5 Changes between 1.0.0f and 1.0.1 [xx XXX xxxx]
7 *) Add support for TLS/DTLS heartbeats.
8 [Robin Seggelmann <seggelmann@fh-muenster.de>]
10 *) Add support for SCTP.
11 [Robin Seggelmann <seggelmann@fh-muenster.de>]
13 *) Improved PRNG seeding for VOS.
14 [Paul Green <Paul.Green@stratus.com>]
16 *) Extensive assembler packs updates, most notably:
18 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
19 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
20 - x86_64: bit-sliced AES implementation;
21 - ARM: NEON support, contemporary platforms optimizations;
22 - s390x: z196 support;
23 - *: GHASH and GF(2^m) multiplication implementations;
27 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
28 (removal of unnecessary code)
29 [Peter Sylvester <peter.sylvester@edelweb.fr>]
31 *) Add TLS key material exporter from RFC 5705.
34 *) Add DTLS-SRTP negotiation from RFC 5764.
37 *) Add Next Protocol Negotiation,
38 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
39 disabled with a no-npn flag to config or Configure. Code donated
41 [Adam Langley <agl@google.com> and Ben Laurie]
43 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
44 NIST-P256, NIST-P521, with constant-time single point multiplication on
45 typical inputs. Compiler support for the nonstandard type __uint128_t is
46 required to use this (present in gcc 4.4 and later, for 64-bit builds).
47 Code made available under Apache License version 2.0.
49 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
50 line to include this in your build of OpenSSL, and run "make depend" (or
51 "make update"). This enables the following EC_METHODs:
53 EC_GFp_nistp224_method()
54 EC_GFp_nistp256_method()
55 EC_GFp_nistp521_method()
57 EC_GROUP_new_by_curve_name() will automatically use these (while
58 EC_GROUP_new_curve_GFp() currently prefers the more flexible
60 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
62 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
63 all platforms. Move ssize_t definition from e_os.h to the public
64 header file e_os2.h as it now appears in public header file cms.h
67 *) New -sigopt option to the ca, req and x509 utilities. Additional
68 signature parameters can be passed using this option and in
72 *) Add RSA PSS signing function. This will generate and set the
73 appropriate AlgorithmIdentifiers for PSS based on those in the
74 corresponding EVP_MD_CTX structure. No application support yet.
77 *) Support for companion algorithm specific ASN1 signing routines.
78 New function ASN1_item_sign_ctx() signs a pre-initialised
79 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
80 the appropriate parameters.
83 *) Add new algorithm specific ASN1 verification initialisation function
84 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
85 handling will be the same no matter what EVP_PKEY_METHOD is used.
86 Add a PSS handler to support verification of PSS signatures: checked
87 against a number of sample certificates.
90 *) Add signature printing for PSS. Add PSS OIDs.
91 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
93 *) Add algorithm specific signature printing. An individual ASN1 method
94 can now print out signatures instead of the standard hex dump.
96 More complex signatures (e.g. PSS) can print out more meaningful
97 information. Include DSA version that prints out the signature
101 *) Password based recipient info support for CMS library: implementing
105 *) Split password based encryption into PBES2 and PBKDF2 functions. This
106 neatly separates the code into cipher and PBE sections and is required
107 for some algorithms that split PBES2 into separate pieces (such as
111 *) Session-handling fixes:
112 - Fix handling of connections that are resuming with a session ID,
113 but also support Session Tickets.
114 - Fix a bug that suppressed issuing of a new ticket if the client
115 presented a ticket with an expired session.
116 - Try to set the ticket lifetime hint to something reasonable.
117 - Make tickets shorter by excluding irrelevant information.
118 - On the client side, don't ignore renewed tickets.
119 [Adam Langley, Bodo Moeller (Google)]
121 *) Fix PSK session representation.
124 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
126 This work was sponsored by Intel.
129 *) Add GCM support to TLS library. Some custom code is needed to split
130 the IV between the fixed (from PRF) and explicit (from TLS record)
131 portions. This adds all GCM ciphersuites supported by RFC5288 and
132 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
133 add a special AESGCM string for GCM only.
136 *) Expand range of ctrls for AES GCM. Permit setting invocation
137 field on decrypt and retrieval of invocation field only on encrypt.
140 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
141 As required by RFC5289 these ciphersuites cannot be used if for
142 versions of TLS earlier than 1.2.
145 *) For FIPS capable OpenSSL interpret a NULL default public key method
146 as unset and return the appopriate default but do *not* set the default.
147 This means we can return the appopriate method in applications that
148 swicth between FIPS and non-FIPS modes.
151 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
152 ENGINE is used then we cannot handle that in the FIPS module so we
153 keep original code iff non-FIPS operations are allowed.
156 *) Add -attime option to openssl utilities.
157 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
159 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
162 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
163 FIPS EC methods unconditionally for now.
166 *) New build option no-ec2m to disable characteristic 2 code.
169 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
170 all cases can be covered as some introduce binary incompatibilities.
173 *) Redirect RSA operations to FIPS module including keygen,
174 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
177 *) Add similar low level API blocking to ciphers.
180 *) Low level digest APIs are not approved in FIPS mode: any attempt
181 to use these will cause a fatal error. Applications that *really* want
182 to use them can use the private_* version instead.
185 *) Redirect cipher operations to FIPS module for FIPS builds.
188 *) Redirect digest operations to FIPS module for FIPS builds.
191 *) Update build system to add "fips" flag which will link in fipscanister.o
192 for static and shared library builds embedding a signature if needed.
195 *) Output TLS supported curves in preference order instead of numerical
196 order. This is currently hardcoded for the highest order curves first.
197 This should be configurable so applications can judge speed vs strength.
200 *) Add TLS v1.2 server support for client authentication.
203 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
207 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
208 FIPS modules versions.
211 *) Add TLS v1.2 client side support for client authentication. Keep cache
212 of handshake records longer as we don't know the hash algorithm to use
213 until after the certificate request message is received.
216 *) Initial TLS v1.2 client support. Add a default signature algorithms
217 extension including all the algorithms we support. Parse new signature
218 format in client key exchange. Relax some ECC signing restrictions for
219 TLS v1.2 as indicated in RFC5246.
222 *) Add server support for TLS v1.2 signature algorithms extension. Switch
223 to new signature format when needed using client digest preference.
224 All server ciphersuites should now work correctly in TLS v1.2. No client
225 support yet and no support for client certificates.
228 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
229 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
230 ciphersuites. At present only RSA key exchange ciphersuites work with
231 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
232 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
233 and version checking.
236 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
237 with this defined it will not be affected by any changes to ssl internal
238 structures. Add several utility functions to allow openssl application
239 to work with OPENSSL_NO_SSL_INTERN defined.
243 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
245 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
248 *) Permit abbreviated handshakes when renegotiating using the function
249 SSL_renegotiate_abbreviated().
250 [Robin Seggelmann <seggelmann@fh-muenster.de>]
252 *) Add call to ENGINE_register_all_complete() to
253 ENGINE_load_builtin_engines(), so some implementations get used
254 automatically instead of needing explicit application support.
257 *) Add support for TLS key exporter as described in RFC5705.
258 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
260 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
261 a few changes are required:
263 Add SSL_OP_NO_TLSv1_1 flag.
265 Update version checking logic to handle version 1.1.
266 Add explicit IV handling (ported from DTLS code).
267 Add command line options to s_client/s_server.
270 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
272 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
273 of the Vaudenay padding oracle attack on CBC mode encryption
274 which enables an efficient plaintext recovery attack against
275 the OpenSSL implementation of DTLS. Their attack exploits timing
276 differences arising during decryption processing. A research
277 paper describing this attack can be found at:
278 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
279 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
280 Security Group at Royal Holloway, University of London
281 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
282 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
283 for preparing the fix. (CVE-2011-4108)
284 [Robin Seggelmann, Michael Tuexen]
286 *) Clear bytes used for block padding of SSL 3.0 records.
288 [Adam Langley (Google)]
290 *) Only allow one SGC handshake restart for SSL/TLS. (CVE-2011-4619)
291 [Adam Langley (Google)]
293 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
294 [Andrey Kulikov <amdeich@gmail.com>]
296 *) Prevent malformed RFC3779 data triggering an assertion failure.
297 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
298 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
299 [Rob Austein <sra@hactrn.net>]
301 *) Improved PRNG seeding for VOS.
302 [Paul Green <Paul.Green@stratus.com>]
304 *) Fix ssl_ciph.c set-up race.
305 [Adam Langley (Google)]
307 *) Fix spurious failures in ecdsatest.c.
308 [Emilia Käsper (Google)]
310 *) Fix the BIO_f_buffer() implementation (which was mixing different
311 interpretations of the '..._len' fields).
312 [Adam Langley (Google)]
314 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
315 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
316 threads won't reuse the same blinding coefficients.
318 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
319 lock to call BN_BLINDING_invert_ex, and avoids one use of
320 BN_BLINDING_update for each BN_BLINDING structure (previously,
321 the last update always remained unused).
322 [Emilia Käsper (Google)]
324 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
325 [Bob Buckholz (Google)]
327 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
329 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
330 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
331 [Kaspar Brand <ossl@velox.ch>]
333 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
334 for multi-threaded use of ECDH. (CVE-2011-3210)
335 [Adam Langley (Google)]
337 *) Fix x509_name_ex_d2i memory leak on bad inputs.
340 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
341 signature public key algorithm by using OID xref utilities instead.
342 Before this you could only use some ECC ciphersuites with SHA1 only.
345 *) Add protection against ECDSA timing attacks as mentioned in the paper
346 by Billy Bob Brumley and Nicola Tuveri, see:
348 http://eprint.iacr.org/2011/232.pdf
350 [Billy Bob Brumley and Nicola Tuveri]
352 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
354 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
355 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
357 *) Fix bug in string printing code: if *any* escaping is enabled we must
358 escape the escape character (backslash) or the resulting string is
362 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
364 *) Disable code workaround for ancient and obsolete Netscape browsers
365 and servers: an attacker can use it in a ciphersuite downgrade attack.
366 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
369 *) Fixed J-PAKE implementation error, originally discovered by
370 Sebastien Martini, further info and confirmation from Stefan
371 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
374 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
376 *) Fix extension code to avoid race conditions which can result in a buffer
377 overrun vulnerability: resumed sessions must not be modified as they can
378 be shared by multiple threads. CVE-2010-3864
381 *) Fix WIN32 build system to correctly link an ENGINE directory into
385 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
387 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
389 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
391 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
393 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
394 context. The operation can be customised via the ctrl mechanism in
395 case ENGINEs want to include additional functionality.
398 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
401 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
402 output hashes compatible with older versions of OpenSSL.
403 [Willy Weisz <weisz@vcpc.univie.ac.at>]
405 *) Fix compression algorithm handling: if resuming a session use the
406 compression algorithm of the resumed session instead of determining
407 it from client hello again. Don't allow server to change algorithm.
410 *) Add load_crls() function to apps tidying load_certs() too. Add option
411 to verify utility to allow additional CRLs to be included.
414 *) Update OCSP request code to permit adding custom headers to the request:
415 some responders need this.
418 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
420 [Julia Lawall <julia@diku.dk>]
422 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
423 needlessly dereferenced structures, used obsolete functions and
424 didn't handle all updated verify codes correctly.
427 *) Disable MD2 in the default configuration.
430 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
431 indicate the initial BIO being pushed or popped. This makes it possible
432 to determine whether the BIO is the one explicitly called or as a result
433 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
434 it handles reference counts correctly and doesn't zero out the I/O bio
435 when it is not being explicitly popped. WARNING: applications which
436 included workarounds for the old buggy behaviour will need to be modified
437 or they could free up already freed BIOs.
440 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
441 renaming to all platforms (within the 0.9.8 branch, this was
442 done conditionally on Netware platforms to avoid a name clash).
443 [Guenter <lists@gknw.net>]
445 *) Add ECDHE and PSK support to DTLS.
446 [Michael Tuexen <tuexen@fh-muenster.de>]
448 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
452 *) Add "missing" function EVP_MD_flags() (without this the only way to
453 retrieve a digest flags is by accessing the structure directly. Update
454 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
455 or cipher is registered as in the "from" argument. Print out all
456 registered digests in the dgst usage message instead of manually
457 attempting to work them out.
460 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
461 this allows the use of compression and extensions. Change default cipher
462 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
463 by default unless an application cipher string requests it.
466 *) Alter match criteria in PKCS12_parse(). It used to try to use local
467 key ids to find matching certificates and keys but some PKCS#12 files
468 don't follow the (somewhat unwritten) rules and this strategy fails.
469 Now just gather all certificates together and the first private key
470 then look for the first certificate that matches the key.
473 *) Support use of registered digest and cipher names for dgst and cipher
474 commands instead of having to add each one as a special case. So now
481 openssl dgst -sha256 foo
483 and this works for ENGINE based algorithms too.
487 *) Update Gost ENGINE to support parameter files.
488 [Victor B. Wagner <vitus@cryptocom.ru>]
490 *) Support GeneralizedTime in ca utility.
491 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
493 *) Enhance the hash format used for certificate directory links. The new
494 form uses the canonical encoding (meaning equivalent names will work
495 even if they aren't identical) and uses SHA1 instead of MD5. This form
496 is incompatible with the older format and as a result c_rehash should
497 be used to rebuild symbolic links.
500 *) Make PKCS#8 the default write format for private keys, replacing the
501 traditional format. This form is standardised, more secure and doesn't
502 include an implicit MD5 dependency.
505 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
506 committed to OpenSSL should pass this lot as a minimum.
509 *) Add session ticket override functionality for use by EAP-FAST.
510 [Jouni Malinen <j@w1.fi>]
512 *) Modify HMAC functions to return a value. Since these can be implemented
513 in an ENGINE errors can occur.
516 *) Type-checked OBJ_bsearch_ex.
519 *) Type-checked OBJ_bsearch. Also some constification necessitated
520 by type-checking. Still to come: TXT_DB, bsearch(?),
521 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
525 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
526 seconds to a tm structure directly, instead of going through OS
527 specific date routines. This avoids any issues with OS routines such
528 as the year 2038 bug. New *_adj() functions for ASN1 time structures
529 and X509_time_adj_ex() to cover the extended range. The existing
530 X509_time_adj() is still usable and will no longer have any date issues.
533 *) Delta CRL support. New use deltas option which will attempt to locate
534 and search any appropriate delta CRLs available.
536 This work was sponsored by Google.
539 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
540 code and add additional score elements. Validate alternate CRL paths
541 as part of the CRL checking and indicate a new error "CRL path validation
542 error" in this case. Applications wanting additional details can use
543 the verify callback and check the new "parent" field. If this is not
544 NULL CRL path validation is taking place. Existing applications wont
545 see this because it requires extended CRL support which is off by
548 This work was sponsored by Google.
551 *) Support for freshest CRL extension.
553 This work was sponsored by Google.
556 *) Initial indirect CRL support. Currently only supported in the CRLs
557 passed directly and not via lookup. Process certificate issuer
558 CRL entry extension and lookup CRL entries by bother issuer name
559 and serial number. Check and process CRL issuer entry in IDP extension.
561 This work was sponsored by Google.
564 *) Add support for distinct certificate and CRL paths. The CRL issuer
565 certificate is validated separately in this case. Only enabled if
566 an extended CRL support flag is set: this flag will enable additional
567 CRL functionality in future.
569 This work was sponsored by Google.
572 *) Add support for policy mappings extension.
574 This work was sponsored by Google.
577 *) Fixes to pathlength constraint, self issued certificate handling,
578 policy processing to align with RFC3280 and PKITS tests.
580 This work was sponsored by Google.
583 *) Support for name constraints certificate extension. DN, email, DNS
584 and URI types are currently supported.
586 This work was sponsored by Google.
589 *) To cater for systems that provide a pointer-based thread ID rather
590 than numeric, deprecate the current numeric thread ID mechanism and
591 replace it with a structure and associated callback type. This
592 mechanism allows a numeric "hash" to be extracted from a thread ID in
593 either case, and on platforms where pointers are larger than 'long',
594 mixing is done to help ensure the numeric 'hash' is usable even if it
595 can't be guaranteed unique. The default mechanism is to use "&errno"
596 as a pointer-based thread ID to distinguish between threads.
598 Applications that want to provide their own thread IDs should now use
599 CRYPTO_THREADID_set_callback() to register a callback that will call
600 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
602 Note that ERR_remove_state() is now deprecated, because it is tied
603 to the assumption that thread IDs are numeric. ERR_remove_state(0)
604 to free the current thread's error state should be replaced by
605 ERR_remove_thread_state(NULL).
607 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
608 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
609 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
610 application was previously providing a numeric thread callback that
611 was inappropriate for distinguishing threads, then uniqueness might
612 have been obtained with &errno that happened immediately in the
613 intermediate development versions of OpenSSL; this is no longer the
614 case, the numeric thread callback will now override the automatic use
616 [Geoff Thorpe, with help from Bodo Moeller]
618 *) Initial support for different CRL issuing certificates. This covers a
619 simple case where the self issued certificates in the chain exist and
620 the real CRL issuer is higher in the existing chain.
622 This work was sponsored by Google.
625 *) Removed effectively defunct crypto/store from the build.
628 *) Revamp of STACK to provide stronger type-checking. Still to come:
629 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
630 ASN1_STRING, CONF_VALUE.
633 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
634 RAM on SSL connections. This option can save about 34k per idle SSL.
637 *) Revamp of LHASH to provide stronger type-checking. Still to come:
638 STACK, TXT_DB, bsearch, qsort.
641 *) Initial support for Cryptographic Message Syntax (aka CMS) based
642 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
643 support for data, signedData, compressedData, digestedData and
644 encryptedData, envelopedData types included. Scripts to check against
645 RFC4134 examples draft and interop and consistency checks of many
646 content types and variants.
649 *) Add options to enc utility to support use of zlib compression BIO.
652 *) Extend mk1mf to support importing of options and assembly language
653 files from Configure script, currently only included in VC-WIN32.
654 The assembly language rules can now optionally generate the source
655 files from the associated perl scripts.
658 *) Implement remaining functionality needed to support GOST ciphersuites.
659 Interop testing has been performed using CryptoPro implementations.
660 [Victor B. Wagner <vitus@cryptocom.ru>]
662 *) s390x assembler pack.
665 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
669 *) Implement Opaque PRF Input TLS extension as specified in
670 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
671 official specification yet and no extension type assignment by
672 IANA exists, this extension (for now) will have to be explicitly
673 enabled when building OpenSSL by providing the extension number
674 to use. For example, specify an option
676 -DTLSEXT_TYPE_opaque_prf_input=0x9527
678 to the "config" or "Configure" script to enable the extension,
679 assuming extension number 0x9527 (which is a completely arbitrary
680 and unofficial assignment based on the MD5 hash of the Internet
681 Draft). Note that by doing so, you potentially lose
682 interoperability with other TLS implementations since these might
683 be using the same extension number for other purposes.
685 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
686 opaque PRF input value to use in the handshake. This will create
687 an interal copy of the length-'len' string at 'src', and will
688 return non-zero for success.
690 To get more control and flexibility, provide a callback function
693 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
694 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
698 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
701 Callback function 'cb' will be called in handshakes, and is
702 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
703 Argument 'arg' is for application purposes (the value as given to
704 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
705 be provided to the callback function). The callback function
706 has to return non-zero to report success: usually 1 to use opaque
707 PRF input just if possible, or 2 to enforce use of the opaque PRF
708 input. In the latter case, the library will abort the handshake
709 if opaque PRF input is not successfully negotiated.
711 Arguments 'peerinput' and 'len' given to the callback function
712 will always be NULL and 0 in the case of a client. A server will
713 see the client's opaque PRF input through these variables if
714 available (NULL and 0 otherwise). Note that if the server
715 provides an opaque PRF input, the length must be the same as the
716 length of the client's opaque PRF input.
718 Note that the callback function will only be called when creating
719 a new session (session resumption can resume whatever was
720 previously negotiated), and will not be called in SSL 2.0
721 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
722 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
723 for applications that need to enforce opaque PRF input.
727 *) Update ssl code to support digests other than SHA1+MD5 for handshake
730 [Victor B. Wagner <vitus@cryptocom.ru>]
732 *) Add RFC4507 support to OpenSSL. This includes the corrections in
733 RFC4507bis. The encrypted ticket format is an encrypted encoded
734 SSL_SESSION structure, that way new session features are automatically
737 If a client application caches session in an SSL_SESSION structure
738 support is transparent because tickets are now stored in the encoded
741 The SSL_CTX structure automatically generates keys for ticket
742 protection in servers so again support should be possible
743 with no application modification.
745 If a client or server wishes to disable RFC4507 support then the option
746 SSL_OP_NO_TICKET can be set.
748 Add a TLS extension debugging callback to allow the contents of any client
749 or server extensions to be examined.
751 This work was sponsored by Google.
754 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
755 OpenSSL should now compile cleanly on gcc 4.2
756 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
758 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
759 support including streaming MAC support: this is required for GOST
761 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
763 *) Add option -stream to use PKCS#7 streaming in smime utility. New
764 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
765 to output in BER and PEM format.
768 *) Experimental support for use of HMAC via EVP_PKEY interface. This
769 allows HMAC to be handled via the EVP_DigestSign*() interface. The
770 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
771 ENGINE support for HMAC keys which are unextractable. New -mac and
772 -macopt options to dgst utility.
775 *) New option -sigopt to dgst utility. Update dgst to use
776 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
777 alternative signing paramaters such as X9.31 or PSS in the dgst
781 *) Change ssl_cipher_apply_rule(), the internal function that does
782 the work each time a ciphersuite string requests enabling
783 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
784 removing ("!foo+bar") a class of ciphersuites: Now it maintains
785 the order of disabled ciphersuites such that those ciphersuites
786 that most recently went from enabled to disabled not only stay
787 in order with respect to each other, but also have higher priority
788 than other disabled ciphersuites the next time ciphersuites are
791 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
792 the same ciphersuites as with "HIGH" alone, but in a specific
793 order where the PSK ciphersuites come first (since they are the
794 most recently disabled ciphersuites when "HIGH" is parsed).
796 Also, change ssl_create_cipher_list() (using this new
797 funcionality) such that between otherwise identical
798 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
802 *) Change ssl_create_cipher_list() so that it automatically
803 arranges the ciphersuites in reasonable order before starting
804 to process the rule string. Thus, the definition for "DEFAULT"
805 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
806 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
807 This makes it much easier to arrive at a reasonable default order
808 in applications for which anonymous ciphers are OK (meaning
809 that you can't actually use DEFAULT).
810 [Bodo Moeller; suggested by Victor Duchovni]
812 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
813 processing) into multiple integers instead of setting
814 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
815 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
816 (These masks as well as the individual bit definitions are hidden
817 away into the non-exported interface ssl/ssl_locl.h, so this
818 change to the definition of the SSL_CIPHER structure shouldn't
819 affect applications.) This give us more bits for each of these
820 categories, so there is no longer a need to coagulate AES128 and
821 AES256 into a single algorithm bit, and to coagulate Camellia128
822 and Camellia256 into a single algorithm bit, which has led to all
825 Thus, among other things, the kludge introduced in 0.9.7m and
826 0.9.8e for masking out AES256 independently of AES128 or masking
827 out Camellia256 independently of AES256 is not needed here in 0.9.9.
829 With the change, we also introduce new ciphersuite aliases that
830 so far were missing: "AES128", "AES256", "CAMELLIA128", and
834 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
835 Use the leftmost N bytes of the signature input if the input is
836 larger than the prime q (with N being the size in bytes of q).
839 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
840 it yet and it is largely untested.
843 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
846 *) Initial incomplete changes to avoid need for function casts in OpenSSL
847 some compilers (gcc 4.2 and later) reject their use. Safestack is
848 reimplemented. Update ASN1 to avoid use of legacy functions.
851 *) Win32/64 targets are linked with Winsock2.
854 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
855 to external functions. This can be used to increase CRL handling
856 efficiency especially when CRLs are very large by (for example) storing
857 the CRL revoked certificates in a database.
860 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
861 new CRLs added to a directory can be used. New command line option
862 -verify_return_error to s_client and s_server. This causes real errors
863 to be returned by the verify callback instead of carrying on no matter
864 what. This reflects the way a "real world" verify callback would behave.
867 *) GOST engine, supporting several GOST algorithms and public key formats.
868 Kindly donated by Cryptocom.
871 *) Partial support for Issuing Distribution Point CRL extension. CRLs
872 partitioned by DP are handled but no indirect CRL or reason partitioning
873 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
874 selected via a scoring technique which handles IDP and AKID in CRLs.
877 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
878 will ultimately be used for all verify operations: this will remove the
879 X509_STORE dependency on certificate verification and allow alternative
880 lookup methods. X509_STORE based implementations of these two callbacks.
883 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
884 Modify get_crl() to find a valid (unexpired) CRL if possible.
887 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
888 this would be called X509_CRL_cmp() but that name is already used by
889 a function that just compares CRL issuer names. Cache several CRL
890 extensions in X509_CRL structure and cache CRLDP in X509.
893 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
894 this maps equivalent X509_NAME structures into a consistent structure.
895 Name comparison can then be performed rapidly using memcmp().
898 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
902 *) Allow digests to supply their own micalg string for S/MIME type using
903 the ctrl EVP_MD_CTRL_MICALG.
906 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
907 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
908 ctrl. It can then customise the structure before and/or after signing
912 *) New function OBJ_add_sigid() to allow application defined signature OIDs
913 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
914 to free up any added signature OIDs.
917 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
918 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
919 digest and cipher tables. New options added to openssl utility:
920 list-message-digest-algorithms and list-cipher-algorithms.
923 *) Change the array representation of binary polynomials: the list
924 of degrees of non-zero coefficients is now terminated with -1.
925 Previously it was terminated with 0, which was also part of the
926 value; thus, the array representation was not applicable to
927 polynomials where t^0 has coefficient zero. This change makes
928 the array representation useful in a more general context.
931 *) Various modifications and fixes to SSL/TLS cipher string
932 handling. For ECC, the code now distinguishes between fixed ECDH
933 with RSA certificates on the one hand and with ECDSA certificates
934 on the other hand, since these are separate ciphersuites. The
935 unused code for Fortezza ciphersuites has been removed.
937 For consistency with EDH, ephemeral ECDH is now called "EECDH"
938 (not "ECDHE"). For consistency with the code for DH
939 certificates, use of ECDH certificates is now considered ECDH
940 authentication, not RSA or ECDSA authentication (the latter is
941 merely the CA's signing algorithm and not actively used in the
944 The temporary ciphersuite alias "ECCdraft" is no longer
945 available, and ECC ciphersuites are no longer excluded from "ALL"
946 and "DEFAULT". The following aliases now exist for RFC 4492
947 ciphersuites, most of these by analogy with the DH case:
949 kECDHr - ECDH cert, signed with RSA
950 kECDHe - ECDH cert, signed with ECDSA
951 kECDH - ECDH cert (signed with either RSA or ECDSA)
952 kEECDH - ephemeral ECDH
953 ECDH - ECDH cert or ephemeral ECDH
959 AECDH - anonymous ECDH
960 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
964 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
965 Use correct micalg parameters depending on digest(s) in signed message.
968 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
969 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
972 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
973 an engine to register a method. Add ENGINE lookups for methods and
974 functional reference processing.
977 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
978 EVP_{Sign,Verify}* which allow an application to customise the signature
982 *) New -resign option to smime utility. This adds one or more signers
983 to an existing PKCS#7 signedData structure. Also -md option to use an
984 alternative message digest algorithm for signing.
987 *) Tidy up PKCS#7 routines and add new functions to make it easier to
988 create PKCS7 structures containing multiple signers. Update smime
989 application to support multiple signers.
992 *) New -macalg option to pkcs12 utility to allow setting of an alternative
996 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
997 Reorganize PBE internals to lookup from a static table using NIDs,
998 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
999 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1000 PRF which will be automatically used with PBES2.
1003 *) Replace the algorithm specific calls to generate keys in "req" with the
1007 *) Update PKCS#7 enveloped data routines to use new API. This is now
1008 supported by any public key method supporting the encrypt operation. A
1009 ctrl is added to allow the public key algorithm to examine or modify
1010 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1014 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1015 a default digest type to use. In most cases this will be SHA1 but some
1016 algorithms (such as GOST) need to specify an alternative digest. The
1017 return value indicates how strong the prefernce is 1 means optional and
1018 2 is mandatory (that is it is the only supported type). Modify
1019 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1020 use the default md. Update openssl utilities to use the default digest
1021 type for signing if it is not explicitly indicated.
1024 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1025 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1026 signing method from the key type. This effectively removes the link
1027 between digests and public key types.
1030 *) Add an OID cross reference table and utility functions. Its purpose is to
1031 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1032 rsaEncryption. This will allow some of the algorithm specific hackery
1033 needed to use the correct OID to be removed.
1036 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1037 structures for PKCS7_sign(). They are now set up by the relevant public
1041 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1044 *) Add support for key derivation (agreement) in the API, DH method and
1048 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1049 public and private key formats. As a side effect these add additional
1050 command line functionality not previously available: DSA signatures can be
1051 generated and verified using pkeyutl and DH key support and generation in
1056 [Oliver Tappe <zooey@hirschkaefer.de>]
1058 *) New make target "install_html_docs" installs HTML renditions of the
1060 [Oliver Tappe <zooey@hirschkaefer.de>]
1062 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1063 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1064 support key and parameter generation and add initial key generation
1065 functionality for RSA.
1068 *) Add functions for main EVP_PKEY_method operations. The undocumented
1069 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1070 EVP_PKEY_{encrypt,decrypt}_old.
1073 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1074 key API, doesn't do much yet.
1077 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1078 public key algorithms. New option to openssl utility:
1079 "list-public-key-algorithms" to print out info.
1082 *) Implement the Supported Elliptic Curves Extension for
1083 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1086 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1087 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1090 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1091 utilities such as rsa, dsa, dsaparam etc except they process any key
1095 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1096 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1097 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1101 *) Initial support for pluggable public key ASN1.
1102 De-spaghettify the public key ASN1 handling. Move public and private
1103 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1104 algorithm specific handling to a single module within the relevant
1105 algorithm directory. Add functions to allow (near) opaque processing
1106 of public and private key structures.
1109 *) Implement the Supported Point Formats Extension for
1110 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1113 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1114 for the psk identity [hint] and the psk callback functions to the
1115 SSL_SESSION, SSL and SSL_CTX structure.
1118 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1122 SSL_CTX_use_psk_identity_hint
1123 SSL_get_psk_identity_hint
1124 SSL_get_psk_identity
1125 SSL_use_psk_identity_hint
1127 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1129 *) Add RFC 3161 compliant time stamp request creation, response generation
1130 and response verification functionality.
1131 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1133 *) Add initial support for TLS extensions, specifically for the server_name
1134 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1135 have new members for a host name. The SSL data structure has an
1136 additional member SSL_CTX *initial_ctx so that new sessions can be
1137 stored in that context to allow for session resumption, even after the
1138 SSL has been switched to a new SSL_CTX in reaction to a client's
1139 server_name extension.
1141 New functions (subject to change):
1143 SSL_get_servername()
1144 SSL_get_servername_type()
1147 New CTRL codes and macros (subject to change):
1149 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1150 - SSL_CTX_set_tlsext_servername_callback()
1151 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1152 - SSL_CTX_set_tlsext_servername_arg()
1153 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1155 openssl s_client has a new '-servername ...' option.
1157 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1158 '-key2 ...', '-servername_fatal' (subject to change). This allows
1159 testing the HostName extension for a specific single host name ('-cert'
1160 and '-key' remain fallbacks for handshakes without HostName
1161 negotiation). If the unrecogninzed_name alert has to be sent, this by
1162 default is a warning; it becomes fatal with the '-servername_fatal'
1165 [Peter Sylvester, Remy Allais, Christophe Renou]
1167 *) Whirlpool hash implementation is added.
1170 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1171 bn(64,32). Because of instruction set limitations it doesn't have
1172 any negative impact on performance. This was done mostly in order
1173 to make it possible to share assembler modules, such as bn_mul_mont
1174 implementations, between 32- and 64-bit builds without hassle.
1177 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1178 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1182 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1183 dedicated Montgomery multiplication procedure, is introduced.
1184 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1185 "64-bit" performance on certain 32-bit targets.
1188 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1189 in SSL structures. New SSL ctrl to set maximum send fragment size.
1190 Save memory by seeting the I/O buffer sizes dynamically instead of
1191 using the maximum available value.
1194 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1195 in addition to the text details.
1198 *) Very, very preliminary EXPERIMENTAL support for printing of general
1199 ASN1 structures. This currently produces rather ugly output and doesn't
1200 handle several customised structures at all.
1203 *) Integrated support for PVK file format and some related formats such
1204 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1205 these in the 'rsa' and 'dsa' utilities.
1208 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1211 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1212 place for the (very old) "NETSCAPE" format certificates which are now
1213 handled using new ASN1 code equivalents.
1216 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1217 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1218 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1221 *) Modify CRL distribution points extension code to print out previously
1222 unsupported fields. Enhance extension setting code to allow setting of
1226 *) Add print and set support for Issuing Distribution Point CRL extension.
1229 *) Change 'Configure' script to enable Camellia by default.
1232 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1234 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1235 of the Vaudenay padding oracle attack on CBC mode encryption
1236 which enables an efficient plaintext recovery attack against
1237 the OpenSSL implementation of DTLS. Their attack exploits timing
1238 differences arising during decryption processing. A research
1239 paper describing this attack can be found at:
1240 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1241 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1242 Security Group at Royal Holloway, University of London
1243 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1244 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1245 for preparing the fix. (CVE-2011-4108)
1246 [Robin Seggelmann, Michael Tuexen]
1248 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1249 [Ben Laurie, Kasper <ekasper@google.com>]
1251 *) Clear bytes used for block padding of SSL 3.0 records.
1253 [Adam Langley (Google)]
1255 *) Only allow one SGC handshake restart for SSL/TLS. (CVE-2011-4619)
1256 [Adam Langley (Google)]
1258 *) Prevent malformed RFC3779 data triggering an assertion failure.
1259 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1260 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1261 [Rob Austein <sra@hactrn.net>]
1263 *) Fix ssl_ciph.c set-up race.
1264 [Adam Langley (Google)]
1266 *) Fix spurious failures in ecdsatest.c.
1267 [Emilia Käsper (Google)]
1269 *) Fix the BIO_f_buffer() implementation (which was mixing different
1270 interpretations of the '..._len' fields).
1271 [Adam Langley (Google)]
1273 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1274 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1275 threads won't reuse the same blinding coefficients.
1277 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1278 lock to call BN_BLINDING_invert_ex, and avoids one use of
1279 BN_BLINDING_update for each BN_BLINDING structure (previously,
1280 the last update always remained unused).
1281 [Emilia Käsper (Google)]
1283 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1284 for multi-threaded use of ECDH.
1285 [Adam Langley (Google)]
1287 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1290 *) Add protection against ECDSA timing attacks as mentioned in the paper
1291 by Billy Bob Brumley and Nicola Tuveri, see:
1293 http://eprint.iacr.org/2011/232.pdf
1295 [Billy Bob Brumley and Nicola Tuveri]
1297 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1299 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1300 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1302 *) Fix bug in string printing code: if *any* escaping is enabled we must
1303 escape the escape character (backslash) or the resulting string is
1307 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1309 *) Disable code workaround for ancient and obsolete Netscape browsers
1310 and servers: an attacker can use it in a ciphersuite downgrade attack.
1311 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1314 *) Fixed J-PAKE implementation error, originally discovered by
1315 Sebastien Martini, further info and confirmation from Stefan
1316 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1319 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1321 *) Fix extension code to avoid race conditions which can result in a buffer
1322 overrun vulnerability: resumed sessions must not be modified as they can
1323 be shared by multiple threads. CVE-2010-3864
1326 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1329 *) Don't reencode certificate when calculating signature: cache and use
1330 the original encoding instead. This makes signature verification of
1331 some broken encodings work correctly.
1334 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1335 is also one of the inputs.
1336 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1338 *) Don't repeatedly append PBE algorithms to table if they already exist.
1339 Sort table on each new add. This effectively makes the table read only
1340 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1344 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1346 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1349 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1350 access or freeing data twice (CVE-2010-0742)
1351 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1353 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1354 common in certificates and some applications which only call
1355 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1359 Reduce copying into .apps and .test in makevms.com
1360 Don't try to use blank CA certificate in CA.com
1361 Allow use of C files from original directories in maketests.com
1362 [Steven M. Schweda" <sms@antinode.info>]
1364 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1366 *) When rejecting SSL/TLS records due to an incorrect version number, never
1367 update s->server with a new major version number. As of
1368 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1369 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1370 the previous behavior could result in a read attempt at NULL when
1371 receiving specific incorrect SSL/TLS records once record payload
1372 protection is active. (CVE-2010-0740)
1373 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1375 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1376 could be crashed if the relevant tables were not present (e.g. chrooted).
1377 [Tomas Hoger <thoger@redhat.com>]
1379 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1381 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1382 [Martin Olsson, Neel Mehta]
1384 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1385 accommodate for stack sorting, always a write lock!).
1388 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1389 excessive delays in the RAND_poll(): over a minute. As a workaround
1390 include a time check in the inner Heap32Next loop too.
1393 *) The code that handled flushing of data in SSL/TLS originally used the
1394 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1395 the problem outlined in PR#1949. The fix suggested there however can
1396 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1397 of Apache). So instead simplify the code to flush unconditionally.
1398 This should be fine since flushing with no data to flush is a no op.
1401 *) Handle TLS versions 2.0 and later properly and correctly use the
1402 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1403 off ancient servers have a habit of sticking around for a while...
1406 *) Modify compression code so it frees up structures without using the
1407 ex_data callbacks. This works around a problem where some applications
1408 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1409 restarting) then use compression (e.g. SSL with compression) later.
1410 This results in significant per-connection memory leaks and
1411 has caused some security issues including CVE-2008-1678 and
1415 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1416 change when encrypting or decrypting.
1419 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1420 connect and renegotiate with servers which do not support RI.
1421 Until RI is more widely deployed this option is enabled by default.
1424 *) Add "missing" ssl ctrls to clear options and mode.
1427 *) If client attempts to renegotiate and doesn't support RI respond with
1428 a no_renegotiation alert as required by RFC5746. Some renegotiating
1429 TLS clients will continue a connection gracefully when they receive
1430 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1431 waiting for a server hello which it will never receive. Now we treat a
1432 received no_renegotiation alert as a fatal error. This is because
1433 applications requesting a renegotiation might well expect it to succeed
1434 and would have no code in place to handle the server denying it so the
1435 only safe thing to do is to terminate the connection.
1438 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1439 peer supports secure renegotiation and 0 otherwise. Print out peer
1440 renegotiation support in s_client/s_server.
1443 *) Replace the highly broken and deprecated SPKAC certification method with
1444 the updated NID creation version. This should correctly handle UTF8.
1447 *) Implement RFC5746. Re-enable renegotiation but require the extension
1448 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1449 turns out to be a bad idea. It has been replaced by
1450 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1451 SSL_CTX_set_options(). This is really not recommended unless you
1452 know what you are doing.
1453 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1455 *) Fixes to stateless session resumption handling. Use initial_ctx when
1456 issuing and attempting to decrypt tickets in case it has changed during
1457 servername handling. Use a non-zero length session ID when attempting
1458 stateless session resumption: this makes it possible to determine if
1459 a resumption has occurred immediately after receiving server hello
1460 (several places in OpenSSL subtly assume this) instead of later in
1464 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1465 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1466 fixes for a few places where the return code is not checked
1468 [Julia Lawall <julia@diku.dk>]
1470 *) Add --strict-warnings option to Configure script to include devteam
1471 warnings in other configurations.
1474 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1475 makes it possible to install openssl libraries in locations which
1476 have names other than "lib", for example "/usr/lib64" which some
1478 [Steve Henson, based on patch from Jeremy Utley]
1480 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1481 X690 8.9.12 and can produce some misleading textual output of OIDs.
1482 [Steve Henson, reported by Dan Kaminsky]
1484 *) Delete MD2 from algorithm tables. This follows the recommendation in
1485 several standards that it is not used in new applications due to
1486 several cryptographic weaknesses. For binary compatibility reasons
1487 the MD2 API is still compiled in by default.
1490 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1494 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1495 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1497 [Guenter <lists@gknw.net>]
1499 *) Fix the server certificate chain building code to use X509_verify_cert(),
1500 it used to have an ad-hoc builder which was unable to cope with anything
1501 other than a simple chain.
1502 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1504 *) Don't check self signed certificate signatures in X509_verify_cert()
1505 by default (a flag can override this): it just wastes time without
1506 adding any security. As a useful side effect self signed root CAs
1507 with non-FIPS digests are now usable in FIPS mode.
1510 *) In dtls1_process_out_of_seq_message() the check if the current message
1511 is already buffered was missing. For every new message was memory
1512 allocated, allowing an attacker to perform an denial of service attack
1513 with sending out of seq handshake messages until there is no memory
1514 left. Additionally every future messege was buffered, even if the
1515 sequence number made no sense and would be part of another handshake.
1516 So only messages with sequence numbers less than 10 in advance will be
1517 buffered. (CVE-2009-1378)
1518 [Robin Seggelmann, discovered by Daniel Mentz]
1520 *) Records are buffered if they arrive with a future epoch to be
1521 processed after finishing the corresponding handshake. There is
1522 currently no limitation to this buffer allowing an attacker to perform
1523 a DOS attack with sending records with future epochs until there is no
1524 memory left. This patch adds the pqueue_size() function to detemine
1525 the size of a buffer and limits the record buffer to 100 entries.
1527 [Robin Seggelmann, discovered by Daniel Mentz]
1529 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1530 parent structure is freed. (CVE-2009-1379)
1533 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1534 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1537 [Ilya O. <vrghost@gmail.com>]
1539 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1541 *) Disable renegotiation completely - this fixes a severe security
1542 problem (CVE-2009-3555) at the cost of breaking all
1543 renegotiation. Renegotiation can be re-enabled by setting
1544 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1545 run-time. This is really not recommended unless you know what
1549 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1551 *) Don't set val to NULL when freeing up structures, it is freed up by
1552 underlying code. If sizeof(void *) > sizeof(long) this can result in
1553 zeroing past the valid field. (CVE-2009-0789)
1554 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1556 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1557 checked correctly. This would allow some invalid signed attributes to
1558 appear to verify correctly. (CVE-2009-0591)
1559 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1561 *) Reject UniversalString and BMPString types with invalid lengths. This
1562 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1563 a legal length. (CVE-2009-0590)
1566 *) Set S/MIME signing as the default purpose rather than setting it
1567 unconditionally. This allows applications to override it at the store
1571 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1572 to handle some structures.
1575 *) Improve efficiency of mem_gets: don't search whole buffer each time
1577 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1579 *) New -hex option for openssl rand.
1582 *) Print out UTF8String and NumericString when parsing ASN1.
1585 *) Support NumericString type for name components.
1588 *) Allow CC in the environment to override the automatically chosen
1589 compiler. Note that nothing is done to ensure flags work with the
1593 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1595 *) Properly check EVP_VerifyFinal() and similar return values
1597 [Ben Laurie, Bodo Moeller, Google Security Team]
1599 *) Enable TLS extensions by default.
1602 *) Allow the CHIL engine to be loaded, whether the application is
1603 multithreaded or not. (This does not release the developer from the
1604 obligation to set up the dynamic locking callbacks.)
1605 [Sander Temme <sander@temme.net>]
1607 *) Use correct exit code if there is an error in dgst command.
1608 [Steve Henson; problem pointed out by Roland Dirlewanger]
1610 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1611 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1614 *) Add experimental JPAKE support, including demo authentication in
1615 s_client and s_server.
1618 *) Set the comparison function in v3_addr_canonize().
1619 [Rob Austein <sra@hactrn.net>]
1621 *) Add support for XMPP STARTTLS in s_client.
1622 [Philip Paeps <philip@freebsd.org>]
1624 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1625 to ensure that even with this option, only ciphersuites in the
1626 server's preference list will be accepted. (Note that the option
1627 applies only when resuming a session, so the earlier behavior was
1628 just about the algorithm choice for symmetric cryptography.)
1631 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1633 *) Fix NULL pointer dereference if a DTLS server received
1634 ChangeCipherSpec as first record (CVE-2009-1386).
1637 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1638 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1641 *) The fix in 0.9.8c that supposedly got rid of unsafe
1642 double-checked locking was incomplete for RSA blinding,
1643 addressing just one layer of what turns out to have been
1644 doubly unsafe triple-checked locking.
1646 So now fix this for real by retiring the MONT_HELPER macro
1647 in crypto/rsa/rsa_eay.c.
1649 [Bodo Moeller; problem pointed out by Marius Schilder]
1651 *) Various precautionary measures:
1653 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1655 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1656 (NB: This would require knowledge of the secret session ticket key
1657 to exploit, in which case you'd be SOL either way.)
1659 - Change bn_nist.c so that it will properly handle input BIGNUMs
1660 outside the expected range.
1662 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1665 [Neel Mehta, Bodo Moeller]
1667 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1668 the load fails. Useful for distros.
1669 [Ben Laurie and the FreeBSD team]
1671 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1674 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1677 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1679 This work was sponsored by Logica.
1682 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1683 keystores. Support for SSL/TLS client authentication too.
1684 Not compiled unless enable-capieng specified to Configure.
1686 This work was sponsored by Logica.
1689 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1690 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1691 attribute creation routines such as certifcate requests and PKCS#12
1695 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1697 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1698 handshake which could lead to a cilent crash as found using the
1699 Codenomicon TLS test suite (CVE-2008-1672)
1700 [Steve Henson, Mark Cox]
1702 *) Fix double free in TLS server name extensions which could lead to
1703 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1706 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1708 Clear the error queue to ensure that error entries left from
1709 older function calls do not interfere with the correct operation.
1710 [Lutz Jaenicke, Erik de Castro Lopo]
1712 *) Remove root CA certificates of commercial CAs:
1714 The OpenSSL project does not recommend any specific CA and does not
1715 have any policy with respect to including or excluding any CA.
1716 Therefore it does not make any sense to ship an arbitrary selection
1717 of root CA certificates with the OpenSSL software.
1720 *) RSA OAEP patches to fix two separate invalid memory reads.
1721 The first one involves inputs when 'lzero' is greater than
1722 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1723 before the beginning of from). The second one involves inputs where
1724 the 'db' section contains nothing but zeroes (there is a one-byte
1725 invalid read after the end of 'db').
1726 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1728 *) Partial backport from 0.9.9-dev:
1730 Introduce bn_mul_mont (dedicated Montgomery multiplication
1731 procedure) as a candidate for BIGNUM assembler implementation.
1732 While 0.9.9-dev uses assembler for various architectures, only
1733 x86_64 is available by default here in the 0.9.8 branch, and
1734 32-bit x86 is available through a compile-time setting.
1736 To try the 32-bit x86 assembler implementation, use Configure
1737 option "enable-montasm" (which exists only for this backport).
1739 As "enable-montasm" for 32-bit x86 disclaims code stability
1740 anyway, in this constellation we activate additional code
1741 backported from 0.9.9-dev for further performance improvements,
1742 namely BN_from_montgomery_word. (To enable this otherwise,
1743 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1745 [Andy Polyakov (backport partially by Bodo Moeller)]
1747 *) Add TLS session ticket callback. This allows an application to set
1748 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1749 values. This is useful for key rollover for example where several key
1750 sets may exist with different names.
1753 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1754 This was broken until now in 0.9.8 releases, such that the only way
1755 a registered ENGINE could be used (assuming it initialises
1756 successfully on the host) was to explicitly set it as the default
1757 for the relevant algorithms. This is in contradiction with 0.9.7
1758 behaviour and the documentation. With this fix, when an ENGINE is
1759 registered into a given algorithm's table of implementations, the
1760 'uptodate' flag is reset so that auto-discovery will be used next
1761 time a new context for that algorithm attempts to select an
1763 [Ian Lister (tweaked by Geoff Thorpe)]
1765 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1766 implemention in the following ways:
1768 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1771 Lack of BER streaming support means one pass streaming processing is
1772 only supported if data is detached: setting the streaming flag is
1773 ignored for embedded content.
1775 CMS support is disabled by default and must be explicitly enabled
1776 with the enable-cms configuration option.
1779 *) Update the GMP engine glue to do direct copies between BIGNUM and
1780 mpz_t when openssl and GMP use the same limb size. Otherwise the
1781 existing "conversion via a text string export" trick is still used.
1782 [Paul Sheer <paulsheer@gmail.com>]
1784 *) Zlib compression BIO. This is a filter BIO which compressed and
1785 uncompresses any data passed through it.
1788 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1789 RFC3394 compatible AES key wrapping.
1792 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1793 sets string data without copying. X509_ALGOR_set0() and
1794 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1795 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1796 from an X509_ATTRIBUTE structure optionally checking it occurs only
1797 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1801 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1802 to get the expected BN_FLG_CONSTTIME behavior.
1803 [Bodo Moeller (Google)]
1807 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1808 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1809 - added some more tests to do_tests.pl
1810 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1811 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1812 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1813 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1814 - various changes to netware.pl to enable gcc-cross builds on Win32
1816 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1817 - various changes to fix missing prototype warnings
1818 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1819 - added AES, WHIRLPOOL and CPUID assembler code to build files
1820 - added missing AES assembler make rules to mk1mf.pl
1821 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1822 [Guenter Knauf <eflash@gmx.net>]
1824 *) Implement certificate status request TLS extension defined in RFC3546.
1825 A client can set the appropriate parameters and receive the encoded
1826 OCSP response via a callback. A server can query the supplied parameters
1827 and set the encoded OCSP response in the callback. Add simplified examples
1828 to s_client and s_server.
1831 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1833 *) Fix various bugs:
1834 + Binary incompatibility of ssl_ctx_st structure
1835 + DTLS interoperation with non-compliant servers
1836 + Don't call get_session_cb() without proposed session
1837 + Fix ia64 assembler code
1838 [Andy Polyakov, Steve Henson]
1840 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1842 *) DTLS Handshake overhaul. There were longstanding issues with
1843 OpenSSL DTLS implementation, which were making it impossible for
1844 RFC 4347 compliant client to communicate with OpenSSL server.
1845 Unfortunately just fixing these incompatibilities would "cut off"
1846 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1847 server keeps tolerating non RFC compliant syntax. The opposite is
1848 not true, 0.9.8f client can not communicate with earlier server.
1849 This update even addresses CVE-2007-4995.
1852 *) Changes to avoid need for function casts in OpenSSL: some compilers
1853 (gcc 4.2 and later) reject their use.
1854 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1857 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1858 RFC4507bis. The encrypted ticket format is an encrypted encoded
1859 SSL_SESSION structure, that way new session features are automatically
1862 If a client application caches session in an SSL_SESSION structure
1863 support is transparent because tickets are now stored in the encoded
1866 The SSL_CTX structure automatically generates keys for ticket
1867 protection in servers so again support should be possible
1868 with no application modification.
1870 If a client or server wishes to disable RFC4507 support then the option
1871 SSL_OP_NO_TICKET can be set.
1873 Add a TLS extension debugging callback to allow the contents of any client
1874 or server extensions to be examined.
1876 This work was sponsored by Google.
1879 *) Add initial support for TLS extensions, specifically for the server_name
1880 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1881 have new members for a host name. The SSL data structure has an
1882 additional member SSL_CTX *initial_ctx so that new sessions can be
1883 stored in that context to allow for session resumption, even after the
1884 SSL has been switched to a new SSL_CTX in reaction to a client's
1885 server_name extension.
1887 New functions (subject to change):
1889 SSL_get_servername()
1890 SSL_get_servername_type()
1893 New CTRL codes and macros (subject to change):
1895 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1896 - SSL_CTX_set_tlsext_servername_callback()
1897 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1898 - SSL_CTX_set_tlsext_servername_arg()
1899 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1901 openssl s_client has a new '-servername ...' option.
1903 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1904 '-key2 ...', '-servername_fatal' (subject to change). This allows
1905 testing the HostName extension for a specific single host name ('-cert'
1906 and '-key' remain fallbacks for handshakes without HostName
1907 negotiation). If the unrecogninzed_name alert has to be sent, this by
1908 default is a warning; it becomes fatal with the '-servername_fatal'
1911 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1913 *) Add AES and SSE2 assembly language support to VC++ build.
1916 *) Mitigate attack on final subtraction in Montgomery reduction.
1919 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1920 (which previously caused an internal error).
1923 *) Squeeze another 10% out of IGE mode when in != out.
1926 *) AES IGE mode speedup.
1927 [Dean Gaudet (Google)]
1929 *) Add the Korean symmetric 128-bit cipher SEED (see
1930 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1931 add SEED ciphersuites from RFC 4162:
1933 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1934 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1935 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1936 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1938 To minimize changes between patchlevels in the OpenSSL 0.9.8
1939 series, SEED remains excluded from compilation unless OpenSSL
1940 is configured with 'enable-seed'.
1941 [KISA, Bodo Moeller]
1943 *) Mitigate branch prediction attacks, which can be practical if a
1944 single processor is shared, allowing a spy process to extract
1945 information. For detailed background information, see
1946 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1947 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1948 and Necessary Software Countermeasures"). The core of the change
1949 are new versions BN_div_no_branch() and
1950 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1951 respectively, which are slower, but avoid the security-relevant
1952 conditional branches. These are automatically called by BN_div()
1953 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1954 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1955 remove a conditional branch.
1957 BN_FLG_CONSTTIME is the new name for the previous
1958 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1959 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1960 in the exponent causes BN_mod_exp_mont() to use the alternative
1961 implementation in BN_mod_exp_mont_consttime().) The old name
1962 remains as a deprecated alias.
1964 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1965 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1966 constant-time implementations for more than just exponentiation.
1967 Here too the old name is kept as a deprecated alias.
1969 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1970 the BN_BLINDING structure gets an independent copy of the
1971 modulus. This means that the previous "BIGNUM *m" argument to
1972 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1973 essentially becomes "const BIGNUM *m", although we can't actually
1974 change this in the header file before 0.9.9. It allows
1975 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1976 enable BN_FLG_CONSTTIME.
1978 [Matthew D Wood (Intel Corp)]
1980 *) In the SSL/TLS server implementation, be strict about session ID
1981 context matching (which matters if an application uses a single
1982 external cache for different purposes). Previously,
1983 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1984 set. This did ensure strict client verification, but meant that,
1985 with applications using a single external cache for quite
1986 different requirements, clients could circumvent ciphersuite
1987 restrictions for a given session ID context by starting a session
1988 in a different context.
1991 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1992 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1993 authentication-only ciphersuites.
1996 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1997 not complete and could lead to a possible single byte overflow
1998 (CVE-2007-5135) [Ben Laurie]
2000 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2002 *) Since AES128 and AES256 (and similarly Camellia128 and
2003 Camellia256) share a single mask bit in the logic of
2004 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2005 kludge to work properly if AES128 is available and AES256 isn't
2006 (or if Camellia128 is available and Camellia256 isn't).
2009 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2010 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2011 When a point or a seed is encoded in a BIT STRING, we need to
2012 prevent the removal of trailing zero bits to get the proper DER
2013 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2014 of a NamedBitList, for which trailing 0 bits need to be removed.)
2017 *) Have SSL/TLS server implementation tolerate "mismatched" record
2018 protocol version while receiving ClientHello even if the
2019 ClientHello is fragmented. (The server can't insist on the
2020 particular protocol version it has chosen before the ServerHello
2021 message has informed the client about his choice.)
2024 *) Add RFC 3779 support.
2025 [Rob Austein for ARIN, Ben Laurie]
2027 *) Load error codes if they are not already present instead of using a
2028 static variable. This allows them to be cleanly unloaded and reloaded.
2029 Improve header file function name parsing.
2032 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2033 or CAPABILITY handshake as required by RFCs.
2036 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2038 *) Introduce limits to prevent malicious keys being able to
2039 cause a denial of service. (CVE-2006-2940)
2040 [Steve Henson, Bodo Moeller]
2042 *) Fix ASN.1 parsing of certain invalid structures that can result
2043 in a denial of service. (CVE-2006-2937) [Steve Henson]
2045 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2046 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2048 *) Fix SSL client code which could crash if connecting to a
2049 malicious SSLv2 server. (CVE-2006-4343)
2050 [Tavis Ormandy and Will Drewry, Google Security Team]
2052 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2053 match only those. Before that, "AES256-SHA" would be interpreted
2054 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2055 the same strength classification in 0.9.7h) as we currently only
2056 have a single AES bit in the ciphersuite description bitmap.
2057 That change, however, also applied to ciphersuite strings such as
2058 "RC4-MD5" that intentionally matched multiple ciphersuites --
2059 namely, SSL 2.0 ciphersuites in addition to the more common ones
2060 from SSL 3.0/TLS 1.0.
2062 So we change the selection algorithm again: Naming an explicit
2063 ciphersuite selects this one ciphersuite, and any other similar
2064 ciphersuite (same bitmap) from *other* protocol versions.
2065 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2066 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2068 Since SSL 2.0 does not have any ciphersuites for which the
2069 128/256 bit distinction would be relevant, this works for now.
2070 The proper fix will be to use different bits for AES128 and
2071 AES256, which would have avoided the problems from the beginning;
2072 however, bits are scarce, so we can only do this in a new release
2073 (not just a patchlevel) when we can change the SSL_CIPHER
2074 definition to split the single 'unsigned long mask' bitmap into
2075 multiple values to extend the available space.
2079 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2081 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2082 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2084 *) Add AES IGE and biIGE modes.
2087 *) Change the Unix randomness entropy gathering to use poll() when
2088 possible instead of select(), since the latter has some
2089 undesirable limitations.
2090 [Darryl Miles via Richard Levitte and Bodo Moeller]
2092 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2093 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2094 cannot be implicitly activated as part of, e.g., the "AES" alias.
2095 However, please upgrade to OpenSSL 0.9.9[-dev] for
2096 non-experimental use of the ECC ciphersuites to get TLS extension
2097 support, which is required for curve and point format negotiation
2098 to avoid potential handshake problems.
2101 *) Disable rogue ciphersuites:
2103 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2104 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2105 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2107 The latter two were purportedly from
2108 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2111 Also deactivate the remaining ciphersuites from
2112 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2113 unofficial, and the ID has long expired.
2116 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2117 dual-core machines) and other potential thread-safety issues.
2120 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2121 versions), which is now available for royalty-free use
2122 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2123 Also, add Camellia TLS ciphersuites from RFC 4132.
2125 To minimize changes between patchlevels in the OpenSSL 0.9.8
2126 series, Camellia remains excluded from compilation unless OpenSSL
2127 is configured with 'enable-camellia'.
2130 *) Disable the padding bug check when compression is in use. The padding
2131 bug check assumes the first packet is of even length, this is not
2132 necessarily true if compresssion is enabled and can result in false
2133 positives causing handshake failure. The actual bug test is ancient
2134 code so it is hoped that implementations will either have fixed it by
2135 now or any which still have the bug do not support compression.
2138 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2140 *) When applying a cipher rule check to see if string match is an explicit
2141 cipher suite and only match that one cipher suite if it is.
2144 *) Link in manifests for VC++ if needed.
2145 [Austin Ziegler <halostatue@gmail.com>]
2147 *) Update support for ECC-based TLS ciphersuites according to
2148 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2149 TLS extensions, which are supported starting with the 0.9.9
2150 branch, not in the OpenSSL 0.9.8 branch).
2153 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2154 opaque EVP_CIPHER_CTX handling.
2157 *) Fixes and enhancements to zlib compression code. We now only use
2158 "zlib1.dll" and use the default __cdecl calling convention on Win32
2159 to conform with the standards mentioned here:
2160 http://www.zlib.net/DLL_FAQ.txt
2161 Static zlib linking now works on Windows and the new --with-zlib-include
2162 --with-zlib-lib options to Configure can be used to supply the location
2163 of the headers and library. Gracefully handle case where zlib library
2167 *) Several fixes and enhancements to the OID generation code. The old code
2168 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2169 handle numbers larger than ULONG_MAX, truncated printing and had a
2170 non standard OBJ_obj2txt() behaviour.
2173 *) Add support for building of engines under engine/ as shared libraries
2174 under VC++ build system.
2177 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2178 Hopefully, we will not see any false combination of paths any more.
2181 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2183 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2184 (part of SSL_OP_ALL). This option used to disable the
2185 countermeasure against man-in-the-middle protocol-version
2186 rollback in the SSL 2.0 server implementation, which is a bad
2187 idea. (CVE-2005-2969)
2189 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2190 for Information Security, National Institute of Advanced Industrial
2191 Science and Technology [AIST], Japan)]
2193 *) Add two function to clear and return the verify parameter flags.
2196 *) Keep cipherlists sorted in the source instead of sorting them at
2197 runtime, thus removing the need for a lock.
2200 *) Avoid some small subgroup attacks in Diffie-Hellman.
2201 [Nick Mathewson and Ben Laurie]
2203 *) Add functions for well-known primes.
2206 *) Extended Windows CE support.
2207 [Satoshi Nakamura and Andy Polyakov]
2209 *) Initialize SSL_METHOD structures at compile time instead of during
2210 runtime, thus removing the need for a lock.
2213 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2214 attempting to decrypt each encrypted key in turn. Add support to
2218 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2220 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2223 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2226 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2227 key into the same file any more.
2230 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2233 *) Add -utf8 command line and config file option to 'ca'.
2234 [Stefan <stf@udoma.org]
2236 *) Removed the macro des_crypt(), as it seems to conflict with some
2237 libraries. Use DES_crypt().
2240 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2241 involves renaming the source and generated shared-libs for
2242 both. The engines will accept the corrected or legacy ids
2243 ('ncipher' and '4758_cca' respectively) when binding. NB,
2244 this only applies when building 'shared'.
2245 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2247 *) Add attribute functions to EVP_PKEY structure. Modify
2248 PKCS12_create() to recognize a CSP name attribute and
2249 use it. Make -CSP option work again in pkcs12 utility.
2252 *) Add new functionality to the bn blinding code:
2253 - automatic re-creation of the BN_BLINDING parameters after
2254 a fixed number of uses (currently 32)
2255 - add new function for parameter creation
2256 - introduce flags to control the update behaviour of the
2257 BN_BLINDING parameters
2258 - hide BN_BLINDING structure
2259 Add a second BN_BLINDING slot to the RSA structure to improve
2260 performance when a single RSA object is shared among several
2264 *) Add support for DTLS.
2265 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2267 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2268 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2271 *) Remove buggy and incompletet DH cert support from
2272 ssl/ssl_rsa.c and ssl/s3_both.c
2275 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2276 the apps/openssl applications.
2279 *) Compile clean with "-Wall -Wmissing-prototypes
2280 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2281 DEBUG_SAFESTACK must also be set.
2284 *) Change ./Configure so that certain algorithms can be disabled by default.
2285 The new counterpiece to "no-xxx" is "enable-xxx".
2287 The patented RC5 and MDC2 algorithms will now be disabled unless
2288 "enable-rc5" and "enable-mdc2", respectively, are specified.
2290 (IDEA remains enabled despite being patented. This is because IDEA
2291 is frequently required for interoperability, and there is no license
2292 fee for non-commercial use. As before, "no-idea" can be used to
2293 avoid this algorithm.)
2297 *) Add processing of proxy certificates (see RFC 3820). This work was
2298 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2299 EGEE (Enabling Grids for E-science in Europe).
2302 *) RC4 performance overhaul on modern architectures/implementations, such
2303 as Intel P4, IA-64 and AMD64.
2306 *) New utility extract-section.pl. This can be used specify an alternative
2307 section number in a pod file instead of having to treat each file as
2308 a separate case in Makefile. This can be done by adding two lines to the
2311 =for comment openssl_section:XXX
2313 The blank line is mandatory.
2317 *) New arguments -certform, -keyform and -pass for s_client and s_server
2318 to allow alternative format key and certificate files and passphrase
2322 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2323 update associated structures and add various utility functions.
2325 Add new policy related verify parameters, include policy checking in
2326 standard verify code. Enhance 'smime' application with extra parameters
2327 to support policy checking and print out.
2330 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2331 Nehemiah processors. These extensions support AES encryption in hardware
2332 as well as RNG (though RNG support is currently disabled).
2333 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2335 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2338 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2339 [Andy Polyakov and a number of other people]
2341 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2342 implementation contributed by IBM.
2343 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2345 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2346 exponent rather than 'unsigned long'. There is a corresponding change to
2347 the new 'rsa_keygen' element of the RSA_METHOD structure.
2348 [Jelte Jansen, Geoff Thorpe]
2350 *) Functionality for creating the initial serial number file is now
2351 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2353 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2354 number file to 1, which is bound to cause problems. To avoid
2355 the problems while respecting compatibility between different 0.9.7
2356 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2357 CA.pl for serial number initialization. With the new release 0.9.8,
2358 we can fix the problem directly in the 'ca' utility.)
2361 *) Reduced header interdepencies by declaring more opaque objects in
2362 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2363 give fewer recursive includes, which could break lazy source code - so
2364 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2365 developers should define this symbol when building and using openssl to
2366 ensure they track the recommended behaviour, interfaces, [etc], but
2367 backwards-compatible behaviour prevails when this isn't defined.
2370 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2373 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2374 This will generate a random key of the appropriate length based on the
2375 cipher context. The EVP_CIPHER can provide its own random key generation
2376 routine to support keys of a specific form. This is used in the des and
2377 3des routines to generate a key of the correct parity. Update S/MIME
2378 code to use new functions and hence generate correct parity DES keys.
2379 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2380 valid (weak or incorrect parity).
2383 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2384 as looking them up. This is useful when the verified structure may contain
2385 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2386 present unless the new PKCS7_NO_CRL flag is asserted.
2389 *) Extend ASN1 oid configuration module. It now additionally accepts the
2392 shortName = some long name, 1.2.3.4
2395 *) Reimplemented the BN_CTX implementation. There is now no more static
2396 limitation on the number of variables it can handle nor the depth of the
2397 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2398 information can now expand as required, and rather than having a single
2399 static array of bignums, BN_CTX now uses a linked-list of such arrays
2400 allowing it to expand on demand whilst maintaining the usefulness of
2401 BN_CTX's "bundling".
2404 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2405 to allow all RSA operations to function using a single BN_CTX.
2408 *) Preliminary support for certificate policy evaluation and checking. This
2409 is initially intended to pass the tests outlined in "Conformance Testing
2410 of Relying Party Client Certificate Path Processing Logic" v1.07.
2413 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2414 remained unused and not that useful. A variety of other little bignum
2415 tweaks and fixes have also been made continuing on from the audit (see
2419 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2420 associated ASN1, EVP and SSL functions and old ASN1 macros.
2423 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2424 and this should never fail. So the return value from the use of
2425 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2426 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2429 *) BN_CTX_get() should return zero-valued bignums, providing the same
2430 initialised value as BN_new().
2431 [Geoff Thorpe, suggested by Ulf Möller]
2433 *) Support for inhibitAnyPolicy certificate extension.
2436 *) An audit of the BIGNUM code is underway, for which debugging code is
2437 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2438 is considered valid when processing BIGNUMs, and causes execution to
2439 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2440 further steps are taken to deliberately pollute unused data in BIGNUM
2441 structures to try and expose faulty code further on. For now, openssl will
2442 (in its default mode of operation) continue to tolerate the inconsistent
2443 forms that it has tolerated in the past, but authors and packagers should
2444 consider trying openssl and their own applications when compiled with
2445 these debugging symbols defined. It will help highlight potential bugs in
2446 their own code, and will improve the test coverage for OpenSSL itself. At
2447 some point, these tighter rules will become openssl's default to improve
2448 maintainability, though the assert()s and other overheads will remain only
2449 in debugging configurations. See bn.h for more details.
2450 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2452 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2453 that can only be obtained through BN_CTX_new() (which implicitly
2454 initialises it). The presence of this function only made it possible
2455 to overwrite an existing structure (and cause memory leaks).
2458 *) Because of the callback-based approach for implementing LHASH as a
2459 template type, lh_insert() adds opaque objects to hash-tables and
2460 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2461 to clean up those corresponding objects before destroying the hash table
2462 (and losing the object pointers). So some over-zealous constifications in
2463 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2464 objects as "const" and the lh_doall[_arg] callback wrappers are not
2465 prototyped to have "const" restrictions on the object pointers they are
2466 given (and so aren't required to cast them away any more).
2469 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2470 (speed) prefers to use its own implementation. The two implementations
2471 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2472 its object type properly exposed (MS_TM) instead of casting to/from "char
2473 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2474 aren't necessarily the greatest nomenclatures - but this is what was used
2475 internally to the implementation so I've used that for now.
2478 *) Ensure that deprecated functions do not get compiled when
2479 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2480 the self-tests were still using deprecated key-generation functions so
2481 these have been updated also.
2484 *) Reorganise PKCS#7 code to separate the digest location functionality
2485 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2486 New function PKCS7_set_digest() to set the digest type for PKCS#7
2487 digestedData type. Add additional code to correctly generate the
2488 digestedData type and add support for this type in PKCS7 initialization
2492 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2493 structure of type "other".
2496 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2497 sure the loop does correctly stop and breaking ("division by zero")
2498 modulus operations are not performed. The (pre-generated) prime
2499 table crypto/bn/bn_prime.h was already correct, but it could not be
2500 re-generated on some platforms because of the "division by zero"
2501 situation in the script.
2502 [Ralf S. Engelschall]
2504 *) Update support for ECC-based TLS ciphersuites according to
2505 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2506 SHA-1 now is only used for "small" curves (where the
2507 representation of a field element takes up to 24 bytes); for
2508 larger curves, the field element resulting from ECDH is directly
2509 used as premaster secret.
2510 [Douglas Stebila (Sun Microsystems Laboratories)]
2512 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2513 curve secp160r1 to the tests.
2514 [Douglas Stebila (Sun Microsystems Laboratories)]
2516 *) Add the possibility to load symbols globally with DSO.
2517 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2519 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2520 control of the error stack.
2523 *) Add support for STORE in ENGINE.
2526 *) Add the STORE type. The intention is to provide a common interface
2527 to certificate and key stores, be they simple file-based stores, or
2528 HSM-type store, or LDAP stores, or...
2529 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2532 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2533 pass a list of arguments to any function as well as provide a way
2534 for a function to pass data back to the caller.
2537 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2538 works like BUF_strdup() but can be used to duplicate a portion of
2539 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2543 *) Add the function sk_find_ex() which works like sk_find(), but will
2544 return an index to an element even if an exact match couldn't be
2545 found. The index is guaranteed to point at the element where the
2546 searched-for key would be inserted to preserve sorting order.
2549 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2550 takes an extra flags argument for optional functionality. Currently,
2551 the following flags are defined:
2553 OBJ_BSEARCH_VALUE_ON_NOMATCH
2554 This one gets OBJ_bsearch_ex() to return a pointer to the first
2555 element where the comparing function returns a negative or zero
2558 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2559 This one gets OBJ_bsearch_ex() to return a pointer to the first
2560 element where the comparing function returns zero. This is useful
2561 if there are more than one element where the comparing function
2565 *) Make it possible to create self-signed certificates with 'openssl ca'
2566 in such a way that the self-signed certificate becomes part of the
2567 CA database and uses the same mechanisms for serial number generation
2568 as all other certificate signing. The new flag '-selfsign' enables
2569 this functionality. Adapt CA.sh and CA.pl.in.
2572 *) Add functionality to check the public key of a certificate request
2573 against a given private. This is useful to check that a certificate
2574 request can be signed by that key (self-signing).
2577 *) Make it possible to have multiple active certificates with the same
2578 subject in the CA index file. This is done only if the keyword
2579 'unique_subject' is set to 'no' in the main CA section (default
2580 if 'CA_default') of the configuration file. The value is saved
2581 with the database itself in a separate index attribute file,
2582 named like the index file with '.attr' appended to the name.
2585 *) Generate muti valued AVAs using '+' notation in config files for
2589 *) Support for nameConstraints certificate extension.
2592 *) Support for policyConstraints certificate extension.
2595 *) Support for policyMappings certificate extension.
2598 *) Make sure the default DSA_METHOD implementation only uses its
2599 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2600 and change its own handlers to be NULL so as to remove unnecessary
2601 indirection. This lets alternative implementations fallback to the
2602 default implementation more easily.
2605 *) Support for directoryName in GeneralName related extensions
2609 *) Make it possible to link applications using Makefile.shared.
2610 Make that possible even when linking against static libraries!
2613 *) Support for single pass processing for S/MIME signing. This now
2614 means that S/MIME signing can be done from a pipe, in addition
2615 cleartext signing (multipart/signed type) is effectively streaming
2616 and the signed data does not need to be all held in memory.
2618 This is done with a new flag PKCS7_STREAM. When this flag is set
2619 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2620 is done after the data is output (and digests calculated) in
2621 SMIME_write_PKCS7().
2624 *) Add full support for -rpath/-R, both in shared libraries and
2625 applications, at least on the platforms where it's known how
2629 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2630 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2631 will now compute a table of multiples of the generator that
2632 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2633 faster (notably in the case of a single point multiplication,
2634 scalar * generator).
2635 [Nils Larsch, Bodo Moeller]
2637 *) IPv6 support for certificate extensions. The various extensions
2638 which use the IP:a.b.c.d can now take IPv6 addresses using the
2639 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2643 *) Added an ENGINE that implements RSA by performing private key
2644 exponentiations with the GMP library. The conversions to and from
2645 GMP's mpz_t format aren't optimised nor are any montgomery forms
2646 cached, and on x86 it appears OpenSSL's own performance has caught up.
2647 However there are likely to be other architectures where GMP could
2648 provide a boost. This ENGINE is not built in by default, but it can be
2649 specified at Configure time and should be accompanied by the necessary
2650 linker additions, eg;
2651 ./config -DOPENSSL_USE_GMP -lgmp
2654 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2655 testing availability of engines with "-t" - the old behaviour is
2656 produced by increasing the feature's verbosity with "-tt".
2659 *) ECDSA routines: under certain error conditions uninitialized BN objects
2660 could be freed. Solution: make sure initialization is performed early
2661 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2665 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2666 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2667 software implementations. For DSA and DH, parameter generation can
2668 also be overriden by providing the appropriate method callbacks.
2671 *) Change the "progress" mechanism used in key-generation and
2672 primality testing to functions that take a new BN_GENCB pointer in
2673 place of callback/argument pairs. The new API functions have "_ex"
2674 postfixes and the older functions are reimplemented as wrappers for
2675 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2676 declarations of the old functions to help (graceful) attempts to
2677 migrate to the new functions. Also, the new key-generation API
2678 functions operate on a caller-supplied key-structure and return
2679 success/failure rather than returning a key or NULL - this is to
2680 help make "keygen" another member function of RSA_METHOD etc.
2682 Example for using the new callback interface:
2684 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2688 BN_GENCB_set(&my_cb, my_callback, my_arg);
2690 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2691 /* For the meaning of a, b in calls to my_callback(), see the
2692 * documentation of the function that calls the callback.
2693 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2694 * my_callback should return 1 if it wants BN_is_prime_ex()
2695 * to continue, or 0 to stop.
2700 *) Change the ZLIB compression method to be stateful, and make it
2701 available to TLS with the number defined in
2702 draft-ietf-tls-compression-04.txt.
2705 *) Add the ASN.1 structures and functions for CertificatePair, which
2706 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2708 CertificatePair ::= SEQUENCE {
2709 forward [0] Certificate OPTIONAL,
2710 reverse [1] Certificate OPTIONAL,
2711 -- at least one of the pair shall be present -- }
2713 Also implement the PEM functions to read and write certificate
2714 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2716 This needed to be defined, mostly for the sake of the LDAP
2717 attribute crossCertificatePair, but may prove useful elsewhere as
2721 *) Make it possible to inhibit symlinking of shared libraries in
2722 Makefile.shared, for Cygwin's sake.
2725 *) Extend the BIGNUM API by creating a function
2726 void BN_set_negative(BIGNUM *a, int neg);
2727 and a macro that behave like
2728 int BN_is_negative(const BIGNUM *a);
2730 to avoid the need to access 'a->neg' directly in applications.
2733 *) Implement fast modular reduction for pseudo-Mersenne primes
2734 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2735 EC_GROUP_new_curve_GFp() will now automatically use this
2737 [Nils Larsch <nla@trustcenter.de>]
2739 *) Add new lock type (CRYPTO_LOCK_BN).
2742 *) Change the ENGINE framework to automatically load engines
2743 dynamically from specific directories unless they could be
2744 found to already be built in or loaded. Move all the
2745 current engines except for the cryptodev one to a new
2747 The engines in engines/ are built as shared libraries if
2748 the "shared" options was given to ./Configure or ./config.
2749 Otherwise, they are inserted in libcrypto.a.
2750 /usr/local/ssl/engines is the default directory for dynamic
2751 engines, but that can be overriden at configure time through
2752 the usual use of --prefix and/or --openssldir, and at run
2753 time with the environment variable OPENSSL_ENGINES.
2754 [Geoff Thorpe and Richard Levitte]
2756 *) Add Makefile.shared, a helper makefile to build shared
2757 libraries. Addapt Makefile.org.
2760 *) Add version info to Win32 DLLs.
2761 [Peter 'Luna' Runestig" <peter@runestig.com>]
2763 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2764 can be added using this API to created arbitrary PKCS#12
2765 files while avoiding the low level API.
2767 New options to PKCS12_create(), key or cert can be NULL and
2768 will then be omitted from the output file. The encryption
2769 algorithm NIDs can be set to -1 for no encryption, the mac
2770 iteration count can be set to 0 to omit the mac.
2772 Enhance pkcs12 utility by making the -nokeys and -nocerts
2773 options work when creating a PKCS#12 file. New option -nomac
2774 to omit the mac, NONE can be set for an encryption algorithm.
2775 New code is modified to use the enhanced PKCS12_create()
2776 instead of the low level API.
2779 *) Extend ASN1 encoder to support indefinite length constructed
2780 encoding. This can output sequences tags and octet strings in
2781 this form. Modify pk7_asn1.c to support indefinite length
2782 encoding. This is experimental and needs additional code to
2783 be useful, such as an ASN1 bio and some enhanced streaming
2786 Extend template encode functionality so that tagging is passed
2787 down to the template encoder.
2790 *) Let 'openssl req' fail if an argument to '-newkey' is not
2791 recognized instead of using RSA as a default.
2794 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2795 As these are not official, they are not included in "ALL";
2796 the "ECCdraft" ciphersuite group alias can be used to select them.
2797 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2799 *) Add ECDH engine support.
2800 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2802 *) Add ECDH in new directory crypto/ecdh/.
2803 [Douglas Stebila (Sun Microsystems Laboratories)]
2805 *) Let BN_rand_range() abort with an error after 100 iterations
2806 without success (which indicates a broken PRNG).
2809 *) Change BN_mod_sqrt() so that it verifies that the input value
2810 is really the square of the return value. (Previously,
2811 BN_mod_sqrt would show GIGO behaviour.)
2814 *) Add named elliptic curves over binary fields from X9.62, SECG,
2815 and WAP/WTLS; add OIDs that were still missing.
2817 [Sheueling Chang Shantz and Douglas Stebila
2818 (Sun Microsystems Laboratories)]
2820 *) Extend the EC library for elliptic curves over binary fields
2821 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2824 EC_GF2m_simple_method
2828 EC_GROUP_new_curve_GF2m
2829 EC_GROUP_set_curve_GF2m
2830 EC_GROUP_get_curve_GF2m
2831 EC_POINT_set_affine_coordinates_GF2m
2832 EC_POINT_get_affine_coordinates_GF2m
2833 EC_POINT_set_compressed_coordinates_GF2m
2835 Point compression for binary fields is disabled by default for
2836 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2839 As binary polynomials are represented as BIGNUMs, various members
2840 of the EC_GROUP and EC_POINT data structures can be shared
2841 between the implementations for prime fields and binary fields;
2842 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2843 are essentially identical to their ..._GFp counterparts.
2844 (For simplicity, the '..._GFp' prefix has been dropped from
2845 various internal method names.)
2847 An internal 'field_div' method (similar to 'field_mul' and
2848 'field_sqr') has been added; this is used only for binary fields.
2850 [Sheueling Chang Shantz and Douglas Stebila
2851 (Sun Microsystems Laboratories)]
2853 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2854 through methods ('mul', 'precompute_mult').
2856 The generic implementations (now internally called 'ec_wNAF_mul'
2857 and 'ec_wNAF_precomputed_mult') remain the default if these
2858 methods are undefined.
2860 [Sheueling Chang Shantz and Douglas Stebila
2861 (Sun Microsystems Laboratories)]
2863 *) New function EC_GROUP_get_degree, which is defined through
2864 EC_METHOD. For curves over prime fields, this returns the bit
2865 length of the modulus.
2867 [Sheueling Chang Shantz and Douglas Stebila
2868 (Sun Microsystems Laboratories)]
2870 *) New functions EC_GROUP_dup, EC_POINT_dup.
2871 (These simply call ..._new and ..._copy).
2873 [Sheueling Chang Shantz and Douglas Stebila
2874 (Sun Microsystems Laboratories)]
2876 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2877 Polynomials are represented as BIGNUMs (where the sign bit is not
2878 used) in the following functions [macros]:
2881 BN_GF2m_sub [= BN_GF2m_add]
2882 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2883 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2884 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2886 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2887 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2888 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2889 BN_GF2m_cmp [= BN_ucmp]
2891 (Note that only the 'mod' functions are actually for fields GF(2^m).
2892 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2894 For some functions, an the irreducible polynomial defining a
2895 field can be given as an 'unsigned int[]' with strictly
2896 decreasing elements giving the indices of those bits that are set;
2897 i.e., p[] represents the polynomial
2898 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2900 p[0] > p[1] > ... > p[k] = 0.
2901 This applies to the following functions:
2906 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2907 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2909 BN_GF2m_mod_sqrt_arr
2910 BN_GF2m_mod_solve_quad_arr
2914 Conversion can be performed by the following functions:
2919 bntest.c has additional tests for binary polynomial arithmetic.
2921 Two implementations for BN_GF2m_mod_div() are available.
2922 The default algorithm simply uses BN_GF2m_mod_inv() and
2923 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2924 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2925 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2927 [Sheueling Chang Shantz and Douglas Stebila
2928 (Sun Microsystems Laboratories)]
2930 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2931 functionality is disabled at compile-time.
2932 [Douglas Stebila <douglas.stebila@sun.com>]
2934 *) Change default behaviour of 'openssl asn1parse' so that more
2935 information is visible when viewing, e.g., a certificate:
2937 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2938 mode the content of non-printable OCTET STRINGs is output in a
2939 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2940 avoid the appearance of a printable string.
2941 [Nils Larsch <nla@trustcenter.de>]
2943 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2945 EC_GROUP_set_asn1_flag()
2946 EC_GROUP_get_asn1_flag()
2947 EC_GROUP_set_point_conversion_form()
2948 EC_GROUP_get_point_conversion_form()
2949 These control ASN1 encoding details:
2950 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2951 has been set to OPENSSL_EC_NAMED_CURVE.
2952 - Points are encoded in uncompressed form by default; options for
2953 asn1_for are as for point2oct, namely
2954 POINT_CONVERSION_COMPRESSED
2955 POINT_CONVERSION_UNCOMPRESSED
2956 POINT_CONVERSION_HYBRID
2958 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2961 EC_GROUP_get0_seed()
2962 EC_GROUP_get_seed_len()
2963 This is used only for ASN1 purposes (so far).
2964 [Nils Larsch <nla@trustcenter.de>]
2966 *) Add 'field_type' member to EC_METHOD, which holds the NID
2967 of the appropriate field type OID. The new function
2968 EC_METHOD_get_field_type() returns this value.
2969 [Nils Larsch <nla@trustcenter.de>]
2974 EC_POINT_point2hex()
2975 EC_POINT_hex2point()
2976 providing useful interfaces to EC_POINT_point2oct() and
2977 EC_POINT_oct2point().
2978 [Nils Larsch <nla@trustcenter.de>]
2980 *) Change internals of the EC library so that the functions
2981 EC_GROUP_set_generator()
2982 EC_GROUP_get_generator()
2983 EC_GROUP_get_order()
2984 EC_GROUP_get_cofactor()
2985 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2986 to methods, which would lead to unnecessary code duplication when
2987 adding different types of curves.
2988 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2990 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2991 arithmetic, and such that modified wNAFs are generated
2992 (which avoid length expansion in many cases).
2995 *) Add a function EC_GROUP_check_discriminant() (defined via
2996 EC_METHOD) that verifies that the curve discriminant is non-zero.
2998 Add a function EC_GROUP_check() that makes some sanity tests
2999 on a EC_GROUP, its generator and order. This includes
3000 EC_GROUP_check_discriminant().
3001 [Nils Larsch <nla@trustcenter.de>]
3003 *) Add ECDSA in new directory crypto/ecdsa/.
3005 Add applications 'openssl ecparam' and 'openssl ecdsa'
3006 (these are based on 'openssl dsaparam' and 'openssl dsa').
3008 ECDSA support is also included in various other files across the
3009 library. Most notably,
3010 - 'openssl req' now has a '-newkey ecdsa:file' option;
3011 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3012 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3013 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3014 them suitable for ECDSA where domain parameters must be
3015 extracted before the specific public key;
3016 - ECDSA engine support has been added.
3017 [Nils Larsch <nla@trustcenter.de>]
3019 *) Include some named elliptic curves, and add OIDs from X9.62,
3020 SECG, and WAP/WTLS. Each curve can be obtained from the new
3022 EC_GROUP_new_by_curve_name(),
3023 and the list of available named curves can be obtained with
3024 EC_get_builtin_curves().
3025 Also add a 'curve_name' member to EC_GROUP objects, which can be
3027 EC_GROUP_set_curve_name()
3028 EC_GROUP_get_curve_name()
3029 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3031 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3032 was actually never needed) and in BN_mul(). The removal in BN_mul()
3033 required a small change in bn_mul_part_recursive() and the addition
3034 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3035 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3036 bn_sub_words() and bn_add_words() except they take arrays with
3040 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3042 *) Cleanse PEM buffers before freeing them since they may contain
3044 [Benjamin Bennett <ben@psc.edu>]
3046 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3047 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3048 authentication-only ciphersuites.
3051 *) Since AES128 and AES256 share a single mask bit in the logic of
3052 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3053 kludge to work properly if AES128 is available and AES256 isn't.
3056 *) Expand security boundary to match 1.1.1 module.
3059 *) Remove redundant features: hash file source, editing of test vectors
3060 modify fipsld to use external fips_premain.c signature.
3063 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3064 run algorithm test programs.
3067 *) Make algorithm test programs more tolerant of whitespace.
3070 *) Have SSL/TLS server implementation tolerate "mismatched" record
3071 protocol version while receiving ClientHello even if the
3072 ClientHello is fragmented. (The server can't insist on the
3073 particular protocol version it has chosen before the ServerHello
3074 message has informed the client about his choice.)
3077 *) Load error codes if they are not already present instead of using a
3078 static variable. This allows them to be cleanly unloaded and reloaded.
3081 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3083 *) Introduce limits to prevent malicious keys being able to
3084 cause a denial of service. (CVE-2006-2940)
3085 [Steve Henson, Bodo Moeller]
3087 *) Fix ASN.1 parsing of certain invalid structures that can result
3088 in a denial of service. (CVE-2006-2937) [Steve Henson]
3090 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3091 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3093 *) Fix SSL client code which could crash if connecting to a
3094 malicious SSLv2 server. (CVE-2006-4343)
3095 [Tavis Ormandy and Will Drewry, Google Security Team]
3097 *) Change ciphersuite string processing so that an explicit
3098 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3099 will no longer include "AES128-SHA"), and any other similar
3100 ciphersuite (same bitmap) from *other* protocol versions (so that
3101 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3102 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3103 changes from 0.9.8b and 0.9.8d.
3106 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3108 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3109 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3111 *) Change the Unix randomness entropy gathering to use poll() when
3112 possible instead of select(), since the latter has some
3113 undesirable limitations.
3114 [Darryl Miles via Richard Levitte and Bodo Moeller]
3116 *) Disable rogue ciphersuites:
3118 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3119 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3120 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3122 The latter two were purportedly from
3123 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3126 Also deactive the remaining ciphersuites from
3127 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3128 unofficial, and the ID has long expired.
3131 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3132 dual-core machines) and other potential thread-safety issues.
3135 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3137 *) Adapt fipsld and the build system to link against the validated FIPS
3138 module in FIPS mode.
3141 *) Fixes for VC++ 2005 build under Windows.
3144 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3145 from a Windows bash shell such as MSYS. It is autodetected from the
3146 "config" script when run from a VC++ environment. Modify standard VC++
3147 build to use fipscanister.o from the GNU make build.
3150 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3152 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3153 The value now differs depending on if you build for FIPS or not.
3154 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3155 safely run with a non-FIPSed libcrypto, as it may crash because of
3156 the difference induced by this change.
3159 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3161 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3162 (part of SSL_OP_ALL). This option used to disable the
3163 countermeasure against man-in-the-middle protocol-version
3164 rollback in the SSL 2.0 server implementation, which is a bad
3165 idea. (CVE-2005-2969)
3167 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3168 for Information Security, National Institute of Advanced Industrial
3169 Science and Technology [AIST], Japan)]
3171 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3172 mainly for FIPS compliance and not fully integrated at this stage.
3175 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3176 the exponentiation using a fixed-length exponent. (Otherwise,
3177 the information leaked through timing could expose the secret key
3178 after many signatures; cf. Bleichenbacher's attack on DSA with
3182 *) Make a new fixed-window mod_exp implementation the default for
3183 RSA, DSA, and DH private-key operations so that the sequence of
3184 squares and multiplies and the memory access pattern are
3185 independent of the particular secret key. This will mitigate
3186 cache-timing and potential related attacks.
3188 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3189 and this is automatically used by BN_mod_exp_mont() if the new flag
3190 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3191 will use this BN flag for private exponents unless the flag
3192 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3193 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3195 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3197 *) Change the client implementation for SSLv23_method() and
3198 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3199 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3200 (Previously, the SSL 2.0 backwards compatible Client Hello
3201 message format would be used even with SSL_OP_NO_SSLv2.)
3204 *) Add support for smime-type MIME parameter in S/MIME messages which some
3208 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3209 a threadsafe manner. Modify rsa code to use new function and add calls
3210 to dsa and dh code (which had race conditions before).
3213 *) Include the fixed error library code in the C error file definitions
3214 instead of fixing them up at runtime. This keeps the error code
3215 structures constant.
3218 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3220 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3223 *) Fixes for newer kerberos headers. NB: the casts are needed because
3224 the 'length' field is signed on one version and unsigned on another
3225 with no (?) obvious way to tell the difference, without these VC++
3226 complains. Also the "definition" of FAR (blank) is no longer included
3227 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3228 some needed definitions.
3231 *) Undo Cygwin change.
3234 *) Added support for proxy certificates according to RFC 3820.
3235 Because they may be a security thread to unaware applications,
3236 they must be explicitely allowed in run-time. See
3237 docs/HOWTO/proxy_certificates.txt for further information.
3240 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3242 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3243 server and client random values. Previously
3244 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3245 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3247 This change has negligible security impact because:
3249 1. Server and client random values still have 24 bytes of pseudo random
3252 2. Server and client random values are sent in the clear in the initial
3255 3. The master secret is derived using the premaster secret (48 bytes in
3256 size for static RSA ciphersuites) as well as client server and random
3259 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3262 [Stephen Henson, reported by UK NISCC]
3264 *) Use Windows randomness collection on Cygwin.
3267 *) Fix hang in EGD/PRNGD query when communication socket is closed
3268 prematurely by EGD/PRNGD.
3269 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3271 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3274 *) Back-port of selected performance improvements from development
3275 branch, as well as improved support for PowerPC platforms.
3278 *) Add lots of checks for memory allocation failure, error codes to indicate
3279 failure and freeing up memory if a failure occurs.
3280 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3282 *) Add new -passin argument to dgst.
3285 *) Perform some character comparisons of different types in X509_NAME_cmp:
3286 this is needed for some certificates that reencode DNs into UTF8Strings
3287 (in violation of RFC3280) and can't or wont issue name rollover
3291 *) Make an explicit check during certificate validation to see that
3292 the CA setting in each certificate on the chain is correct. As a
3293 side effect always do the following basic checks on extensions,
3294 not just when there's an associated purpose to the check:
3296 - if there is an unhandled critical extension (unless the user
3297 has chosen to ignore this fault)
3298 - if the path length has been exceeded (if one is set at all)
3299 - that certain extensions fit the associated purpose (if one has
3303 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3305 *) Avoid a race condition when CRLs are checked in a multi threaded
3306 environment. This would happen due to the reordering of the revoked
3307 entries during signature checking and serial number lookup. Now the
3308 encoding is cached and the serial number sort performed under a lock.
3309 Add new STACK function sk_is_sorted().
3312 *) Add Delta CRL to the extension code.
3315 *) Various fixes to s3_pkt.c so alerts are sent properly.
3316 [David Holmes <d.holmes@f5.com>]
3318 *) Reduce the chances of duplicate issuer name and serial numbers (in
3319 violation of RFC3280) using the OpenSSL certificate creation utilities.
3320 This is done by creating a random 64 bit value for the initial serial
3321 number when a serial number file is created or when a self signed
3322 certificate is created using 'openssl req -x509'. The initial serial
3323 number file is created using 'openssl x509 -next_serial' in CA.pl
3324 rather than being initialized to 1.
3327 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3329 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3330 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3331 [Joe Orton, Steve Henson]
3333 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3335 [Joe Orton, Steve Henson]
3337 *) Make it possible to have multiple active certificates with the same
3338 subject in the CA index file. This is done only if the keyword
3339 'unique_subject' is set to 'no' in the main CA section (default
3340 if 'CA_default') of the configuration file. The value is saved
3341 with the database itself in a separate index attribute file,
3342 named like the index file with '.attr' appended to the name.
3345 *) X509 verify fixes. Disable broken certificate workarounds when
3346 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3347 keyUsage extension present. Don't accept CRLs with unhandled critical
3348 extensions: since verify currently doesn't process CRL extensions this
3349 rejects a CRL with *any* critical extensions. Add new verify error codes
3353 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3354 A clarification of RFC2560 will require the use of OCTET STRINGs and
3355 some implementations cannot handle the current raw format. Since OpenSSL
3356 copies and compares OCSP nonces as opaque blobs without any attempt at
3357 parsing them this should not create any compatibility issues.
3360 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3361 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3362 this HMAC (and other) operations are several times slower than OpenSSL
3366 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3367 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3369 *) Use the correct content when signing type "other".
3372 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3374 *) Fix various bugs revealed by running the NISCC test suite:
3376 Stop out of bounds reads in the ASN1 code when presented with
3377 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3379 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3381 If verify callback ignores invalid public key errors don't try to check
3382 certificate signature with the NULL public key.
3386 *) New -ignore_err option in ocsp application to stop the server
3387 exiting on the first error in a request.
3390 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3391 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3395 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3396 extra data after the compression methods not only for TLS 1.0
3397 but also for SSL 3.0 (as required by the specification).
3398 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3400 *) Change X509_certificate_type() to mark the key as exported/exportable
3401 when it's 512 *bits* long, not 512 bytes.
3404 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3405 blocks during encryption.
3408 *) Various fixes to base64 BIO and non blocking I/O. On write
3409 flushes were not handled properly if the BIO retried. On read
3410 data was not being buffered properly and had various logic bugs.
3411 This also affects blocking I/O when the data being decoded is a
3415 *) Various S/MIME bugfixes and compatibility changes:
3416 output correct application/pkcs7 MIME type if
3417 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3418 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3419 of files as .eml work). Correctly handle very long lines in MIME
3423 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3425 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3426 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3427 a protocol version number mismatch like a decryption error
3428 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3431 *) Turn on RSA blinding by default in the default implementation
3432 to avoid a timing attack. Applications that don't want it can call
3433 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3434 They would be ill-advised to do so in most cases.
3435 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3437 *) Change RSA blinding code so that it works when the PRNG is not
3438 seeded (in this case, the secret RSA exponent is abused as
3439 an unpredictable seed -- if it is not unpredictable, there
3440 is no point in blinding anyway). Make RSA blinding thread-safe
3441 by remembering the creator's thread ID in rsa->blinding and
3442 having all other threads use local one-time blinding factors
3443 (this requires more computation than sharing rsa->blinding, but
3444 avoids excessive locking; and if an RSA object is not shared
3445 between threads, blinding will still be very fast).
3448 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3449 ENGINE as defaults for all supported algorithms irrespective of
3450 the 'flags' parameter. 'flags' is now honoured, so applications
3451 should make sure they are passing it correctly.
3454 *) Target "mingw" now allows native Windows code to be generated in
3455 the Cygwin environment as well as with the MinGW compiler.
3458 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3460 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3461 via timing by performing a MAC computation even if incorrrect
3462 block cipher padding has been found. This is a countermeasure
3463 against active attacks where the attacker has to distinguish
3464 between bad padding and a MAC verification error. (CVE-2003-0078)
3466 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3467 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3468 Martin Vuagnoux (EPFL, Ilion)]
3470 *) Make the no-err option work as intended. The intention with no-err
3471 is not to have the whole error stack handling routines removed from
3472 libcrypto, it's only intended to remove all the function name and
3473 reason texts, thereby removing some of the footprint that may not
3474 be interesting if those errors aren't displayed anyway.
3476 NOTE: it's still possible for any application or module to have it's
3477 own set of error texts inserted. The routines are there, just not
3478 used by default when no-err is given.
3481 *) Add support for FreeBSD on IA64.
3482 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3484 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3485 Kerberos function mit_des_cbc_cksum(). Before this change,
3486 the value returned by DES_cbc_cksum() was like the one from
3487 mit_des_cbc_cksum(), except the bytes were swapped.
3488 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3490 *) Allow an application to disable the automatic SSL chain building.
3491 Before this a rather primitive chain build was always performed in
3492 ssl3_output_cert_chain(): an application had no way to send the
3493 correct chain if the automatic operation produced an incorrect result.
3495 Now the chain builder is disabled if either:
3497 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3499 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3501 The reasoning behind this is that an application would not want the
3502 auto chain building to take place if extra chain certificates are
3503 present and it might also want a means of sending no additional
3504 certificates (for example the chain has two certificates and the
3508 *) Add the possibility to build without the ENGINE framework.
3509 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3511 *) Under Win32 gmtime() can return NULL: check return value in
3512 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3515 *) DSA routines: under certain error conditions uninitialized BN objects
3516 could be freed. Solution: make sure initialization is performed early
3517 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3518 Nils Larsch <nla@trustcenter.de> via PR#459)
3521 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3522 checked on reconnect on the client side, therefore session resumption
3523 could still fail with a "ssl session id is different" error. This
3524 behaviour is masked when SSL_OP_ALL is used due to
3525 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3526 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3527 followup to PR #377.
3530 *) IA-32 assembler support enhancements: unified ELF targets, support
3531 for SCO/Caldera platforms, fix for Cygwin shared build.
3534 *) Add support for FreeBSD on sparc64. As a consequence, support for
3535 FreeBSD on non-x86 processors is separate from x86 processors on
3536 the config script, much like the NetBSD support.
3537 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3539 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3541 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3544 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3545 code (06) was taken as the first octet of the session ID and the last
3546 octet was ignored consequently. As a result SSLv2 client side session
3547 caching could not have worked due to the session ID mismatch between
3549 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3553 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3554 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3558 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3559 seems that in spite of existing for more than a year, many application
3560 author have done nothing to provide the necessary callbacks, which
3561 means that this particular engine will not work properly anywhere.
3562 This is a very unfortunate situation which forces us, in the name
3563 of usability, to give the hw_ncipher.c a static lock, which is part
3565 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3566 appear in 0.9.8 or later. We EXPECT application authors to have
3567 dealt properly with this when 0.9.8 is released (unless we actually
3568 make such changes in the libcrypto locking code that changes will
3569 have to be made anyway).
3572 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3573 octets have been read, EOF or an error occurs. Without this change
3574 some truncated ASN1 structures will not produce an error.
3577 *) Disable Heimdal support, since it hasn't been fully implemented.
3578 Still give the possibility to force the use of Heimdal, but with
3579 warnings and a request that patches get sent to openssl-dev.
3582 *) Add the VC-CE target, introduce the WINCE sysname, and add
3583 INSTALL.WCE and appropriate conditionals to make it build.
3584 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3586 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3587 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3588 edit numbers of the version.
3589 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3591 *) Introduce safe string copy and catenation functions
3592 (BUF_strlcpy() and BUF_strlcat()).
3593 [Ben Laurie (CHATS) and Richard Levitte]
3595 *) Avoid using fixed-size buffers for one-line DNs.
3596 [Ben Laurie (CHATS)]
3598 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3599 resizing buffers containing secrets, and use where appropriate.
3600 [Ben Laurie (CHATS)]
3602 *) Avoid using fixed size buffers for configuration file location.
3603 [Ben Laurie (CHATS)]
3605 *) Avoid filename truncation for various CA files.
3606 [Ben Laurie (CHATS)]
3608 *) Use sizeof in preference to magic numbers.
3609 [Ben Laurie (CHATS)]
3611 *) Avoid filename truncation in cert requests.
3612 [Ben Laurie (CHATS)]
3614 *) Add assertions to check for (supposedly impossible) buffer
3616 [Ben Laurie (CHATS)]
3618 *) Don't cache truncated DNS entries in the local cache (this could
3619 potentially lead to a spoofing attack).
3620 [Ben Laurie (CHATS)]
3622 *) Fix various buffers to be large enough for hex/decimal
3623 representations in a platform independent manner.
3624 [Ben Laurie (CHATS)]
3626 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3627 resizing buffers containing secrets, and use where appropriate.
3628 [Ben Laurie (CHATS)]
3630 *) Add BIO_indent() to avoid much slightly worrying code to do
3632 [Ben Laurie (CHATS)]
3634 *) Convert sprintf()/BIO_puts() to BIO_printf().
3635 [Ben Laurie (CHATS)]
3637 *) buffer_gets() could terminate with the buffer only half
3639 [Ben Laurie (CHATS)]
3641 *) Add assertions to prevent user-supplied crypto functions from
3642 overflowing internal buffers by having large block sizes, etc.
3643 [Ben Laurie (CHATS)]
3645 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3647 [Ben Laurie (CHATS)]
3649 *) Eliminate unused copy of key in RC4.
3650 [Ben Laurie (CHATS)]
3652 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3653 [Ben Laurie (CHATS)]
3655 *) Fix off-by-one error in EGD path.
3656 [Ben Laurie (CHATS)]
3658 *) If RANDFILE path is too long, ignore instead of truncating.
3659 [Ben Laurie (CHATS)]
3661 *) Eliminate unused and incorrectly sized X.509 structure
3663 [Ben Laurie (CHATS)]
3665 *) Eliminate unused and dangerous function knumber().
3666 [Ben Laurie (CHATS)]
3668 *) Eliminate unused and dangerous structure, KSSL_ERR.
3669 [Ben Laurie (CHATS)]
3671 *) Protect against overlong session ID context length in an encoded
3672 session object. Since these are local, this does not appear to be
3674 [Ben Laurie (CHATS)]
3676 *) Change from security patch (see 0.9.6e below) that did not affect
3677 the 0.9.6 release series:
3679 Remote buffer overflow in SSL3 protocol - an attacker could
3680 supply an oversized master key in Kerberos-enabled versions.
3682 [Ben Laurie (CHATS)]
3684 *) Change the SSL kerb5 codes to match RFC 2712.
3687 *) Make -nameopt work fully for req and add -reqopt switch.
3688 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3690 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3691 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3693 *) Make sure tests can be performed even if the corresponding algorithms
3694 have been removed entirely. This was also the last step to make
3695 OpenSSL compilable with DJGPP under all reasonable conditions.
3696 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3698 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3699 to allow version independent disabling of normally unselected ciphers,
3700 which may be activated as a side-effect of selecting a single cipher.
3702 (E.g., cipher list string "RSA" enables ciphersuites that are left
3703 out of "ALL" because they do not provide symmetric encryption.
3704 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3705 [Lutz Jaenicke, Bodo Moeller]
3707 *) Add appropriate support for separate platform-dependent build
3708 directories. The recommended way to make a platform-dependent
3709 build directory is the following (tested on Linux), maybe with
3712 # Place yourself outside of the OpenSSL source tree. In
3713 # this example, the environment variable OPENSSL_SOURCE
3714 # is assumed to contain the absolute OpenSSL source directory.
3715 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3716 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3717 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3718 mkdir -p `dirname $F`
3719 ln -s $OPENSSL_SOURCE/$F $F
3722 To be absolutely sure not to disturb the source tree, a "make clean"
3723 is a good thing. If it isn't successfull, don't worry about it,
3724 it probably means the source directory is very clean.
3727 *) Make sure any ENGINE control commands make local copies of string
3728 pointers passed to them whenever necessary. Otherwise it is possible
3729 the caller may have overwritten (or deallocated) the original string
3730 data when a later ENGINE operation tries to use the stored values.
3731 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3733 *) Improve diagnostics in file reading and command-line digests.
3734 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3736 *) Add AES modes CFB and OFB to the object database. Correct an
3737 error in AES-CFB decryption.
3740 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3741 allows existing EVP_CIPHER_CTX structures to be reused after
3742 calling EVP_*Final(). This behaviour is used by encryption
3743 BIOs and some applications. This has the side effect that
3744 applications must explicitly clean up cipher contexts with
3745 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3748 *) Check the values of dna and dnb in bn_mul_recursive before calling
3749 bn_mul_comba (a non zero value means the a or b arrays do not contain
3750 n2 elements) and fallback to bn_mul_normal if either is not zero.
3753 *) Fix escaping of non-ASCII characters when using the -subj option
3754 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3757 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3758 form for "surname", serialNumber has no short form.
3759 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3760 therefore remove "mail" short name for "internet 7".
3761 The OID for unique identifiers in X509 certificates is
3762 x500UniqueIdentifier, not uniqueIdentifier.
3763 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3766 *) Add an "init" command to the ENGINE config module and auto initialize
3767 ENGINEs. Without any "init" command the ENGINE will be initialized
3768 after all ctrl commands have been executed on it. If init=1 the
3769 ENGINE is initailized at that point (ctrls before that point are run
3770 on the uninitialized ENGINE and after on the initialized one). If
3771 init=0 then the ENGINE will not be iniatialized at all.
3774 *) Fix the 'app_verify_callback' interface so that the user-defined
3775 argument is actually passed to the callback: In the
3776 SSL_CTX_set_cert_verify_callback() prototype, the callback
3777 declaration has been changed from
3780 int (*cb)(X509_STORE_CTX *,void *);
3781 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3782 i=s->ctx->app_verify_callback(&ctx)
3783 has been changed into
3784 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3786 To update applications using SSL_CTX_set_cert_verify_callback(),
3787 a dummy argument can be added to their callback functions.
3788 [D. K. Smetters <smetters@parc.xerox.com>]
3790 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3791 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3793 *) Add and OPENSSL_LOAD_CONF define which will cause
3794 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3795 This allows older applications to transparently support certain
3796 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3797 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3798 load the config file and OPENSSL_add_all_algorithms_conf() which will
3799 always load it have also been added.
3802 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3803 Adjust NIDs and EVP layer.
3804 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3806 *) Config modules support in openssl utility.
3808 Most commands now load modules from the config file,
3809 though in a few (such as version) this isn't done
3810 because it couldn't be used for anything.
3812 In the case of ca and req the config file used is
3813 the same as the utility itself: that is the -config
3814 command line option can be used to specify an
3818 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3819 use "openssl_conf" if filename is NULL use default openssl config file.
3822 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3823 config section name. Add a new flag to tolerate a missing config file
3824 and move code to CONF_modules_load_file().
3827 *) Support for crypto accelerator cards from Accelerated Encryption
3828 Processing, www.aep.ie. (Use engine 'aep')
3829 The support was copied from 0.9.6c [engine] and adapted/corrected
3830 to work with the new engine framework.
3831 [AEP Inc. and Richard Levitte]
3833 *) Support for SureWare crypto accelerator cards from Baltimore
3834 Technologies. (Use engine 'sureware')
3835 The support was copied from 0.9.6c [engine] and adapted
3836 to work with the new engine framework.
3839 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3840 make the newer ENGINE framework commands for the CHIL engine work.
3841 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3843 *) Make it possible to produce shared libraries on ReliantUNIX.
3844 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3846 *) Add the configuration target debug-linux-ppro.
3847 Make 'openssl rsa' use the general key loading routines
3848 implemented in apps.c, and make those routines able to
3849 handle the key format FORMAT_NETSCAPE and the variant
3851 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3853 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3854 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3856 *) Add -keyform to rsautl, and document -engine.
3857 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3859 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3860 BIO_R_NO_SUCH_FILE error code rather than the generic
3861 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3864 *) Add new functions
3866 ERR_peek_last_error_line
3867 ERR_peek_last_error_line_data.
3868 These are similar to
3871 ERR_peek_error_line_data,
3872 but report on the latest error recorded rather than the first one
3873 still in the error queue.
3874 [Ben Laurie, Bodo Moeller]
3876 *) default_algorithms option in ENGINE config module. This allows things
3878 default_algorithms = ALL
3879 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3882 *) Prelminary ENGINE config module.
3885 *) New experimental application configuration code.
3888 *) Change the AES code to follow the same name structure as all other
3889 symmetric ciphers, and behave the same way. Move everything to
3890 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3891 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3893 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3894 [Ben Laurie and Theo de Raadt]
3896 *) Add option to output public keys in req command.
3897 [Massimiliano Pala madwolf@openca.org]
3899 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3900 (up to about 10% better than before for P-192 and P-224).
3903 *) New functions/macros
3905 SSL_CTX_set_msg_callback(ctx, cb)
3906 SSL_CTX_set_msg_callback_arg(ctx, arg)
3907 SSL_set_msg_callback(ssl, cb)
3908 SSL_set_msg_callback_arg(ssl, arg)
3910 to request calling a callback function
3912 void cb(int write_p, int version, int content_type,
3913 const void *buf, size_t len, SSL *ssl, void *arg)
3915 whenever a protocol message has been completely received
3916 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3917 protocol version according to which the SSL library interprets
3918 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3919 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3920 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3921 specification (change_cipher_spec(20), alert(21), handshake(22)).
3922 'buf' and 'len' point to the actual message, 'ssl' to the
3923 SSL object, and 'arg' is the application-defined value set by
3924 SSL[_CTX]_set_msg_callback_arg().
3926 'openssl s_client' and 'openssl s_server' have new '-msg' options
3927 to enable a callback that displays all protocol messages.
3930 *) Change the shared library support so shared libraries are built as
3931 soon as the corresponding static library is finished, and thereby get
3932 openssl and the test programs linked against the shared library.
3933 This still only happens when the keyword "shard" has been given to
3934 the configuration scripts.
3936 NOTE: shared library support is still an experimental thing, and
3937 backward binary compatibility is still not guaranteed.
3938 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3940 *) Add support for Subject Information Access extension.
3941 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3943 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3944 additional bytes when new memory had to be allocated, not just
3945 when reusing an existing buffer.
3948 *) New command line and configuration option 'utf8' for the req command.
3949 This allows field values to be specified as UTF8 strings.
3952 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3953 runs for the former and machine-readable output for the latter.
3956 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3957 of the e-mail address in the DN (i.e., it will go into a certificate
3958 extension only). The new configuration file option 'email_in_dn = no'
3959 has the same effect.
3960 [Massimiliano Pala madwolf@openca.org]
3962 *) Change all functions with names starting with des_ to be starting
3963 with DES_ instead. Add wrappers that are compatible with libdes,
3964 but are named _ossl_old_des_*. Finally, add macros that map the
3965 des_* symbols to the corresponding _ossl_old_des_* if libdes
3966 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3967 desired, the des_* symbols will be mapped to DES_*, with one
3970 Since we provide two compatibility mappings, the user needs to
3971 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3972 compatibility is desired. The default (i.e., when that macro
3973 isn't defined) is OpenSSL 0.9.6c compatibility.
3975 There are also macros that enable and disable the support of old
3976 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3977 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3978 are defined, the default will apply: to support the old des routines.
3980 In either case, one must include openssl/des.h to get the correct
3981 definitions. Do not try to just include openssl/des_old.h, that
3984 NOTE: This is a major break of an old API into a new one. Software
3985 authors are encouraged to switch to the DES_ style functions. Some
3986 time in the future, des_old.h and the libdes compatibility functions
3987 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3988 default), and then completely removed.
3991 *) Test for certificates which contain unsupported critical extensions.
3992 If such a certificate is found during a verify operation it is
3993 rejected by default: this behaviour can be overridden by either
3994 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3995 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3996 X509_supported_extension() has also been added which returns 1 if a
3997 particular extension is supported.
4000 *) Modify the behaviour of EVP cipher functions in similar way to digests
4001 to retain compatibility with existing code.
4004 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4005 compatibility with existing code. In particular the 'ctx' parameter does
4006 not have to be to be initialized before the call to EVP_DigestInit() and
4007 it is tidied up after a call to EVP_DigestFinal(). New function
4008 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4009 EVP_MD_CTX_copy() changed to not require the destination to be
4010 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4011 requires the destination to be valid.
4013 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4014 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4017 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4018 so that complete 'Handshake' protocol structures are kept in memory
4019 instead of overwriting 'msg_type' and 'length' with 'body' data.
4022 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4023 [Massimo Santin via Richard Levitte]
4025 *) Major restructuring to the underlying ENGINE code. This includes
4026 reduction of linker bloat, separation of pure "ENGINE" manipulation
4027 (initialisation, etc) from functionality dealing with implementations
4028 of specific crypto iterfaces. This change also introduces integrated
4029 support for symmetric ciphers and digest implementations - so ENGINEs
4030 can now accelerate these by providing EVP_CIPHER and EVP_MD
4031 implementations of their own. This is detailed in crypto/engine/README
4032 as it couldn't be adequately described here. However, there are a few
4033 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4034 were changed in the original introduction of ENGINE code have now
4035 reverted back - the hooking from this code to ENGINE is now a good
4036 deal more passive and at run-time, operations deal directly with
4037 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4038 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4039 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4040 they were not being used by the framework as there is no concept of a
4041 BIGNUM_METHOD and they could not be generalised to the new
4042 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4043 ENGINE_cpy() has been removed as it cannot be consistently defined in
4047 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4050 *) Change mkdef.pl to sort symbols that get the same entry number,
4051 and make sure the automatically generated functions ERR_load_*
4052 become part of libeay.num as well.
4055 *) New function SSL_renegotiate_pending(). This returns true once
4056 renegotiation has been requested (either SSL_renegotiate() call
4057 or HelloRequest/ClientHello receveived from the peer) and becomes
4058 false once a handshake has been completed.
4059 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4060 sends a HelloRequest, but does not ensure that a handshake takes
4061 place. SSL_renegotiate_pending() is useful for checking if the
4062 client has followed the request.)
4065 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4066 By default, clients may request session resumption even during
4067 renegotiation (if session ID contexts permit); with this option,
4068 session resumption is possible only in the first handshake.
4070 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4071 more bits available for options that should not be part of
4072 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4075 *) Add some demos for certificate and certificate request creation.
4078 *) Make maximum certificate chain size accepted from the peer application
4079 settable (SSL*_get/set_max_cert_list()), as proposed by
4080 "Douglas E. Engert" <deengert@anl.gov>.
4083 *) Add support for shared libraries for Unixware-7
4084 (Boyd Lynn Gerber <gerberb@zenez.com>).
4087 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4088 be done prior to destruction. Use this to unload error strings from
4089 ENGINEs that load their own error strings. NB: This adds two new API
4090 functions to "get" and "set" this destroy handler in an ENGINE.
4093 *) Alter all existing ENGINE implementations (except "openssl" and
4094 "openbsd") to dynamically instantiate their own error strings. This
4095 makes them more flexible to be built both as statically-linked ENGINEs
4096 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4097 Also, add stub code to each that makes building them as self-contained
4098 shared-libraries easier (see README.ENGINE).
4101 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4102 implementations into applications that are completely implemented in
4103 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4104 commands that can be used to configure what shared-library to load and
4105 to control aspects of the way it is handled. Also, made an update to
4106 the README.ENGINE file that brings its information up-to-date and
4107 provides some information and instructions on the "dynamic" ENGINE
4108 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4111 *) Make it possible to unload ranges of ERR strings with a new
4112 "ERR_unload_strings" function.
4115 *) Add a copy() function to EVP_MD.
4118 *) Make EVP_MD routines take a context pointer instead of just the
4119 md_data void pointer.
4122 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4123 that the digest can only process a single chunk of data
4124 (typically because it is provided by a piece of
4125 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4126 is only going to provide a single chunk of data, and hence the
4127 framework needn't accumulate the data for oneshot drivers.
4130 *) As with "ERR", make it possible to replace the underlying "ex_data"
4131 functions. This change also alters the storage and management of global
4132 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4133 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4134 index counters. The API functions that use this state have been changed
4135 to take a "class_index" rather than pointers to the class's local STACK
4136 and counter, and there is now an API function to dynamically create new
4137 classes. This centralisation allows us to (a) plug a lot of the
4138 thread-safety problems that existed, and (b) makes it possible to clean
4139 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4140 such data would previously have always leaked in application code and
4141 workarounds were in place to make the memory debugging turn a blind eye
4142 to it. Application code that doesn't use this new function will still
4143 leak as before, but their memory debugging output will announce it now
4144 rather than letting it slide.
4146 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4147 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4148 has a return value to indicate success or failure.
4151 *) Make it possible to replace the underlying "ERR" functions such that the
4152 global state (2 LHASH tables and 2 locks) is only used by the "default"
4153 implementation. This change also adds two functions to "get" and "set"
4154 the implementation prior to it being automatically set the first time
4155 any other ERR function takes place. Ie. an application can call "get",
4156 pass the return value to a module it has just loaded, and that module
4157 can call its own "set" function using that value. This means the
4158 module's "ERR" operations will use (and modify) the error state in the
4159 application and not in its own statically linked copy of OpenSSL code.
4162 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4163 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4164 the operation, and provides a more encapsulated way for external code
4165 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4166 to use these functions rather than manually incrementing the counts.
4168 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4171 *) Add EVP test program.
4174 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4177 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4178 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4179 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4180 These allow a CRL to be built without having to access X509_CRL fields
4181 directly. Modify 'ca' application to use new functions.
4184 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4185 bug workarounds. Rollback attack detection is a security feature.
4186 The problem will only arise on OpenSSL servers when TLSv1 is not
4187 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4188 Software authors not wanting to support TLSv1 will have special reasons
4189 for their choice and can explicitly enable this option.
4190 [Bodo Moeller, Lutz Jaenicke]
4192 *) Rationalise EVP so it can be extended: don't include a union of
4193 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4194 (similar to those existing for EVP_CIPHER_CTX).
4199 EVP_MD_CTX_init(&md); /* new function call */
4200 EVP_DigestInit(&md, EVP_sha1());
4201 EVP_DigestUpdate(&md, in, len);
4202 EVP_DigestFinal(&md, out, NULL);
4203 EVP_MD_CTX_cleanup(&md); /* new function call */
4207 *) Make DES key schedule conform to the usual scheme, as well as
4208 correcting its structure. This means that calls to DES functions
4209 now have to pass a pointer to a des_key_schedule instead of a
4210 plain des_key_schedule (which was actually always a pointer
4213 des_key_schedule ks;
4215 des_set_key_checked(..., &ks);
4216 des_ncbc_encrypt(..., &ks, ...);
4218 (Note that a later change renames 'des_...' into 'DES_...'.)
4221 *) Initial reduction of linker bloat: the use of some functions, such as
4222 PEM causes large amounts of unused functions to be linked in due to
4223 poor organisation. For example pem_all.c contains every PEM function
4224 which has a knock on effect of linking in large amounts of (unused)
4225 ASN1 code. Grouping together similar functions and splitting unrelated
4226 functions prevents this.
4229 *) Cleanup of EVP macros.
4232 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4233 correct _ecb suffix.
4236 *) Add initial OCSP responder support to ocsp application. The
4237 revocation information is handled using the text based index
4238 use by the ca application. The responder can either handle
4239 requests generated internally, supplied in files (for example
4240 via a CGI script) or using an internal minimal server.
4243 *) Add configuration choices to get zlib compression for TLS.
4246 *) Changes to Kerberos SSL for RFC 2712 compliance:
4247 1. Implemented real KerberosWrapper, instead of just using
4248 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4249 2. Implemented optional authenticator field of KerberosWrapper.
4251 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4252 and authenticator structs; see crypto/krb5/.
4254 Generalized Kerberos calls to support multiple Kerberos libraries.
4255 [Vern Staats <staatsvr@asc.hpc.mil>,
4256 Jeffrey Altman <jaltman@columbia.edu>
4257 via Richard Levitte]
4259 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4260 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4261 values for each of the key sizes rather than having just
4262 parameters (and 'speed' generating keys each time).
4265 *) Speed up EVP routines.
4268 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4269 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4270 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4271 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4273 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4274 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4275 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4278 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4280 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4283 *) Added the OS2-EMX target.
4284 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4286 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4287 to support NCONF routines in extension code. New function CONF_set_nconf()
4288 to allow functions which take an NCONF to also handle the old LHASH
4289 structure: this means that the old CONF compatible routines can be
4290 retained (in particular wrt extensions) without having to duplicate the
4291 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4294 *) Enhance the general user interface with mechanisms for inner control
4295 and with possibilities to have yes/no kind of prompts.
4298 *) Change all calls to low level digest routines in the library and
4299 applications to use EVP. Add missing calls to HMAC_cleanup() and
4300 don't assume HMAC_CTX can be copied using memcpy().
4301 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4303 *) Add the possibility to control engines through control names but with
4304 arbitrary arguments instead of just a string.
4305 Change the key loaders to take a UI_METHOD instead of a callback
4306 function pointer. NOTE: this breaks binary compatibility with earlier
4307 versions of OpenSSL [engine].
4308 Adapt the nCipher code for these new conditions and add a card insertion
4312 *) Enhance the general user interface with mechanisms to better support
4313 dialog box interfaces, application-defined prompts, the possibility
4314 to use defaults (for example default passwords from somewhere else)
4315 and interrupts/cancellations.
4318 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4319 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4322 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4323 tidy up some unnecessarily weird code in 'sk_new()').
4324 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4326 *) Change the key loading routines for ENGINEs to use the same kind
4327 callback (pem_password_cb) as all other routines that need this
4331 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4332 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4333 than this minimum value is recommended.
4336 *) New random seeder for OpenVMS, using the system process statistics
4337 that are easily reachable.
4340 *) Windows apparently can't transparently handle global
4341 variables defined in DLLs. Initialisations such as:
4343 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4345 wont compile. This is used by the any applications that need to
4346 declare their own ASN1 modules. This was fixed by adding the option
4347 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4348 needed for static libraries under Win32.
4351 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4352 setting of purpose and trust fields. New X509_STORE trust and
4353 purpose functions and tidy up setting in other SSL functions.
4356 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4357 structure. These are inherited by X509_STORE_CTX when it is
4358 initialised. This allows various defaults to be set in the
4359 X509_STORE structure (such as flags for CRL checking and custom
4360 purpose or trust settings) for functions which only use X509_STORE_CTX
4361 internally such as S/MIME.
4363 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4364 trust settings if they are not set in X509_STORE. This allows X509_STORE
4365 purposes and trust (in S/MIME for example) to override any set by default.
4367 Add command line options for CRL checking to smime, s_client and s_server
4371 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4372 are set then the CRL is looked up in the X509_STORE structure and
4373 its validity and signature checked, then if the certificate is found
4374 in the CRL the verify fails with a revoked error.
4376 Various new CRL related callbacks added to X509_STORE_CTX structure.
4378 Command line options added to 'verify' application to support this.
4380 This needs some additional work, such as being able to handle multiple
4381 CRLs with different times, extension based lookup (rather than just
4382 by subject name) and ultimately more complete V2 CRL extension
4386 *) Add a general user interface API (crypto/ui/). This is designed
4387 to replace things like des_read_password and friends (backward
4388 compatibility functions using this new API are provided).
4389 The purpose is to remove prompting functions from the DES code
4390 section as well as provide for prompting through dialog boxes in
4391 a window system and the like.
4394 *) Add "ex_data" support to ENGINE so implementations can add state at a
4395 per-structure level rather than having to store it globally.
4398 *) Make it possible for ENGINE structures to be copied when retrieved by
4399 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4400 This causes the "original" ENGINE structure to act like a template,
4401 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4402 operational state can be localised to each ENGINE structure, despite the
4403 fact they all share the same "methods". New ENGINE structures returned in
4404 this case have no functional references and the return value is the single
4405 structural reference. This matches the single structural reference returned
4406 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4410 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4411 needs to match any other type at all we need to manually clear the
4415 *) Changes to the "openssl engine" utility to include;
4416 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4417 about an ENGINE's available control commands.
4418 - executing control commands from command line arguments using the
4419 '-pre' and '-post' switches. '-post' is only used if '-t' is
4420 specified and the ENGINE is successfully initialised. The syntax for
4421 the individual commands are colon-separated, for example;
4422 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4425 *) New dynamic control command support for ENGINEs. ENGINEs can now
4426 declare their own commands (numbers), names (strings), descriptions,
4427 and input types for run-time discovery by calling applications. A
4428 subset of these commands are implicitly classed as "executable"
4429 depending on their input type, and only these can be invoked through
4430 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4431 can be based on user input, config files, etc). The distinction is
4432 that "executable" commands cannot return anything other than a boolean
4433 result and can only support numeric or string input, whereas some
4434 discoverable commands may only be for direct use through
4435 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4436 pointers, or other custom uses. The "executable" commands are to
4437 support parameterisations of ENGINE behaviour that can be
4438 unambiguously defined by ENGINEs and used consistently across any
4439 OpenSSL-based application. Commands have been added to all the
4440 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4441 control over shared-library paths without source code alterations.
4444 *) Changed all ENGINE implementations to dynamically allocate their
4445 ENGINEs rather than declaring them statically. Apart from this being
4446 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4447 this also allows the implementations to compile without using the
4448 internal engine_int.h header.
4451 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4452 'const' value. Any code that should be able to modify a RAND_METHOD
4453 should already have non-const pointers to it (ie. they should only
4454 modify their own ones).
4457 *) Made a variety of little tweaks to the ENGINE code.
4458 - "atalla" and "ubsec" string definitions were moved from header files
4459 to C code. "nuron" string definitions were placed in variables
4460 rather than hard-coded - allowing parameterisation of these values
4461 later on via ctrl() commands.
4462 - Removed unused "#if 0"'d code.
4463 - Fixed engine list iteration code so it uses ENGINE_free() to release
4464 structural references.
4465 - Constified the RAND_METHOD element of ENGINE structures.
4466 - Constified various get/set functions as appropriate and added
4467 missing functions (including a catch-all ENGINE_cpy that duplicates
4468 all ENGINE values onto a new ENGINE except reference counts/state).
4469 - Removed NULL parameter checks in get/set functions. Setting a method
4470 or function to NULL is a way of cancelling out a previously set
4471 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4472 and doesn't justify the extra error symbols and code.
4473 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4474 flags from engine_int.h to engine.h.
4475 - Changed prototypes for ENGINE handler functions (init(), finish(),
4476 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4479 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4480 to the algorithm using long division. The binary algorithm can be
4481 used only if the modulus is odd. On 32-bit systems, it is faster
4482 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4483 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4484 up to 450 bits. In 64-bit environments, the binary algorithm
4485 appears to be advantageous for much longer moduli; here we use it
4486 for moduli up to 2048 bits.
4489 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4490 could not support the combine flag in choice fields.
4493 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4494 extensions from a certificate request to the certificate.
4497 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4498 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4499 file: this allows the display of the certificate about to be
4500 signed to be customised, to allow certain fields to be included
4501 or excluded and extension details. The old system didn't display
4502 multicharacter strings properly, omitted fields not in the policy
4503 and couldn't display additional details such as extensions.
4506 *) Function EC_POINTs_mul for multiple scalar multiplication
4507 of an arbitrary number of elliptic curve points
4508 \sum scalars[i]*points[i],
4509 optionally including the generator defined for the EC_GROUP:
4510 scalar*generator + \sum scalars[i]*points[i].
4512 EC_POINT_mul is a simple wrapper function for the typical case
4513 that the point list has just one item (besides the optional
4517 *) First EC_METHODs for curves over GF(p):
4519 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4520 operations and provides various method functions that can also
4521 operate with faster implementations of modular arithmetic.
4523 EC_GFp_mont_method() reuses most functions that are part of
4524 EC_GFp_simple_method, but uses Montgomery arithmetic.
4526 [Bodo Moeller; point addition and point doubling
4527 implementation directly derived from source code provided by
4528 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4530 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4531 crypto/ec/ec_lib.c):
4533 Curves are EC_GROUP objects (with an optional group generator)
4534 based on EC_METHODs that are built into the library.
4536 Points are EC_POINT objects based on EC_GROUP objects.
4538 Most of the framework would be able to handle curves over arbitrary
4539 finite fields, but as there are no obvious types for fields other
4540 than GF(p), some functions are limited to that for now.
4543 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4544 that the file contains a complete HTTP response.
4547 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4548 change the def and num file printf format specifier from "%-40sXXX"
4549 to "%-39s XXX". The latter will always guarantee a space after the
4550 field while the former will cause them to run together if the field
4551 is 40 of more characters long.
4554 *) Constify the cipher and digest 'method' functions and structures
4555 and modify related functions to take constant EVP_MD and EVP_CIPHER
4559 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4560 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4563 *) Modify EVP_Digest*() routines so they now return values. Although the
4564 internal software routines can never fail additional hardware versions
4568 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4570 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4571 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4574 ERR_R_NESTED_ASN1_ERROR
4576 ERR_R_MISSING_ASN1_EOS
4577 were 4 .. 9, conflicting with
4578 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4580 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4581 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4583 Add new error code 'ERR_R_INTERNAL_ERROR'.
4586 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4590 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4591 sets the subject name for a new request or supersedes the
4592 subject name in a given request. Formats that can be parsed are
4593 'CN=Some Name, OU=myOU, C=IT'
4595 'CN=Some Name/OU=myOU/C=IT'.
4597 Add options '-batch' and '-verbose' to 'openssl req'.
4598 [Massimiliano Pala <madwolf@hackmasters.net>]
4600 *) Introduce the possibility to access global variables through
4601 functions on platform were that's the best way to handle exporting
4602 global variables in shared libraries. To enable this functionality,
4603 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4604 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4605 is normally done by Configure or something similar).
4607 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4608 in the source file (foo.c) like this:
4610 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4611 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4613 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4614 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4616 OPENSSL_DECLARE_GLOBAL(int,foo);
4617 #define foo OPENSSL_GLOBAL_REF(foo)
4618 OPENSSL_DECLARE_GLOBAL(double,bar);
4619 #define bar OPENSSL_GLOBAL_REF(bar)
4621 The #defines are very important, and therefore so is including the
4622 header file everywhere where the defined globals are used.
4624 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4625 of ASN.1 items, but that structure is a bit different.
4627 The largest change is in util/mkdef.pl which has been enhanced with
4628 better and easier to understand logic to choose which symbols should
4629 go into the Windows .def files as well as a number of fixes and code
4630 cleanup (among others, algorithm keywords are now sorted
4631 lexicographically to avoid constant rewrites).
4634 *) In BN_div() keep a copy of the sign of 'num' before writing the
4635 result to 'rm' because if rm==num the value will be overwritten
4636 and produce the wrong result if 'num' is negative: this caused
4637 problems with BN_mod() and BN_nnmod().
4640 *) Function OCSP_request_verify(). This checks the signature on an
4641 OCSP request and verifies the signer certificate. The signer
4642 certificate is just checked for a generic purpose and OCSP request
4646 *) Add OCSP_check_validity() function to check the validity of OCSP
4647 responses. OCSP responses are prepared in real time and may only
4648 be a few seconds old. Simply checking that the current time lies
4649 between thisUpdate and nextUpdate max reject otherwise valid responses
4650 caused by either OCSP responder or client clock inaccuracy. Instead
4651 we allow thisUpdate and nextUpdate to fall within a certain period of
4652 the current time. The age of the response can also optionally be
4653 checked. Two new options -validity_period and -status_age added to
4657 *) If signature or public key algorithm is unrecognized print out its
4658 OID rather that just UNKNOWN.
4661 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4662 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4663 ID to be generated from the issuer certificate alone which can then be
4664 passed to OCSP_id_issuer_cmp().
4667 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4668 ASN1 modules to export functions returning ASN1_ITEM pointers
4669 instead of the ASN1_ITEM structures themselves. This adds several
4670 new macros which allow the underlying ASN1 function/structure to
4671 be accessed transparently. As a result code should not use ASN1_ITEM
4672 references directly (such as &X509_it) but instead use the relevant
4673 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4674 use of the new ASN1 code on platforms where exporting structures
4675 is problematical (for example in shared libraries) but exporting
4676 functions returning pointers to structures is not.
4679 *) Add support for overriding the generation of SSL/TLS session IDs.
4680 These callbacks can be registered either in an SSL_CTX or per SSL.
4681 The purpose of this is to allow applications to control, if they wish,
4682 the arbitrary values chosen for use as session IDs, particularly as it
4683 can be useful for session caching in multiple-server environments. A
4684 command-line switch for testing this (and any client code that wishes
4685 to use such a feature) has been added to "s_server".
4686 [Geoff Thorpe, Lutz Jaenicke]
4688 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4689 of the form '#if defined(...) || defined(...) || ...' and
4690 '#if !defined(...) && !defined(...) && ...'. This also avoids
4691 the growing number of special cases it was previously handling.
4694 *) Make all configuration macros available for application by making
4695 sure they are available in opensslconf.h, by giving them names starting
4696 with "OPENSSL_" to avoid conflicts with other packages and by making
4697 sure e_os2.h will cover all platform-specific cases together with
4699 Additionally, it is now possible to define configuration/platform-
4700 specific names (called "system identities"). In the C code, these
4701 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4702 macro with the name beginning with "OPENSSL_SYS_", which is determined
4703 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4707 *) New option -set_serial to 'req' and 'x509' this allows the serial
4708 number to use to be specified on the command line. Previously self
4709 signed certificates were hard coded with serial number 0 and the
4710 CA options of 'x509' had to use a serial number in a file which was
4714 *) New options to 'ca' utility to support V2 CRL entry extensions.
4715 Currently CRL reason, invalidity date and hold instruction are
4716 supported. Add new CRL extensions to V3 code and some new objects.
4719 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4720 disable standard block padding (aka PKCS#5 padding) in the EVP
4721 API, which was previously mandatory. This means that the data is
4722 not padded in any way and so the total length much be a multiple
4723 of the block size, otherwise an error occurs.
4726 *) Initial (incomplete) OCSP SSL support.
4729 *) New function OCSP_parse_url(). This splits up a URL into its host,
4730 port and path components: primarily to parse OCSP URLs. New -url
4731 option to ocsp utility.
4734 *) New nonce behavior. The return value of OCSP_check_nonce() now
4735 reflects the various checks performed. Applications can decide
4736 whether to tolerate certain situations such as an absent nonce
4737 in a response when one was present in a request: the ocsp application
4738 just prints out a warning. New function OCSP_add1_basic_nonce()
4739 this is to allow responders to include a nonce in a response even if
4740 the request is nonce-less.
4743 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4744 skipped when using openssl x509 multiple times on a single input file,
4745 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4748 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4749 set string type: to handle setting ASN1_TIME structures. Fix ca
4750 utility to correctly initialize revocation date of CRLs.
4753 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4754 the clients preferred ciphersuites and rather use its own preferences.
4755 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4756 Internet Explorer by ensuring unchanged hash method during stepup.
4757 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4760 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4761 to aes and add a new 'exist' option to print out symbols that don't
4765 *) Additional options to ocsp utility to allow flags to be set and
4766 additional certificates supplied.
4769 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4770 OCSP client a number of certificate to only verify the response
4774 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4775 handle the new API. Currently only ECB, CBC modes supported. Add new
4778 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4779 Encryption Standard (AES) Ciphersuites for Transport Layer
4780 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4781 not enabled by default and were not part of the "ALL" ciphersuite
4782 alias because they were not yet official; they could be
4783 explicitly requested by specifying the "AESdraft" ciphersuite
4784 group alias. In the final release of OpenSSL 0.9.7, the group
4785 alias is called "AES" and is part of "ALL".)
4786 [Ben Laurie, Steve Henson, Bodo Moeller]
4788 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4789 request to response.
4792 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4793 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4794 extract information from a certificate request. OCSP_response_create()
4795 creates a response and optionally adds a basic response structure.
4796 OCSP_basic_add1_status() adds a complete single response to a basic
4797 response and returns the OCSP_SINGLERESP structure just added (to allow
4798 extensions to be included for example). OCSP_basic_add1_cert() adds a
4799 certificate to a basic response and OCSP_basic_sign() signs a basic
4800 response with various flags. New helper functions ASN1_TIME_check()
4801 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4802 (converts ASN1_TIME to GeneralizedTime).
4805 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4806 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4807 structure from a certificate. X509_pubkey_digest() digests the public_key
4808 contents: this is used in various key identifiers.
4811 *) Make sk_sort() tolerate a NULL argument.
4812 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4814 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4815 passed by the function are trusted implicitly. If any of them signed the
4816 response then it is assumed to be valid and is not verified.
4819 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4820 to data. This was previously part of the PKCS7 ASN1 code. This
4821 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4822 [Steve Henson, reported by Kenneth R. Robinette
4823 <support@securenetterm.com>]
4825 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4826 routines: without these tracing memory leaks is very painful.
4827 Fix leaks in PKCS12 and PKCS7 routines.
4830 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4831 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4832 effectively meant GeneralizedTime would never be used. Now it
4833 is initialised to -1 but X509_time_adj() now has to check the value
4834 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4835 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4836 [Steve Henson, reported by Kenneth R. Robinette
4837 <support@securenetterm.com>]
4839 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4840 result in a zero length in the ASN1_INTEGER structure which was
4841 not consistent with the structure when d2i_ASN1_INTEGER() was used
4842 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4843 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4844 where it did not print out a minus for negative ASN1_INTEGER.
4847 *) Add summary printout to ocsp utility. The various functions which
4848 convert status values to strings have been renamed to:
4849 OCSP_response_status_str(), OCSP_cert_status_str() and
4850 OCSP_crl_reason_str() and are no longer static. New options
4851 to verify nonce values and to disable verification. OCSP response
4852 printout format cleaned up.
4855 *) Add additional OCSP certificate checks. These are those specified
4856 in RFC2560. This consists of two separate checks: the CA of the
4857 certificate being checked must either be the OCSP signer certificate
4858 or the issuer of the OCSP signer certificate. In the latter case the
4859 OCSP signer certificate must contain the OCSP signing extended key
4860 usage. This check is performed by attempting to match the OCSP
4861 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4862 in the OCSP_CERTID structures of the response.
4865 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4866 and related routines. This uses the standard OpenSSL certificate
4867 verify routines to perform initial checks (just CA validity) and
4868 to obtain the certificate chain. Then additional checks will be
4869 performed on the chain. Currently the root CA is checked to see
4870 if it is explicitly trusted for OCSP signing. This is used to set
4871 a root CA as a global signing root: that is any certificate that
4872 chains to that CA is an acceptable OCSP signing certificate.
4875 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4876 extensions from a separate configuration file.
4877 As when reading extensions from the main configuration file,
4878 the '-extensions ...' option may be used for specifying the
4880 [Massimiliano Pala <madwolf@comune.modena.it>]
4882 *) New OCSP utility. Allows OCSP requests to be generated or
4883 read. The request can be sent to a responder and the output
4884 parsed, outputed or printed in text form. Not complete yet:
4885 still needs to check the OCSP response validity.
4888 *) New subcommands for 'openssl ca':
4889 'openssl ca -status <serial>' prints the status of the cert with
4890 the given serial number (according to the index file).
4891 'openssl ca -updatedb' updates the expiry status of certificates
4893 [Massimiliano Pala <madwolf@comune.modena.it>]
4895 *) New '-newreq-nodes' command option to CA.pl. This is like
4896 '-newreq', but calls 'openssl req' with the '-nodes' option
4897 so that the resulting key is not encrypted.
4898 [Damien Miller <djm@mindrot.org>]
4900 *) New configuration for the GNU Hurd.
4901 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4903 *) Initial code to implement OCSP basic response verify. This
4904 is currently incomplete. Currently just finds the signer's
4905 certificate and verifies the signature on the response.
4908 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4909 value of OPENSSLDIR. This is available via the new '-d' option
4910 to 'openssl version', and is also included in 'openssl version -a'.
4913 *) Allowing defining memory allocation callbacks that will be given
4914 file name and line number information in additional arguments
4915 (a const char* and an int). The basic functionality remains, as
4916 well as the original possibility to just replace malloc(),
4917 realloc() and free() by functions that do not know about these
4918 additional arguments. To register and find out the current
4919 settings for extended allocation functions, the following
4920 functions are provided:
4922 CRYPTO_set_mem_ex_functions
4923 CRYPTO_set_locked_mem_ex_functions
4924 CRYPTO_get_mem_ex_functions
4925 CRYPTO_get_locked_mem_ex_functions
4927 These work the same way as CRYPTO_set_mem_functions and friends.
4928 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4929 extended allocation function is enabled.
4930 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4931 a conventional allocation function is enabled.
4932 [Richard Levitte, Bodo Moeller]
4934 *) Finish off removing the remaining LHASH function pointer casts.
4935 There should no longer be any prototype-casting required when using
4936 the LHASH abstraction, and any casts that remain are "bugs". See
4937 the callback types and macros at the head of lhash.h for details
4938 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4941 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4942 If /dev/[u]random devices are not available or do not return enough
4943 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4945 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4946 /etc/entropy will be queried once each in this sequence, quering stops
4947 when enough entropy was collected without querying more sockets.
4950 *) Change the Unix RAND_poll() variant to be able to poll several
4951 random devices, as specified by DEVRANDOM, until a sufficient amount
4952 of data has been collected. We spend at most 10 ms on each file
4953 (select timeout) and read in non-blocking mode. DEVRANDOM now
4954 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4955 (previously it was just the string "/dev/urandom"), so on typical
4956 platforms the 10 ms delay will never occur.
4957 Also separate out the Unix variant to its own file, rand_unix.c.
4958 For VMS, there's a currently-empty rand_vms.c.
4961 *) Move OCSP client related routines to ocsp_cl.c. These
4962 provide utility functions which an application needing
4963 to issue a request to an OCSP responder and analyse the
4964 response will typically need: as opposed to those which an
4965 OCSP responder itself would need which will be added later.
4967 OCSP_request_sign() signs an OCSP request with an API similar
4968 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4969 response. OCSP_response_get1_basic() extracts basic response
4970 from response. OCSP_resp_find_status(): finds and extracts status
4971 information from an OCSP_CERTID structure (which will be created
4972 when the request structure is built). These are built from lower
4973 level functions which work on OCSP_SINGLERESP structures but
4974 wont normally be used unless the application wishes to examine
4975 extensions in the OCSP response for example.
4977 Replace nonce routines with a pair of functions.
4978 OCSP_request_add1_nonce() adds a nonce value and optionally
4979 generates a random value. OCSP_check_nonce() checks the
4980 validity of the nonce in an OCSP response.
4983 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4984 This doesn't copy the supplied OCSP_CERTID and avoids the
4985 need to free up the newly created id. Change return type
4986 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4987 This can then be used to add extensions to the request.
4988 Deleted OCSP_request_new(), since most of its functionality
4989 is now in OCSP_REQUEST_new() (and the case insensitive name
4990 clash) apart from the ability to set the request name which
4991 will be added elsewhere.
4994 *) Update OCSP API. Remove obsolete extensions argument from
4995 various functions. Extensions are now handled using the new
4996 OCSP extension code. New simple OCSP HTTP function which
4997 can be used to send requests and parse the response.
5000 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5001 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5002 uses the special reorder version of SET OF to sort the attributes
5003 and reorder them to match the encoded order. This resolves a long
5004 standing problem: a verify on a PKCS7 structure just after signing
5005 it used to fail because the attribute order did not match the
5006 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5007 it uses the received order. This is necessary to tolerate some broken
5008 software that does not order SET OF. This is handled by encoding
5009 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5010 to produce the required SET OF.
5013 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5014 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5015 files to get correct declarations of the ASN.1 item variables.
5018 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5019 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5020 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5021 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5022 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5023 ASN1_ITEM and no wrapper functions.
5026 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5027 replace the old function pointer based I/O routines. Change most of
5028 the *_d2i_bio() and *_d2i_fp() functions to use these.
5031 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5032 lines, recognice more "algorithms" that can be deselected, and make
5033 it complain about algorithm deselection that isn't recognised.
5036 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5037 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5038 to use new functions. Add NO_ASN1_OLD which can be set to remove
5039 some old style ASN1 functions: this can be used to determine if old
5040 code will still work when these eventually go away.
5043 *) New extension functions for OCSP structures, these follow the
5044 same conventions as certificates and CRLs.
5047 *) New function X509V3_add1_i2d(). This automatically encodes and
5048 adds an extension. Its behaviour can be customised with various
5049 flags to append, replace or delete. Various wrappers added for
5050 certifcates and CRLs.
5053 *) Fix to avoid calling the underlying ASN1 print routine when
5054 an extension cannot be parsed. Correct a typo in the
5055 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5058 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5059 entries for variables.
5062 *) Add functionality to apps/openssl.c for detecting locking
5063 problems: As the program is single-threaded, all we have
5064 to do is register a locking callback using an array for
5065 storing which locks are currently held by the program.
5068 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5069 SSL_get_ex_data_X509_STORE_idx(), which is used in
5070 ssl_verify_cert_chain() and thus can be called at any time
5071 during TLS/SSL handshakes so that thread-safety is essential.
5072 Unfortunately, the ex_data design is not at all suited
5073 for multi-threaded use, so it probably should be abolished.
5076 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5077 [Broadcom, tweaked and integrated by Geoff Thorpe]
5079 *) Move common extension printing code to new function
5080 X509V3_print_extensions(). Reorganise OCSP print routines and
5081 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5084 *) New function X509_signature_print() to remove duplication in some
5088 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5089 set (this was treated exactly the same as SET OF previously). This
5090 is used to reorder the STACK representing the structure to match the
5091 encoding. This will be used to get round a problem where a PKCS7
5092 structure which was signed could not be verified because the STACK
5093 order did not reflect the encoded order.
5096 *) Reimplement the OCSP ASN1 module using the new code.
5099 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5100 for its ASN1 operations. The old style function pointers still exist
5101 for now but they will eventually go away.
5104 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5105 completely replaces the old ASN1 functionality with a table driven
5106 encoder and decoder which interprets an ASN1_ITEM structure describing
5107 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5108 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5109 has also been converted to the new form.
5112 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5113 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5114 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5115 for negative moduli.
5118 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5119 of not touching the result's sign bit.
5122 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5126 *) Changed the LHASH code to use prototypes for callbacks, and created
5127 macros to declare and implement thin (optionally static) functions
5128 that provide type-safety and avoid function pointer casting for the
5129 type-specific callbacks.
5132 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5134 [Veers Staats <staatsvr@asc.hpc.mil>,
5135 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5137 *) Reformat the FAQ so the different questions and answers can be divided
5138 in sections depending on the subject.
5141 *) Have the zlib compression code load ZLIB.DLL dynamically under
5145 *) New function BN_mod_sqrt for computing square roots modulo a prime
5146 (using the probabilistic Tonelli-Shanks algorithm unless
5147 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5148 be handled deterministically).
5149 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5151 *) Make BN_mod_inverse faster by explicitly handling small quotients
5152 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5153 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5156 *) New function BN_kronecker.
5159 *) Fix BN_gcd so that it works on negative inputs; the result is
5160 positive unless both parameters are zero.
5161 Previously something reasonably close to an infinite loop was
5162 possible because numbers could be growing instead of shrinking
5163 in the implementation of Euclid's algorithm.
5166 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5167 sign of the number in question.
5169 Fix BN_is_word(a,w) to work correctly for w == 0.
5171 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5172 because its test if the absolute value of 'a' equals 'w'.
5173 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5174 it exists mostly for use in the implementations of BN_is_zero(),
5175 BN_is_one(), and BN_is_word().
5178 *) New function BN_swap.
5181 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5182 the exponentiation functions are more likely to produce reasonable
5183 results on negative inputs.
5186 *) Change BN_mod_mul so that the result is always non-negative.
5187 Previously, it could be negative if one of the factors was negative;
5188 I don't think anyone really wanted that behaviour.
5191 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5192 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5193 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5194 and add new functions:
5203 BN_mod_lshift1_quick
5207 These functions always generate non-negative results.
5209 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5210 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5212 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5213 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5214 be reduced modulo m.
5215 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5218 The following entry accidentily appeared in the CHANGES file
5219 distributed with OpenSSL 0.9.7. The modifications described in
5220 it do *not* apply to OpenSSL 0.9.7.
5222 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5223 was actually never needed) and in BN_mul(). The removal in BN_mul()
5224 required a small change in bn_mul_part_recursive() and the addition
5225 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5226 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5227 bn_sub_words() and bn_add_words() except they take arrays with
5232 *) In 'openssl passwd', verify passwords read from the terminal
5233 unless the '-salt' option is used (which usually means that
5234 verification would just waste user's time since the resulting
5235 hash is going to be compared with some given password hash)
5236 or the new '-noverify' option is used.
5238 This is an incompatible change, but it does not affect
5239 non-interactive use of 'openssl passwd' (passwords on the command
5240 line, '-stdin' option, '-in ...' option) and thus should not
5244 *) Remove all references to RSAref, since there's no more need for it.
5247 *) Make DSO load along a path given through an environment variable
5248 (SHLIB_PATH) with shl_load().
5251 *) Constify the ENGINE code as a result of BIGNUM constification.
5252 Also constify the RSA code and most things related to it. In a
5253 few places, most notable in the depth of the ASN.1 code, ugly
5254 casts back to non-const were required (to be solved at a later
5258 *) Make it so the openssl application has all engines loaded by default.
5261 *) Constify the BIGNUM routines a little more.
5264 *) Add the following functions:
5266 ENGINE_load_cswift()
5268 ENGINE_load_atalla()
5270 ENGINE_load_builtin_engines()
5272 That way, an application can itself choose if external engines that
5273 are built-in in OpenSSL shall ever be used or not. The benefit is
5274 that applications won't have to be linked with libdl or other dso
5275 libraries unless it's really needed.
5277 Changed 'openssl engine' to load all engines on demand.
5278 Changed the engine header files to avoid the duplication of some
5279 declarations (they differed!).
5282 *) 'openssl engine' can now list capabilities.
5285 *) Better error reporting in 'openssl engine'.
5288 *) Never call load_dh_param(NULL) in s_server.
5291 *) Add engine application. It can currently list engines by name and
5292 identity, and test if they are actually available.
5295 *) Improve RPM specification file by forcing symbolic linking and making
5296 sure the installed documentation is also owned by root.root.
5297 [Damien Miller <djm@mindrot.org>]
5299 *) Give the OpenSSL applications more possibilities to make use of
5300 keys (public as well as private) handled by engines.
5303 *) Add OCSP code that comes from CertCo.
5306 *) Add VMS support for the Rijndael code.
5309 *) Added untested support for Nuron crypto accelerator.
5312 *) Add support for external cryptographic devices. This code was
5313 previously distributed separately as the "engine" branch.
5314 [Geoff Thorpe, Richard Levitte]
5316 *) Rework the filename-translation in the DSO code. It is now possible to
5317 have far greater control over how a "name" is turned into a filename
5318 depending on the operating environment and any oddities about the
5319 different shared library filenames on each system.
5322 *) Support threads on FreeBSD-elf in Configure.
5325 *) Fix for SHA1 assembly problem with MASM: it produces
5326 warnings about corrupt line number information when assembling
5327 with debugging information. This is caused by the overlapping
5329 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5332 NCONF_get_number() has no error checking at all. As a replacement,
5333 NCONF_get_number_e() is defined (_e for "error checking") and is
5334 promoted strongly. The old NCONF_get_number is kept around for
5335 binary backward compatibility.
5336 Make it possible for methods to load from something other than a BIO,
5337 by providing a function pointer that is given a name instead of a BIO.
5338 For example, this could be used to load configuration data from an
5342 *) Fix for non blocking accept BIOs. Added new I/O special reason
5343 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5344 with non blocking I/O was not possible because no retry code was
5345 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5349 *) Added the beginnings of Rijndael support.
5352 *) Fix for bug in DirectoryString mask setting. Add support for
5353 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5354 to allow certificate printing to more controllable, additional
5355 'certopt' option to 'x509' to allow new printing options to be
5359 *) Clean old EAY MD5 hack from e_os.h.
5362 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5364 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5365 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5366 [Joe Orton, Steve Henson]
5368 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5370 *) Fix additional bug revealed by the NISCC test suite:
5372 Stop bug triggering large recursion when presented with
5373 certain ASN.1 tags (CVE-2003-0851)
5376 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5378 *) Fix various bugs revealed by running the NISCC test suite:
5380 Stop out of bounds reads in the ASN1 code when presented with
5381 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5383 If verify callback ignores invalid public key errors don't try to check
5384 certificate signature with the NULL public key.
5388 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5389 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5393 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5394 extra data after the compression methods not only for TLS 1.0
5395 but also for SSL 3.0 (as required by the specification).
5396 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5398 *) Change X509_certificate_type() to mark the key as exported/exportable
5399 when it's 512 *bits* long, not 512 bytes.
5402 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5404 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5405 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5406 a protocol version number mismatch like a decryption error
5407 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5410 *) Turn on RSA blinding by default in the default implementation
5411 to avoid a timing attack. Applications that don't want it can call
5412 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5413 They would be ill-advised to do so in most cases.
5414 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5416 *) Change RSA blinding code so that it works when the PRNG is not
5417 seeded (in this case, the secret RSA exponent is abused as
5418 an unpredictable seed -- if it is not unpredictable, there
5419 is no point in blinding anyway). Make RSA blinding thread-safe
5420 by remembering the creator's thread ID in rsa->blinding and
5421 having all other threads use local one-time blinding factors
5422 (this requires more computation than sharing rsa->blinding, but
5423 avoids excessive locking; and if an RSA object is not shared
5424 between threads, blinding will still be very fast).
5427 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5429 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5430 via timing by performing a MAC computation even if incorrrect
5431 block cipher padding has been found. This is a countermeasure
5432 against active attacks where the attacker has to distinguish
5433 between bad padding and a MAC verification error. (CVE-2003-0078)
5435 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5436 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5437 Martin Vuagnoux (EPFL, Ilion)]
5439 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5441 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5442 memory from it's contents. This is done with a counter that will
5443 place alternating values in each byte. This can be used to solve
5444 two issues: 1) the removal of calls to memset() by highly optimizing
5445 compilers, and 2) cleansing with other values than 0, since those can
5446 be read through on certain media, for example a swap space on disk.
5449 *) Bugfix: client side session caching did not work with external caching,
5450 because the session->cipher setting was not restored when reloading
5451 from the external cache. This problem was masked, when
5452 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5453 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5456 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5457 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5458 [Zeev Lieber <zeev-l@yahoo.com>]
5460 *) Undo an undocumented change introduced in 0.9.6e which caused
5461 repeated calls to OpenSSL_add_all_ciphers() and
5462 OpenSSL_add_all_digests() to be ignored, even after calling
5466 *) Change the default configuration reader to deal with last line not
5467 being properly terminated.
5470 *) Change X509_NAME_cmp() so it applies the special rules on handling
5471 DN values that are of type PrintableString, as well as RDNs of type
5472 emailAddress where the value has the type ia5String.
5473 [stefank@valicert.com via Richard Levitte]
5475 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5476 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5477 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5478 the bitwise-OR of the two for use by the majority of applications
5479 wanting this behaviour, and update the docs. The documented
5480 behaviour and actual behaviour were inconsistent and had been
5481 changing anyway, so this is more a bug-fix than a behavioural
5483 [Geoff Thorpe, diagnosed by Nadav Har'El]
5485 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5486 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5489 *) Fix initialization code race conditions in
5490 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5491 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5492 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5493 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5494 ssl2_get_cipher_by_char(),
5495 ssl3_get_cipher_by_char().
5496 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5498 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5499 the cached sessions are flushed, as the remove_cb() might use ex_data
5500 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5501 (see [openssl.org #212]).
5502 [Geoff Thorpe, Lutz Jaenicke]
5504 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5505 length, instead of the encoding length to d2i_ASN1_OBJECT.
5508 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5510 *) [In 0.9.6g-engine release:]
5511 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5512 [Lynn Gazis <lgazis@rainbow.com>]
5514 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5516 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5517 and get fix the header length calculation.
5518 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5519 Alon Kantor <alonk@checkpoint.com> (and others),
5522 *) Use proper error handling instead of 'assertions' in buffer
5523 overflow checks added in 0.9.6e. This prevents DoS (the
5524 assertions could call abort()).
5525 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5527 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5529 *) Add various sanity checks to asn1_get_length() to reject
5530 the ASN1 length bytes if they exceed sizeof(long), will appear
5531 negative or the content length exceeds the length of the
5533 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5535 *) Fix cipher selection routines: ciphers without encryption had no flags
5536 for the cipher strength set and where therefore not handled correctly
5537 by the selection routines (PR #130).
5540 *) Fix EVP_dsa_sha macro.
5544 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5545 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5546 that was added in OpenSSL 0.9.6d.
5548 As the countermeasure turned out to be incompatible with some
5549 broken SSL implementations, the new option is part of SSL_OP_ALL.
5550 SSL_OP_ALL is usually employed when compatibility with weird SSL
5551 implementations is desired (e.g. '-bugs' option to 's_client' and
5552 's_server'), so the new option is automatically set in many
5556 *) Changes in security patch:
5558 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5559 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5560 Air Force Materiel Command, USAF, under agreement number
5563 *) Add various sanity checks to asn1_get_length() to reject
5564 the ASN1 length bytes if they exceed sizeof(long), will appear
5565 negative or the content length exceeds the length of the
5566 supplied buffer. (CVE-2002-0659)
5567 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5569 *) Assertions for various potential buffer overflows, not known to
5571 [Ben Laurie (CHATS)]
5573 *) Various temporary buffers to hold ASCII versions of integers were
5574 too small for 64 bit platforms. (CVE-2002-0655)
5575 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5577 *) Remote buffer overflow in SSL3 protocol - an attacker could
5578 supply an oversized session ID to a client. (CVE-2002-0656)
5579 [Ben Laurie (CHATS)]
5581 *) Remote buffer overflow in SSL2 protocol - an attacker could
5582 supply an oversized client master key. (CVE-2002-0656)
5583 [Ben Laurie (CHATS)]
5585 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5587 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5588 encoded as NULL) with id-dsa-with-sha1.
5589 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5591 *) Check various X509_...() return values in apps/req.c.
5592 [Nils Larsch <nla@trustcenter.de>]
5594 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5595 an end-of-file condition would erronously be flagged, when the CRLF
5596 was just at the end of a processed block. The bug was discovered when
5597 processing data through a buffering memory BIO handing the data to a
5598 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5599 <ptsekov@syntrex.com> and Nedelcho Stanev.
5602 *) Implement a countermeasure against a vulnerability recently found
5603 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5604 before application data chunks to avoid the use of known IVs
5605 with data potentially chosen by the attacker.
5608 *) Fix length checks in ssl3_get_client_hello().
5611 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5612 to prevent ssl3_read_internal() from incorrectly assuming that
5613 ssl3_read_bytes() found application data while handshake
5614 processing was enabled when in fact s->s3->in_read_app_data was
5615 merely automatically cleared during the initial handshake.
5616 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5618 *) Fix object definitions for Private and Enterprise: they were not
5619 recognized in their shortname (=lowercase) representation. Extend
5620 obj_dat.pl to issue an error when using undefined keywords instead
5621 of silently ignoring the problem (Svenning Sorensen
5622 <sss@sss.dnsalias.net>).
5625 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5626 generators, i.e. generators other than 2 and 5. (Previously, the
5627 code did not properly initialise the 'add' and 'rem' values to
5628 BN_generate_prime().)
5630 In the new general case, we do not insist that 'generator' is
5631 actually a primitive root: This requirement is rather pointless;
5632 a generator of the order-q subgroup is just as good, if not
5636 *) Map new X509 verification errors to alerts. Discovered and submitted by
5637 Tom Wu <tom@arcot.com>.
5640 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5641 returning non-zero before the data has been completely received
5642 when using non-blocking I/O.
5643 [Bodo Moeller; problem pointed out by John Hughes]
5645 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5646 [Ben Laurie, Lutz Jaenicke]
5648 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5649 Yoram Zahavi <YoramZ@gilian.com>).
5652 *) Add information about CygWin 1.3 and on, and preserve proper
5653 configuration for the versions before that.
5654 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5656 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5657 check whether we deal with a copy of a session and do not delete from
5658 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5659 <izhar@checkpoint.com>.
5662 *) Do not store session data into the internal session cache, if it
5663 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5664 flag is set). Proposed by Aslam <aslam@funk.com>.
5667 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5671 *) [In 0.9.6d-engine release:]
5672 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5673 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5675 *) Add the configuration target linux-s390x.
5676 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5678 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5679 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5680 variable as an indication that a ClientHello message has been
5681 received. As the flag value will be lost between multiple
5682 invocations of ssl3_accept when using non-blocking I/O, the
5683 function may not be aware that a handshake has actually taken
5684 place, thus preventing a new session from being added to the
5687 To avoid this problem, we now set s->new_session to 2 instead of
5688 using a local variable.
5689 [Lutz Jaenicke, Bodo Moeller]
5691 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5692 if the SSL_R_LENGTH_MISMATCH error is detected.
5693 [Geoff Thorpe, Bodo Moeller]
5695 *) New 'shared_ldflag' column in Configure platform table.
5698 *) Fix EVP_CIPHER_mode macro.
5699 ["Dan S. Camper" <dan@bti.net>]
5701 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5702 type, we must throw them away by setting rr->length to 0.
5703 [D P Chang <dpc@qualys.com>]
5705 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5707 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5708 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5709 worked incorrectly for those cases where range = 10..._2 and
5710 3*range is two bits longer than range.)
5713 *) Only add signing time to PKCS7 structures if it is not already
5717 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5718 OBJ_ld_ce should be OBJ_id_ce.
5719 Also some ip-pda OIDs in crypto/objects/objects.txt were
5720 incorrect (cf. RFC 3039).
5721 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5723 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5724 returns early because it has nothing to do.
5725 [Andy Schneider <andy.schneider@bjss.co.uk>]
5727 *) [In 0.9.6c-engine release:]
5728 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5729 [Andy Schneider <andy.schneider@bjss.co.uk>]
5731 *) [In 0.9.6c-engine release:]
5732 Add support for Cryptographic Appliance's keyserver technology.
5733 (Use engine 'keyclient')
5734 [Cryptographic Appliances and Geoff Thorpe]
5736 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5737 is called via tools/c89.sh because arguments have to be
5738 rearranged (all '-L' options must appear before the first object
5740 [Richard Shapiro <rshapiro@abinitio.com>]
5742 *) [In 0.9.6c-engine release:]
5743 Add support for Broadcom crypto accelerator cards, backported
5745 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5747 *) [In 0.9.6c-engine release:]
5748 Add support for SureWare crypto accelerator cards from
5749 Baltimore Technologies. (Use engine 'sureware')
5750 [Baltimore Technologies and Mark Cox]
5752 *) [In 0.9.6c-engine release:]
5753 Add support for crypto accelerator cards from Accelerated
5754 Encryption Processing, www.aep.ie. (Use engine 'aep')
5755 [AEP Inc. and Mark Cox]
5757 *) Add a configuration entry for gcc on UnixWare.
5758 [Gary Benson <gbenson@redhat.com>]
5760 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5761 messages are stored in a single piece (fixed-length part and
5762 variable-length part combined) and fix various bugs found on the way.
5765 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5766 instead. BIO_gethostbyname() does not know what timeouts are
5767 appropriate, so entries would stay in cache even when they have
5769 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5771 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5772 faced with a pathologically small ClientHello fragment that does
5773 not contain client_version: Instead of aborting with an error,
5774 simply choose the highest available protocol version (i.e.,
5775 TLS 1.0 unless it is disabled). In practice, ClientHello
5776 messages are never sent like this, but this change gives us
5777 strictly correct behaviour at least for TLS.
5780 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5781 never resets s->method to s->ctx->method when called from within
5782 one of the SSL handshake functions.
5783 [Bodo Moeller; problem pointed out by Niko Baric]
5785 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5786 (sent using the client's version number) if client_version is
5787 smaller than the protocol version in use. Also change
5788 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5789 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5790 the client will at least see that alert.
5793 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5797 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5798 client receives HelloRequest while in a handshake.
5799 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5801 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5802 should end in 'break', not 'goto end' which circuments various
5803 cleanups done in state SSL_ST_OK. But session related stuff
5804 must be disabled for SSL_ST_OK in the case that we just sent a
5807 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5808 before just sending a HelloRequest.
5809 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5811 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5812 reveal whether illegal block cipher padding was found or a MAC
5813 verification error occured. (Neither SSLerr() codes nor alerts
5814 are directly visible to potential attackers, but the information
5815 may leak via logfiles.)
5817 Similar changes are not required for the SSL 2.0 implementation
5818 because the number of padding bytes is sent in clear for SSL 2.0,
5819 and the extra bytes are just ignored. However ssl/s2_pkt.c
5820 failed to verify that the purported number of padding bytes is in
5824 *) Add OpenUNIX-8 support including shared libraries
5825 (Boyd Lynn Gerber <gerberb@zenez.com>).
5828 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5829 'wristwatch attack' using huge encoding parameters (cf.
5830 James H. Manger's CRYPTO 2001 paper). Note that the
5831 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5832 encoding parameters and hence was not vulnerable.
5835 *) BN_sqr() bug fix.
5836 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5838 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5839 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5840 followed by modular reduction.
5841 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5843 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5844 equivalent based on BN_pseudo_rand() instead of BN_rand().
5847 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5848 This function was broken, as the check for a new client hello message
5849 to handle SGC did not allow these large messages.
5850 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5853 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5856 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5857 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5860 *) Rework the configuration and shared library support for Tru64 Unix.
5861 The configuration part makes use of modern compiler features and
5862 still retains old compiler behavior for those that run older versions
5863 of the OS. The shared library support part includes a variant that
5864 uses the RPATH feature, and is available through the special
5865 configuration target "alpha-cc-rpath", which will never be selected
5867 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5869 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5870 with the same message size as in ssl3_get_certificate_request().
5871 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5872 messages might inadvertently be reject as too long.
5873 [Petr Lampa <lampa@fee.vutbr.cz>]
5875 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5878 *) Modified SSL library such that the verify_callback that has been set
5879 specificly for an SSL object with SSL_set_verify() is actually being
5880 used. Before the change, a verify_callback set with this function was
5881 ignored and the verify_callback() set in the SSL_CTX at the time of
5882 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5883 to allow the necessary settings.
5886 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5887 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5888 done automatically (in contradiction to the requirements of the C
5889 standard). This made problems when used from OpenSSH.
5892 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5893 dh->length and always used
5895 BN_rand_range(priv_key, dh->p).
5897 BN_rand_range() is not necessary for Diffie-Hellman, and this
5898 specific range makes Diffie-Hellman unnecessarily inefficient if
5899 dh->length (recommended exponent length) is much smaller than the
5900 length of dh->p. We could use BN_rand_range() if the order of
5901 the subgroup was stored in the DH structure, but we only have
5906 BN_rand(priv_key, l, ...)
5908 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5914 RSA_eay_public_encrypt
5915 RSA_eay_private_decrypt
5916 RSA_eay_private_encrypt (signing)
5917 RSA_eay_public_decrypt (signature verification)
5919 (default implementations for RSA_public_encrypt,
5920 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5921 always reject numbers >= n.
5924 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5925 to synchronize access to 'locking_thread'. This is necessary on
5926 systems where access to 'locking_thread' (an 'unsigned long'
5927 variable) is not atomic.
5930 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5931 *before* setting the 'crypto_lock_rand' flag. The previous code had
5932 a race condition if 0 is a valid thread ID.
5933 [Travis Vitek <vitek@roguewave.com>]
5935 *) Add support for shared libraries under Irix.
5936 [Albert Chin-A-Young <china@thewrittenword.com>]
5938 *) Add configuration option to build on Linux on both big-endian and
5940 [Ralf Baechle <ralf@uni-koblenz.de>]
5942 *) Add the possibility to create shared libraries on HP-UX.
5945 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5947 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5948 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5949 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5950 PRNG state recovery was possible based on the output of
5951 one PRNG request appropriately sized to gain knowledge on
5952 'md' followed by enough consecutive 1-byte PRNG requests
5953 to traverse all of 'state'.
5955 1. When updating 'md_local' (the current thread's copy of 'md')
5956 during PRNG output generation, hash all of the previous
5957 'md_local' value, not just the half used for PRNG output.
5959 2. Make the number of bytes from 'state' included into the hash
5960 independent from the number of PRNG bytes requested.
5962 The first measure alone would be sufficient to avoid
5963 Markku-Juhani's attack. (Actually it had never occurred
5964 to me that the half of 'md_local' used for chaining was the
5965 half from which PRNG output bytes were taken -- I had always
5966 assumed that the secret half would be used.) The second
5967 measure makes sure that additional data from 'state' is never
5968 mixed into 'md_local' in small portions; this heuristically
5969 further strengthens the PRNG.
5972 *) Fix crypto/bn/asm/mips3.s.
5975 *) When only the key is given to "enc", the IV is undefined. Print out
5976 an error message in this case.
5979 *) Handle special case when X509_NAME is empty in X509 printing routines.
5982 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5983 positive and less than q.
5986 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5987 used: it isn't thread safe and the add_lock_callback should handle
5989 [Paul Rose <Paul.Rose@bridge.com>]
5991 *) Verify that incoming data obeys the block size in
5992 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5996 [Ulf Möller, Bodo Möller]
5998 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5999 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6000 when fixing the server behaviour for backwards-compatible 'client
6001 hello' messages. (Note that the attack is impractical against
6002 SSL 3.0 and TLS 1.0 anyway because length and version checking
6003 means that the probability of guessing a valid ciphertext is
6004 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6007 Before 0.9.5, the countermeasure (hide the error by generating a
6008 random 'decryption result') did not work properly because
6009 ERR_clear_error() was missing, meaning that SSL_get_error() would
6010 detect the supposedly ignored error.
6012 Both problems are now fixed.
6015 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6016 (previously it was 1024).
6019 *) Fix for compatibility mode trust settings: ignore trust settings
6020 unless some valid trust or reject settings are present.
6023 *) Fix for blowfish EVP: its a variable length cipher.
6026 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6027 parameters in DSA public key structures and return an error in the
6028 DSA routines if parameters are absent.
6031 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6032 in the current directory if neither $RANDFILE nor $HOME was set.
6033 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6034 caused some confusion to Windows users who haven't defined $HOME.
6035 Thus RAND_file_name() is changed again: e_os.h can define a
6036 DEFAULT_HOME, which will be used if $HOME is not set.
6037 For Windows, we use "C:"; on other platforms, we still require
6038 environment variables.
6040 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6041 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6042 having multiple threads call RAND_poll() concurrently.
6045 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6046 combination of a flag and a thread ID variable.
6047 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6048 flag), *other* threads can enter ssleay_add_bytes without obeying
6049 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6050 that they do not hold after the first thread unsets add_do_not_lock).
6053 *) Change bctest again: '-x' expressions are not available in all
6057 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6059 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6060 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6062 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6063 the default extension for executables, if any. Also, make the perl
6064 scripts that use symlink() to test if it really exists and use "cp"
6065 if it doesn't. All this made OpenSSL compilable and installable in
6069 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6070 If SEQUENCE is length is indefinite just set c->slen to the total
6071 amount of data available.
6072 [Steve Henson, reported by shige@FreeBSD.org]
6073 [This change does not apply to 0.9.7.]
6075 *) Change bctest to avoid here-documents inside command substitution
6076 (workaround for FreeBSD /bin/sh bug).
6077 For compatibility with Ultrix, avoid shell functions (introduced
6078 in the bctest version that searches along $PATH).
6081 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6082 with des_encrypt() defined on some operating systems, like Solaris
6086 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6087 On the Importance of Eliminating Errors in Cryptographic
6088 Computations, J. Cryptology 14 (2001) 2, 101-119,
6089 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6092 *) MIPS assembler BIGNUM division bug fix.
6095 *) Disabled incorrect Alpha assembler code.
6098 *) Fix PKCS#7 decode routines so they correctly update the length
6099 after reading an EOC for the EXPLICIT tag.
6101 [This change does not apply to 0.9.7.]
6103 *) Fix bug in PKCS#12 key generation routines. This was triggered
6104 if a 3DES key was generated with a 0 initial byte. Include
6105 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6106 (but broken) behaviour.
6109 *) Enhance bctest to search for a working bc along $PATH and print
6111 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6113 *) Fix memory leaks in err.c: free err_data string if necessary;
6114 don't write to the wrong index in ERR_set_error_data.
6117 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6121 *) Replace rdtsc with _emit statements for VC++ version 5.
6122 [Jeremy Cooper <jeremy@baymoo.org>]
6124 *) Make it possible to reuse SSLv2 sessions.
6127 *) In copy_email() check for >= 0 as a return value for
6128 X509_NAME_get_index_by_NID() since 0 is a valid index.
6129 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6131 *) Avoid coredump with unsupported or invalid public keys by checking if
6132 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6133 PKCS7_verify() fails with non detached data.
6136 *) Don't use getenv in library functions when run as setuid/setgid.
6137 New function OPENSSL_issetugid().
6140 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6141 due to incorrect handling of multi-threading:
6143 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6145 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6147 3. Count how many times MemCheck_off() has been called so that
6148 nested use can be treated correctly. This also avoids
6149 inband-signalling in the previous code (which relied on the
6150 assumption that thread ID 0 is impossible).
6153 *) Add "-rand" option also to s_client and s_server.
6156 *) Fix CPU detection on Irix 6.x.
6157 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6158 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6160 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6163 [This change does not apply to 0.9.7.]
6165 *) Use the cached encoding of an X509_NAME structure rather than
6166 copying it. This is apparently the reason for the libsafe "errors"
6167 but the code is actually correct.
6170 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6171 Bleichenbacher's DSA attack.
6172 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6173 to be set and top=0 forces the highest bit to be set; top=-1 is new
6174 and leaves the highest bit random.
6175 [Ulf Moeller, Bodo Moeller]
6177 *) In the NCONF_...-based implementations for CONF_... queries
6178 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6179 a temporary CONF structure with the data component set to NULL
6180 (which gives segmentation faults in lh_retrieve).
6181 Instead, use NULL for the CONF pointer in CONF_get_string and
6182 CONF_get_number (which may use environment variables) and directly
6183 return NULL from CONF_get_section.
6186 *) Fix potential buffer overrun for EBCDIC.
6189 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6190 keyUsage if basicConstraints absent for a CA.
6193 *) Make SMIME_write_PKCS7() write mail header values with a format that
6194 is more generally accepted (no spaces before the semicolon), since
6195 some programs can't parse those values properly otherwise. Also make
6196 sure BIO's that break lines after each write do not create invalid
6200 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6201 macros previously used would not encode an empty SEQUENCE OF
6202 and break the signature.
6204 [This change does not apply to 0.9.7.]
6206 *) Zero the premaster secret after deriving the master secret in
6210 *) Add some EVP_add_digest_alias registrations (as found in
6211 OpenSSL_add_all_digests()) to SSL_library_init()
6212 aka OpenSSL_add_ssl_algorithms(). This provides improved
6213 compatibility with peers using X.509 certificates
6214 with unconventional AlgorithmIdentifier OIDs.
6217 *) Fix for Irix with NO_ASM.
6218 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6220 *) ./config script fixes.
6221 [Ulf Moeller, Richard Levitte]
6223 *) Fix 'openssl passwd -1'.
6226 *) Change PKCS12_key_gen_asc() so it can cope with non null
6227 terminated strings whose length is passed in the passlen
6228 parameter, for example from PEM callbacks. This was done
6229 by adding an extra length parameter to asc2uni().
6230 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6232 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6233 call failed, free the DSA structure.
6236 *) Fix to uni2asc() to cope with zero length Unicode strings.
6237 These are present in some PKCS#12 files.
6240 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6241 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6242 when writing a 32767 byte record.
6243 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6245 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6246 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6248 (RSA objects have a reference count access to which is protected
6249 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6250 so they are meant to be shared between threads.)
6251 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6252 "Reddie, Steven" <Steven.Reddie@ca.com>]
6254 *) Fix a deadlock in CRYPTO_mem_leaks().
6257 *) Use better test patterns in bntest.
6260 *) rand_win.c fix for Borland C.
6263 *) BN_rshift bugfix for n == 0.
6266 *) Add a 'bctest' script that checks for some known 'bc' bugs
6267 so that 'make test' does not abort just because 'bc' is broken.
6270 *) Store verify_result within SSL_SESSION also for client side to
6271 avoid potential security hole. (Re-used sessions on the client side
6272 always resulted in verify_result==X509_V_OK, not using the original
6273 result of the server certificate verification.)
6276 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6277 SSL3_RT_APPLICATION_DATA, return 0.
6278 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6282 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6283 releases, have been re-implemented by renaming the previous
6284 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6285 and ssl3_read_internal, respectively, and adding 'peek' parameters
6286 to them. The new ssl[23]_{read,peek} functions are calls to
6287 ssl[23]_read_internal with the 'peek' flag set appropriately.
6288 A 'peek' parameter has also been added to ssl3_read_bytes, which
6289 does the actual work for ssl3_read_internal.
6292 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6293 the method-specific "init()" handler. Also clean up ex_data after
6294 calling the method-specific "finish()" handler. Previously, this was
6295 happening the other way round.
6298 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6299 The previous value, 12, was not always sufficient for BN_mod_exp().
6302 *) Make sure that shared libraries get the internal name engine with
6303 the full version number and not just 0. This should mark the
6304 shared libraries as not backward compatible. Of course, this should
6305 be changed again when we can guarantee backward binary compatibility.
6308 *) Fix typo in get_cert_by_subject() in by_dir.c
6309 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6311 *) Rework the system to generate shared libraries:
6313 - Make note of the expected extension for the shared libraries and
6314 if there is a need for symbolic links from for example libcrypto.so.0
6315 to libcrypto.so.0.9.7. There is extended info in Configure for
6318 - Make as few rebuilds of the shared libraries as possible.
6320 - Still avoid linking the OpenSSL programs with the shared libraries.
6322 - When installing, install the shared libraries separately from the
6326 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6328 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6329 and not in SSL_clear because the latter is also used by the
6330 accept/connect functions; previously, the settings made by
6331 SSL_set_read_ahead would be lost during the handshake.
6332 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6334 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6335 Previously, it would create entries for disableed algorithms no
6339 *) Added several new manual pages for SSL_* function.
6342 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6344 *) In ssl23_get_client_hello, generate an error message when faced
6345 with an initial SSL 3.0/TLS record that is too small to contain the
6346 first two bytes of the ClientHello message, i.e. client_version.
6347 (Note that this is a pathologic case that probably has never happened
6348 in real life.) The previous approach was to use the version number
6349 from the record header as a substitute; but our protocol choice
6350 should not depend on that one because it is not authenticated
6351 by the Finished messages.
6354 *) More robust randomness gathering functions for Windows.
6355 [Jeffrey Altman <jaltman@columbia.edu>]
6357 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6358 not set then we don't setup the error code for issuer check errors
6359 to avoid possibly overwriting other errors which the callback does
6360 handle. If an application does set the flag then we assume it knows
6361 what it is doing and can handle the new informational codes
6365 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6366 a general "ANY" type, as such it should be able to decode anything
6367 including tagged types. However it didn't check the class so it would
6368 wrongly interpret tagged types in the same way as their universal
6369 counterpart and unknown types were just rejected. Changed so that the
6370 tagged and unknown types are handled in the same way as a SEQUENCE:
6371 that is the encoding is stored intact. There is also a new type
6372 "V_ASN1_OTHER" which is used when the class is not universal, in this
6373 case we have no idea what the actual type is so we just lump them all
6377 *) On VMS, stdout may very well lead to a file that is written to
6378 in a record-oriented fashion. That means that every write() will
6379 write a separate record, which will be read separately by the
6380 programs trying to read from it. This can be very confusing.
6382 The solution is to put a BIO filter in the way that will buffer
6383 text until a linefeed is reached, and then write everything a
6384 line at a time, so every record written will be an actual line,
6385 not chunks of lines and not (usually doesn't happen, but I've
6386 seen it once) several lines in one record. BIO_f_linebuffer() is
6389 Currently, it's a VMS-only method, because that's where it has
6390 been tested well enough.
6393 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6394 it can return incorrect results.
6395 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6396 but it was in 0.9.6-beta[12].)
6399 *) Disable the check for content being present when verifying detached
6400 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6401 include zero length content when signing messages.
6404 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6405 BIO_ctrl (for BIO pairs).
6408 *) Add DSO method for VMS.
6411 *) Bug fix: Montgomery multiplication could produce results with the
6415 *) Add RPM specification openssl.spec and modify it to build three
6416 packages. The default package contains applications, application
6417 documentation and run-time libraries. The devel package contains
6418 include files, static libraries and function documentation. The
6419 doc package contains the contents of the doc directory. The original
6420 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6423 *) Add a large number of documentation files for many SSL routines.
6424 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6426 *) Add a configuration entry for Sony News 4.
6427 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6429 *) Don't set the two most significant bits to one when generating a
6430 random number < q in the DSA library.
6433 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6434 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6435 the underlying transport is blocking) if a handshake took place.
6436 (The default behaviour is needed by applications such as s_client
6437 and s_server that use select() to determine when to use SSL_read;
6438 but for applications that know in advance when to expect data, it
6439 just makes things more complicated.)
6442 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6446 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6447 work better on such systems.
6448 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6450 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6451 Update PKCS12_parse() so it copies the friendlyName and the
6452 keyid to the certificates aux info.
6455 *) Fix bug in PKCS7_verify() which caused an infinite loop
6456 if there was more than one signature.
6457 [Sven Uszpelkat <su@celocom.de>]
6459 *) Major change in util/mkdef.pl to include extra information
6460 about each symbol, as well as presentig variables as well
6461 as functions. This change means that there's n more need
6462 to rebuild the .num files when some algorithms are excluded.
6465 *) Allow the verify time to be set by an application,
6466 rather than always using the current time.
6469 *) Phase 2 verify code reorganisation. The certificate
6470 verify code now looks up an issuer certificate by a
6471 number of criteria: subject name, authority key id
6472 and key usage. It also verifies self signed certificates
6473 by the same criteria. The main comparison function is
6474 X509_check_issued() which performs these checks.
6476 Lot of changes were necessary in order to support this
6477 without completely rewriting the lookup code.
6479 Authority and subject key identifier are now cached.
6481 The LHASH 'certs' is X509_STORE has now been replaced
6482 by a STACK_OF(X509_OBJECT). This is mainly because an
6483 LHASH can't store or retrieve multiple objects with
6484 the same hash value.
6486 As a result various functions (which were all internal
6487 use only) have changed to handle the new X509_STORE
6488 structure. This will break anything that messed round
6489 with X509_STORE internally.
6491 The functions X509_STORE_add_cert() now checks for an
6492 exact match, rather than just subject name.
6494 The X509_STORE API doesn't directly support the retrieval
6495 of multiple certificates matching a given criteria, however
6496 this can be worked round by performing a lookup first
6497 (which will fill the cache with candidate certificates)
6498 and then examining the cache for matches. This is probably
6499 the best we can do without throwing out X509_LOOKUP
6500 entirely (maybe later...).
6502 The X509_VERIFY_CTX structure has been enhanced considerably.
6504 All certificate lookup operations now go via a get_issuer()
6505 callback. Although this currently uses an X509_STORE it
6506 can be replaced by custom lookups. This is a simple way
6507 to bypass the X509_STORE hackery necessary to make this
6508 work and makes it possible to use more efficient techniques
6509 in future. A very simple version which uses a simple
6510 STACK for its trusted certificate store is also provided
6511 using X509_STORE_CTX_trusted_stack().
6513 The verify_cb() and verify() callbacks now have equivalents
6514 in the X509_STORE_CTX structure.
6516 X509_STORE_CTX also has a 'flags' field which can be used
6517 to customise the verify behaviour.
6520 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6521 excludes S/MIME capabilities.
6524 *) When a certificate request is read in keep a copy of the
6525 original encoding of the signed data and use it when outputing
6526 again. Signatures then use the original encoding rather than
6527 a decoded, encoded version which may cause problems if the
6528 request is improperly encoded.
6531 *) For consistency with other BIO_puts implementations, call
6532 buffer_write(b, ...) directly in buffer_puts instead of calling
6535 In BIO_puts, increment b->num_write as in BIO_write.
6536 [Peter.Sylvester@EdelWeb.fr]
6538 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6539 BN_zero, we may not return a BIGNUM with an array consisting of
6543 *) Avoid calling abort() from within the library when problems are
6544 detected, except if preprocessor symbols have been defined
6545 (such as REF_CHECK, BN_DEBUG etc.).
6548 *) New openssl application 'rsautl'. This utility can be
6549 used for low level RSA operations. DER public key
6550 BIO/fp routines also added.
6553 *) New Configure entry and patches for compiling on QNX 4.
6554 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6556 *) A demo state-machine implementation was sponsored by
6557 Nuron (http://www.nuron.com/) and is now available in
6558 demos/state_machine.
6561 *) New options added to the 'dgst' utility for signature
6562 generation and verification.
6565 *) Unrecognized PKCS#7 content types are now handled via a
6566 catch all ASN1_TYPE structure. This allows unsupported
6567 types to be stored as a "blob" and an application can
6568 encode and decode it manually.
6571 *) Fix various signed/unsigned issues to make a_strex.c
6573 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6575 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6576 length if passed a buffer. ASN1_INTEGER_to_BN failed
6577 if passed a NULL BN and its argument was negative.
6578 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6580 *) Modification to PKCS#7 encoding routines to output definite
6581 length encoding. Since currently the whole structures are in
6582 memory there's not real point in using indefinite length
6583 constructed encoding. However if OpenSSL is compiled with
6584 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6587 *) Added BIO_vprintf() and BIO_vsnprintf().
6590 *) Added more prefixes to parse for in the the strings written
6591 through a logging bio, to cover all the levels that are available
6592 through syslog. The prefixes are now:
6594 PANIC, EMERG, EMR => LOG_EMERG
6595 ALERT, ALR => LOG_ALERT
6596 CRIT, CRI => LOG_CRIT
6597 ERROR, ERR => LOG_ERR
6598 WARNING, WARN, WAR => LOG_WARNING
6599 NOTICE, NOTE, NOT => LOG_NOTICE
6600 INFO, INF => LOG_INFO
6601 DEBUG, DBG => LOG_DEBUG
6603 and as before, if none of those prefixes are present at the
6604 beginning of the string, LOG_ERR is chosen.
6606 On Win32, the LOG_* levels are mapped according to this:
6608 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6609 LOG_WARNING => EVENTLOG_WARNING_TYPE
6610 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6614 *) Made it possible to reconfigure with just the configuration
6615 argument "reconf" or "reconfigure". The command line arguments
6616 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6617 and are retrieved from there when reconfiguring.
6621 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6623 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6626 *) The obj_dat.pl script was messing up the sorting of object
6627 names. The reason was that it compared the quoted version
6628 of strings as a result "OCSP" > "OCSP Signing" because
6629 " > SPACE. Changed script to store unquoted versions of
6630 names and add quotes on output. It was also omitting some
6631 names from the lookup table if they were given a default
6632 value (that is if SN is missing it is given the same
6633 value as LN and vice versa), these are now added on the
6634 grounds that if an object has a name we should be able to
6635 look it up. Finally added warning output when duplicate
6636 short or long names are found.
6639 *) Changes needed for Tandem NSK.
6640 [Scott Uroff <scott@xypro.com>]
6642 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6643 RSA_padding_check_SSLv23(), special padding was never detected
6644 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6645 version rollback attacks was not effective.
6647 In s23_clnt.c, don't use special rollback-attack detection padding
6648 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6649 client; similarly, in s23_srvr.c, don't do the rollback check if
6650 SSL 2.0 is the only protocol enabled in the server.
6653 *) Make it possible to get hexdumps of unprintable data with 'openssl
6654 asn1parse'. By implication, the functions ASN1_parse_dump() and
6655 BIO_dump_indent() are added.
6658 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6659 these print out strings and name structures based on various
6660 flags including RFC2253 support and proper handling of
6661 multibyte characters. Added options to the 'x509' utility
6662 to allow the various flags to be set.
6665 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6666 Also change the functions X509_cmp_current_time() and
6667 X509_gmtime_adj() work with an ASN1_TIME structure,
6668 this will enable certificates using GeneralizedTime in validity
6669 dates to be checked.
6672 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6673 negative public key encodings) on by default,
6674 NO_NEG_PUBKEY_BUG can be set to disable it.
6677 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6678 content octets. An i2c_ASN1_OBJECT is unnecessary because
6679 the encoding can be trivially obtained from the structure.
6682 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6683 not read locks (CRYPTO_r_[un]lock).
6686 *) A first attempt at creating official support for shared
6687 libraries through configuration. I've kept it so the
6688 default is static libraries only, and the OpenSSL programs
6689 are always statically linked for now, but there are
6690 preparations for dynamic linking in place.
6691 This has been tested on Linux and Tru64.
6694 *) Randomness polling function for Win9x, as described in:
6695 Peter Gutmann, Software Generation of Practically Strong
6699 *) Fix so PRNG is seeded in req if using an already existing
6703 *) New options to smime application. -inform and -outform
6704 allow alternative formats for the S/MIME message including
6705 PEM and DER. The -content option allows the content to be
6706 specified separately. This should allow things like Netscape
6707 form signing output easier to verify.
6710 *) Fix the ASN1 encoding of tags using the 'long form'.
6713 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6714 STRING types. These convert content octets to and from the
6715 underlying type. The actual tag and length octets are
6716 already assumed to have been read in and checked. These
6717 are needed because all other string types have virtually
6718 identical handling apart from the tag. By having versions
6719 of the ASN1 functions that just operate on content octets
6720 IMPLICIT tagging can be handled properly. It also allows
6721 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6722 and ASN1_INTEGER are identical apart from the tag.
6725 *) Change the handling of OID objects as follows:
6727 - New object identifiers are inserted in objects.txt, following
6728 the syntax given in objects.README.
6729 - objects.pl is used to process obj_mac.num and create a new
6731 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6734 This is currently kind of a hack, and the perl code in objects.pl
6735 isn't very elegant, but it works as I intended. The simplest way
6736 to check that it worked correctly is to look in obj_dat.h and
6737 check the array nid_objs and make sure the objects haven't moved
6738 around (this is important!). Additions are OK, as well as
6739 consistent name changes.
6742 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6745 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6746 The given file adds to whatever has already been seeded into the
6747 random pool through the RANDFILE configuration file option or
6748 environment variable, or the default random state file.
6751 *) mkstack.pl now sorts each macro group into lexical order.
6752 Previously the output order depended on the order the files
6753 appeared in the directory, resulting in needless rewriting
6757 *) Patches to make OpenSSL compile under Win32 again. Mostly
6758 work arounds for the VC++ problem that it treats func() as
6759 func(void). Also stripped out the parts of mkdef.pl that
6760 added extra typesafe functions: these no longer exist.
6763 *) Reorganisation of the stack code. The macros are now all
6764 collected in safestack.h . Each macro is defined in terms of
6765 a "stack macro" of the form SKM_<name>(type, a, b). The
6766 DEBUG_SAFESTACK is now handled in terms of function casts,
6767 this has the advantage of retaining type safety without the
6768 use of additional functions. If DEBUG_SAFESTACK is not defined
6769 then the non typesafe macros are used instead. Also modified the
6770 mkstack.pl script to handle the new form. Needs testing to see
6771 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6772 the default if no major problems. Similar behaviour for ASN1_SET_OF
6773 and PKCS12_STACK_OF.
6776 *) When some versions of IIS use the 'NET' form of private key the
6777 key derivation algorithm is different. Normally MD5(password) is
6778 used as a 128 bit RC4 key. In the modified case
6779 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6780 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6781 as the old Netscape_RSA functions except they have an additional
6782 'sgckey' parameter which uses the modified algorithm. Also added
6783 an -sgckey command line option to the rsa utility. Thanks to
6784 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6785 algorithm to openssl-dev.
6788 *) The evp_local.h macros were using 'c.##kname' which resulted in
6789 invalid expansion on some systems (SCO 5.0.5 for example).
6790 Corrected to 'c.kname'.
6791 [Phillip Porch <root@theporch.com>]
6793 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6794 a STACK of email addresses from a certificate or request, these look
6795 in the subject name and the subject alternative name extensions and
6796 omit any duplicate addresses.
6799 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6800 This makes DSA verification about 2 % faster.
6803 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6804 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6805 plus overhead for 1024 bit moduli).
6806 This makes exponentiations about 0.5 % faster for 1024 bit
6807 exponents (as measured by "openssl speed rsa2048").
6810 *) Rename memory handling macros to avoid conflicts with other
6812 Malloc => OPENSSL_malloc
6813 Malloc_locked => OPENSSL_malloc_locked
6814 Realloc => OPENSSL_realloc
6815 Free => OPENSSL_free
6818 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6819 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6822 *) CygWin32 support.
6823 [John Jarvie <jjarvie@newsguy.com>]
6825 *) The type-safe stack code has been rejigged. It is now only compiled
6826 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6827 by default all type-specific stack functions are "#define"d back to
6828 standard stack functions. This results in more streamlined output
6829 but retains the type-safety checking possibilities of the original
6833 *) The STACK code has been cleaned up, and certain type declarations
6834 that didn't make a lot of sense have been brought in line. This has
6835 also involved a cleanup of sorts in safestack.h to more correctly
6836 map type-safe stack functions onto their plain stack counterparts.
6837 This work has also resulted in a variety of "const"ifications of
6838 lots of the code, especially "_cmp" operations which should normally
6839 be prototyped with "const" parameters anyway.
6842 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6843 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6844 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6845 where all of 'md' is used each time the PRNG is used, but 'state'
6846 is used only indexed by a cyclic counter. As entropy may not be
6847 well distributed from the beginning, 'md' is important as a
6848 chaining variable. However, the output function chains only half
6849 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6850 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6851 in all of 'state' being rewritten, with the new values depending
6852 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6855 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6856 the handshake is continued after ssl_verify_cert_chain();
6857 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6858 can lead to 'unexplainable' connection aborts later.
6859 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6861 *) Major EVP API cipher revision.
6862 Add hooks for extra EVP features. This allows various cipher
6863 parameters to be set in the EVP interface. Support added for variable
6864 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6865 setting of RC2 and RC5 parameters.
6867 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6870 Remove lots of duplicated code from the EVP library. For example *every*
6871 cipher init() function handles the 'iv' in the same way according to the
6872 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6873 for CFB and OFB modes they zero ctx->num.
6875 New functionality allows removal of S/MIME code RC2 hack.
6877 Most of the routines have the same form and so can be declared in terms
6880 By shifting this to the top level EVP_CipherInit() it can be removed from
6881 all individual ciphers. If the cipher wants to handle IVs or keys
6882 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6885 Change lots of functions like EVP_EncryptUpdate() to now return a
6886 value: although software versions of the algorithms cannot fail
6887 any installed hardware versions can.
6890 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6891 this option is set, tolerate broken clients that send the negotiated
6892 protocol version number instead of the requested protocol version
6896 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6897 i.e. non-zero for export ciphersuites, zero otherwise.
6898 Previous versions had this flag inverted, inconsistent with
6899 rsa_tmp_cb (..._TMP_RSA_CB).
6900 [Bodo Moeller; problem reported by Amit Chopra]
6902 *) Add missing DSA library text string. Work around for some IIS
6903 key files with invalid SEQUENCE encoding.
6906 *) Add a document (doc/standards.txt) that list all kinds of standards
6907 and so on that are implemented in OpenSSL.
6910 *) Enhance c_rehash script. Old version would mishandle certificates
6911 with the same subject name hash and wouldn't handle CRLs at all.
6912 Added -fingerprint option to crl utility, to support new c_rehash
6916 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6919 *) Fix for SSL server purpose checking. Server checking was
6920 rejecting certificates which had extended key usage present
6921 but no ssl client purpose.
6922 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6924 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6925 is a little unclear about how a blank password is handled.
6926 Since the password in encoded as a BMPString with terminating
6927 double NULL a zero length password would end up as just the
6928 double NULL. However no password at all is different and is
6929 handled differently in the PKCS#12 key generation code. NS
6930 treats a blank password as zero length. MSIE treats it as no
6931 password on export: but it will try both on import. We now do
6932 the same: PKCS12_parse() tries zero length and no password if
6933 the password is set to "" or NULL (NULL is now a valid password:
6934 it wasn't before) as does the pkcs12 application.
6937 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6938 perror when PEM_read_bio_X509_REQ fails, the error message must
6939 be obtained from the error queue.
6942 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6943 it in ERR_remove_state if appropriate, and change ERR_get_state
6944 accordingly to avoid race conditions (this is necessary because
6945 thread_hash is no longer constant once set).
6948 *) Bugfix for linux-elf makefile.one.
6951 *) RSA_get_default_method() will now cause a default
6952 RSA_METHOD to be chosen if one doesn't exist already.
6953 Previously this was only set during a call to RSA_new()
6954 or RSA_new_method(NULL) meaning it was possible for
6955 RSA_get_default_method() to return NULL.
6958 *) Added native name translation to the existing DSO code
6959 that will convert (if the flag to do so is set) filenames
6960 that are sufficiently small and have no path information
6961 into a canonical native form. Eg. "blah" converted to
6962 "libblah.so" or "blah.dll" etc.
6965 *) New function ERR_error_string_n(e, buf, len) which is like
6966 ERR_error_string(e, buf), but writes at most 'len' bytes
6967 including the 0 terminator. For ERR_error_string_n, 'buf'
6969 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6971 *) CONF library reworked to become more general. A new CONF
6972 configuration file reader "class" is implemented as well as a
6973 new functions (NCONF_*, for "New CONF") to handle it. The now
6974 old CONF_* functions are still there, but are reimplemented to
6975 work in terms of the new functions. Also, a set of functions
6976 to handle the internal storage of the configuration data is
6977 provided to make it easier to write new configuration file
6978 reader "classes" (I can definitely see something reading a
6979 configuration file in XML format, for example), called _CONF_*,
6980 or "the configuration storage API"...
6982 The new configuration file reading functions are:
6984 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6985 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6987 NCONF_default, NCONF_WIN32
6989 NCONF_dump_fp, NCONF_dump_bio
6991 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6992 NCONF_new creates a new CONF object. This works in the same way
6993 as other interfaces in OpenSSL, like the BIO interface.
6994 NCONF_dump_* dump the internal storage of the configuration file,
6995 which is useful for debugging. All other functions take the same
6996 arguments as the old CONF_* functions wth the exception of the
6997 first that must be a `CONF *' instead of a `LHASH *'.
6999 To make it easer to use the new classes with the old CONF_* functions,
7000 the function CONF_set_default_method is provided.
7003 *) Add '-tls1' option to 'openssl ciphers', which was already
7004 mentioned in the documentation but had not been implemented.
7005 (This option is not yet really useful because even the additional
7006 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7009 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7010 OpenSSL-based applications) load shared libraries and bind to
7011 them in a portable way.
7012 [Geoff Thorpe, with contributions from Richard Levitte]
7014 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7016 *) Make sure _lrotl and _lrotr are only used with MSVC.
7018 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7019 (the default implementation of RAND_status).
7021 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7022 to '-clrext' (= clear extensions), as intended and documented.
7023 [Bodo Moeller; inconsistency pointed out by Michael Attili
7026 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7027 was larger than the MD block size.
7028 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7030 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7031 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7032 using the passed key: if the passed key was a private key the result
7033 of X509_print(), for example, would be to print out all the private key
7037 *) des_quad_cksum() byte order bug fix.
7038 [Ulf Möller, using the problem description in krb4-0.9.7, where
7039 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7041 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7043 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7045 *) For easily testing in shell scripts whether some command
7046 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7047 returns with exit code 0 iff no command of the given name is available.
7048 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7049 the output goes to stdout and nothing is printed to stderr.
7050 Additional arguments are always ignored.
7052 Since for each cipher there is a command of the same name,
7053 the 'no-cipher' compilation switches can be tested this way.
7055 ('openssl no-XXX' is not able to detect pseudo-commands such
7056 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7059 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7062 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7063 is set; it will be thrown away anyway because each handshake creates
7065 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7066 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7067 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7068 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7071 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7072 'Q' and 'R' lose their special meanings (quit/renegotiate).
7073 This is part of what -quiet does; unlike -quiet, -ign_eof
7074 does not suppress any output.
7077 *) Add compatibility options to the purpose and trust code. The
7078 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7079 accepts a certificate or CA, this was the previous behaviour,
7080 with all the associated security issues.
7082 X509_TRUST_COMPAT is the old trust behaviour: only and
7083 automatically trust self signed roots in certificate store. A
7084 new trust setting X509_TRUST_DEFAULT is used to specify that
7085 a purpose has no associated trust setting and it should instead
7086 use the value in the default purpose.
7089 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7090 and fix a memory leak.
7093 *) In util/mkerr.pl (which implements 'make errors'), preserve
7094 reason strings from the previous version of the .c file, as
7095 the default to have only downcase letters (and digits) in
7096 automatically generated reasons codes is not always appropriate.
7099 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7100 using strerror. Previously, ERR_reason_error_string() returned
7101 library names as reason strings for SYSerr; but SYSerr is a special
7102 case where small numbers are errno values, not library numbers.
7105 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7106 converts DSA parameters into DH parameters. (When creating parameters,
7107 DSA_generate_parameters is used.)
7110 *) Include 'length' (recommended exponent length) in C code generated
7111 by 'openssl dhparam -C'.
7114 *) The second argument to set_label in perlasm was already being used
7115 so couldn't be used as a "file scope" flag. Moved to third argument
7119 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7120 instead of RAND_bytes for encryption IVs and salts.
7123 *) Include RAND_status() into RAND_METHOD instead of implementing
7124 it only for md_rand.c Otherwise replacing the PRNG by calling
7125 RAND_set_rand_method would be impossible.
7128 *) Don't let DSA_generate_key() enter an infinite loop if the random
7129 number generation fails.
7132 *) New 'rand' application for creating pseudo-random output.
7135 *) Added configuration support for Linux/IA64
7136 [Rolf Haberrecker <rolf@suse.de>]
7138 *) Assembler module support for Mingw32.
7141 *) Shared library support for HPUX (in shlib/).
7142 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7144 *) Shared library support for Solaris gcc.
7145 [Lutz Behnke <behnke@trustcenter.de>]
7147 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7149 *) PKCS7_encrypt() was adding text MIME headers twice because they
7150 were added manually and by SMIME_crlf_copy().
7153 *) In bntest.c don't call BN_rand with zero bits argument.
7154 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7156 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7157 case was implemented. This caused BN_div_recp() to fail occasionally.
7160 *) Add an optional second argument to the set_label() in the perl
7161 assembly language builder. If this argument exists and is set
7162 to 1 it signals that the assembler should use a symbol whose
7163 scope is the entire file, not just the current function. This
7164 is needed with MASM which uses the format label:: for this scope.
7165 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7167 *) Change the ASN1 types so they are typedefs by default. Before
7168 almost all types were #define'd to ASN1_STRING which was causing
7169 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7173 *) Change names of new functions to the new get1/get0 naming
7174 convention: After 'get1', the caller owns a reference count
7175 and has to call ..._free; 'get0' returns a pointer to some
7176 data structure without incrementing reference counters.
7177 (Some of the existing 'get' functions increment a reference
7178 counter, some don't.)
7179 Similarly, 'set1' and 'add1' functions increase reference
7180 counters or duplicate objects.
7183 *) Allow for the possibility of temp RSA key generation failure:
7184 the code used to assume it always worked and crashed on failure.
7187 *) Fix potential buffer overrun problem in BIO_printf().
7188 [Ulf Möller, using public domain code by Patrick Powell; problem
7189 pointed out by David Sacerdote <das33@cornell.edu>]
7191 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7192 RAND_egd() and RAND_status(). In the command line application,
7193 the EGD socket can be specified like a seed file using RANDFILE
7197 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7198 Some CAs (e.g. Verisign) distribute certificates in this form.
7201 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7202 list to exclude them. This means that no special compilation option
7203 is needed to use anonymous DH: it just needs to be included in the
7207 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7208 EVP_MD_type. The old functionality is available in a new macro called
7209 EVP_MD_md(). Change code that uses it and update docs.
7212 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7213 where the 'void *' argument is replaced by a function pointer argument.
7214 Previously 'void *' was abused to point to functions, which works on
7215 many platforms, but is not correct. As these functions are usually
7216 called by macros defined in OpenSSL header files, most source code
7217 should work without changes.
7220 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7221 sections with information on -D... compiler switches used for
7222 compiling the library so that applications can see them. To enable
7223 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7224 must be defined. E.g.,
7225 #define OPENSSL_ALGORITHM_DEFINES
7226 #include <openssl/opensslconf.h>
7227 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7228 [Richard Levitte, Ulf and Bodo Möller]
7230 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7234 *) Change the 'other' type in certificate aux info to a STACK_OF
7235 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7236 the required ASN1 format: arbitrary types determined by an OID.
7239 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7240 argument to 'req'. This is not because the function is newer or
7241 better than others it just uses the work 'NEW' in the certificate
7242 request header lines. Some software needs this.
7245 *) Reorganise password command line arguments: now passwords can be
7246 obtained from various sources. Delete the PEM_cb function and make
7247 it the default behaviour: i.e. if the callback is NULL and the
7248 usrdata argument is not NULL interpret it as a null terminated pass
7249 phrase. If usrdata and the callback are NULL then the pass phrase
7250 is prompted for as usual.
7253 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7254 the support is automatically enabled. The resulting binaries will
7255 autodetect the card and use it if present.
7256 [Ben Laurie and Compaq Inc.]
7258 *) Work around for Netscape hang bug. This sends certificate request
7259 and server done in one record. Since this is perfectly legal in the
7260 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7261 the bugs/SSLv3 entry for more info.
7264 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7267 *) Add -rand argument to smime and pkcs12 applications and read/write
7271 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7274 *) Add command line password options to the remaining applications.
7277 *) Bug fix for BN_div_recp() for numerators with an even number of
7281 *) More tests in bntest.c, and changed test_bn output.
7284 *) ./config recognizes MacOS X now.
7287 *) Bug fix for BN_div() when the first words of num and divsor are
7288 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7291 *) Add support for various broken PKCS#8 formats, and command line
7292 options to produce them.
7295 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7296 get temporary BIGNUMs from a BN_CTX.
7299 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7303 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7304 include a #define from the old name to the new. The original intent
7305 was that statically linked binaries could for example just call
7306 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7307 link with digests. This never worked becayse SSLeay_add_all_digests()
7308 and SSLeay_add_all_ciphers() were in the same source file so calling
7309 one would link with the other. They are now in separate source files.
7312 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7315 *) Use a less unusual form of the Miller-Rabin primality test (it used
7316 a binary algorithm for exponentiation integrated into the Miller-Rabin
7317 loop, our standard modexp algorithms are faster).
7320 *) Support for the EBCDIC character set completed.
7321 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7323 *) Source code cleanups: use const where appropriate, eliminate casts,
7324 use void * instead of char * in lhash.
7327 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7328 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7329 this the server could overwrite ephemeral keys that the client
7333 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7334 using 50 iterations of the Rabin-Miller test.
7336 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7337 iterations of the Rabin-Miller test as required by the appendix
7338 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7339 As BN_is_prime_fasttest includes trial division, DSA parameter
7340 generation becomes much faster.
7342 This implies a change for the callback functions in DSA_is_prime
7343 and DSA_generate_parameters: The callback function is called once
7344 for each positive witness in the Rabin-Miller test, not just
7345 occasionally in the inner loop; and the parameters to the
7346 callback function now provide an iteration count for the outer
7347 loop rather than for the current invocation of the inner loop.
7348 DSA_generate_parameters additionally can call the callback
7349 function with an 'iteration count' of -1, meaning that a
7350 candidate has passed the trial division test (when q is generated
7351 from an application-provided seed, trial division is skipped).
7354 *) New function BN_is_prime_fasttest that optionally does trial
7355 division before starting the Rabin-Miller test and has
7356 an additional BN_CTX * argument (whereas BN_is_prime always
7357 has to allocate at least one BN_CTX).
7358 'callback(1, -1, cb_arg)' is called when a number has passed the
7359 trial division stage.
7362 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7366 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7369 *) New function BN_pseudo_rand().
7372 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7373 bignum version of BN_from_montgomery() with the working code from
7374 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7378 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7379 made it impossible to use the same SSL_SESSION data structure in
7380 SSL2 clients in multiple threads.
7383 *) The return value of RAND_load_file() no longer counts bytes obtained
7384 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7385 to seed the PRNG (previously an explicit byte count was required).
7386 [Ulf Möller, Bodo Möller]
7388 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7389 used (char *) instead of (void *) and had casts all over the place.
7392 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7395 *) Retain source code compatibility for BN_prime_checks macro:
7396 BN_is_prime(..., BN_prime_checks, ...) now uses
7397 BN_prime_checks_for_size to determine the appropriate number of
7398 Rabin-Miller iterations.
7401 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7402 DH_CHECK_P_NOT_SAFE_PRIME.
7403 (Check if this is true? OpenPGP calls them "strong".)
7406 *) Merge the functionality of "dh" and "gendh" programs into a new program
7407 "dhparam". The old programs are retained for now but will handle DH keys
7408 (instead of parameters) in future.
7411 *) Make the ciphers, s_server and s_client programs check the return values
7412 when a new cipher list is set.
7415 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7416 ciphers. Before when the 56bit ciphers were enabled the sorting was
7419 The syntax for the cipher sorting has been extended to support sorting by
7420 cipher-strength (using the strength_bits hard coded in the tables).
7421 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7423 Fix a bug in the cipher-command parser: when supplying a cipher command
7424 string with an "undefined" symbol (neither command nor alphanumeric
7425 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7426 an error is flagged.
7428 Due to the strength-sorting extension, the code of the
7429 ssl_create_cipher_list() function was completely rearranged. I hope that
7430 the readability was also increased :-)
7431 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7433 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7434 for the first serial number and places 2 in the serial number file. This
7435 avoids problems when the root CA is created with serial number zero and
7436 the first user certificate has the same issuer name and serial number
7440 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7441 the new code. Add documentation for this stuff.
7444 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7445 X509_*() to X509at_*() on the grounds that they don't handle X509
7446 structures and behave in an analagous way to the X509v3 functions:
7447 they shouldn't be called directly but wrapper functions should be used
7450 So we also now have some wrapper functions that call the X509at functions
7451 when passed certificate requests. (TO DO: similar things can be done with
7452 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7453 things. Some of these need some d2i or i2d and print functionality
7454 because they handle more complex structures.)
7457 *) Add missing #ifndefs that caused missing symbols when building libssl
7458 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7459 NO_RSA in ssl/s2*.c.
7460 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7462 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7463 has a return value which indicates the quality of the random data
7464 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7465 error queue. New function RAND_pseudo_bytes() generates output that is
7466 guaranteed to be unique but not unpredictable. RAND_add is like
7467 RAND_seed, but takes an extra argument for an entropy estimate
7468 (RAND_seed always assumes full entropy).
7471 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7472 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7473 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7474 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7475 false-positive rate of at most 2^-80 for random input.
7478 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7481 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7482 in the 0.9.5 release), this returns the chain
7483 from an X509_CTX structure with a dup of the stack and all
7484 the X509 reference counts upped: so the stack will exist
7485 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7488 Also make SSL_SESSION_print() print out the verify return
7492 *) Add manpage for the pkcs12 command. Also change the default
7493 behaviour so MAC iteration counts are used unless the new
7494 -nomaciter option is used. This improves file security and
7495 only older versions of MSIE (4.0 for example) need it.
7498 *) Honor the no-xxx Configure options when creating .DEF files.
7501 *) Add PKCS#10 attributes to field table: challengePassword,
7502 unstructuredName and unstructuredAddress. These are taken from
7503 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7504 international characters are used.
7506 More changes to X509_ATTRIBUTE code: allow the setting of types
7507 based on strings. Remove the 'loc' parameter when adding
7508 attributes because these will be a SET OF encoding which is sorted
7512 *) Initial changes to the 'req' utility to allow request generation
7513 automation. This will allow an application to just generate a template
7514 file containing all the field values and have req construct the
7517 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7518 used all over the place including certificate requests and PKCS#7
7519 structures. They are currently handled manually where necessary with
7520 some primitive wrappers for PKCS#7. The new functions behave in a
7521 manner analogous to the X509 extension functions: they allow
7522 attributes to be looked up by NID and added.
7524 Later something similar to the X509V3 code would be desirable to
7525 automatically handle the encoding, decoding and printing of the
7526 more complex types. The string types like challengePassword can
7527 be handled by the string table functions.
7529 Also modified the multi byte string table handling. Now there is
7530 a 'global mask' which masks out certain types. The table itself
7531 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7532 is useful when for example there is only one permissible type
7533 (as in countryName) and using the mask might result in no valid
7537 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7538 SSL_get_peer_finished to allow applications to obtain the latest
7539 Finished messages sent to the peer or expected from the peer,
7540 respectively. (SSL_get_peer_finished is usually the Finished message
7541 actually received from the peer, otherwise the protocol will be aborted.)
7543 As the Finished message are message digests of the complete handshake
7544 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7545 be used for external authentication procedures when the authentication
7546 provided by SSL/TLS is not desired or is not enough.
7549 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7550 the host supports BWX extension and if Compaq C is present on the
7551 $PATH. Just exploiting of the BWX extension results in 20-30%
7552 performance kick for some algorithms, e.g. DES and RC4 to mention
7553 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7557 *) Add support for MS "fast SGC". This is arguably a violation of the
7558 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7559 weak crypto and after checking the certificate is SGC a second one
7560 with strong crypto. MS SGC stops the first handshake after receiving
7561 the server certificate message and sends a second client hello. Since
7562 a server will typically do all the time consuming operations before
7563 expecting any further messages from the client (server key exchange
7564 is the most expensive) there is little difference between the two.
7566 To get OpenSSL to support MS SGC we have to permit a second client
7567 hello message after we have sent server done. In addition we have to
7568 reset the MAC if we do get this second client hello.
7571 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7572 if a DER encoded private key is RSA or DSA traditional format. Changed
7573 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7574 format DER encoded private key. Newer code should use PKCS#8 format which
7575 has the key type encoded in the ASN1 structure. Added DER private key
7576 support to pkcs8 application.
7579 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7580 ciphersuites has been selected (as required by the SSL 3/TLS 1
7581 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7582 is set, we interpret this as a request to violate the specification
7583 (the worst that can happen is a handshake failure, and 'correct'
7584 behaviour would result in a handshake failure anyway).
7587 *) In SSL_CTX_add_session, take into account that there might be multiple
7588 SSL_SESSION structures with the same session ID (e.g. when two threads
7589 concurrently obtain them from an external cache).
7590 The internal cache can handle only one SSL_SESSION with a given ID,
7591 so if there's a conflict, we now throw out the old one to achieve
7595 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7596 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7597 some routines that use cipher OIDs: some ciphers do not have OIDs
7598 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7602 *) Simplify the trust setting structure and code. Now we just have
7603 two sequences of OIDs for trusted and rejected settings. These will
7604 typically have values the same as the extended key usage extension
7605 and any application specific purposes.
7607 The trust checking code now has a default behaviour: it will just
7608 check for an object with the same NID as the passed id. Functions can
7609 be provided to override either the default behaviour or the behaviour
7610 for a given id. SSL client, server and email already have functions
7611 in place for compatibility: they check the NID and also return "trusted"
7612 if the certificate is self signed.
7615 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7616 traditional format into an EVP_PKEY structure.
7619 *) Add a password callback function PEM_cb() which either prompts for
7620 a password if usr_data is NULL or otherwise assumes it is a null
7621 terminated password. Allow passwords to be passed on command line
7622 environment or config files in a few more utilities.
7625 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7626 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7627 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7628 Update documentation.
7631 *) Support for ASN1 "NULL" type. This could be handled before by using
7632 ASN1_TYPE but there wasn't any function that would try to read a NULL
7633 and produce an error if it couldn't. For compatibility we also have
7634 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7635 don't allocate anything because they don't need to.
7638 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7640 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7642 *) Rebuild of the memory allocation routines used by OpenSSL code and
7643 possibly others as well. The purpose is to make an interface that
7644 provide hooks so anyone can build a separate set of allocation and
7645 deallocation routines to be used by OpenSSL, for example memory
7646 pool implementations, or something else, which was previously hard
7647 since Malloc(), Realloc() and Free() were defined as macros having
7648 the values malloc, realloc and free, respectively (except for Win32
7649 compilations). The same is provided for memory debugging code.
7650 OpenSSL already comes with functionality to find memory leaks, but
7651 this gives people a chance to debug other memory problems.
7653 With these changes, a new set of functions and macros have appeared:
7655 CRYPTO_set_mem_debug_functions() [F]
7656 CRYPTO_get_mem_debug_functions() [F]
7657 CRYPTO_dbg_set_options() [F]
7658 CRYPTO_dbg_get_options() [F]
7659 CRYPTO_malloc_debug_init() [M]
7661 The memory debug functions are NULL by default, unless the library
7662 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7663 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7664 gives the standard debugging functions that come with OpenSSL) or
7665 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7666 provided by the library user) must be used. When the standard
7667 debugging functions are used, CRYPTO_dbg_set_options can be used to
7668 request additional information:
7669 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7670 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7672 Also, things like CRYPTO_set_mem_functions will always give the
7673 expected result (the new set of functions is used for allocation
7674 and deallocation) at all times, regardless of platform and compiler
7677 To finish it up, some functions that were never use in any other
7678 way than through macros have a new API and new semantic:
7681 CRYPTO_dbg_realloc()
7684 All macros of value have retained their old syntax.
7685 [Richard Levitte and Bodo Moeller]
7687 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7688 ordering of SMIMECapabilities wasn't in "strength order" and there
7689 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7693 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7694 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7695 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7697 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7698 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7699 functionality to handle multipart/signed properly) and a utility
7700 called 'smime' to call all this stuff. This is based on code I
7701 originally wrote for Celo who have kindly allowed it to be
7702 included in OpenSSL.
7705 *) Add variants des_set_key_checked and des_set_key_unchecked of
7706 des_set_key (aka des_key_sched). Global variable des_check_key
7707 decides which of these is called by des_set_key; this way
7708 des_check_key behaves as it always did, but applications and
7709 the library itself, which was buggy for des_check_key == 1,
7710 have a cleaner way to pick the version they need.
7713 *) New function PKCS12_newpass() which changes the password of a
7717 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7718 dynamic mix. In both cases the ids can be used as an index into the
7719 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7720 functions so they accept a list of the field values and the
7721 application doesn't need to directly manipulate the X509_TRUST
7725 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7729 *) Modify the way the V3 extension code looks up extensions. This now
7730 works in a similar way to the object code: we have some "standard"
7731 extensions in a static table which is searched with OBJ_bsearch()
7732 and the application can add dynamic ones if needed. The file
7733 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7734 updated whenever a new extension is added to the core code and kept
7735 in ext_nid order. There is a simple program 'tabtest.c' which checks
7736 this. New extensions are not added too often so this file can readily
7737 be maintained manually.
7739 There are two big advantages in doing things this way. The extensions
7740 can be looked up immediately and no longer need to be "added" using
7741 X509V3_add_standard_extensions(): this function now does nothing.
7742 [Side note: I get *lots* of email saying the extension code doesn't
7743 work because people forget to call this function]
7744 Also no dynamic allocation is done unless new extensions are added:
7745 so if we don't add custom extensions there is no need to call
7746 X509V3_EXT_cleanup().
7749 *) Modify enc utility's salting as follows: make salting the default. Add a
7750 magic header, so unsalted files fail gracefully instead of just decrypting
7751 to garbage. This is because not salting is a big security hole, so people
7752 should be discouraged from doing it.
7755 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7756 digest to be passed on the command line but it only used this
7757 parameter when signing a certificate. Modified so all relevant
7758 operations are affected by the digest parameter including the
7759 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7760 DSA key was used because it didn't fix the digest.
7763 *) Initial certificate chain verify code. Currently tests the untrusted
7764 certificates for consistency with the verify purpose (which is set
7765 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7767 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7768 this is because it will reject chains with invalid extensions whereas
7769 every previous version of OpenSSL and SSLeay made no checks at all.
7771 Trust code: checks the root CA for the relevant trust settings. Trust
7772 settings have an initial value consistent with the verify purpose: e.g.
7773 if the verify purpose is for SSL client use it expects the CA to be
7774 trusted for SSL client use. However the default value can be changed to
7775 permit custom trust settings: one example of this would be to only trust
7776 certificates from a specific "secure" set of CAs.
7778 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7779 which should be used for version portability: especially since the
7780 verify structure is likely to change more often now.
7782 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7783 to set them. If not set then assume SSL clients will verify SSL servers
7786 Two new options to the verify program: -untrusted allows a set of
7787 untrusted certificates to be passed in and -purpose which sets the
7788 intended purpose of the certificate. If a purpose is set then the
7789 new chain verify code is used to check extension consistency.
7792 *) Support for the authority information access extension.
7795 *) Modify RSA and DSA PEM read routines to transparently handle
7796 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7797 public keys in a format compatible with certificate
7798 SubjectPublicKeyInfo structures. Unfortunately there were already
7799 functions called *_PublicKey_* which used various odd formats so
7800 these are retained for compatibility: however the DSA variants were
7801 never in a public release so they have been deleted. Changed dsa/rsa
7802 utilities to handle the new format: note no releases ever handled public
7803 keys so we should be OK.
7805 The primary motivation for this change is to avoid the same fiasco
7806 that dogs private keys: there are several incompatible private key
7807 formats some of which are standard and some OpenSSL specific and
7808 require various evil hacks to allow partial transparent handling and
7809 even then it doesn't work with DER formats. Given the option anything
7810 other than PKCS#8 should be dumped: but the other formats have to
7811 stay in the name of compatibility.
7813 With public keys and the benefit of hindsight one standard format
7814 is used which works with EVP_PKEY, RSA or DSA structures: though
7815 it clearly returns an error if you try to read the wrong kind of key.
7817 Added a -pubkey option to the 'x509' utility to output the public key.
7818 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7819 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7820 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7821 that do the same as the EVP_PKEY_assign_*() except they up the
7822 reference count of the added key (they don't "swallow" the
7826 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7827 CRLs would fail if the file contained no certificates or no CRLs:
7828 added a new function to read in both types and return the number
7829 read: this means that if none are read it will be an error. The
7830 DER versions of the certificate and CRL reader would always fail
7831 because it isn't possible to mix certificates and CRLs in DER format
7832 without choking one or the other routine. Changed this to just read
7833 a certificate: this is the best we can do. Also modified the code
7834 in apps/verify.c to take notice of return codes: it was previously
7835 attempting to read in certificates from NULL pointers and ignoring
7836 any errors: this is one reason why the cert and CRL reader seemed
7837 to work. It doesn't check return codes from the default certificate
7838 routines: these may well fail if the certificates aren't installed.
7841 *) Code to support otherName option in GeneralName.
7844 *) First update to verify code. Change the verify utility
7845 so it warns if it is passed a self signed certificate:
7846 for consistency with the normal behaviour. X509_verify
7847 has been modified to it will now verify a self signed
7848 certificate if *exactly* the same certificate appears
7849 in the store: it was previously impossible to trust a
7850 single self signed certificate. This means that:
7851 openssl verify ss.pem
7852 now gives a warning about a self signed certificate but
7853 openssl verify -CAfile ss.pem ss.pem
7857 *) For servers, store verify_result in SSL_SESSION data structure
7858 (and add it to external session representation).
7859 This is needed when client certificate verifications fails,
7860 but an application-provided verification callback (set by
7861 SSL_CTX_set_cert_verify_callback) allows accepting the session
7862 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7863 but returns 1): When the session is reused, we have to set
7864 ssl->verify_result to the appropriate error code to avoid
7866 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7868 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7869 case in PKCS7_dataInit() where the signed PKCS7 structure
7870 didn't contain any existing data because it was being created.
7871 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7873 *) Add a salt to the key derivation routines in enc.c. This
7874 forms the first 8 bytes of the encrypted file. Also add a
7875 -S option to allow a salt to be input on the command line.
7878 *) New function X509_cmp(). Oddly enough there wasn't a function
7879 to compare two certificates. We do this by working out the SHA1
7880 hash and comparing that. X509_cmp() will be needed by the trust
7884 *) SSL_get1_session() is like SSL_get_session(), but increments
7885 the reference count in the SSL_SESSION returned.
7886 [Geoff Thorpe <geoff@eu.c2.net>]
7888 *) Fix for 'req': it was adding a null to request attributes.
7889 Also change the X509_LOOKUP and X509_INFO code to handle
7890 certificate auxiliary information.
7893 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7897 *) Add the possibility to add extra information to the memory leak
7898 detecting output, to form tracebacks, showing from where each
7899 allocation was originated: CRYPTO_push_info("constant string") adds
7900 the string plus current file name and line number to a per-thread
7901 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7902 is like calling CYRPTO_pop_info() until the stack is empty.
7903 Also updated memory leak detection code to be multi-thread-safe.
7906 *) Add options -text and -noout to pkcs7 utility and delete the
7907 encryption options which never did anything. Update docs.
7910 *) Add options to some of the utilities to allow the pass phrase
7911 to be included on either the command line (not recommended on
7912 OSes like Unix) or read from the environment. Update the
7913 manpages and fix a few bugs.
7916 *) Add a few manpages for some of the openssl commands.
7919 *) Fix the -revoke option in ca. It was freeing up memory twice,
7920 leaking and not finding already revoked certificates.
7923 *) Extensive changes to support certificate auxiliary information.
7924 This involves the use of X509_CERT_AUX structure and X509_AUX
7925 functions. An X509_AUX function such as PEM_read_X509_AUX()
7926 can still read in a certificate file in the usual way but it
7927 will also read in any additional "auxiliary information". By
7928 doing things this way a fair degree of compatibility can be
7929 retained: existing certificates can have this information added
7930 using the new 'x509' options.
7932 Current auxiliary information includes an "alias" and some trust
7933 settings. The trust settings will ultimately be used in enhanced
7934 certificate chain verification routines: currently a certificate
7935 can only be trusted if it is self signed and then it is trusted
7939 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7940 The problem was that one of the replacement routines had not been working
7941 since SSLeay releases. For now the offending routine has been replaced
7942 with non-optimised assembler. Even so, this now gives around 95%
7943 performance improvement for 1024 bit RSA signs.
7946 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7947 handling. Most clients have the effective key size in bits equal to
7948 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7949 A few however don't do this and instead use the size of the decrypted key
7950 to determine the RC2 key length and the AlgorithmIdentifier to determine
7951 the effective key length. In this case the effective key length can still
7952 be 40 bits but the key length can be 168 bits for example. This is fixed
7953 by manually forcing an RC2 key into the EVP_PKEY structure because the
7954 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7955 the key length and effective key length are equal.
7958 *) Add a bunch of functions that should simplify the creation of
7959 X509_NAME structures. Now you should be able to do:
7960 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7961 and have it automatically work out the correct field type and fill in
7962 the structures. The more adventurous can try:
7963 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7964 and it will (hopefully) work out the correct multibyte encoding.
7967 *) Change the 'req' utility to use the new field handling and multibyte
7968 copy routines. Before the DN field creation was handled in an ad hoc
7969 way in req, ca, and x509 which was rather broken and didn't support
7970 BMPStrings or UTF8Strings. Since some software doesn't implement
7971 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7972 using the dirstring_type option. See the new comment in the default
7973 openssl.cnf for more info.
7976 *) Make crypto/rand/md_rand.c more robust:
7977 - Assure unique random numbers after fork().
7978 - Make sure that concurrent threads access the global counter and
7979 md serializably so that we never lose entropy in them
7980 or use exactly the same state in multiple threads.
7981 Access to the large state is not always serializable because
7982 the additional locking could be a performance killer, and
7983 md should be large enough anyway.
7986 *) New file apps/app_rand.c with commonly needed functionality
7987 for handling the random seed file.
7989 Use the random seed file in some applications that previously did not:
7991 dsaparam -genkey (which also ignored its '-rand' option),
7994 x509 (when signing).
7995 Except on systems with /dev/urandom, it is crucial to have a random
7996 seed file at least for key creation, DSA signing, and for DH exchanges;
7997 for RSA signatures we could do without one.
7999 gendh and gendsa (unlike genrsa) used to read only the first byte
8000 of each file listed in the '-rand' option. The function as previously
8001 found in genrsa is now in app_rand.c and is used by all programs
8002 that support '-rand'.
8005 *) In RAND_write_file, use mode 0600 for creating files;
8006 don't just chmod when it may be too late.
8009 *) Report an error from X509_STORE_load_locations
8010 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8013 *) New function ASN1_mbstring_copy() this copies a string in either
8014 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8015 into an ASN1_STRING type. A mask of permissible types is passed
8016 and it chooses the "minimal" type to use or an error if not type
8020 *) Add function equivalents to the various macros in asn1.h. The old
8021 macros are retained with an M_ prefix. Code inside the library can
8022 use the M_ macros. External code (including the openssl utility)
8023 should *NOT* in order to be "shared library friendly".
8026 *) Add various functions that can check a certificate's extensions
8027 to see if it usable for various purposes such as SSL client,
8028 server or S/MIME and CAs of these types. This is currently
8029 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8030 verification. Also added a -purpose flag to x509 utility to
8031 print out all the purposes.
8034 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8038 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8039 for, obtain and decode and extension and obtain its critical flag.
8040 This allows all the necessary extension code to be handled in a
8041 single function call.
8044 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8045 platforms. See crypto/rc4/rc4_enc.c for further details.
8048 *) New -noout option to asn1parse. This causes no output to be produced
8049 its main use is when combined with -strparse and -out to extract data
8050 from a file (which may not be in ASN.1 format).
8053 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8054 when producing the local key id.
8055 [Richard Levitte <levitte@stacken.kth.se>]
8057 *) New option -dhparam in s_server. This allows a DH parameter file to be
8058 stated explicitly. If it is not stated then it tries the first server
8059 certificate file. The previous behaviour hard coded the filename
8063 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8064 a public key to be input or output. For example:
8065 openssl rsa -in key.pem -pubout -out pubkey.pem
8066 Also added necessary DSA public key functions to handle this.
8069 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8070 in the message. This was handled by allowing
8071 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8072 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8074 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8075 to the end of the strings whereas this didn't. This would cause problems
8076 if strings read with d2i_ASN1_bytes() were later modified.
8077 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8079 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8080 data and it contains EOF it will end up returning an error. This is
8081 caused by input 46 bytes long. The cause is due to the way base64
8082 BIOs find the start of base64 encoded data. They do this by trying a
8083 trial decode on each line until they find one that works. When they
8084 do a flag is set and it starts again knowing it can pass all the
8085 data directly through the decoder. Unfortunately it doesn't reset
8086 the context it uses. This means that if EOF is reached an attempt
8087 is made to pass two EOFs through the context and this causes the
8088 resulting error. This can also cause other problems as well. As is
8089 usual with these problems it takes *ages* to find and the fix is
8090 trivial: move one line.
8091 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8093 *) Ugly workaround to get s_client and s_server working under Windows. The
8094 old code wouldn't work because it needed to select() on sockets and the
8095 tty (for keypresses and to see if data could be written). Win32 only
8096 supports select() on sockets so we select() with a 1s timeout on the
8097 sockets and then see if any characters are waiting to be read, if none
8098 are present then we retry, we also assume we can always write data to
8099 the tty. This isn't nice because the code then blocks until we've
8100 received a complete line of data and it is effectively polling the
8101 keyboard at 1s intervals: however it's quite a bit better than not
8102 working at all :-) A dedicated Windows application might handle this
8103 with an event loop for example.
8106 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8107 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8108 will be called when RSA_sign() and RSA_verify() are used. This is useful
8109 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8110 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8111 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8112 This necessitated the support of an extra signature type NID_md5_sha1
8113 for SSL signatures and modifications to the SSL library to use it instead
8114 of calling RSA_public_decrypt() and RSA_private_encrypt().
8117 *) Add new -verify -CAfile and -CApath options to the crl program, these
8118 will lookup a CRL issuers certificate and verify the signature in a
8119 similar way to the verify program. Tidy up the crl program so it
8120 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8121 less strict. It will now permit CRL extensions even if it is not
8122 a V2 CRL: this will allow it to tolerate some broken CRLs.
8125 *) Initialize all non-automatic variables each time one of the openssl
8126 sub-programs is started (this is necessary as they may be started
8127 multiple times from the "OpenSSL>" prompt).
8128 [Lennart Bang, Bodo Moeller]
8130 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8131 removing all other RSA functionality (this is what NO_RSA does). This
8132 is so (for example) those in the US can disable those operations covered
8133 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8137 *) Non-copying interface to BIO pairs.
8138 (still largely untested)
8141 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8142 ASCII string. This was handled independently in various places before.
8145 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8146 UTF8 strings a character at a time.
8149 *) Use client_version from client hello to select the protocol
8150 (s23_srvr.c) and for RSA client key exchange verification
8151 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8154 *) Add various utility functions to handle SPKACs, these were previously
8155 handled by poking round in the structure internals. Added new function
8156 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8157 print, verify and generate SPKACs. Based on an original idea from
8158 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8161 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8164 *) Allow the config file extension section to be overwritten on the
8165 command line. Based on an original idea from Massimiliano Pala
8166 <madwolf@comune.modena.it>. The new option is called -extensions
8167 and can be applied to ca, req and x509. Also -reqexts to override
8168 the request extensions in req and -crlexts to override the crl extensions
8172 *) Add new feature to the SPKAC handling in ca. Now you can include
8173 the same field multiple times by preceding it by "XXXX." for example:
8176 this is the same syntax as used in the req config file.
8179 *) Allow certificate extensions to be added to certificate requests. These
8180 are specified in a 'req_extensions' option of the req section of the
8181 config file. They can be printed out with the -text option to req but
8182 are otherwise ignored at present.
8185 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8186 data read consists of only the final block it would not decrypted because
8187 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8188 A misplaced 'break' also meant the decrypted final block might not be
8189 copied until the next read.
8192 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8193 a few extra parameters to the DH structure: these will be useful if
8194 for example we want the value of 'q' or implement X9.42 DH.
8197 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8198 provides hooks that allow the default DSA functions or functions on a
8199 "per key" basis to be replaced. This allows hardware acceleration and
8200 hardware key storage to be handled without major modification to the
8201 library. Also added low level modexp hooks and CRYPTO_EX structure and
8202 associated functions.
8205 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8206 as "read only": it can't be written to and the buffer it points to will
8207 not be freed. Reading from a read only BIO is much more efficient than
8208 a normal memory BIO. This was added because there are several times when
8209 an area of memory needs to be read from a BIO. The previous method was
8210 to create a memory BIO and write the data to it, this results in two
8211 copies of the data and an O(n^2) reading algorithm. There is a new
8212 function BIO_new_mem_buf() which creates a read only memory BIO from
8213 an area of memory. Also modified the PKCS#7 routines to use read only
8217 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8218 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8219 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8220 but a retry condition occured while trying to read the rest.
8223 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8224 NID_pkcs7_encrypted by default: this was wrong since this should almost
8225 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8226 the encrypted data type: this is a more sensible place to put it and it
8227 allows the PKCS#12 code to be tidied up that duplicated this
8231 *) Changed obj_dat.pl script so it takes its input and output files on
8232 the command line. This should avoid shell escape redirection problems
8236 *) Initial support for certificate extension requests, these are included
8237 in things like Xenroll certificate requests. Included functions to allow
8238 extensions to be obtained and added.
8241 *) -crlf option to s_client and s_server for sending newlines as
8242 CRLF (as required by many protocols).
8245 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8247 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8248 [Ralf S. Engelschall]
8250 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8251 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8253 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8257 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8258 DH parameters/keys (q is lost during that conversion, but the resulting
8259 DH parameters contain its length).
8261 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8262 much faster than DH_generate_parameters (which creates parameters
8263 where p = 2*q + 1), and also the smaller q makes DH computations
8264 much more efficient (160-bit exponentiation instead of 1024-bit
8265 exponentiation); so this provides a convenient way to support DHE
8266 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8267 utter importance to use
8268 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8270 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8271 when such DH parameters are used, because otherwise small subgroup
8272 attacks may become possible!
8275 *) Avoid memory leak in i2d_DHparams.
8278 *) Allow the -k option to be used more than once in the enc program:
8279 this allows the same encrypted message to be read by multiple recipients.
8282 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8283 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8284 it will always use the numerical form of the OID, even if it has a short
8288 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8289 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8290 otherwise bn_mod_exp was called. In the case of hardware keys for example
8291 no private key components need be present and it might store extra data
8292 in the RSA structure, which cannot be accessed from bn_mod_exp.
8293 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8294 private key operations.
8297 *) Added support for SPARC Linux.
8300 *) pem_password_cb function type incompatibly changed from
8301 typedef int pem_password_cb(char *buf, int size, int rwflag);
8303 ....(char *buf, int size, int rwflag, void *userdata);
8304 so that applications can pass data to their callbacks:
8305 The PEM[_ASN1]_{read,write}... functions and macros now take an
8306 additional void * argument, which is just handed through whenever
8307 the password callback is called.
8308 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8310 New function SSL_CTX_set_default_passwd_cb_userdata.
8312 Compatibility note: As many C implementations push function arguments
8313 onto the stack in reverse order, the new library version is likely to
8314 interoperate with programs that have been compiled with the old
8315 pem_password_cb definition (PEM_whatever takes some data that
8316 happens to be on the stack as its last argument, and the callback
8317 just ignores this garbage); but there is no guarantee whatsoever that
8320 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8321 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8322 problems not only on Windows, but also on some Unix platforms.
8323 To avoid problematic command lines, these definitions are now in an
8324 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8325 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8328 *) MIPS III/IV assembler module is reimplemented.
8331 *) More DES library cleanups: remove references to srand/rand and
8332 delete an unused file.
8335 *) Add support for the the free Netwide assembler (NASM) under Win32,
8336 since not many people have MASM (ml) and it can be hard to obtain.
8337 This is currently experimental but it seems to work OK and pass all
8338 the tests. Check out INSTALL.W32 for info.
8341 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8342 without temporary keys kept an extra copy of the server key,
8343 and connections with temporary keys did not free everything in case
8347 *) New function RSA_check_key and new openssl rsa option -check
8348 for verifying the consistency of RSA keys.
8349 [Ulf Moeller, Bodo Moeller]
8351 *) Various changes to make Win32 compile work:
8352 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8353 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8354 comparison" warnings.
8355 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8358 *) Add a debugging option to PKCS#5 v2 key generation function: when
8359 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8360 derived keys are printed to stderr.
8363 *) Copy the flags in ASN1_STRING_dup().
8364 [Roman E. Pavlov <pre@mo.msk.ru>]
8366 *) The x509 application mishandled signing requests containing DSA
8367 keys when the signing key was also DSA and the parameters didn't match.
8369 It was supposed to omit the parameters when they matched the signing key:
8370 the verifying software was then supposed to automatically use the CA's
8371 parameters if they were absent from the end user certificate.
8373 Omitting parameters is no longer recommended. The test was also
8374 the wrong way round! This was probably due to unusual behaviour in
8375 EVP_cmp_parameters() which returns 1 if the parameters match.
8376 This meant that parameters were omitted when they *didn't* match and
8377 the certificate was useless. Certificates signed with 'ca' didn't have
8379 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8381 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8382 The interface is as follows:
8383 Applications can use
8384 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8385 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8386 "off" is now the default.
8387 The library internally uses
8388 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8389 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8390 to disable memory-checking temporarily.
8392 Some inconsistent states that previously were possible (and were
8393 even the default) are now avoided.
8395 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8396 with each memory chunk allocated; this is occasionally more helpful
8397 than just having a counter.
8399 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8401 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8405 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8406 which largely parallels "options", but is for changing API behaviour,
8407 whereas "options" are about protocol behaviour.
8408 Initial "mode" flags are:
8410 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8411 a single record has been written.
8412 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8413 retries use the same buffer location.
8414 (But all of the contents must be
8418 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8421 *) Fix problems with no-hmac etc.
8422 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8424 *) New functions RSA_get_default_method(), RSA_set_method() and
8425 RSA_get_method(). These allows replacement of RSA_METHODs without having
8426 to mess around with the internals of an RSA structure.
8429 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8430 Also really enable memory leak checks in openssl.c and in some
8432 [Chad C. Mulligan, Bodo Moeller]
8434 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8435 up the length of negative integers. This has now been simplified to just
8436 store the length when it is first determined and use it later, rather
8437 than trying to keep track of where data is copied and updating it to
8439 [Steve Henson, reported by Brien Wheeler
8440 <bwheeler@authentica-security.com>]
8442 *) Add a new function PKCS7_signatureVerify. This allows the verification
8443 of a PKCS#7 signature but with the signing certificate passed to the
8444 function itself. This contrasts with PKCS7_dataVerify which assumes the
8445 certificate is present in the PKCS#7 structure. This isn't always the
8446 case: certificates can be omitted from a PKCS#7 structure and be
8447 distributed by "out of band" means (such as a certificate database).
8450 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8451 function prototypes in pem.h, also change util/mkdef.pl to add the
8452 necessary function names.
8455 *) mk1mf.pl (used by Windows builds) did not properly read the
8456 options set by Configure in the top level Makefile, and Configure
8457 was not even able to write more than one option correctly.
8458 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8461 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8462 file to be loaded from a BIO or FILE pointer. The BIO version will
8463 for example allow memory BIOs to contain config info.
8466 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8467 Whoever hopes to achieve shared-library compatibility across versions
8468 must use this, not the compile-time macro.
8469 (Exercise 0.9.4: Which is the minimum library version required by
8471 Note: All this applies only to multi-threaded programs, others don't
8475 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8476 through a BIO pair triggered the default case, i.e.
8477 SSLerr(...,SSL_R_UNKNOWN_STATE).
8480 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8481 can use the SSL library even if none of the specific BIOs is
8485 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8486 for the encoded length.
8487 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8489 *) Add initial documentation of the X509V3 functions.
8492 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8493 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8494 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8495 secure PKCS#8 private key format with a high iteration count.
8498 *) Fix determination of Perl interpreter: A perl or perl5
8499 _directory_ in $PATH was also accepted as the interpreter.
8500 [Ralf S. Engelschall]
8502 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8503 wrong with it but it was very old and did things like calling
8504 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8508 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8509 to use the new extension code.
8512 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8513 with macros. This should make it easier to change their form, add extra
8514 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8518 *) Add to configuration table a new entry that can specify an alternative
8519 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8520 according to Mark Crispin <MRC@Panda.COM>.
8524 *) DES CBC did not update the IV. Weird.
8527 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8528 Changing the behaviour of the former might break existing programs --
8529 where IV updating is needed, des_ncbc_encrypt can be used.
8532 *) When bntest is run from "make test" it drives bc to check its
8533 calculations, as well as internally checking them. If an internal check
8534 fails, it needs to cause bc to give a non-zero result or make test carries
8535 on without noticing the failure. Fixed.
8538 *) DES library cleanups.
8541 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8542 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8543 ciphers. NOTE: although the key derivation function has been verified
8544 against some published test vectors it has not been extensively tested
8545 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8549 *) Instead of "mkdir -p", which is not fully portable, use new
8550 Perl script "util/mkdir-p.pl".
8553 *) Rewrite the way password based encryption (PBE) is handled. It used to
8554 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8555 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8556 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8557 the 'parameter' field of the AlgorithmIdentifier is passed to the
8558 underlying key generation function so it must do its own ASN1 parsing.
8559 This has also changed the EVP_PBE_CipherInit() function which now has a
8560 'parameter' argument instead of literal salt and iteration count values
8561 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8564 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8565 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8566 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8567 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8568 value was just used as a "magic string" and not used directly its
8569 value doesn't matter.
8572 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8576 *) "linux-sparc64" configuration (ultrapenguin).
8577 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8578 "linux-sparc" configuration.
8579 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8581 *) config now generates no-xxx options for missing ciphers.
8584 *) Support the EBCDIC character set (work in progress).
8585 File ebcdic.c not yet included because it has a different license.
8586 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8588 *) Support BS2000/OSD-POSIX.
8589 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8591 *) Make callbacks for key generation use void * instead of char *.
8594 *) Make S/MIME samples compile (not yet tested).
8597 *) Additional typesafe stacks.
8600 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8604 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8606 *) New configuration variant "sco5-gcc".
8608 *) Updated some demos.
8609 [Sean O Riordain, Wade Scholine]
8611 *) Add missing BIO_free at exit of pkcs12 application.
8614 *) Fix memory leak in conf.c.
8617 *) Updates for Win32 to assembler version of MD5.
8620 *) Set #! path to perl in apps/der_chop to where we found it
8621 instead of using a fixed path.
8624 *) SHA library changes for irix64-mips4-cc.
8627 *) Improvements for VMS support.
8631 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8633 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8634 This also avoids the problems with SC4.2 and unpatched SC5.
8635 [Andy Polyakov <appro@fy.chalmers.se>]
8637 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8638 These are required because of the typesafe stack would otherwise break
8639 existing code. If old code used a structure member which used to be STACK
8640 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8641 sk_num or sk_value it would produce an error because the num, data members
8642 are not present in STACK_OF. Now it just produces a warning. sk_set
8643 replaces the old method of assigning a value to sk_value
8644 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8645 that does this will no longer work (and should use sk_set instead) but
8646 this could be regarded as a "questionable" behaviour anyway.
8649 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8650 correctly handle encrypted S/MIME data.
8653 *) Change type of various DES function arguments from des_cblock
8654 (which means, in function argument declarations, pointer to char)
8655 to des_cblock * (meaning pointer to array with 8 char elements),
8656 which allows the compiler to do more typechecking; it was like
8657 that back in SSLeay, but with lots of ugly casts.
8659 Introduce new type const_des_cblock.
8662 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8663 problems: find RecipientInfo structure that matches recipient certificate
8664 and initialise the ASN1 structures properly based on passed cipher.
8667 *) Belatedly make the BN tests actually check the results.
8670 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8671 to and from BNs: it was completely broken. New compilation option
8672 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8673 key elements as negative integers.
8676 *) Reorganize and speed up MD5.
8677 [Andy Polyakov <appro@fy.chalmers.se>]
8680 [Richard Levitte <richard@levitte.org>]
8682 *) New option -out to asn1parse to allow the parsed structure to be
8683 output to a file. This is most useful when combined with the -strparse
8684 option to examine the output of things like OCTET STRINGS.
8687 *) Make SSL library a little more fool-proof by not requiring any longer
8688 that SSL_set_{accept,connect}_state be called before
8689 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8690 in many applications because usually everything *appeared* to work as
8691 intended anyway -- now it really works as intended).
8694 *) Move openssl.cnf out of lib/.
8697 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8698 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8699 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8700 [Ralf S. Engelschall]
8702 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8703 handle PKCS#7 enveloped data properly.
8704 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8706 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8707 copying pointers. The cert_st handling is changed by this in
8708 various ways (and thus what used to be known as ctx->default_cert
8709 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8710 any longer when s->cert does not give us what we need).
8711 ssl_cert_instantiate becomes obsolete by this change.
8712 As soon as we've got the new code right (possibly it already is?),
8713 we have solved a couple of bugs of the earlier code where s->cert
8714 was used as if it could not have been shared with other SSL structures.
8716 Note that using the SSL API in certain dirty ways now will result
8717 in different behaviour than observed with earlier library versions:
8718 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8719 does not influence s as it used to.
8721 In order to clean up things more thoroughly, inside SSL_SESSION
8722 we don't use CERT any longer, but a new structure SESS_CERT
8723 that holds per-session data (if available); currently, this is
8724 the peer's certificate chain and, for clients, the server's certificate
8725 and temporary key. CERT holds only those values that can have
8726 meaningful defaults in an SSL_CTX.
8729 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8730 from the internal representation. Various PKCS#7 fixes: remove some
8731 evil casts and set the enc_dig_alg field properly based on the signing
8735 *) Allow PKCS#12 password to be set from the command line or the
8736 environment. Let 'ca' get its config file name from the environment
8737 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8741 *) Allow certificate policies extension to use an IA5STRING for the
8742 organization field. This is contrary to the PKIX definition but
8743 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8747 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8748 without disallowing inline assembler and the like for non-pedantic builds.
8751 *) Support Borland C++ builder.
8752 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8757 *) SHA-1 cleanups and performance enhancements.
8758 [Andy Polyakov <appro@fy.chalmers.se>]
8760 *) Sparc v8plus assembler for the bignum library.
8761 [Andy Polyakov <appro@fy.chalmers.se>]
8763 *) Accept any -xxx and +xxx compiler options in Configure.
8766 *) Update HPUX configuration.
8769 *) Add missing sk_<type>_unshift() function to safestack.h
8770 [Ralf S. Engelschall]
8772 *) New function SSL_CTX_use_certificate_chain_file that sets the
8773 "extra_cert"s in addition to the certificate. (This makes sense
8774 only for "PEM" format files, as chains as a whole are not
8778 *) Support verify_depth from the SSL API.
8779 x509_vfy.c had what can be considered an off-by-one-error:
8780 Its depth (which was not part of the external interface)
8781 was actually counting the number of certificates in a chain;
8782 now it really counts the depth.
8785 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8786 instead of X509err, which often resulted in confusing error
8787 messages since the error codes are not globally unique
8788 (e.g. an alleged error in ssl3_accept when a certificate
8789 didn't match the private key).
8791 *) New function SSL_CTX_set_session_id_context that allows to set a default
8792 value (so that you don't need SSL_set_session_id_context for each
8793 connection using the SSL_CTX).
8796 *) OAEP decoding bug fix.
8799 *) Support INSTALL_PREFIX for package builders, as proposed by
8803 *) New Configure options "threads" and "no-threads". For systems
8804 where the proper compiler options are known (currently Solaris
8805 and Linux), "threads" is the default.
8808 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8811 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8812 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8813 such as /usr/local/bin.
8816 *) "make linux-shared" to build shared libraries.
8817 [Niels Poppe <niels@netbox.org>]
8819 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8822 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8823 extension adding in x509 utility.
8826 *) Remove NOPROTO sections and error code comments.
8829 *) Partial rewrite of the DEF file generator to now parse the ANSI
8833 *) New Configure options --prefix=DIR and --openssldir=DIR.
8836 *) Complete rewrite of the error code script(s). It is all now handled
8837 by one script at the top level which handles error code gathering,
8838 header rewriting and C source file generation. It should be much better
8839 than the old method: it now uses a modified version of Ulf's parser to
8840 read the ANSI prototypes in all header files (thus the old K&R definitions
8841 aren't needed for error creation any more) and do a better job of
8842 translating function codes into names. The old 'ASN1 error code imbedded
8843 in a comment' is no longer necessary and it doesn't use .err files which
8844 have now been deleted. Also the error code call doesn't have to appear all
8845 on one line (which resulted in some large lines...).
8848 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8851 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8852 0 (which usually indicates a closed connection), but continue reading.
8855 *) Fix some race conditions.
8858 *) Add support for CRL distribution points extension. Add Certificate
8859 Policies and CRL distribution points documentation.
8862 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8865 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8866 8 of keying material. Merlin has also confirmed interop with this fix
8867 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8868 [Merlin Hughes <merlin@baltimore.ie>]
8870 *) Fix lots of warnings.
8871 [Richard Levitte <levitte@stacken.kth.se>]
8873 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8874 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8875 [Richard Levitte <levitte@stacken.kth.se>]
8877 *) Fix problems with sizeof(long) == 8.
8878 [Andy Polyakov <appro@fy.chalmers.se>]
8880 *) Change functions to ANSI C.
8883 *) Fix typos in error codes.
8884 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8886 *) Remove defunct assembler files from Configure.
8889 *) SPARC v8 assembler BIGNUM implementation.
8890 [Andy Polyakov <appro@fy.chalmers.se>]
8892 *) Support for Certificate Policies extension: both print and set.
8893 Various additions to support the r2i method this uses.
8896 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8897 return a const string when you are expecting an allocated buffer.
8900 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8901 types DirectoryString and DisplayText.
8904 *) Add code to allow r2i extensions to access the configuration database,
8905 add an LHASH database driver and add several ctx helper functions.
8908 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8909 fail when they extended the size of a BIGNUM.
8912 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8913 support typesafe stack.
8916 *) Fix typo in SSL_[gs]et_options().
8917 [Nils Frostberg <nils@medcom.se>]
8919 *) Delete various functions and files that belonged to the (now obsolete)
8920 old X509V3 handling code.
8923 *) New Configure option "rsaref".
8926 *) Don't auto-generate pem.h.
8929 *) Introduce type-safe ASN.1 SETs.
8932 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8933 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8935 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8936 that links with OpenSSL (well at least cause lots of warnings), but fear
8937 not: the conversion is trivial, and it eliminates loads of evil casts. A
8938 few STACKed things have been converted already. Feel free to convert more.
8939 In the fullness of time, I'll do away with the STACK type altogether.
8942 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8943 specified in <certfile> by updating the entry in the index.txt file.
8944 This way one no longer has to edit the index.txt file manually for
8945 revoking a certificate. The -revoke option does the gory details now.
8946 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8948 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8949 `-text' option at all and this way the `-noout -text' combination was
8950 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8951 [Ralf S. Engelschall]
8953 *) Make sure a corresponding plain text error message exists for the
8954 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8955 verify callback function determined that a certificate was revoked.
8956 [Ralf S. Engelschall]
8958 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8959 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8960 all available cipers including rc5, which was forgotten until now.
8961 In order to let the testing shell script know which algorithms
8962 are available, a new (up to now undocumented) command
8963 "openssl list-cipher-commands" is used.
8966 *) Bugfix: s_client occasionally would sleep in select() when
8967 it should have checked SSL_pending() first.
8970 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8971 the raw DSA values prior to ASN.1 encoding.
8974 *) Tweaks to Configure
8975 [Niels Poppe <niels@netbox.org>]
8977 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8981 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8984 *) New config option to avoid instructions that are illegal on the 80386.
8985 The default code is faster, but requires at least a 486.
8988 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8989 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8990 same as SSL2_VERSION anyway.
8993 *) New "-showcerts" option for s_client.
8996 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8997 application. Various cleanups and fixes.
9000 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9001 modify error routines to work internally. Add error codes and PBE init
9002 to library startup routines.
9005 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9006 packing functions to asn1 and evp. Changed function names and error
9007 codes along the way.
9010 *) PKCS12 integration: and so it begins... First of several patches to
9011 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9012 objects to objects.h
9015 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9016 and display support for Thawte strong extranet extension.
9019 *) Add LinuxPPC support.
9020 [Jeff Dubrule <igor@pobox.org>]
9022 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9023 bn_div_words in alpha.s.
9024 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9026 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9027 OAEP isn't supported when OpenSSL is built with RSAref.
9028 [Ulf Moeller <ulf@fitug.de>]
9030 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9031 so they no longer are missing under -DNOPROTO.
9032 [Soren S. Jorvang <soren@t.dk>]
9035 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9037 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9038 doesn't work when the session is reused. Coming soon!
9041 *) Fix a security hole, that allows sessions to be reused in the wrong
9042 context thus bypassing client cert protection! All software that uses
9043 client certs and session caches in multiple contexts NEEDS PATCHING to
9044 allow session reuse! A fuller solution is in the works.
9045 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9047 *) Some more source tree cleanups (removed obsolete files
9048 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9049 permission on "config" script to be executable) and a fix for the INSTALL
9051 [Ulf Moeller <ulf@fitug.de>]
9053 *) Remove some legacy and erroneous uses of malloc, free instead of
9055 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9057 *) Make rsa_oaep_test return non-zero on error.
9058 [Ulf Moeller <ulf@fitug.de>]
9060 *) Add support for native Solaris shared libraries. Configure
9061 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9062 if someone would make that last step automatic.
9063 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9065 *) ctx_size was not built with the right compiler during "make links". Fixed.
9068 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9069 except NULL ciphers". This means the default cipher list will no longer
9070 enable NULL ciphers. They need to be specifically enabled e.g. with
9071 the string "DEFAULT:eNULL".
9074 *) Fix to RSA private encryption routines: if p < q then it would
9075 occasionally produce an invalid result. This will only happen with
9076 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9079 *) Be less restrictive and allow also `perl util/perlpath.pl
9080 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9081 because this way one can also use an interpreter named `perl5' (which is
9082 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9083 installed as `perl').
9084 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9086 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9087 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9089 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9090 advapi32.lib to Win32 build and change the pem test comparision
9091 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9092 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9093 and crypto/des/ede_cbcm_enc.c.
9096 *) DES quad checksum was broken on big-endian architectures. Fixed.
9099 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9100 Win32 test batch file so it (might) work again. The Win32 test batch file
9101 is horrible: I feel ill....
9104 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9105 in e_os.h. Audit of header files to check ANSI and non ANSI
9106 sections: 10 functions were absent from non ANSI section and not exported
9107 from Windows DLLs. Fixed up libeay.num for new functions.
9110 *) Make `openssl version' output lines consistent.
9111 [Ralf S. Engelschall]
9113 *) Fix Win32 symbol export lists for BIO functions: Added
9114 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9115 to ms/libeay{16,32}.def.
9116 [Ralf S. Engelschall]
9118 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9119 fine under Unix and passes some trivial tests I've now added. But the
9120 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9121 added to make sure no one expects that this stuff really works in the
9122 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9123 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9125 [Ralf S. Engelschall]
9127 *) Fix the generation of two part addresses in perl.
9128 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9130 *) Add config entry for Linux on MIPS.
9131 [John Tobey <jtobey@channel1.com>]
9133 *) Make links whenever Configure is run, unless we are on Windoze.
9136 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9137 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9141 *) Add a useful kludge to allow package maintainers to specify compiler and
9142 other platforms details on the command line without having to patch the
9143 Configure script everytime: One now can use ``perl Configure
9144 <id>:<details>'', i.e. platform ids are allowed to have details appended
9145 to them (seperated by colons). This is treated as there would be a static
9146 pre-configured entry in Configure's %table under key <id> with value
9147 <details> and ``perl Configure <id>'' is called. So, when you want to
9148 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9149 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9150 now, which overrides the FreeBSD-elf entry on-the-fly.
9151 [Ralf S. Engelschall]
9153 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9156 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9157 on the `perl Configure ...' command line. This way one can compile
9158 OpenSSL libraries with Position Independent Code (PIC) which is needed
9159 for linking it into DSOs.
9160 [Ralf S. Engelschall]
9162 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9166 *) Cleaned up the LICENSE document: The official contact for any license
9167 questions now is the OpenSSL core team under openssl-core@openssl.org.
9168 And add a paragraph about the dual-license situation to make sure people
9169 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9170 to the OpenSSL toolkit.
9171 [Ralf S. Engelschall]
9173 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9174 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9175 Additonally cleaned up the `make links' target: Remove unnecessary
9176 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9177 to speed processing and no longer clutter the display with confusing
9178 stuff. Instead only the actually done links are displayed.
9179 [Ralf S. Engelschall]
9181 *) Permit null encryption ciphersuites, used for authentication only. It used
9182 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9183 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9187 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9188 signed attributes when verifying signatures (this would break them),
9189 the detached data encoding was wrong and public keys obtained using
9190 X509_get_pubkey() weren't freed.
9193 *) Add text documentation for the BUFFER functions. Also added a work around
9194 to a Win95 console bug. This was triggered by the password read stuff: the
9195 last character typed gets carried over to the next fread(). If you were
9196 generating a new cert request using 'req' for example then the last
9197 character of the passphrase would be CR which would then enter the first
9201 *) Added the new `Includes OpenSSL Cryptography Software' button as
9202 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9203 button and can be used by applications based on OpenSSL to show the
9204 relationship to the OpenSSL project.
9205 [Ralf S. Engelschall]
9207 *) Remove confusing variables in function signatures in files
9208 ssl/ssl_lib.c and ssl/ssl.h.
9209 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9211 *) Don't install bss_file.c under PREFIX/include/
9212 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9214 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9215 functions that return function pointers and has support for NT specific
9216 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9217 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9218 unsigned to signed types: this was killing the Win32 compile.
9221 *) Add new certificate file to stack functions,
9222 SSL_add_dir_cert_subjects_to_stack() and
9223 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9224 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9225 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9226 This means that Apache-SSL and similar packages don't have to mess around
9227 to add as many CAs as they want to the preferred list.
9230 *) Experiment with doxygen documentation. Currently only partially applied to
9232 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9233 openssl.doxy as the configuration file.
9236 *) Get rid of remaining C++-style comments which strict C compilers hate.
9237 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9239 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9240 compiled in by default: it has problems with large keys.
9243 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9244 DH private keys and/or callback functions which directly correspond to
9245 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9246 is needed for applications which have to configure certificates on a
9247 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9249 For the RSA certificate situation is makes no difference, but
9250 for the DSA certificate situation this fixes the "no shared cipher"
9251 problem where the OpenSSL cipher selection procedure failed because the
9252 temporary keys were not overtaken from the context and the API provided
9253 no way to reconfigure them.
9254 The new functions now let applications reconfigure the stuff and they
9255 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9256 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9257 non-public-API function ssl_cert_instantiate() is used as a helper
9258 function and also to reduce code redundancy inside ssl_rsa.c.
9259 [Ralf S. Engelschall]
9261 *) Move s_server -dcert and -dkey options out of the undocumented feature
9262 area because they are useful for the DSA situation and should be
9263 recognized by the users.
9264 [Ralf S. Engelschall]
9266 *) Fix the cipher decision scheme for export ciphers: the export bits are
9267 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9268 SSL_EXP_MASK. So, the original variable has to be used instead of the
9269 already masked variable.
9270 [Richard Levitte <levitte@stacken.kth.se>]
9272 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9273 [Richard Levitte <levitte@stacken.kth.se>]
9275 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9276 from `int' to `unsigned int' because it's a length and initialized by
9277 EVP_DigestFinal() which expects an `unsigned int *'.
9278 [Richard Levitte <levitte@stacken.kth.se>]
9280 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9281 script. Instead use the usual Shell->Perl transition trick.
9282 [Ralf S. Engelschall]
9284 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9285 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9286 -noout -modulus' as it's already the case for `openssl rsa -noout
9287 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9288 currently the public key is printed (a decision which was already done by
9289 `openssl dsa -modulus' in the past) which serves a similar purpose.
9290 Additionally the NO_RSA no longer completely removes the whole -modulus
9291 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9293 [Ralf S. Engelschall]
9295 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9296 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9297 [Arne Ansper <arne@ats.cyber.ee>]
9299 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9300 to be added. Now both 'req' and 'ca' can use new objects defined in the
9304 *) Add cool BIO that does syslog (or event log on NT).
9305 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9307 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9308 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9309 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9310 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9313 *) Add preliminary config info for new extension code.
9316 *) Make RSA_NO_PADDING really use no padding.
9317 [Ulf Moeller <ulf@fitug.de>]
9319 *) Generate errors when private/public key check is done.
9322 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9323 for some CRL extensions and new objects added.
9326 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9327 key usage extension and fuller support for authority key id.
9330 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9331 padding method for RSA, which is recommended for new applications in PKCS
9332 #1 v2.0 (RFC 2437, October 1998).
9333 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9334 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9335 against Bleichbacher's attack on RSA.
9336 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9339 *) Updates to the new SSL compression code
9340 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9342 *) Fix so that the version number in the master secret, when passed
9343 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9344 (because the server will not accept higher), that the version number
9345 is 0x03,0x01, not 0x03,0x00
9346 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9348 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9349 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9350 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9353 *) Support for RAW extensions where an arbitrary extension can be
9354 created by including its DER encoding. See apps/openssl.cnf for
9358 *) Make sure latest Perl versions don't interpret some generated C array
9359 code as Perl array code in the crypto/err/err_genc.pl script.
9360 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9362 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9363 not many people have the assembler. Various Win32 compilation fixes and
9364 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9368 *) Modify configure script 'Configure' to automatically create crypto/date.h
9369 file under Win32 and also build pem.h from pem.org. New script
9370 util/mkfiles.pl to create the MINFO file on environments that can't do a
9371 'make files': perl util/mkfiles.pl >MINFO should work.
9374 *) Major rework of DES function declarations, in the pursuit of correctness
9375 and purity. As a result, many evil casts evaporated, and some weirdness,
9376 too. You may find this causes warnings in your code. Zapping your evil
9377 casts will probably fix them. Mostly.
9380 *) Fix for a typo in asn1.h. Bug fix to object creation script
9381 obj_dat.pl. It considered a zero in an object definition to mean
9382 "end of object": none of the objects in objects.h have any zeros
9383 so it wasn't spotted.
9384 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9386 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9387 Masking (CBCM). In the absence of test vectors, the best I have been able
9388 to do is check that the decrypt undoes the encrypt, so far. Send me test
9389 vectors if you have them.
9392 *) Correct calculation of key length for export ciphers (too much space was
9393 allocated for null ciphers). This has not been tested!
9396 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9397 message is now correct (it understands "crypto" and "ssl" on its
9398 command line). There is also now an "update" option. This will update
9399 the util/ssleay.num and util/libeay.num files with any new functions.
9401 perl util/mkdef.pl crypto ssl update
9402 it will update them.
9405 *) Overhauled the Perl interface (perl/*):
9406 - ported BN stuff to OpenSSL's different BN library
9407 - made the perl/ source tree CVS-aware
9408 - renamed the package from SSLeay to OpenSSL (the files still contain
9409 their history because I've copied them in the repository)
9410 - removed obsolete files (the test scripts will be replaced
9411 by better Test::Harness variants in the future)
9412 [Ralf S. Engelschall]
9414 *) First cut for a very conservative source tree cleanup:
9415 1. merge various obsolete readme texts into doc/ssleay.txt
9416 where we collect the old documents and readme texts.
9417 2. remove the first part of files where I'm already sure that we no
9418 longer need them because of three reasons: either they are just temporary
9419 files which were left by Eric or they are preserved original files where
9420 I've verified that the diff is also available in the CVS via "cvs diff
9421 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9422 the crypto/md/ stuff).
9423 [Ralf S. Engelschall]
9425 *) More extension code. Incomplete support for subject and issuer alt
9426 name, issuer and authority key id. Change the i2v function parameters
9427 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9428 what that's for :-) Fix to ASN1 macro which messed up
9429 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9432 *) Preliminary support for ENUMERATED type. This is largely copied from the
9436 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9437 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9439 *) Make sure `make rehash' target really finds the `openssl' program.
9440 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9442 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9443 like to hear about it if this slows down other processors.
9446 *) Add CygWin32 platform information to Configure script.
9447 [Alan Batie <batie@aahz.jf.intel.com>]
9449 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9450 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9452 *) New program nseq to manipulate netscape certificate sequences
9455 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9459 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9460 but the BN code had some problems that would cause failures when
9461 doing certificate verification and some other functions.
9462 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9464 *) Add ASN1 and PEM code to support netscape certificate sequences.
9467 *) Add ASN1 and PEM code to support netscape certificate sequences.
9470 *) Add several PKIX and private extended key usage OIDs.
9473 *) Modify the 'ca' program to handle the new extension code. Modify
9474 openssl.cnf for new extension format, add comments.
9477 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9478 and add a sample to openssl.cnf so req -x509 now adds appropriate
9482 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9483 error code, add initial support to X509_print() and x509 application.
9486 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9487 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9488 stuff is currently isolated and isn't even compiled yet.
9491 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9492 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9493 Removed the versions check from X509 routines when loading extensions:
9494 this allows certain broken certificates that don't set the version
9495 properly to be processed.
9498 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9499 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9500 can still be regenerated with "make depend".
9503 *) Spelling mistake in C version of CAST-128.
9504 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9506 *) Changes to the error generation code. The perl script err-code.pl
9507 now reads in the old error codes and retains the old numbers, only
9508 adding new ones if necessary. It also only changes the .err files if new
9509 codes are added. The makefiles have been modified to only insert errors
9510 when needed (to avoid needlessly modifying header files). This is done
9511 by only inserting errors if the .err file is newer than the auto generated
9512 C file. To rebuild all the error codes from scratch (the old behaviour)
9513 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9514 or delete all the .err files.
9517 *) CAST-128 was incorrectly implemented for short keys. The C version has
9518 been fixed, but is untested. The assembler versions are also fixed, but
9519 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9520 to regenerate it if needed.
9521 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9522 Hagino <itojun@kame.net>]
9524 *) File was opened incorrectly in randfile.c.
9525 [Ulf Möller <ulf@fitug.de>]
9527 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9528 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9529 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9530 al: it's just almost always a UTCTime. Note this patch adds new error
9531 codes so do a "make errors" if there are problems.
9534 *) Correct Linux 1 recognition in config.
9535 [Ulf Möller <ulf@fitug.de>]
9537 *) Remove pointless MD5 hash when using DSA keys in ca.
9538 [Anonymous <nobody@replay.com>]
9540 *) Generate an error if given an empty string as a cert directory. Also
9541 generate an error if handed NULL (previously returned 0 to indicate an
9542 error, but didn't set one).
9543 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9545 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9548 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9549 parameters. This was causing a warning which killed off the Win32 compile.
9552 *) Remove C++ style comments from crypto/bn/bn_local.h.
9553 [Neil Costigan <neil.costigan@celocom.com>]
9555 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9556 based on a text string, looking up short and long names and finally
9557 "dot" format. The "dot" format stuff didn't work. Added new function
9558 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9559 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9560 OID is not part of the table.
9563 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9564 X509_LOOKUP_by_alias().
9567 *) Sort openssl functions by name.
9570 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9571 encryption from sample DSA keys (in case anyone is interested the password
9575 *) Make _all_ *_free functions accept a NULL pointer.
9576 [Frans Heymans <fheymans@isaserver.be>]
9578 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9580 [Anonymous <nobody@replay.com>]
9582 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9583 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9585 *) Don't blow it for numeric -newkey arguments to apps/req.
9586 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9588 *) Temp key "for export" tests were wrong in s3_srvr.c.
9589 [Anonymous <nobody@replay.com>]
9591 *) Add prototype for temp key callback functions
9592 SSL_CTX_set_tmp_{rsa,dh}_callback().
9595 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9596 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9599 *) X509_name_add_entry() freed the wrong thing after an error.
9600 [Arne Ansper <arne@ats.cyber.ee>]
9602 *) rsa_eay.c would attempt to free a NULL context.
9603 [Arne Ansper <arne@ats.cyber.ee>]
9605 *) BIO_s_socket() had a broken should_retry() on Windoze.
9606 [Arne Ansper <arne@ats.cyber.ee>]
9608 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9609 [Arne Ansper <arne@ats.cyber.ee>]
9611 *) Make sure the already existing X509_STORE->depth variable is initialized
9612 in X509_STORE_new(), but document the fact that this variable is still
9613 unused in the certificate verification process.
9614 [Ralf S. Engelschall]
9616 *) Fix the various library and apps files to free up pkeys obtained from
9617 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9620 *) Fix reference counting in X509_PUBKEY_get(). This makes
9621 demos/maurice/example2.c work, amongst others, probably.
9622 [Steve Henson and Ben Laurie]
9624 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9625 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9626 are no longer created. This way we have a single and consistent command
9627 line interface `openssl <command>', similar to `cvs <command>'.
9628 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9630 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9631 BIT STRING wrapper always have zero unused bits.
9634 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9637 *) Make the top-level INSTALL documentation easier to understand.
9640 *) Makefiles updated to exit if an error occurs in a sub-directory
9641 make (including if user presses ^C) [Paul Sutton]
9643 *) Make Montgomery context stuff explicit in RSA data structure.
9646 *) Fix build order of pem and err to allow for generated pem.h.
9649 *) Fix renumbering bug in X509_NAME_delete_entry().
9652 *) Enhanced the err-ins.pl script so it makes the error library number
9653 global and can add a library name. This is needed for external ASN1 and
9654 other error libraries.
9657 *) Fixed sk_insert which never worked properly.
9660 *) Fix ASN1 macros so they can handle indefinite length construted
9661 EXPLICIT tags. Some non standard certificates use these: they can now
9665 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9666 into a single doc/ssleay.txt bundle. This way the information is still
9667 preserved but no longer messes up this directory. Now it's new room for
9668 the new set of documenation files.
9669 [Ralf S. Engelschall]
9671 *) SETs were incorrectly DER encoded. This was a major pain, because they
9672 shared code with SEQUENCEs, which aren't coded the same. This means that
9673 almost everything to do with SETs or SEQUENCEs has either changed name or
9674 number of arguments.
9675 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9677 *) Fix test data to work with the above.
9680 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9681 was already fixed by Eric for 0.9.1 it seems.
9682 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9684 *) Autodetect FreeBSD3.
9687 *) Fix various bugs in Configure. This affects the following platforms:
9691 unixware-2.0-pentium
9695 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9696 before they are needed.
9699 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9703 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9705 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9706 changed SSLeay to OpenSSL in version strings.
9707 [Ralf S. Engelschall]
9709 *) Some fixups to the top-level documents.
9712 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9713 because the symlink to include/ was missing.
9714 [Ralf S. Engelschall]
9716 *) Incorporated the popular no-RSA/DSA-only patches
9717 which allow to compile a RSA-free SSLeay.
9718 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9720 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9721 when "ssleay" is still not found.
9722 [Ralf S. Engelschall]
9724 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9725 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9727 *) Updated the README file.
9728 [Ralf S. Engelschall]
9730 *) Added various .cvsignore files in the CVS repository subdirs
9731 to make a "cvs update" really silent.
9732 [Ralf S. Engelschall]
9734 *) Recompiled the error-definition header files and added
9735 missing symbols to the Win32 linker tables.
9736 [Ralf S. Engelschall]
9738 *) Cleaned up the top-level documents;
9739 o new files: CHANGES and LICENSE
9740 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9741 o merged COPYRIGHT into LICENSE
9742 o removed obsolete TODO file
9743 o renamed MICROSOFT to INSTALL.W32
9744 [Ralf S. Engelschall]
9746 *) Removed dummy files from the 0.9.1b source tree:
9747 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9748 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9749 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9750 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9751 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9752 [Ralf S. Engelschall]
9754 *) Added various platform portability fixes.
9757 *) The Genesis of the OpenSSL rpject:
9758 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9759 Young and Tim J. Hudson created while they were working for C2Net until
9761 [The OpenSSL Project]
9764 Changes between 0.9.0b and 0.9.1b [not released]
9766 *) Updated a few CA certificates under certs/
9769 *) Changed some BIGNUM api stuff.
9772 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9773 DGUX x86, Linux Alpha, etc.
9776 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9777 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9781 *) Add -strparse option to asn1pars program which parses nested
9783 [Dr Stephen Henson <shenson@bigfoot.com>]
9785 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9788 *) DSA fix for "ca" program.
9791 *) Added "-genkey" option to "dsaparam" program.
9794 *) Added RIPE MD160 (rmd160) message digest.
9797 *) Added -a (all) option to "ssleay version" command.
9800 *) Added PLATFORM define which is the id given to Configure.
9803 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9806 *) Extended the ASN.1 parser routines.
9809 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9812 *) Added a BN_CTX to the BN library.
9815 *) Fixed the weak key values in DES library
9818 *) Changed API in EVP library for cipher aliases.
9821 *) Added support for RC2/64bit cipher.
9824 *) Converted the lhash library to the crypto/mem.c functions.
9827 *) Added more recognized ASN.1 object ids.
9830 *) Added more RSA padding checks for SSL/TLS.
9833 *) Added BIO proxy/filter functionality.
9836 *) Added extra_certs to SSL_CTX which can be used
9837 send extra CA certificates to the client in the CA cert chain sending
9838 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9841 *) Now Fortezza is denied in the authentication phase because
9842 this is key exchange mechanism is not supported by SSLeay at all.
9845 *) Additional PKCS1 checks.
9848 *) Support the string "TLSv1" for all TLS v1 ciphers.
9851 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9852 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9855 *) Fixed a few memory leaks.
9858 *) Fixed various code and comment typos.
9861 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9862 bytes sent in the client random.
9863 [Edward Bishop <ebishop@spyglass.com>]