimore Technologies and Mark Cox]
4341 *) [In 0.9.6c-engine release:]
4342 Add support for crypto accelerator cards from Accelerated
4343 Encryption Processing, www.aep.ie. (Use engine 'aep')
4344 [AEP Inc. and Mark Cox]
4346 *) Add a configuration entry for gcc on UnixWare.
4347 [Gary Benson <gbenson@redhat.com>]
4349 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4350 messages are stored in a single piece (fixed-length part and
4351 variable-length part combined) and fix various bugs found on the way.
4354 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4355 instead. BIO_gethostbyname() does not know what timeouts are
4356 appropriate, so entries would stay in cache even when they have
4358 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4360 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4361 faced with a pathologically small ClientHello fragment that does
4362 not contain client_version: Instead of aborting with an error,
4363 simply choose the highest available protocol version (i.e.,
4364 TLS 1.0 unless it is disabled). In practice, ClientHello
4365 messages are never sent like this, but this change gives us
4366 strictly correct behaviour at least for TLS.
4369 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4370 never resets s->method to s->ctx->method when called from within
4371 one of the SSL handshake functions.
4372 [Bodo Moeller; problem pointed out by Niko Baric]
4374 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4375 (sent using the client's version number) if client_version is
4376 smaller than the protocol version in use. Also change
4377 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4378 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4379 the client will at least see that alert.
4382 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4386 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4387 client receives HelloRequest while in a handshake.
4388 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4390 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4391 should end in 'break', not 'goto end' which circuments various
4392 cleanups done in state SSL_ST_OK. But session related stuff
4393 must be disabled for SSL_ST_OK in the case that we just sent a
4396 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4397 before just sending a HelloRequest.
4398 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4400 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4401 reveal whether illegal block cipher padding was found or a MAC
4402 verification error occured. (Neither SSLerr() codes nor alerts
4403 are directly visible to potential attackers, but the information
4404 may leak via logfiles.)
4406 Similar changes are not required for the SSL 2.0 implementation
4407 because the number of padding bytes is sent in clear for SSL 2.0,
4408 and the extra bytes are just ignored. However ssl/s2_pkt.c
4409 failed to verify that the purported number of padding bytes is in
4413 *) Add OpenUNIX-8 support including shared libraries
4414 (Boyd Lynn Gerber <gerberb@zenez.com>).
4417 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4418 'wristwatch attack' using huge encoding parameters (cf.
4419 James H. Manger's CRYPTO 2001 paper). Note that the
4420 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4421 encoding parameters and hence was not vulnerable.
4424 *) BN_sqr() bug fix.
4425 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4427 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4428 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4429 followed by modular reduction.
4430 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4432 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4433 equivalent based on BN_pseudo_rand() instead of BN_rand().
4436 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4437 This function was broken, as the check for a new client hello message
4438 to handle SGC did not allow these large messages.
4439 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4442 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4445 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4446 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4449 *) Rework the configuration and shared library support for Tru64 Unix.
4450 The configuration part makes use of modern compiler features and
4451 still retains old compiler behavior for those that run older versions
4452 of the OS. The shared library support part includes a variant that
4453 uses the RPATH feature, and is available through the special
4454 configuration target "alpha-cc-rpath", which will never be selected
4456 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4458 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4459 with the same message size as in ssl3_get_certificate_request().
4460 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4461 messages might inadvertently be reject as too long.
4462 [Petr Lampa <lampa@fee.vutbr.cz>]
4464 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4467 *) Modified SSL library such that the verify_callback that has been set
4468 specificly for an SSL object with SSL_set_verify() is actually being
4469 used. Before the change, a verify_callback set with this function was
4470 ignored and the verify_callback() set in the SSL_CTX at the time of
4471 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4472 to allow the necessary settings.
4475 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4476 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4477 done automatically (in contradiction to the requirements of the C
4478 standard). This made problems when used from OpenSSH.
4481 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4482 dh->length and always used
4484 BN_rand_range(priv_key, dh->p).
4486 BN_rand_range() is not necessary for Diffie-Hellman, and this
4487 specific range makes Diffie-Hellman unnecessarily inefficient if
4488 dh->length (recommended exponent length) is much smaller than the
4489 length of dh->p. We could use BN_rand_range() if the order of
4490 the subgroup was stored in the DH structure, but we only have
4495 BN_rand(priv_key, l, ...)
4497 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4503 RSA_eay_public_encrypt
4504 RSA_eay_private_decrypt
4505 RSA_eay_private_encrypt (signing)
4506 RSA_eay_public_decrypt (signature verification)
4508 (default implementations for RSA_public_encrypt,
4509 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4510 always reject numbers >= n.
4513 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4514 to synchronize access to 'locking_thread'. This is necessary on
4515 systems where access to 'locking_thread' (an 'unsigned long'
4516 variable) is not atomic.
4519 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4520 *before* setting the 'crypto_lock_rand' flag. The previous code had
4521 a race condition if 0 is a valid thread ID.
4522 [Travis Vitek <vitek@roguewave.com>]
4524 *) Add support for shared libraries under Irix.
4525 [Albert Chin-A-Young <china@thewrittenword.com>]
4527 *) Add configuration option to build on Linux on both big-endian and
4529 [Ralf Baechle <ralf@uni-koblenz.de>]
4531 *) Add the possibility to create shared libraries on HP-UX.
4534 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4536 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4537 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4538 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4539 PRNG state recovery was possible based on the output of
4540 one PRNG request appropriately sized to gain knowledge on
4541 'md' followed by enough consecutive 1-byte PRNG requests
4542 to traverse all of 'state'.
4544 1. When updating 'md_local' (the current thread's copy of 'md')
4545 during PRNG output generation, hash all of the previous
4546 'md_local' value, not just the half used for PRNG output.
4548 2. Make the number of bytes from 'state' included into the hash
4549 independent from the number of PRNG bytes requested.
4551 The first measure alone would be sufficient to avoid
4552 Markku-Juhani's attack. (Actually it had never occurred
4553 to me that the half of 'md_local' used for chaining was the
4554 half from which PRNG output bytes were taken -- I had always
4555 assumed that the secret half would be used.) The second
4556 measure makes sure that additional data from 'state' is never
4557 mixed into 'md_local' in small portions; this heuristically
4558 further strengthens the PRNG.
4561 *) Fix crypto/bn/asm/mips3.s.
4564 *) When only the key is given to "enc", the IV is undefined. Print out
4565 an error message in this case.
4568 *) Handle special case when X509_NAME is empty in X509 printing routines.
4571 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4572 positive and less than q.
4575 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4576 used: it isn't thread safe and the add_lock_callback should handle
4578 [Paul Rose <Paul.Rose@bridge.com>]
4580 *) Verify that incoming data obeys the block size in
4581 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4585 [Ulf Möller, Bodo Möller]
4587 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4588 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4589 when fixing the server behaviour for backwards-compatible 'client
4590 hello' messages. (Note that the attack is impractical against
4591 SSL 3.0 and TLS 1.0 anyway because length and version checking
4592 means that the probability of guessing a valid ciphertext is
4593 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4596 Before 0.9.5, the countermeasure (hide the error by generating a
4597 random 'decryption result') did not work properly because
4598 ERR_clear_error() was missing, meaning that SSL_get_error() would
4599 detect the supposedly ignored error.
4601 Both problems are now fixed.
4604 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4605 (previously it was 1024).
4608 *) Fix for compatibility mode trust settings: ignore trust settings
4609 unless some valid trust or reject settings are present.
4612 *) Fix for blowfish EVP: its a variable length cipher.
4615 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4616 parameters in DSA public key structures and return an error in the
4617 DSA routines if parameters are absent.
4620 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4621 in the current directory if neither $RANDFILE nor $HOME was set.
4622 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4623 caused some confusion to Windows users who haven't defined $HOME.
4624 Thus RAND_file_name() is changed again: e_os.h can define a
4625 DEFAULT_HOME, which will be used if $HOME is not set.
4626 For Windows, we use "C:"; on other platforms, we still require
4627 environment variables.
4629 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4630 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4631 having multiple threads call RAND_poll() concurrently.
4634 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4635 combination of a flag and a thread ID variable.
4636 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4637 flag), *other* threads can enter ssleay_add_bytes without obeying
4638 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4639 that they do not hold after the first thread unsets add_do_not_lock).
4642 *) Change bctest again: '-x' expressions are not available in all
4646 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4648 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4649 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4651 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4652 the default extension for executables, if any. Also, make the perl
4653 scripts that use symlink() to test if it really exists and use "cp"
4654 if it doesn't. All this made OpenSSL compilable and installable in
4658 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4659 If SEQUENCE is length is indefinite just set c->slen to the total
4660 amount of data available.
4661 [Steve Henson, reported by shige@FreeBSD.org]
4662 [This change does not apply to 0.9.7.]
4664 *) Change bctest to avoid here-documents inside command substitution
4665 (workaround for FreeBSD /bin/sh bug).
4666 For compatibility with Ultrix, avoid shell functions (introduced
4667 in the bctest version that searches along $PATH).
4670 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4671 with des_encrypt() defined on some operating systems, like Solaris
4675 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4676 On the Importance of Eliminating Errors in Cryptographic
4677 Computations, J. Cryptology 14 (2001) 2, 101-119,
4678 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4681 *) MIPS assembler BIGNUM division bug fix.
4684 *) Disabled incorrect Alpha assembler code.
4687 *) Fix PKCS#7 decode routines so they correctly update the length
4688 after reading an EOC for the EXPLICIT tag.
4690 [This change does not apply to 0.9.7.]
4692 *) Fix bug in PKCS#12 key generation routines. This was triggered
4693 if a 3DES key was generated with a 0 initial byte. Include
4694 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4695 (but broken) behaviour.
4698 *) Enhance bctest to search for a working bc along $PATH and print
4700 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4702 *) Fix memory leaks in err.c: free err_data string if necessary;
4703 don't write to the wrong index in ERR_set_error_data.
4706 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4710 *) Replace rdtsc with _emit statements for VC++ version 5.
4711 [Jeremy Cooper <jeremy@baymoo.org>]
4713 *) Make it possible to reuse SSLv2 sessions.
4716 *) In copy_email() check for >= 0 as a return value for
4717 X509_NAME_get_index_by_NID() since 0 is a valid index.
4718 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4720 *) Avoid coredump with unsupported or invalid public keys by checking if
4721 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4722 PKCS7_verify() fails with non detached data.
4725 *) Don't use getenv in library functions when run as setuid/setgid.
4726 New function OPENSSL_issetugid().
4729 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4730 due to incorrect handling of multi-threading:
4732 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4734 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4736 3. Count how many times MemCheck_off() has been called so that
4737 nested use can be treated correctly. This also avoids
4738 inband-signalling in the previous code (which relied on the
4739 assumption that thread ID 0 is impossible).
4742 *) Add "-rand" option also to s_client and s_server.
4745 *) Fix CPU detection on Irix 6.x.
4746 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4747 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4749 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4752 [This change does not apply to 0.9.7.]
4754 *) Use the cached encoding of an X509_NAME structure rather than
4755 copying it. This is apparently the reason for the libsafe "errors"
4756 but the code is actually correct.
4759 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4760 Bleichenbacher's DSA attack.
4761 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4762 to be set and top=0 forces the highest bit to be set; top=-1 is new
4763 and leaves the highest bit random.
4764 [Ulf Moeller, Bodo Moeller]
4766 *) In the NCONF_...-based implementations for CONF_... queries
4767 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4768 a temporary CONF structure with the data component set to NULL
4769 (which gives segmentation faults in lh_retrieve).
4770 Instead, use NULL for the CONF pointer in CONF_get_string and
4771 CONF_get_number (which may use environment variables) and directly
4772 return NULL from CONF_get_section.
4775 *) Fix potential buffer overrun for EBCDIC.
4778 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4779 keyUsage if basicConstraints absent for a CA.
4782 *) Make SMIME_write_PKCS7() write mail header values with a format that
4783 is more generally accepted (no spaces before the semicolon), since
4784 some programs can't parse those values properly otherwise. Also make
4785 sure BIO's that break lines after each write do not create invalid
4789 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4790 macros previously used would not encode an empty SEQUENCE OF
4791 and break the signature.
4793 [This change does not apply to 0.9.7.]
4795 *) Zero the premaster secret after deriving the master secret in
4799 *) Add some EVP_add_digest_alias registrations (as found in
4800 OpenSSL_add_all_digests()) to SSL_library_init()
4801 aka OpenSSL_add_ssl_algorithms(). This provides improved
4802 compatibility with peers using X.509 certificates
4803 with unconventional AlgorithmIdentifier OIDs.
4806 *) Fix for Irix with NO_ASM.
4807 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4809 *) ./config script fixes.
4810 [Ulf Moeller, Richard Levitte]
4812 *) Fix 'openssl passwd -1'.
4815 *) Change PKCS12_key_gen_asc() so it can cope with non null
4816 terminated strings whose length is passed in the passlen
4817 parameter, for example from PEM callbacks. This was done
4818 by adding an extra length parameter to asc2uni().
4819 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4821 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4822 call failed, free the DSA structure.
4825 *) Fix to uni2asc() to cope with zero length Unicode strings.
4826 These are present in some PKCS#12 files.
4829 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4830 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4831 when writing a 32767 byte record.
4832 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4834 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4835 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4837 (RSA objects have a reference count access to which is protected
4838 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4839 so they are meant to be shared between threads.)
4840 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4841 "Reddie, Steven" <Steven.Reddie@ca.com>]
4843 *) Fix a deadlock in CRYPTO_mem_leaks().
4846 *) Use better test patterns in bntest.
4849 *) rand_win.c fix for Borland C.
4852 *) BN_rshift bugfix for n == 0.
4855 *) Add a 'bctest' script that checks for some known 'bc' bugs
4856 so that 'make test' does not abort just because 'bc' is broken.
4859 *) Store verify_result within SSL_SESSION also for client side to
4860 avoid potential security hole. (Re-used sessions on the client side
4861 always resulted in verify_result==X509_V_OK, not using the original
4862 result of the server certificate verification.)
4865 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4866 SSL3_RT_APPLICATION_DATA, return 0.
4867 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4871 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4872 releases, have been re-implemented by renaming the previous
4873 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4874 and ssl3_read_internal, respectively, and adding 'peek' parameters
4875 to them. The new ssl[23]_{read,peek} functions are calls to
4876 ssl[23]_read_internal with the 'peek' flag set appropriately.
4877 A 'peek' parameter has also been added to ssl3_read_bytes, which
4878 does the actual work for ssl3_read_internal.
4881 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4882 the method-specific "init()" handler. Also clean up ex_data after
4883 calling the method-specific "finish()" handler. Previously, this was
4884 happening the other way round.
4887 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4888 The previous value, 12, was not always sufficient for BN_mod_exp().
4891 *) Make sure that shared libraries get the internal name engine with
4892 the full version number and not just 0. This should mark the
4893 shared libraries as not backward compatible. Of course, this should
4894 be changed again when we can guarantee backward binary compatibility.
4897 *) Fix typo in get_cert_by_subject() in by_dir.c
4898 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4900 *) Rework the system to generate shared libraries:
4902 - Make note of the expected extension for the shared libraries and
4903 if there is a need for symbolic links from for example libcrypto.so.0
4904 to libcrypto.so.0.9.7. There is extended info in Configure for
4907 - Make as few rebuilds of the shared libraries as possible.
4909 - Still avoid linking the OpenSSL programs with the shared libraries.
4911 - When installing, install the shared libraries separately from the
4915 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4917 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4918 and not in SSL_clear because the latter is also used by the
4919 accept/connect functions; previously, the settings made by
4920 SSL_set_read_ahead would be lost during the handshake.
4921 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4923 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4924 Previously, it would create entries for disableed algorithms no
4928 *) Added several new manual pages for SSL_* function.
4931 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4933 *) In ssl23_get_client_hello, generate an error message when faced
4934 with an initial SSL 3.0/TLS record that is too small to contain the
4935 first two bytes of the ClientHello message, i.e. client_version.
4936 (Note that this is a pathologic case that probably has never happened
4937 in real life.) The previous approach was to use the version number
4938 from the record header as a substitute; but our protocol choice
4939 should not depend on that one because it is not authenticated
4940 by the Finished messages.
4943 *) More robust randomness gathering functions for Windows.
4944 [Jeffrey Altman <jaltman@columbia.edu>]
4946 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4947 not set then we don't setup the error code for issuer check errors
4948 to avoid possibly overwriting other errors which the callback does
4949 handle. If an application does set the flag then we assume it knows
4950 what it is doing and can handle the new informational codes
4954 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4955 a general "ANY" type, as such it should be able to decode anything
4956 including tagged types. However it didn't check the class so it would
4957 wrongly interpret tagged types in the same way as their universal
4958 counterpart and unknown types were just rejected. Changed so that the
4959 tagged and unknown types are handled in the same way as a SEQUENCE:
4960 that is the encoding is stored intact. There is also a new type
4961 "V_ASN1_OTHER" which is used when the class is not universal, in this
4962 case we have no idea what the actual type is so we just lump them all
4966 *) On VMS, stdout may very well lead to a file that is written to
4967 in a record-oriented fashion. That means that every write() will
4968 write a separate record, which will be read separately by the
4969 programs trying to read from it. This can be very confusing.
4971 The solution is to put a BIO filter in the way that will buffer
4972 text until a linefeed is reached, and then write everything a
4973 line at a time, so every record written will be an actual line,
4974 not chunks of lines and not (usually doesn't happen, but I've
4975 seen it once) several lines in one record. BIO_f_linebuffer() is
4978 Currently, it's a VMS-only method, because that's where it has
4979 been tested well enough.
4982 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4983 it can return incorrect results.
4984 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4985 but it was in 0.9.6-beta[12].)
4988 *) Disable the check for content being present when verifying detached
4989 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4990 include zero length content when signing messages.
4993 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4994 BIO_ctrl (for BIO pairs).
4997 *) Add DSO method for VMS.
5000 *) Bug fix: Montgomery multiplication could produce results with the
5004 *) Add RPM specification openssl.spec and modify it to build three
5005 packages. The default package contains applications, application
5006 documentation and run-time libraries. The devel package contains
5007 include files, static libraries and function documentation. The
5008 doc package contains the contents of the doc directory. The original
5009 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5012 *) Add a large number of documentation files for many SSL routines.
5013 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5015 *) Add a configuration entry for Sony News 4.
5016 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5018 *) Don't set the two most significant bits to one when generating a
5019 random number < q in the DSA library.
5022 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5023 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5024 the underlying transport is blocking) if a handshake took place.
5025 (The default behaviour is needed by applications such as s_client
5026 and s_server that use select() to determine when to use SSL_read;
5027 but for applications that know in advance when to expect data, it
5028 just makes things more complicated.)
5031 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5035 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5036 work better on such systems.
5037 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5039 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5040 Update PKCS12_parse() so it copies the friendlyName and the
5041 keyid to the certificates aux info.
5044 *) Fix bug in PKCS7_verify() which caused an infinite loop
5045 if there was more than one signature.
5046 [Sven Uszpelkat <su@celocom.de>]
5048 *) Major change in util/mkdef.pl to include extra information
5049 about each symbol, as well as presentig variables as well
5050 as functions. This change means that there's n more need
5051 to rebuild the .num files when some algorithms are excluded.
5054 *) Allow the verify time to be set by an application,
5055 rather than always using the current time.
5058 *) Phase 2 verify code reorganisation. The certificate
5059 verify code now looks up an issuer certificate by a
5060 number of criteria: subject name, authority key id
5061 and key usage. It also verifies self signed certificates
5062 by the same criteria. The main comparison function is
5063 X509_check_issued() which performs these checks.
5065 Lot of changes were necessary in order to support this
5066 without completely rewriting the lookup code.
5068 Authority and subject key identifier are now cached.
5070 The LHASH 'certs' is X509_STORE has now been replaced
5071 by a STACK_OF(X509_OBJECT). This is mainly because an
5072 LHASH can't store or retrieve multiple objects with
5073 the same hash value.
5075 As a result various functions (which were all internal
5076 use only) have changed to handle the new X509_STORE
5077 structure. This will break anything that messed round
5078 with X509_STORE internally.
5080 The functions X509_STORE_add_cert() now checks for an
5081 exact match, rather than just subject name.
5083 The X509_STORE API doesn't directly support the retrieval
5084 of multiple certificates matching a given criteria, however
5085 this can be worked round by performing a lookup first
5086 (which will fill the cache with candidate certificates)
5087 and then examining the cache for matches. This is probably
5088 the best we can do without throwing out X509_LOOKUP
5089 entirely (maybe later...).
5091 The X509_VERIFY_CTX structure has been enhanced considerably.
5093 All certificate lookup operations now go via a get_issuer()
5094 callback. Although this currently uses an X509_STORE it
5095 can be replaced by custom lookups. This is a simple way
5096 to bypass the X509_STORE hackery necessary to make this
5097 work and makes it possible to use more efficient techniques
5098 in future. A very simple version which uses a simple
5099 STACK for its trusted certificate store is also provided
5100 using X509_STORE_CTX_trusted_stack().
5102 The verify_cb() and verify() callbacks now have equivalents
5103 in the X509_STORE_CTX structure.
5105 X509_STORE_CTX also has a 'flags' field which can be used
5106 to customise the verify behaviour.
5109 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5110 excludes S/MIME capabilities.
5113 *) When a certificate request is read in keep a copy of the
5114 original encoding of the signed data and use it when outputing
5115 again. Signatures then use the original encoding rather than
5116 a decoded, encoded version which may cause problems if the
5117 request is improperly encoded.
5120 *) For consistency with other BIO_puts implementations, call
5121 buffer_write(b, ...) directly in buffer_puts instead of calling
5124 In BIO_puts, increment b->num_write as in BIO_write.
5125 [Peter.Sylvester@EdelWeb.fr]
5127 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5128 BN_zero, we may not return a BIGNUM with an array consisting of
5132 *) Avoid calling abort() from within the library when problems are
5133 detected, except if preprocessor symbols have been defined
5134 (such as REF_CHECK, BN_DEBUG etc.).
5137 *) New openssl application 'rsautl'. This utility can be
5138 used for low level RSA operations. DER public key
5139 BIO/fp routines also added.
5142 *) New Configure entry and patches for compiling on QNX 4.
5143 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5145 *) A demo state-machine implementation was sponsored by
5146 Nuron (http://www.nuron.com/) and is now available in
5147 demos/state_machine.
5150 *) New options added to the 'dgst' utility for signature
5151 generation and verification.
5154 *) Unrecognized PKCS#7 content types are now handled via a
5155 catch all ASN1_TYPE structure. This allows unsupported
5156 types to be stored as a "blob" and an application can
5157 encode and decode it manually.
5160 *) Fix various signed/unsigned issues to make a_strex.c
5162 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5164 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5165 length if passed a buffer. ASN1_INTEGER_to_BN failed
5166 if passed a NULL BN and its argument was negative.
5167 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5169 *) Modification to PKCS#7 encoding routines to output definite
5170 length encoding. Since currently the whole structures are in
5171 memory there's not real point in using indefinite length
5172 constructed encoding. However if OpenSSL is compiled with
5173 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5176 *) Added BIO_vprintf() and BIO_vsnprintf().
5179 *) Added more prefixes to parse for in the the strings written
5180 through a logging bio, to cover all the levels that are available
5181 through syslog. The prefixes are now:
5183 PANIC, EMERG, EMR => LOG_EMERG
5184 ALERT, ALR => LOG_ALERT
5185 CRIT, CRI => LOG_CRIT
5186 ERROR, ERR => LOG_ERR
5187 WARNING, WARN, WAR => LOG_WARNING
5188 NOTICE, NOTE, NOT => LOG_NOTICE
5189 INFO, INF => LOG_INFO
5190 DEBUG, DBG => LOG_DEBUG
5192 and as before, if none of those prefixes are present at the
5193 beginning of the string, LOG_ERR is chosen.
5195 On Win32, the LOG_* levels are mapped according to this:
5197 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5198 LOG_WARNING => EVENTLOG_WARNING_TYPE
5199 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5203 *) Made it possible to reconfigure with just the configuration
5204 argument "reconf" or "reconfigure". The command line arguments
5205 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5206 and are retrieved from there when reconfiguring.
5210 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5212 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5215 *) The obj_dat.pl script was messing up the sorting of object
5216 names. The reason was that it compared the quoted version
5217 of strings as a result "OCSP" > "OCSP Signing" because
5218 " > SPACE. Changed script to store unquoted versions of
5219 names and add quotes on output. It was also omitting some
5220 names from the lookup table if they were given a default
5221 value (that is if SN is missing it is given the same
5222 value as LN and vice versa), these are now added on the
5223 grounds that if an object has a name we should be able to
5224 look it up. Finally added warning output when duplicate
5225 short or long names are found.
5228 *) Changes needed for Tandem NSK.
5229 [Scott Uroff <scott@xypro.com>]
5231 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5232 RSA_padding_check_SSLv23(), special padding was never detected
5233 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5234 version rollback attacks was not effective.
5236 In s23_clnt.c, don't use special rollback-attack detection padding
5237 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5238 client; similarly, in s23_srvr.c, don't do the rollback check if
5239 SSL 2.0 is the only protocol enabled in the server.
5242 *) Make it possible to get hexdumps of unprintable data with 'openssl
5243 asn1parse'. By implication, the functions ASN1_parse_dump() and
5244 BIO_dump_indent() are added.
5247 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5248 these print out strings and name structures based on various
5249 flags including RFC2253 support and proper handling of
5250 multibyte characters. Added options to the 'x509' utility
5251 to allow the various flags to be set.
5254 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5255 Also change the functions X509_cmp_current_time() and
5256 X509_gmtime_adj() work with an ASN1_TIME structure,
5257 this will enable certificates using GeneralizedTime in validity
5258 dates to be checked.
5261 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5262 negative public key encodings) on by default,
5263 NO_NEG_PUBKEY_BUG can be set to disable it.
5266 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5267 content octets. An i2c_ASN1_OBJECT is unnecessary because
5268 the encoding can be trivially obtained from the structure.
5271 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5272 not read locks (CRYPTO_r_[un]lock).
5275 *) A first attempt at creating official support for shared
5276 libraries through configuration. I've kept it so the
5277 default is static libraries only, and the OpenSSL programs
5278 are always statically linked for now, but there are
5279 preparations for dynamic linking in place.
5280 This has been tested on Linux and Tru64.
5283 *) Randomness polling function for Win9x, as described in:
5284 Peter Gutmann, Software Generation of Practically Strong
5288 *) Fix so PRNG is seeded in req if using an already existing
5292 *) New options to smime application. -inform and -outform
5293 allow alternative formats for the S/MIME message including
5294 PEM and DER. The -content option allows the content to be
5295 specified separately. This should allow things like Netscape
5296 form signing output easier to verify.
5299 *) Fix the ASN1 encoding of tags using the 'long form'.
5302 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5303 STRING types. These convert content octets to and from the
5304 underlying type. The actual tag and length octets are
5305 already assumed to have been read in and checked. These
5306 are needed because all other string types have virtually
5307 identical handling apart from the tag. By having versions
5308 of the ASN1 functions that just operate on content octets
5309 IMPLICIT tagging can be handled properly. It also allows
5310 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5311 and ASN1_INTEGER are identical apart from the tag.
5314 *) Change the handling of OID objects as follows:
5316 - New object identifiers are inserted in objects.txt, following
5317 the syntax given in objects.README.
5318 - objects.pl is used to process obj_mac.num and create a new
5320 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5323 This is currently kind of a hack, and the perl code in objects.pl
5324 isn't very elegant, but it works as I intended. The simplest way
5325 to check that it worked correctly is to look in obj_dat.h and
5326 check the array nid_objs and make sure the objects haven't moved
5327 around (this is important!). Additions are OK, as well as
5328 consistent name changes.
5331 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5334 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5335 The given file adds to whatever has already been seeded into the
5336 random pool through the RANDFILE configuration file option or
5337 environment variable, or the default random state file.
5340 *) mkstack.pl now sorts each macro group into lexical order.
5341 Previously the output order depended on the order the files
5342 appeared in the directory, resulting in needless rewriting
5346 *) Patches to make OpenSSL compile under Win32 again. Mostly
5347 work arounds for the VC++ problem that it treats func() as
5348 func(void). Also stripped out the parts of mkdef.pl that
5349 added extra typesafe functions: these no longer exist.
5352 *) Reorganisation of the stack code. The macros are now all
5353 collected in safestack.h . Each macro is defined in terms of
5354 a "stack macro" of the form SKM_<name>(type, a, b). The
5355 DEBUG_SAFESTACK is now handled in terms of function casts,
5356 this has the advantage of retaining type safety without the
5357 use of additional functions. If DEBUG_SAFESTACK is not defined
5358 then the non typesafe macros are used instead. Also modified the
5359 mkstack.pl script to handle the new form. Needs testing to see
5360 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5361 the default if no major problems. Similar behaviour for ASN1_SET_OF
5362 and PKCS12_STACK_OF.
5365 *) When some versions of IIS use the 'NET' form of private key the
5366 key derivation algorithm is different. Normally MD5(password) is
5367 used as a 128 bit RC4 key. In the modified case
5368 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5369 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5370 as the old Netscape_RSA functions except they have an additional
5371 'sgckey' parameter which uses the modified algorithm. Also added
5372 an -sgckey command line option to the rsa utility. Thanks to
5373 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5374 algorithm to openssl-dev.
5377 *) The evp_local.h macros were using 'c.##kname' which resulted in
5378 invalid expansion on some systems (SCO 5.0.5 for example).
5379 Corrected to 'c.kname'.
5380 [Phillip Porch <root@theporch.com>]
5382 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5383 a STACK of email addresses from a certificate or request, these look
5384 in the subject name and the subject alternative name extensions and
5385 omit any duplicate addresses.
5388 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5389 This makes DSA verification about 2 % faster.
5392 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5393 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5394 plus overhead for 1024 bit moduli).
5395 This makes exponentiations about 0.5 % faster for 1024 bit
5396 exponents (as measured by "openssl speed rsa2048").
5399 *) Rename memory handling macros to avoid conflicts with other
5401 Malloc => OPENSSL_malloc
5402 Malloc_locked => OPENSSL_malloc_locked
5403 Realloc => OPENSSL_realloc
5404 Free => OPENSSL_free
5407 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5408 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5411 *) CygWin32 support.
5412 [John Jarvie <jjarvie@newsguy.com>]
5414 *) The type-safe stack code has been rejigged. It is now only compiled
5415 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5416 by default all type-specific stack functions are "#define"d back to
5417 standard stack functions. This results in more streamlined output
5418 but retains the type-safety checking possibilities of the original
5422 *) The STACK code has been cleaned up, and certain type declarations
5423 that didn't make a lot of sense have been brought in line. This has
5424 also involved a cleanup of sorts in safestack.h to more correctly
5425 map type-safe stack functions onto their plain stack counterparts.
5426 This work has also resulted in a variety of "const"ifications of
5427 lots of the code, especially "_cmp" operations which should normally
5428 be prototyped with "const" parameters anyway.
5431 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5432 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5433 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5434 where all of 'md' is used each time the PRNG is used, but 'state'
5435 is used only indexed by a cyclic counter. As entropy may not be
5436 well distributed from the beginning, 'md' is important as a
5437 chaining variable. However, the output function chains only half
5438 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5439 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5440 in all of 'state' being rewritten, with the new values depending
5441 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5444 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5445 the handshake is continued after ssl_verify_cert_chain();
5446 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5447 can lead to 'unexplainable' connection aborts later.
5448 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5450 *) Major EVP API cipher revision.
5451 Add hooks for extra EVP features. This allows various cipher
5452 parameters to be set in the EVP interface. Support added for variable
5453 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5454 setting of RC2 and RC5 parameters.
5456 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5459 Remove lots of duplicated code from the EVP library. For example *every*
5460 cipher init() function handles the 'iv' in the same way according to the
5461 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5462 for CFB and OFB modes they zero ctx->num.
5464 New functionality allows removal of S/MIME code RC2 hack.
5466 Most of the routines have the same form and so can be declared in terms
5469 By shifting this to the top level EVP_CipherInit() it can be removed from
5470 all individual ciphers. If the cipher wants to handle IVs or keys
5471 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5474 Change lots of functions like EVP_EncryptUpdate() to now return a
5475 value: although software versions of the algorithms cannot fail
5476 any installed hardware versions can.
5479 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5480 this option is set, tolerate broken clients that send the negotiated
5481 protocol version number instead of the requested protocol version
5485 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5486 i.e. non-zero for export ciphersuites, zero otherwise.
5487 Previous versions had this flag inverted, inconsistent with
5488 rsa_tmp_cb (..._TMP_RSA_CB).
5489 [Bodo Moeller; problem reported by Amit Chopra]
5491 *) Add missing DSA library text string. Work around for some IIS
5492 key files with invalid SEQUENCE encoding.
5495 *) Add a document (doc/standards.txt) that list all kinds of standards
5496 and so on that are implemented in OpenSSL.
5499 *) Enhance c_rehash script. Old version would mishandle certificates
5500 with the same subject name hash and wouldn't handle CRLs at all.
5501 Added -fingerprint option to crl utility, to support new c_rehash
5505 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5508 *) Fix for SSL server purpose checking. Server checking was
5509 rejecting certificates which had extended key usage present
5510 but no ssl client purpose.
5511 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5513 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5514 is a little unclear about how a blank password is handled.
5515 Since the password in encoded as a BMPString with terminating
5516 double NULL a zero length password would end up as just the
5517 double NULL. However no password at all is different and is
5518 handled differently in the PKCS#12 key generation code. NS
5519 treats a blank password as zero length. MSIE treats it as no
5520 password on export: but it will try both on import. We now do
5521 the same: PKCS12_parse() tries zero length and no password if
5522 the password is set to "" or NULL (NULL is now a valid password:
5523 it wasn't before) as does the pkcs12 application.
5526 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5527 perror when PEM_read_bio_X509_REQ fails, the error message must
5528 be obtained from the error queue.
5531 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5532 it in ERR_remove_state if appropriate, and change ERR_get_state
5533 accordingly to avoid race conditions (this is necessary because
5534 thread_hash is no longer constant once set).
5537 *) Bugfix for linux-elf makefile.one.
5540 *) RSA_get_default_method() will now cause a default
5541 RSA_METHOD to be chosen if one doesn't exist already.
5542 Previously this was only set during a call to RSA_new()
5543 or RSA_new_method(NULL) meaning it was possible for
5544 RSA_get_default_method() to return NULL.
5547 *) Added native name translation to the existing DSO code
5548 that will convert (if the flag to do so is set) filenames
5549 that are sufficiently small and have no path information
5550 into a canonical native form. Eg. "blah" converted to
5551 "libblah.so" or "blah.dll" etc.
5554 *) New function ERR_error_string_n(e, buf, len) which is like
5555 ERR_error_string(e, buf), but writes at most 'len' bytes
5556 including the 0 terminator. For ERR_error_string_n, 'buf'
5558 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5560 *) CONF library reworked to become more general. A new CONF
5561 configuration file reader "class" is implemented as well as a
5562 new functions (NCONF_*, for "New CONF") to handle it. The now
5563 old CONF_* functions are still there, but are reimplemented to
5564 work in terms of the new functions. Also, a set of functions
5565 to handle the internal storage of the configuration data is
5566 provided to make it easier to write new configuration file
5567 reader "classes" (I can definitely see something reading a
5568 configuration file in XML format, for example), called _CONF_*,
5569 or "the configuration storage API"...
5571 The new configuration file reading functions are:
5573 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5574 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5576 NCONF_default, NCONF_WIN32
5578 NCONF_dump_fp, NCONF_dump_bio
5580 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5581 NCONF_new creates a new CONF object. This works in the same way
5582 as other interfaces in OpenSSL, like the BIO interface.
5583 NCONF_dump_* dump the internal storage of the configuration file,
5584 which is useful for debugging. All other functions take the same
5585 arguments as the old CONF_* functions wth the exception of the
5586 first that must be a `CONF *' instead of a `LHASH *'.
5588 To make it easer to use the new classes with the old CONF_* functions,
5589 the function CONF_set_default_method is provided.
5592 *) Add '-tls1' option to 'openssl ciphers', which was already
5593 mentioned in the documentation but had not been implemented.
5594 (This option is not yet really useful because even the additional
5595 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5598 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5599 OpenSSL-based applications) load shared libraries and bind to
5600 them in a portable way.
5601 [Geoff Thorpe, with contributions from Richard Levitte]
5603 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5605 *) Make sure _lrotl and _lrotr are only used with MSVC.
5607 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5608 (the default implementation of RAND_status).
5610 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5611 to '-clrext' (= clear extensions), as intended and documented.
5612 [Bodo Moeller; inconsistency pointed out by Michael Attili
5615 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5616 was larger than the MD block size.
5617 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5619 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5620 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5621 using the passed key: if the passed key was a private key the result
5622 of X509_print(), for example, would be to print out all the private key
5626 *) des_quad_cksum() byte order bug fix.
5627 [Ulf Möller, using the problem description in krb4-0.9.7, where
5628 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5630 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5632 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5634 *) For easily testing in shell scripts whether some command
5635 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5636 returns with exit code 0 iff no command of the given name is available.
5637 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5638 the output goes to stdout and nothing is printed to stderr.
5639 Additional arguments are always ignored.
5641 Since for each cipher there is a command of the same name,
5642 the 'no-cipher' compilation switches can be tested this way.
5644 ('openssl no-XXX' is not able to detect pseudo-commands such
5645 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5648 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5651 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5652 is set; it will be thrown away anyway because each handshake creates
5654 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5655 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5656 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5657 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5660 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5661 'Q' and 'R' lose their special meanings (quit/renegotiate).
5662 This is part of what -quiet does; unlike -quiet, -ign_eof
5663 does not suppress any output.
5666 *) Add compatibility options to the purpose and trust code. The
5667 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5668 accepts a certificate or CA, this was the previous behaviour,
5669 with all the associated security issues.
5671 X509_TRUST_COMPAT is the old trust behaviour: only and
5672 automatically trust self signed roots in certificate store. A
5673 new trust setting X509_TRUST_DEFAULT is used to specify that
5674 a purpose has no associated trust setting and it should instead
5675 use the value in the default purpose.
5678 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5679 and fix a memory leak.
5682 *) In util/mkerr.pl (which implements 'make errors'), preserve
5683 reason strings from the previous version of the .c file, as
5684 the default to have only downcase letters (and digits) in
5685 automatically generated reasons codes is not always appropriate.
5688 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5689 using strerror. Previously, ERR_reason_error_string() returned
5690 library names as reason strings for SYSerr; but SYSerr is a special
5691 case where small numbers are errno values, not library numbers.
5694 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5695 converts DSA parameters into DH parameters. (When creating parameters,
5696 DSA_generate_parameters is used.)
5699 *) Include 'length' (recommended exponent length) in C code generated
5700 by 'openssl dhparam -C'.
5703 *) The second argument to set_label in perlasm was already being used
5704 so couldn't be used as a "file scope" flag. Moved to third argument
5708 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5709 instead of RAND_bytes for encryption IVs and salts.
5712 *) Include RAND_status() into RAND_METHOD instead of implementing
5713 it only for md_rand.c Otherwise replacing the PRNG by calling
5714 RAND_set_rand_method would be impossible.
5717 *) Don't let DSA_generate_key() enter an infinite loop if the random
5718 number generation fails.
5721 *) New 'rand' application for creating pseudo-random output.
5724 *) Added configuration support for Linux/IA64
5725 [Rolf Haberrecker <rolf@suse.de>]
5727 *) Assembler module support for Mingw32.
5730 *) Shared library support for HPUX (in shlib/).
5731 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5733 *) Shared library support for Solaris gcc.
5734 [Lutz Behnke <behnke@trustcenter.de>]
5736 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5738 *) PKCS7_encrypt() was adding text MIME headers twice because they
5739 were added manually and by SMIME_crlf_copy().
5742 *) In bntest.c don't call BN_rand with zero bits argument.
5743 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5745 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5746 case was implemented. This caused BN_div_recp() to fail occasionally.
5749 *) Add an optional second argument to the set_label() in the perl
5750 assembly language builder. If this argument exists and is set
5751 to 1 it signals that the assembler should use a symbol whose
5752 scope is the entire file, not just the current function. This
5753 is needed with MASM which uses the format label:: for this scope.
5754 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5756 *) Change the ASN1 types so they are typedefs by default. Before
5757 almost all types were #define'd to ASN1_STRING which was causing
5758 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5762 *) Change names of new functions to the new get1/get0 naming
5763 convention: After 'get1', the caller owns a reference count
5764 and has to call ..._free; 'get0' returns a pointer to some
5765 data structure without incrementing reference counters.
5766 (Some of the existing 'get' functions increment a reference
5767 counter, some don't.)
5768 Similarly, 'set1' and 'add1' functions increase reference
5769 counters or duplicate objects.
5772 *) Allow for the possibility of temp RSA key generation failure:
5773 the code used to assume it always worked and crashed on failure.
5776 *) Fix potential buffer overrun problem in BIO_printf().
5777 [Ulf Möller, using public domain code by Patrick Powell; problem
5778 pointed out by David Sacerdote <das33@cornell.edu>]
5780 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5781 RAND_egd() and RAND_status(). In the command line application,
5782 the EGD socket can be specified like a seed file using RANDFILE
5786 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5787 Some CAs (e.g. Verisign) distribute certificates in this form.
5790 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5791 list to exclude them. This means that no special compilation option
5792 is needed to use anonymous DH: it just needs to be included in the
5796 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5797 EVP_MD_type. The old functionality is available in a new macro called
5798 EVP_MD_md(). Change code that uses it and update docs.
5801 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5802 where the 'void *' argument is replaced by a function pointer argument.
5803 Previously 'void *' was abused to point to functions, which works on
5804 many platforms, but is not correct. As these functions are usually
5805 called by macros defined in OpenSSL header files, most source code
5806 should work without changes.
5809 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5810 sections with information on -D... compiler switches used for
5811 compiling the library so that applications can see them. To enable
5812 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5813 must be defined. E.g.,
5814 #define OPENSSL_ALGORITHM_DEFINES
5815 #include <openssl/opensslconf.h>
5816 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5817 [Richard Levitte, Ulf and Bodo Möller]
5819 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5823 *) Change the 'other' type in certificate aux info to a STACK_OF
5824 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5825 the required ASN1 format: arbitrary types determined by an OID.
5828 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5829 argument to 'req'. This is not because the function is newer or
5830 better than others it just uses the work 'NEW' in the certificate
5831 request header lines. Some software needs this.
5834 *) Reorganise password command line arguments: now passwords can be
5835 obtained from various sources. Delete the PEM_cb function and make
5836 it the default behaviour: i.e. if the callback is NULL and the
5837 usrdata argument is not NULL interpret it as a null terminated pass
5838 phrase. If usrdata and the callback are NULL then the pass phrase
5839 is prompted for as usual.
5842 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5843 the support is automatically enabled. The resulting binaries will
5844 autodetect the card and use it if present.
5845 [Ben Laurie and Compaq Inc.]
5847 *) Work around for Netscape hang bug. This sends certificate request
5848 and server done in one record. Since this is perfectly legal in the
5849 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5850 the bugs/SSLv3 entry for more info.
5853 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5856 *) Add -rand argument to smime and pkcs12 applications and read/write
5860 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5863 *) Add command line password options to the remaining applications.
5866 *) Bug fix for BN_div_recp() for numerators with an even number of
5870 *) More tests in bntest.c, and changed test_bn output.
5873 *) ./config recognizes MacOS X now.
5876 *) Bug fix for BN_div() when the first words of num and divsor are
5877 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5880 *) Add support for various broken PKCS#8 formats, and command line
5881 options to produce them.
5884 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5885 get temporary BIGNUMs from a BN_CTX.
5888 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5892 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5893 include a #define from the old name to the new. The original intent
5894 was that statically linked binaries could for example just call
5895 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5896 link with digests. This never worked becayse SSLeay_add_all_digests()
5897 and SSLeay_add_all_ciphers() were in the same source file so calling
5898 one would link with the other. They are now in separate source files.
5901 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5904 *) Use a less unusual form of the Miller-Rabin primality test (it used
5905 a binary algorithm for exponentiation integrated into the Miller-Rabin
5906 loop, our standard modexp algorithms are faster).
5909 *) Support for the EBCDIC character set completed.
5910 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5912 *) Source code cleanups: use const where appropriate, eliminate casts,
5913 use void * instead of char * in lhash.
5916 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5917 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5918 this the server could overwrite ephemeral keys that the client
5922 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5923 using 50 iterations of the Rabin-Miller test.
5925 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5926 iterations of the Rabin-Miller test as required by the appendix
5927 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5928 As BN_is_prime_fasttest includes trial division, DSA parameter
5929 generation becomes much faster.
5931 This implies a change for the callback functions in DSA_is_prime
5932 and DSA_generate_parameters: The callback function is called once
5933 for each positive witness in the Rabin-Miller test, not just
5934 occasionally in the inner loop; and the parameters to the
5935 callback function now provide an iteration count for the outer
5936 loop rather than for the current invocation of the inner loop.
5937 DSA_generate_parameters additionally can call the callback
5938 function with an 'iteration count' of -1, meaning that a
5939 candidate has passed the trial division test (when q is generated
5940 from an application-provided seed, trial division is skipped).
5943 *) New function BN_is_prime_fasttest that optionally does trial
5944 division before starting the Rabin-Miller test and has
5945 an additional BN_CTX * argument (whereas BN_is_prime always
5946 has to allocate at least one BN_CTX).
5947 'callback(1, -1, cb_arg)' is called when a number has passed the
5948 trial division stage.
5951 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5955 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5958 *) New function BN_pseudo_rand().
5961 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5962 bignum version of BN_from_montgomery() with the working code from
5963 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5967 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5968 made it impossible to use the same SSL_SESSION data structure in
5969 SSL2 clients in multiple threads.
5972 *) The return value of RAND_load_file() no longer counts bytes obtained
5973 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5974 to seed the PRNG (previously an explicit byte count was required).
5975 [Ulf Möller, Bodo Möller]
5977 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5978 used (char *) instead of (void *) and had casts all over the place.
5981 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5984 *) Retain source code compatibility for BN_prime_checks macro:
5985 BN_is_prime(..., BN_prime_checks, ...) now uses
5986 BN_prime_checks_for_size to determine the appropriate number of
5987 Rabin-Miller iterations.
5990 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5991 DH_CHECK_P_NOT_SAFE_PRIME.
5992 (Check if this is true? OpenPGP calls them "strong".)
5995 *) Merge the functionality of "dh" and "gendh" programs into a new program
5996 "dhparam". The old programs are retained for now but will handle DH keys
5997 (instead of parameters) in future.
6000 *) Make the ciphers, s_server and s_client programs check the return values
6001 when a new cipher list is set.
6004 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6005 ciphers. Before when the 56bit ciphers were enabled the sorting was
6008 The syntax for the cipher sorting has been extended to support sorting by
6009 cipher-strength (using the strength_bits hard coded in the tables).
6010 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6012 Fix a bug in the cipher-command parser: when supplying a cipher command
6013 string with an "undefined" symbol (neither command nor alphanumeric
6014 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6015 an error is flagged.
6017 Due to the strength-sorting extension, the code of the
6018 ssl_create_cipher_list() function was completely rearranged. I hope that
6019 the readability was also increased :-)
6020 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6022 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6023 for the first serial number and places 2 in the serial number file. This
6024 avoids problems when the root CA is created with serial number zero and
6025 the first user certificate has the same issuer name and serial number
6029 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6030 the new code. Add documentation for this stuff.
6033 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6034 X509_*() to X509at_*() on the grounds that they don't handle X509
6035 structures and behave in an analagous way to the X509v3 functions:
6036 they shouldn't be called directly but wrapper functions should be used
6039 So we also now have some wrapper functions that call the X509at functions
6040 when passed certificate requests. (TO DO: similar things can be done with
6041 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6042 things. Some of these need some d2i or i2d and print functionality
6043 because they handle more complex structures.)
6046 *) Add missing #ifndefs that caused missing symbols when building libssl
6047 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6048 NO_RSA in ssl/s2*.c.
6049 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6051 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6052 has a return value which indicates the quality of the random data
6053 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6054 error queue. New function RAND_pseudo_bytes() generates output that is
6055 guaranteed to be unique but not unpredictable. RAND_add is like
6056 RAND_seed, but takes an extra argument for an entropy estimate
6057 (RAND_seed always assumes full entropy).
6060 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6061 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6062 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6063 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6064 false-positive rate of at most 2^-80 for random input.
6067 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6070 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6071 in the 0.9.5 release), this returns the chain
6072 from an X509_CTX structure with a dup of the stack and all
6073 the X509 reference counts upped: so the stack will exist
6074 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6077 Also make SSL_SESSION_print() print out the verify return
6081 *) Add manpage for the pkcs12 command. Also change the default
6082 behaviour so MAC iteration counts are used unless the new
6083 -nomaciter option is used. This improves file security and
6084 only older versions of MSIE (4.0 for example) need it.
6087 *) Honor the no-xxx Configure options when creating .DEF files.
6090 *) Add PKCS#10 attributes to field table: challengePassword,
6091 unstructuredName and unstructuredAddress. These are taken from
6092 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6093 international characters are used.
6095 More changes to X509_ATTRIBUTE code: allow the setting of types
6096 based on strings. Remove the 'loc' parameter when adding
6097 attributes because these will be a SET OF encoding which is sorted
6101 *) Initial changes to the 'req' utility to allow request generation
6102 automation. This will allow an application to just generate a template
6103 file containing all the field values and have req construct the
6106 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6107 used all over the place including certificate requests and PKCS#7
6108 structures. They are currently handled manually where necessary with
6109 some primitive wrappers for PKCS#7. The new functions behave in a
6110 manner analogous to the X509 extension functions: they allow
6111 attributes to be looked up by NID and added.
6113 Later something similar to the X509V3 code would be desirable to
6114 automatically handle the encoding, decoding and printing of the
6115 more complex types. The string types like challengePassword can
6116 be handled by the string table functions.
6118 Also modified the multi byte string table handling. Now there is
6119 a 'global mask' which masks out certain types. The table itself
6120 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6121 is useful when for example there is only one permissible type
6122 (as in countryName) and using the mask might result in no valid
6126 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6127 SSL_get_peer_finished to allow applications to obtain the latest
6128 Finished messages sent to the peer or expected from the peer,
6129 respectively. (SSL_get_peer_finished is usually the Finished message
6130 actually received from the peer, otherwise the protocol will be aborted.)
6132 As the Finished message are message digests of the complete handshake
6133 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6134 be used for external authentication procedures when the authentication
6135 provided by SSL/TLS is not desired or is not enough.
6138 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6139 the host supports BWX extension and if Compaq C is present on the
6140 $PATH. Just exploiting of the BWX extension results in 20-30%
6141 performance kick for some algorithms, e.g. DES and RC4 to mention
6142 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6146 *) Add support for MS "fast SGC". This is arguably a violation of the
6147 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6148 weak crypto and after checking the certificate is SGC a second one
6149 with strong crypto. MS SGC stops the first handshake after receiving
6150 the server certificate message and sends a second client hello. Since
6151 a server will typically do all the time consuming operations before
6152 expecting any further messages from the client (server key exchange
6153 is the most expensive) there is little difference between the two.
6155 To get OpenSSL to support MS SGC we have to permit a second client
6156 hello message after we have sent server done. In addition we have to
6157 reset the MAC if we do get this second client hello.
6160 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6161 if a DER encoded private key is RSA or DSA traditional format. Changed
6162 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6163 format DER encoded private key. Newer code should use PKCS#8 format which
6164 has the key type encoded in the ASN1 structure. Added DER private key
6165 support to pkcs8 application.
6168 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6169 ciphersuites has been selected (as required by the SSL 3/TLS 1
6170 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6171 is set, we interpret this as a request to violate the specification
6172 (the worst that can happen is a handshake failure, and 'correct'
6173 behaviour would result in a handshake failure anyway).
6176 *) In SSL_CTX_add_session, take into account that there might be multiple
6177 SSL_SESSION structures with the same session ID (e.g. when two threads
6178 concurrently obtain them from an external cache).
6179 The internal cache can handle only one SSL_SESSION with a given ID,
6180 so if there's a conflict, we now throw out the old one to achieve
6184 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6185 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6186 some routines that use cipher OIDs: some ciphers do not have OIDs
6187 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6191 *) Simplify the trust setting structure and code. Now we just have
6192 two sequences of OIDs for trusted and rejected settings. These will
6193 typically have values the same as the extended key usage extension
6194 and any application specific purposes.
6196 The trust checking code now has a default behaviour: it will just
6197 check for an object with the same NID as the passed id. Functions can
6198 be provided to override either the default behaviour or the behaviour
6199 for a given id. SSL client, server and email already have functions
6200 in place for compatibility: they check the NID and also return "trusted"
6201 if the certificate is self signed.
6204 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6205 traditional format into an EVP_PKEY structure.
6208 *) Add a password callback function PEM_cb() which either prompts for
6209 a password if usr_data is NULL or otherwise assumes it is a null
6210 terminated password. Allow passwords to be passed on command line
6211 environment or config files in a few more utilities.
6214 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6215 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6216 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6217 Update documentation.
6220 *) Support for ASN1 "NULL" type. This could be handled before by using
6221 ASN1_TYPE but there wasn't any function that would try to read a NULL
6222 and produce an error if it couldn't. For compatibility we also have
6223 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6224 don't allocate anything because they don't need to.
6227 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6229 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6231 *) Rebuild of the memory allocation routines used by OpenSSL code and
6232 possibly others as well. The purpose is to make an interface that
6233 provide hooks so anyone can build a separate set of allocation and
6234 deallocation routines to be used by OpenSSL, for example memory
6235 pool implementations, or something else, which was previously hard
6236 since Malloc(), Realloc() and Free() were defined as macros having
6237 the values malloc, realloc and free, respectively (except for Win32
6238 compilations). The same is provided for memory debugging code.
6239 OpenSSL already comes with functionality to find memory leaks, but
6240 this gives people a chance to debug other memory problems.
6242 With these changes, a new set of functions and macros have appeared:
6244 CRYPTO_set_mem_debug_functions() [F]
6245 CRYPTO_get_mem_debug_functions() [F]
6246 CRYPTO_dbg_set_options() [F]
6247 CRYPTO_dbg_get_options() [F]
6248 CRYPTO_malloc_debug_init() [M]
6250 The memory debug functions are NULL by default, unless the library
6251 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6252 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6253 gives the standard debugging functions that come with OpenSSL) or
6254 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6255 provided by the library user) must be used. When the standard
6256 debugging functions are used, CRYPTO_dbg_set_options can be used to
6257 request additional information:
6258 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6259 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6261 Also, things like CRYPTO_set_mem_functions will always give the
6262 expected result (the new set of functions is used for allocation
6263 and deallocation) at all times, regardless of platform and compiler
6266 To finish it up, some functions that were never use in any other
6267 way than through macros have a new API and new semantic:
6270 CRYPTO_dbg_realloc()
6273 All macros of value have retained their old syntax.
6274 [Richard Levitte and Bodo Moeller]
6276 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6277 ordering of SMIMECapabilities wasn't in "strength order" and there
6278 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6282 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6283 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6284 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6286 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6287 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6288 functionality to handle multipart/signed properly) and a utility
6289 called 'smime' to call all this stuff. This is based on code I
6290 originally wrote for Celo who have kindly allowed it to be
6291 included in OpenSSL.
6294 *) Add variants des_set_key_checked and des_set_key_unchecked of
6295 des_set_key (aka des_key_sched). Global variable des_check_key
6296 decides which of these is called by des_set_key; this way
6297 des_check_key behaves as it always did, but applications and
6298 the library itself, which was buggy for des_check_key == 1,
6299 have a cleaner way to pick the version they need.
6302 *) New function PKCS12_newpass() which changes the password of a
6306 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6307 dynamic mix. In both cases the ids can be used as an index into the
6308 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6309 functions so they accept a list of the field values and the
6310 application doesn't need to directly manipulate the X509_TRUST
6314 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6318 *) Modify the way the V3 extension code looks up extensions. This now
6319 works in a similar way to the object code: we have some "standard"
6320 extensions in a static table which is searched with OBJ_bsearch()
6321 and the application can add dynamic ones if needed. The file
6322 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6323 updated whenever a new extension is added to the core code and kept
6324 in ext_nid order. There is a simple program 'tabtest.c' which checks
6325 this. New extensions are not added too often so this file can readily
6326 be maintained manually.
6328 There are two big advantages in doing things this way. The extensions
6329 can be looked up immediately and no longer need to be "added" using
6330 X509V3_add_standard_extensions(): this function now does nothing.
6331 [Side note: I get *lots* of email saying the extension code doesn't
6332 work because people forget to call this function]
6333 Also no dynamic allocation is done unless new extensions are added:
6334 so if we don't add custom extensions there is no need to call
6335 X509V3_EXT_cleanup().
6338 *) Modify enc utility's salting as follows: make salting the default. Add a
6339 magic header, so unsalted files fail gracefully instead of just decrypting
6340 to garbage. This is because not salting is a big security hole, so people
6341 should be discouraged from doing it.
6344 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6345 digest to be passed on the command line but it only used this
6346 parameter when signing a certificate. Modified so all relevant
6347 operations are affected by the digest parameter including the
6348 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6349 DSA key was used because it didn't fix the digest.
6352 *) Initial certificate chain verify code. Currently tests the untrusted
6353 certificates for consistency with the verify purpose (which is set
6354 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6356 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6357 this is because it will reject chains with invalid extensions whereas
6358 every previous version of OpenSSL and SSLeay made no checks at all.
6360 Trust code: checks the root CA for the relevant trust settings. Trust
6361 settings have an initial value consistent with the verify purpose: e.g.
6362 if the verify purpose is for SSL client use it expects the CA to be
6363 trusted for SSL client use. However the default value can be changed to
6364 permit custom trust settings: one example of this would be to only trust
6365 certificates from a specific "secure" set of CAs.
6367 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6368 which should be used for version portability: especially since the
6369 verify structure is likely to change more often now.
6371 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6372 to set them. If not set then assume SSL clients will verify SSL servers
6375 Two new options to the verify program: -untrusted allows a set of
6376 untrusted certificates to be passed in and -purpose which sets the
6377 intended purpose of the certificate. If a purpose is set then the
6378 new chain verify code is used to check extension consistency.
6381 *) Support for the authority information access extension.
6384 *) Modify RSA and DSA PEM read routines to transparently handle
6385 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6386 public keys in a format compatible with certificate
6387 SubjectPublicKeyInfo structures. Unfortunately there were already
6388 functions called *_PublicKey_* which used various odd formats so
6389 these are retained for compatibility: however the DSA variants were
6390 never in a public release so they have been deleted. Changed dsa/rsa
6391 utilities to handle the new format: note no releases ever handled public
6392 keys so we should be OK.
6394 The primary motivation for this change is to avoid the same fiasco
6395 that dogs private keys: there are several incompatible private key
6396 formats some of which are standard and some OpenSSL specific and
6397 require various evil hacks to allow partial transparent handling and
6398 even then it doesn't work with DER formats. Given the option anything
6399 other than PKCS#8 should be dumped: but the other formats have to
6400 stay in the name of compatibility.
6402 With public keys and the benefit of hindsight one standard format
6403 is used which works with EVP_PKEY, RSA or DSA structures: though
6404 it clearly returns an error if you try to read the wrong kind of key.
6406 Added a -pubkey option to the 'x509' utility to output the public key.
6407 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6408 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6409 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6410 that do the same as the EVP_PKEY_assign_*() except they up the
6411 reference count of the added key (they don't "swallow" the
6415 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6416 CRLs would fail if the file contained no certificates or no CRLs:
6417 added a new function to read in both types and return the number
6418 read: this means that if none are read it will be an error. The
6419 DER versions of the certificate and CRL reader would always fail
6420 because it isn't possible to mix certificates and CRLs in DER format
6421 without choking one or the other routine. Changed this to just read
6422 a certificate: this is the best we can do. Also modified the code
6423 in apps/verify.c to take notice of return codes: it was previously
6424 attempting to read in certificates from NULL pointers and ignoring
6425 any errors: this is one reason why the cert and CRL reader seemed
6426 to work. It doesn't check return codes from the default certificate
6427 routines: these may well fail if the certificates aren't installed.
6430 *) Code to support otherName option in GeneralName.
6433 *) First update to verify code. Change the verify utility
6434 so it warns if it is passed a self signed certificate:
6435 for consistency with the normal behaviour. X509_verify
6436 has been modified to it will now verify a self signed
6437 certificate if *exactly* the same certificate appears
6438 in the store: it was previously impossible to trust a
6439 single self signed certificate. This means that:
6440 openssl verify ss.pem
6441 now gives a warning about a self signed certificate but
6442 openssl verify -CAfile ss.pem ss.pem
6446 *) For servers, store verify_result in SSL_SESSION data structure
6447 (and add it to external session representation).
6448 This is needed when client certificate verifications fails,
6449 but an application-provided verification callback (set by
6450 SSL_CTX_set_cert_verify_callback) allows accepting the session
6451 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6452 but returns 1): When the session is reused, we have to set
6453 ssl->verify_result to the appropriate error code to avoid
6455 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6457 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6458 case in PKCS7_dataInit() where the signed PKCS7 structure
6459 didn't contain any existing data because it was being created.
6460 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6462 *) Add a salt to the key derivation routines in enc.c. This
6463 forms the first 8 bytes of the encrypted file. Also add a
6464 -S option to allow a salt to be input on the command line.
6467 *) New function X509_cmp(). Oddly enough there wasn't a function
6468 to compare two certificates. We do this by working out the SHA1
6469 hash and comparing that. X509_cmp() will be needed by the trust
6473 *) SSL_get1_session() is like SSL_get_session(), but increments
6474 the reference count in the SSL_SESSION returned.
6475 [Geoff Thorpe <geoff@eu.c2.net>]
6477 *) Fix for 'req': it was adding a null to request attributes.
6478 Also change the X509_LOOKUP and X509_INFO code to handle
6479 certificate auxiliary information.
6482 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6486 *) Add the possibility to add extra information to the memory leak
6487 detecting output, to form tracebacks, showing from where each
6488 allocation was originated: CRYPTO_push_info("constant string") adds
6489 the string plus current file name and line number to a per-thread
6490 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6491 is like calling CYRPTO_pop_info() until the stack is empty.
6492 Also updated memory leak detection code to be multi-thread-safe.
6495 *) Add options -text and -noout to pkcs7 utility and delete the
6496 encryption options which never did anything. Update docs.
6499 *) Add options to some of the utilities to allow the pass phrase
6500 to be included on either the command line (not recommended on
6501 OSes like Unix) or read from the environment. Update the
6502 manpages and fix a few bugs.
6505 *) Add a few manpages for some of the openssl commands.
6508 *) Fix the -revoke option in ca. It was freeing up memory twice,
6509 leaking and not finding already revoked certificates.
6512 *) Extensive changes to support certificate auxiliary information.
6513 This involves the use of X509_CERT_AUX structure and X509_AUX
6514 functions. An X509_AUX function such as PEM_read_X509_AUX()
6515 can still read in a certificate file in the usual way but it
6516 will also read in any additional "auxiliary information". By
6517 doing things this way a fair degree of compatibility can be
6518 retained: existing certificates can have this information added
6519 using the new 'x509' options.
6521 Current auxiliary information includes an "alias" and some trust
6522 settings. The trust settings will ultimately be used in enhanced
6523 certificate chain verification routines: currently a certificate
6524 can only be trusted if it is self signed and then it is trusted
6528 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6529 The problem was that one of the replacement routines had not been working
6530 since SSLeay releases. For now the offending routine has been replaced
6531 with non-optimised assembler. Even so, this now gives around 95%
6532 performance improvement for 1024 bit RSA signs.
6535 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6536 handling. Most clients have the effective key size in bits equal to
6537 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6538 A few however don't do this and instead use the size of the decrypted key
6539 to determine the RC2 key length and the AlgorithmIdentifier to determine
6540 the effective key length. In this case the effective key length can still
6541 be 40 bits but the key length can be 168 bits for example. This is fixed
6542 by manually forcing an RC2 key into the EVP_PKEY structure because the
6543 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6544 the key length and effective key length are equal.
6547 *) Add a bunch of functions that should simplify the creation of
6548 X509_NAME structures. Now you should be able to do:
6549 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6550 and have it automatically work out the correct field type and fill in
6551 the structures. The more adventurous can try:
6552 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6553 and it will (hopefully) work out the correct multibyte encoding.
6556 *) Change the 'req' utility to use the new field handling and multibyte
6557 copy routines. Before the DN field creation was handled in an ad hoc
6558 way in req, ca, and x509 which was rather broken and didn't support
6559 BMPStrings or UTF8Strings. Since some software doesn't implement
6560 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6561 using the dirstring_type option. See the new comment in the default
6562 openssl.cnf for more info.
6565 *) Make crypto/rand/md_rand.c more robust:
6566 - Assure unique random numbers after fork().
6567 - Make sure that concurrent threads access the global counter and
6568 md serializably so that we never lose entropy in them
6569 or use exactly the same state in multiple threads.
6570 Access to the large state is not always serializable because
6571 the additional locking could be a performance killer, and
6572 md should be large enough anyway.
6575 *) New file apps/app_rand.c with commonly needed functionality
6576 for handling the random seed file.
6578 Use the random seed file in some applications that previously did not:
6580 dsaparam -genkey (which also ignored its '-rand' option),
6583 x509 (when signing).
6584 Except on systems with /dev/urandom, it is crucial to have a random
6585 seed file at least for key creation, DSA signing, and for DH exchanges;
6586 for RSA signatures we could do without one.
6588 gendh and gendsa (unlike genrsa) used to read only the first byte
6589 of each file listed in the '-rand' option. The function as previously
6590 found in genrsa is now in app_rand.c and is used by all programs
6591 that support '-rand'.
6594 *) In RAND_write_file, use mode 0600 for creating files;
6595 don't just chmod when it may be too late.
6598 *) Report an error from X509_STORE_load_locations
6599 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6602 *) New function ASN1_mbstring_copy() this copies a string in either
6603 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6604 into an ASN1_STRING type. A mask of permissible types is passed
6605 and it chooses the "minimal" type to use or an error if not type
6609 *) Add function equivalents to the various macros in asn1.h. The old
6610 macros are retained with an M_ prefix. Code inside the library can
6611 use the M_ macros. External code (including the openssl utility)
6612 should *NOT* in order to be "shared library friendly".
6615 *) Add various functions that can check a certificate's extensions
6616 to see if it usable for various purposes such as SSL client,
6617 server or S/MIME and CAs of these types. This is currently
6618 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6619 verification. Also added a -purpose flag to x509 utility to
6620 print out all the purposes.
6623 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6627 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6628 for, obtain and decode and extension and obtain its critical flag.
6629 This allows all the necessary extension code to be handled in a
6630 single function call.
6633 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6634 platforms. See crypto/rc4/rc4_enc.c for further details.
6637 *) New -noout option to asn1parse. This causes no output to be produced
6638 its main use is when combined with -strparse and -out to extract data
6639 from a file (which may not be in ASN.1 format).
6642 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6643 when producing the local key id.
6644 [Richard Levitte <levitte@stacken.kth.se>]
6646 *) New option -dhparam in s_server. This allows a DH parameter file to be
6647 stated explicitly. If it is not stated then it tries the first server
6648 certificate file. The previous behaviour hard coded the filename
6652 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6653 a public key to be input or output. For example:
6654 openssl rsa -in key.pem -pubout -out pubkey.pem
6655 Also added necessary DSA public key functions to handle this.
6658 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6659 in the message. This was handled by allowing
6660 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6661 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6663 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6664 to the end of the strings whereas this didn't. This would cause problems
6665 if strings read with d2i_ASN1_bytes() were later modified.
6666 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6668 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6669 data and it contains EOF it will end up returning an error. This is
6670 caused by input 46 bytes long. The cause is due to the way base64
6671 BIOs find the start of base64 encoded data. They do this by trying a
6672 trial decode on each line until they find one that works. When they
6673 do a flag is set and it starts again knowing it can pass all the
6674 data directly through the decoder. Unfortunately it doesn't reset
6675 the context it uses. This means that if EOF is reached an attempt
6676 is made to pass two EOFs through the context and this causes the
6677 resulting error. This can also cause other problems as well. As is
6678 usual with these problems it takes *ages* to find and the fix is
6679 trivial: move one line.
6680 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6682 *) Ugly workaround to get s_client and s_server working under Windows. The
6683 old code wouldn't work because it needed to select() on sockets and the
6684 tty (for keypresses and to see if data could be written). Win32 only
6685 supports select() on sockets so we select() with a 1s timeout on the
6686 sockets and then see if any characters are waiting to be read, if none
6687 are present then we retry, we also assume we can always write data to
6688 the tty. This isn't nice because the code then blocks until we've
6689 received a complete line of data and it is effectively polling the
6690 keyboard at 1s intervals: however it's quite a bit better than not
6691 working at all :-) A dedicated Windows application might handle this
6692 with an event loop for example.
6695 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6696 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6697 will be called when RSA_sign() and RSA_verify() are used. This is useful
6698 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6699 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6700 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6701 This necessitated the support of an extra signature type NID_md5_sha1
6702 for SSL signatures and modifications to the SSL library to use it instead
6703 of calling RSA_public_decrypt() and RSA_private_encrypt().
6706 *) Add new -verify -CAfile and -CApath options to the crl program, these
6707 will lookup a CRL issuers certificate and verify the signature in a
6708 similar way to the verify program. Tidy up the crl program so it
6709 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6710 less strict. It will now permit CRL extensions even if it is not
6711 a V2 CRL: this will allow it to tolerate some broken CRLs.
6714 *) Initialize all non-automatic variables each time one of the openssl
6715 sub-programs is started (this is necessary as they may be started
6716 multiple times from the "OpenSSL>" prompt).
6717 [Lennart Bang, Bodo Moeller]
6719 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6720 removing all other RSA functionality (this is what NO_RSA does). This
6721 is so (for example) those in the US can disable those operations covered
6722 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6726 *) Non-copying interface to BIO pairs.
6727 (still largely untested)
6730 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6731 ASCII string. This was handled independently in various places before.
6734 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6735 UTF8 strings a character at a time.
6738 *) Use client_version from client hello to select the protocol
6739 (s23_srvr.c) and for RSA client key exchange verification
6740 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6743 *) Add various utility functions to handle SPKACs, these were previously
6744 handled by poking round in the structure internals. Added new function
6745 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6746 print, verify and generate SPKACs. Based on an original idea from
6747 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6750 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6753 *) Allow the config file extension section to be overwritten on the
6754 command line. Based on an original idea from Massimiliano Pala
6755 <madwolf@comune.modena.it>. The new option is called -extensions
6756 and can be applied to ca, req and x509. Also -reqexts to override
6757 the request extensions in req and -crlexts to override the crl extensions
6761 *) Add new feature to the SPKAC handling in ca. Now you can include
6762 the same field multiple times by preceding it by "XXXX." for example:
6765 this is the same syntax as used in the req config file.
6768 *) Allow certificate extensions to be added to certificate requests. These
6769 are specified in a 'req_extensions' option of the req section of the
6770 config file. They can be printed out with the -text option to req but
6771 are otherwise ignored at present.
6774 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6775 data read consists of only the final block it would not decrypted because
6776 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6777 A misplaced 'break' also meant the decrypted final block might not be
6778 copied until the next read.
6781 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6782 a few extra parameters to the DH structure: these will be useful if
6783 for example we want the value of 'q' or implement X9.42 DH.
6786 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6787 provides hooks that allow the default DSA functions or functions on a
6788 "per key" basis to be replaced. This allows hardware acceleration and
6789 hardware key storage to be handled without major modification to the
6790 library. Also added low level modexp hooks and CRYPTO_EX structure and
6791 associated functions.
6794 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6795 as "read only": it can't be written to and the buffer it points to will
6796 not be freed. Reading from a read only BIO is much more efficient than
6797 a normal memory BIO. This was added because there are several times when
6798 an area of memory needs to be read from a BIO. The previous method was
6799 to create a memory BIO and write the data to it, this results in two
6800 copies of the data and an O(n^2) reading algorithm. There is a new
6801 function BIO_new_mem_buf() which creates a read only memory BIO from
6802 an area of memory. Also modified the PKCS#7 routines to use read only
6806 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6807 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6808 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6809 but a retry condition occured while trying to read the rest.
6812 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6813 NID_pkcs7_encrypted by default: this was wrong since this should almost
6814 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6815 the encrypted data type: this is a more sensible place to put it and it
6816 allows the PKCS#12 code to be tidied up that duplicated this
6820 *) Changed obj_dat.pl script so it takes its input and output files on
6821 the command line. This should avoid shell escape redirection problems
6825 *) Initial support for certificate extension requests, these are included
6826 in things like Xenroll certificate requests. Included functions to allow
6827 extensions to be obtained and added.
6830 *) -crlf option to s_client and s_server for sending newlines as
6831 CRLF (as required by many protocols).
6834 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6836 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6837 [Ralf S. Engelschall]
6839 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6840 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6842 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6846 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6847 DH parameters/keys (q is lost during that conversion, but the resulting
6848 DH parameters contain its length).
6850 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6851 much faster than DH_generate_parameters (which creates parameters
6852 where p = 2*q + 1), and also the smaller q makes DH computations
6853 much more efficient (160-bit exponentiation instead of 1024-bit
6854 exponentiation); so this provides a convenient way to support DHE
6855 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6856 utter importance to use
6857 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6859 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6860 when such DH parameters are used, because otherwise small subgroup
6861 attacks may become possible!
6864 *) Avoid memory leak in i2d_DHparams.
6867 *) Allow the -k option to be used more than once in the enc program:
6868 this allows the same encrypted message to be read by multiple recipients.
6871 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6872 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6873 it will always use the numerical form of the OID, even if it has a short
6877 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6878 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6879 otherwise bn_mod_exp was called. In the case of hardware keys for example
6880 no private key components need be present and it might store extra data
6881 in the RSA structure, which cannot be accessed from bn_mod_exp.
6882 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6883 private key operations.
6886 *) Added support for SPARC Linux.
6889 *) pem_password_cb function type incompatibly changed from
6890 typedef int pem_password_cb(char *buf, int size, int rwflag);
6892 ....(char *buf, int size, int rwflag, void *userdata);
6893 so that applications can pass data to their callbacks:
6894 The PEM[_ASN1]_{read,write}... functions and macros now take an
6895 additional void * argument, which is just handed through whenever
6896 the password callback is called.
6897 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6899 New function SSL_CTX_set_default_passwd_cb_userdata.
6901 Compatibility note: As many C implementations push function arguments
6902 onto the stack in reverse order, the new library version is likely to
6903 interoperate with programs that have been compiled with the old
6904 pem_password_cb definition (PEM_whatever takes some data that
6905 happens to be on the stack as its last argument, and the callback
6906 just ignores this garbage); but there is no guarantee whatsoever that
6909 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6910 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6911 problems not only on Windows, but also on some Unix platforms.
6912 To avoid problematic command lines, these definitions are now in an
6913 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6914 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6917 *) MIPS III/IV assembler module is reimplemented.
6920 *) More DES library cleanups: remove references to srand/rand and
6921 delete an unused file.
6924 *) Add support for the the free Netwide assembler (NASM) under Win32,
6925 since not many people have MASM (ml) and it can be hard to obtain.
6926 This is currently experimental but it seems to work OK and pass all
6927 the tests. Check out INSTALL.W32 for info.
6930 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6931 without temporary keys kept an extra copy of the server key,
6932 and connections with temporary keys did not free everything in case
6936 *) New function RSA_check_key and new openssl rsa option -check
6937 for verifying the consistency of RSA keys.
6938 [Ulf Moeller, Bodo Moeller]
6940 *) Various changes to make Win32 compile work:
6941 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6942 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6943 comparison" warnings.
6944 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6947 *) Add a debugging option to PKCS#5 v2 key generation function: when
6948 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6949 derived keys are printed to stderr.
6952 *) Copy the flags in ASN1_STRING_dup().
6953 [Roman E. Pavlov <pre@mo.msk.ru>]
6955 *) The x509 application mishandled signing requests containing DSA
6956 keys when the signing key was also DSA and the parameters didn't match.
6958 It was supposed to omit the parameters when they matched the signing key:
6959 the verifying software was then supposed to automatically use the CA's
6960 parameters if they were absent from the end user certificate.
6962 Omitting parameters is no longer recommended. The test was also
6963 the wrong way round! This was probably due to unusual behaviour in
6964 EVP_cmp_parameters() which returns 1 if the parameters match.
6965 This meant that parameters were omitted when they *didn't* match and
6966 the certificate was useless. Certificates signed with 'ca' didn't have
6968 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6970 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6971 The interface is as follows:
6972 Applications can use
6973 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6974 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6975 "off" is now the default.
6976 The library internally uses
6977 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6978 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6979 to disable memory-checking temporarily.
6981 Some inconsistent states that previously were possible (and were
6982 even the default) are now avoided.
6984 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6985 with each memory chunk allocated; this is occasionally more helpful
6986 than just having a counter.
6988 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6990 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6994 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6995 which largely parallels "options", but is for changing API behaviour,
6996 whereas "options" are about protocol behaviour.
6997 Initial "mode" flags are:
6999 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7000 a single record has been written.
7001 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7002 retries use the same buffer location.
7003 (But all of the contents must be
7007 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7010 *) Fix problems with no-hmac etc.
7011 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7013 *) New functions RSA_get_default_method(), RSA_set_method() and
7014 RSA_get_method(). These allows replacement of RSA_METHODs without having
7015 to mess around with the internals of an RSA structure.
7018 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7019 Also really enable memory leak checks in openssl.c and in some
7021 [Chad C. Mulligan, Bodo Moeller]
7023 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7024 up the length of negative integers. This has now been simplified to just
7025 store the length when it is first determined and use it later, rather
7026 than trying to keep track of where data is copied and updating it to
7028 [Steve Henson, reported by Brien Wheeler
7029 <bwheeler@authentica-security.com>]
7031 *) Add a new function PKCS7_signatureVerify. This allows the verification
7032 of a PKCS#7 signature but with the signing certificate passed to the
7033 function itself. This contrasts with PKCS7_dataVerify which assumes the
7034 certificate is present in the PKCS#7 structure. This isn't always the
7035 case: certificates can be omitted from a PKCS#7 structure and be
7036 distributed by "out of band" means (such as a certificate database).
7039 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7040 function prototypes in pem.h, also change util/mkdef.pl to add the
7041 necessary function names.
7044 *) mk1mf.pl (used by Windows builds) did not properly read the
7045 options set by Configure in the top level Makefile, and Configure
7046 was not even able to write more than one option correctly.
7047 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7050 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7051 file to be loaded from a BIO or FILE pointer. The BIO version will
7052 for example allow memory BIOs to contain config info.
7055 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7056 Whoever hopes to achieve shared-library compatibility across versions
7057 must use this, not the compile-time macro.
7058 (Exercise 0.9.4: Which is the minimum library version required by
7060 Note: All this applies only to multi-threaded programs, others don't
7064 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7065 through a BIO pair triggered the default case, i.e.
7066 SSLerr(...,SSL_R_UNKNOWN_STATE).
7069 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7070 can use the SSL library even if none of the specific BIOs is
7074 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7075 for the encoded length.
7076 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7078 *) Add initial documentation of the X509V3 functions.
7081 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7082 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7083 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7084 secure PKCS#8 private key format with a high iteration count.
7087 *) Fix determination of Perl interpreter: A perl or perl5
7088 _directory_ in $PATH was also accepted as the interpreter.
7089 [Ralf S. Engelschall]
7091 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7092 wrong with it but it was very old and did things like calling
7093 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7097 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7098 to use the new extension code.
7101 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7102 with macros. This should make it easier to change their form, add extra
7103 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7107 *) Add to configuration table a new entry that can specify an alternative
7108 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7109 according to Mark Crispin <MRC@Panda.COM>.
7113 *) DES CBC did not update the IV. Weird.
7116 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7117 Changing the behaviour of the former might break existing programs --
7118 where IV updating is needed, des_ncbc_encrypt can be used.
7121 *) When bntest is run from "make test" it drives bc to check its
7122 calculations, as well as internally checking them. If an internal check
7123 fails, it needs to cause bc to give a non-zero result or make test carries
7124 on without noticing the failure. Fixed.
7127 *) DES library cleanups.
7130 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7131 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7132 ciphers. NOTE: although the key derivation function has been verified
7133 against some published test vectors it has not been extensively tested
7134 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7138 *) Instead of "mkdir -p", which is not fully portable, use new
7139 Perl script "util/mkdir-p.pl".
7142 *) Rewrite the way password based encryption (PBE) is handled. It used to
7143 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7144 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7145 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7146 the 'parameter' field of the AlgorithmIdentifier is passed to the
7147 underlying key generation function so it must do its own ASN1 parsing.
7148 This has also changed the EVP_PBE_CipherInit() function which now has a
7149 'parameter' argument instead of literal salt and iteration count values
7150 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7153 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7154 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7155 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7156 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7157 value was just used as a "magic string" and not used directly its
7158 value doesn't matter.
7161 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7165 *) "linux-sparc64" configuration (ultrapenguin).
7166 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7167 "linux-sparc" configuration.
7168 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7170 *) config now generates no-xxx options for missing ciphers.
7173 *) Support the EBCDIC character set (work in progress).
7174 File ebcdic.c not yet included because it has a different license.
7175 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7177 *) Support BS2000/OSD-POSIX.
7178 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7180 *) Make callbacks for key generation use void * instead of char *.
7183 *) Make S/MIME samples compile (not yet tested).
7186 *) Additional typesafe stacks.
7189 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7193 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7195 *) New configuration variant "sco5-gcc".
7197 *) Updated some demos.
7198 [Sean O Riordain, Wade Scholine]
7200 *) Add missing BIO_free at exit of pkcs12 application.
7203 *) Fix memory leak in conf.c.
7206 *) Updates for Win32 to assembler version of MD5.
7209 *) Set #! path to perl in apps/der_chop to where we found it
7210 instead of using a fixed path.
7213 *) SHA library changes for irix64-mips4-cc.
7216 *) Improvements for VMS support.
7220 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7222 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7223 This also avoids the problems with SC4.2 and unpatched SC5.
7224 [Andy Polyakov <appro@fy.chalmers.se>]
7226 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7227 These are required because of the typesafe stack would otherwise break
7228 existing code. If old code used a structure member which used to be STACK
7229 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7230 sk_num or sk_value it would produce an error because the num, data members
7231 are not present in STACK_OF. Now it just produces a warning. sk_set
7232 replaces the old method of assigning a value to sk_value
7233 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7234 that does this will no longer work (and should use sk_set instead) but
7235 this could be regarded as a "questionable" behaviour anyway.
7238 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7239 correctly handle encrypted S/MIME data.
7242 *) Change type of various DES function arguments from des_cblock
7243 (which means, in function argument declarations, pointer to char)
7244 to des_cblock * (meaning pointer to array with 8 char elements),
7245 which allows the compiler to do more typechecking; it was like
7246 that back in SSLeay, but with lots of ugly casts.
7248 Introduce new type const_des_cblock.
7251 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7252 problems: find RecipientInfo structure that matches recipient certificate
7253 and initialise the ASN1 structures properly based on passed cipher.
7256 *) Belatedly make the BN tests actually check the results.
7259 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7260 to and from BNs: it was completely broken. New compilation option
7261 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7262 key elements as negative integers.
7265 *) Reorganize and speed up MD5.
7266 [Andy Polyakov <appro@fy.chalmers.se>]
7269 [Richard Levitte <richard@levitte.org>]
7271 *) New option -out to asn1parse to allow the parsed structure to be
7272 output to a file. This is most useful when combined with the -strparse
7273 option to examine the output of things like OCTET STRINGS.
7276 *) Make SSL library a little more fool-proof by not requiring any longer
7277 that SSL_set_{accept,connect}_state be called before
7278 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7279 in many applications because usually everything *appeared* to work as
7280 intended anyway -- now it really works as intended).
7283 *) Move openssl.cnf out of lib/.
7286 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7287 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7288 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7289 [Ralf S. Engelschall]
7291 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7292 handle PKCS#7 enveloped data properly.
7293 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7295 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7296 copying pointers. The cert_st handling is changed by this in
7297 various ways (and thus what used to be known as ctx->default_cert
7298 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7299 any longer when s->cert does not give us what we need).
7300 ssl_cert_instantiate becomes obsolete by this change.
7301 As soon as we've got the new code right (possibly it already is?),
7302 we have solved a couple of bugs of the earlier code where s->cert
7303 was used as if it could not have been shared with other SSL structures.
7305 Note that using the SSL API in certain dirty ways now will result
7306 in different behaviour than observed with earlier library versions:
7307 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7308 does not influence s as it used to.
7310 In order to clean up things more thoroughly, inside SSL_SESSION
7311 we don't use CERT any longer, but a new structure SESS_CERT
7312 that holds per-session data (if available); currently, this is
7313 the peer's certificate chain and, for clients, the server's certificate
7314 and temporary key. CERT holds only those values that can have
7315 meaningful defaults in an SSL_CTX.
7318 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7319 from the internal representation. Various PKCS#7 fixes: remove some
7320 evil casts and set the enc_dig_alg field properly based on the signing
7324 *) Allow PKCS#12 password to be set from the command line or the
7325 environment. Let 'ca' get its config file name from the environment
7326 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7330 *) Allow certificate policies extension to use an IA5STRING for the
7331 organization field. This is contrary to the PKIX definition but
7332 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7336 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7337 without disallowing inline assembler and the like for non-pedantic builds.
7340 *) Support Borland C++ builder.
7341 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7346 *) SHA-1 cleanups and performance enhancements.
7347 [Andy Polyakov <appro@fy.chalmers.se>]
7349 *) Sparc v8plus assembler for the bignum library.
7350 [Andy Polyakov <appro@fy.chalmers.se>]
7352 *) Accept any -xxx and +xxx compiler options in Configure.
7355 *) Update HPUX configuration.
7358 *) Add missing sk_<type>_unshift() function to safestack.h
7359 [Ralf S. Engelschall]
7361 *) New function SSL_CTX_use_certificate_chain_file that sets the
7362 "extra_cert"s in addition to the certificate. (This makes sense
7363 only for "PEM" format files, as chains as a whole are not
7367 *) Support verify_depth from the SSL API.
7368 x509_vfy.c had what can be considered an off-by-one-error:
7369 Its depth (which was not part of the external interface)
7370 was actually counting the number of certificates in a chain;
7371 now it really counts the depth.
7374 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7375 instead of X509err, which often resulted in confusing error
7376 messages since the error codes are not globally unique
7377 (e.g. an alleged error in ssl3_accept when a certificate
7378 didn't match the private key).
7380 *) New function SSL_CTX_set_session_id_context that allows to set a default
7381 value (so that you don't need SSL_set_session_id_context for each
7382 connection using the SSL_CTX).
7385 *) OAEP decoding bug fix.
7388 *) Support INSTALL_PREFIX for package builders, as proposed by
7392 *) New Configure options "threads" and "no-threads". For systems
7393 where the proper compiler options are known (currently Solaris
7394 and Linux), "threads" is the default.
7397 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7400 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7401 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7402 such as /usr/local/bin.
7405 *) "make linux-shared" to build shared libraries.
7406 [Niels Poppe <niels@netbox.org>]
7408 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7411 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7412 extension adding in x509 utility.
7415 *) Remove NOPROTO sections and error code comments.
7418 *) Partial rewrite of the DEF file generator to now parse the ANSI
7422 *) New Configure options --prefix=DIR and --openssldir=DIR.
7425 *) Complete rewrite of the error code script(s). It is all now handled
7426 by one script at the top level which handles error code gathering,
7427 header rewriting and C source file generation. It should be much better
7428 than the old method: it now uses a modified version of Ulf's parser to
7429 read the ANSI prototypes in all header files (thus the old K&R definitions
7430 aren't needed for error creation any more) and do a better job of
7431 translating function codes into names. The old 'ASN1 error code imbedded
7432 in a comment' is no longer necessary and it doesn't use .err files which
7433 have now been deleted. Also the error code call doesn't have to appear all
7434 on one line (which resulted in some large lines...).
7437 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7440 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7441 0 (which usually indicates a closed connection), but continue reading.
7444 *) Fix some race conditions.
7447 *) Add support for CRL distribution points extension. Add Certificate
7448 Policies and CRL distribution points documentation.
7451 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7454 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7455 8 of keying material. Merlin has also confirmed interop with this fix
7456 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7457 [Merlin Hughes <merlin@baltimore.ie>]
7459 *) Fix lots of warnings.
7460 [Richard Levitte <levitte@stacken.kth.se>]
7462 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7463 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7464 [Richard Levitte <levitte@stacken.kth.se>]
7466 *) Fix problems with sizeof(long) == 8.
7467 [Andy Polyakov <appro@fy.chalmers.se>]
7469 *) Change functions to ANSI C.
7472 *) Fix typos in error codes.
7473 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7475 *) Remove defunct assembler files from Configure.
7478 *) SPARC v8 assembler BIGNUM implementation.
7479 [Andy Polyakov <appro@fy.chalmers.se>]
7481 *) Support for Certificate Policies extension: both print and set.
7482 Various additions to support the r2i method this uses.
7485 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7486 return a const string when you are expecting an allocated buffer.
7489 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7490 types DirectoryString and DisplayText.
7493 *) Add code to allow r2i extensions to access the configuration database,
7494 add an LHASH database driver and add several ctx helper functions.
7497 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7498 fail when they extended the size of a BIGNUM.
7501 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7502 support typesafe stack.
7505 *) Fix typo in SSL_[gs]et_options().
7506 [Nils Frostberg <nils@medcom.se>]
7508 *) Delete various functions and files that belonged to the (now obsolete)
7509 old X509V3 handling code.
7512 *) New Configure option "rsaref".
7515 *) Don't auto-generate pem.h.
7518 *) Introduce type-safe ASN.1 SETs.
7521 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7522 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7524 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7525 that links with OpenSSL (well at least cause lots of warnings), but fear
7526 not: the conversion is trivial, and it eliminates loads of evil casts. A
7527 few STACKed things have been converted already. Feel free to convert more.
7528 In the fullness of time, I'll do away with the STACK type altogether.
7531 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7532 specified in <certfile> by updating the entry in the index.txt file.
7533 This way one no longer has to edit the index.txt file manually for
7534 revoking a certificate. The -revoke option does the gory details now.
7535 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7537 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7538 `-text' option at all and this way the `-noout -text' combination was
7539 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7540 [Ralf S. Engelschall]
7542 *) Make sure a corresponding plain text error message exists for the
7543 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7544 verify callback function determined that a certificate was revoked.
7545 [Ralf S. Engelschall]
7547 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7548 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7549 all available cipers including rc5, which was forgotten until now.
7550 In order to let the testing shell script know which algorithms
7551 are available, a new (up to now undocumented) command
7552 "openssl list-cipher-commands" is used.
7555 *) Bugfix: s_client occasionally would sleep in select() when
7556 it should have checked SSL_pending() first.
7559 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7560 the raw DSA values prior to ASN.1 encoding.
7563 *) Tweaks to Configure
7564 [Niels Poppe <niels@netbox.org>]
7566 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7570 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7573 *) New config option to avoid instructions that are illegal on the 80386.
7574 The default code is faster, but requires at least a 486.
7577 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7578 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7579 same as SSL2_VERSION anyway.
7582 *) New "-showcerts" option for s_client.
7585 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7586 application. Various cleanups and fixes.
7589 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7590 modify error routines to work internally. Add error codes and PBE init
7591 to library startup routines.
7594 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7595 packing functions to asn1 and evp. Changed function names and error
7596 codes along the way.
7599 *) PKCS12 integration: and so it begins... First of several patches to
7600 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7601 objects to objects.h
7604 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7605 and display support for Thawte strong extranet extension.
7608 *) Add LinuxPPC support.
7609 [Jeff Dubrule <igor@pobox.org>]
7611 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7612 bn_div_words in alpha.s.
7613 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7615 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7616 OAEP isn't supported when OpenSSL is built with RSAref.
7617 [Ulf Moeller <ulf@fitug.de>]
7619 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7620 so they no longer are missing under -DNOPROTO.
7621 [Soren S. Jorvang <soren@t.dk>]
7624 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7626 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7627 doesn't work when the session is reused. Coming soon!
7630 *) Fix a security hole, that allows sessions to be reused in the wrong
7631 context thus bypassing client cert protection! All software that uses
7632 client certs and session caches in multiple contexts NEEDS PATCHING to
7633 allow session reuse! A fuller solution is in the works.
7634 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7636 *) Some more source tree cleanups (removed obsolete files
7637 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7638 permission on "config" script to be executable) and a fix for the INSTALL
7640 [Ulf Moeller <ulf@fitug.de>]
7642 *) Remove some legacy and erroneous uses of malloc, free instead of
7644 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7646 *) Make rsa_oaep_test return non-zero on error.
7647 [Ulf Moeller <ulf@fitug.de>]
7649 *) Add support for native Solaris shared libraries. Configure
7650 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7651 if someone would make that last step automatic.
7652 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7654 *) ctx_size was not built with the right compiler during "make links". Fixed.
7657 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7658 except NULL ciphers". This means the default cipher list will no longer
7659 enable NULL ciphers. They need to be specifically enabled e.g. with
7660 the string "DEFAULT:eNULL".
7663 *) Fix to RSA private encryption routines: if p < q then it would
7664 occasionally produce an invalid result. This will only happen with
7665 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7668 *) Be less restrictive and allow also `perl util/perlpath.pl
7669 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7670 because this way one can also use an interpreter named `perl5' (which is
7671 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7672 installed as `perl').
7673 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7675 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7676 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7678 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7679 advapi32.lib to Win32 build and change the pem test comparision
7680 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7681 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7682 and crypto/des/ede_cbcm_enc.c.
7685 *) DES quad checksum was broken on big-endian architectures. Fixed.
7688 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7689 Win32 test batch file so it (might) work again. The Win32 test batch file
7690 is horrible: I feel ill....
7693 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7694 in e_os.h. Audit of header files to check ANSI and non ANSI
7695 sections: 10 functions were absent from non ANSI section and not exported
7696 from Windows DLLs. Fixed up libeay.num for new functions.
7699 *) Make `openssl version' output lines consistent.
7700 [Ralf S. Engelschall]
7702 *) Fix Win32 symbol export lists for BIO functions: Added
7703 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7704 to ms/libeay{16,32}.def.
7705 [Ralf S. Engelschall]
7707 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7708 fine under Unix and passes some trivial tests I've now added. But the
7709 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7710 added to make sure no one expects that this stuff really works in the
7711 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7712 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7714 [Ralf S. Engelschall]
7716 *) Fix the generation of two part addresses in perl.
7717 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7719 *) Add config entry for Linux on MIPS.
7720 [John Tobey <jtobey@channel1.com>]
7722 *) Make links whenever Configure is run, unless we are on Windoze.
7725 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7726 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7730 *) Add a useful kludge to allow package maintainers to specify compiler and
7731 other platforms details on the command line without having to patch the
7732 Configure script everytime: One now can use ``perl Configure
7733 <id>:<details>'', i.e. platform ids are allowed to have details appended
7734 to them (seperated by colons). This is treated as there would be a static
7735 pre-configured entry in Configure's %table under key <id> with value
7736 <details> and ``perl Configure <id>'' is called. So, when you want to
7737 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7738 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7739 now, which overrides the FreeBSD-elf entry on-the-fly.
7740 [Ralf S. Engelschall]
7742 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7745 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7746 on the `perl Configure ...' command line. This way one can compile
7747 OpenSSL libraries with Position Independent Code (PIC) which is needed
7748 for linking it into DSOs.
7749 [Ralf S. Engelschall]
7751 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7755 *) Cleaned up the LICENSE document: The official contact for any license
7756 questions now is the OpenSSL core team under openssl-core@openssl.org.
7757 And add a paragraph about the dual-license situation to make sure people
7758 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7759 to the OpenSSL toolkit.
7760 [Ralf S. Engelschall]
7762 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7763 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7764 Additonally cleaned up the `make links' target: Remove unnecessary
7765 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7766 to speed processing and no longer clutter the display with confusing
7767 stuff. Instead only the actually done links are displayed.
7768 [Ralf S. Engelschall]
7770 *) Permit null encryption ciphersuites, used for authentication only. It used
7771 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7772 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7776 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7777 signed attributes when verifying signatures (this would break them),
7778 the detached data encoding was wrong and public keys obtained using
7779 X509_get_pubkey() weren't freed.
7782 *) Add text documentation for the BUFFER functions. Also added a work around
7783 to a Win95 console bug. This was triggered by the password read stuff: the
7784 last character typed gets carried over to the next fread(). If you were
7785 generating a new cert request using 'req' for example then the last
7786 character of the passphrase would be CR which would then enter the first
7790 *) Added the new `Includes OpenSSL Cryptography Software' button as
7791 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7792 button and can be used by applications based on OpenSSL to show the
7793 relationship to the OpenSSL project.
7794 [Ralf S. Engelschall]
7796 *) Remove confusing variables in function signatures in files
7797 ssl/ssl_lib.c and ssl/ssl.h.
7798 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7800 *) Don't install bss_file.c under PREFIX/include/
7801 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7803 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7804 functions that return function pointers and has support for NT specific
7805 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7806 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7807 unsigned to signed types: this was killing the Win32 compile.
7810 *) Add new certificate file to stack functions,
7811 SSL_add_dir_cert_subjects_to_stack() and
7812 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7813 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7814 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7815 This means that Apache-SSL and similar packages don't have to mess around
7816 to add as many CAs as they want to the preferred list.
7819 *) Experiment with doxygen documentation. Currently only partially applied to
7821 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7822 openssl.doxy as the configuration file.
7825 *) Get rid of remaining C++-style comments which strict C compilers hate.
7826 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7828 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7829 compiled in by default: it has problems with large keys.
7832 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7833 DH private keys and/or callback functions which directly correspond to
7834 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7835 is needed for applications which have to configure certificates on a
7836 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7838 For the RSA certificate situation is makes no difference, but
7839 for the DSA certificate situation this fixes the "no shared cipher"
7840 problem where the OpenSSL cipher selection procedure failed because the
7841 temporary keys were not overtaken from the context and the API provided
7842 no way to reconfigure them.
7843 The new functions now let applications reconfigure the stuff and they
7844 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7845 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7846 non-public-API function ssl_cert_instantiate() is used as a helper
7847 function and also to reduce code redundancy inside ssl_rsa.c.
7848 [Ralf S. Engelschall]
7850 *) Move s_server -dcert and -dkey options out of the undocumented feature
7851 area because they are useful for the DSA situation and should be
7852 recognized by the users.
7853 [Ralf S. Engelschall]
7855 *) Fix the cipher decision scheme for export ciphers: the export bits are
7856 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7857 SSL_EXP_MASK. So, the original variable has to be used instead of the
7858 already masked variable.
7859 [Richard Levitte <levitte@stacken.kth.se>]
7861 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7862 [Richard Levitte <levitte@stacken.kth.se>]
7864 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7865 from `int' to `unsigned int' because it's a length and initialized by
7866 EVP_DigestFinal() which expects an `unsigned int *'.
7867 [Richard Levitte <levitte@stacken.kth.se>]
7869 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7870 script. Instead use the usual Shell->Perl transition trick.
7871 [Ralf S. Engelschall]
7873 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7874 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7875 -noout -modulus' as it's already the case for `openssl rsa -noout
7876 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7877 currently the public key is printed (a decision which was already done by
7878 `openssl dsa -modulus' in the past) which serves a similar purpose.
7879 Additionally the NO_RSA no longer completely removes the whole -modulus
7880 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7882 [Ralf S. Engelschall]
7884 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7885 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7886 [Arne Ansper <arne@ats.cyber.ee>]
7888 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7889 to be added. Now both 'req' and 'ca' can use new objects defined in the
7893 *) Add cool BIO that does syslog (or event log on NT).
7894 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7896 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7897 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7898 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7899 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7902 *) Add preliminary config info for new extension code.
7905 *) Make RSA_NO_PADDING really use no padding.
7906 [Ulf Moeller <ulf@fitug.de>]
7908 *) Generate errors when private/public key check is done.
7911 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7912 for some CRL extensions and new objects added.
7915 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7916 key usage extension and fuller support for authority key id.
7919 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7920 padding method for RSA, which is recommended for new applications in PKCS
7921 #1 v2.0 (RFC 2437, October 1998).
7922 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7923 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7924 against Bleichbacher's attack on RSA.
7925 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7928 *) Updates to the new SSL compression code
7929 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7931 *) Fix so that the version number in the master secret, when passed
7932 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7933 (because the server will not accept higher), that the version number
7934 is 0x03,0x01, not 0x03,0x00
7935 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7937 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7938 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7939 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7942 *) Support for RAW extensions where an arbitrary extension can be
7943 created by including its DER encoding. See apps/openssl.cnf for
7947 *) Make sure latest Perl versions don't interpret some generated C array
7948 code as Perl array code in the crypto/err/err_genc.pl script.
7949 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7951 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7952 not many people have the assembler. Various Win32 compilation fixes and
7953 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7957 *) Modify configure script 'Configure' to automatically create crypto/date.h
7958 file under Win32 and also build pem.h from pem.org. New script
7959 util/mkfiles.pl to create the MINFO file on environments that can't do a
7960 'make files': perl util/mkfiles.pl >MINFO should work.
7963 *) Major rework of DES function declarations, in the pursuit of correctness
7964 and purity. As a result, many evil casts evaporated, and some weirdness,
7965 too. You may find this causes warnings in your code. Zapping your evil
7966 casts will probably fix them. Mostly.
7969 *) Fix for a typo in asn1.h. Bug fix to object creation script
7970 obj_dat.pl. It considered a zero in an object definition to mean
7971 "end of object": none of the objects in objects.h have any zeros
7972 so it wasn't spotted.
7973 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7975 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7976 Masking (CBCM). In the absence of test vectors, the best I have been able
7977 to do is check that the decrypt undoes the encrypt, so far. Send me test
7978 vectors if you have them.
7981 *) Correct calculation of key length for export ciphers (too much space was
7982 allocated for null ciphers). This has not been tested!
7985 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7986 message is now correct (it understands "crypto" and "ssl" on its
7987 command line). There is also now an "update" option. This will update
7988 the util/ssleay.num and util/libeay.num files with any new functions.
7990 perl util/mkdef.pl crypto ssl update
7991 it will update them.
7994 *) Overhauled the Perl interface (perl/*):
7995 - ported BN stuff to OpenSSL's different BN library
7996 - made the perl/ source tree CVS-aware
7997 - renamed the package from SSLeay to OpenSSL (the files still contain
7998 their history because I've copied them in the repository)
7999 - removed obsolete files (the test scripts will be replaced
8000 by better Test::Harness variants in the future)
8001 [Ralf S. Engelschall]
8003 *) First cut for a very conservative source tree cleanup:
8004 1. merge various obsolete readme texts into doc/ssleay.txt
8005 where we collect the old documents and readme texts.
8006 2. remove the first part of files where I'm already sure that we no
8007 longer need them because of three reasons: either they are just temporary
8008 files which were left by Eric or they are preserved original files where
8009 I've verified that the diff is also available in the CVS via "cvs diff
8010 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8011 the crypto/md/ stuff).
8012 [Ralf S. Engelschall]
8014 *) More extension code. Incomplete support for subject and issuer alt
8015 name, issuer and authority key id. Change the i2v function parameters
8016 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8017 what that's for :-) Fix to ASN1 macro which messed up
8018 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8021 *) Preliminary support for ENUMERATED type. This is largely copied from the
8025 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8026 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8028 *) Make sure `make rehash' target really finds the `openssl' program.
8029 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8031 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8032 like to hear about it if this slows down other processors.
8035 *) Add CygWin32 platform information to Configure script.
8036 [Alan Batie <batie@aahz.jf.intel.com>]
8038 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8039 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8041 *) New program nseq to manipulate netscape certificate sequences
8044 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8048 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8049 but the BN code had some problems that would cause failures when
8050 doing certificate verification and some other functions.
8051 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8053 *) Add ASN1 and PEM code to support netscape certificate sequences.
8056 *) Add ASN1 and PEM code to support netscape certificate sequences.
8059 *) Add several PKIX and private extended key usage OIDs.
8062 *) Modify the 'ca' program to handle the new extension code. Modify
8063 openssl.cnf for new extension format, add comments.
8066 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8067 and add a sample to openssl.cnf so req -x509 now adds appropriate
8071 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8072 error code, add initial support to X509_print() and x509 application.
8075 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8076 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8077 stuff is currently isolated and isn't even compiled yet.
8080 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8081 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8082 Removed the versions check from X509 routines when loading extensions:
8083 this allows certain broken certificates that don't set the version
8084 properly to be processed.
8087 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8088 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8089 can still be regenerated with "make depend".
8092 *) Spelling mistake in C version of CAST-128.
8093 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8095 *) Changes to the error generation code. The perl script err-code.pl
8096 now reads in the old error codes and retains the old numbers, only
8097 adding new ones if necessary. It also only changes the .err files if new
8098 codes are added. The makefiles have been modified to only insert errors
8099 when needed (to avoid needlessly modifying header files). This is done
8100 by only inserting errors if the .err file is newer than the auto generated
8101 C file. To rebuild all the error codes from scratch (the old behaviour)
8102 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8103 or delete all the .err files.
8106 *) CAST-128 was incorrectly implemented for short keys. The C version has
8107 been fixed, but is untested. The assembler versions are also fixed, but
8108 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8109 to regenerate it if needed.
8110 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8111 Hagino <itojun@kame.net>]
8113 *) File was opened incorrectly in randfile.c.
8114 [Ulf Möller <ulf@fitug.de>]
8116 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8117 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8118 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8119 al: it's just almost always a UTCTime. Note this patch adds new error
8120 codes so do a "make errors" if there are problems.
8123 *) Correct Linux 1 recognition in config.
8124 [Ulf Möller <ulf@fitug.de>]
8126 *) Remove pointless MD5 hash when using DSA keys in ca.
8127 [Anonymous <nobody@replay.com>]
8129 *) Generate an error if given an empty string as a cert directory. Also
8130 generate an error if handed NULL (previously returned 0 to indicate an
8131 error, but didn't set one).
8132 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8134 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8137 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8138 parameters. This was causing a warning which killed off the Win32 compile.
8141 *) Remove C++ style comments from crypto/bn/bn_local.h.
8142 [Neil Costigan <neil.costigan@celocom.com>]
8144 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8145 based on a text string, looking up short and long names and finally
8146 "dot" format. The "dot" format stuff didn't work. Added new function
8147 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8148 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8149 OID is not part of the table.
8152 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8153 X509_LOOKUP_by_alias().
8156 *) Sort openssl functions by name.
8159 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8160 encryption from sample DSA keys (in case anyone is interested the password
8164 *) Make _all_ *_free functions accept a NULL pointer.
8165 [Frans Heymans <fheymans@isaserver.be>]
8167 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8169 [Anonymous <nobody@replay.com>]
8171 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8172 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8174 *) Don't blow it for numeric -newkey arguments to apps/req.
8175 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8177 *) Temp key "for export" tests were wrong in s3_srvr.c.
8178 [Anonymous <nobody@replay.com>]
8180 *) Add prototype for temp key callback functions
8181 SSL_CTX_set_tmp_{rsa,dh}_callback().
8184 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8185 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8188 *) X509_name_add_entry() freed the wrong thing after an error.
8189 [Arne Ansper <arne@ats.cyber.ee>]
8191 *) rsa_eay.c would attempt to free a NULL context.
8192 [Arne Ansper <arne@ats.cyber.ee>]
8194 *) BIO_s_socket() had a broken should_retry() on Windoze.
8195 [Arne Ansper <arne@ats.cyber.ee>]
8197 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8198 [Arne Ansper <arne@ats.cyber.ee>]
8200 *) Make sure the already existing X509_STORE->depth variable is initialized
8201 in X509_STORE_new(), but document the fact that this variable is still
8202 unused in the certificate verification process.
8203 [Ralf S. Engelschall]
8205 *) Fix the various library and apps files to free up pkeys obtained from
8206 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8209 *) Fix reference counting in X509_PUBKEY_get(). This makes
8210 demos/maurice/example2.c work, amongst others, probably.
8211 [Steve Henson and Ben Laurie]
8213 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8214 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8215 are no longer created. This way we have a single and consistent command
8216 line interface `openssl <command>', similar to `cvs <command>'.
8217 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8219 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8220 BIT STRING wrapper always have zero unused bits.
8223 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8226 *) Make the top-level INSTALL documentation easier to understand.
8229 *) Makefiles updated to exit if an error occurs in a sub-directory
8230 make (including if user presses ^C) [Paul Sutton]
8232 *) Make Montgomery context stuff explicit in RSA data structure.
8235 *) Fix build order of pem and err to allow for generated pem.h.
8238 *) Fix renumbering bug in X509_NAME_delete_entry().
8241 *) Enhanced the err-ins.pl script so it makes the error library number
8242 global and can add a library name. This is needed for external ASN1 and
8243 other error libraries.
8246 *) Fixed sk_insert which never worked properly.
8249 *) Fix ASN1 macros so they can handle indefinite length construted
8250 EXPLICIT tags. Some non standard certificates use these: they can now
8254 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8255 into a single doc/ssleay.txt bundle. This way the information is still
8256 preserved but no longer messes up this directory. Now it's new room for
8257 the new set of documenation files.
8258 [Ralf S. Engelschall]
8260 *) SETs were incorrectly DER encoded. This was a major pain, because they
8261 shared code with SEQUENCEs, which aren't coded the same. This means that
8262 almost everything to do with SETs or SEQUENCEs has either changed name or
8263 number of arguments.
8264 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8266 *) Fix test data to work with the above.
8269 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8270 was already fixed by Eric for 0.9.1 it seems.
8271 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8273 *) Autodetect FreeBSD3.
8276 *) Fix various bugs in Configure. This affects the following platforms:
8280 unixware-2.0-pentium
8284 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8285 before they are needed.
8288 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8292 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8294 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8295 changed SSLeay to OpenSSL in version strings.
8296 [Ralf S. Engelschall]
8298 *) Some fixups to the top-level documents.
8301 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8302 because the symlink to include/ was missing.
8303 [Ralf S. Engelschall]
8305 *) Incorporated the popular no-RSA/DSA-only patches
8306 which allow to compile a RSA-free SSLeay.
8307 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8309 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8310 when "ssleay" is still not found.
8311 [Ralf S. Engelschall]
8313 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8314 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8316 *) Updated the README file.
8317 [Ralf S. Engelschall]
8319 *) Added various .cvsignore files in the CVS repository subdirs
8320 to make a "cvs update" really silent.
8321 [Ralf S. Engelschall]
8323 *) Recompiled the error-definition header files and added
8324 missing symbols to the Win32 linker tables.
8325 [Ralf S. Engelschall]
8327 *) Cleaned up the top-level documents;
8328 o new files: CHANGES and LICENSE
8329 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8330 o merged COPYRIGHT into LICENSE
8331 o removed obsolete TODO file
8332 o renamed MICROSOFT to INSTALL.W32
8333 [Ralf S. Engelschall]
8335 *) Removed dummy files from the 0.9.1b source tree:
8336 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8337 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8338 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8339 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8340 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8341 [Ralf S. Engelschall]
8343 *) Added various platform portability fixes.
8346 *) The Genesis of the OpenSSL rpject:
8347 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8348 Young and Tim J. Hudson created while they were working for C2Net until
8350 [The OpenSSL Project]
8353 Changes between 0.9.0b and 0.9.1b [not released]
8355 *) Updated a few CA certificates under certs/
8358 *) Changed some BIGNUM api stuff.
8361 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8362 DGUX x86, Linux Alpha, etc.
8365 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8366 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8370 *) Add -strparse option to asn1pars program which parses nested
8372 [Dr Stephen Henson <shenson@bigfoot.com>]
8374 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8377 *) DSA fix for "ca" program.
8380 *) Added "-genkey" option to "dsaparam" program.
8383 *) Added RIPE MD160 (rmd160) message digest.
8386 *) Added -a (all) option to "ssleay version" command.
8389 *) Added PLATFORM define which is the id given to Configure.
8392 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8395 *) Extended the ASN.1 parser routines.
8398 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8401 *) Added a BN_CTX to the BN library.
8404 *) Fixed the weak key values in DES library
8407 *) Changed API in EVP library for cipher aliases.
8410 *) Added support for RC2/64bit cipher.
8413 *) Converted the lhash library to the crypto/mem.c functions.
8416 *) Added more recognized ASN.1 object ids.
8419 *) Added more RSA padding checks for SSL/TLS.
8422 *) Added BIO proxy/filter functionality.
8425 *) Added extra_certs to SSL_CTX which can be used
8426 send extra CA certificates to the client in the CA cert chain sending
8427 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8430 *) Now Fortezza is denied in the authentication phase because
8431 this is key exchange mechanism is not supported by SSLeay at all.
8434 *) Additional PKCS1 checks.
8437 *) Support the string "TLSv1" for all TLS v1 ciphers.
8440 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8441 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8444 *) Fixed a few memory leaks.
8447 *) Fixed various code and comment typos.
8450 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8451 bytes sent in the client random.
8452 [Edward Bishop <ebishop@spyglass.com>]