5 Changes between 1.0.0f and 1.0.1 [xx XXX xxxx]
7 *) Add TLS key material exporter from RFC 5705.
10 *) Add DTLS-SRTP negotiation from RFC 5764.
13 *) Add Next Protocol Negotiation,
14 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
15 disabled with a no-npn flag to config or Configure. Code donated
17 [Adam Langley <agl@google.com> and Ben Laurie]
19 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
20 NIST-P256, NIST-P521, with constant-time single point multiplication on
21 typical inputs. Compiler support for the nonstandard type __uint128_t is
22 required to use this (present in gcc 4.4 and later, for 64-bit builds).
23 Code made available under Apache License version 2.0.
25 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
26 line to include this in your build of OpenSSL, and run "make depend" (or
27 "make update"). This enables the following EC_METHODs:
29 EC_GFp_nistp224_method()
30 EC_GFp_nistp256_method()
31 EC_GFp_nistp521_method()
33 EC_GROUP_new_by_curve_name() will automatically use these (while
34 EC_GROUP_new_curve_GFp() currently prefers the more flexible
36 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
38 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
39 all platforms. Move ssize_t definition from e_os.h to the public
40 header file e_os2.h as it now appears in public header file cms.h
43 *) New -sigopt option to the ca, req and x509 utilities. Additional
44 signature parameters can be passed using this option and in
48 *) Add RSA PSS signing function. This will generate and set the
49 appropriate AlgorithmIdentifiers for PSS based on those in the
50 corresponding EVP_MD_CTX structure. No application support yet.
53 *) Support for companion algorithm specific ASN1 signing routines.
54 New function ASN1_item_sign_ctx() signs a pre-initialised
55 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
56 the appropriate parameters.
59 *) Add new algorithm specific ASN1 verification initialisation function
60 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
61 handling will be the same no matter what EVP_PKEY_METHOD is used.
62 Add a PSS handler to support verification of PSS signatures: checked
63 against a number of sample certificates.
66 *) Add signature printing for PSS. Add PSS OIDs.
67 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
69 *) Add algorithm specific signature printing. An individual ASN1 method
70 can now print out signatures instead of the standard hex dump.
72 More complex signatures (e.g. PSS) can print out more meaningful
73 information. Include DSA version that prints out the signature
77 *) Password based recipient info support for CMS library: implementing
81 *) Split password based encryption into PBES2 and PBKDF2 functions. This
82 neatly separates the code into cipher and PBE sections and is required
83 for some algorithms that split PBES2 into separate pieces (such as
87 *) Session-handling fixes:
88 - Fix handling of connections that are resuming with a session ID,
89 but also support Session Tickets.
90 - Fix a bug that suppressed issuing of a new ticket if the client
91 presented a ticket with an expired session.
92 - Try to set the ticket lifetime hint to something reasonable.
93 - Make tickets shorter by excluding irrelevant information.
94 - On the client side, don't ignore renewed tickets.
95 [Adam Langley, Bodo Moeller (Google)]
97 *) Fix PSK session representation.
100 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
102 This work was sponsored by Intel.
105 *) Add GCM support to TLS library. Some custom code is needed to split
106 the IV between the fixed (from PRF) and explicit (from TLS record)
107 portions. This adds all GCM ciphersuites supported by RFC5288 and
108 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
109 add a special AESGCM string for GCM only.
112 *) Expand range of ctrls for AES GCM. Permit setting invocation
113 field on decrypt and retrieval of invocation field only on encrypt.
116 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
117 As required by RFC5289 these ciphersuites cannot be used if for
118 versions of TLS earlier than 1.2.
121 *) For FIPS capable OpenSSL interpret a NULL default public key method
122 as unset and return the appopriate default but do *not* set the default.
123 This means we can return the appopriate method in applications that
124 swicth between FIPS and non-FIPS modes.
127 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
128 ENGINE is used then we cannot handle that in the FIPS module so we
129 keep original code iff non-FIPS operations are allowed.
132 *) Add -attime option to openssl verify.
133 [Peter Eckersley <pde@eff.org> and Ben Laurie]
135 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
138 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
139 FIPS EC methods unconditionally for now.
142 *) New build option no-ec2m to disable characteristic 2 code.
145 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
146 all cases can be covered as some introduce binary incompatibilities.
149 *) Redirect RSA operations to FIPS module including keygen,
150 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
153 *) Add similar low level API blocking to ciphers.
156 *) Low level digest APIs are not approved in FIPS mode: any attempt
157 to use these will cause a fatal error. Applications that *really* want
158 to use them can use the private_* version instead.
161 *) Redirect cipher operations to FIPS module for FIPS builds.
164 *) Redirect digest operations to FIPS module for FIPS builds.
167 *) Update build system to add "fips" flag which will link in fipscanister.o
168 for static and shared library builds embedding a signature if needed.
171 *) Output TLS supported curves in preference order instead of numerical
172 order. This is currently hardcoded for the highest order curves first.
173 This should be configurable so applications can judge speed vs strength.
176 *) Add TLS v1.2 server support for client authentication.
179 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
183 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
184 FIPS modules versions.
187 *) Add TLS v1.2 client side support for client authentication. Keep cache
188 of handshake records longer as we don't know the hash algorithm to use
189 until after the certificate request message is received.
192 *) Initial TLS v1.2 client support. Add a default signature algorithms
193 extension including all the algorithms we support. Parse new signature
194 format in client key exchange. Relax some ECC signing restrictions for
195 TLS v1.2 as indicated in RFC5246.
198 *) Add server support for TLS v1.2 signature algorithms extension. Switch
199 to new signature format when needed using client digest preference.
200 All server ciphersuites should now work correctly in TLS v1.2. No client
201 support yet and no support for client certificates.
204 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
205 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
206 ciphersuites. At present only RSA key exchange ciphersuites work with
207 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
208 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
209 and version checking.
212 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
213 with this defined it will not be affected by any changes to ssl internal
214 structures. Add several utility functions to allow openssl application
215 to work with OPENSSL_NO_SSL_INTERN defined.
219 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
221 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
224 *) Permit abbreviated handshakes when renegotiating using the function
225 SSL_renegotiate_abbreviated().
226 [Robin Seggelmann <seggelmann@fh-muenster.de>]
228 *) Add call to ENGINE_register_all_complete() to
229 ENGINE_load_builtin_engines(), so some implementations get used
230 automatically instead of needing explicit application support.
233 *) Add support for TLS key exporter as described in RFC5705.
234 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
236 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
237 a few changes are required:
239 Add SSL_OP_NO_TLSv1_1 flag.
241 Update version checking logic to handle version 1.1.
242 Add explicit IV handling (ported from DTLS code).
243 Add command line options to s_client/s_server.
246 Changes between 1.0.0e and 1.0.0f [xx XXX xxxx]
248 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
249 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
250 threads won't reuse the same blinding coefficients.
252 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
253 lock to call BN_BLINDING_invert_ex, and avoids one use of
254 BN_BLINDING_update for each BN_BLINDING structure (previously,
255 the last update always remained unused).
256 [Emilia Käsper (Google)]
258 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
259 [Bob Buckholz (Google)]
261 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
263 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
264 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
265 [Kaspar Brand <ossl@velox.ch>]
267 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
268 for multi-threaded use of ECDH. (CVE-2011-3210)
269 [Adam Langley (Google)]
271 *) Fix x509_name_ex_d2i memory leak on bad inputs.
274 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
275 signature public key algorithm by using OID xref utilities instead.
276 Before this you could only use some ECC ciphersuites with SHA1 only.
279 *) Add protection against ECDSA timing attacks as mentioned in the paper
280 by Billy Bob Brumley and Nicola Tuveri, see:
282 http://eprint.iacr.org/2011/232.pdf
284 [Billy Bob Brumley and Nicola Tuveri]
286 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
288 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
289 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
291 *) Fix bug in string printing code: if *any* escaping is enabled we must
292 escape the escape character (backslash) or the resulting string is
296 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
298 *) Disable code workaround for ancient and obsolete Netscape browsers
299 and servers: an attacker can use it in a ciphersuite downgrade attack.
300 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
303 *) Fixed J-PAKE implementation error, originally discovered by
304 Sebastien Martini, further info and confirmation from Stefan
305 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
308 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
310 *) Fix extension code to avoid race conditions which can result in a buffer
311 overrun vulnerability: resumed sessions must not be modified as they can
312 be shared by multiple threads. CVE-2010-3864
315 *) Fix WIN32 build system to correctly link an ENGINE directory into
319 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
321 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
323 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
325 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
327 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
328 context. The operation can be customised via the ctrl mechanism in
329 case ENGINEs want to include additional functionality.
332 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
335 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
336 output hashes compatible with older versions of OpenSSL.
337 [Willy Weisz <weisz@vcpc.univie.ac.at>]
339 *) Fix compression algorithm handling: if resuming a session use the
340 compression algorithm of the resumed session instead of determining
341 it from client hello again. Don't allow server to change algorithm.
344 *) Add load_crls() function to apps tidying load_certs() too. Add option
345 to verify utility to allow additional CRLs to be included.
348 *) Update OCSP request code to permit adding custom headers to the request:
349 some responders need this.
352 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
354 [Julia Lawall <julia@diku.dk>]
356 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
357 needlessly dereferenced structures, used obsolete functions and
358 didn't handle all updated verify codes correctly.
361 *) Disable MD2 in the default configuration.
364 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
365 indicate the initial BIO being pushed or popped. This makes it possible
366 to determine whether the BIO is the one explicitly called or as a result
367 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
368 it handles reference counts correctly and doesn't zero out the I/O bio
369 when it is not being explicitly popped. WARNING: applications which
370 included workarounds for the old buggy behaviour will need to be modified
371 or they could free up already freed BIOs.
374 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
375 renaming to all platforms (within the 0.9.8 branch, this was
376 done conditionally on Netware platforms to avoid a name clash).
377 [Guenter <lists@gknw.net>]
379 *) Add ECDHE and PSK support to DTLS.
380 [Michael Tuexen <tuexen@fh-muenster.de>]
382 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
386 *) Add "missing" function EVP_MD_flags() (without this the only way to
387 retrieve a digest flags is by accessing the structure directly. Update
388 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
389 or cipher is registered as in the "from" argument. Print out all
390 registered digests in the dgst usage message instead of manually
391 attempting to work them out.
394 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
395 this allows the use of compression and extensions. Change default cipher
396 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
397 by default unless an application cipher string requests it.
400 *) Alter match criteria in PKCS12_parse(). It used to try to use local
401 key ids to find matching certificates and keys but some PKCS#12 files
402 don't follow the (somewhat unwritten) rules and this strategy fails.
403 Now just gather all certificates together and the first private key
404 then look for the first certificate that matches the key.
407 *) Support use of registered digest and cipher names for dgst and cipher
408 commands instead of having to add each one as a special case. So now
415 openssl dgst -sha256 foo
417 and this works for ENGINE based algorithms too.
421 *) Update Gost ENGINE to support parameter files.
422 [Victor B. Wagner <vitus@cryptocom.ru>]
424 *) Support GeneralizedTime in ca utility.
425 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
427 *) Enhance the hash format used for certificate directory links. The new
428 form uses the canonical encoding (meaning equivalent names will work
429 even if they aren't identical) and uses SHA1 instead of MD5. This form
430 is incompatible with the older format and as a result c_rehash should
431 be used to rebuild symbolic links.
434 *) Make PKCS#8 the default write format for private keys, replacing the
435 traditional format. This form is standardised, more secure and doesn't
436 include an implicit MD5 dependency.
439 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
440 committed to OpenSSL should pass this lot as a minimum.
443 *) Add session ticket override functionality for use by EAP-FAST.
444 [Jouni Malinen <j@w1.fi>]
446 *) Modify HMAC functions to return a value. Since these can be implemented
447 in an ENGINE errors can occur.
450 *) Type-checked OBJ_bsearch_ex.
453 *) Type-checked OBJ_bsearch. Also some constification necessitated
454 by type-checking. Still to come: TXT_DB, bsearch(?),
455 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
459 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
460 seconds to a tm structure directly, instead of going through OS
461 specific date routines. This avoids any issues with OS routines such
462 as the year 2038 bug. New *_adj() functions for ASN1 time structures
463 and X509_time_adj_ex() to cover the extended range. The existing
464 X509_time_adj() is still usable and will no longer have any date issues.
467 *) Delta CRL support. New use deltas option which will attempt to locate
468 and search any appropriate delta CRLs available.
470 This work was sponsored by Google.
473 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
474 code and add additional score elements. Validate alternate CRL paths
475 as part of the CRL checking and indicate a new error "CRL path validation
476 error" in this case. Applications wanting additional details can use
477 the verify callback and check the new "parent" field. If this is not
478 NULL CRL path validation is taking place. Existing applications wont
479 see this because it requires extended CRL support which is off by
482 This work was sponsored by Google.
485 *) Support for freshest CRL extension.
487 This work was sponsored by Google.
490 *) Initial indirect CRL support. Currently only supported in the CRLs
491 passed directly and not via lookup. Process certificate issuer
492 CRL entry extension and lookup CRL entries by bother issuer name
493 and serial number. Check and process CRL issuer entry in IDP extension.
495 This work was sponsored by Google.
498 *) Add support for distinct certificate and CRL paths. The CRL issuer
499 certificate is validated separately in this case. Only enabled if
500 an extended CRL support flag is set: this flag will enable additional
501 CRL functionality in future.
503 This work was sponsored by Google.
506 *) Add support for policy mappings extension.
508 This work was sponsored by Google.
511 *) Fixes to pathlength constraint, self issued certificate handling,
512 policy processing to align with RFC3280 and PKITS tests.
514 This work was sponsored by Google.
517 *) Support for name constraints certificate extension. DN, email, DNS
518 and URI types are currently supported.
520 This work was sponsored by Google.
523 *) To cater for systems that provide a pointer-based thread ID rather
524 than numeric, deprecate the current numeric thread ID mechanism and
525 replace it with a structure and associated callback type. This
526 mechanism allows a numeric "hash" to be extracted from a thread ID in
527 either case, and on platforms where pointers are larger than 'long',
528 mixing is done to help ensure the numeric 'hash' is usable even if it
529 can't be guaranteed unique. The default mechanism is to use "&errno"
530 as a pointer-based thread ID to distinguish between threads.
532 Applications that want to provide their own thread IDs should now use
533 CRYPTO_THREADID_set_callback() to register a callback that will call
534 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
536 Note that ERR_remove_state() is now deprecated, because it is tied
537 to the assumption that thread IDs are numeric. ERR_remove_state(0)
538 to free the current thread's error state should be replaced by
539 ERR_remove_thread_state(NULL).
541 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
542 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
543 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
544 application was previously providing a numeric thread callback that
545 was inappropriate for distinguishing threads, then uniqueness might
546 have been obtained with &errno that happened immediately in the
547 intermediate development versions of OpenSSL; this is no longer the
548 case, the numeric thread callback will now override the automatic use
550 [Geoff Thorpe, with help from Bodo Moeller]
552 *) Initial support for different CRL issuing certificates. This covers a
553 simple case where the self issued certificates in the chain exist and
554 the real CRL issuer is higher in the existing chain.
556 This work was sponsored by Google.
559 *) Removed effectively defunct crypto/store from the build.
562 *) Revamp of STACK to provide stronger type-checking. Still to come:
563 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
564 ASN1_STRING, CONF_VALUE.
567 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
568 RAM on SSL connections. This option can save about 34k per idle SSL.
571 *) Revamp of LHASH to provide stronger type-checking. Still to come:
572 STACK, TXT_DB, bsearch, qsort.
575 *) Initial support for Cryptographic Message Syntax (aka CMS) based
576 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
577 support for data, signedData, compressedData, digestedData and
578 encryptedData, envelopedData types included. Scripts to check against
579 RFC4134 examples draft and interop and consistency checks of many
580 content types and variants.
583 *) Add options to enc utility to support use of zlib compression BIO.
586 *) Extend mk1mf to support importing of options and assembly language
587 files from Configure script, currently only included in VC-WIN32.
588 The assembly language rules can now optionally generate the source
589 files from the associated perl scripts.
592 *) Implement remaining functionality needed to support GOST ciphersuites.
593 Interop testing has been performed using CryptoPro implementations.
594 [Victor B. Wagner <vitus@cryptocom.ru>]
596 *) s390x assembler pack.
599 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
603 *) Implement Opaque PRF Input TLS extension as specified in
604 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
605 official specification yet and no extension type assignment by
606 IANA exists, this extension (for now) will have to be explicitly
607 enabled when building OpenSSL by providing the extension number
608 to use. For example, specify an option
610 -DTLSEXT_TYPE_opaque_prf_input=0x9527
612 to the "config" or "Configure" script to enable the extension,
613 assuming extension number 0x9527 (which is a completely arbitrary
614 and unofficial assignment based on the MD5 hash of the Internet
615 Draft). Note that by doing so, you potentially lose
616 interoperability with other TLS implementations since these might
617 be using the same extension number for other purposes.
619 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
620 opaque PRF input value to use in the handshake. This will create
621 an interal copy of the length-'len' string at 'src', and will
622 return non-zero for success.
624 To get more control and flexibility, provide a callback function
627 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
628 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
632 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
635 Callback function 'cb' will be called in handshakes, and is
636 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
637 Argument 'arg' is for application purposes (the value as given to
638 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
639 be provided to the callback function). The callback function
640 has to return non-zero to report success: usually 1 to use opaque
641 PRF input just if possible, or 2 to enforce use of the opaque PRF
642 input. In the latter case, the library will abort the handshake
643 if opaque PRF input is not successfully negotiated.
645 Arguments 'peerinput' and 'len' given to the callback function
646 will always be NULL and 0 in the case of a client. A server will
647 see the client's opaque PRF input through these variables if
648 available (NULL and 0 otherwise). Note that if the server
649 provides an opaque PRF input, the length must be the same as the
650 length of the client's opaque PRF input.
652 Note that the callback function will only be called when creating
653 a new session (session resumption can resume whatever was
654 previously negotiated), and will not be called in SSL 2.0
655 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
656 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
657 for applications that need to enforce opaque PRF input.
661 *) Update ssl code to support digests other than SHA1+MD5 for handshake
664 [Victor B. Wagner <vitus@cryptocom.ru>]
666 *) Add RFC4507 support to OpenSSL. This includes the corrections in
667 RFC4507bis. The encrypted ticket format is an encrypted encoded
668 SSL_SESSION structure, that way new session features are automatically
671 If a client application caches session in an SSL_SESSION structure
672 support is transparent because tickets are now stored in the encoded
675 The SSL_CTX structure automatically generates keys for ticket
676 protection in servers so again support should be possible
677 with no application modification.
679 If a client or server wishes to disable RFC4507 support then the option
680 SSL_OP_NO_TICKET can be set.
682 Add a TLS extension debugging callback to allow the contents of any client
683 or server extensions to be examined.
685 This work was sponsored by Google.
688 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
689 OpenSSL should now compile cleanly on gcc 4.2
690 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
692 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
693 support including streaming MAC support: this is required for GOST
695 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
697 *) Add option -stream to use PKCS#7 streaming in smime utility. New
698 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
699 to output in BER and PEM format.
702 *) Experimental support for use of HMAC via EVP_PKEY interface. This
703 allows HMAC to be handled via the EVP_DigestSign*() interface. The
704 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
705 ENGINE support for HMAC keys which are unextractable. New -mac and
706 -macopt options to dgst utility.
709 *) New option -sigopt to dgst utility. Update dgst to use
710 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
711 alternative signing paramaters such as X9.31 or PSS in the dgst
715 *) Change ssl_cipher_apply_rule(), the internal function that does
716 the work each time a ciphersuite string requests enabling
717 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
718 removing ("!foo+bar") a class of ciphersuites: Now it maintains
719 the order of disabled ciphersuites such that those ciphersuites
720 that most recently went from enabled to disabled not only stay
721 in order with respect to each other, but also have higher priority
722 than other disabled ciphersuites the next time ciphersuites are
725 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
726 the same ciphersuites as with "HIGH" alone, but in a specific
727 order where the PSK ciphersuites come first (since they are the
728 most recently disabled ciphersuites when "HIGH" is parsed).
730 Also, change ssl_create_cipher_list() (using this new
731 funcionality) such that between otherwise identical
732 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
736 *) Change ssl_create_cipher_list() so that it automatically
737 arranges the ciphersuites in reasonable order before starting
738 to process the rule string. Thus, the definition for "DEFAULT"
739 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
740 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
741 This makes it much easier to arrive at a reasonable default order
742 in applications for which anonymous ciphers are OK (meaning
743 that you can't actually use DEFAULT).
744 [Bodo Moeller; suggested by Victor Duchovni]
746 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
747 processing) into multiple integers instead of setting
748 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
749 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
750 (These masks as well as the individual bit definitions are hidden
751 away into the non-exported interface ssl/ssl_locl.h, so this
752 change to the definition of the SSL_CIPHER structure shouldn't
753 affect applications.) This give us more bits for each of these
754 categories, so there is no longer a need to coagulate AES128 and
755 AES256 into a single algorithm bit, and to coagulate Camellia128
756 and Camellia256 into a single algorithm bit, which has led to all
759 Thus, among other things, the kludge introduced in 0.9.7m and
760 0.9.8e for masking out AES256 independently of AES128 or masking
761 out Camellia256 independently of AES256 is not needed here in 0.9.9.
763 With the change, we also introduce new ciphersuite aliases that
764 so far were missing: "AES128", "AES256", "CAMELLIA128", and
768 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
769 Use the leftmost N bytes of the signature input if the input is
770 larger than the prime q (with N being the size in bytes of q).
773 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
774 it yet and it is largely untested.
777 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
780 *) Initial incomplete changes to avoid need for function casts in OpenSSL
781 some compilers (gcc 4.2 and later) reject their use. Safestack is
782 reimplemented. Update ASN1 to avoid use of legacy functions.
785 *) Win32/64 targets are linked with Winsock2.
788 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
789 to external functions. This can be used to increase CRL handling
790 efficiency especially when CRLs are very large by (for example) storing
791 the CRL revoked certificates in a database.
794 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
795 new CRLs added to a directory can be used. New command line option
796 -verify_return_error to s_client and s_server. This causes real errors
797 to be returned by the verify callback instead of carrying on no matter
798 what. This reflects the way a "real world" verify callback would behave.
801 *) GOST engine, supporting several GOST algorithms and public key formats.
802 Kindly donated by Cryptocom.
805 *) Partial support for Issuing Distribution Point CRL extension. CRLs
806 partitioned by DP are handled but no indirect CRL or reason partitioning
807 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
808 selected via a scoring technique which handles IDP and AKID in CRLs.
811 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
812 will ultimately be used for all verify operations: this will remove the
813 X509_STORE dependency on certificate verification and allow alternative
814 lookup methods. X509_STORE based implementations of these two callbacks.
817 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
818 Modify get_crl() to find a valid (unexpired) CRL if possible.
821 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
822 this would be called X509_CRL_cmp() but that name is already used by
823 a function that just compares CRL issuer names. Cache several CRL
824 extensions in X509_CRL structure and cache CRLDP in X509.
827 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
828 this maps equivalent X509_NAME structures into a consistent structure.
829 Name comparison can then be performed rapidly using memcmp().
832 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
836 *) Allow digests to supply their own micalg string for S/MIME type using
837 the ctrl EVP_MD_CTRL_MICALG.
840 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
841 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
842 ctrl. It can then customise the structure before and/or after signing
846 *) New function OBJ_add_sigid() to allow application defined signature OIDs
847 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
848 to free up any added signature OIDs.
851 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
852 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
853 digest and cipher tables. New options added to openssl utility:
854 list-message-digest-algorithms and list-cipher-algorithms.
857 *) Change the array representation of binary polynomials: the list
858 of degrees of non-zero coefficients is now terminated with -1.
859 Previously it was terminated with 0, which was also part of the
860 value; thus, the array representation was not applicable to
861 polynomials where t^0 has coefficient zero. This change makes
862 the array representation useful in a more general context.
865 *) Various modifications and fixes to SSL/TLS cipher string
866 handling. For ECC, the code now distinguishes between fixed ECDH
867 with RSA certificates on the one hand and with ECDSA certificates
868 on the other hand, since these are separate ciphersuites. The
869 unused code for Fortezza ciphersuites has been removed.
871 For consistency with EDH, ephemeral ECDH is now called "EECDH"
872 (not "ECDHE"). For consistency with the code for DH
873 certificates, use of ECDH certificates is now considered ECDH
874 authentication, not RSA or ECDSA authentication (the latter is
875 merely the CA's signing algorithm and not actively used in the
878 The temporary ciphersuite alias "ECCdraft" is no longer
879 available, and ECC ciphersuites are no longer excluded from "ALL"
880 and "DEFAULT". The following aliases now exist for RFC 4492
881 ciphersuites, most of these by analogy with the DH case:
883 kECDHr - ECDH cert, signed with RSA
884 kECDHe - ECDH cert, signed with ECDSA
885 kECDH - ECDH cert (signed with either RSA or ECDSA)
886 kEECDH - ephemeral ECDH
887 ECDH - ECDH cert or ephemeral ECDH
893 AECDH - anonymous ECDH
894 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
898 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
899 Use correct micalg parameters depending on digest(s) in signed message.
902 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
903 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
906 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
907 an engine to register a method. Add ENGINE lookups for methods and
908 functional reference processing.
911 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
912 EVP_{Sign,Verify}* which allow an application to customise the signature
916 *) New -resign option to smime utility. This adds one or more signers
917 to an existing PKCS#7 signedData structure. Also -md option to use an
918 alternative message digest algorithm for signing.
921 *) Tidy up PKCS#7 routines and add new functions to make it easier to
922 create PKCS7 structures containing multiple signers. Update smime
923 application to support multiple signers.
926 *) New -macalg option to pkcs12 utility to allow setting of an alternative
930 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
931 Reorganize PBE internals to lookup from a static table using NIDs,
932 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
933 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
934 PRF which will be automatically used with PBES2.
937 *) Replace the algorithm specific calls to generate keys in "req" with the
941 *) Update PKCS#7 enveloped data routines to use new API. This is now
942 supported by any public key method supporting the encrypt operation. A
943 ctrl is added to allow the public key algorithm to examine or modify
944 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
948 *) Add a ctrl to asn1 method to allow a public key algorithm to express
949 a default digest type to use. In most cases this will be SHA1 but some
950 algorithms (such as GOST) need to specify an alternative digest. The
951 return value indicates how strong the prefernce is 1 means optional and
952 2 is mandatory (that is it is the only supported type). Modify
953 ASN1_item_sign() to accept a NULL digest argument to indicate it should
954 use the default md. Update openssl utilities to use the default digest
955 type for signing if it is not explicitly indicated.
958 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
959 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
960 signing method from the key type. This effectively removes the link
961 between digests and public key types.
964 *) Add an OID cross reference table and utility functions. Its purpose is to
965 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
966 rsaEncryption. This will allow some of the algorithm specific hackery
967 needed to use the correct OID to be removed.
970 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
971 structures for PKCS7_sign(). They are now set up by the relevant public
975 *) Add provisional EC pkey method with support for ECDSA and ECDH.
978 *) Add support for key derivation (agreement) in the API, DH method and
982 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
983 public and private key formats. As a side effect these add additional
984 command line functionality not previously available: DSA signatures can be
985 generated and verified using pkeyutl and DH key support and generation in
990 [Oliver Tappe <zooey@hirschkaefer.de>]
992 *) New make target "install_html_docs" installs HTML renditions of the
994 [Oliver Tappe <zooey@hirschkaefer.de>]
996 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
997 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
998 support key and parameter generation and add initial key generation
999 functionality for RSA.
1002 *) Add functions for main EVP_PKEY_method operations. The undocumented
1003 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1004 EVP_PKEY_{encrypt,decrypt}_old.
1007 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1008 key API, doesn't do much yet.
1011 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1012 public key algorithms. New option to openssl utility:
1013 "list-public-key-algorithms" to print out info.
1016 *) Implement the Supported Elliptic Curves Extension for
1017 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1020 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1021 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1024 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1025 utilities such as rsa, dsa, dsaparam etc except they process any key
1029 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1030 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1031 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1035 *) Initial support for pluggable public key ASN1.
1036 De-spaghettify the public key ASN1 handling. Move public and private
1037 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1038 algorithm specific handling to a single module within the relevant
1039 algorithm directory. Add functions to allow (near) opaque processing
1040 of public and private key structures.
1043 *) Implement the Supported Point Formats Extension for
1044 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1047 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1048 for the psk identity [hint] and the psk callback functions to the
1049 SSL_SESSION, SSL and SSL_CTX structure.
1052 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1056 SSL_CTX_use_psk_identity_hint
1057 SSL_get_psk_identity_hint
1058 SSL_get_psk_identity
1059 SSL_use_psk_identity_hint
1061 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1063 *) Add RFC 3161 compliant time stamp request creation, response generation
1064 and response verification functionality.
1065 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1067 *) Add initial support for TLS extensions, specifically for the server_name
1068 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1069 have new members for a host name. The SSL data structure has an
1070 additional member SSL_CTX *initial_ctx so that new sessions can be
1071 stored in that context to allow for session resumption, even after the
1072 SSL has been switched to a new SSL_CTX in reaction to a client's
1073 server_name extension.
1075 New functions (subject to change):
1077 SSL_get_servername()
1078 SSL_get_servername_type()
1081 New CTRL codes and macros (subject to change):
1083 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1084 - SSL_CTX_set_tlsext_servername_callback()
1085 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1086 - SSL_CTX_set_tlsext_servername_arg()
1087 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1089 openssl s_client has a new '-servername ...' option.
1091 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1092 '-key2 ...', '-servername_fatal' (subject to change). This allows
1093 testing the HostName extension for a specific single host name ('-cert'
1094 and '-key' remain fallbacks for handshakes without HostName
1095 negotiation). If the unrecogninzed_name alert has to be sent, this by
1096 default is a warning; it becomes fatal with the '-servername_fatal'
1099 [Peter Sylvester, Remy Allais, Christophe Renou]
1101 *) Whirlpool hash implementation is added.
1104 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1105 bn(64,32). Because of instruction set limitations it doesn't have
1106 any negative impact on performance. This was done mostly in order
1107 to make it possible to share assembler modules, such as bn_mul_mont
1108 implementations, between 32- and 64-bit builds without hassle.
1111 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1112 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1116 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1117 dedicated Montgomery multiplication procedure, is introduced.
1118 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1119 "64-bit" performance on certain 32-bit targets.
1122 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1123 in SSL structures. New SSL ctrl to set maximum send fragment size.
1124 Save memory by seeting the I/O buffer sizes dynamically instead of
1125 using the maximum available value.
1128 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1129 in addition to the text details.
1132 *) Very, very preliminary EXPERIMENTAL support for printing of general
1133 ASN1 structures. This currently produces rather ugly output and doesn't
1134 handle several customised structures at all.
1137 *) Integrated support for PVK file format and some related formats such
1138 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1139 these in the 'rsa' and 'dsa' utilities.
1142 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1145 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1146 place for the (very old) "NETSCAPE" format certificates which are now
1147 handled using new ASN1 code equivalents.
1150 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1151 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1152 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1155 *) Modify CRL distribution points extension code to print out previously
1156 unsupported fields. Enhance extension setting code to allow setting of
1160 *) Add print and set support for Issuing Distribution Point CRL extension.
1163 *) Change 'Configure' script to enable Camellia by default.
1166 Changes between 0.9.8r and 0.9.8s [xx XXX xxxx]
1168 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1169 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1170 threads won't reuse the same blinding coefficients.
1172 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1173 lock to call BN_BLINDING_invert_ex, and avoids one use of
1174 BN_BLINDING_update for each BN_BLINDING structure (previously,
1175 the last update always remained unused).
1176 [Emilia Käsper (Google)]
1178 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1179 for multi-threaded use of ECDH.
1180 [Adam Langley (Google)]
1182 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1185 *) Add protection against ECDSA timing attacks as mentioned in the paper
1186 by Billy Bob Brumley and Nicola Tuveri, see:
1188 http://eprint.iacr.org/2011/232.pdf
1190 [Billy Bob Brumley and Nicola Tuveri]
1192 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1194 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1195 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1197 *) Fix bug in string printing code: if *any* escaping is enabled we must
1198 escape the escape character (backslash) or the resulting string is
1202 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1204 *) Disable code workaround for ancient and obsolete Netscape browsers
1205 and servers: an attacker can use it in a ciphersuite downgrade attack.
1206 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1209 *) Fixed J-PAKE implementation error, originally discovered by
1210 Sebastien Martini, further info and confirmation from Stefan
1211 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1214 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1216 *) Fix extension code to avoid race conditions which can result in a buffer
1217 overrun vulnerability: resumed sessions must not be modified as they can
1218 be shared by multiple threads. CVE-2010-3864
1221 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1224 *) Don't reencode certificate when calculating signature: cache and use
1225 the original encoding instead. This makes signature verification of
1226 some broken encodings work correctly.
1229 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1230 is also one of the inputs.
1231 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1233 *) Don't repeatedly append PBE algorithms to table if they already exist.
1234 Sort table on each new add. This effectively makes the table read only
1235 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1239 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1241 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1244 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1245 access or freeing data twice (CVE-2010-0742)
1246 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1248 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1249 common in certificates and some applications which only call
1250 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1254 Reduce copying into .apps and .test in makevms.com
1255 Don't try to use blank CA certificate in CA.com
1256 Allow use of C files from original directories in maketests.com
1257 [Steven M. Schweda" <sms@antinode.info>]
1259 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1261 *) When rejecting SSL/TLS records due to an incorrect version number, never
1262 update s->server with a new major version number. As of
1263 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1264 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1265 the previous behavior could result in a read attempt at NULL when
1266 receiving specific incorrect SSL/TLS records once record payload
1267 protection is active. (CVE-2010-0740)
1268 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1270 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1271 could be crashed if the relevant tables were not present (e.g. chrooted).
1272 [Tomas Hoger <thoger@redhat.com>]
1274 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1276 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1277 [Martin Olsson, Neel Mehta]
1279 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1280 accommodate for stack sorting, always a write lock!).
1283 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1284 excessive delays in the RAND_poll(): over a minute. As a workaround
1285 include a time check in the inner Heap32Next loop too.
1288 *) The code that handled flushing of data in SSL/TLS originally used the
1289 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1290 the problem outlined in PR#1949. The fix suggested there however can
1291 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1292 of Apache). So instead simplify the code to flush unconditionally.
1293 This should be fine since flushing with no data to flush is a no op.
1296 *) Handle TLS versions 2.0 and later properly and correctly use the
1297 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1298 off ancient servers have a habit of sticking around for a while...
1301 *) Modify compression code so it frees up structures without using the
1302 ex_data callbacks. This works around a problem where some applications
1303 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1304 restarting) then use compression (e.g. SSL with compression) later.
1305 This results in significant per-connection memory leaks and
1306 has caused some security issues including CVE-2008-1678 and
1310 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1311 change when encrypting or decrypting.
1314 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1315 connect and renegotiate with servers which do not support RI.
1316 Until RI is more widely deployed this option is enabled by default.
1319 *) Add "missing" ssl ctrls to clear options and mode.
1322 *) If client attempts to renegotiate and doesn't support RI respond with
1323 a no_renegotiation alert as required by RFC5746. Some renegotiating
1324 TLS clients will continue a connection gracefully when they receive
1325 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1326 waiting for a server hello which it will never receive. Now we treat a
1327 received no_renegotiation alert as a fatal error. This is because
1328 applications requesting a renegotiation might well expect it to succeed
1329 and would have no code in place to handle the server denying it so the
1330 only safe thing to do is to terminate the connection.
1333 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1334 peer supports secure renegotiation and 0 otherwise. Print out peer
1335 renegotiation support in s_client/s_server.
1338 *) Replace the highly broken and deprecated SPKAC certification method with
1339 the updated NID creation version. This should correctly handle UTF8.
1342 *) Implement RFC5746. Re-enable renegotiation but require the extension
1343 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1344 turns out to be a bad idea. It has been replaced by
1345 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1346 SSL_CTX_set_options(). This is really not recommended unless you
1347 know what you are doing.
1348 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1350 *) Fixes to stateless session resumption handling. Use initial_ctx when
1351 issuing and attempting to decrypt tickets in case it has changed during
1352 servername handling. Use a non-zero length session ID when attempting
1353 stateless session resumption: this makes it possible to determine if
1354 a resumption has occurred immediately after receiving server hello
1355 (several places in OpenSSL subtly assume this) instead of later in
1359 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1360 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1361 fixes for a few places where the return code is not checked
1363 [Julia Lawall <julia@diku.dk>]
1365 *) Add --strict-warnings option to Configure script to include devteam
1366 warnings in other configurations.
1369 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1370 makes it possible to install openssl libraries in locations which
1371 have names other than "lib", for example "/usr/lib64" which some
1373 [Steve Henson, based on patch from Jeremy Utley]
1375 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1376 X690 8.9.12 and can produce some misleading textual output of OIDs.
1377 [Steve Henson, reported by Dan Kaminsky]
1379 *) Delete MD2 from algorithm tables. This follows the recommendation in
1380 several standards that it is not used in new applications due to
1381 several cryptographic weaknesses. For binary compatibility reasons
1382 the MD2 API is still compiled in by default.
1385 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1389 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1390 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1392 [Guenter <lists@gknw.net>]
1394 *) Fix the server certificate chain building code to use X509_verify_cert(),
1395 it used to have an ad-hoc builder which was unable to cope with anything
1396 other than a simple chain.
1397 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1399 *) Don't check self signed certificate signatures in X509_verify_cert()
1400 by default (a flag can override this): it just wastes time without
1401 adding any security. As a useful side effect self signed root CAs
1402 with non-FIPS digests are now usable in FIPS mode.
1405 *) In dtls1_process_out_of_seq_message() the check if the current message
1406 is already buffered was missing. For every new message was memory
1407 allocated, allowing an attacker to perform an denial of service attack
1408 with sending out of seq handshake messages until there is no memory
1409 left. Additionally every future messege was buffered, even if the
1410 sequence number made no sense and would be part of another handshake.
1411 So only messages with sequence numbers less than 10 in advance will be
1412 buffered. (CVE-2009-1378)
1413 [Robin Seggelmann, discovered by Daniel Mentz]
1415 *) Records are buffered if they arrive with a future epoch to be
1416 processed after finishing the corresponding handshake. There is
1417 currently no limitation to this buffer allowing an attacker to perform
1418 a DOS attack with sending records with future epochs until there is no
1419 memory left. This patch adds the pqueue_size() function to detemine
1420 the size of a buffer and limits the record buffer to 100 entries.
1422 [Robin Seggelmann, discovered by Daniel Mentz]
1424 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1425 parent structure is freed. (CVE-2009-1379)
1428 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1429 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1432 [Ilya O. <vrghost@gmail.com>]
1434 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1436 *) Disable renegotiation completely - this fixes a severe security
1437 problem (CVE-2009-3555) at the cost of breaking all
1438 renegotiation. Renegotiation can be re-enabled by setting
1439 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1440 run-time. This is really not recommended unless you know what
1444 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1446 *) Don't set val to NULL when freeing up structures, it is freed up by
1447 underlying code. If sizeof(void *) > sizeof(long) this can result in
1448 zeroing past the valid field. (CVE-2009-0789)
1449 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1451 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1452 checked correctly. This would allow some invalid signed attributes to
1453 appear to verify correctly. (CVE-2009-0591)
1454 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1456 *) Reject UniversalString and BMPString types with invalid lengths. This
1457 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1458 a legal length. (CVE-2009-0590)
1461 *) Set S/MIME signing as the default purpose rather than setting it
1462 unconditionally. This allows applications to override it at the store
1466 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1467 to handle some structures.
1470 *) Improve efficiency of mem_gets: don't search whole buffer each time
1472 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1474 *) New -hex option for openssl rand.
1477 *) Print out UTF8String and NumericString when parsing ASN1.
1480 *) Support NumericString type for name components.
1483 *) Allow CC in the environment to override the automatically chosen
1484 compiler. Note that nothing is done to ensure flags work with the
1488 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1490 *) Properly check EVP_VerifyFinal() and similar return values
1492 [Ben Laurie, Bodo Moeller, Google Security Team]
1494 *) Enable TLS extensions by default.
1497 *) Allow the CHIL engine to be loaded, whether the application is
1498 multithreaded or not. (This does not release the developer from the
1499 obligation to set up the dynamic locking callbacks.)
1500 [Sander Temme <sander@temme.net>]
1502 *) Use correct exit code if there is an error in dgst command.
1503 [Steve Henson; problem pointed out by Roland Dirlewanger]
1505 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1506 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1509 *) Add experimental JPAKE support, including demo authentication in
1510 s_client and s_server.
1513 *) Set the comparison function in v3_addr_canonize().
1514 [Rob Austein <sra@hactrn.net>]
1516 *) Add support for XMPP STARTTLS in s_client.
1517 [Philip Paeps <philip@freebsd.org>]
1519 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1520 to ensure that even with this option, only ciphersuites in the
1521 server's preference list will be accepted. (Note that the option
1522 applies only when resuming a session, so the earlier behavior was
1523 just about the algorithm choice for symmetric cryptography.)
1526 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1528 *) Fix NULL pointer dereference if a DTLS server received
1529 ChangeCipherSpec as first record (CVE-2009-1386).
1532 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1533 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1536 *) The fix in 0.9.8c that supposedly got rid of unsafe
1537 double-checked locking was incomplete for RSA blinding,
1538 addressing just one layer of what turns out to have been
1539 doubly unsafe triple-checked locking.
1541 So now fix this for real by retiring the MONT_HELPER macro
1542 in crypto/rsa/rsa_eay.c.
1544 [Bodo Moeller; problem pointed out by Marius Schilder]
1546 *) Various precautionary measures:
1548 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1550 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1551 (NB: This would require knowledge of the secret session ticket key
1552 to exploit, in which case you'd be SOL either way.)
1554 - Change bn_nist.c so that it will properly handle input BIGNUMs
1555 outside the expected range.
1557 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1560 [Neel Mehta, Bodo Moeller]
1562 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1563 the load fails. Useful for distros.
1564 [Ben Laurie and the FreeBSD team]
1566 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1569 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1572 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1574 This work was sponsored by Logica.
1577 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1578 keystores. Support for SSL/TLS client authentication too.
1579 Not compiled unless enable-capieng specified to Configure.
1581 This work was sponsored by Logica.
1584 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1585 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1586 attribute creation routines such as certifcate requests and PKCS#12
1590 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1592 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1593 handshake which could lead to a cilent crash as found using the
1594 Codenomicon TLS test suite (CVE-2008-1672)
1595 [Steve Henson, Mark Cox]
1597 *) Fix double free in TLS server name extensions which could lead to
1598 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1601 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1603 Clear the error queue to ensure that error entries left from
1604 older function calls do not interfere with the correct operation.
1605 [Lutz Jaenicke, Erik de Castro Lopo]
1607 *) Remove root CA certificates of commercial CAs:
1609 The OpenSSL project does not recommend any specific CA and does not
1610 have any policy with respect to including or excluding any CA.
1611 Therefore it does not make any sense to ship an arbitrary selection
1612 of root CA certificates with the OpenSSL software.
1615 *) RSA OAEP patches to fix two separate invalid memory reads.
1616 The first one involves inputs when 'lzero' is greater than
1617 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1618 before the beginning of from). The second one involves inputs where
1619 the 'db' section contains nothing but zeroes (there is a one-byte
1620 invalid read after the end of 'db').
1621 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1623 *) Partial backport from 0.9.9-dev:
1625 Introduce bn_mul_mont (dedicated Montgomery multiplication
1626 procedure) as a candidate for BIGNUM assembler implementation.
1627 While 0.9.9-dev uses assembler for various architectures, only
1628 x86_64 is available by default here in the 0.9.8 branch, and
1629 32-bit x86 is available through a compile-time setting.
1631 To try the 32-bit x86 assembler implementation, use Configure
1632 option "enable-montasm" (which exists only for this backport).
1634 As "enable-montasm" for 32-bit x86 disclaims code stability
1635 anyway, in this constellation we activate additional code
1636 backported from 0.9.9-dev for further performance improvements,
1637 namely BN_from_montgomery_word. (To enable this otherwise,
1638 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1640 [Andy Polyakov (backport partially by Bodo Moeller)]
1642 *) Add TLS session ticket callback. This allows an application to set
1643 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1644 values. This is useful for key rollover for example where several key
1645 sets may exist with different names.
1648 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1649 This was broken until now in 0.9.8 releases, such that the only way
1650 a registered ENGINE could be used (assuming it initialises
1651 successfully on the host) was to explicitly set it as the default
1652 for the relevant algorithms. This is in contradiction with 0.9.7
1653 behaviour and the documentation. With this fix, when an ENGINE is
1654 registered into a given algorithm's table of implementations, the
1655 'uptodate' flag is reset so that auto-discovery will be used next
1656 time a new context for that algorithm attempts to select an
1658 [Ian Lister (tweaked by Geoff Thorpe)]
1660 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1661 implemention in the following ways:
1663 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1666 Lack of BER streaming support means one pass streaming processing is
1667 only supported if data is detached: setting the streaming flag is
1668 ignored for embedded content.
1670 CMS support is disabled by default and must be explicitly enabled
1671 with the enable-cms configuration option.
1674 *) Update the GMP engine glue to do direct copies between BIGNUM and
1675 mpz_t when openssl and GMP use the same limb size. Otherwise the
1676 existing "conversion via a text string export" trick is still used.
1677 [Paul Sheer <paulsheer@gmail.com>]
1679 *) Zlib compression BIO. This is a filter BIO which compressed and
1680 uncompresses any data passed through it.
1683 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1684 RFC3394 compatible AES key wrapping.
1687 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1688 sets string data without copying. X509_ALGOR_set0() and
1689 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1690 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1691 from an X509_ATTRIBUTE structure optionally checking it occurs only
1692 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1696 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1697 to get the expected BN_FLG_CONSTTIME behavior.
1698 [Bodo Moeller (Google)]
1702 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1703 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1704 - added some more tests to do_tests.pl
1705 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1706 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1707 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1708 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1709 - various changes to netware.pl to enable gcc-cross builds on Win32
1711 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1712 - various changes to fix missing prototype warnings
1713 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1714 - added AES, WHIRLPOOL and CPUID assembler code to build files
1715 - added missing AES assembler make rules to mk1mf.pl
1716 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1717 [Guenter Knauf <eflash@gmx.net>]
1719 *) Implement certificate status request TLS extension defined in RFC3546.
1720 A client can set the appropriate parameters and receive the encoded
1721 OCSP response via a callback. A server can query the supplied parameters
1722 and set the encoded OCSP response in the callback. Add simplified examples
1723 to s_client and s_server.
1726 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1728 *) Fix various bugs:
1729 + Binary incompatibility of ssl_ctx_st structure
1730 + DTLS interoperation with non-compliant servers
1731 + Don't call get_session_cb() without proposed session
1732 + Fix ia64 assembler code
1733 [Andy Polyakov, Steve Henson]
1735 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1737 *) DTLS Handshake overhaul. There were longstanding issues with
1738 OpenSSL DTLS implementation, which were making it impossible for
1739 RFC 4347 compliant client to communicate with OpenSSL server.
1740 Unfortunately just fixing these incompatibilities would "cut off"
1741 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1742 server keeps tolerating non RFC compliant syntax. The opposite is
1743 not true, 0.9.8f client can not communicate with earlier server.
1744 This update even addresses CVE-2007-4995.
1747 *) Changes to avoid need for function casts in OpenSSL: some compilers
1748 (gcc 4.2 and later) reject their use.
1749 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1752 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1753 RFC4507bis. The encrypted ticket format is an encrypted encoded
1754 SSL_SESSION structure, that way new session features are automatically
1757 If a client application caches session in an SSL_SESSION structure
1758 support is transparent because tickets are now stored in the encoded
1761 The SSL_CTX structure automatically generates keys for ticket
1762 protection in servers so again support should be possible
1763 with no application modification.
1765 If a client or server wishes to disable RFC4507 support then the option
1766 SSL_OP_NO_TICKET can be set.
1768 Add a TLS extension debugging callback to allow the contents of any client
1769 or server extensions to be examined.
1771 This work was sponsored by Google.
1774 *) Add initial support for TLS extensions, specifically for the server_name
1775 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1776 have new members for a host name. The SSL data structure has an
1777 additional member SSL_CTX *initial_ctx so that new sessions can be
1778 stored in that context to allow for session resumption, even after the
1779 SSL has been switched to a new SSL_CTX in reaction to a client's
1780 server_name extension.
1782 New functions (subject to change):
1784 SSL_get_servername()
1785 SSL_get_servername_type()
1788 New CTRL codes and macros (subject to change):
1790 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1791 - SSL_CTX_set_tlsext_servername_callback()
1792 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1793 - SSL_CTX_set_tlsext_servername_arg()
1794 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1796 openssl s_client has a new '-servername ...' option.
1798 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1799 '-key2 ...', '-servername_fatal' (subject to change). This allows
1800 testing the HostName extension for a specific single host name ('-cert'
1801 and '-key' remain fallbacks for handshakes without HostName
1802 negotiation). If the unrecogninzed_name alert has to be sent, this by
1803 default is a warning; it becomes fatal with the '-servername_fatal'
1806 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1808 *) Add AES and SSE2 assembly language support to VC++ build.
1811 *) Mitigate attack on final subtraction in Montgomery reduction.
1814 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1815 (which previously caused an internal error).
1818 *) Squeeze another 10% out of IGE mode when in != out.
1821 *) AES IGE mode speedup.
1822 [Dean Gaudet (Google)]
1824 *) Add the Korean symmetric 128-bit cipher SEED (see
1825 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1826 add SEED ciphersuites from RFC 4162:
1828 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1829 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1830 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1831 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1833 To minimize changes between patchlevels in the OpenSSL 0.9.8
1834 series, SEED remains excluded from compilation unless OpenSSL
1835 is configured with 'enable-seed'.
1836 [KISA, Bodo Moeller]
1838 *) Mitigate branch prediction attacks, which can be practical if a
1839 single processor is shared, allowing a spy process to extract
1840 information. For detailed background information, see
1841 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1842 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1843 and Necessary Software Countermeasures"). The core of the change
1844 are new versions BN_div_no_branch() and
1845 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1846 respectively, which are slower, but avoid the security-relevant
1847 conditional branches. These are automatically called by BN_div()
1848 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1849 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1850 remove a conditional branch.
1852 BN_FLG_CONSTTIME is the new name for the previous
1853 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1854 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1855 in the exponent causes BN_mod_exp_mont() to use the alternative
1856 implementation in BN_mod_exp_mont_consttime().) The old name
1857 remains as a deprecated alias.
1859 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1860 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1861 constant-time implementations for more than just exponentiation.
1862 Here too the old name is kept as a deprecated alias.
1864 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1865 the BN_BLINDING structure gets an independent copy of the
1866 modulus. This means that the previous "BIGNUM *m" argument to
1867 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1868 essentially becomes "const BIGNUM *m", although we can't actually
1869 change this in the header file before 0.9.9. It allows
1870 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1871 enable BN_FLG_CONSTTIME.
1873 [Matthew D Wood (Intel Corp)]
1875 *) In the SSL/TLS server implementation, be strict about session ID
1876 context matching (which matters if an application uses a single
1877 external cache for different purposes). Previously,
1878 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1879 set. This did ensure strict client verification, but meant that,
1880 with applications using a single external cache for quite
1881 different requirements, clients could circumvent ciphersuite
1882 restrictions for a given session ID context by starting a session
1883 in a different context.
1886 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1887 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1888 authentication-only ciphersuites.
1891 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1892 not complete and could lead to a possible single byte overflow
1893 (CVE-2007-5135) [Ben Laurie]
1895 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1897 *) Since AES128 and AES256 (and similarly Camellia128 and
1898 Camellia256) share a single mask bit in the logic of
1899 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1900 kludge to work properly if AES128 is available and AES256 isn't
1901 (or if Camellia128 is available and Camellia256 isn't).
1904 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1905 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1906 When a point or a seed is encoded in a BIT STRING, we need to
1907 prevent the removal of trailing zero bits to get the proper DER
1908 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1909 of a NamedBitList, for which trailing 0 bits need to be removed.)
1912 *) Have SSL/TLS server implementation tolerate "mismatched" record
1913 protocol version while receiving ClientHello even if the
1914 ClientHello is fragmented. (The server can't insist on the
1915 particular protocol version it has chosen before the ServerHello
1916 message has informed the client about his choice.)
1919 *) Add RFC 3779 support.
1920 [Rob Austein for ARIN, Ben Laurie]
1922 *) Load error codes if they are not already present instead of using a
1923 static variable. This allows them to be cleanly unloaded and reloaded.
1924 Improve header file function name parsing.
1927 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1928 or CAPABILITY handshake as required by RFCs.
1931 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1933 *) Introduce limits to prevent malicious keys being able to
1934 cause a denial of service. (CVE-2006-2940)
1935 [Steve Henson, Bodo Moeller]
1937 *) Fix ASN.1 parsing of certain invalid structures that can result
1938 in a denial of service. (CVE-2006-2937) [Steve Henson]
1940 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1941 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1943 *) Fix SSL client code which could crash if connecting to a
1944 malicious SSLv2 server. (CVE-2006-4343)
1945 [Tavis Ormandy and Will Drewry, Google Security Team]
1947 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1948 match only those. Before that, "AES256-SHA" would be interpreted
1949 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1950 the same strength classification in 0.9.7h) as we currently only
1951 have a single AES bit in the ciphersuite description bitmap.
1952 That change, however, also applied to ciphersuite strings such as
1953 "RC4-MD5" that intentionally matched multiple ciphersuites --
1954 namely, SSL 2.0 ciphersuites in addition to the more common ones
1955 from SSL 3.0/TLS 1.0.
1957 So we change the selection algorithm again: Naming an explicit
1958 ciphersuite selects this one ciphersuite, and any other similar
1959 ciphersuite (same bitmap) from *other* protocol versions.
1960 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1961 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1963 Since SSL 2.0 does not have any ciphersuites for which the
1964 128/256 bit distinction would be relevant, this works for now.
1965 The proper fix will be to use different bits for AES128 and
1966 AES256, which would have avoided the problems from the beginning;
1967 however, bits are scarce, so we can only do this in a new release
1968 (not just a patchlevel) when we can change the SSL_CIPHER
1969 definition to split the single 'unsigned long mask' bitmap into
1970 multiple values to extend the available space.
1974 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1976 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1977 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1979 *) Add AES IGE and biIGE modes.
1982 *) Change the Unix randomness entropy gathering to use poll() when
1983 possible instead of select(), since the latter has some
1984 undesirable limitations.
1985 [Darryl Miles via Richard Levitte and Bodo Moeller]
1987 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1988 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1989 cannot be implicitly activated as part of, e.g., the "AES" alias.
1990 However, please upgrade to OpenSSL 0.9.9[-dev] for
1991 non-experimental use of the ECC ciphersuites to get TLS extension
1992 support, which is required for curve and point format negotiation
1993 to avoid potential handshake problems.
1996 *) Disable rogue ciphersuites:
1998 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1999 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2000 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2002 The latter two were purportedly from
2003 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2006 Also deactivate the remaining ciphersuites from
2007 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2008 unofficial, and the ID has long expired.
2011 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2012 dual-core machines) and other potential thread-safety issues.
2015 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2016 versions), which is now available for royalty-free use
2017 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2018 Also, add Camellia TLS ciphersuites from RFC 4132.
2020 To minimize changes between patchlevels in the OpenSSL 0.9.8
2021 series, Camellia remains excluded from compilation unless OpenSSL
2022 is configured with 'enable-camellia'.
2025 *) Disable the padding bug check when compression is in use. The padding
2026 bug check assumes the first packet is of even length, this is not
2027 necessarily true if compresssion is enabled and can result in false
2028 positives causing handshake failure. The actual bug test is ancient
2029 code so it is hoped that implementations will either have fixed it by
2030 now or any which still have the bug do not support compression.
2033 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2035 *) When applying a cipher rule check to see if string match is an explicit
2036 cipher suite and only match that one cipher suite if it is.
2039 *) Link in manifests for VC++ if needed.
2040 [Austin Ziegler <halostatue@gmail.com>]
2042 *) Update support for ECC-based TLS ciphersuites according to
2043 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2044 TLS extensions, which are supported starting with the 0.9.9
2045 branch, not in the OpenSSL 0.9.8 branch).
2048 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2049 opaque EVP_CIPHER_CTX handling.
2052 *) Fixes and enhancements to zlib compression code. We now only use
2053 "zlib1.dll" and use the default __cdecl calling convention on Win32
2054 to conform with the standards mentioned here:
2055 http://www.zlib.net/DLL_FAQ.txt
2056 Static zlib linking now works on Windows and the new --with-zlib-include
2057 --with-zlib-lib options to Configure can be used to supply the location
2058 of the headers and library. Gracefully handle case where zlib library
2062 *) Several fixes and enhancements to the OID generation code. The old code
2063 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2064 handle numbers larger than ULONG_MAX, truncated printing and had a
2065 non standard OBJ_obj2txt() behaviour.
2068 *) Add support for building of engines under engine/ as shared libraries
2069 under VC++ build system.
2072 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2073 Hopefully, we will not see any false combination of paths any more.
2076 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2078 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2079 (part of SSL_OP_ALL). This option used to disable the
2080 countermeasure against man-in-the-middle protocol-version
2081 rollback in the SSL 2.0 server implementation, which is a bad
2082 idea. (CVE-2005-2969)
2084 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2085 for Information Security, National Institute of Advanced Industrial
2086 Science and Technology [AIST], Japan)]
2088 *) Add two function to clear and return the verify parameter flags.
2091 *) Keep cipherlists sorted in the source instead of sorting them at
2092 runtime, thus removing the need for a lock.
2095 *) Avoid some small subgroup attacks in Diffie-Hellman.
2096 [Nick Mathewson and Ben Laurie]
2098 *) Add functions for well-known primes.
2101 *) Extended Windows CE support.
2102 [Satoshi Nakamura and Andy Polyakov]
2104 *) Initialize SSL_METHOD structures at compile time instead of during
2105 runtime, thus removing the need for a lock.
2108 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2109 attempting to decrypt each encrypted key in turn. Add support to
2113 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2115 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2118 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2121 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2122 key into the same file any more.
2125 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2128 *) Add -utf8 command line and config file option to 'ca'.
2129 [Stefan <stf@udoma.org]
2131 *) Removed the macro des_crypt(), as it seems to conflict with some
2132 libraries. Use DES_crypt().
2135 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2136 involves renaming the source and generated shared-libs for
2137 both. The engines will accept the corrected or legacy ids
2138 ('ncipher' and '4758_cca' respectively) when binding. NB,
2139 this only applies when building 'shared'.
2140 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2142 *) Add attribute functions to EVP_PKEY structure. Modify
2143 PKCS12_create() to recognize a CSP name attribute and
2144 use it. Make -CSP option work again in pkcs12 utility.
2147 *) Add new functionality to the bn blinding code:
2148 - automatic re-creation of the BN_BLINDING parameters after
2149 a fixed number of uses (currently 32)
2150 - add new function for parameter creation
2151 - introduce flags to control the update behaviour of the
2152 BN_BLINDING parameters
2153 - hide BN_BLINDING structure
2154 Add a second BN_BLINDING slot to the RSA structure to improve
2155 performance when a single RSA object is shared among several
2159 *) Add support for DTLS.
2160 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2162 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2163 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2166 *) Remove buggy and incompletet DH cert support from
2167 ssl/ssl_rsa.c and ssl/s3_both.c
2170 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2171 the apps/openssl applications.
2174 *) Compile clean with "-Wall -Wmissing-prototypes
2175 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2176 DEBUG_SAFESTACK must also be set.
2179 *) Change ./Configure so that certain algorithms can be disabled by default.
2180 The new counterpiece to "no-xxx" is "enable-xxx".
2182 The patented RC5 and MDC2 algorithms will now be disabled unless
2183 "enable-rc5" and "enable-mdc2", respectively, are specified.
2185 (IDEA remains enabled despite being patented. This is because IDEA
2186 is frequently required for interoperability, and there is no license
2187 fee for non-commercial use. As before, "no-idea" can be used to
2188 avoid this algorithm.)
2192 *) Add processing of proxy certificates (see RFC 3820). This work was
2193 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2194 EGEE (Enabling Grids for E-science in Europe).
2197 *) RC4 performance overhaul on modern architectures/implementations, such
2198 as Intel P4, IA-64 and AMD64.
2201 *) New utility extract-section.pl. This can be used specify an alternative
2202 section number in a pod file instead of having to treat each file as
2203 a separate case in Makefile. This can be done by adding two lines to the
2206 =for comment openssl_section:XXX
2208 The blank line is mandatory.
2212 *) New arguments -certform, -keyform and -pass for s_client and s_server
2213 to allow alternative format key and certificate files and passphrase
2217 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2218 update associated structures and add various utility functions.
2220 Add new policy related verify parameters, include policy checking in
2221 standard verify code. Enhance 'smime' application with extra parameters
2222 to support policy checking and print out.
2225 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2226 Nehemiah processors. These extensions support AES encryption in hardware
2227 as well as RNG (though RNG support is currently disabled).
2228 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2230 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2233 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2234 [Andy Polyakov and a number of other people]
2236 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2237 implementation contributed by IBM.
2238 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2240 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2241 exponent rather than 'unsigned long'. There is a corresponding change to
2242 the new 'rsa_keygen' element of the RSA_METHOD structure.
2243 [Jelte Jansen, Geoff Thorpe]
2245 *) Functionality for creating the initial serial number file is now
2246 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2248 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2249 number file to 1, which is bound to cause problems. To avoid
2250 the problems while respecting compatibility between different 0.9.7
2251 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2252 CA.pl for serial number initialization. With the new release 0.9.8,
2253 we can fix the problem directly in the 'ca' utility.)
2256 *) Reduced header interdepencies by declaring more opaque objects in
2257 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2258 give fewer recursive includes, which could break lazy source code - so
2259 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2260 developers should define this symbol when building and using openssl to
2261 ensure they track the recommended behaviour, interfaces, [etc], but
2262 backwards-compatible behaviour prevails when this isn't defined.
2265 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2268 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2269 This will generate a random key of the appropriate length based on the
2270 cipher context. The EVP_CIPHER can provide its own random key generation
2271 routine to support keys of a specific form. This is used in the des and
2272 3des routines to generate a key of the correct parity. Update S/MIME
2273 code to use new functions and hence generate correct parity DES keys.
2274 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2275 valid (weak or incorrect parity).
2278 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2279 as looking them up. This is useful when the verified structure may contain
2280 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2281 present unless the new PKCS7_NO_CRL flag is asserted.
2284 *) Extend ASN1 oid configuration module. It now additionally accepts the
2287 shortName = some long name, 1.2.3.4
2290 *) Reimplemented the BN_CTX implementation. There is now no more static
2291 limitation on the number of variables it can handle nor the depth of the
2292 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2293 information can now expand as required, and rather than having a single
2294 static array of bignums, BN_CTX now uses a linked-list of such arrays
2295 allowing it to expand on demand whilst maintaining the usefulness of
2296 BN_CTX's "bundling".
2299 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2300 to allow all RSA operations to function using a single BN_CTX.
2303 *) Preliminary support for certificate policy evaluation and checking. This
2304 is initially intended to pass the tests outlined in "Conformance Testing
2305 of Relying Party Client Certificate Path Processing Logic" v1.07.
2308 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2309 remained unused and not that useful. A variety of other little bignum
2310 tweaks and fixes have also been made continuing on from the audit (see
2314 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2315 associated ASN1, EVP and SSL functions and old ASN1 macros.
2318 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2319 and this should never fail. So the return value from the use of
2320 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2321 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2324 *) BN_CTX_get() should return zero-valued bignums, providing the same
2325 initialised value as BN_new().
2326 [Geoff Thorpe, suggested by Ulf Möller]
2328 *) Support for inhibitAnyPolicy certificate extension.
2331 *) An audit of the BIGNUM code is underway, for which debugging code is
2332 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2333 is considered valid when processing BIGNUMs, and causes execution to
2334 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2335 further steps are taken to deliberately pollute unused data in BIGNUM
2336 structures to try and expose faulty code further on. For now, openssl will
2337 (in its default mode of operation) continue to tolerate the inconsistent
2338 forms that it has tolerated in the past, but authors and packagers should
2339 consider trying openssl and their own applications when compiled with
2340 these debugging symbols defined. It will help highlight potential bugs in
2341 their own code, and will improve the test coverage for OpenSSL itself. At
2342 some point, these tighter rules will become openssl's default to improve
2343 maintainability, though the assert()s and other overheads will remain only
2344 in debugging configurations. See bn.h for more details.
2345 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2347 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2348 that can only be obtained through BN_CTX_new() (which implicitly
2349 initialises it). The presence of this function only made it possible
2350 to overwrite an existing structure (and cause memory leaks).
2353 *) Because of the callback-based approach for implementing LHASH as a
2354 template type, lh_insert() adds opaque objects to hash-tables and
2355 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2356 to clean up those corresponding objects before destroying the hash table
2357 (and losing the object pointers). So some over-zealous constifications in
2358 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2359 objects as "const" and the lh_doall[_arg] callback wrappers are not
2360 prototyped to have "const" restrictions on the object pointers they are
2361 given (and so aren't required to cast them away any more).
2364 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2365 (speed) prefers to use its own implementation. The two implementations
2366 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2367 its object type properly exposed (MS_TM) instead of casting to/from "char
2368 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2369 aren't necessarily the greatest nomenclatures - but this is what was used
2370 internally to the implementation so I've used that for now.
2373 *) Ensure that deprecated functions do not get compiled when
2374 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2375 the self-tests were still using deprecated key-generation functions so
2376 these have been updated also.
2379 *) Reorganise PKCS#7 code to separate the digest location functionality
2380 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2381 New function PKCS7_set_digest() to set the digest type for PKCS#7
2382 digestedData type. Add additional code to correctly generate the
2383 digestedData type and add support for this type in PKCS7 initialization
2387 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2388 structure of type "other".
2391 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2392 sure the loop does correctly stop and breaking ("division by zero")
2393 modulus operations are not performed. The (pre-generated) prime
2394 table crypto/bn/bn_prime.h was already correct, but it could not be
2395 re-generated on some platforms because of the "division by zero"
2396 situation in the script.
2397 [Ralf S. Engelschall]
2399 *) Update support for ECC-based TLS ciphersuites according to
2400 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2401 SHA-1 now is only used for "small" curves (where the
2402 representation of a field element takes up to 24 bytes); for
2403 larger curves, the field element resulting from ECDH is directly
2404 used as premaster secret.
2405 [Douglas Stebila (Sun Microsystems Laboratories)]
2407 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2408 curve secp160r1 to the tests.
2409 [Douglas Stebila (Sun Microsystems Laboratories)]
2411 *) Add the possibility to load symbols globally with DSO.
2412 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2414 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2415 control of the error stack.
2418 *) Add support for STORE in ENGINE.
2421 *) Add the STORE type. The intention is to provide a common interface
2422 to certificate and key stores, be they simple file-based stores, or
2423 HSM-type store, or LDAP stores, or...
2424 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2427 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2428 pass a list of arguments to any function as well as provide a way
2429 for a function to pass data back to the caller.
2432 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2433 works like BUF_strdup() but can be used to duplicate a portion of
2434 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2438 *) Add the function sk_find_ex() which works like sk_find(), but will
2439 return an index to an element even if an exact match couldn't be
2440 found. The index is guaranteed to point at the element where the
2441 searched-for key would be inserted to preserve sorting order.
2444 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2445 takes an extra flags argument for optional functionality. Currently,
2446 the following flags are defined:
2448 OBJ_BSEARCH_VALUE_ON_NOMATCH
2449 This one gets OBJ_bsearch_ex() to return a pointer to the first
2450 element where the comparing function returns a negative or zero
2453 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2454 This one gets OBJ_bsearch_ex() to return a pointer to the first
2455 element where the comparing function returns zero. This is useful
2456 if there are more than one element where the comparing function
2460 *) Make it possible to create self-signed certificates with 'openssl ca'
2461 in such a way that the self-signed certificate becomes part of the
2462 CA database and uses the same mechanisms for serial number generation
2463 as all other certificate signing. The new flag '-selfsign' enables
2464 this functionality. Adapt CA.sh and CA.pl.in.
2467 *) Add functionality to check the public key of a certificate request
2468 against a given private. This is useful to check that a certificate
2469 request can be signed by that key (self-signing).
2472 *) Make it possible to have multiple active certificates with the same
2473 subject in the CA index file. This is done only if the keyword
2474 'unique_subject' is set to 'no' in the main CA section (default
2475 if 'CA_default') of the configuration file. The value is saved
2476 with the database itself in a separate index attribute file,
2477 named like the index file with '.attr' appended to the name.
2480 *) Generate muti valued AVAs using '+' notation in config files for
2484 *) Support for nameConstraints certificate extension.
2487 *) Support for policyConstraints certificate extension.
2490 *) Support for policyMappings certificate extension.
2493 *) Make sure the default DSA_METHOD implementation only uses its
2494 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2495 and change its own handlers to be NULL so as to remove unnecessary
2496 indirection. This lets alternative implementations fallback to the
2497 default implementation more easily.
2500 *) Support for directoryName in GeneralName related extensions
2504 *) Make it possible to link applications using Makefile.shared.
2505 Make that possible even when linking against static libraries!
2508 *) Support for single pass processing for S/MIME signing. This now
2509 means that S/MIME signing can be done from a pipe, in addition
2510 cleartext signing (multipart/signed type) is effectively streaming
2511 and the signed data does not need to be all held in memory.
2513 This is done with a new flag PKCS7_STREAM. When this flag is set
2514 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2515 is done after the data is output (and digests calculated) in
2516 SMIME_write_PKCS7().
2519 *) Add full support for -rpath/-R, both in shared libraries and
2520 applications, at least on the platforms where it's known how
2524 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2525 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2526 will now compute a table of multiples of the generator that
2527 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2528 faster (notably in the case of a single point multiplication,
2529 scalar * generator).
2530 [Nils Larsch, Bodo Moeller]
2532 *) IPv6 support for certificate extensions. The various extensions
2533 which use the IP:a.b.c.d can now take IPv6 addresses using the
2534 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2538 *) Added an ENGINE that implements RSA by performing private key
2539 exponentiations with the GMP library. The conversions to and from
2540 GMP's mpz_t format aren't optimised nor are any montgomery forms
2541 cached, and on x86 it appears OpenSSL's own performance has caught up.
2542 However there are likely to be other architectures where GMP could
2543 provide a boost. This ENGINE is not built in by default, but it can be
2544 specified at Configure time and should be accompanied by the necessary
2545 linker additions, eg;
2546 ./config -DOPENSSL_USE_GMP -lgmp
2549 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2550 testing availability of engines with "-t" - the old behaviour is
2551 produced by increasing the feature's verbosity with "-tt".
2554 *) ECDSA routines: under certain error conditions uninitialized BN objects
2555 could be freed. Solution: make sure initialization is performed early
2556 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2560 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2561 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2562 software implementations. For DSA and DH, parameter generation can
2563 also be overriden by providing the appropriate method callbacks.
2566 *) Change the "progress" mechanism used in key-generation and
2567 primality testing to functions that take a new BN_GENCB pointer in
2568 place of callback/argument pairs. The new API functions have "_ex"
2569 postfixes and the older functions are reimplemented as wrappers for
2570 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2571 declarations of the old functions to help (graceful) attempts to
2572 migrate to the new functions. Also, the new key-generation API
2573 functions operate on a caller-supplied key-structure and return
2574 success/failure rather than returning a key or NULL - this is to
2575 help make "keygen" another member function of RSA_METHOD etc.
2577 Example for using the new callback interface:
2579 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2583 BN_GENCB_set(&my_cb, my_callback, my_arg);
2585 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2586 /* For the meaning of a, b in calls to my_callback(), see the
2587 * documentation of the function that calls the callback.
2588 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2589 * my_callback should return 1 if it wants BN_is_prime_ex()
2590 * to continue, or 0 to stop.
2595 *) Change the ZLIB compression method to be stateful, and make it
2596 available to TLS with the number defined in
2597 draft-ietf-tls-compression-04.txt.
2600 *) Add the ASN.1 structures and functions for CertificatePair, which
2601 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2603 CertificatePair ::= SEQUENCE {
2604 forward [0] Certificate OPTIONAL,
2605 reverse [1] Certificate OPTIONAL,
2606 -- at least one of the pair shall be present -- }
2608 Also implement the PEM functions to read and write certificate
2609 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2611 This needed to be defined, mostly for the sake of the LDAP
2612 attribute crossCertificatePair, but may prove useful elsewhere as
2616 *) Make it possible to inhibit symlinking of shared libraries in
2617 Makefile.shared, for Cygwin's sake.
2620 *) Extend the BIGNUM API by creating a function
2621 void BN_set_negative(BIGNUM *a, int neg);
2622 and a macro that behave like
2623 int BN_is_negative(const BIGNUM *a);
2625 to avoid the need to access 'a->neg' directly in applications.
2628 *) Implement fast modular reduction for pseudo-Mersenne primes
2629 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2630 EC_GROUP_new_curve_GFp() will now automatically use this
2632 [Nils Larsch <nla@trustcenter.de>]
2634 *) Add new lock type (CRYPTO_LOCK_BN).
2637 *) Change the ENGINE framework to automatically load engines
2638 dynamically from specific directories unless they could be
2639 found to already be built in or loaded. Move all the
2640 current engines except for the cryptodev one to a new
2642 The engines in engines/ are built as shared libraries if
2643 the "shared" options was given to ./Configure or ./config.
2644 Otherwise, they are inserted in libcrypto.a.
2645 /usr/local/ssl/engines is the default directory for dynamic
2646 engines, but that can be overriden at configure time through
2647 the usual use of --prefix and/or --openssldir, and at run
2648 time with the environment variable OPENSSL_ENGINES.
2649 [Geoff Thorpe and Richard Levitte]
2651 *) Add Makefile.shared, a helper makefile to build shared
2652 libraries. Addapt Makefile.org.
2655 *) Add version info to Win32 DLLs.
2656 [Peter 'Luna' Runestig" <peter@runestig.com>]
2658 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2659 can be added using this API to created arbitrary PKCS#12
2660 files while avoiding the low level API.
2662 New options to PKCS12_create(), key or cert can be NULL and
2663 will then be omitted from the output file. The encryption
2664 algorithm NIDs can be set to -1 for no encryption, the mac
2665 iteration count can be set to 0 to omit the mac.
2667 Enhance pkcs12 utility by making the -nokeys and -nocerts
2668 options work when creating a PKCS#12 file. New option -nomac
2669 to omit the mac, NONE can be set for an encryption algorithm.
2670 New code is modified to use the enhanced PKCS12_create()
2671 instead of the low level API.
2674 *) Extend ASN1 encoder to support indefinite length constructed
2675 encoding. This can output sequences tags and octet strings in
2676 this form. Modify pk7_asn1.c to support indefinite length
2677 encoding. This is experimental and needs additional code to
2678 be useful, such as an ASN1 bio and some enhanced streaming
2681 Extend template encode functionality so that tagging is passed
2682 down to the template encoder.
2685 *) Let 'openssl req' fail if an argument to '-newkey' is not
2686 recognized instead of using RSA as a default.
2689 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2690 As these are not official, they are not included in "ALL";
2691 the "ECCdraft" ciphersuite group alias can be used to select them.
2692 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2694 *) Add ECDH engine support.
2695 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2697 *) Add ECDH in new directory crypto/ecdh/.
2698 [Douglas Stebila (Sun Microsystems Laboratories)]
2700 *) Let BN_rand_range() abort with an error after 100 iterations
2701 without success (which indicates a broken PRNG).
2704 *) Change BN_mod_sqrt() so that it verifies that the input value
2705 is really the square of the return value. (Previously,
2706 BN_mod_sqrt would show GIGO behaviour.)
2709 *) Add named elliptic curves over binary fields from X9.62, SECG,
2710 and WAP/WTLS; add OIDs that were still missing.
2712 [Sheueling Chang Shantz and Douglas Stebila
2713 (Sun Microsystems Laboratories)]
2715 *) Extend the EC library for elliptic curves over binary fields
2716 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2719 EC_GF2m_simple_method
2723 EC_GROUP_new_curve_GF2m
2724 EC_GROUP_set_curve_GF2m
2725 EC_GROUP_get_curve_GF2m
2726 EC_POINT_set_affine_coordinates_GF2m
2727 EC_POINT_get_affine_coordinates_GF2m
2728 EC_POINT_set_compressed_coordinates_GF2m
2730 Point compression for binary fields is disabled by default for
2731 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2734 As binary polynomials are represented as BIGNUMs, various members
2735 of the EC_GROUP and EC_POINT data structures can be shared
2736 between the implementations for prime fields and binary fields;
2737 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2738 are essentially identical to their ..._GFp counterparts.
2739 (For simplicity, the '..._GFp' prefix has been dropped from
2740 various internal method names.)
2742 An internal 'field_div' method (similar to 'field_mul' and
2743 'field_sqr') has been added; this is used only for binary fields.
2745 [Sheueling Chang Shantz and Douglas Stebila
2746 (Sun Microsystems Laboratories)]
2748 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2749 through methods ('mul', 'precompute_mult').
2751 The generic implementations (now internally called 'ec_wNAF_mul'
2752 and 'ec_wNAF_precomputed_mult') remain the default if these
2753 methods are undefined.
2755 [Sheueling Chang Shantz and Douglas Stebila
2756 (Sun Microsystems Laboratories)]
2758 *) New function EC_GROUP_get_degree, which is defined through
2759 EC_METHOD. For curves over prime fields, this returns the bit
2760 length of the modulus.
2762 [Sheueling Chang Shantz and Douglas Stebila
2763 (Sun Microsystems Laboratories)]
2765 *) New functions EC_GROUP_dup, EC_POINT_dup.
2766 (These simply call ..._new and ..._copy).
2768 [Sheueling Chang Shantz and Douglas Stebila
2769 (Sun Microsystems Laboratories)]
2771 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2772 Polynomials are represented as BIGNUMs (where the sign bit is not
2773 used) in the following functions [macros]:
2776 BN_GF2m_sub [= BN_GF2m_add]
2777 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2778 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2779 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2781 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2782 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2783 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2784 BN_GF2m_cmp [= BN_ucmp]
2786 (Note that only the 'mod' functions are actually for fields GF(2^m).
2787 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2789 For some functions, an the irreducible polynomial defining a
2790 field can be given as an 'unsigned int[]' with strictly
2791 decreasing elements giving the indices of those bits that are set;
2792 i.e., p[] represents the polynomial
2793 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2795 p[0] > p[1] > ... > p[k] = 0.
2796 This applies to the following functions:
2801 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2802 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2804 BN_GF2m_mod_sqrt_arr
2805 BN_GF2m_mod_solve_quad_arr
2809 Conversion can be performed by the following functions:
2814 bntest.c has additional tests for binary polynomial arithmetic.
2816 Two implementations for BN_GF2m_mod_div() are available.
2817 The default algorithm simply uses BN_GF2m_mod_inv() and
2818 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2819 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2820 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2822 [Sheueling Chang Shantz and Douglas Stebila
2823 (Sun Microsystems Laboratories)]
2825 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2826 functionality is disabled at compile-time.
2827 [Douglas Stebila <douglas.stebila@sun.com>]
2829 *) Change default behaviour of 'openssl asn1parse' so that more
2830 information is visible when viewing, e.g., a certificate:
2832 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2833 mode the content of non-printable OCTET STRINGs is output in a
2834 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2835 avoid the appearance of a printable string.
2836 [Nils Larsch <nla@trustcenter.de>]
2838 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2840 EC_GROUP_set_asn1_flag()
2841 EC_GROUP_get_asn1_flag()
2842 EC_GROUP_set_point_conversion_form()
2843 EC_GROUP_get_point_conversion_form()
2844 These control ASN1 encoding details:
2845 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2846 has been set to OPENSSL_EC_NAMED_CURVE.
2847 - Points are encoded in uncompressed form by default; options for
2848 asn1_for are as for point2oct, namely
2849 POINT_CONVERSION_COMPRESSED
2850 POINT_CONVERSION_UNCOMPRESSED
2851 POINT_CONVERSION_HYBRID
2853 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2856 EC_GROUP_get0_seed()
2857 EC_GROUP_get_seed_len()
2858 This is used only for ASN1 purposes (so far).
2859 [Nils Larsch <nla@trustcenter.de>]
2861 *) Add 'field_type' member to EC_METHOD, which holds the NID
2862 of the appropriate field type OID. The new function
2863 EC_METHOD_get_field_type() returns this value.
2864 [Nils Larsch <nla@trustcenter.de>]
2869 EC_POINT_point2hex()
2870 EC_POINT_hex2point()
2871 providing useful interfaces to EC_POINT_point2oct() and
2872 EC_POINT_oct2point().
2873 [Nils Larsch <nla@trustcenter.de>]
2875 *) Change internals of the EC library so that the functions
2876 EC_GROUP_set_generator()
2877 EC_GROUP_get_generator()
2878 EC_GROUP_get_order()
2879 EC_GROUP_get_cofactor()
2880 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2881 to methods, which would lead to unnecessary code duplication when
2882 adding different types of curves.
2883 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2885 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2886 arithmetic, and such that modified wNAFs are generated
2887 (which avoid length expansion in many cases).
2890 *) Add a function EC_GROUP_check_discriminant() (defined via
2891 EC_METHOD) that verifies that the curve discriminant is non-zero.
2893 Add a function EC_GROUP_check() that makes some sanity tests
2894 on a EC_GROUP, its generator and order. This includes
2895 EC_GROUP_check_discriminant().
2896 [Nils Larsch <nla@trustcenter.de>]
2898 *) Add ECDSA in new directory crypto/ecdsa/.
2900 Add applications 'openssl ecparam' and 'openssl ecdsa'
2901 (these are based on 'openssl dsaparam' and 'openssl dsa').
2903 ECDSA support is also included in various other files across the
2904 library. Most notably,
2905 - 'openssl req' now has a '-newkey ecdsa:file' option;
2906 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2907 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2908 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2909 them suitable for ECDSA where domain parameters must be
2910 extracted before the specific public key;
2911 - ECDSA engine support has been added.
2912 [Nils Larsch <nla@trustcenter.de>]
2914 *) Include some named elliptic curves, and add OIDs from X9.62,
2915 SECG, and WAP/WTLS. Each curve can be obtained from the new
2917 EC_GROUP_new_by_curve_name(),
2918 and the list of available named curves can be obtained with
2919 EC_get_builtin_curves().
2920 Also add a 'curve_name' member to EC_GROUP objects, which can be
2922 EC_GROUP_set_curve_name()
2923 EC_GROUP_get_curve_name()
2924 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2926 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2927 was actually never needed) and in BN_mul(). The removal in BN_mul()
2928 required a small change in bn_mul_part_recursive() and the addition
2929 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2930 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2931 bn_sub_words() and bn_add_words() except they take arrays with
2935 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2937 *) Cleanse PEM buffers before freeing them since they may contain
2939 [Benjamin Bennett <ben@psc.edu>]
2941 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2942 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2943 authentication-only ciphersuites.
2946 *) Since AES128 and AES256 share a single mask bit in the logic of
2947 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2948 kludge to work properly if AES128 is available and AES256 isn't.
2951 *) Expand security boundary to match 1.1.1 module.
2954 *) Remove redundant features: hash file source, editing of test vectors
2955 modify fipsld to use external fips_premain.c signature.
2958 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2959 run algorithm test programs.
2962 *) Make algorithm test programs more tolerant of whitespace.
2965 *) Have SSL/TLS server implementation tolerate "mismatched" record
2966 protocol version while receiving ClientHello even if the
2967 ClientHello is fragmented. (The server can't insist on the
2968 particular protocol version it has chosen before the ServerHello
2969 message has informed the client about his choice.)
2972 *) Load error codes if they are not already present instead of using a
2973 static variable. This allows them to be cleanly unloaded and reloaded.
2976 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2978 *) Introduce limits to prevent malicious keys being able to
2979 cause a denial of service. (CVE-2006-2940)
2980 [Steve Henson, Bodo Moeller]
2982 *) Fix ASN.1 parsing of certain invalid structures that can result
2983 in a denial of service. (CVE-2006-2937) [Steve Henson]
2985 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2986 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2988 *) Fix SSL client code which could crash if connecting to a
2989 malicious SSLv2 server. (CVE-2006-4343)
2990 [Tavis Ormandy and Will Drewry, Google Security Team]
2992 *) Change ciphersuite string processing so that an explicit
2993 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2994 will no longer include "AES128-SHA"), and any other similar
2995 ciphersuite (same bitmap) from *other* protocol versions (so that
2996 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2997 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2998 changes from 0.9.8b and 0.9.8d.
3001 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3003 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3004 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3006 *) Change the Unix randomness entropy gathering to use poll() when
3007 possible instead of select(), since the latter has some
3008 undesirable limitations.
3009 [Darryl Miles via Richard Levitte and Bodo Moeller]
3011 *) Disable rogue ciphersuites:
3013 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3014 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3015 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3017 The latter two were purportedly from
3018 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3021 Also deactive the remaining ciphersuites from
3022 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3023 unofficial, and the ID has long expired.
3026 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3027 dual-core machines) and other potential thread-safety issues.
3030 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3032 *) Adapt fipsld and the build system to link against the validated FIPS
3033 module in FIPS mode.
3036 *) Fixes for VC++ 2005 build under Windows.
3039 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3040 from a Windows bash shell such as MSYS. It is autodetected from the
3041 "config" script when run from a VC++ environment. Modify standard VC++
3042 build to use fipscanister.o from the GNU make build.
3045 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3047 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3048 The value now differs depending on if you build for FIPS or not.
3049 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3050 safely run with a non-FIPSed libcrypto, as it may crash because of
3051 the difference induced by this change.
3054 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3056 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3057 (part of SSL_OP_ALL). This option used to disable the
3058 countermeasure against man-in-the-middle protocol-version
3059 rollback in the SSL 2.0 server implementation, which is a bad
3060 idea. (CVE-2005-2969)
3062 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3063 for Information Security, National Institute of Advanced Industrial
3064 Science and Technology [AIST], Japan)]
3066 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3067 mainly for FIPS compliance and not fully integrated at this stage.
3070 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3071 the exponentiation using a fixed-length exponent. (Otherwise,
3072 the information leaked through timing could expose the secret key
3073 after many signatures; cf. Bleichenbacher's attack on DSA with
3077 *) Make a new fixed-window mod_exp implementation the default for
3078 RSA, DSA, and DH private-key operations so that the sequence of
3079 squares and multiplies and the memory access pattern are
3080 independent of the particular secret key. This will mitigate
3081 cache-timing and potential related attacks.
3083 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3084 and this is automatically used by BN_mod_exp_mont() if the new flag
3085 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3086 will use this BN flag for private exponents unless the flag
3087 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3088 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3090 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3092 *) Change the client implementation for SSLv23_method() and
3093 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3094 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3095 (Previously, the SSL 2.0 backwards compatible Client Hello
3096 message format would be used even with SSL_OP_NO_SSLv2.)
3099 *) Add support for smime-type MIME parameter in S/MIME messages which some
3103 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3104 a threadsafe manner. Modify rsa code to use new function and add calls
3105 to dsa and dh code (which had race conditions before).
3108 *) Include the fixed error library code in the C error file definitions
3109 instead of fixing them up at runtime. This keeps the error code
3110 structures constant.
3113 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3115 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3118 *) Fixes for newer kerberos headers. NB: the casts are needed because
3119 the 'length' field is signed on one version and unsigned on another
3120 with no (?) obvious way to tell the difference, without these VC++
3121 complains. Also the "definition" of FAR (blank) is no longer included
3122 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3123 some needed definitions.
3126 *) Undo Cygwin change.
3129 *) Added support for proxy certificates according to RFC 3820.
3130 Because they may be a security thread to unaware applications,
3131 they must be explicitely allowed in run-time. See
3132 docs/HOWTO/proxy_certificates.txt for further information.
3135 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3137 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3138 server and client random values. Previously
3139 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3140 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3142 This change has negligible security impact because:
3144 1. Server and client random values still have 24 bytes of pseudo random
3147 2. Server and client random values are sent in the clear in the initial
3150 3. The master secret is derived using the premaster secret (48 bytes in
3151 size for static RSA ciphersuites) as well as client server and random
3154 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3157 [Stephen Henson, reported by UK NISCC]
3159 *) Use Windows randomness collection on Cygwin.
3162 *) Fix hang in EGD/PRNGD query when communication socket is closed
3163 prematurely by EGD/PRNGD.
3164 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3166 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3169 *) Back-port of selected performance improvements from development
3170 branch, as well as improved support for PowerPC platforms.
3173 *) Add lots of checks for memory allocation failure, error codes to indicate
3174 failure and freeing up memory if a failure occurs.
3175 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3177 *) Add new -passin argument to dgst.
3180 *) Perform some character comparisons of different types in X509_NAME_cmp:
3181 this is needed for some certificates that reencode DNs into UTF8Strings
3182 (in violation of RFC3280) and can't or wont issue name rollover
3186 *) Make an explicit check during certificate validation to see that
3187 the CA setting in each certificate on the chain is correct. As a
3188 side effect always do the following basic checks on extensions,
3189 not just when there's an associated purpose to the check:
3191 - if there is an unhandled critical extension (unless the user
3192 has chosen to ignore this fault)
3193 - if the path length has been exceeded (if one is set at all)
3194 - that certain extensions fit the associated purpose (if one has
3198 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3200 *) Avoid a race condition when CRLs are checked in a multi threaded
3201 environment. This would happen due to the reordering of the revoked
3202 entries during signature checking and serial number lookup. Now the
3203 encoding is cached and the serial number sort performed under a lock.
3204 Add new STACK function sk_is_sorted().
3207 *) Add Delta CRL to the extension code.
3210 *) Various fixes to s3_pkt.c so alerts are sent properly.
3211 [David Holmes <d.holmes@f5.com>]
3213 *) Reduce the chances of duplicate issuer name and serial numbers (in
3214 violation of RFC3280) using the OpenSSL certificate creation utilities.
3215 This is done by creating a random 64 bit value for the initial serial
3216 number when a serial number file is created or when a self signed
3217 certificate is created using 'openssl req -x509'. The initial serial
3218 number file is created using 'openssl x509 -next_serial' in CA.pl
3219 rather than being initialized to 1.
3222 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3224 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3225 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3226 [Joe Orton, Steve Henson]
3228 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3230 [Joe Orton, Steve Henson]
3232 *) Make it possible to have multiple active certificates with the same
3233 subject in the CA index file. This is done only if the keyword
3234 'unique_subject' is set to 'no' in the main CA section (default
3235 if 'CA_default') of the configuration file. The value is saved
3236 with the database itself in a separate index attribute file,
3237 named like the index file with '.attr' appended to the name.
3240 *) X509 verify fixes. Disable broken certificate workarounds when
3241 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3242 keyUsage extension present. Don't accept CRLs with unhandled critical
3243 extensions: since verify currently doesn't process CRL extensions this
3244 rejects a CRL with *any* critical extensions. Add new verify error codes
3248 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3249 A clarification of RFC2560 will require the use of OCTET STRINGs and
3250 some implementations cannot handle the current raw format. Since OpenSSL
3251 copies and compares OCSP nonces as opaque blobs without any attempt at
3252 parsing them this should not create any compatibility issues.
3255 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3256 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3257 this HMAC (and other) operations are several times slower than OpenSSL
3261 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3262 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3264 *) Use the correct content when signing type "other".
3267 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3269 *) Fix various bugs revealed by running the NISCC test suite:
3271 Stop out of bounds reads in the ASN1 code when presented with
3272 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3274 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3276 If verify callback ignores invalid public key errors don't try to check
3277 certificate signature with the NULL public key.
3281 *) New -ignore_err option in ocsp application to stop the server
3282 exiting on the first error in a request.
3285 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3286 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3290 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3291 extra data after the compression methods not only for TLS 1.0
3292 but also for SSL 3.0 (as required by the specification).
3293 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3295 *) Change X509_certificate_type() to mark the key as exported/exportable
3296 when it's 512 *bits* long, not 512 bytes.
3299 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3300 blocks during encryption.
3303 *) Various fixes to base64 BIO and non blocking I/O. On write
3304 flushes were not handled properly if the BIO retried. On read
3305 data was not being buffered properly and had various logic bugs.
3306 This also affects blocking I/O when the data being decoded is a
3310 *) Various S/MIME bugfixes and compatibility changes:
3311 output correct application/pkcs7 MIME type if
3312 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3313 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3314 of files as .eml work). Correctly handle very long lines in MIME
3318 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3320 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3321 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3322 a protocol version number mismatch like a decryption error
3323 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3326 *) Turn on RSA blinding by default in the default implementation
3327 to avoid a timing attack. Applications that don't want it can call
3328 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3329 They would be ill-advised to do so in most cases.
3330 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3332 *) Change RSA blinding code so that it works when the PRNG is not
3333 seeded (in this case, the secret RSA exponent is abused as
3334 an unpredictable seed -- if it is not unpredictable, there
3335 is no point in blinding anyway). Make RSA blinding thread-safe
3336 by remembering the creator's thread ID in rsa->blinding and
3337 having all other threads use local one-time blinding factors
3338 (this requires more computation than sharing rsa->blinding, but
3339 avoids excessive locking; and if an RSA object is not shared
3340 between threads, blinding will still be very fast).
3343 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3344 ENGINE as defaults for all supported algorithms irrespective of
3345 the 'flags' parameter. 'flags' is now honoured, so applications
3346 should make sure they are passing it correctly.
3349 *) Target "mingw" now allows native Windows code to be generated in
3350 the Cygwin environment as well as with the MinGW compiler.
3353 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3355 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3356 via timing by performing a MAC computation even if incorrrect
3357 block cipher padding has been found. This is a countermeasure
3358 against active attacks where the attacker has to distinguish
3359 between bad padding and a MAC verification error. (CVE-2003-0078)
3361 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3362 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3363 Martin Vuagnoux (EPFL, Ilion)]
3365 *) Make the no-err option work as intended. The intention with no-err
3366 is not to have the whole error stack handling routines removed from
3367 libcrypto, it's only intended to remove all the function name and
3368 reason texts, thereby removing some of the footprint that may not
3369 be interesting if those errors aren't displayed anyway.
3371 NOTE: it's still possible for any application or module to have it's
3372 own set of error texts inserted. The routines are there, just not
3373 used by default when no-err is given.
3376 *) Add support for FreeBSD on IA64.
3377 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3379 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3380 Kerberos function mit_des_cbc_cksum(). Before this change,
3381 the value returned by DES_cbc_cksum() was like the one from
3382 mit_des_cbc_cksum(), except the bytes were swapped.
3383 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3385 *) Allow an application to disable the automatic SSL chain building.
3386 Before this a rather primitive chain build was always performed in
3387 ssl3_output_cert_chain(): an application had no way to send the
3388 correct chain if the automatic operation produced an incorrect result.
3390 Now the chain builder is disabled if either:
3392 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3394 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3396 The reasoning behind this is that an application would not want the
3397 auto chain building to take place if extra chain certificates are
3398 present and it might also want a means of sending no additional
3399 certificates (for example the chain has two certificates and the
3403 *) Add the possibility to build without the ENGINE framework.
3404 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3406 *) Under Win32 gmtime() can return NULL: check return value in
3407 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3410 *) DSA routines: under certain error conditions uninitialized BN objects
3411 could be freed. Solution: make sure initialization is performed early
3412 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3413 Nils Larsch <nla@trustcenter.de> via PR#459)
3416 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3417 checked on reconnect on the client side, therefore session resumption
3418 could still fail with a "ssl session id is different" error. This
3419 behaviour is masked when SSL_OP_ALL is used due to
3420 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3421 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3422 followup to PR #377.
3425 *) IA-32 assembler support enhancements: unified ELF targets, support
3426 for SCO/Caldera platforms, fix for Cygwin shared build.
3429 *) Add support for FreeBSD on sparc64. As a consequence, support for
3430 FreeBSD on non-x86 processors is separate from x86 processors on
3431 the config script, much like the NetBSD support.
3432 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3434 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3436 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3439 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3440 code (06) was taken as the first octet of the session ID and the last
3441 octet was ignored consequently. As a result SSLv2 client side session
3442 caching could not have worked due to the session ID mismatch between
3444 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3448 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3449 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3453 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3454 seems that in spite of existing for more than a year, many application
3455 author have done nothing to provide the necessary callbacks, which
3456 means that this particular engine will not work properly anywhere.
3457 This is a very unfortunate situation which forces us, in the name
3458 of usability, to give the hw_ncipher.c a static lock, which is part
3460 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3461 appear in 0.9.8 or later. We EXPECT application authors to have
3462 dealt properly with this when 0.9.8 is released (unless we actually
3463 make such changes in the libcrypto locking code that changes will
3464 have to be made anyway).
3467 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3468 octets have been read, EOF or an error occurs. Without this change
3469 some truncated ASN1 structures will not produce an error.
3472 *) Disable Heimdal support, since it hasn't been fully implemented.
3473 Still give the possibility to force the use of Heimdal, but with
3474 warnings and a request that patches get sent to openssl-dev.
3477 *) Add the VC-CE target, introduce the WINCE sysname, and add
3478 INSTALL.WCE and appropriate conditionals to make it build.
3479 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3481 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3482 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3483 edit numbers of the version.
3484 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3486 *) Introduce safe string copy and catenation functions
3487 (BUF_strlcpy() and BUF_strlcat()).
3488 [Ben Laurie (CHATS) and Richard Levitte]
3490 *) Avoid using fixed-size buffers for one-line DNs.
3491 [Ben Laurie (CHATS)]
3493 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3494 resizing buffers containing secrets, and use where appropriate.
3495 [Ben Laurie (CHATS)]
3497 *) Avoid using fixed size buffers for configuration file location.
3498 [Ben Laurie (CHATS)]
3500 *) Avoid filename truncation for various CA files.
3501 [Ben Laurie (CHATS)]
3503 *) Use sizeof in preference to magic numbers.
3504 [Ben Laurie (CHATS)]
3506 *) Avoid filename truncation in cert requests.
3507 [Ben Laurie (CHATS)]
3509 *) Add assertions to check for (supposedly impossible) buffer
3511 [Ben Laurie (CHATS)]
3513 *) Don't cache truncated DNS entries in the local cache (this could
3514 potentially lead to a spoofing attack).
3515 [Ben Laurie (CHATS)]
3517 *) Fix various buffers to be large enough for hex/decimal
3518 representations in a platform independent manner.
3519 [Ben Laurie (CHATS)]
3521 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3522 resizing buffers containing secrets, and use where appropriate.
3523 [Ben Laurie (CHATS)]
3525 *) Add BIO_indent() to avoid much slightly worrying code to do
3527 [Ben Laurie (CHATS)]
3529 *) Convert sprintf()/BIO_puts() to BIO_printf().
3530 [Ben Laurie (CHATS)]
3532 *) buffer_gets() could terminate with the buffer only half
3534 [Ben Laurie (CHATS)]
3536 *) Add assertions to prevent user-supplied crypto functions from
3537 overflowing internal buffers by having large block sizes, etc.
3538 [Ben Laurie (CHATS)]
3540 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3542 [Ben Laurie (CHATS)]
3544 *) Eliminate unused copy of key in RC4.
3545 [Ben Laurie (CHATS)]
3547 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3548 [Ben Laurie (CHATS)]
3550 *) Fix off-by-one error in EGD path.
3551 [Ben Laurie (CHATS)]
3553 *) If RANDFILE path is too long, ignore instead of truncating.
3554 [Ben Laurie (CHATS)]
3556 *) Eliminate unused and incorrectly sized X.509 structure
3558 [Ben Laurie (CHATS)]
3560 *) Eliminate unused and dangerous function knumber().
3561 [Ben Laurie (CHATS)]
3563 *) Eliminate unused and dangerous structure, KSSL_ERR.
3564 [Ben Laurie (CHATS)]
3566 *) Protect against overlong session ID context length in an encoded
3567 session object. Since these are local, this does not appear to be
3569 [Ben Laurie (CHATS)]
3571 *) Change from security patch (see 0.9.6e below) that did not affect
3572 the 0.9.6 release series:
3574 Remote buffer overflow in SSL3 protocol - an attacker could
3575 supply an oversized master key in Kerberos-enabled versions.
3577 [Ben Laurie (CHATS)]
3579 *) Change the SSL kerb5 codes to match RFC 2712.
3582 *) Make -nameopt work fully for req and add -reqopt switch.
3583 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3585 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3586 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3588 *) Make sure tests can be performed even if the corresponding algorithms
3589 have been removed entirely. This was also the last step to make
3590 OpenSSL compilable with DJGPP under all reasonable conditions.
3591 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3593 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3594 to allow version independent disabling of normally unselected ciphers,
3595 which may be activated as a side-effect of selecting a single cipher.
3597 (E.g., cipher list string "RSA" enables ciphersuites that are left
3598 out of "ALL" because they do not provide symmetric encryption.
3599 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3600 [Lutz Jaenicke, Bodo Moeller]
3602 *) Add appropriate support for separate platform-dependent build
3603 directories. The recommended way to make a platform-dependent
3604 build directory is the following (tested on Linux), maybe with
3607 # Place yourself outside of the OpenSSL source tree. In
3608 # this example, the environment variable OPENSSL_SOURCE
3609 # is assumed to contain the absolute OpenSSL source directory.
3610 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3611 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3612 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3613 mkdir -p `dirname $F`
3614 ln -s $OPENSSL_SOURCE/$F $F
3617 To be absolutely sure not to disturb the source tree, a "make clean"
3618 is a good thing. If it isn't successfull, don't worry about it,
3619 it probably means the source directory is very clean.
3622 *) Make sure any ENGINE control commands make local copies of string
3623 pointers passed to them whenever necessary. Otherwise it is possible
3624 the caller may have overwritten (or deallocated) the original string
3625 data when a later ENGINE operation tries to use the stored values.
3626 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3628 *) Improve diagnostics in file reading and command-line digests.
3629 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3631 *) Add AES modes CFB and OFB to the object database. Correct an
3632 error in AES-CFB decryption.
3635 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3636 allows existing EVP_CIPHER_CTX structures to be reused after
3637 calling EVP_*Final(). This behaviour is used by encryption
3638 BIOs and some applications. This has the side effect that
3639 applications must explicitly clean up cipher contexts with
3640 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3643 *) Check the values of dna and dnb in bn_mul_recursive before calling
3644 bn_mul_comba (a non zero value means the a or b arrays do not contain
3645 n2 elements) and fallback to bn_mul_normal if either is not zero.
3648 *) Fix escaping of non-ASCII characters when using the -subj option
3649 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3652 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3653 form for "surname", serialNumber has no short form.
3654 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3655 therefore remove "mail" short name for "internet 7".
3656 The OID for unique identifiers in X509 certificates is
3657 x500UniqueIdentifier, not uniqueIdentifier.
3658 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3661 *) Add an "init" command to the ENGINE config module and auto initialize
3662 ENGINEs. Without any "init" command the ENGINE will be initialized
3663 after all ctrl commands have been executed on it. If init=1 the
3664 ENGINE is initailized at that point (ctrls before that point are run
3665 on the uninitialized ENGINE and after on the initialized one). If
3666 init=0 then the ENGINE will not be iniatialized at all.
3669 *) Fix the 'app_verify_callback' interface so that the user-defined
3670 argument is actually passed to the callback: In the
3671 SSL_CTX_set_cert_verify_callback() prototype, the callback
3672 declaration has been changed from
3675 int (*cb)(X509_STORE_CTX *,void *);
3676 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3677 i=s->ctx->app_verify_callback(&ctx)
3678 has been changed into
3679 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3681 To update applications using SSL_CTX_set_cert_verify_callback(),
3682 a dummy argument can be added to their callback functions.
3683 [D. K. Smetters <smetters@parc.xerox.com>]
3685 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3686 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3688 *) Add and OPENSSL_LOAD_CONF define which will cause
3689 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3690 This allows older applications to transparently support certain
3691 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3692 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3693 load the config file and OPENSSL_add_all_algorithms_conf() which will
3694 always load it have also been added.
3697 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3698 Adjust NIDs and EVP layer.
3699 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3701 *) Config modules support in openssl utility.
3703 Most commands now load modules from the config file,
3704 though in a few (such as version) this isn't done
3705 because it couldn't be used for anything.
3707 In the case of ca and req the config file used is
3708 the same as the utility itself: that is the -config
3709 command line option can be used to specify an
3713 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3714 use "openssl_conf" if filename is NULL use default openssl config file.
3717 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3718 config section name. Add a new flag to tolerate a missing config file
3719 and move code to CONF_modules_load_file().
3722 *) Support for crypto accelerator cards from Accelerated Encryption
3723 Processing, www.aep.ie. (Use engine 'aep')
3724 The support was copied from 0.9.6c [engine] and adapted/corrected
3725 to work with the new engine framework.
3726 [AEP Inc. and Richard Levitte]
3728 *) Support for SureWare crypto accelerator cards from Baltimore
3729 Technologies. (Use engine 'sureware')
3730 The support was copied from 0.9.6c [engine] and adapted
3731 to work with the new engine framework.
3734 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3735 make the newer ENGINE framework commands for the CHIL engine work.
3736 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3738 *) Make it possible to produce shared libraries on ReliantUNIX.
3739 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3741 *) Add the configuration target debug-linux-ppro.
3742 Make 'openssl rsa' use the general key loading routines
3743 implemented in apps.c, and make those routines able to
3744 handle the key format FORMAT_NETSCAPE and the variant
3746 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3748 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3749 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3751 *) Add -keyform to rsautl, and document -engine.
3752 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3754 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3755 BIO_R_NO_SUCH_FILE error code rather than the generic
3756 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3759 *) Add new functions
3761 ERR_peek_last_error_line
3762 ERR_peek_last_error_line_data.
3763 These are similar to
3766 ERR_peek_error_line_data,
3767 but report on the latest error recorded rather than the first one
3768 still in the error queue.
3769 [Ben Laurie, Bodo Moeller]
3771 *) default_algorithms option in ENGINE config module. This allows things
3773 default_algorithms = ALL
3774 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3777 *) Prelminary ENGINE config module.
3780 *) New experimental application configuration code.
3783 *) Change the AES code to follow the same name structure as all other
3784 symmetric ciphers, and behave the same way. Move everything to
3785 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3786 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3788 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3789 [Ben Laurie and Theo de Raadt]
3791 *) Add option to output public keys in req command.
3792 [Massimiliano Pala madwolf@openca.org]
3794 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3795 (up to about 10% better than before for P-192 and P-224).
3798 *) New functions/macros
3800 SSL_CTX_set_msg_callback(ctx, cb)
3801 SSL_CTX_set_msg_callback_arg(ctx, arg)
3802 SSL_set_msg_callback(ssl, cb)
3803 SSL_set_msg_callback_arg(ssl, arg)
3805 to request calling a callback function
3807 void cb(int write_p, int version, int content_type,
3808 const void *buf, size_t len, SSL *ssl, void *arg)
3810 whenever a protocol message has been completely received
3811 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3812 protocol version according to which the SSL library interprets
3813 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3814 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3815 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3816 specification (change_cipher_spec(20), alert(21), handshake(22)).
3817 'buf' and 'len' point to the actual message, 'ssl' to the
3818 SSL object, and 'arg' is the application-defined value set by
3819 SSL[_CTX]_set_msg_callback_arg().
3821 'openssl s_client' and 'openssl s_server' have new '-msg' options
3822 to enable a callback that displays all protocol messages.
3825 *) Change the shared library support so shared libraries are built as
3826 soon as the corresponding static library is finished, and thereby get
3827 openssl and the test programs linked against the shared library.
3828 This still only happens when the keyword "shard" has been given to
3829 the configuration scripts.
3831 NOTE: shared library support is still an experimental thing, and
3832 backward binary compatibility is still not guaranteed.
3833 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3835 *) Add support for Subject Information Access extension.
3836 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3838 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3839 additional bytes when new memory had to be allocated, not just
3840 when reusing an existing buffer.
3843 *) New command line and configuration option 'utf8' for the req command.
3844 This allows field values to be specified as UTF8 strings.
3847 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3848 runs for the former and machine-readable output for the latter.
3851 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3852 of the e-mail address in the DN (i.e., it will go into a certificate
3853 extension only). The new configuration file option 'email_in_dn = no'
3854 has the same effect.
3855 [Massimiliano Pala madwolf@openca.org]
3857 *) Change all functions with names starting with des_ to be starting
3858 with DES_ instead. Add wrappers that are compatible with libdes,
3859 but are named _ossl_old_des_*. Finally, add macros that map the
3860 des_* symbols to the corresponding _ossl_old_des_* if libdes
3861 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3862 desired, the des_* symbols will be mapped to DES_*, with one
3865 Since we provide two compatibility mappings, the user needs to
3866 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3867 compatibility is desired. The default (i.e., when that macro
3868 isn't defined) is OpenSSL 0.9.6c compatibility.
3870 There are also macros that enable and disable the support of old
3871 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3872 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3873 are defined, the default will apply: to support the old des routines.
3875 In either case, one must include openssl/des.h to get the correct
3876 definitions. Do not try to just include openssl/des_old.h, that
3879 NOTE: This is a major break of an old API into a new one. Software
3880 authors are encouraged to switch to the DES_ style functions. Some
3881 time in the future, des_old.h and the libdes compatibility functions
3882 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3883 default), and then completely removed.
3886 *) Test for certificates which contain unsupported critical extensions.
3887 If such a certificate is found during a verify operation it is
3888 rejected by default: this behaviour can be overridden by either
3889 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3890 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3891 X509_supported_extension() has also been added which returns 1 if a
3892 particular extension is supported.
3895 *) Modify the behaviour of EVP cipher functions in similar way to digests
3896 to retain compatibility with existing code.
3899 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3900 compatibility with existing code. In particular the 'ctx' parameter does
3901 not have to be to be initialized before the call to EVP_DigestInit() and
3902 it is tidied up after a call to EVP_DigestFinal(). New function
3903 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3904 EVP_MD_CTX_copy() changed to not require the destination to be
3905 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3906 requires the destination to be valid.
3908 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3909 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3912 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3913 so that complete 'Handshake' protocol structures are kept in memory
3914 instead of overwriting 'msg_type' and 'length' with 'body' data.
3917 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3918 [Massimo Santin via Richard Levitte]
3920 *) Major restructuring to the underlying ENGINE code. This includes
3921 reduction of linker bloat, separation of pure "ENGINE" manipulation
3922 (initialisation, etc) from functionality dealing with implementations
3923 of specific crypto iterfaces. This change also introduces integrated
3924 support for symmetric ciphers and digest implementations - so ENGINEs
3925 can now accelerate these by providing EVP_CIPHER and EVP_MD
3926 implementations of their own. This is detailed in crypto/engine/README
3927 as it couldn't be adequately described here. However, there are a few
3928 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3929 were changed in the original introduction of ENGINE code have now
3930 reverted back - the hooking from this code to ENGINE is now a good
3931 deal more passive and at run-time, operations deal directly with
3932 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3933 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3934 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3935 they were not being used by the framework as there is no concept of a
3936 BIGNUM_METHOD and they could not be generalised to the new
3937 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3938 ENGINE_cpy() has been removed as it cannot be consistently defined in
3942 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3945 *) Change mkdef.pl to sort symbols that get the same entry number,
3946 and make sure the automatically generated functions ERR_load_*
3947 become part of libeay.num as well.
3950 *) New function SSL_renegotiate_pending(). This returns true once
3951 renegotiation has been requested (either SSL_renegotiate() call
3952 or HelloRequest/ClientHello receveived from the peer) and becomes
3953 false once a handshake has been completed.
3954 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3955 sends a HelloRequest, but does not ensure that a handshake takes
3956 place. SSL_renegotiate_pending() is useful for checking if the
3957 client has followed the request.)
3960 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3961 By default, clients may request session resumption even during
3962 renegotiation (if session ID contexts permit); with this option,
3963 session resumption is possible only in the first handshake.
3965 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3966 more bits available for options that should not be part of
3967 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3970 *) Add some demos for certificate and certificate request creation.
3973 *) Make maximum certificate chain size accepted from the peer application
3974 settable (SSL*_get/set_max_cert_list()), as proposed by
3975 "Douglas E. Engert" <deengert@anl.gov>.
3978 *) Add support for shared libraries for Unixware-7
3979 (Boyd Lynn Gerber <gerberb@zenez.com>).
3982 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3983 be done prior to destruction. Use this to unload error strings from
3984 ENGINEs that load their own error strings. NB: This adds two new API
3985 functions to "get" and "set" this destroy handler in an ENGINE.
3988 *) Alter all existing ENGINE implementations (except "openssl" and
3989 "openbsd") to dynamically instantiate their own error strings. This
3990 makes them more flexible to be built both as statically-linked ENGINEs
3991 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3992 Also, add stub code to each that makes building them as self-contained
3993 shared-libraries easier (see README.ENGINE).
3996 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3997 implementations into applications that are completely implemented in
3998 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3999 commands that can be used to configure what shared-library to load and
4000 to control aspects of the way it is handled. Also, made an update to
4001 the README.ENGINE file that brings its information up-to-date and
4002 provides some information and instructions on the "dynamic" ENGINE
4003 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4006 *) Make it possible to unload ranges of ERR strings with a new
4007 "ERR_unload_strings" function.
4010 *) Add a copy() function to EVP_MD.
4013 *) Make EVP_MD routines take a context pointer instead of just the
4014 md_data void pointer.
4017 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4018 that the digest can only process a single chunk of data
4019 (typically because it is provided by a piece of
4020 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4021 is only going to provide a single chunk of data, and hence the
4022 framework needn't accumulate the data for oneshot drivers.
4025 *) As with "ERR", make it possible to replace the underlying "ex_data"
4026 functions. This change also alters the storage and management of global
4027 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4028 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4029 index counters. The API functions that use this state have been changed
4030 to take a "class_index" rather than pointers to the class's local STACK
4031 and counter, and there is now an API function to dynamically create new
4032 classes. This centralisation allows us to (a) plug a lot of the
4033 thread-safety problems that existed, and (b) makes it possible to clean
4034 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4035 such data would previously have always leaked in application code and
4036 workarounds were in place to make the memory debugging turn a blind eye
4037 to it. Application code that doesn't use this new function will still
4038 leak as before, but their memory debugging output will announce it now
4039 rather than letting it slide.
4041 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4042 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4043 has a return value to indicate success or failure.
4046 *) Make it possible to replace the underlying "ERR" functions such that the
4047 global state (2 LHASH tables and 2 locks) is only used by the "default"
4048 implementation. This change also adds two functions to "get" and "set"
4049 the implementation prior to it being automatically set the first time
4050 any other ERR function takes place. Ie. an application can call "get",
4051 pass the return value to a module it has just loaded, and that module
4052 can call its own "set" function using that value. This means the
4053 module's "ERR" operations will use (and modify) the error state in the
4054 application and not in its own statically linked copy of OpenSSL code.
4057 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4058 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4059 the operation, and provides a more encapsulated way for external code
4060 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4061 to use these functions rather than manually incrementing the counts.
4063 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4066 *) Add EVP test program.
4069 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4072 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4073 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4074 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4075 These allow a CRL to be built without having to access X509_CRL fields
4076 directly. Modify 'ca' application to use new functions.
4079 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4080 bug workarounds. Rollback attack detection is a security feature.
4081 The problem will only arise on OpenSSL servers when TLSv1 is not
4082 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4083 Software authors not wanting to support TLSv1 will have special reasons
4084 for their choice and can explicitly enable this option.
4085 [Bodo Moeller, Lutz Jaenicke]
4087 *) Rationalise EVP so it can be extended: don't include a union of
4088 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4089 (similar to those existing for EVP_CIPHER_CTX).
4094 EVP_MD_CTX_init(&md); /* new function call */
4095 EVP_DigestInit(&md, EVP_sha1());
4096 EVP_DigestUpdate(&md, in, len);
4097 EVP_DigestFinal(&md, out, NULL);
4098 EVP_MD_CTX_cleanup(&md); /* new function call */
4102 *) Make DES key schedule conform to the usual scheme, as well as
4103 correcting its structure. This means that calls to DES functions
4104 now have to pass a pointer to a des_key_schedule instead of a
4105 plain des_key_schedule (which was actually always a pointer
4108 des_key_schedule ks;
4110 des_set_key_checked(..., &ks);
4111 des_ncbc_encrypt(..., &ks, ...);
4113 (Note that a later change renames 'des_...' into 'DES_...'.)
4116 *) Initial reduction of linker bloat: the use of some functions, such as
4117 PEM causes large amounts of unused functions to be linked in due to
4118 poor organisation. For example pem_all.c contains every PEM function
4119 which has a knock on effect of linking in large amounts of (unused)
4120 ASN1 code. Grouping together similar functions and splitting unrelated
4121 functions prevents this.
4124 *) Cleanup of EVP macros.
4127 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4128 correct _ecb suffix.
4131 *) Add initial OCSP responder support to ocsp application. The
4132 revocation information is handled using the text based index
4133 use by the ca application. The responder can either handle
4134 requests generated internally, supplied in files (for example
4135 via a CGI script) or using an internal minimal server.
4138 *) Add configuration choices to get zlib compression for TLS.
4141 *) Changes to Kerberos SSL for RFC 2712 compliance:
4142 1. Implemented real KerberosWrapper, instead of just using
4143 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4144 2. Implemented optional authenticator field of KerberosWrapper.
4146 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4147 and authenticator structs; see crypto/krb5/.
4149 Generalized Kerberos calls to support multiple Kerberos libraries.
4150 [Vern Staats <staatsvr@asc.hpc.mil>,
4151 Jeffrey Altman <jaltman@columbia.edu>
4152 via Richard Levitte]
4154 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4155 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4156 values for each of the key sizes rather than having just
4157 parameters (and 'speed' generating keys each time).
4160 *) Speed up EVP routines.
4163 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4164 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4165 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4166 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4168 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4169 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4170 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4173 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4175 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4178 *) Added the OS2-EMX target.
4179 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4181 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4182 to support NCONF routines in extension code. New function CONF_set_nconf()
4183 to allow functions which take an NCONF to also handle the old LHASH
4184 structure: this means that the old CONF compatible routines can be
4185 retained (in particular wrt extensions) without having to duplicate the
4186 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4189 *) Enhance the general user interface with mechanisms for inner control
4190 and with possibilities to have yes/no kind of prompts.
4193 *) Change all calls to low level digest routines in the library and
4194 applications to use EVP. Add missing calls to HMAC_cleanup() and
4195 don't assume HMAC_CTX can be copied using memcpy().
4196 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4198 *) Add the possibility to control engines through control names but with
4199 arbitrary arguments instead of just a string.
4200 Change the key loaders to take a UI_METHOD instead of a callback
4201 function pointer. NOTE: this breaks binary compatibility with earlier
4202 versions of OpenSSL [engine].
4203 Adapt the nCipher code for these new conditions and add a card insertion
4207 *) Enhance the general user interface with mechanisms to better support
4208 dialog box interfaces, application-defined prompts, the possibility
4209 to use defaults (for example default passwords from somewhere else)
4210 and interrupts/cancellations.
4213 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4214 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4217 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4218 tidy up some unnecessarily weird code in 'sk_new()').
4219 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4221 *) Change the key loading routines for ENGINEs to use the same kind
4222 callback (pem_password_cb) as all other routines that need this
4226 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4227 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4228 than this minimum value is recommended.
4231 *) New random seeder for OpenVMS, using the system process statistics
4232 that are easily reachable.
4235 *) Windows apparently can't transparently handle global
4236 variables defined in DLLs. Initialisations such as:
4238 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4240 wont compile. This is used by the any applications that need to
4241 declare their own ASN1 modules. This was fixed by adding the option
4242 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4243 needed for static libraries under Win32.
4246 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4247 setting of purpose and trust fields. New X509_STORE trust and
4248 purpose functions and tidy up setting in other SSL functions.
4251 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4252 structure. These are inherited by X509_STORE_CTX when it is
4253 initialised. This allows various defaults to be set in the
4254 X509_STORE structure (such as flags for CRL checking and custom
4255 purpose or trust settings) for functions which only use X509_STORE_CTX
4256 internally such as S/MIME.
4258 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4259 trust settings if they are not set in X509_STORE. This allows X509_STORE
4260 purposes and trust (in S/MIME for example) to override any set by default.
4262 Add command line options for CRL checking to smime, s_client and s_server
4266 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4267 are set then the CRL is looked up in the X509_STORE structure and
4268 its validity and signature checked, then if the certificate is found
4269 in the CRL the verify fails with a revoked error.
4271 Various new CRL related callbacks added to X509_STORE_CTX structure.
4273 Command line options added to 'verify' application to support this.
4275 This needs some additional work, such as being able to handle multiple
4276 CRLs with different times, extension based lookup (rather than just
4277 by subject name) and ultimately more complete V2 CRL extension
4281 *) Add a general user interface API (crypto/ui/). This is designed
4282 to replace things like des_read_password and friends (backward
4283 compatibility functions using this new API are provided).
4284 The purpose is to remove prompting functions from the DES code
4285 section as well as provide for prompting through dialog boxes in
4286 a window system and the like.
4289 *) Add "ex_data" support to ENGINE so implementations can add state at a
4290 per-structure level rather than having to store it globally.
4293 *) Make it possible for ENGINE structures to be copied when retrieved by
4294 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4295 This causes the "original" ENGINE structure to act like a template,
4296 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4297 operational state can be localised to each ENGINE structure, despite the
4298 fact they all share the same "methods". New ENGINE structures returned in
4299 this case have no functional references and the return value is the single
4300 structural reference. This matches the single structural reference returned
4301 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4305 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4306 needs to match any other type at all we need to manually clear the
4310 *) Changes to the "openssl engine" utility to include;
4311 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4312 about an ENGINE's available control commands.
4313 - executing control commands from command line arguments using the
4314 '-pre' and '-post' switches. '-post' is only used if '-t' is
4315 specified and the ENGINE is successfully initialised. The syntax for
4316 the individual commands are colon-separated, for example;
4317 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4320 *) New dynamic control command support for ENGINEs. ENGINEs can now
4321 declare their own commands (numbers), names (strings), descriptions,
4322 and input types for run-time discovery by calling applications. A
4323 subset of these commands are implicitly classed as "executable"
4324 depending on their input type, and only these can be invoked through
4325 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4326 can be based on user input, config files, etc). The distinction is
4327 that "executable" commands cannot return anything other than a boolean
4328 result and can only support numeric or string input, whereas some
4329 discoverable commands may only be for direct use through
4330 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4331 pointers, or other custom uses. The "executable" commands are to
4332 support parameterisations of ENGINE behaviour that can be
4333 unambiguously defined by ENGINEs and used consistently across any
4334 OpenSSL-based application. Commands have been added to all the
4335 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4336 control over shared-library paths without source code alterations.
4339 *) Changed all ENGINE implementations to dynamically allocate their
4340 ENGINEs rather than declaring them statically. Apart from this being
4341 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4342 this also allows the implementations to compile without using the
4343 internal engine_int.h header.
4346 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4347 'const' value. Any code that should be able to modify a RAND_METHOD
4348 should already have non-const pointers to it (ie. they should only
4349 modify their own ones).
4352 *) Made a variety of little tweaks to the ENGINE code.
4353 - "atalla" and "ubsec" string definitions were moved from header files
4354 to C code. "nuron" string definitions were placed in variables
4355 rather than hard-coded - allowing parameterisation of these values
4356 later on via ctrl() commands.
4357 - Removed unused "#if 0"'d code.
4358 - Fixed engine list iteration code so it uses ENGINE_free() to release
4359 structural references.
4360 - Constified the RAND_METHOD element of ENGINE structures.
4361 - Constified various get/set functions as appropriate and added
4362 missing functions (including a catch-all ENGINE_cpy that duplicates
4363 all ENGINE values onto a new ENGINE except reference counts/state).
4364 - Removed NULL parameter checks in get/set functions. Setting a method
4365 or function to NULL is a way of cancelling out a previously set
4366 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4367 and doesn't justify the extra error symbols and code.
4368 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4369 flags from engine_int.h to engine.h.
4370 - Changed prototypes for ENGINE handler functions (init(), finish(),
4371 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4374 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4375 to the algorithm using long division. The binary algorithm can be
4376 used only if the modulus is odd. On 32-bit systems, it is faster
4377 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4378 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4379 up to 450 bits. In 64-bit environments, the binary algorithm
4380 appears to be advantageous for much longer moduli; here we use it
4381 for moduli up to 2048 bits.
4384 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4385 could not support the combine flag in choice fields.
4388 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4389 extensions from a certificate request to the certificate.
4392 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4393 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4394 file: this allows the display of the certificate about to be
4395 signed to be customised, to allow certain fields to be included
4396 or excluded and extension details. The old system didn't display
4397 multicharacter strings properly, omitted fields not in the policy
4398 and couldn't display additional details such as extensions.
4401 *) Function EC_POINTs_mul for multiple scalar multiplication
4402 of an arbitrary number of elliptic curve points
4403 \sum scalars[i]*points[i],
4404 optionally including the generator defined for the EC_GROUP:
4405 scalar*generator + \sum scalars[i]*points[i].
4407 EC_POINT_mul is a simple wrapper function for the typical case
4408 that the point list has just one item (besides the optional
4412 *) First EC_METHODs for curves over GF(p):
4414 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4415 operations and provides various method functions that can also
4416 operate with faster implementations of modular arithmetic.
4418 EC_GFp_mont_method() reuses most functions that are part of
4419 EC_GFp_simple_method, but uses Montgomery arithmetic.
4421 [Bodo Moeller; point addition and point doubling
4422 implementation directly derived from source code provided by
4423 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4425 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4426 crypto/ec/ec_lib.c):
4428 Curves are EC_GROUP objects (with an optional group generator)
4429 based on EC_METHODs that are built into the library.
4431 Points are EC_POINT objects based on EC_GROUP objects.
4433 Most of the framework would be able to handle curves over arbitrary
4434 finite fields, but as there are no obvious types for fields other
4435 than GF(p), some functions are limited to that for now.
4438 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4439 that the file contains a complete HTTP response.
4442 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4443 change the def and num file printf format specifier from "%-40sXXX"
4444 to "%-39s XXX". The latter will always guarantee a space after the
4445 field while the former will cause them to run together if the field
4446 is 40 of more characters long.
4449 *) Constify the cipher and digest 'method' functions and structures
4450 and modify related functions to take constant EVP_MD and EVP_CIPHER
4454 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4455 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4458 *) Modify EVP_Digest*() routines so they now return values. Although the
4459 internal software routines can never fail additional hardware versions
4463 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4465 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4466 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4469 ERR_R_NESTED_ASN1_ERROR
4471 ERR_R_MISSING_ASN1_EOS
4472 were 4 .. 9, conflicting with
4473 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4475 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4476 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4478 Add new error code 'ERR_R_INTERNAL_ERROR'.
4481 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4485 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4486 sets the subject name for a new request or supersedes the
4487 subject name in a given request. Formats that can be parsed are
4488 'CN=Some Name, OU=myOU, C=IT'
4490 'CN=Some Name/OU=myOU/C=IT'.
4492 Add options '-batch' and '-verbose' to 'openssl req'.
4493 [Massimiliano Pala <madwolf@hackmasters.net>]
4495 *) Introduce the possibility to access global variables through
4496 functions on platform were that's the best way to handle exporting
4497 global variables in shared libraries. To enable this functionality,
4498 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4499 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4500 is normally done by Configure or something similar).
4502 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4503 in the source file (foo.c) like this:
4505 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4506 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4508 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4509 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4511 OPENSSL_DECLARE_GLOBAL(int,foo);
4512 #define foo OPENSSL_GLOBAL_REF(foo)
4513 OPENSSL_DECLARE_GLOBAL(double,bar);
4514 #define bar OPENSSL_GLOBAL_REF(bar)
4516 The #defines are very important, and therefore so is including the
4517 header file everywhere where the defined globals are used.
4519 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4520 of ASN.1 items, but that structure is a bit different.
4522 The largest change is in util/mkdef.pl which has been enhanced with
4523 better and easier to understand logic to choose which symbols should
4524 go into the Windows .def files as well as a number of fixes and code
4525 cleanup (among others, algorithm keywords are now sorted
4526 lexicographically to avoid constant rewrites).
4529 *) In BN_div() keep a copy of the sign of 'num' before writing the
4530 result to 'rm' because if rm==num the value will be overwritten
4531 and produce the wrong result if 'num' is negative: this caused
4532 problems with BN_mod() and BN_nnmod().
4535 *) Function OCSP_request_verify(). This checks the signature on an
4536 OCSP request and verifies the signer certificate. The signer
4537 certificate is just checked for a generic purpose and OCSP request
4541 *) Add OCSP_check_validity() function to check the validity of OCSP
4542 responses. OCSP responses are prepared in real time and may only
4543 be a few seconds old. Simply checking that the current time lies
4544 between thisUpdate and nextUpdate max reject otherwise valid responses
4545 caused by either OCSP responder or client clock inaccuracy. Instead
4546 we allow thisUpdate and nextUpdate to fall within a certain period of
4547 the current time. The age of the response can also optionally be
4548 checked. Two new options -validity_period and -status_age added to
4552 *) If signature or public key algorithm is unrecognized print out its
4553 OID rather that just UNKNOWN.
4556 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4557 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4558 ID to be generated from the issuer certificate alone which can then be
4559 passed to OCSP_id_issuer_cmp().
4562 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4563 ASN1 modules to export functions returning ASN1_ITEM pointers
4564 instead of the ASN1_ITEM structures themselves. This adds several
4565 new macros which allow the underlying ASN1 function/structure to
4566 be accessed transparently. As a result code should not use ASN1_ITEM
4567 references directly (such as &X509_it) but instead use the relevant
4568 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4569 use of the new ASN1 code on platforms where exporting structures
4570 is problematical (for example in shared libraries) but exporting
4571 functions returning pointers to structures is not.
4574 *) Add support for overriding the generation of SSL/TLS session IDs.
4575 These callbacks can be registered either in an SSL_CTX or per SSL.
4576 The purpose of this is to allow applications to control, if they wish,
4577 the arbitrary values chosen for use as session IDs, particularly as it
4578 can be useful for session caching in multiple-server environments. A
4579 command-line switch for testing this (and any client code that wishes
4580 to use such a feature) has been added to "s_server".
4581 [Geoff Thorpe, Lutz Jaenicke]
4583 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4584 of the form '#if defined(...) || defined(...) || ...' and
4585 '#if !defined(...) && !defined(...) && ...'. This also avoids
4586 the growing number of special cases it was previously handling.
4589 *) Make all configuration macros available for application by making
4590 sure they are available in opensslconf.h, by giving them names starting
4591 with "OPENSSL_" to avoid conflicts with other packages and by making
4592 sure e_os2.h will cover all platform-specific cases together with
4594 Additionally, it is now possible to define configuration/platform-
4595 specific names (called "system identities"). In the C code, these
4596 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4597 macro with the name beginning with "OPENSSL_SYS_", which is determined
4598 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4602 *) New option -set_serial to 'req' and 'x509' this allows the serial
4603 number to use to be specified on the command line. Previously self
4604 signed certificates were hard coded with serial number 0 and the
4605 CA options of 'x509' had to use a serial number in a file which was
4609 *) New options to 'ca' utility to support V2 CRL entry extensions.
4610 Currently CRL reason, invalidity date and hold instruction are
4611 supported. Add new CRL extensions to V3 code and some new objects.
4614 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4615 disable standard block padding (aka PKCS#5 padding) in the EVP
4616 API, which was previously mandatory. This means that the data is
4617 not padded in any way and so the total length much be a multiple
4618 of the block size, otherwise an error occurs.
4621 *) Initial (incomplete) OCSP SSL support.
4624 *) New function OCSP_parse_url(). This splits up a URL into its host,
4625 port and path components: primarily to parse OCSP URLs. New -url
4626 option to ocsp utility.
4629 *) New nonce behavior. The return value of OCSP_check_nonce() now
4630 reflects the various checks performed. Applications can decide
4631 whether to tolerate certain situations such as an absent nonce
4632 in a response when one was present in a request: the ocsp application
4633 just prints out a warning. New function OCSP_add1_basic_nonce()
4634 this is to allow responders to include a nonce in a response even if
4635 the request is nonce-less.
4638 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4639 skipped when using openssl x509 multiple times on a single input file,
4640 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4643 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4644 set string type: to handle setting ASN1_TIME structures. Fix ca
4645 utility to correctly initialize revocation date of CRLs.
4648 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4649 the clients preferred ciphersuites and rather use its own preferences.
4650 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4651 Internet Explorer by ensuring unchanged hash method during stepup.
4652 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4655 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4656 to aes and add a new 'exist' option to print out symbols that don't
4660 *) Additional options to ocsp utility to allow flags to be set and
4661 additional certificates supplied.
4664 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4665 OCSP client a number of certificate to only verify the response
4669 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4670 handle the new API. Currently only ECB, CBC modes supported. Add new
4673 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4674 Encryption Standard (AES) Ciphersuites for Transport Layer
4675 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4676 not enabled by default and were not part of the "ALL" ciphersuite
4677 alias because they were not yet official; they could be
4678 explicitly requested by specifying the "AESdraft" ciphersuite
4679 group alias. In the final release of OpenSSL 0.9.7, the group
4680 alias is called "AES" and is part of "ALL".)
4681 [Ben Laurie, Steve Henson, Bodo Moeller]
4683 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4684 request to response.
4687 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4688 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4689 extract information from a certificate request. OCSP_response_create()
4690 creates a response and optionally adds a basic response structure.
4691 OCSP_basic_add1_status() adds a complete single response to a basic
4692 response and returns the OCSP_SINGLERESP structure just added (to allow
4693 extensions to be included for example). OCSP_basic_add1_cert() adds a
4694 certificate to a basic response and OCSP_basic_sign() signs a basic
4695 response with various flags. New helper functions ASN1_TIME_check()
4696 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4697 (converts ASN1_TIME to GeneralizedTime).
4700 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4701 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4702 structure from a certificate. X509_pubkey_digest() digests the public_key
4703 contents: this is used in various key identifiers.
4706 *) Make sk_sort() tolerate a NULL argument.
4707 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4709 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4710 passed by the function are trusted implicitly. If any of them signed the
4711 response then it is assumed to be valid and is not verified.
4714 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4715 to data. This was previously part of the PKCS7 ASN1 code. This
4716 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4717 [Steve Henson, reported by Kenneth R. Robinette
4718 <support@securenetterm.com>]
4720 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4721 routines: without these tracing memory leaks is very painful.
4722 Fix leaks in PKCS12 and PKCS7 routines.
4725 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4726 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4727 effectively meant GeneralizedTime would never be used. Now it
4728 is initialised to -1 but X509_time_adj() now has to check the value
4729 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4730 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4731 [Steve Henson, reported by Kenneth R. Robinette
4732 <support@securenetterm.com>]
4734 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4735 result in a zero length in the ASN1_INTEGER structure which was
4736 not consistent with the structure when d2i_ASN1_INTEGER() was used
4737 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4738 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4739 where it did not print out a minus for negative ASN1_INTEGER.
4742 *) Add summary printout to ocsp utility. The various functions which
4743 convert status values to strings have been renamed to:
4744 OCSP_response_status_str(), OCSP_cert_status_str() and
4745 OCSP_crl_reason_str() and are no longer static. New options
4746 to verify nonce values and to disable verification. OCSP response
4747 printout format cleaned up.
4750 *) Add additional OCSP certificate checks. These are those specified
4751 in RFC2560. This consists of two separate checks: the CA of the
4752 certificate being checked must either be the OCSP signer certificate
4753 or the issuer of the OCSP signer certificate. In the latter case the
4754 OCSP signer certificate must contain the OCSP signing extended key
4755 usage. This check is performed by attempting to match the OCSP
4756 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4757 in the OCSP_CERTID structures of the response.
4760 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4761 and related routines. This uses the standard OpenSSL certificate
4762 verify routines to perform initial checks (just CA validity) and
4763 to obtain the certificate chain. Then additional checks will be
4764 performed on the chain. Currently the root CA is checked to see
4765 if it is explicitly trusted for OCSP signing. This is used to set
4766 a root CA as a global signing root: that is any certificate that
4767 chains to that CA is an acceptable OCSP signing certificate.
4770 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4771 extensions from a separate configuration file.
4772 As when reading extensions from the main configuration file,
4773 the '-extensions ...' option may be used for specifying the
4775 [Massimiliano Pala <madwolf@comune.modena.it>]
4777 *) New OCSP utility. Allows OCSP requests to be generated or
4778 read. The request can be sent to a responder and the output
4779 parsed, outputed or printed in text form. Not complete yet:
4780 still needs to check the OCSP response validity.
4783 *) New subcommands for 'openssl ca':
4784 'openssl ca -status <serial>' prints the status of the cert with
4785 the given serial number (according to the index file).
4786 'openssl ca -updatedb' updates the expiry status of certificates
4788 [Massimiliano Pala <madwolf@comune.modena.it>]
4790 *) New '-newreq-nodes' command option to CA.pl. This is like
4791 '-newreq', but calls 'openssl req' with the '-nodes' option
4792 so that the resulting key is not encrypted.
4793 [Damien Miller <djm@mindrot.org>]
4795 *) New configuration for the GNU Hurd.
4796 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4798 *) Initial code to implement OCSP basic response verify. This
4799 is currently incomplete. Currently just finds the signer's
4800 certificate and verifies the signature on the response.
4803 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4804 value of OPENSSLDIR. This is available via the new '-d' option
4805 to 'openssl version', and is also included in 'openssl version -a'.
4808 *) Allowing defining memory allocation callbacks that will be given
4809 file name and line number information in additional arguments
4810 (a const char* and an int). The basic functionality remains, as
4811 well as the original possibility to just replace malloc(),
4812 realloc() and free() by functions that do not know about these
4813 additional arguments. To register and find out the current
4814 settings for extended allocation functions, the following
4815 functions are provided:
4817 CRYPTO_set_mem_ex_functions
4818 CRYPTO_set_locked_mem_ex_functions
4819 CRYPTO_get_mem_ex_functions
4820 CRYPTO_get_locked_mem_ex_functions
4822 These work the same way as CRYPTO_set_mem_functions and friends.
4823 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4824 extended allocation function is enabled.
4825 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4826 a conventional allocation function is enabled.
4827 [Richard Levitte, Bodo Moeller]
4829 *) Finish off removing the remaining LHASH function pointer casts.
4830 There should no longer be any prototype-casting required when using
4831 the LHASH abstraction, and any casts that remain are "bugs". See
4832 the callback types and macros at the head of lhash.h for details
4833 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4836 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4837 If /dev/[u]random devices are not available or do not return enough
4838 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4840 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4841 /etc/entropy will be queried once each in this sequence, quering stops
4842 when enough entropy was collected without querying more sockets.
4845 *) Change the Unix RAND_poll() variant to be able to poll several
4846 random devices, as specified by DEVRANDOM, until a sufficient amount
4847 of data has been collected. We spend at most 10 ms on each file
4848 (select timeout) and read in non-blocking mode. DEVRANDOM now
4849 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4850 (previously it was just the string "/dev/urandom"), so on typical
4851 platforms the 10 ms delay will never occur.
4852 Also separate out the Unix variant to its own file, rand_unix.c.
4853 For VMS, there's a currently-empty rand_vms.c.
4856 *) Move OCSP client related routines to ocsp_cl.c. These
4857 provide utility functions which an application needing
4858 to issue a request to an OCSP responder and analyse the
4859 response will typically need: as opposed to those which an
4860 OCSP responder itself would need which will be added later.
4862 OCSP_request_sign() signs an OCSP request with an API similar
4863 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4864 response. OCSP_response_get1_basic() extracts basic response
4865 from response. OCSP_resp_find_status(): finds and extracts status
4866 information from an OCSP_CERTID structure (which will be created
4867 when the request structure is built). These are built from lower
4868 level functions which work on OCSP_SINGLERESP structures but
4869 wont normally be used unless the application wishes to examine
4870 extensions in the OCSP response for example.
4872 Replace nonce routines with a pair of functions.
4873 OCSP_request_add1_nonce() adds a nonce value and optionally
4874 generates a random value. OCSP_check_nonce() checks the
4875 validity of the nonce in an OCSP response.
4878 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4879 This doesn't copy the supplied OCSP_CERTID and avoids the
4880 need to free up the newly created id. Change return type
4881 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4882 This can then be used to add extensions to the request.
4883 Deleted OCSP_request_new(), since most of its functionality
4884 is now in OCSP_REQUEST_new() (and the case insensitive name
4885 clash) apart from the ability to set the request name which
4886 will be added elsewhere.
4889 *) Update OCSP API. Remove obsolete extensions argument from
4890 various functions. Extensions are now handled using the new
4891 OCSP extension code. New simple OCSP HTTP function which
4892 can be used to send requests and parse the response.
4895 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4896 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4897 uses the special reorder version of SET OF to sort the attributes
4898 and reorder them to match the encoded order. This resolves a long
4899 standing problem: a verify on a PKCS7 structure just after signing
4900 it used to fail because the attribute order did not match the
4901 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4902 it uses the received order. This is necessary to tolerate some broken
4903 software that does not order SET OF. This is handled by encoding
4904 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4905 to produce the required SET OF.
4908 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4909 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4910 files to get correct declarations of the ASN.1 item variables.
4913 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4914 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4915 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4916 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4917 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4918 ASN1_ITEM and no wrapper functions.
4921 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4922 replace the old function pointer based I/O routines. Change most of
4923 the *_d2i_bio() and *_d2i_fp() functions to use these.
4926 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4927 lines, recognice more "algorithms" that can be deselected, and make
4928 it complain about algorithm deselection that isn't recognised.
4931 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4932 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4933 to use new functions. Add NO_ASN1_OLD which can be set to remove
4934 some old style ASN1 functions: this can be used to determine if old
4935 code will still work when these eventually go away.
4938 *) New extension functions for OCSP structures, these follow the
4939 same conventions as certificates and CRLs.
4942 *) New function X509V3_add1_i2d(). This automatically encodes and
4943 adds an extension. Its behaviour can be customised with various
4944 flags to append, replace or delete. Various wrappers added for
4945 certifcates and CRLs.
4948 *) Fix to avoid calling the underlying ASN1 print routine when
4949 an extension cannot be parsed. Correct a typo in the
4950 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4953 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4954 entries for variables.
4957 *) Add functionality to apps/openssl.c for detecting locking
4958 problems: As the program is single-threaded, all we have
4959 to do is register a locking callback using an array for
4960 storing which locks are currently held by the program.
4963 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4964 SSL_get_ex_data_X509_STORE_idx(), which is used in
4965 ssl_verify_cert_chain() and thus can be called at any time
4966 during TLS/SSL handshakes so that thread-safety is essential.
4967 Unfortunately, the ex_data design is not at all suited
4968 for multi-threaded use, so it probably should be abolished.
4971 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4972 [Broadcom, tweaked and integrated by Geoff Thorpe]
4974 *) Move common extension printing code to new function
4975 X509V3_print_extensions(). Reorganise OCSP print routines and
4976 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4979 *) New function X509_signature_print() to remove duplication in some
4983 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4984 set (this was treated exactly the same as SET OF previously). This
4985 is used to reorder the STACK representing the structure to match the
4986 encoding. This will be used to get round a problem where a PKCS7
4987 structure which was signed could not be verified because the STACK
4988 order did not reflect the encoded order.
4991 *) Reimplement the OCSP ASN1 module using the new code.
4994 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4995 for its ASN1 operations. The old style function pointers still exist
4996 for now but they will eventually go away.
4999 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5000 completely replaces the old ASN1 functionality with a table driven
5001 encoder and decoder which interprets an ASN1_ITEM structure describing
5002 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5003 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5004 has also been converted to the new form.
5007 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5008 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5009 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5010 for negative moduli.
5013 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5014 of not touching the result's sign bit.
5017 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5021 *) Changed the LHASH code to use prototypes for callbacks, and created
5022 macros to declare and implement thin (optionally static) functions
5023 that provide type-safety and avoid function pointer casting for the
5024 type-specific callbacks.
5027 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5029 [Veers Staats <staatsvr@asc.hpc.mil>,
5030 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5032 *) Reformat the FAQ so the different questions and answers can be divided
5033 in sections depending on the subject.
5036 *) Have the zlib compression code load ZLIB.DLL dynamically under
5040 *) New function BN_mod_sqrt for computing square roots modulo a prime
5041 (using the probabilistic Tonelli-Shanks algorithm unless
5042 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5043 be handled deterministically).
5044 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5046 *) Make BN_mod_inverse faster by explicitly handling small quotients
5047 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5048 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5051 *) New function BN_kronecker.
5054 *) Fix BN_gcd so that it works on negative inputs; the result is
5055 positive unless both parameters are zero.
5056 Previously something reasonably close to an infinite loop was
5057 possible because numbers could be growing instead of shrinking
5058 in the implementation of Euclid's algorithm.
5061 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5062 sign of the number in question.
5064 Fix BN_is_word(a,w) to work correctly for w == 0.
5066 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5067 because its test if the absolute value of 'a' equals 'w'.
5068 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5069 it exists mostly for use in the implementations of BN_is_zero(),
5070 BN_is_one(), and BN_is_word().
5073 *) New function BN_swap.
5076 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5077 the exponentiation functions are more likely to produce reasonable
5078 results on negative inputs.
5081 *) Change BN_mod_mul so that the result is always non-negative.
5082 Previously, it could be negative if one of the factors was negative;
5083 I don't think anyone really wanted that behaviour.
5086 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5087 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5088 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5089 and add new functions:
5098 BN_mod_lshift1_quick
5102 These functions always generate non-negative results.
5104 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5105 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5107 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5108 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5109 be reduced modulo m.
5110 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5113 The following entry accidentily appeared in the CHANGES file
5114 distributed with OpenSSL 0.9.7. The modifications described in
5115 it do *not* apply to OpenSSL 0.9.7.
5117 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5118 was actually never needed) and in BN_mul(). The removal in BN_mul()
5119 required a small change in bn_mul_part_recursive() and the addition
5120 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5121 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5122 bn_sub_words() and bn_add_words() except they take arrays with
5127 *) In 'openssl passwd', verify passwords read from the terminal
5128 unless the '-salt' option is used (which usually means that
5129 verification would just waste user's time since the resulting
5130 hash is going to be compared with some given password hash)
5131 or the new '-noverify' option is used.
5133 This is an incompatible change, but it does not affect
5134 non-interactive use of 'openssl passwd' (passwords on the command
5135 line, '-stdin' option, '-in ...' option) and thus should not
5139 *) Remove all references to RSAref, since there's no more need for it.
5142 *) Make DSO load along a path given through an environment variable
5143 (SHLIB_PATH) with shl_load().
5146 *) Constify the ENGINE code as a result of BIGNUM constification.
5147 Also constify the RSA code and most things related to it. In a
5148 few places, most notable in the depth of the ASN.1 code, ugly
5149 casts back to non-const were required (to be solved at a later
5153 *) Make it so the openssl application has all engines loaded by default.
5156 *) Constify the BIGNUM routines a little more.
5159 *) Add the following functions:
5161 ENGINE_load_cswift()
5163 ENGINE_load_atalla()
5165 ENGINE_load_builtin_engines()
5167 That way, an application can itself choose if external engines that
5168 are built-in in OpenSSL shall ever be used or not. The benefit is
5169 that applications won't have to be linked with libdl or other dso
5170 libraries unless it's really needed.
5172 Changed 'openssl engine' to load all engines on demand.
5173 Changed the engine header files to avoid the duplication of some
5174 declarations (they differed!).
5177 *) 'openssl engine' can now list capabilities.
5180 *) Better error reporting in 'openssl engine'.
5183 *) Never call load_dh_param(NULL) in s_server.
5186 *) Add engine application. It can currently list engines by name and
5187 identity, and test if they are actually available.
5190 *) Improve RPM specification file by forcing symbolic linking and making
5191 sure the installed documentation is also owned by root.root.
5192 [Damien Miller <djm@mindrot.org>]
5194 *) Give the OpenSSL applications more possibilities to make use of
5195 keys (public as well as private) handled by engines.
5198 *) Add OCSP code that comes from CertCo.
5201 *) Add VMS support for the Rijndael code.
5204 *) Added untested support for Nuron crypto accelerator.
5207 *) Add support for external cryptographic devices. This code was
5208 previously distributed separately as the "engine" branch.
5209 [Geoff Thorpe, Richard Levitte]
5211 *) Rework the filename-translation in the DSO code. It is now possible to
5212 have far greater control over how a "name" is turned into a filename
5213 depending on the operating environment and any oddities about the
5214 different shared library filenames on each system.
5217 *) Support threads on FreeBSD-elf in Configure.
5220 *) Fix for SHA1 assembly problem with MASM: it produces
5221 warnings about corrupt line number information when assembling
5222 with debugging information. This is caused by the overlapping
5224 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5227 NCONF_get_number() has no error checking at all. As a replacement,
5228 NCONF_get_number_e() is defined (_e for "error checking") and is
5229 promoted strongly. The old NCONF_get_number is kept around for
5230 binary backward compatibility.
5231 Make it possible for methods to load from something other than a BIO,
5232 by providing a function pointer that is given a name instead of a BIO.
5233 For example, this could be used to load configuration data from an
5237 *) Fix for non blocking accept BIOs. Added new I/O special reason
5238 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5239 with non blocking I/O was not possible because no retry code was
5240 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5244 *) Added the beginnings of Rijndael support.
5247 *) Fix for bug in DirectoryString mask setting. Add support for
5248 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5249 to allow certificate printing to more controllable, additional
5250 'certopt' option to 'x509' to allow new printing options to be
5254 *) Clean old EAY MD5 hack from e_os.h.
5257 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5259 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5260 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5261 [Joe Orton, Steve Henson]
5263 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5265 *) Fix additional bug revealed by the NISCC test suite:
5267 Stop bug triggering large recursion when presented with
5268 certain ASN.1 tags (CVE-2003-0851)
5271 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5273 *) Fix various bugs revealed by running the NISCC test suite:
5275 Stop out of bounds reads in the ASN1 code when presented with
5276 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5278 If verify callback ignores invalid public key errors don't try to check
5279 certificate signature with the NULL public key.
5283 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5284 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5288 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5289 extra data after the compression methods not only for TLS 1.0
5290 but also for SSL 3.0 (as required by the specification).
5291 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5293 *) Change X509_certificate_type() to mark the key as exported/exportable
5294 when it's 512 *bits* long, not 512 bytes.
5297 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5299 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5300 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5301 a protocol version number mismatch like a decryption error
5302 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5305 *) Turn on RSA blinding by default in the default implementation
5306 to avoid a timing attack. Applications that don't want it can call
5307 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5308 They would be ill-advised to do so in most cases.
5309 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5311 *) Change RSA blinding code so that it works when the PRNG is not
5312 seeded (in this case, the secret RSA exponent is abused as
5313 an unpredictable seed -- if it is not unpredictable, there
5314 is no point in blinding anyway). Make RSA blinding thread-safe
5315 by remembering the creator's thread ID in rsa->blinding and
5316 having all other threads use local one-time blinding factors
5317 (this requires more computation than sharing rsa->blinding, but
5318 avoids excessive locking; and if an RSA object is not shared
5319 between threads, blinding will still be very fast).
5322 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5324 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5325 via timing by performing a MAC computation even if incorrrect
5326 block cipher padding has been found. This is a countermeasure
5327 against active attacks where the attacker has to distinguish
5328 between bad padding and a MAC verification error. (CVE-2003-0078)
5330 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5331 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5332 Martin Vuagnoux (EPFL, Ilion)]
5334 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5336 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5337 memory from it's contents. This is done with a counter that will
5338 place alternating values in each byte. This can be used to solve
5339 two issues: 1) the removal of calls to memset() by highly optimizing
5340 compilers, and 2) cleansing with other values than 0, since those can
5341 be read through on certain media, for example a swap space on disk.
5344 *) Bugfix: client side session caching did not work with external caching,
5345 because the session->cipher setting was not restored when reloading
5346 from the external cache. This problem was masked, when
5347 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5348 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5351 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5352 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5353 [Zeev Lieber <zeev-l@yahoo.com>]
5355 *) Undo an undocumented change introduced in 0.9.6e which caused
5356 repeated calls to OpenSSL_add_all_ciphers() and
5357 OpenSSL_add_all_digests() to be ignored, even after calling
5361 *) Change the default configuration reader to deal with last line not
5362 being properly terminated.
5365 *) Change X509_NAME_cmp() so it applies the special rules on handling
5366 DN values that are of type PrintableString, as well as RDNs of type
5367 emailAddress where the value has the type ia5String.
5368 [stefank@valicert.com via Richard Levitte]
5370 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5371 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5372 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5373 the bitwise-OR of the two for use by the majority of applications
5374 wanting this behaviour, and update the docs. The documented
5375 behaviour and actual behaviour were inconsistent and had been
5376 changing anyway, so this is more a bug-fix than a behavioural
5378 [Geoff Thorpe, diagnosed by Nadav Har'El]
5380 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5381 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5384 *) Fix initialization code race conditions in
5385 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5386 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5387 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5388 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5389 ssl2_get_cipher_by_char(),
5390 ssl3_get_cipher_by_char().
5391 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5393 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5394 the cached sessions are flushed, as the remove_cb() might use ex_data
5395 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5396 (see [openssl.org #212]).
5397 [Geoff Thorpe, Lutz Jaenicke]
5399 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5400 length, instead of the encoding length to d2i_ASN1_OBJECT.
5403 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5405 *) [In 0.9.6g-engine release:]
5406 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5407 [Lynn Gazis <lgazis@rainbow.com>]
5409 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5411 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5412 and get fix the header length calculation.
5413 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5414 Alon Kantor <alonk@checkpoint.com> (and others),
5417 *) Use proper error handling instead of 'assertions' in buffer
5418 overflow checks added in 0.9.6e. This prevents DoS (the
5419 assertions could call abort()).
5420 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5422 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5424 *) Add various sanity checks to asn1_get_length() to reject
5425 the ASN1 length bytes if they exceed sizeof(long), will appear
5426 negative or the content length exceeds the length of the
5428 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5430 *) Fix cipher selection routines: ciphers without encryption had no flags
5431 for the cipher strength set and where therefore not handled correctly
5432 by the selection routines (PR #130).
5435 *) Fix EVP_dsa_sha macro.
5439 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5440 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5441 that was added in OpenSSL 0.9.6d.
5443 As the countermeasure turned out to be incompatible with some
5444 broken SSL implementations, the new option is part of SSL_OP_ALL.
5445 SSL_OP_ALL is usually employed when compatibility with weird SSL
5446 implementations is desired (e.g. '-bugs' option to 's_client' and
5447 's_server'), so the new option is automatically set in many
5451 *) Changes in security patch:
5453 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5454 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5455 Air Force Materiel Command, USAF, under agreement number
5458 *) Add various sanity checks to asn1_get_length() to reject
5459 the ASN1 length bytes if they exceed sizeof(long), will appear
5460 negative or the content length exceeds the length of the
5461 supplied buffer. (CVE-2002-0659)
5462 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5464 *) Assertions for various potential buffer overflows, not known to
5466 [Ben Laurie (CHATS)]
5468 *) Various temporary buffers to hold ASCII versions of integers were
5469 too small for 64 bit platforms. (CVE-2002-0655)
5470 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5472 *) Remote buffer overflow in SSL3 protocol - an attacker could
5473 supply an oversized session ID to a client. (CVE-2002-0656)
5474 [Ben Laurie (CHATS)]
5476 *) Remote buffer overflow in SSL2 protocol - an attacker could
5477 supply an oversized client master key. (CVE-2002-0656)
5478 [Ben Laurie (CHATS)]
5480 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5482 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5483 encoded as NULL) with id-dsa-with-sha1.
5484 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5486 *) Check various X509_...() return values in apps/req.c.
5487 [Nils Larsch <nla@trustcenter.de>]
5489 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5490 an end-of-file condition would erronously be flagged, when the CRLF
5491 was just at the end of a processed block. The bug was discovered when
5492 processing data through a buffering memory BIO handing the data to a
5493 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5494 <ptsekov@syntrex.com> and Nedelcho Stanev.
5497 *) Implement a countermeasure against a vulnerability recently found
5498 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5499 before application data chunks to avoid the use of known IVs
5500 with data potentially chosen by the attacker.
5503 *) Fix length checks in ssl3_get_client_hello().
5506 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5507 to prevent ssl3_read_internal() from incorrectly assuming that
5508 ssl3_read_bytes() found application data while handshake
5509 processing was enabled when in fact s->s3->in_read_app_data was
5510 merely automatically cleared during the initial handshake.
5511 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5513 *) Fix object definitions for Private and Enterprise: they were not
5514 recognized in their shortname (=lowercase) representation. Extend
5515 obj_dat.pl to issue an error when using undefined keywords instead
5516 of silently ignoring the problem (Svenning Sorensen
5517 <sss@sss.dnsalias.net>).
5520 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5521 generators, i.e. generators other than 2 and 5. (Previously, the
5522 code did not properly initialise the 'add' and 'rem' values to
5523 BN_generate_prime().)
5525 In the new general case, we do not insist that 'generator' is
5526 actually a primitive root: This requirement is rather pointless;
5527 a generator of the order-q subgroup is just as good, if not
5531 *) Map new X509 verification errors to alerts. Discovered and submitted by
5532 Tom Wu <tom@arcot.com>.
5535 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5536 returning non-zero before the data has been completely received
5537 when using non-blocking I/O.
5538 [Bodo Moeller; problem pointed out by John Hughes]
5540 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5541 [Ben Laurie, Lutz Jaenicke]
5543 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5544 Yoram Zahavi <YoramZ@gilian.com>).
5547 *) Add information about CygWin 1.3 and on, and preserve proper
5548 configuration for the versions before that.
5549 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5551 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5552 check whether we deal with a copy of a session and do not delete from
5553 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5554 <izhar@checkpoint.com>.
5557 *) Do not store session data into the internal session cache, if it
5558 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5559 flag is set). Proposed by Aslam <aslam@funk.com>.
5562 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5566 *) [In 0.9.6d-engine release:]
5567 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5568 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5570 *) Add the configuration target linux-s390x.
5571 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5573 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5574 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5575 variable as an indication that a ClientHello message has been
5576 received. As the flag value will be lost between multiple
5577 invocations of ssl3_accept when using non-blocking I/O, the
5578 function may not be aware that a handshake has actually taken
5579 place, thus preventing a new session from being added to the
5582 To avoid this problem, we now set s->new_session to 2 instead of
5583 using a local variable.
5584 [Lutz Jaenicke, Bodo Moeller]
5586 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5587 if the SSL_R_LENGTH_MISMATCH error is detected.
5588 [Geoff Thorpe, Bodo Moeller]
5590 *) New 'shared_ldflag' column in Configure platform table.
5593 *) Fix EVP_CIPHER_mode macro.
5594 ["Dan S. Camper" <dan@bti.net>]
5596 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5597 type, we must throw them away by setting rr->length to 0.
5598 [D P Chang <dpc@qualys.com>]
5600 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5602 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5603 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5604 worked incorrectly for those cases where range = 10..._2 and
5605 3*range is two bits longer than range.)
5608 *) Only add signing time to PKCS7 structures if it is not already
5612 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5613 OBJ_ld_ce should be OBJ_id_ce.
5614 Also some ip-pda OIDs in crypto/objects/objects.txt were
5615 incorrect (cf. RFC 3039).
5616 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5618 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5619 returns early because it has nothing to do.
5620 [Andy Schneider <andy.schneider@bjss.co.uk>]
5622 *) [In 0.9.6c-engine release:]
5623 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5624 [Andy Schneider <andy.schneider@bjss.co.uk>]
5626 *) [In 0.9.6c-engine release:]
5627 Add support for Cryptographic Appliance's keyserver technology.
5628 (Use engine 'keyclient')
5629 [Cryptographic Appliances and Geoff Thorpe]
5631 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5632 is called via tools/c89.sh because arguments have to be
5633 rearranged (all '-L' options must appear before the first object
5635 [Richard Shapiro <rshapiro@abinitio.com>]
5637 *) [In 0.9.6c-engine release:]
5638 Add support for Broadcom crypto accelerator cards, backported
5640 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5642 *) [In 0.9.6c-engine release:]
5643 Add support for SureWare crypto accelerator cards from
5644 Baltimore Technologies. (Use engine 'sureware')
5645 [Baltimore Technologies and Mark Cox]
5647 *) [In 0.9.6c-engine release:]
5648 Add support for crypto accelerator cards from Accelerated
5649 Encryption Processing, www.aep.ie. (Use engine 'aep')
5650 [AEP Inc. and Mark Cox]
5652 *) Add a configuration entry for gcc on UnixWare.
5653 [Gary Benson <gbenson@redhat.com>]
5655 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5656 messages are stored in a single piece (fixed-length part and
5657 variable-length part combined) and fix various bugs found on the way.
5660 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5661 instead. BIO_gethostbyname() does not know what timeouts are
5662 appropriate, so entries would stay in cache even when they have
5664 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5666 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5667 faced with a pathologically small ClientHello fragment that does
5668 not contain client_version: Instead of aborting with an error,
5669 simply choose the highest available protocol version (i.e.,
5670 TLS 1.0 unless it is disabled). In practice, ClientHello
5671 messages are never sent like this, but this change gives us
5672 strictly correct behaviour at least for TLS.
5675 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5676 never resets s->method to s->ctx->method when called from within
5677 one of the SSL handshake functions.
5678 [Bodo Moeller; problem pointed out by Niko Baric]
5680 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5681 (sent using the client's version number) if client_version is
5682 smaller than the protocol version in use. Also change
5683 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5684 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5685 the client will at least see that alert.
5688 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5692 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5693 client receives HelloRequest while in a handshake.
5694 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5696 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5697 should end in 'break', not 'goto end' which circuments various
5698 cleanups done in state SSL_ST_OK. But session related stuff
5699 must be disabled for SSL_ST_OK in the case that we just sent a
5702 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5703 before just sending a HelloRequest.
5704 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5706 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5707 reveal whether illegal block cipher padding was found or a MAC
5708 verification error occured. (Neither SSLerr() codes nor alerts
5709 are directly visible to potential attackers, but the information
5710 may leak via logfiles.)
5712 Similar changes are not required for the SSL 2.0 implementation
5713 because the number of padding bytes is sent in clear for SSL 2.0,
5714 and the extra bytes are just ignored. However ssl/s2_pkt.c
5715 failed to verify that the purported number of padding bytes is in
5719 *) Add OpenUNIX-8 support including shared libraries
5720 (Boyd Lynn Gerber <gerberb@zenez.com>).
5723 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5724 'wristwatch attack' using huge encoding parameters (cf.
5725 James H. Manger's CRYPTO 2001 paper). Note that the
5726 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5727 encoding parameters and hence was not vulnerable.
5730 *) BN_sqr() bug fix.
5731 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5733 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5734 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5735 followed by modular reduction.
5736 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5738 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5739 equivalent based on BN_pseudo_rand() instead of BN_rand().
5742 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5743 This function was broken, as the check for a new client hello message
5744 to handle SGC did not allow these large messages.
5745 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5748 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5751 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5752 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5755 *) Rework the configuration and shared library support for Tru64 Unix.
5756 The configuration part makes use of modern compiler features and
5757 still retains old compiler behavior for those that run older versions
5758 of the OS. The shared library support part includes a variant that
5759 uses the RPATH feature, and is available through the special
5760 configuration target "alpha-cc-rpath", which will never be selected
5762 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5764 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5765 with the same message size as in ssl3_get_certificate_request().
5766 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5767 messages might inadvertently be reject as too long.
5768 [Petr Lampa <lampa@fee.vutbr.cz>]
5770 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5773 *) Modified SSL library such that the verify_callback that has been set
5774 specificly for an SSL object with SSL_set_verify() is actually being
5775 used. Before the change, a verify_callback set with this function was
5776 ignored and the verify_callback() set in the SSL_CTX at the time of
5777 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5778 to allow the necessary settings.
5781 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5782 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5783 done automatically (in contradiction to the requirements of the C
5784 standard). This made problems when used from OpenSSH.
5787 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5788 dh->length and always used
5790 BN_rand_range(priv_key, dh->p).
5792 BN_rand_range() is not necessary for Diffie-Hellman, and this
5793 specific range makes Diffie-Hellman unnecessarily inefficient if
5794 dh->length (recommended exponent length) is much smaller than the
5795 length of dh->p. We could use BN_rand_range() if the order of
5796 the subgroup was stored in the DH structure, but we only have
5801 BN_rand(priv_key, l, ...)
5803 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5809 RSA_eay_public_encrypt
5810 RSA_eay_private_decrypt
5811 RSA_eay_private_encrypt (signing)
5812 RSA_eay_public_decrypt (signature verification)
5814 (default implementations for RSA_public_encrypt,
5815 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5816 always reject numbers >= n.
5819 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5820 to synchronize access to 'locking_thread'. This is necessary on
5821 systems where access to 'locking_thread' (an 'unsigned long'
5822 variable) is not atomic.
5825 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5826 *before* setting the 'crypto_lock_rand' flag. The previous code had
5827 a race condition if 0 is a valid thread ID.
5828 [Travis Vitek <vitek@roguewave.com>]
5830 *) Add support for shared libraries under Irix.
5831 [Albert Chin-A-Young <china@thewrittenword.com>]
5833 *) Add configuration option to build on Linux on both big-endian and
5835 [Ralf Baechle <ralf@uni-koblenz.de>]
5837 *) Add the possibility to create shared libraries on HP-UX.
5840 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5842 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5843 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5844 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5845 PRNG state recovery was possible based on the output of
5846 one PRNG request appropriately sized to gain knowledge on
5847 'md' followed by enough consecutive 1-byte PRNG requests
5848 to traverse all of 'state'.
5850 1. When updating 'md_local' (the current thread's copy of 'md')
5851 during PRNG output generation, hash all of the previous
5852 'md_local' value, not just the half used for PRNG output.
5854 2. Make the number of bytes from 'state' included into the hash
5855 independent from the number of PRNG bytes requested.
5857 The first measure alone would be sufficient to avoid
5858 Markku-Juhani's attack. (Actually it had never occurred
5859 to me that the half of 'md_local' used for chaining was the
5860 half from which PRNG output bytes were taken -- I had always
5861 assumed that the secret half would be used.) The second
5862 measure makes sure that additional data from 'state' is never
5863 mixed into 'md_local' in small portions; this heuristically
5864 further strengthens the PRNG.
5867 *) Fix crypto/bn/asm/mips3.s.
5870 *) When only the key is given to "enc", the IV is undefined. Print out
5871 an error message in this case.
5874 *) Handle special case when X509_NAME is empty in X509 printing routines.
5877 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5878 positive and less than q.
5881 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5882 used: it isn't thread safe and the add_lock_callback should handle
5884 [Paul Rose <Paul.Rose@bridge.com>]
5886 *) Verify that incoming data obeys the block size in
5887 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5891 [Ulf Möller, Bodo Möller]
5893 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5894 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5895 when fixing the server behaviour for backwards-compatible 'client
5896 hello' messages. (Note that the attack is impractical against
5897 SSL 3.0 and TLS 1.0 anyway because length and version checking
5898 means that the probability of guessing a valid ciphertext is
5899 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5902 Before 0.9.5, the countermeasure (hide the error by generating a
5903 random 'decryption result') did not work properly because
5904 ERR_clear_error() was missing, meaning that SSL_get_error() would
5905 detect the supposedly ignored error.
5907 Both problems are now fixed.
5910 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5911 (previously it was 1024).
5914 *) Fix for compatibility mode trust settings: ignore trust settings
5915 unless some valid trust or reject settings are present.
5918 *) Fix for blowfish EVP: its a variable length cipher.
5921 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5922 parameters in DSA public key structures and return an error in the
5923 DSA routines if parameters are absent.
5926 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5927 in the current directory if neither $RANDFILE nor $HOME was set.
5928 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5929 caused some confusion to Windows users who haven't defined $HOME.
5930 Thus RAND_file_name() is changed again: e_os.h can define a
5931 DEFAULT_HOME, which will be used if $HOME is not set.
5932 For Windows, we use "C:"; on other platforms, we still require
5933 environment variables.
5935 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5936 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5937 having multiple threads call RAND_poll() concurrently.
5940 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5941 combination of a flag and a thread ID variable.
5942 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5943 flag), *other* threads can enter ssleay_add_bytes without obeying
5944 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5945 that they do not hold after the first thread unsets add_do_not_lock).
5948 *) Change bctest again: '-x' expressions are not available in all
5952 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5954 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5955 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5957 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5958 the default extension for executables, if any. Also, make the perl
5959 scripts that use symlink() to test if it really exists and use "cp"
5960 if it doesn't. All this made OpenSSL compilable and installable in
5964 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5965 If SEQUENCE is length is indefinite just set c->slen to the total
5966 amount of data available.
5967 [Steve Henson, reported by shige@FreeBSD.org]
5968 [This change does not apply to 0.9.7.]
5970 *) Change bctest to avoid here-documents inside command substitution
5971 (workaround for FreeBSD /bin/sh bug).
5972 For compatibility with Ultrix, avoid shell functions (introduced
5973 in the bctest version that searches along $PATH).
5976 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5977 with des_encrypt() defined on some operating systems, like Solaris
5981 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5982 On the Importance of Eliminating Errors in Cryptographic
5983 Computations, J. Cryptology 14 (2001) 2, 101-119,
5984 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5987 *) MIPS assembler BIGNUM division bug fix.
5990 *) Disabled incorrect Alpha assembler code.
5993 *) Fix PKCS#7 decode routines so they correctly update the length
5994 after reading an EOC for the EXPLICIT tag.
5996 [This change does not apply to 0.9.7.]
5998 *) Fix bug in PKCS#12 key generation routines. This was triggered
5999 if a 3DES key was generated with a 0 initial byte. Include
6000 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6001 (but broken) behaviour.
6004 *) Enhance bctest to search for a working bc along $PATH and print
6006 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6008 *) Fix memory leaks in err.c: free err_data string if necessary;
6009 don't write to the wrong index in ERR_set_error_data.
6012 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6016 *) Replace rdtsc with _emit statements for VC++ version 5.
6017 [Jeremy Cooper <jeremy@baymoo.org>]
6019 *) Make it possible to reuse SSLv2 sessions.
6022 *) In copy_email() check for >= 0 as a return value for
6023 X509_NAME_get_index_by_NID() since 0 is a valid index.
6024 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6026 *) Avoid coredump with unsupported or invalid public keys by checking if
6027 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6028 PKCS7_verify() fails with non detached data.
6031 *) Don't use getenv in library functions when run as setuid/setgid.
6032 New function OPENSSL_issetugid().
6035 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6036 due to incorrect handling of multi-threading:
6038 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6040 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6042 3. Count how many times MemCheck_off() has been called so that
6043 nested use can be treated correctly. This also avoids
6044 inband-signalling in the previous code (which relied on the
6045 assumption that thread ID 0 is impossible).
6048 *) Add "-rand" option also to s_client and s_server.
6051 *) Fix CPU detection on Irix 6.x.
6052 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6053 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6055 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6058 [This change does not apply to 0.9.7.]
6060 *) Use the cached encoding of an X509_NAME structure rather than
6061 copying it. This is apparently the reason for the libsafe "errors"
6062 but the code is actually correct.
6065 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6066 Bleichenbacher's DSA attack.
6067 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6068 to be set and top=0 forces the highest bit to be set; top=-1 is new
6069 and leaves the highest bit random.
6070 [Ulf Moeller, Bodo Moeller]
6072 *) In the NCONF_...-based implementations for CONF_... queries
6073 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6074 a temporary CONF structure with the data component set to NULL
6075 (which gives segmentation faults in lh_retrieve).
6076 Instead, use NULL for the CONF pointer in CONF_get_string and
6077 CONF_get_number (which may use environment variables) and directly
6078 return NULL from CONF_get_section.
6081 *) Fix potential buffer overrun for EBCDIC.
6084 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6085 keyUsage if basicConstraints absent for a CA.
6088 *) Make SMIME_write_PKCS7() write mail header values with a format that
6089 is more generally accepted (no spaces before the semicolon), since
6090 some programs can't parse those values properly otherwise. Also make
6091 sure BIO's that break lines after each write do not create invalid
6095 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6096 macros previously used would not encode an empty SEQUENCE OF
6097 and break the signature.
6099 [This change does not apply to 0.9.7.]
6101 *) Zero the premaster secret after deriving the master secret in
6105 *) Add some EVP_add_digest_alias registrations (as found in
6106 OpenSSL_add_all_digests()) to SSL_library_init()
6107 aka OpenSSL_add_ssl_algorithms(). This provides improved
6108 compatibility with peers using X.509 certificates
6109 with unconventional AlgorithmIdentifier OIDs.
6112 *) Fix for Irix with NO_ASM.
6113 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6115 *) ./config script fixes.
6116 [Ulf Moeller, Richard Levitte]
6118 *) Fix 'openssl passwd -1'.
6121 *) Change PKCS12_key_gen_asc() so it can cope with non null
6122 terminated strings whose length is passed in the passlen
6123 parameter, for example from PEM callbacks. This was done
6124 by adding an extra length parameter to asc2uni().
6125 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6127 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6128 call failed, free the DSA structure.
6131 *) Fix to uni2asc() to cope with zero length Unicode strings.
6132 These are present in some PKCS#12 files.
6135 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6136 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6137 when writing a 32767 byte record.
6138 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6140 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6141 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6143 (RSA objects have a reference count access to which is protected
6144 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6145 so they are meant to be shared between threads.)
6146 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6147 "Reddie, Steven" <Steven.Reddie@ca.com>]
6149 *) Fix a deadlock in CRYPTO_mem_leaks().
6152 *) Use better test patterns in bntest.
6155 *) rand_win.c fix for Borland C.
6158 *) BN_rshift bugfix for n == 0.
6161 *) Add a 'bctest' script that checks for some known 'bc' bugs
6162 so that 'make test' does not abort just because 'bc' is broken.
6165 *) Store verify_result within SSL_SESSION also for client side to
6166 avoid potential security hole. (Re-used sessions on the client side
6167 always resulted in verify_result==X509_V_OK, not using the original
6168 result of the server certificate verification.)
6171 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6172 SSL3_RT_APPLICATION_DATA, return 0.
6173 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6177 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6178 releases, have been re-implemented by renaming the previous
6179 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6180 and ssl3_read_internal, respectively, and adding 'peek' parameters
6181 to them. The new ssl[23]_{read,peek} functions are calls to
6182 ssl[23]_read_internal with the 'peek' flag set appropriately.
6183 A 'peek' parameter has also been added to ssl3_read_bytes, which
6184 does the actual work for ssl3_read_internal.
6187 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6188 the method-specific "init()" handler. Also clean up ex_data after
6189 calling the method-specific "finish()" handler. Previously, this was
6190 happening the other way round.
6193 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6194 The previous value, 12, was not always sufficient for BN_mod_exp().
6197 *) Make sure that shared libraries get the internal name engine with
6198 the full version number and not just 0. This should mark the
6199 shared libraries as not backward compatible. Of course, this should
6200 be changed again when we can guarantee backward binary compatibility.
6203 *) Fix typo in get_cert_by_subject() in by_dir.c
6204 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6206 *) Rework the system to generate shared libraries:
6208 - Make note of the expected extension for the shared libraries and
6209 if there is a need for symbolic links from for example libcrypto.so.0
6210 to libcrypto.so.0.9.7. There is extended info in Configure for
6213 - Make as few rebuilds of the shared libraries as possible.
6215 - Still avoid linking the OpenSSL programs with the shared libraries.
6217 - When installing, install the shared libraries separately from the
6221 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6223 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6224 and not in SSL_clear because the latter is also used by the
6225 accept/connect functions; previously, the settings made by
6226 SSL_set_read_ahead would be lost during the handshake.
6227 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6229 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6230 Previously, it would create entries for disableed algorithms no
6234 *) Added several new manual pages for SSL_* function.
6237 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6239 *) In ssl23_get_client_hello, generate an error message when faced
6240 with an initial SSL 3.0/TLS record that is too small to contain the
6241 first two bytes of the ClientHello message, i.e. client_version.
6242 (Note that this is a pathologic case that probably has never happened
6243 in real life.) The previous approach was to use the version number
6244 from the record header as a substitute; but our protocol choice
6245 should not depend on that one because it is not authenticated
6246 by the Finished messages.
6249 *) More robust randomness gathering functions for Windows.
6250 [Jeffrey Altman <jaltman@columbia.edu>]
6252 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6253 not set then we don't setup the error code for issuer check errors
6254 to avoid possibly overwriting other errors which the callback does
6255 handle. If an application does set the flag then we assume it knows
6256 what it is doing and can handle the new informational codes
6260 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6261 a general "ANY" type, as such it should be able to decode anything
6262 including tagged types. However it didn't check the class so it would
6263 wrongly interpret tagged types in the same way as their universal
6264 counterpart and unknown types were just rejected. Changed so that the
6265 tagged and unknown types are handled in the same way as a SEQUENCE:
6266 that is the encoding is stored intact. There is also a new type
6267 "V_ASN1_OTHER" which is used when the class is not universal, in this
6268 case we have no idea what the actual type is so we just lump them all
6272 *) On VMS, stdout may very well lead to a file that is written to
6273 in a record-oriented fashion. That means that every write() will
6274 write a separate record, which will be read separately by the
6275 programs trying to read from it. This can be very confusing.
6277 The solution is to put a BIO filter in the way that will buffer
6278 text until a linefeed is reached, and then write everything a
6279 line at a time, so every record written will be an actual line,
6280 not chunks of lines and not (usually doesn't happen, but I've
6281 seen it once) several lines in one record. BIO_f_linebuffer() is
6284 Currently, it's a VMS-only method, because that's where it has
6285 been tested well enough.
6288 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6289 it can return incorrect results.
6290 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6291 but it was in 0.9.6-beta[12].)
6294 *) Disable the check for content being present when verifying detached
6295 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6296 include zero length content when signing messages.
6299 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6300 BIO_ctrl (for BIO pairs).
6303 *) Add DSO method for VMS.
6306 *) Bug fix: Montgomery multiplication could produce results with the
6310 *) Add RPM specification openssl.spec and modify it to build three
6311 packages. The default package contains applications, application
6312 documentation and run-time libraries. The devel package contains
6313 include files, static libraries and function documentation. The
6314 doc package contains the contents of the doc directory. The original
6315 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6318 *) Add a large number of documentation files for many SSL routines.
6319 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6321 *) Add a configuration entry for Sony News 4.
6322 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6324 *) Don't set the two most significant bits to one when generating a
6325 random number < q in the DSA library.
6328 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6329 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6330 the underlying transport is blocking) if a handshake took place.
6331 (The default behaviour is needed by applications such as s_client
6332 and s_server that use select() to determine when to use SSL_read;
6333 but for applications that know in advance when to expect data, it
6334 just makes things more complicated.)
6337 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6341 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6342 work better on such systems.
6343 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6345 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6346 Update PKCS12_parse() so it copies the friendlyName and the
6347 keyid to the certificates aux info.
6350 *) Fix bug in PKCS7_verify() which caused an infinite loop
6351 if there was more than one signature.
6352 [Sven Uszpelkat <su@celocom.de>]
6354 *) Major change in util/mkdef.pl to include extra information
6355 about each symbol, as well as presentig variables as well
6356 as functions. This change means that there's n more need
6357 to rebuild the .num files when some algorithms are excluded.
6360 *) Allow the verify time to be set by an application,
6361 rather than always using the current time.
6364 *) Phase 2 verify code reorganisation. The certificate
6365 verify code now looks up an issuer certificate by a
6366 number of criteria: subject name, authority key id
6367 and key usage. It also verifies self signed certificates
6368 by the same criteria. The main comparison function is
6369 X509_check_issued() which performs these checks.
6371 Lot of changes were necessary in order to support this
6372 without completely rewriting the lookup code.
6374 Authority and subject key identifier are now cached.
6376 The LHASH 'certs' is X509_STORE has now been replaced
6377 by a STACK_OF(X509_OBJECT). This is mainly because an
6378 LHASH can't store or retrieve multiple objects with
6379 the same hash value.
6381 As a result various functions (which were all internal
6382 use only) have changed to handle the new X509_STORE
6383 structure. This will break anything that messed round
6384 with X509_STORE internally.
6386 The functions X509_STORE_add_cert() now checks for an
6387 exact match, rather than just subject name.
6389 The X509_STORE API doesn't directly support the retrieval
6390 of multiple certificates matching a given criteria, however
6391 this can be worked round by performing a lookup first
6392 (which will fill the cache with candidate certificates)
6393 and then examining the cache for matches. This is probably
6394 the best we can do without throwing out X509_LOOKUP
6395 entirely (maybe later...).
6397 The X509_VERIFY_CTX structure has been enhanced considerably.
6399 All certificate lookup operations now go via a get_issuer()
6400 callback. Although this currently uses an X509_STORE it
6401 can be replaced by custom lookups. This is a simple way
6402 to bypass the X509_STORE hackery necessary to make this
6403 work and makes it possible to use more efficient techniques
6404 in future. A very simple version which uses a simple
6405 STACK for its trusted certificate store is also provided
6406 using X509_STORE_CTX_trusted_stack().
6408 The verify_cb() and verify() callbacks now have equivalents
6409 in the X509_STORE_CTX structure.
6411 X509_STORE_CTX also has a 'flags' field which can be used
6412 to customise the verify behaviour.
6415 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6416 excludes S/MIME capabilities.
6419 *) When a certificate request is read in keep a copy of the
6420 original encoding of the signed data and use it when outputing
6421 again. Signatures then use the original encoding rather than
6422 a decoded, encoded version which may cause problems if the
6423 request is improperly encoded.
6426 *) For consistency with other BIO_puts implementations, call
6427 buffer_write(b, ...) directly in buffer_puts instead of calling
6430 In BIO_puts, increment b->num_write as in BIO_write.
6431 [Peter.Sylvester@EdelWeb.fr]
6433 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6434 BN_zero, we may not return a BIGNUM with an array consisting of
6438 *) Avoid calling abort() from within the library when problems are
6439 detected, except if preprocessor symbols have been defined
6440 (such as REF_CHECK, BN_DEBUG etc.).
6443 *) New openssl application 'rsautl'. This utility can be
6444 used for low level RSA operations. DER public key
6445 BIO/fp routines also added.
6448 *) New Configure entry and patches for compiling on QNX 4.
6449 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6451 *) A demo state-machine implementation was sponsored by
6452 Nuron (http://www.nuron.com/) and is now available in
6453 demos/state_machine.
6456 *) New options added to the 'dgst' utility for signature
6457 generation and verification.
6460 *) Unrecognized PKCS#7 content types are now handled via a
6461 catch all ASN1_TYPE structure. This allows unsupported
6462 types to be stored as a "blob" and an application can
6463 encode and decode it manually.
6466 *) Fix various signed/unsigned issues to make a_strex.c
6468 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6470 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6471 length if passed a buffer. ASN1_INTEGER_to_BN failed
6472 if passed a NULL BN and its argument was negative.
6473 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6475 *) Modification to PKCS#7 encoding routines to output definite
6476 length encoding. Since currently the whole structures are in
6477 memory there's not real point in using indefinite length
6478 constructed encoding. However if OpenSSL is compiled with
6479 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6482 *) Added BIO_vprintf() and BIO_vsnprintf().
6485 *) Added more prefixes to parse for in the the strings written
6486 through a logging bio, to cover all the levels that are available
6487 through syslog. The prefixes are now:
6489 PANIC, EMERG, EMR => LOG_EMERG
6490 ALERT, ALR => LOG_ALERT
6491 CRIT, CRI => LOG_CRIT
6492 ERROR, ERR => LOG_ERR
6493 WARNING, WARN, WAR => LOG_WARNING
6494 NOTICE, NOTE, NOT => LOG_NOTICE
6495 INFO, INF => LOG_INFO
6496 DEBUG, DBG => LOG_DEBUG
6498 and as before, if none of those prefixes are present at the
6499 beginning of the string, LOG_ERR is chosen.
6501 On Win32, the LOG_* levels are mapped according to this:
6503 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6504 LOG_WARNING => EVENTLOG_WARNING_TYPE
6505 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6509 *) Made it possible to reconfigure with just the configuration
6510 argument "reconf" or "reconfigure". The command line arguments
6511 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6512 and are retrieved from there when reconfiguring.
6516 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6518 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6521 *) The obj_dat.pl script was messing up the sorting of object
6522 names. The reason was that it compared the quoted version
6523 of strings as a result "OCSP" > "OCSP Signing" because
6524 " > SPACE. Changed script to store unquoted versions of
6525 names and add quotes on output. It was also omitting some
6526 names from the lookup table if they were given a default
6527 value (that is if SN is missing it is given the same
6528 value as LN and vice versa), these are now added on the
6529 grounds that if an object has a name we should be able to
6530 look it up. Finally added warning output when duplicate
6531 short or long names are found.
6534 *) Changes needed for Tandem NSK.
6535 [Scott Uroff <scott@xypro.com>]
6537 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6538 RSA_padding_check_SSLv23(), special padding was never detected
6539 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6540 version rollback attacks was not effective.
6542 In s23_clnt.c, don't use special rollback-attack detection padding
6543 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6544 client; similarly, in s23_srvr.c, don't do the rollback check if
6545 SSL 2.0 is the only protocol enabled in the server.
6548 *) Make it possible to get hexdumps of unprintable data with 'openssl
6549 asn1parse'. By implication, the functions ASN1_parse_dump() and
6550 BIO_dump_indent() are added.
6553 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6554 these print out strings and name structures based on various
6555 flags including RFC2253 support and proper handling of
6556 multibyte characters. Added options to the 'x509' utility
6557 to allow the various flags to be set.
6560 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6561 Also change the functions X509_cmp_current_time() and
6562 X509_gmtime_adj() work with an ASN1_TIME structure,
6563 this will enable certificates using GeneralizedTime in validity
6564 dates to be checked.
6567 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6568 negative public key encodings) on by default,
6569 NO_NEG_PUBKEY_BUG can be set to disable it.
6572 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6573 content octets. An i2c_ASN1_OBJECT is unnecessary because
6574 the encoding can be trivially obtained from the structure.
6577 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6578 not read locks (CRYPTO_r_[un]lock).
6581 *) A first attempt at creating official support for shared
6582 libraries through configuration. I've kept it so the
6583 default is static libraries only, and the OpenSSL programs
6584 are always statically linked for now, but there are
6585 preparations for dynamic linking in place.
6586 This has been tested on Linux and Tru64.
6589 *) Randomness polling function for Win9x, as described in:
6590 Peter Gutmann, Software Generation of Practically Strong
6594 *) Fix so PRNG is seeded in req if using an already existing
6598 *) New options to smime application. -inform and -outform
6599 allow alternative formats for the S/MIME message including
6600 PEM and DER. The -content option allows the content to be
6601 specified separately. This should allow things like Netscape
6602 form signing output easier to verify.
6605 *) Fix the ASN1 encoding of tags using the 'long form'.
6608 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6609 STRING types. These convert content octets to and from the
6610 underlying type. The actual tag and length octets are
6611 already assumed to have been read in and checked. These
6612 are needed because all other string types have virtually
6613 identical handling apart from the tag. By having versions
6614 of the ASN1 functions that just operate on content octets
6615 IMPLICIT tagging can be handled properly. It also allows
6616 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6617 and ASN1_INTEGER are identical apart from the tag.
6620 *) Change the handling of OID objects as follows:
6622 - New object identifiers are inserted in objects.txt, following
6623 the syntax given in objects.README.
6624 - objects.pl is used to process obj_mac.num and create a new
6626 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6629 This is currently kind of a hack, and the perl code in objects.pl
6630 isn't very elegant, but it works as I intended. The simplest way
6631 to check that it worked correctly is to look in obj_dat.h and
6632 check the array nid_objs and make sure the objects haven't moved
6633 around (this is important!). Additions are OK, as well as
6634 consistent name changes.
6637 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6640 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6641 The given file adds to whatever has already been seeded into the
6642 random pool through the RANDFILE configuration file option or
6643 environment variable, or the default random state file.
6646 *) mkstack.pl now sorts each macro group into lexical order.
6647 Previously the output order depended on the order the files
6648 appeared in the directory, resulting in needless rewriting
6652 *) Patches to make OpenSSL compile under Win32 again. Mostly
6653 work arounds for the VC++ problem that it treats func() as
6654 func(void). Also stripped out the parts of mkdef.pl that
6655 added extra typesafe functions: these no longer exist.
6658 *) Reorganisation of the stack code. The macros are now all
6659 collected in safestack.h . Each macro is defined in terms of
6660 a "stack macro" of the form SKM_<name>(type, a, b). The
6661 DEBUG_SAFESTACK is now handled in terms of function casts,
6662 this has the advantage of retaining type safety without the
6663 use of additional functions. If DEBUG_SAFESTACK is not defined
6664 then the non typesafe macros are used instead. Also modified the
6665 mkstack.pl script to handle the new form. Needs testing to see
6666 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6667 the default if no major problems. Similar behaviour for ASN1_SET_OF
6668 and PKCS12_STACK_OF.
6671 *) When some versions of IIS use the 'NET' form of private key the
6672 key derivation algorithm is different. Normally MD5(password) is
6673 used as a 128 bit RC4 key. In the modified case
6674 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6675 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6676 as the old Netscape_RSA functions except they have an additional
6677 'sgckey' parameter which uses the modified algorithm. Also added
6678 an -sgckey command line option to the rsa utility. Thanks to
6679 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6680 algorithm to openssl-dev.
6683 *) The evp_local.h macros were using 'c.##kname' which resulted in
6684 invalid expansion on some systems (SCO 5.0.5 for example).
6685 Corrected to 'c.kname'.
6686 [Phillip Porch <root@theporch.com>]
6688 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6689 a STACK of email addresses from a certificate or request, these look
6690 in the subject name and the subject alternative name extensions and
6691 omit any duplicate addresses.
6694 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6695 This makes DSA verification about 2 % faster.
6698 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6699 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6700 plus overhead for 1024 bit moduli).
6701 This makes exponentiations about 0.5 % faster for 1024 bit
6702 exponents (as measured by "openssl speed rsa2048").
6705 *) Rename memory handling macros to avoid conflicts with other
6707 Malloc => OPENSSL_malloc
6708 Malloc_locked => OPENSSL_malloc_locked
6709 Realloc => OPENSSL_realloc
6710 Free => OPENSSL_free
6713 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6714 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6717 *) CygWin32 support.
6718 [John Jarvie <jjarvie@newsguy.com>]
6720 *) The type-safe stack code has been rejigged. It is now only compiled
6721 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6722 by default all type-specific stack functions are "#define"d back to
6723 standard stack functions. This results in more streamlined output
6724 but retains the type-safety checking possibilities of the original
6728 *) The STACK code has been cleaned up, and certain type declarations
6729 that didn't make a lot of sense have been brought in line. This has
6730 also involved a cleanup of sorts in safestack.h to more correctly
6731 map type-safe stack functions onto their plain stack counterparts.
6732 This work has also resulted in a variety of "const"ifications of
6733 lots of the code, especially "_cmp" operations which should normally
6734 be prototyped with "const" parameters anyway.
6737 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6738 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6739 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6740 where all of 'md' is used each time the PRNG is used, but 'state'
6741 is used only indexed by a cyclic counter. As entropy may not be
6742 well distributed from the beginning, 'md' is important as a
6743 chaining variable. However, the output function chains only half
6744 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6745 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6746 in all of 'state' being rewritten, with the new values depending
6747 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6750 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6751 the handshake is continued after ssl_verify_cert_chain();
6752 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6753 can lead to 'unexplainable' connection aborts later.
6754 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6756 *) Major EVP API cipher revision.
6757 Add hooks for extra EVP features. This allows various cipher
6758 parameters to be set in the EVP interface. Support added for variable
6759 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6760 setting of RC2 and RC5 parameters.
6762 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6765 Remove lots of duplicated code from the EVP library. For example *every*
6766 cipher init() function handles the 'iv' in the same way according to the
6767 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6768 for CFB and OFB modes they zero ctx->num.
6770 New functionality allows removal of S/MIME code RC2 hack.
6772 Most of the routines have the same form and so can be declared in terms
6775 By shifting this to the top level EVP_CipherInit() it can be removed from
6776 all individual ciphers. If the cipher wants to handle IVs or keys
6777 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6780 Change lots of functions like EVP_EncryptUpdate() to now return a
6781 value: although software versions of the algorithms cannot fail
6782 any installed hardware versions can.
6785 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6786 this option is set, tolerate broken clients that send the negotiated
6787 protocol version number instead of the requested protocol version
6791 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6792 i.e. non-zero for export ciphersuites, zero otherwise.
6793 Previous versions had this flag inverted, inconsistent with
6794 rsa_tmp_cb (..._TMP_RSA_CB).
6795 [Bodo Moeller; problem reported by Amit Chopra]
6797 *) Add missing DSA library text string. Work around for some IIS
6798 key files with invalid SEQUENCE encoding.
6801 *) Add a document (doc/standards.txt) that list all kinds of standards
6802 and so on that are implemented in OpenSSL.
6805 *) Enhance c_rehash script. Old version would mishandle certificates
6806 with the same subject name hash and wouldn't handle CRLs at all.
6807 Added -fingerprint option to crl utility, to support new c_rehash
6811 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6814 *) Fix for SSL server purpose checking. Server checking was
6815 rejecting certificates which had extended key usage present
6816 but no ssl client purpose.
6817 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6819 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6820 is a little unclear about how a blank password is handled.
6821 Since the password in encoded as a BMPString with terminating
6822 double NULL a zero length password would end up as just the
6823 double NULL. However no password at all is different and is
6824 handled differently in the PKCS#12 key generation code. NS
6825 treats a blank password as zero length. MSIE treats it as no
6826 password on export: but it will try both on import. We now do
6827 the same: PKCS12_parse() tries zero length and no password if
6828 the password is set to "" or NULL (NULL is now a valid password:
6829 it wasn't before) as does the pkcs12 application.
6832 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6833 perror when PEM_read_bio_X509_REQ fails, the error message must
6834 be obtained from the error queue.
6837 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6838 it in ERR_remove_state if appropriate, and change ERR_get_state
6839 accordingly to avoid race conditions (this is necessary because
6840 thread_hash is no longer constant once set).
6843 *) Bugfix for linux-elf makefile.one.
6846 *) RSA_get_default_method() will now cause a default
6847 RSA_METHOD to be chosen if one doesn't exist already.
6848 Previously this was only set during a call to RSA_new()
6849 or RSA_new_method(NULL) meaning it was possible for
6850 RSA_get_default_method() to return NULL.
6853 *) Added native name translation to the existing DSO code
6854 that will convert (if the flag to do so is set) filenames
6855 that are sufficiently small and have no path information
6856 into a canonical native form. Eg. "blah" converted to
6857 "libblah.so" or "blah.dll" etc.
6860 *) New function ERR_error_string_n(e, buf, len) which is like
6861 ERR_error_string(e, buf), but writes at most 'len' bytes
6862 including the 0 terminator. For ERR_error_string_n, 'buf'
6864 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6866 *) CONF library reworked to become more general. A new CONF
6867 configuration file reader "class" is implemented as well as a
6868 new functions (NCONF_*, for "New CONF") to handle it. The now
6869 old CONF_* functions are still there, but are reimplemented to
6870 work in terms of the new functions. Also, a set of functions
6871 to handle the internal storage of the configuration data is
6872 provided to make it easier to write new configuration file
6873 reader "classes" (I can definitely see something reading a
6874 configuration file in XML format, for example), called _CONF_*,
6875 or "the configuration storage API"...
6877 The new configuration file reading functions are:
6879 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6880 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6882 NCONF_default, NCONF_WIN32
6884 NCONF_dump_fp, NCONF_dump_bio
6886 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6887 NCONF_new creates a new CONF object. This works in the same way
6888 as other interfaces in OpenSSL, like the BIO interface.
6889 NCONF_dump_* dump the internal storage of the configuration file,
6890 which is useful for debugging. All other functions take the same
6891 arguments as the old CONF_* functions wth the exception of the
6892 first that must be a `CONF *' instead of a `LHASH *'.
6894 To make it easer to use the new classes with the old CONF_* functions,
6895 the function CONF_set_default_method is provided.
6898 *) Add '-tls1' option to 'openssl ciphers', which was already
6899 mentioned in the documentation but had not been implemented.
6900 (This option is not yet really useful because even the additional
6901 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6904 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6905 OpenSSL-based applications) load shared libraries and bind to
6906 them in a portable way.
6907 [Geoff Thorpe, with contributions from Richard Levitte]
6909 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6911 *) Make sure _lrotl and _lrotr are only used with MSVC.
6913 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6914 (the default implementation of RAND_status).
6916 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6917 to '-clrext' (= clear extensions), as intended and documented.
6918 [Bodo Moeller; inconsistency pointed out by Michael Attili
6921 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6922 was larger than the MD block size.
6923 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6925 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6926 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6927 using the passed key: if the passed key was a private key the result
6928 of X509_print(), for example, would be to print out all the private key
6932 *) des_quad_cksum() byte order bug fix.
6933 [Ulf Möller, using the problem description in krb4-0.9.7, where
6934 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6936 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6938 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6940 *) For easily testing in shell scripts whether some command
6941 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6942 returns with exit code 0 iff no command of the given name is available.
6943 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6944 the output goes to stdout and nothing is printed to stderr.
6945 Additional arguments are always ignored.
6947 Since for each cipher there is a command of the same name,
6948 the 'no-cipher' compilation switches can be tested this way.
6950 ('openssl no-XXX' is not able to detect pseudo-commands such
6951 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6954 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6957 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6958 is set; it will be thrown away anyway because each handshake creates
6960 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6961 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6962 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6963 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6966 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6967 'Q' and 'R' lose their special meanings (quit/renegotiate).
6968 This is part of what -quiet does; unlike -quiet, -ign_eof
6969 does not suppress any output.
6972 *) Add compatibility options to the purpose and trust code. The
6973 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6974 accepts a certificate or CA, this was the previous behaviour,
6975 with all the associated security issues.
6977 X509_TRUST_COMPAT is the old trust behaviour: only and
6978 automatically trust self signed roots in certificate store. A
6979 new trust setting X509_TRUST_DEFAULT is used to specify that
6980 a purpose has no associated trust setting and it should instead
6981 use the value in the default purpose.
6984 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6985 and fix a memory leak.
6988 *) In util/mkerr.pl (which implements 'make errors'), preserve
6989 reason strings from the previous version of the .c file, as
6990 the default to have only downcase letters (and digits) in
6991 automatically generated reasons codes is not always appropriate.
6994 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6995 using strerror. Previously, ERR_reason_error_string() returned
6996 library names as reason strings for SYSerr; but SYSerr is a special
6997 case where small numbers are errno values, not library numbers.
7000 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7001 converts DSA parameters into DH parameters. (When creating parameters,
7002 DSA_generate_parameters is used.)
7005 *) Include 'length' (recommended exponent length) in C code generated
7006 by 'openssl dhparam -C'.
7009 *) The second argument to set_label in perlasm was already being used
7010 so couldn't be used as a "file scope" flag. Moved to third argument
7014 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7015 instead of RAND_bytes for encryption IVs and salts.
7018 *) Include RAND_status() into RAND_METHOD instead of implementing
7019 it only for md_rand.c Otherwise replacing the PRNG by calling
7020 RAND_set_rand_method would be impossible.
7023 *) Don't let DSA_generate_key() enter an infinite loop if the random
7024 number generation fails.
7027 *) New 'rand' application for creating pseudo-random output.
7030 *) Added configuration support for Linux/IA64
7031 [Rolf Haberrecker <rolf@suse.de>]
7033 *) Assembler module support for Mingw32.
7036 *) Shared library support for HPUX (in shlib/).
7037 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7039 *) Shared library support for Solaris gcc.
7040 [Lutz Behnke <behnke@trustcenter.de>]
7042 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7044 *) PKCS7_encrypt() was adding text MIME headers twice because they
7045 were added manually and by SMIME_crlf_copy().
7048 *) In bntest.c don't call BN_rand with zero bits argument.
7049 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7051 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7052 case was implemented. This caused BN_div_recp() to fail occasionally.
7055 *) Add an optional second argument to the set_label() in the perl
7056 assembly language builder. If this argument exists and is set
7057 to 1 it signals that the assembler should use a symbol whose
7058 scope is the entire file, not just the current function. This
7059 is needed with MASM which uses the format label:: for this scope.
7060 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7062 *) Change the ASN1 types so they are typedefs by default. Before
7063 almost all types were #define'd to ASN1_STRING which was causing
7064 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7068 *) Change names of new functions to the new get1/get0 naming
7069 convention: After 'get1', the caller owns a reference count
7070 and has to call ..._free; 'get0' returns a pointer to some
7071 data structure without incrementing reference counters.
7072 (Some of the existing 'get' functions increment a reference
7073 counter, some don't.)
7074 Similarly, 'set1' and 'add1' functions increase reference
7075 counters or duplicate objects.
7078 *) Allow for the possibility of temp RSA key generation failure:
7079 the code used to assume it always worked and crashed on failure.
7082 *) Fix potential buffer overrun problem in BIO_printf().
7083 [Ulf Möller, using public domain code by Patrick Powell; problem
7084 pointed out by David Sacerdote <das33@cornell.edu>]
7086 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7087 RAND_egd() and RAND_status(). In the command line application,
7088 the EGD socket can be specified like a seed file using RANDFILE
7092 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7093 Some CAs (e.g. Verisign) distribute certificates in this form.
7096 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7097 list to exclude them. This means that no special compilation option
7098 is needed to use anonymous DH: it just needs to be included in the
7102 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7103 EVP_MD_type. The old functionality is available in a new macro called
7104 EVP_MD_md(). Change code that uses it and update docs.
7107 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7108 where the 'void *' argument is replaced by a function pointer argument.
7109 Previously 'void *' was abused to point to functions, which works on
7110 many platforms, but is not correct. As these functions are usually
7111 called by macros defined in OpenSSL header files, most source code
7112 should work without changes.
7115 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7116 sections with information on -D... compiler switches used for
7117 compiling the library so that applications can see them. To enable
7118 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7119 must be defined. E.g.,
7120 #define OPENSSL_ALGORITHM_DEFINES
7121 #include <openssl/opensslconf.h>
7122 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7123 [Richard Levitte, Ulf and Bodo Möller]
7125 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7129 *) Change the 'other' type in certificate aux info to a STACK_OF
7130 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7131 the required ASN1 format: arbitrary types determined by an OID.
7134 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7135 argument to 'req'. This is not because the function is newer or
7136 better than others it just uses the work 'NEW' in the certificate
7137 request header lines. Some software needs this.
7140 *) Reorganise password command line arguments: now passwords can be
7141 obtained from various sources. Delete the PEM_cb function and make
7142 it the default behaviour: i.e. if the callback is NULL and the
7143 usrdata argument is not NULL interpret it as a null terminated pass
7144 phrase. If usrdata and the callback are NULL then the pass phrase
7145 is prompted for as usual.
7148 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7149 the support is automatically enabled. The resulting binaries will
7150 autodetect the card and use it if present.
7151 [Ben Laurie and Compaq Inc.]
7153 *) Work around for Netscape hang bug. This sends certificate request
7154 and server done in one record. Since this is perfectly legal in the
7155 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7156 the bugs/SSLv3 entry for more info.
7159 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7162 *) Add -rand argument to smime and pkcs12 applications and read/write
7166 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7169 *) Add command line password options to the remaining applications.
7172 *) Bug fix for BN_div_recp() for numerators with an even number of
7176 *) More tests in bntest.c, and changed test_bn output.
7179 *) ./config recognizes MacOS X now.
7182 *) Bug fix for BN_div() when the first words of num and divsor are
7183 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7186 *) Add support for various broken PKCS#8 formats, and command line
7187 options to produce them.
7190 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7191 get temporary BIGNUMs from a BN_CTX.
7194 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7198 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7199 include a #define from the old name to the new. The original intent
7200 was that statically linked binaries could for example just call
7201 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7202 link with digests. This never worked becayse SSLeay_add_all_digests()
7203 and SSLeay_add_all_ciphers() were in the same source file so calling
7204 one would link with the other. They are now in separate source files.
7207 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7210 *) Use a less unusual form of the Miller-Rabin primality test (it used
7211 a binary algorithm for exponentiation integrated into the Miller-Rabin
7212 loop, our standard modexp algorithms are faster).
7215 *) Support for the EBCDIC character set completed.
7216 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7218 *) Source code cleanups: use const where appropriate, eliminate casts,
7219 use void * instead of char * in lhash.
7222 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7223 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7224 this the server could overwrite ephemeral keys that the client
7228 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7229 using 50 iterations of the Rabin-Miller test.
7231 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7232 iterations of the Rabin-Miller test as required by the appendix
7233 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7234 As BN_is_prime_fasttest includes trial division, DSA parameter
7235 generation becomes much faster.
7237 This implies a change for the callback functions in DSA_is_prime
7238 and DSA_generate_parameters: The callback function is called once
7239 for each positive witness in the Rabin-Miller test, not just
7240 occasionally in the inner loop; and the parameters to the
7241 callback function now provide an iteration count for the outer
7242 loop rather than for the current invocation of the inner loop.
7243 DSA_generate_parameters additionally can call the callback
7244 function with an 'iteration count' of -1, meaning that a
7245 candidate has passed the trial division test (when q is generated
7246 from an application-provided seed, trial division is skipped).
7249 *) New function BN_is_prime_fasttest that optionally does trial
7250 division before starting the Rabin-Miller test and has
7251 an additional BN_CTX * argument (whereas BN_is_prime always
7252 has to allocate at least one BN_CTX).
7253 'callback(1, -1, cb_arg)' is called when a number has passed the
7254 trial division stage.
7257 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7261 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7264 *) New function BN_pseudo_rand().
7267 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7268 bignum version of BN_from_montgomery() with the working code from
7269 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7273 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7274 made it impossible to use the same SSL_SESSION data structure in
7275 SSL2 clients in multiple threads.
7278 *) The return value of RAND_load_file() no longer counts bytes obtained
7279 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7280 to seed the PRNG (previously an explicit byte count was required).
7281 [Ulf Möller, Bodo Möller]
7283 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7284 used (char *) instead of (void *) and had casts all over the place.
7287 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7290 *) Retain source code compatibility for BN_prime_checks macro:
7291 BN_is_prime(..., BN_prime_checks, ...) now uses
7292 BN_prime_checks_for_size to determine the appropriate number of
7293 Rabin-Miller iterations.
7296 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7297 DH_CHECK_P_NOT_SAFE_PRIME.
7298 (Check if this is true? OpenPGP calls them "strong".)
7301 *) Merge the functionality of "dh" and "gendh" programs into a new program
7302 "dhparam". The old programs are retained for now but will handle DH keys
7303 (instead of parameters) in future.
7306 *) Make the ciphers, s_server and s_client programs check the return values
7307 when a new cipher list is set.
7310 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7311 ciphers. Before when the 56bit ciphers were enabled the sorting was
7314 The syntax for the cipher sorting has been extended to support sorting by
7315 cipher-strength (using the strength_bits hard coded in the tables).
7316 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7318 Fix a bug in the cipher-command parser: when supplying a cipher command
7319 string with an "undefined" symbol (neither command nor alphanumeric
7320 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7321 an error is flagged.
7323 Due to the strength-sorting extension, the code of the
7324 ssl_create_cipher_list() function was completely rearranged. I hope that
7325 the readability was also increased :-)
7326 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7328 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7329 for the first serial number and places 2 in the serial number file. This
7330 avoids problems when the root CA is created with serial number zero and
7331 the first user certificate has the same issuer name and serial number
7335 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7336 the new code. Add documentation for this stuff.
7339 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7340 X509_*() to X509at_*() on the grounds that they don't handle X509
7341 structures and behave in an analagous way to the X509v3 functions:
7342 they shouldn't be called directly but wrapper functions should be used
7345 So we also now have some wrapper functions that call the X509at functions
7346 when passed certificate requests. (TO DO: similar things can be done with
7347 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7348 things. Some of these need some d2i or i2d and print functionality
7349 because they handle more complex structures.)
7352 *) Add missing #ifndefs that caused missing symbols when building libssl
7353 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7354 NO_RSA in ssl/s2*.c.
7355 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7357 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7358 has a return value which indicates the quality of the random data
7359 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7360 error queue. New function RAND_pseudo_bytes() generates output that is
7361 guaranteed to be unique but not unpredictable. RAND_add is like
7362 RAND_seed, but takes an extra argument for an entropy estimate
7363 (RAND_seed always assumes full entropy).
7366 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7367 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7368 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7369 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7370 false-positive rate of at most 2^-80 for random input.
7373 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7376 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7377 in the 0.9.5 release), this returns the chain
7378 from an X509_CTX structure with a dup of the stack and all
7379 the X509 reference counts upped: so the stack will exist
7380 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7383 Also make SSL_SESSION_print() print out the verify return
7387 *) Add manpage for the pkcs12 command. Also change the default
7388 behaviour so MAC iteration counts are used unless the new
7389 -nomaciter option is used. This improves file security and
7390 only older versions of MSIE (4.0 for example) need it.
7393 *) Honor the no-xxx Configure options when creating .DEF files.
7396 *) Add PKCS#10 attributes to field table: challengePassword,
7397 unstructuredName and unstructuredAddress. These are taken from
7398 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7399 international characters are used.
7401 More changes to X509_ATTRIBUTE code: allow the setting of types
7402 based on strings. Remove the 'loc' parameter when adding
7403 attributes because these will be a SET OF encoding which is sorted
7407 *) Initial changes to the 'req' utility to allow request generation
7408 automation. This will allow an application to just generate a template
7409 file containing all the field values and have req construct the
7412 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7413 used all over the place including certificate requests and PKCS#7
7414 structures. They are currently handled manually where necessary with
7415 some primitive wrappers for PKCS#7. The new functions behave in a
7416 manner analogous to the X509 extension functions: they allow
7417 attributes to be looked up by NID and added.
7419 Later something similar to the X509V3 code would be desirable to
7420 automatically handle the encoding, decoding and printing of the
7421 more complex types. The string types like challengePassword can
7422 be handled by the string table functions.
7424 Also modified the multi byte string table handling. Now there is
7425 a 'global mask' which masks out certain types. The table itself
7426 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7427 is useful when for example there is only one permissible type
7428 (as in countryName) and using the mask might result in no valid
7432 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7433 SSL_get_peer_finished to allow applications to obtain the latest
7434 Finished messages sent to the peer or expected from the peer,
7435 respectively. (SSL_get_peer_finished is usually the Finished message
7436 actually received from the peer, otherwise the protocol will be aborted.)
7438 As the Finished message are message digests of the complete handshake
7439 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7440 be used for external authentication procedures when the authentication
7441 provided by SSL/TLS is not desired or is not enough.
7444 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7445 the host supports BWX extension and if Compaq C is present on the
7446 $PATH. Just exploiting of the BWX extension results in 20-30%
7447 performance kick for some algorithms, e.g. DES and RC4 to mention
7448 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7452 *) Add support for MS "fast SGC". This is arguably a violation of the
7453 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7454 weak crypto and after checking the certificate is SGC a second one
7455 with strong crypto. MS SGC stops the first handshake after receiving
7456 the server certificate message and sends a second client hello. Since
7457 a server will typically do all the time consuming operations before
7458 expecting any further messages from the client (server key exchange
7459 is the most expensive) there is little difference between the two.
7461 To get OpenSSL to support MS SGC we have to permit a second client
7462 hello message after we have sent server done. In addition we have to
7463 reset the MAC if we do get this second client hello.
7466 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7467 if a DER encoded private key is RSA or DSA traditional format. Changed
7468 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7469 format DER encoded private key. Newer code should use PKCS#8 format which
7470 has the key type encoded in the ASN1 structure. Added DER private key
7471 support to pkcs8 application.
7474 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7475 ciphersuites has been selected (as required by the SSL 3/TLS 1
7476 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7477 is set, we interpret this as a request to violate the specification
7478 (the worst that can happen is a handshake failure, and 'correct'
7479 behaviour would result in a handshake failure anyway).
7482 *) In SSL_CTX_add_session, take into account that there might be multiple
7483 SSL_SESSION structures with the same session ID (e.g. when two threads
7484 concurrently obtain them from an external cache).
7485 The internal cache can handle only one SSL_SESSION with a given ID,
7486 so if there's a conflict, we now throw out the old one to achieve
7490 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7491 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7492 some routines that use cipher OIDs: some ciphers do not have OIDs
7493 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7497 *) Simplify the trust setting structure and code. Now we just have
7498 two sequences of OIDs for trusted and rejected settings. These will
7499 typically have values the same as the extended key usage extension
7500 and any application specific purposes.
7502 The trust checking code now has a default behaviour: it will just
7503 check for an object with the same NID as the passed id. Functions can
7504 be provided to override either the default behaviour or the behaviour
7505 for a given id. SSL client, server and email already have functions
7506 in place for compatibility: they check the NID and also return "trusted"
7507 if the certificate is self signed.
7510 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7511 traditional format into an EVP_PKEY structure.
7514 *) Add a password callback function PEM_cb() which either prompts for
7515 a password if usr_data is NULL or otherwise assumes it is a null
7516 terminated password. Allow passwords to be passed on command line
7517 environment or config files in a few more utilities.
7520 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7521 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7522 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7523 Update documentation.
7526 *) Support for ASN1 "NULL" type. This could be handled before by using
7527 ASN1_TYPE but there wasn't any function that would try to read a NULL
7528 and produce an error if it couldn't. For compatibility we also have
7529 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7530 don't allocate anything because they don't need to.
7533 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7535 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7537 *) Rebuild of the memory allocation routines used by OpenSSL code and
7538 possibly others as well. The purpose is to make an interface that
7539 provide hooks so anyone can build a separate set of allocation and
7540 deallocation routines to be used by OpenSSL, for example memory
7541 pool implementations, or something else, which was previously hard
7542 since Malloc(), Realloc() and Free() were defined as macros having
7543 the values malloc, realloc and free, respectively (except for Win32
7544 compilations). The same is provided for memory debugging code.
7545 OpenSSL already comes with functionality to find memory leaks, but
7546 this gives people a chance to debug other memory problems.
7548 With these changes, a new set of functions and macros have appeared:
7550 CRYPTO_set_mem_debug_functions() [F]
7551 CRYPTO_get_mem_debug_functions() [F]
7552 CRYPTO_dbg_set_options() [F]
7553 CRYPTO_dbg_get_options() [F]
7554 CRYPTO_malloc_debug_init() [M]
7556 The memory debug functions are NULL by default, unless the library
7557 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7558 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7559 gives the standard debugging functions that come with OpenSSL) or
7560 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7561 provided by the library user) must be used. When the standard
7562 debugging functions are used, CRYPTO_dbg_set_options can be used to
7563 request additional information:
7564 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7565 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7567 Also, things like CRYPTO_set_mem_functions will always give the
7568 expected result (the new set of functions is used for allocation
7569 and deallocation) at all times, regardless of platform and compiler
7572 To finish it up, some functions that were never use in any other
7573 way than through macros have a new API and new semantic:
7576 CRYPTO_dbg_realloc()
7579 All macros of value have retained their old syntax.
7580 [Richard Levitte and Bodo Moeller]
7582 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7583 ordering of SMIMECapabilities wasn't in "strength order" and there
7584 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7588 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7589 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7590 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7592 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7593 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7594 functionality to handle multipart/signed properly) and a utility
7595 called 'smime' to call all this stuff. This is based on code I
7596 originally wrote for Celo who have kindly allowed it to be
7597 included in OpenSSL.
7600 *) Add variants des_set_key_checked and des_set_key_unchecked of
7601 des_set_key (aka des_key_sched). Global variable des_check_key
7602 decides which of these is called by des_set_key; this way
7603 des_check_key behaves as it always did, but applications and
7604 the library itself, which was buggy for des_check_key == 1,
7605 have a cleaner way to pick the version they need.
7608 *) New function PKCS12_newpass() which changes the password of a
7612 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7613 dynamic mix. In both cases the ids can be used as an index into the
7614 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7615 functions so they accept a list of the field values and the
7616 application doesn't need to directly manipulate the X509_TRUST
7620 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7624 *) Modify the way the V3 extension code looks up extensions. This now
7625 works in a similar way to the object code: we have some "standard"
7626 extensions in a static table which is searched with OBJ_bsearch()
7627 and the application can add dynamic ones if needed. The file
7628 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7629 updated whenever a new extension is added to the core code and kept
7630 in ext_nid order. There is a simple program 'tabtest.c' which checks
7631 this. New extensions are not added too often so this file can readily
7632 be maintained manually.
7634 There are two big advantages in doing things this way. The extensions
7635 can be looked up immediately and no longer need to be "added" using
7636 X509V3_add_standard_extensions(): this function now does nothing.
7637 [Side note: I get *lots* of email saying the extension code doesn't
7638 work because people forget to call this function]
7639 Also no dynamic allocation is done unless new extensions are added:
7640 so if we don't add custom extensions there is no need to call
7641 X509V3_EXT_cleanup().
7644 *) Modify enc utility's salting as follows: make salting the default. Add a
7645 magic header, so unsalted files fail gracefully instead of just decrypting
7646 to garbage. This is because not salting is a big security hole, so people
7647 should be discouraged from doing it.
7650 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7651 digest to be passed on the command line but it only used this
7652 parameter when signing a certificate. Modified so all relevant
7653 operations are affected by the digest parameter including the
7654 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7655 DSA key was used because it didn't fix the digest.
7658 *) Initial certificate chain verify code. Currently tests the untrusted
7659 certificates for consistency with the verify purpose (which is set
7660 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7662 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7663 this is because it will reject chains with invalid extensions whereas
7664 every previous version of OpenSSL and SSLeay made no checks at all.
7666 Trust code: checks the root CA for the relevant trust settings. Trust
7667 settings have an initial value consistent with the verify purpose: e.g.
7668 if the verify purpose is for SSL client use it expects the CA to be
7669 trusted for SSL client use. However the default value can be changed to
7670 permit custom trust settings: one example of this would be to only trust
7671 certificates from a specific "secure" set of CAs.
7673 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7674 which should be used for version portability: especially since the
7675 verify structure is likely to change more often now.
7677 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7678 to set them. If not set then assume SSL clients will verify SSL servers
7681 Two new options to the verify program: -untrusted allows a set of
7682 untrusted certificates to be passed in and -purpose which sets the
7683 intended purpose of the certificate. If a purpose is set then the
7684 new chain verify code is used to check extension consistency.
7687 *) Support for the authority information access extension.
7690 *) Modify RSA and DSA PEM read routines to transparently handle
7691 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7692 public keys in a format compatible with certificate
7693 SubjectPublicKeyInfo structures. Unfortunately there were already
7694 functions called *_PublicKey_* which used various odd formats so
7695 these are retained for compatibility: however the DSA variants were
7696 never in a public release so they have been deleted. Changed dsa/rsa
7697 utilities to handle the new format: note no releases ever handled public
7698 keys so we should be OK.
7700 The primary motivation for this change is to avoid the same fiasco
7701 that dogs private keys: there are several incompatible private key
7702 formats some of which are standard and some OpenSSL specific and
7703 require various evil hacks to allow partial transparent handling and
7704 even then it doesn't work with DER formats. Given the option anything
7705 other than PKCS#8 should be dumped: but the other formats have to
7706 stay in the name of compatibility.
7708 With public keys and the benefit of hindsight one standard format
7709 is used which works with EVP_PKEY, RSA or DSA structures: though
7710 it clearly returns an error if you try to read the wrong kind of key.
7712 Added a -pubkey option to the 'x509' utility to output the public key.
7713 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7714 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7715 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7716 that do the same as the EVP_PKEY_assign_*() except they up the
7717 reference count of the added key (they don't "swallow" the
7721 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7722 CRLs would fail if the file contained no certificates or no CRLs:
7723 added a new function to read in both types and return the number
7724 read: this means that if none are read it will be an error. The
7725 DER versions of the certificate and CRL reader would always fail
7726 because it isn't possible to mix certificates and CRLs in DER format
7727 without choking one or the other routine. Changed this to just read
7728 a certificate: this is the best we can do. Also modified the code
7729 in apps/verify.c to take notice of return codes: it was previously
7730 attempting to read in certificates from NULL pointers and ignoring
7731 any errors: this is one reason why the cert and CRL reader seemed
7732 to work. It doesn't check return codes from the default certificate
7733 routines: these may well fail if the certificates aren't installed.
7736 *) Code to support otherName option in GeneralName.
7739 *) First update to verify code. Change the verify utility
7740 so it warns if it is passed a self signed certificate:
7741 for consistency with the normal behaviour. X509_verify
7742 has been modified to it will now verify a self signed
7743 certificate if *exactly* the same certificate appears
7744 in the store: it was previously impossible to trust a
7745 single self signed certificate. This means that:
7746 openssl verify ss.pem
7747 now gives a warning about a self signed certificate but
7748 openssl verify -CAfile ss.pem ss.pem
7752 *) For servers, store verify_result in SSL_SESSION data structure
7753 (and add it to external session representation).
7754 This is needed when client certificate verifications fails,
7755 but an application-provided verification callback (set by
7756 SSL_CTX_set_cert_verify_callback) allows accepting the session
7757 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7758 but returns 1): When the session is reused, we have to set
7759 ssl->verify_result to the appropriate error code to avoid
7761 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7763 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7764 case in PKCS7_dataInit() where the signed PKCS7 structure
7765 didn't contain any existing data because it was being created.
7766 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7768 *) Add a salt to the key derivation routines in enc.c. This
7769 forms the first 8 bytes of the encrypted file. Also add a
7770 -S option to allow a salt to be input on the command line.
7773 *) New function X509_cmp(). Oddly enough there wasn't a function
7774 to compare two certificates. We do this by working out the SHA1
7775 hash and comparing that. X509_cmp() will be needed by the trust
7779 *) SSL_get1_session() is like SSL_get_session(), but increments
7780 the reference count in the SSL_SESSION returned.
7781 [Geoff Thorpe <geoff@eu.c2.net>]
7783 *) Fix for 'req': it was adding a null to request attributes.
7784 Also change the X509_LOOKUP and X509_INFO code to handle
7785 certificate auxiliary information.
7788 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7792 *) Add the possibility to add extra information to the memory leak
7793 detecting output, to form tracebacks, showing from where each
7794 allocation was originated: CRYPTO_push_info("constant string") adds
7795 the string plus current file name and line number to a per-thread
7796 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7797 is like calling CYRPTO_pop_info() until the stack is empty.
7798 Also updated memory leak detection code to be multi-thread-safe.
7801 *) Add options -text and -noout to pkcs7 utility and delete the
7802 encryption options which never did anything. Update docs.
7805 *) Add options to some of the utilities to allow the pass phrase
7806 to be included on either the command line (not recommended on
7807 OSes like Unix) or read from the environment. Update the
7808 manpages and fix a few bugs.
7811 *) Add a few manpages for some of the openssl commands.
7814 *) Fix the -revoke option in ca. It was freeing up memory twice,
7815 leaking and not finding already revoked certificates.
7818 *) Extensive changes to support certificate auxiliary information.
7819 This involves the use of X509_CERT_AUX structure and X509_AUX
7820 functions. An X509_AUX function such as PEM_read_X509_AUX()
7821 can still read in a certificate file in the usual way but it
7822 will also read in any additional "auxiliary information". By
7823 doing things this way a fair degree of compatibility can be
7824 retained: existing certificates can have this information added
7825 using the new 'x509' options.
7827 Current auxiliary information includes an "alias" and some trust
7828 settings. The trust settings will ultimately be used in enhanced
7829 certificate chain verification routines: currently a certificate
7830 can only be trusted if it is self signed and then it is trusted
7834 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7835 The problem was that one of the replacement routines had not been working
7836 since SSLeay releases. For now the offending routine has been replaced
7837 with non-optimised assembler. Even so, this now gives around 95%
7838 performance improvement for 1024 bit RSA signs.
7841 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7842 handling. Most clients have the effective key size in bits equal to
7843 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7844 A few however don't do this and instead use the size of the decrypted key
7845 to determine the RC2 key length and the AlgorithmIdentifier to determine
7846 the effective key length. In this case the effective key length can still
7847 be 40 bits but the key length can be 168 bits for example. This is fixed
7848 by manually forcing an RC2 key into the EVP_PKEY structure because the
7849 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7850 the key length and effective key length are equal.
7853 *) Add a bunch of functions that should simplify the creation of
7854 X509_NAME structures. Now you should be able to do:
7855 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7856 and have it automatically work out the correct field type and fill in
7857 the structures. The more adventurous can try:
7858 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7859 and it will (hopefully) work out the correct multibyte encoding.
7862 *) Change the 'req' utility to use the new field handling and multibyte
7863 copy routines. Before the DN field creation was handled in an ad hoc
7864 way in req, ca, and x509 which was rather broken and didn't support
7865 BMPStrings or UTF8Strings. Since some software doesn't implement
7866 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7867 using the dirstring_type option. See the new comment in the default
7868 openssl.cnf for more info.
7871 *) Make crypto/rand/md_rand.c more robust:
7872 - Assure unique random numbers after fork().
7873 - Make sure that concurrent threads access the global counter and
7874 md serializably so that we never lose entropy in them
7875 or use exactly the same state in multiple threads.
7876 Access to the large state is not always serializable because
7877 the additional locking could be a performance killer, and
7878 md should be large enough anyway.
7881 *) New file apps/app_rand.c with commonly needed functionality
7882 for handling the random seed file.
7884 Use the random seed file in some applications that previously did not:
7886 dsaparam -genkey (which also ignored its '-rand' option),
7889 x509 (when signing).
7890 Except on systems with /dev/urandom, it is crucial to have a random
7891 seed file at least for key creation, DSA signing, and for DH exchanges;
7892 for RSA signatures we could do without one.
7894 gendh and gendsa (unlike genrsa) used to read only the first byte
7895 of each file listed in the '-rand' option. The function as previously
7896 found in genrsa is now in app_rand.c and is used by all programs
7897 that support '-rand'.
7900 *) In RAND_write_file, use mode 0600 for creating files;
7901 don't just chmod when it may be too late.
7904 *) Report an error from X509_STORE_load_locations
7905 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7908 *) New function ASN1_mbstring_copy() this copies a string in either
7909 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7910 into an ASN1_STRING type. A mask of permissible types is passed
7911 and it chooses the "minimal" type to use or an error if not type
7915 *) Add function equivalents to the various macros in asn1.h. The old
7916 macros are retained with an M_ prefix. Code inside the library can
7917 use the M_ macros. External code (including the openssl utility)
7918 should *NOT* in order to be "shared library friendly".
7921 *) Add various functions that can check a certificate's extensions
7922 to see if it usable for various purposes such as SSL client,
7923 server or S/MIME and CAs of these types. This is currently
7924 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7925 verification. Also added a -purpose flag to x509 utility to
7926 print out all the purposes.
7929 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7933 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7934 for, obtain and decode and extension and obtain its critical flag.
7935 This allows all the necessary extension code to be handled in a
7936 single function call.
7939 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7940 platforms. See crypto/rc4/rc4_enc.c for further details.
7943 *) New -noout option to asn1parse. This causes no output to be produced
7944 its main use is when combined with -strparse and -out to extract data
7945 from a file (which may not be in ASN.1 format).
7948 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7949 when producing the local key id.
7950 [Richard Levitte <levitte@stacken.kth.se>]
7952 *) New option -dhparam in s_server. This allows a DH parameter file to be
7953 stated explicitly. If it is not stated then it tries the first server
7954 certificate file. The previous behaviour hard coded the filename
7958 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7959 a public key to be input or output. For example:
7960 openssl rsa -in key.pem -pubout -out pubkey.pem
7961 Also added necessary DSA public key functions to handle this.
7964 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7965 in the message. This was handled by allowing
7966 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7967 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7969 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7970 to the end of the strings whereas this didn't. This would cause problems
7971 if strings read with d2i_ASN1_bytes() were later modified.
7972 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7974 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7975 data and it contains EOF it will end up returning an error. This is
7976 caused by input 46 bytes long. The cause is due to the way base64
7977 BIOs find the start of base64 encoded data. They do this by trying a
7978 trial decode on each line until they find one that works. When they
7979 do a flag is set and it starts again knowing it can pass all the
7980 data directly through the decoder. Unfortunately it doesn't reset
7981 the context it uses. This means that if EOF is reached an attempt
7982 is made to pass two EOFs through the context and this causes the
7983 resulting error. This can also cause other problems as well. As is
7984 usual with these problems it takes *ages* to find and the fix is
7985 trivial: move one line.
7986 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7988 *) Ugly workaround to get s_client and s_server working under Windows. The
7989 old code wouldn't work because it needed to select() on sockets and the
7990 tty (for keypresses and to see if data could be written). Win32 only
7991 supports select() on sockets so we select() with a 1s timeout on the
7992 sockets and then see if any characters are waiting to be read, if none
7993 are present then we retry, we also assume we can always write data to
7994 the tty. This isn't nice because the code then blocks until we've
7995 received a complete line of data and it is effectively polling the
7996 keyboard at 1s intervals: however it's quite a bit better than not
7997 working at all :-) A dedicated Windows application might handle this
7998 with an event loop for example.
8001 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8002 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8003 will be called when RSA_sign() and RSA_verify() are used. This is useful
8004 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8005 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8006 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8007 This necessitated the support of an extra signature type NID_md5_sha1
8008 for SSL signatures and modifications to the SSL library to use it instead
8009 of calling RSA_public_decrypt() and RSA_private_encrypt().
8012 *) Add new -verify -CAfile and -CApath options to the crl program, these
8013 will lookup a CRL issuers certificate and verify the signature in a
8014 similar way to the verify program. Tidy up the crl program so it
8015 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8016 less strict. It will now permit CRL extensions even if it is not
8017 a V2 CRL: this will allow it to tolerate some broken CRLs.
8020 *) Initialize all non-automatic variables each time one of the openssl
8021 sub-programs is started (this is necessary as they may be started
8022 multiple times from the "OpenSSL>" prompt).
8023 [Lennart Bang, Bodo Moeller]
8025 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8026 removing all other RSA functionality (this is what NO_RSA does). This
8027 is so (for example) those in the US can disable those operations covered
8028 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8032 *) Non-copying interface to BIO pairs.
8033 (still largely untested)
8036 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8037 ASCII string. This was handled independently in various places before.
8040 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8041 UTF8 strings a character at a time.
8044 *) Use client_version from client hello to select the protocol
8045 (s23_srvr.c) and for RSA client key exchange verification
8046 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8049 *) Add various utility functions to handle SPKACs, these were previously
8050 handled by poking round in the structure internals. Added new function
8051 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8052 print, verify and generate SPKACs. Based on an original idea from
8053 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8056 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8059 *) Allow the config file extension section to be overwritten on the
8060 command line. Based on an original idea from Massimiliano Pala
8061 <madwolf@comune.modena.it>. The new option is called -extensions
8062 and can be applied to ca, req and x509. Also -reqexts to override
8063 the request extensions in req and -crlexts to override the crl extensions
8067 *) Add new feature to the SPKAC handling in ca. Now you can include
8068 the same field multiple times by preceding it by "XXXX." for example:
8071 this is the same syntax as used in the req config file.
8074 *) Allow certificate extensions to be added to certificate requests. These
8075 are specified in a 'req_extensions' option of the req section of the
8076 config file. They can be printed out with the -text option to req but
8077 are otherwise ignored at present.
8080 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8081 data read consists of only the final block it would not decrypted because
8082 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8083 A misplaced 'break' also meant the decrypted final block might not be
8084 copied until the next read.
8087 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8088 a few extra parameters to the DH structure: these will be useful if
8089 for example we want the value of 'q' or implement X9.42 DH.
8092 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8093 provides hooks that allow the default DSA functions or functions on a
8094 "per key" basis to be replaced. This allows hardware acceleration and
8095 hardware key storage to be handled without major modification to the
8096 library. Also added low level modexp hooks and CRYPTO_EX structure and
8097 associated functions.
8100 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8101 as "read only": it can't be written to and the buffer it points to will
8102 not be freed. Reading from a read only BIO is much more efficient than
8103 a normal memory BIO. This was added because there are several times when
8104 an area of memory needs to be read from a BIO. The previous method was
8105 to create a memory BIO and write the data to it, this results in two
8106 copies of the data and an O(n^2) reading algorithm. There is a new
8107 function BIO_new_mem_buf() which creates a read only memory BIO from
8108 an area of memory. Also modified the PKCS#7 routines to use read only
8112 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8113 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8114 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8115 but a retry condition occured while trying to read the rest.
8118 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8119 NID_pkcs7_encrypted by default: this was wrong since this should almost
8120 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8121 the encrypted data type: this is a more sensible place to put it and it
8122 allows the PKCS#12 code to be tidied up that duplicated this
8126 *) Changed obj_dat.pl script so it takes its input and output files on
8127 the command line. This should avoid shell escape redirection problems
8131 *) Initial support for certificate extension requests, these are included
8132 in things like Xenroll certificate requests. Included functions to allow
8133 extensions to be obtained and added.
8136 *) -crlf option to s_client and s_server for sending newlines as
8137 CRLF (as required by many protocols).
8140 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8142 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8143 [Ralf S. Engelschall]
8145 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8146 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8148 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8152 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8153 DH parameters/keys (q is lost during that conversion, but the resulting
8154 DH parameters contain its length).
8156 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8157 much faster than DH_generate_parameters (which creates parameters
8158 where p = 2*q + 1), and also the smaller q makes DH computations
8159 much more efficient (160-bit exponentiation instead of 1024-bit
8160 exponentiation); so this provides a convenient way to support DHE
8161 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8162 utter importance to use
8163 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8165 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8166 when such DH parameters are used, because otherwise small subgroup
8167 attacks may become possible!
8170 *) Avoid memory leak in i2d_DHparams.
8173 *) Allow the -k option to be used more than once in the enc program:
8174 this allows the same encrypted message to be read by multiple recipients.
8177 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8178 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8179 it will always use the numerical form of the OID, even if it has a short
8183 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8184 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8185 otherwise bn_mod_exp was called. In the case of hardware keys for example
8186 no private key components need be present and it might store extra data
8187 in the RSA structure, which cannot be accessed from bn_mod_exp.
8188 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8189 private key operations.
8192 *) Added support for SPARC Linux.
8195 *) pem_password_cb function type incompatibly changed from
8196 typedef int pem_password_cb(char *buf, int size, int rwflag);
8198 ....(char *buf, int size, int rwflag, void *userdata);
8199 so that applications can pass data to their callbacks:
8200 The PEM[_ASN1]_{read,write}... functions and macros now take an
8201 additional void * argument, which is just handed through whenever
8202 the password callback is called.
8203 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8205 New function SSL_CTX_set_default_passwd_cb_userdata.
8207 Compatibility note: As many C implementations push function arguments
8208 onto the stack in reverse order, the new library version is likely to
8209 interoperate with programs that have been compiled with the old
8210 pem_password_cb definition (PEM_whatever takes some data that
8211 happens to be on the stack as its last argument, and the callback
8212 just ignores this garbage); but there is no guarantee whatsoever that
8215 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8216 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8217 problems not only on Windows, but also on some Unix platforms.
8218 To avoid problematic command lines, these definitions are now in an
8219 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8220 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8223 *) MIPS III/IV assembler module is reimplemented.
8226 *) More DES library cleanups: remove references to srand/rand and
8227 delete an unused file.
8230 *) Add support for the the free Netwide assembler (NASM) under Win32,
8231 since not many people have MASM (ml) and it can be hard to obtain.
8232 This is currently experimental but it seems to work OK and pass all
8233 the tests. Check out INSTALL.W32 for info.
8236 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8237 without temporary keys kept an extra copy of the server key,
8238 and connections with temporary keys did not free everything in case
8242 *) New function RSA_check_key and new openssl rsa option -check
8243 for verifying the consistency of RSA keys.
8244 [Ulf Moeller, Bodo Moeller]
8246 *) Various changes to make Win32 compile work:
8247 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8248 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8249 comparison" warnings.
8250 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8253 *) Add a debugging option to PKCS#5 v2 key generation function: when
8254 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8255 derived keys are printed to stderr.
8258 *) Copy the flags in ASN1_STRING_dup().
8259 [Roman E. Pavlov <pre@mo.msk.ru>]
8261 *) The x509 application mishandled signing requests containing DSA
8262 keys when the signing key was also DSA and the parameters didn't match.
8264 It was supposed to omit the parameters when they matched the signing key:
8265 the verifying software was then supposed to automatically use the CA's
8266 parameters if they were absent from the end user certificate.
8268 Omitting parameters is no longer recommended. The test was also
8269 the wrong way round! This was probably due to unusual behaviour in
8270 EVP_cmp_parameters() which returns 1 if the parameters match.
8271 This meant that parameters were omitted when they *didn't* match and
8272 the certificate was useless. Certificates signed with 'ca' didn't have
8274 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8276 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8277 The interface is as follows:
8278 Applications can use
8279 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8280 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8281 "off" is now the default.
8282 The library internally uses
8283 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8284 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8285 to disable memory-checking temporarily.
8287 Some inconsistent states that previously were possible (and were
8288 even the default) are now avoided.
8290 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8291 with each memory chunk allocated; this is occasionally more helpful
8292 than just having a counter.
8294 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8296 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8300 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8301 which largely parallels "options", but is for changing API behaviour,
8302 whereas "options" are about protocol behaviour.
8303 Initial "mode" flags are:
8305 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8306 a single record has been written.
8307 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8308 retries use the same buffer location.
8309 (But all of the contents must be
8313 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8316 *) Fix problems with no-hmac etc.
8317 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8319 *) New functions RSA_get_default_method(), RSA_set_method() and
8320 RSA_get_method(). These allows replacement of RSA_METHODs without having
8321 to mess around with the internals of an RSA structure.
8324 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8325 Also really enable memory leak checks in openssl.c and in some
8327 [Chad C. Mulligan, Bodo Moeller]
8329 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8330 up the length of negative integers. This has now been simplified to just
8331 store the length when it is first determined and use it later, rather
8332 than trying to keep track of where data is copied and updating it to
8334 [Steve Henson, reported by Brien Wheeler
8335 <bwheeler@authentica-security.com>]
8337 *) Add a new function PKCS7_signatureVerify. This allows the verification
8338 of a PKCS#7 signature but with the signing certificate passed to the
8339 function itself. This contrasts with PKCS7_dataVerify which assumes the
8340 certificate is present in the PKCS#7 structure. This isn't always the
8341 case: certificates can be omitted from a PKCS#7 structure and be
8342 distributed by "out of band" means (such as a certificate database).
8345 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8346 function prototypes in pem.h, also change util/mkdef.pl to add the
8347 necessary function names.
8350 *) mk1mf.pl (used by Windows builds) did not properly read the
8351 options set by Configure in the top level Makefile, and Configure
8352 was not even able to write more than one option correctly.
8353 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8356 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8357 file to be loaded from a BIO or FILE pointer. The BIO version will
8358 for example allow memory BIOs to contain config info.
8361 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8362 Whoever hopes to achieve shared-library compatibility across versions
8363 must use this, not the compile-time macro.
8364 (Exercise 0.9.4: Which is the minimum library version required by
8366 Note: All this applies only to multi-threaded programs, others don't
8370 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8371 through a BIO pair triggered the default case, i.e.
8372 SSLerr(...,SSL_R_UNKNOWN_STATE).
8375 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8376 can use the SSL library even if none of the specific BIOs is
8380 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8381 for the encoded length.
8382 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8384 *) Add initial documentation of the X509V3 functions.
8387 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8388 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8389 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8390 secure PKCS#8 private key format with a high iteration count.
8393 *) Fix determination of Perl interpreter: A perl or perl5
8394 _directory_ in $PATH was also accepted as the interpreter.
8395 [Ralf S. Engelschall]
8397 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8398 wrong with it but it was very old and did things like calling
8399 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8403 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8404 to use the new extension code.
8407 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8408 with macros. This should make it easier to change their form, add extra
8409 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8413 *) Add to configuration table a new entry that can specify an alternative
8414 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8415 according to Mark Crispin <MRC@Panda.COM>.
8419 *) DES CBC did not update the IV. Weird.
8422 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8423 Changing the behaviour of the former might break existing programs --
8424 where IV updating is needed, des_ncbc_encrypt can be used.
8427 *) When bntest is run from "make test" it drives bc to check its
8428 calculations, as well as internally checking them. If an internal check
8429 fails, it needs to cause bc to give a non-zero result or make test carries
8430 on without noticing the failure. Fixed.
8433 *) DES library cleanups.
8436 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8437 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8438 ciphers. NOTE: although the key derivation function has been verified
8439 against some published test vectors it has not been extensively tested
8440 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8444 *) Instead of "mkdir -p", which is not fully portable, use new
8445 Perl script "util/mkdir-p.pl".
8448 *) Rewrite the way password based encryption (PBE) is handled. It used to
8449 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8450 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8451 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8452 the 'parameter' field of the AlgorithmIdentifier is passed to the
8453 underlying key generation function so it must do its own ASN1 parsing.
8454 This has also changed the EVP_PBE_CipherInit() function which now has a
8455 'parameter' argument instead of literal salt and iteration count values
8456 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8459 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8460 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8461 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8462 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8463 value was just used as a "magic string" and not used directly its
8464 value doesn't matter.
8467 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8471 *) "linux-sparc64" configuration (ultrapenguin).
8472 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8473 "linux-sparc" configuration.
8474 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8476 *) config now generates no-xxx options for missing ciphers.
8479 *) Support the EBCDIC character set (work in progress).
8480 File ebcdic.c not yet included because it has a different license.
8481 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8483 *) Support BS2000/OSD-POSIX.
8484 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8486 *) Make callbacks for key generation use void * instead of char *.
8489 *) Make S/MIME samples compile (not yet tested).
8492 *) Additional typesafe stacks.
8495 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8499 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8501 *) New configuration variant "sco5-gcc".
8503 *) Updated some demos.
8504 [Sean O Riordain, Wade Scholine]
8506 *) Add missing BIO_free at exit of pkcs12 application.
8509 *) Fix memory leak in conf.c.
8512 *) Updates for Win32 to assembler version of MD5.
8515 *) Set #! path to perl in apps/der_chop to where we found it
8516 instead of using a fixed path.
8519 *) SHA library changes for irix64-mips4-cc.
8522 *) Improvements for VMS support.
8526 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8528 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8529 This also avoids the problems with SC4.2 and unpatched SC5.
8530 [Andy Polyakov <appro@fy.chalmers.se>]
8532 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8533 These are required because of the typesafe stack would otherwise break
8534 existing code. If old code used a structure member which used to be STACK
8535 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8536 sk_num or sk_value it would produce an error because the num, data members
8537 are not present in STACK_OF. Now it just produces a warning. sk_set
8538 replaces the old method of assigning a value to sk_value
8539 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8540 that does this will no longer work (and should use sk_set instead) but
8541 this could be regarded as a "questionable" behaviour anyway.
8544 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8545 correctly handle encrypted S/MIME data.
8548 *) Change type of various DES function arguments from des_cblock
8549 (which means, in function argument declarations, pointer to char)
8550 to des_cblock * (meaning pointer to array with 8 char elements),
8551 which allows the compiler to do more typechecking; it was like
8552 that back in SSLeay, but with lots of ugly casts.
8554 Introduce new type const_des_cblock.
8557 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8558 problems: find RecipientInfo structure that matches recipient certificate
8559 and initialise the ASN1 structures properly based on passed cipher.
8562 *) Belatedly make the BN tests actually check the results.
8565 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8566 to and from BNs: it was completely broken. New compilation option
8567 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8568 key elements as negative integers.
8571 *) Reorganize and speed up MD5.
8572 [Andy Polyakov <appro@fy.chalmers.se>]
8575 [Richard Levitte <richard@levitte.org>]
8577 *) New option -out to asn1parse to allow the parsed structure to be
8578 output to a file. This is most useful when combined with the -strparse
8579 option to examine the output of things like OCTET STRINGS.
8582 *) Make SSL library a little more fool-proof by not requiring any longer
8583 that SSL_set_{accept,connect}_state be called before
8584 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8585 in many applications because usually everything *appeared* to work as
8586 intended anyway -- now it really works as intended).
8589 *) Move openssl.cnf out of lib/.
8592 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8593 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8594 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8595 [Ralf S. Engelschall]
8597 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8598 handle PKCS#7 enveloped data properly.
8599 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8601 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8602 copying pointers. The cert_st handling is changed by this in
8603 various ways (and thus what used to be known as ctx->default_cert
8604 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8605 any longer when s->cert does not give us what we need).
8606 ssl_cert_instantiate becomes obsolete by this change.
8607 As soon as we've got the new code right (possibly it already is?),
8608 we have solved a couple of bugs of the earlier code where s->cert
8609 was used as if it could not have been shared with other SSL structures.
8611 Note that using the SSL API in certain dirty ways now will result
8612 in different behaviour than observed with earlier library versions:
8613 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8614 does not influence s as it used to.
8616 In order to clean up things more thoroughly, inside SSL_SESSION
8617 we don't use CERT any longer, but a new structure SESS_CERT
8618 that holds per-session data (if available); currently, this is
8619 the peer's certificate chain and, for clients, the server's certificate
8620 and temporary key. CERT holds only those values that can have
8621 meaningful defaults in an SSL_CTX.
8624 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8625 from the internal representation. Various PKCS#7 fixes: remove some
8626 evil casts and set the enc_dig_alg field properly based on the signing
8630 *) Allow PKCS#12 password to be set from the command line or the
8631 environment. Let 'ca' get its config file name from the environment
8632 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8636 *) Allow certificate policies extension to use an IA5STRING for the
8637 organization field. This is contrary to the PKIX definition but
8638 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8642 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8643 without disallowing inline assembler and the like for non-pedantic builds.
8646 *) Support Borland C++ builder.
8647 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8652 *) SHA-1 cleanups and performance enhancements.
8653 [Andy Polyakov <appro@fy.chalmers.se>]
8655 *) Sparc v8plus assembler for the bignum library.
8656 [Andy Polyakov <appro@fy.chalmers.se>]
8658 *) Accept any -xxx and +xxx compiler options in Configure.
8661 *) Update HPUX configuration.
8664 *) Add missing sk_<type>_unshift() function to safestack.h
8665 [Ralf S. Engelschall]
8667 *) New function SSL_CTX_use_certificate_chain_file that sets the
8668 "extra_cert"s in addition to the certificate. (This makes sense
8669 only for "PEM" format files, as chains as a whole are not
8673 *) Support verify_depth from the SSL API.
8674 x509_vfy.c had what can be considered an off-by-one-error:
8675 Its depth (which was not part of the external interface)
8676 was actually counting the number of certificates in a chain;
8677 now it really counts the depth.
8680 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8681 instead of X509err, which often resulted in confusing error
8682 messages since the error codes are not globally unique
8683 (e.g. an alleged error in ssl3_accept when a certificate
8684 didn't match the private key).
8686 *) New function SSL_CTX_set_session_id_context that allows to set a default
8687 value (so that you don't need SSL_set_session_id_context for each
8688 connection using the SSL_CTX).
8691 *) OAEP decoding bug fix.
8694 *) Support INSTALL_PREFIX for package builders, as proposed by
8698 *) New Configure options "threads" and "no-threads". For systems
8699 where the proper compiler options are known (currently Solaris
8700 and Linux), "threads" is the default.
8703 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8706 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8707 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8708 such as /usr/local/bin.
8711 *) "make linux-shared" to build shared libraries.
8712 [Niels Poppe <niels@netbox.org>]
8714 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8717 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8718 extension adding in x509 utility.
8721 *) Remove NOPROTO sections and error code comments.
8724 *) Partial rewrite of the DEF file generator to now parse the ANSI
8728 *) New Configure options --prefix=DIR and --openssldir=DIR.
8731 *) Complete rewrite of the error code script(s). It is all now handled
8732 by one script at the top level which handles error code gathering,
8733 header rewriting and C source file generation. It should be much better
8734 than the old method: it now uses a modified version of Ulf's parser to
8735 read the ANSI prototypes in all header files (thus the old K&R definitions
8736 aren't needed for error creation any more) and do a better job of
8737 translating function codes into names. The old 'ASN1 error code imbedded
8738 in a comment' is no longer necessary and it doesn't use .err files which
8739 have now been deleted. Also the error code call doesn't have to appear all
8740 on one line (which resulted in some large lines...).
8743 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8746 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8747 0 (which usually indicates a closed connection), but continue reading.
8750 *) Fix some race conditions.
8753 *) Add support for CRL distribution points extension. Add Certificate
8754 Policies and CRL distribution points documentation.
8757 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8760 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8761 8 of keying material. Merlin has also confirmed interop with this fix
8762 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8763 [Merlin Hughes <merlin@baltimore.ie>]
8765 *) Fix lots of warnings.
8766 [Richard Levitte <levitte@stacken.kth.se>]
8768 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8769 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8770 [Richard Levitte <levitte@stacken.kth.se>]
8772 *) Fix problems with sizeof(long) == 8.
8773 [Andy Polyakov <appro@fy.chalmers.se>]
8775 *) Change functions to ANSI C.
8778 *) Fix typos in error codes.
8779 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8781 *) Remove defunct assembler files from Configure.
8784 *) SPARC v8 assembler BIGNUM implementation.
8785 [Andy Polyakov <appro@fy.chalmers.se>]
8787 *) Support for Certificate Policies extension: both print and set.
8788 Various additions to support the r2i method this uses.
8791 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8792 return a const string when you are expecting an allocated buffer.
8795 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8796 types DirectoryString and DisplayText.
8799 *) Add code to allow r2i extensions to access the configuration database,
8800 add an LHASH database driver and add several ctx helper functions.
8803 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8804 fail when they extended the size of a BIGNUM.
8807 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8808 support typesafe stack.
8811 *) Fix typo in SSL_[gs]et_options().
8812 [Nils Frostberg <nils@medcom.se>]
8814 *) Delete various functions and files that belonged to the (now obsolete)
8815 old X509V3 handling code.
8818 *) New Configure option "rsaref".
8821 *) Don't auto-generate pem.h.
8824 *) Introduce type-safe ASN.1 SETs.
8827 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8828 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8830 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8831 that links with OpenSSL (well at least cause lots of warnings), but fear
8832 not: the conversion is trivial, and it eliminates loads of evil casts. A
8833 few STACKed things have been converted already. Feel free to convert more.
8834 In the fullness of time, I'll do away with the STACK type altogether.
8837 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8838 specified in <certfile> by updating the entry in the index.txt file.
8839 This way one no longer has to edit the index.txt file manually for
8840 revoking a certificate. The -revoke option does the gory details now.
8841 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8843 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8844 `-text' option at all and this way the `-noout -text' combination was
8845 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8846 [Ralf S. Engelschall]
8848 *) Make sure a corresponding plain text error message exists for the
8849 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8850 verify callback function determined that a certificate was revoked.
8851 [Ralf S. Engelschall]
8853 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8854 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8855 all available cipers including rc5, which was forgotten until now.
8856 In order to let the testing shell script know which algorithms
8857 are available, a new (up to now undocumented) command
8858 "openssl list-cipher-commands" is used.
8861 *) Bugfix: s_client occasionally would sleep in select() when
8862 it should have checked SSL_pending() first.
8865 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8866 the raw DSA values prior to ASN.1 encoding.
8869 *) Tweaks to Configure
8870 [Niels Poppe <niels@netbox.org>]
8872 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8876 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8879 *) New config option to avoid instructions that are illegal on the 80386.
8880 The default code is faster, but requires at least a 486.
8883 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8884 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8885 same as SSL2_VERSION anyway.
8888 *) New "-showcerts" option for s_client.
8891 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8892 application. Various cleanups and fixes.
8895 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8896 modify error routines to work internally. Add error codes and PBE init
8897 to library startup routines.
8900 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8901 packing functions to asn1 and evp. Changed function names and error
8902 codes along the way.
8905 *) PKCS12 integration: and so it begins... First of several patches to
8906 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8907 objects to objects.h
8910 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8911 and display support for Thawte strong extranet extension.
8914 *) Add LinuxPPC support.
8915 [Jeff Dubrule <igor@pobox.org>]
8917 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8918 bn_div_words in alpha.s.
8919 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8921 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8922 OAEP isn't supported when OpenSSL is built with RSAref.
8923 [Ulf Moeller <ulf@fitug.de>]
8925 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8926 so they no longer are missing under -DNOPROTO.
8927 [Soren S. Jorvang <soren@t.dk>]
8930 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8932 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8933 doesn't work when the session is reused. Coming soon!
8936 *) Fix a security hole, that allows sessions to be reused in the wrong
8937 context thus bypassing client cert protection! All software that uses
8938 client certs and session caches in multiple contexts NEEDS PATCHING to
8939 allow session reuse! A fuller solution is in the works.
8940 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8942 *) Some more source tree cleanups (removed obsolete files
8943 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8944 permission on "config" script to be executable) and a fix for the INSTALL
8946 [Ulf Moeller <ulf@fitug.de>]
8948 *) Remove some legacy and erroneous uses of malloc, free instead of
8950 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8952 *) Make rsa_oaep_test return non-zero on error.
8953 [Ulf Moeller <ulf@fitug.de>]
8955 *) Add support for native Solaris shared libraries. Configure
8956 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8957 if someone would make that last step automatic.
8958 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8960 *) ctx_size was not built with the right compiler during "make links". Fixed.
8963 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8964 except NULL ciphers". This means the default cipher list will no longer
8965 enable NULL ciphers. They need to be specifically enabled e.g. with
8966 the string "DEFAULT:eNULL".
8969 *) Fix to RSA private encryption routines: if p < q then it would
8970 occasionally produce an invalid result. This will only happen with
8971 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8974 *) Be less restrictive and allow also `perl util/perlpath.pl
8975 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8976 because this way one can also use an interpreter named `perl5' (which is
8977 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8978 installed as `perl').
8979 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8981 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8982 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8984 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8985 advapi32.lib to Win32 build and change the pem test comparision
8986 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8987 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8988 and crypto/des/ede_cbcm_enc.c.
8991 *) DES quad checksum was broken on big-endian architectures. Fixed.
8994 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8995 Win32 test batch file so it (might) work again. The Win32 test batch file
8996 is horrible: I feel ill....
8999 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9000 in e_os.h. Audit of header files to check ANSI and non ANSI
9001 sections: 10 functions were absent from non ANSI section and not exported
9002 from Windows DLLs. Fixed up libeay.num for new functions.
9005 *) Make `openssl version' output lines consistent.
9006 [Ralf S. Engelschall]
9008 *) Fix Win32 symbol export lists for BIO functions: Added
9009 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9010 to ms/libeay{16,32}.def.
9011 [Ralf S. Engelschall]
9013 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9014 fine under Unix and passes some trivial tests I've now added. But the
9015 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9016 added to make sure no one expects that this stuff really works in the
9017 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9018 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9020 [Ralf S. Engelschall]
9022 *) Fix the generation of two part addresses in perl.
9023 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9025 *) Add config entry for Linux on MIPS.
9026 [John Tobey <jtobey@channel1.com>]
9028 *) Make links whenever Configure is run, unless we are on Windoze.
9031 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9032 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9036 *) Add a useful kludge to allow package maintainers to specify compiler and
9037 other platforms details on the command line without having to patch the
9038 Configure script everytime: One now can use ``perl Configure
9039 <id>:<details>'', i.e. platform ids are allowed to have details appended
9040 to them (seperated by colons). This is treated as there would be a static
9041 pre-configured entry in Configure's %table under key <id> with value
9042 <details> and ``perl Configure <id>'' is called. So, when you want to
9043 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9044 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9045 now, which overrides the FreeBSD-elf entry on-the-fly.
9046 [Ralf S. Engelschall]
9048 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9051 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9052 on the `perl Configure ...' command line. This way one can compile
9053 OpenSSL libraries with Position Independent Code (PIC) which is needed
9054 for linking it into DSOs.
9055 [Ralf S. Engelschall]
9057 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9061 *) Cleaned up the LICENSE document: The official contact for any license
9062 questions now is the OpenSSL core team under openssl-core@openssl.org.
9063 And add a paragraph about the dual-license situation to make sure people
9064 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9065 to the OpenSSL toolkit.
9066 [Ralf S. Engelschall]
9068 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9069 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9070 Additonally cleaned up the `make links' target: Remove unnecessary
9071 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9072 to speed processing and no longer clutter the display with confusing
9073 stuff. Instead only the actually done links are displayed.
9074 [Ralf S. Engelschall]
9076 *) Permit null encryption ciphersuites, used for authentication only. It used
9077 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9078 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9082 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9083 signed attributes when verifying signatures (this would break them),
9084 the detached data encoding was wrong and public keys obtained using
9085 X509_get_pubkey() weren't freed.
9088 *) Add text documentation for the BUFFER functions. Also added a work around
9089 to a Win95 console bug. This was triggered by the password read stuff: the
9090 last character typed gets carried over to the next fread(). If you were
9091 generating a new cert request using 'req' for example then the last
9092 character of the passphrase would be CR which would then enter the first
9096 *) Added the new `Includes OpenSSL Cryptography Software' button as
9097 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9098 button and can be used by applications based on OpenSSL to show the
9099 relationship to the OpenSSL project.
9100 [Ralf S. Engelschall]
9102 *) Remove confusing variables in function signatures in files
9103 ssl/ssl_lib.c and ssl/ssl.h.
9104 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9106 *) Don't install bss_file.c under PREFIX/include/
9107 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9109 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9110 functions that return function pointers and has support for NT specific
9111 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9112 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9113 unsigned to signed types: this was killing the Win32 compile.
9116 *) Add new certificate file to stack functions,
9117 SSL_add_dir_cert_subjects_to_stack() and
9118 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9119 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9120 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9121 This means that Apache-SSL and similar packages don't have to mess around
9122 to add as many CAs as they want to the preferred list.
9125 *) Experiment with doxygen documentation. Currently only partially applied to
9127 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9128 openssl.doxy as the configuration file.
9131 *) Get rid of remaining C++-style comments which strict C compilers hate.
9132 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9134 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9135 compiled in by default: it has problems with large keys.
9138 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9139 DH private keys and/or callback functions which directly correspond to
9140 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9141 is needed for applications which have to configure certificates on a
9142 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9144 For the RSA certificate situation is makes no difference, but
9145 for the DSA certificate situation this fixes the "no shared cipher"
9146 problem where the OpenSSL cipher selection procedure failed because the
9147 temporary keys were not overtaken from the context and the API provided
9148 no way to reconfigure them.
9149 The new functions now let applications reconfigure the stuff and they
9150 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9151 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9152 non-public-API function ssl_cert_instantiate() is used as a helper
9153 function and also to reduce code redundancy inside ssl_rsa.c.
9154 [Ralf S. Engelschall]
9156 *) Move s_server -dcert and -dkey options out of the undocumented feature
9157 area because they are useful for the DSA situation and should be
9158 recognized by the users.
9159 [Ralf S. Engelschall]
9161 *) Fix the cipher decision scheme for export ciphers: the export bits are
9162 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9163 SSL_EXP_MASK. So, the original variable has to be used instead of the
9164 already masked variable.
9165 [Richard Levitte <levitte@stacken.kth.se>]
9167 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9168 [Richard Levitte <levitte@stacken.kth.se>]
9170 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9171 from `int' to `unsigned int' because it's a length and initialized by
9172 EVP_DigestFinal() which expects an `unsigned int *'.
9173 [Richard Levitte <levitte@stacken.kth.se>]
9175 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9176 script. Instead use the usual Shell->Perl transition trick.
9177 [Ralf S. Engelschall]
9179 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9180 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9181 -noout -modulus' as it's already the case for `openssl rsa -noout
9182 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9183 currently the public key is printed (a decision which was already done by
9184 `openssl dsa -modulus' in the past) which serves a similar purpose.
9185 Additionally the NO_RSA no longer completely removes the whole -modulus
9186 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9188 [Ralf S. Engelschall]
9190 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9191 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9192 [Arne Ansper <arne@ats.cyber.ee>]
9194 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9195 to be added. Now both 'req' and 'ca' can use new objects defined in the
9199 *) Add cool BIO that does syslog (or event log on NT).
9200 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9202 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9203 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9204 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9205 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9208 *) Add preliminary config info for new extension code.
9211 *) Make RSA_NO_PADDING really use no padding.
9212 [Ulf Moeller <ulf@fitug.de>]
9214 *) Generate errors when private/public key check is done.
9217 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9218 for some CRL extensions and new objects added.
9221 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9222 key usage extension and fuller support for authority key id.
9225 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9226 padding method for RSA, which is recommended for new applications in PKCS
9227 #1 v2.0 (RFC 2437, October 1998).
9228 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9229 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9230 against Bleichbacher's attack on RSA.
9231 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9234 *) Updates to the new SSL compression code
9235 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9237 *) Fix so that the version number in the master secret, when passed
9238 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9239 (because the server will not accept higher), that the version number
9240 is 0x03,0x01, not 0x03,0x00
9241 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9243 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9244 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9245 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9248 *) Support for RAW extensions where an arbitrary extension can be
9249 created by including its DER encoding. See apps/openssl.cnf for
9253 *) Make sure latest Perl versions don't interpret some generated C array
9254 code as Perl array code in the crypto/err/err_genc.pl script.
9255 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9257 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9258 not many people have the assembler. Various Win32 compilation fixes and
9259 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9263 *) Modify configure script 'Configure' to automatically create crypto/date.h
9264 file under Win32 and also build pem.h from pem.org. New script
9265 util/mkfiles.pl to create the MINFO file on environments that can't do a
9266 'make files': perl util/mkfiles.pl >MINFO should work.
9269 *) Major rework of DES function declarations, in the pursuit of correctness
9270 and purity. As a result, many evil casts evaporated, and some weirdness,
9271 too. You may find this causes warnings in your code. Zapping your evil
9272 casts will probably fix them. Mostly.
9275 *) Fix for a typo in asn1.h. Bug fix to object creation script
9276 obj_dat.pl. It considered a zero in an object definition to mean
9277 "end of object": none of the objects in objects.h have any zeros
9278 so it wasn't spotted.
9279 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9281 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9282 Masking (CBCM). In the absence of test vectors, the best I have been able
9283 to do is check that the decrypt undoes the encrypt, so far. Send me test
9284 vectors if you have them.
9287 *) Correct calculation of key length for export ciphers (too much space was
9288 allocated for null ciphers). This has not been tested!
9291 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9292 message is now correct (it understands "crypto" and "ssl" on its
9293 command line). There is also now an "update" option. This will update
9294 the util/ssleay.num and util/libeay.num files with any new functions.
9296 perl util/mkdef.pl crypto ssl update
9297 it will update them.
9300 *) Overhauled the Perl interface (perl/*):
9301 - ported BN stuff to OpenSSL's different BN library
9302 - made the perl/ source tree CVS-aware
9303 - renamed the package from SSLeay to OpenSSL (the files still contain
9304 their history because I've copied them in the repository)
9305 - removed obsolete files (the test scripts will be replaced
9306 by better Test::Harness variants in the future)
9307 [Ralf S. Engelschall]
9309 *) First cut for a very conservative source tree cleanup:
9310 1. merge various obsolete readme texts into doc/ssleay.txt
9311 where we collect the old documents and readme texts.
9312 2. remove the first part of files where I'm already sure that we no
9313 longer need them because of three reasons: either they are just temporary
9314 files which were left by Eric or they are preserved original files where
9315 I've verified that the diff is also available in the CVS via "cvs diff
9316 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9317 the crypto/md/ stuff).
9318 [Ralf S. Engelschall]
9320 *) More extension code. Incomplete support for subject and issuer alt
9321 name, issuer and authority key id. Change the i2v function parameters
9322 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9323 what that's for :-) Fix to ASN1 macro which messed up
9324 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9327 *) Preliminary support for ENUMERATED type. This is largely copied from the
9331 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9332 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9334 *) Make sure `make rehash' target really finds the `openssl' program.
9335 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9337 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9338 like to hear about it if this slows down other processors.
9341 *) Add CygWin32 platform information to Configure script.
9342 [Alan Batie <batie@aahz.jf.intel.com>]
9344 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9345 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9347 *) New program nseq to manipulate netscape certificate sequences
9350 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9354 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9355 but the BN code had some problems that would cause failures when
9356 doing certificate verification and some other functions.
9357 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9359 *) Add ASN1 and PEM code to support netscape certificate sequences.
9362 *) Add ASN1 and PEM code to support netscape certificate sequences.
9365 *) Add several PKIX and private extended key usage OIDs.
9368 *) Modify the 'ca' program to handle the new extension code. Modify
9369 openssl.cnf for new extension format, add comments.
9372 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9373 and add a sample to openssl.cnf so req -x509 now adds appropriate
9377 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9378 error code, add initial support to X509_print() and x509 application.
9381 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9382 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9383 stuff is currently isolated and isn't even compiled yet.
9386 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9387 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9388 Removed the versions check from X509 routines when loading extensions:
9389 this allows certain broken certificates that don't set the version
9390 properly to be processed.
9393 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9394 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9395 can still be regenerated with "make depend".
9398 *) Spelling mistake in C version of CAST-128.
9399 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9401 *) Changes to the error generation code. The perl script err-code.pl
9402 now reads in the old error codes and retains the old numbers, only
9403 adding new ones if necessary. It also only changes the .err files if new
9404 codes are added. The makefiles have been modified to only insert errors
9405 when needed (to avoid needlessly modifying header files). This is done
9406 by only inserting errors if the .err file is newer than the auto generated
9407 C file. To rebuild all the error codes from scratch (the old behaviour)
9408 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9409 or delete all the .err files.
9412 *) CAST-128 was incorrectly implemented for short keys. The C version has
9413 been fixed, but is untested. The assembler versions are also fixed, but
9414 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9415 to regenerate it if needed.
9416 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9417 Hagino <itojun@kame.net>]
9419 *) File was opened incorrectly in randfile.c.
9420 [Ulf Möller <ulf@fitug.de>]
9422 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9423 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9424 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9425 al: it's just almost always a UTCTime. Note this patch adds new error
9426 codes so do a "make errors" if there are problems.
9429 *) Correct Linux 1 recognition in config.
9430 [Ulf Möller <ulf@fitug.de>]
9432 *) Remove pointless MD5 hash when using DSA keys in ca.
9433 [Anonymous <nobody@replay.com>]
9435 *) Generate an error if given an empty string as a cert directory. Also
9436 generate an error if handed NULL (previously returned 0 to indicate an
9437 error, but didn't set one).
9438 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9440 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9443 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9444 parameters. This was causing a warning which killed off the Win32 compile.
9447 *) Remove C++ style comments from crypto/bn/bn_local.h.
9448 [Neil Costigan <neil.costigan@celocom.com>]
9450 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9451 based on a text string, looking up short and long names and finally
9452 "dot" format. The "dot" format stuff didn't work. Added new function
9453 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9454 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9455 OID is not part of the table.
9458 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9459 X509_LOOKUP_by_alias().
9462 *) Sort openssl functions by name.
9465 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9466 encryption from sample DSA keys (in case anyone is interested the password
9470 *) Make _all_ *_free functions accept a NULL pointer.
9471 [Frans Heymans <fheymans@isaserver.be>]
9473 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9475 [Anonymous <nobody@replay.com>]
9477 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9478 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9480 *) Don't blow it for numeric -newkey arguments to apps/req.
9481 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9483 *) Temp key "for export" tests were wrong in s3_srvr.c.
9484 [Anonymous <nobody@replay.com>]
9486 *) Add prototype for temp key callback functions
9487 SSL_CTX_set_tmp_{rsa,dh}_callback().
9490 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9491 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9494 *) X509_name_add_entry() freed the wrong thing after an error.
9495 [Arne Ansper <arne@ats.cyber.ee>]
9497 *) rsa_eay.c would attempt to free a NULL context.
9498 [Arne Ansper <arne@ats.cyber.ee>]
9500 *) BIO_s_socket() had a broken should_retry() on Windoze.
9501 [Arne Ansper <arne@ats.cyber.ee>]
9503 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9504 [Arne Ansper <arne@ats.cyber.ee>]
9506 *) Make sure the already existing X509_STORE->depth variable is initialized
9507 in X509_STORE_new(), but document the fact that this variable is still
9508 unused in the certificate verification process.
9509 [Ralf S. Engelschall]
9511 *) Fix the various library and apps files to free up pkeys obtained from
9512 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9515 *) Fix reference counting in X509_PUBKEY_get(). This makes
9516 demos/maurice/example2.c work, amongst others, probably.
9517 [Steve Henson and Ben Laurie]
9519 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9520 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9521 are no longer created. This way we have a single and consistent command
9522 line interface `openssl <command>', similar to `cvs <command>'.
9523 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9525 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9526 BIT STRING wrapper always have zero unused bits.
9529 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9532 *) Make the top-level INSTALL documentation easier to understand.
9535 *) Makefiles updated to exit if an error occurs in a sub-directory
9536 make (including if user presses ^C) [Paul Sutton]
9538 *) Make Montgomery context stuff explicit in RSA data structure.
9541 *) Fix build order of pem and err to allow for generated pem.h.
9544 *) Fix renumbering bug in X509_NAME_delete_entry().
9547 *) Enhanced the err-ins.pl script so it makes the error library number
9548 global and can add a library name. This is needed for external ASN1 and
9549 other error libraries.
9552 *) Fixed sk_insert which never worked properly.
9555 *) Fix ASN1 macros so they can handle indefinite length construted
9556 EXPLICIT tags. Some non standard certificates use these: they can now
9560 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9561 into a single doc/ssleay.txt bundle. This way the information is still
9562 preserved but no longer messes up this directory. Now it's new room for
9563 the new set of documenation files.
9564 [Ralf S. Engelschall]
9566 *) SETs were incorrectly DER encoded. This was a major pain, because they
9567 shared code with SEQUENCEs, which aren't coded the same. This means that
9568 almost everything to do with SETs or SEQUENCEs has either changed name or
9569 number of arguments.
9570 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9572 *) Fix test data to work with the above.
9575 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9576 was already fixed by Eric for 0.9.1 it seems.
9577 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9579 *) Autodetect FreeBSD3.
9582 *) Fix various bugs in Configure. This affects the following platforms:
9586 unixware-2.0-pentium
9590 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9591 before they are needed.
9594 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9598 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9600 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9601 changed SSLeay to OpenSSL in version strings.
9602 [Ralf S. Engelschall]
9604 *) Some fixups to the top-level documents.
9607 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9608 because the symlink to include/ was missing.
9609 [Ralf S. Engelschall]
9611 *) Incorporated the popular no-RSA/DSA-only patches
9612 which allow to compile a RSA-free SSLeay.
9613 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9615 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9616 when "ssleay" is still not found.
9617 [Ralf S. Engelschall]
9619 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9620 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9622 *) Updated the README file.
9623 [Ralf S. Engelschall]
9625 *) Added various .cvsignore files in the CVS repository subdirs
9626 to make a "cvs update" really silent.
9627 [Ralf S. Engelschall]
9629 *) Recompiled the error-definition header files and added
9630 missing symbols to the Win32 linker tables.
9631 [Ralf S. Engelschall]
9633 *) Cleaned up the top-level documents;
9634 o new files: CHANGES and LICENSE
9635 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9636 o merged COPYRIGHT into LICENSE
9637 o removed obsolete TODO file
9638 o renamed MICROSOFT to INSTALL.W32
9639 [Ralf S. Engelschall]
9641 *) Removed dummy files from the 0.9.1b source tree:
9642 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9643 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9644 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9645 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9646 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9647 [Ralf S. Engelschall]
9649 *) Added various platform portability fixes.
9652 *) The Genesis of the OpenSSL rpject:
9653 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9654 Young and Tim J. Hudson created while they were working for C2Net until
9656 [The OpenSSL Project]
9659 Changes between 0.9.0b and 0.9.1b [not released]
9661 *) Updated a few CA certificates under certs/
9664 *) Changed some BIGNUM api stuff.
9667 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9668 DGUX x86, Linux Alpha, etc.
9671 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9672 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9676 *) Add -strparse option to asn1pars program which parses nested
9678 [Dr Stephen Henson <shenson@bigfoot.com>]
9680 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9683 *) DSA fix for "ca" program.
9686 *) Added "-genkey" option to "dsaparam" program.
9689 *) Added RIPE MD160 (rmd160) message digest.
9692 *) Added -a (all) option to "ssleay version" command.
9695 *) Added PLATFORM define which is the id given to Configure.
9698 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9701 *) Extended the ASN.1 parser routines.
9704 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9707 *) Added a BN_CTX to the BN library.
9710 *) Fixed the weak key values in DES library
9713 *) Changed API in EVP library for cipher aliases.
9716 *) Added support for RC2/64bit cipher.
9719 *) Converted the lhash library to the crypto/mem.c functions.
9722 *) Added more recognized ASN.1 object ids.
9725 *) Added more RSA padding checks for SSL/TLS.
9728 *) Added BIO proxy/filter functionality.
9731 *) Added extra_certs to SSL_CTX which can be used
9732 send extra CA certificates to the client in the CA cert chain sending
9733 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9736 *) Now Fortezza is denied in the authentication phase because
9737 this is key exchange mechanism is not supported by SSLeay at all.
9740 *) Additional PKCS1 checks.
9743 *) Support the string "TLSv1" for all TLS v1 ciphers.
9746 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9747 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9750 *) Fixed a few memory leaks.
9753 *) Fixed various code and comment typos.
9756 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9757 bytes sent in the client random.
9758 [Edward Bishop <ebishop@spyglass.com>]