5 Changes between 1.0.0f and 1.0.1 [xx XXX xxxx]
7 *) Extensive assembler packs updates, most notably:
9 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
10 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
11 - x86_64: bit-sliced AES implementation;
12 - ARM: NEON support, contemporary platforms optimizations;
13 - s390x: z196 support;
14 - *: GHASH and GF(2^m) multiplication implementations;
18 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
19 (removal of unnecessary code)
20 [Peter Sylvester <peter.sylvester@edelweb.fr>]
22 *) Add TLS key material exporter from RFC 5705.
25 *) Add DTLS-SRTP negotiation from RFC 5764.
28 *) Add Next Protocol Negotiation,
29 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
30 disabled with a no-npn flag to config or Configure. Code donated
32 [Adam Langley <agl@google.com> and Ben Laurie]
34 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
35 NIST-P256, NIST-P521, with constant-time single point multiplication on
36 typical inputs. Compiler support for the nonstandard type __uint128_t is
37 required to use this (present in gcc 4.4 and later, for 64-bit builds).
38 Code made available under Apache License version 2.0.
40 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
41 line to include this in your build of OpenSSL, and run "make depend" (or
42 "make update"). This enables the following EC_METHODs:
44 EC_GFp_nistp224_method()
45 EC_GFp_nistp256_method()
46 EC_GFp_nistp521_method()
48 EC_GROUP_new_by_curve_name() will automatically use these (while
49 EC_GROUP_new_curve_GFp() currently prefers the more flexible
51 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
53 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
54 all platforms. Move ssize_t definition from e_os.h to the public
55 header file e_os2.h as it now appears in public header file cms.h
58 *) New -sigopt option to the ca, req and x509 utilities. Additional
59 signature parameters can be passed using this option and in
63 *) Add RSA PSS signing function. This will generate and set the
64 appropriate AlgorithmIdentifiers for PSS based on those in the
65 corresponding EVP_MD_CTX structure. No application support yet.
68 *) Support for companion algorithm specific ASN1 signing routines.
69 New function ASN1_item_sign_ctx() signs a pre-initialised
70 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
71 the appropriate parameters.
74 *) Add new algorithm specific ASN1 verification initialisation function
75 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
76 handling will be the same no matter what EVP_PKEY_METHOD is used.
77 Add a PSS handler to support verification of PSS signatures: checked
78 against a number of sample certificates.
81 *) Add signature printing for PSS. Add PSS OIDs.
82 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
84 *) Add algorithm specific signature printing. An individual ASN1 method
85 can now print out signatures instead of the standard hex dump.
87 More complex signatures (e.g. PSS) can print out more meaningful
88 information. Include DSA version that prints out the signature
92 *) Password based recipient info support for CMS library: implementing
96 *) Split password based encryption into PBES2 and PBKDF2 functions. This
97 neatly separates the code into cipher and PBE sections and is required
98 for some algorithms that split PBES2 into separate pieces (such as
102 *) Session-handling fixes:
103 - Fix handling of connections that are resuming with a session ID,
104 but also support Session Tickets.
105 - Fix a bug that suppressed issuing of a new ticket if the client
106 presented a ticket with an expired session.
107 - Try to set the ticket lifetime hint to something reasonable.
108 - Make tickets shorter by excluding irrelevant information.
109 - On the client side, don't ignore renewed tickets.
110 [Adam Langley, Bodo Moeller (Google)]
112 *) Fix PSK session representation.
115 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
117 This work was sponsored by Intel.
120 *) Add GCM support to TLS library. Some custom code is needed to split
121 the IV between the fixed (from PRF) and explicit (from TLS record)
122 portions. This adds all GCM ciphersuites supported by RFC5288 and
123 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
124 add a special AESGCM string for GCM only.
127 *) Expand range of ctrls for AES GCM. Permit setting invocation
128 field on decrypt and retrieval of invocation field only on encrypt.
131 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
132 As required by RFC5289 these ciphersuites cannot be used if for
133 versions of TLS earlier than 1.2.
136 *) For FIPS capable OpenSSL interpret a NULL default public key method
137 as unset and return the appopriate default but do *not* set the default.
138 This means we can return the appopriate method in applications that
139 swicth between FIPS and non-FIPS modes.
142 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
143 ENGINE is used then we cannot handle that in the FIPS module so we
144 keep original code iff non-FIPS operations are allowed.
147 *) Add -attime option to openssl utilities.
148 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
150 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
153 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
154 FIPS EC methods unconditionally for now.
157 *) New build option no-ec2m to disable characteristic 2 code.
160 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
161 all cases can be covered as some introduce binary incompatibilities.
164 *) Redirect RSA operations to FIPS module including keygen,
165 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
168 *) Add similar low level API blocking to ciphers.
171 *) Low level digest APIs are not approved in FIPS mode: any attempt
172 to use these will cause a fatal error. Applications that *really* want
173 to use them can use the private_* version instead.
176 *) Redirect cipher operations to FIPS module for FIPS builds.
179 *) Redirect digest operations to FIPS module for FIPS builds.
182 *) Update build system to add "fips" flag which will link in fipscanister.o
183 for static and shared library builds embedding a signature if needed.
186 *) Output TLS supported curves in preference order instead of numerical
187 order. This is currently hardcoded for the highest order curves first.
188 This should be configurable so applications can judge speed vs strength.
191 *) Add TLS v1.2 server support for client authentication.
194 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
198 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
199 FIPS modules versions.
202 *) Add TLS v1.2 client side support for client authentication. Keep cache
203 of handshake records longer as we don't know the hash algorithm to use
204 until after the certificate request message is received.
207 *) Initial TLS v1.2 client support. Add a default signature algorithms
208 extension including all the algorithms we support. Parse new signature
209 format in client key exchange. Relax some ECC signing restrictions for
210 TLS v1.2 as indicated in RFC5246.
213 *) Add server support for TLS v1.2 signature algorithms extension. Switch
214 to new signature format when needed using client digest preference.
215 All server ciphersuites should now work correctly in TLS v1.2. No client
216 support yet and no support for client certificates.
219 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
220 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
221 ciphersuites. At present only RSA key exchange ciphersuites work with
222 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
223 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
224 and version checking.
227 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
228 with this defined it will not be affected by any changes to ssl internal
229 structures. Add several utility functions to allow openssl application
230 to work with OPENSSL_NO_SSL_INTERN defined.
234 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
236 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
239 *) Permit abbreviated handshakes when renegotiating using the function
240 SSL_renegotiate_abbreviated().
241 [Robin Seggelmann <seggelmann@fh-muenster.de>]
243 *) Add call to ENGINE_register_all_complete() to
244 ENGINE_load_builtin_engines(), so some implementations get used
245 automatically instead of needing explicit application support.
248 *) Add support for TLS key exporter as described in RFC5705.
249 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
251 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
252 a few changes are required:
254 Add SSL_OP_NO_TLSv1_1 flag.
256 Update version checking logic to handle version 1.1.
257 Add explicit IV handling (ported from DTLS code).
258 Add command line options to s_client/s_server.
261 Changes between 1.0.0e and 1.0.0f [xx XXX xxxx]
263 *) Fix ssl_ciph.c set-up race.
264 [Adam Langley (Google)]
266 *) Fix spurious failures in ecdsatest.c.
267 [Emilia Käsper (Google)]
269 *) Fix the BIO_f_buffer() implementation (which was mixing different
270 interpretations of the '..._len' fields).
271 [Adam Langley (Google)]
273 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
274 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
275 threads won't reuse the same blinding coefficients.
277 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
278 lock to call BN_BLINDING_invert_ex, and avoids one use of
279 BN_BLINDING_update for each BN_BLINDING structure (previously,
280 the last update always remained unused).
281 [Emilia Käsper (Google)]
283 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
284 [Bob Buckholz (Google)]
286 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
288 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
289 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
290 [Kaspar Brand <ossl@velox.ch>]
292 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
293 for multi-threaded use of ECDH. (CVE-2011-3210)
294 [Adam Langley (Google)]
296 *) Fix x509_name_ex_d2i memory leak on bad inputs.
299 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
300 signature public key algorithm by using OID xref utilities instead.
301 Before this you could only use some ECC ciphersuites with SHA1 only.
304 *) Add protection against ECDSA timing attacks as mentioned in the paper
305 by Billy Bob Brumley and Nicola Tuveri, see:
307 http://eprint.iacr.org/2011/232.pdf
309 [Billy Bob Brumley and Nicola Tuveri]
311 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
313 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
314 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
316 *) Fix bug in string printing code: if *any* escaping is enabled we must
317 escape the escape character (backslash) or the resulting string is
321 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
323 *) Disable code workaround for ancient and obsolete Netscape browsers
324 and servers: an attacker can use it in a ciphersuite downgrade attack.
325 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
328 *) Fixed J-PAKE implementation error, originally discovered by
329 Sebastien Martini, further info and confirmation from Stefan
330 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
333 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
335 *) Fix extension code to avoid race conditions which can result in a buffer
336 overrun vulnerability: resumed sessions must not be modified as they can
337 be shared by multiple threads. CVE-2010-3864
340 *) Fix WIN32 build system to correctly link an ENGINE directory into
344 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
346 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
348 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
350 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
352 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
353 context. The operation can be customised via the ctrl mechanism in
354 case ENGINEs want to include additional functionality.
357 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
360 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
361 output hashes compatible with older versions of OpenSSL.
362 [Willy Weisz <weisz@vcpc.univie.ac.at>]
364 *) Fix compression algorithm handling: if resuming a session use the
365 compression algorithm of the resumed session instead of determining
366 it from client hello again. Don't allow server to change algorithm.
369 *) Add load_crls() function to apps tidying load_certs() too. Add option
370 to verify utility to allow additional CRLs to be included.
373 *) Update OCSP request code to permit adding custom headers to the request:
374 some responders need this.
377 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
379 [Julia Lawall <julia@diku.dk>]
381 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
382 needlessly dereferenced structures, used obsolete functions and
383 didn't handle all updated verify codes correctly.
386 *) Disable MD2 in the default configuration.
389 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
390 indicate the initial BIO being pushed or popped. This makes it possible
391 to determine whether the BIO is the one explicitly called or as a result
392 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
393 it handles reference counts correctly and doesn't zero out the I/O bio
394 when it is not being explicitly popped. WARNING: applications which
395 included workarounds for the old buggy behaviour will need to be modified
396 or they could free up already freed BIOs.
399 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
400 renaming to all platforms (within the 0.9.8 branch, this was
401 done conditionally on Netware platforms to avoid a name clash).
402 [Guenter <lists@gknw.net>]
404 *) Add ECDHE and PSK support to DTLS.
405 [Michael Tuexen <tuexen@fh-muenster.de>]
407 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
411 *) Add "missing" function EVP_MD_flags() (without this the only way to
412 retrieve a digest flags is by accessing the structure directly. Update
413 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
414 or cipher is registered as in the "from" argument. Print out all
415 registered digests in the dgst usage message instead of manually
416 attempting to work them out.
419 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
420 this allows the use of compression and extensions. Change default cipher
421 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
422 by default unless an application cipher string requests it.
425 *) Alter match criteria in PKCS12_parse(). It used to try to use local
426 key ids to find matching certificates and keys but some PKCS#12 files
427 don't follow the (somewhat unwritten) rules and this strategy fails.
428 Now just gather all certificates together and the first private key
429 then look for the first certificate that matches the key.
432 *) Support use of registered digest and cipher names for dgst and cipher
433 commands instead of having to add each one as a special case. So now
440 openssl dgst -sha256 foo
442 and this works for ENGINE based algorithms too.
446 *) Update Gost ENGINE to support parameter files.
447 [Victor B. Wagner <vitus@cryptocom.ru>]
449 *) Support GeneralizedTime in ca utility.
450 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
452 *) Enhance the hash format used for certificate directory links. The new
453 form uses the canonical encoding (meaning equivalent names will work
454 even if they aren't identical) and uses SHA1 instead of MD5. This form
455 is incompatible with the older format and as a result c_rehash should
456 be used to rebuild symbolic links.
459 *) Make PKCS#8 the default write format for private keys, replacing the
460 traditional format. This form is standardised, more secure and doesn't
461 include an implicit MD5 dependency.
464 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
465 committed to OpenSSL should pass this lot as a minimum.
468 *) Add session ticket override functionality for use by EAP-FAST.
469 [Jouni Malinen <j@w1.fi>]
471 *) Modify HMAC functions to return a value. Since these can be implemented
472 in an ENGINE errors can occur.
475 *) Type-checked OBJ_bsearch_ex.
478 *) Type-checked OBJ_bsearch. Also some constification necessitated
479 by type-checking. Still to come: TXT_DB, bsearch(?),
480 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
484 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
485 seconds to a tm structure directly, instead of going through OS
486 specific date routines. This avoids any issues with OS routines such
487 as the year 2038 bug. New *_adj() functions for ASN1 time structures
488 and X509_time_adj_ex() to cover the extended range. The existing
489 X509_time_adj() is still usable and will no longer have any date issues.
492 *) Delta CRL support. New use deltas option which will attempt to locate
493 and search any appropriate delta CRLs available.
495 This work was sponsored by Google.
498 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
499 code and add additional score elements. Validate alternate CRL paths
500 as part of the CRL checking and indicate a new error "CRL path validation
501 error" in this case. Applications wanting additional details can use
502 the verify callback and check the new "parent" field. If this is not
503 NULL CRL path validation is taking place. Existing applications wont
504 see this because it requires extended CRL support which is off by
507 This work was sponsored by Google.
510 *) Support for freshest CRL extension.
512 This work was sponsored by Google.
515 *) Initial indirect CRL support. Currently only supported in the CRLs
516 passed directly and not via lookup. Process certificate issuer
517 CRL entry extension and lookup CRL entries by bother issuer name
518 and serial number. Check and process CRL issuer entry in IDP extension.
520 This work was sponsored by Google.
523 *) Add support for distinct certificate and CRL paths. The CRL issuer
524 certificate is validated separately in this case. Only enabled if
525 an extended CRL support flag is set: this flag will enable additional
526 CRL functionality in future.
528 This work was sponsored by Google.
531 *) Add support for policy mappings extension.
533 This work was sponsored by Google.
536 *) Fixes to pathlength constraint, self issued certificate handling,
537 policy processing to align with RFC3280 and PKITS tests.
539 This work was sponsored by Google.
542 *) Support for name constraints certificate extension. DN, email, DNS
543 and URI types are currently supported.
545 This work was sponsored by Google.
548 *) To cater for systems that provide a pointer-based thread ID rather
549 than numeric, deprecate the current numeric thread ID mechanism and
550 replace it with a structure and associated callback type. This
551 mechanism allows a numeric "hash" to be extracted from a thread ID in
552 either case, and on platforms where pointers are larger than 'long',
553 mixing is done to help ensure the numeric 'hash' is usable even if it
554 can't be guaranteed unique. The default mechanism is to use "&errno"
555 as a pointer-based thread ID to distinguish between threads.
557 Applications that want to provide their own thread IDs should now use
558 CRYPTO_THREADID_set_callback() to register a callback that will call
559 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
561 Note that ERR_remove_state() is now deprecated, because it is tied
562 to the assumption that thread IDs are numeric. ERR_remove_state(0)
563 to free the current thread's error state should be replaced by
564 ERR_remove_thread_state(NULL).
566 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
567 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
568 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
569 application was previously providing a numeric thread callback that
570 was inappropriate for distinguishing threads, then uniqueness might
571 have been obtained with &errno that happened immediately in the
572 intermediate development versions of OpenSSL; this is no longer the
573 case, the numeric thread callback will now override the automatic use
575 [Geoff Thorpe, with help from Bodo Moeller]
577 *) Initial support for different CRL issuing certificates. This covers a
578 simple case where the self issued certificates in the chain exist and
579 the real CRL issuer is higher in the existing chain.
581 This work was sponsored by Google.
584 *) Removed effectively defunct crypto/store from the build.
587 *) Revamp of STACK to provide stronger type-checking. Still to come:
588 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
589 ASN1_STRING, CONF_VALUE.
592 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
593 RAM on SSL connections. This option can save about 34k per idle SSL.
596 *) Revamp of LHASH to provide stronger type-checking. Still to come:
597 STACK, TXT_DB, bsearch, qsort.
600 *) Initial support for Cryptographic Message Syntax (aka CMS) based
601 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
602 support for data, signedData, compressedData, digestedData and
603 encryptedData, envelopedData types included. Scripts to check against
604 RFC4134 examples draft and interop and consistency checks of many
605 content types and variants.
608 *) Add options to enc utility to support use of zlib compression BIO.
611 *) Extend mk1mf to support importing of options and assembly language
612 files from Configure script, currently only included in VC-WIN32.
613 The assembly language rules can now optionally generate the source
614 files from the associated perl scripts.
617 *) Implement remaining functionality needed to support GOST ciphersuites.
618 Interop testing has been performed using CryptoPro implementations.
619 [Victor B. Wagner <vitus@cryptocom.ru>]
621 *) s390x assembler pack.
624 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
628 *) Implement Opaque PRF Input TLS extension as specified in
629 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
630 official specification yet and no extension type assignment by
631 IANA exists, this extension (for now) will have to be explicitly
632 enabled when building OpenSSL by providing the extension number
633 to use. For example, specify an option
635 -DTLSEXT_TYPE_opaque_prf_input=0x9527
637 to the "config" or "Configure" script to enable the extension,
638 assuming extension number 0x9527 (which is a completely arbitrary
639 and unofficial assignment based on the MD5 hash of the Internet
640 Draft). Note that by doing so, you potentially lose
641 interoperability with other TLS implementations since these might
642 be using the same extension number for other purposes.
644 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
645 opaque PRF input value to use in the handshake. This will create
646 an interal copy of the length-'len' string at 'src', and will
647 return non-zero for success.
649 To get more control and flexibility, provide a callback function
652 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
653 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
657 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
660 Callback function 'cb' will be called in handshakes, and is
661 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
662 Argument 'arg' is for application purposes (the value as given to
663 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
664 be provided to the callback function). The callback function
665 has to return non-zero to report success: usually 1 to use opaque
666 PRF input just if possible, or 2 to enforce use of the opaque PRF
667 input. In the latter case, the library will abort the handshake
668 if opaque PRF input is not successfully negotiated.
670 Arguments 'peerinput' and 'len' given to the callback function
671 will always be NULL and 0 in the case of a client. A server will
672 see the client's opaque PRF input through these variables if
673 available (NULL and 0 otherwise). Note that if the server
674 provides an opaque PRF input, the length must be the same as the
675 length of the client's opaque PRF input.
677 Note that the callback function will only be called when creating
678 a new session (session resumption can resume whatever was
679 previously negotiated), and will not be called in SSL 2.0
680 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
681 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
682 for applications that need to enforce opaque PRF input.
686 *) Update ssl code to support digests other than SHA1+MD5 for handshake
689 [Victor B. Wagner <vitus@cryptocom.ru>]
691 *) Add RFC4507 support to OpenSSL. This includes the corrections in
692 RFC4507bis. The encrypted ticket format is an encrypted encoded
693 SSL_SESSION structure, that way new session features are automatically
696 If a client application caches session in an SSL_SESSION structure
697 support is transparent because tickets are now stored in the encoded
700 The SSL_CTX structure automatically generates keys for ticket
701 protection in servers so again support should be possible
702 with no application modification.
704 If a client or server wishes to disable RFC4507 support then the option
705 SSL_OP_NO_TICKET can be set.
707 Add a TLS extension debugging callback to allow the contents of any client
708 or server extensions to be examined.
710 This work was sponsored by Google.
713 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
714 OpenSSL should now compile cleanly on gcc 4.2
715 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
717 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
718 support including streaming MAC support: this is required for GOST
720 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
722 *) Add option -stream to use PKCS#7 streaming in smime utility. New
723 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
724 to output in BER and PEM format.
727 *) Experimental support for use of HMAC via EVP_PKEY interface. This
728 allows HMAC to be handled via the EVP_DigestSign*() interface. The
729 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
730 ENGINE support for HMAC keys which are unextractable. New -mac and
731 -macopt options to dgst utility.
734 *) New option -sigopt to dgst utility. Update dgst to use
735 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
736 alternative signing paramaters such as X9.31 or PSS in the dgst
740 *) Change ssl_cipher_apply_rule(), the internal function that does
741 the work each time a ciphersuite string requests enabling
742 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
743 removing ("!foo+bar") a class of ciphersuites: Now it maintains
744 the order of disabled ciphersuites such that those ciphersuites
745 that most recently went from enabled to disabled not only stay
746 in order with respect to each other, but also have higher priority
747 than other disabled ciphersuites the next time ciphersuites are
750 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
751 the same ciphersuites as with "HIGH" alone, but in a specific
752 order where the PSK ciphersuites come first (since they are the
753 most recently disabled ciphersuites when "HIGH" is parsed).
755 Also, change ssl_create_cipher_list() (using this new
756 funcionality) such that between otherwise identical
757 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
761 *) Change ssl_create_cipher_list() so that it automatically
762 arranges the ciphersuites in reasonable order before starting
763 to process the rule string. Thus, the definition for "DEFAULT"
764 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
765 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
766 This makes it much easier to arrive at a reasonable default order
767 in applications for which anonymous ciphers are OK (meaning
768 that you can't actually use DEFAULT).
769 [Bodo Moeller; suggested by Victor Duchovni]
771 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
772 processing) into multiple integers instead of setting
773 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
774 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
775 (These masks as well as the individual bit definitions are hidden
776 away into the non-exported interface ssl/ssl_locl.h, so this
777 change to the definition of the SSL_CIPHER structure shouldn't
778 affect applications.) This give us more bits for each of these
779 categories, so there is no longer a need to coagulate AES128 and
780 AES256 into a single algorithm bit, and to coagulate Camellia128
781 and Camellia256 into a single algorithm bit, which has led to all
784 Thus, among other things, the kludge introduced in 0.9.7m and
785 0.9.8e for masking out AES256 independently of AES128 or masking
786 out Camellia256 independently of AES256 is not needed here in 0.9.9.
788 With the change, we also introduce new ciphersuite aliases that
789 so far were missing: "AES128", "AES256", "CAMELLIA128", and
793 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
794 Use the leftmost N bytes of the signature input if the input is
795 larger than the prime q (with N being the size in bytes of q).
798 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
799 it yet and it is largely untested.
802 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
805 *) Initial incomplete changes to avoid need for function casts in OpenSSL
806 some compilers (gcc 4.2 and later) reject their use. Safestack is
807 reimplemented. Update ASN1 to avoid use of legacy functions.
810 *) Win32/64 targets are linked with Winsock2.
813 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
814 to external functions. This can be used to increase CRL handling
815 efficiency especially when CRLs are very large by (for example) storing
816 the CRL revoked certificates in a database.
819 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
820 new CRLs added to a directory can be used. New command line option
821 -verify_return_error to s_client and s_server. This causes real errors
822 to be returned by the verify callback instead of carrying on no matter
823 what. This reflects the way a "real world" verify callback would behave.
826 *) GOST engine, supporting several GOST algorithms and public key formats.
827 Kindly donated by Cryptocom.
830 *) Partial support for Issuing Distribution Point CRL extension. CRLs
831 partitioned by DP are handled but no indirect CRL or reason partitioning
832 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
833 selected via a scoring technique which handles IDP and AKID in CRLs.
836 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
837 will ultimately be used for all verify operations: this will remove the
838 X509_STORE dependency on certificate verification and allow alternative
839 lookup methods. X509_STORE based implementations of these two callbacks.
842 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
843 Modify get_crl() to find a valid (unexpired) CRL if possible.
846 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
847 this would be called X509_CRL_cmp() but that name is already used by
848 a function that just compares CRL issuer names. Cache several CRL
849 extensions in X509_CRL structure and cache CRLDP in X509.
852 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
853 this maps equivalent X509_NAME structures into a consistent structure.
854 Name comparison can then be performed rapidly using memcmp().
857 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
861 *) Allow digests to supply their own micalg string for S/MIME type using
862 the ctrl EVP_MD_CTRL_MICALG.
865 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
866 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
867 ctrl. It can then customise the structure before and/or after signing
871 *) New function OBJ_add_sigid() to allow application defined signature OIDs
872 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
873 to free up any added signature OIDs.
876 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
877 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
878 digest and cipher tables. New options added to openssl utility:
879 list-message-digest-algorithms and list-cipher-algorithms.
882 *) Change the array representation of binary polynomials: the list
883 of degrees of non-zero coefficients is now terminated with -1.
884 Previously it was terminated with 0, which was also part of the
885 value; thus, the array representation was not applicable to
886 polynomials where t^0 has coefficient zero. This change makes
887 the array representation useful in a more general context.
890 *) Various modifications and fixes to SSL/TLS cipher string
891 handling. For ECC, the code now distinguishes between fixed ECDH
892 with RSA certificates on the one hand and with ECDSA certificates
893 on the other hand, since these are separate ciphersuites. The
894 unused code for Fortezza ciphersuites has been removed.
896 For consistency with EDH, ephemeral ECDH is now called "EECDH"
897 (not "ECDHE"). For consistency with the code for DH
898 certificates, use of ECDH certificates is now considered ECDH
899 authentication, not RSA or ECDSA authentication (the latter is
900 merely the CA's signing algorithm and not actively used in the
903 The temporary ciphersuite alias "ECCdraft" is no longer
904 available, and ECC ciphersuites are no longer excluded from "ALL"
905 and "DEFAULT". The following aliases now exist for RFC 4492
906 ciphersuites, most of these by analogy with the DH case:
908 kECDHr - ECDH cert, signed with RSA
909 kECDHe - ECDH cert, signed with ECDSA
910 kECDH - ECDH cert (signed with either RSA or ECDSA)
911 kEECDH - ephemeral ECDH
912 ECDH - ECDH cert or ephemeral ECDH
918 AECDH - anonymous ECDH
919 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
923 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
924 Use correct micalg parameters depending on digest(s) in signed message.
927 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
928 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
931 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
932 an engine to register a method. Add ENGINE lookups for methods and
933 functional reference processing.
936 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
937 EVP_{Sign,Verify}* which allow an application to customise the signature
941 *) New -resign option to smime utility. This adds one or more signers
942 to an existing PKCS#7 signedData structure. Also -md option to use an
943 alternative message digest algorithm for signing.
946 *) Tidy up PKCS#7 routines and add new functions to make it easier to
947 create PKCS7 structures containing multiple signers. Update smime
948 application to support multiple signers.
951 *) New -macalg option to pkcs12 utility to allow setting of an alternative
955 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
956 Reorganize PBE internals to lookup from a static table using NIDs,
957 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
958 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
959 PRF which will be automatically used with PBES2.
962 *) Replace the algorithm specific calls to generate keys in "req" with the
966 *) Update PKCS#7 enveloped data routines to use new API. This is now
967 supported by any public key method supporting the encrypt operation. A
968 ctrl is added to allow the public key algorithm to examine or modify
969 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
973 *) Add a ctrl to asn1 method to allow a public key algorithm to express
974 a default digest type to use. In most cases this will be SHA1 but some
975 algorithms (such as GOST) need to specify an alternative digest. The
976 return value indicates how strong the prefernce is 1 means optional and
977 2 is mandatory (that is it is the only supported type). Modify
978 ASN1_item_sign() to accept a NULL digest argument to indicate it should
979 use the default md. Update openssl utilities to use the default digest
980 type for signing if it is not explicitly indicated.
983 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
984 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
985 signing method from the key type. This effectively removes the link
986 between digests and public key types.
989 *) Add an OID cross reference table and utility functions. Its purpose is to
990 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
991 rsaEncryption. This will allow some of the algorithm specific hackery
992 needed to use the correct OID to be removed.
995 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
996 structures for PKCS7_sign(). They are now set up by the relevant public
1000 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1003 *) Add support for key derivation (agreement) in the API, DH method and
1007 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1008 public and private key formats. As a side effect these add additional
1009 command line functionality not previously available: DSA signatures can be
1010 generated and verified using pkeyutl and DH key support and generation in
1015 [Oliver Tappe <zooey@hirschkaefer.de>]
1017 *) New make target "install_html_docs" installs HTML renditions of the
1019 [Oliver Tappe <zooey@hirschkaefer.de>]
1021 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1022 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1023 support key and parameter generation and add initial key generation
1024 functionality for RSA.
1027 *) Add functions for main EVP_PKEY_method operations. The undocumented
1028 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1029 EVP_PKEY_{encrypt,decrypt}_old.
1032 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1033 key API, doesn't do much yet.
1036 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1037 public key algorithms. New option to openssl utility:
1038 "list-public-key-algorithms" to print out info.
1041 *) Implement the Supported Elliptic Curves Extension for
1042 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1045 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1046 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1049 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1050 utilities such as rsa, dsa, dsaparam etc except they process any key
1054 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1055 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1056 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1060 *) Initial support for pluggable public key ASN1.
1061 De-spaghettify the public key ASN1 handling. Move public and private
1062 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1063 algorithm specific handling to a single module within the relevant
1064 algorithm directory. Add functions to allow (near) opaque processing
1065 of public and private key structures.
1068 *) Implement the Supported Point Formats Extension for
1069 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1072 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1073 for the psk identity [hint] and the psk callback functions to the
1074 SSL_SESSION, SSL and SSL_CTX structure.
1077 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1081 SSL_CTX_use_psk_identity_hint
1082 SSL_get_psk_identity_hint
1083 SSL_get_psk_identity
1084 SSL_use_psk_identity_hint
1086 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1088 *) Add RFC 3161 compliant time stamp request creation, response generation
1089 and response verification functionality.
1090 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1092 *) Add initial support for TLS extensions, specifically for the server_name
1093 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1094 have new members for a host name. The SSL data structure has an
1095 additional member SSL_CTX *initial_ctx so that new sessions can be
1096 stored in that context to allow for session resumption, even after the
1097 SSL has been switched to a new SSL_CTX in reaction to a client's
1098 server_name extension.
1100 New functions (subject to change):
1102 SSL_get_servername()
1103 SSL_get_servername_type()
1106 New CTRL codes and macros (subject to change):
1108 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1109 - SSL_CTX_set_tlsext_servername_callback()
1110 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1111 - SSL_CTX_set_tlsext_servername_arg()
1112 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1114 openssl s_client has a new '-servername ...' option.
1116 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1117 '-key2 ...', '-servername_fatal' (subject to change). This allows
1118 testing the HostName extension for a specific single host name ('-cert'
1119 and '-key' remain fallbacks for handshakes without HostName
1120 negotiation). If the unrecogninzed_name alert has to be sent, this by
1121 default is a warning; it becomes fatal with the '-servername_fatal'
1124 [Peter Sylvester, Remy Allais, Christophe Renou]
1126 *) Whirlpool hash implementation is added.
1129 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1130 bn(64,32). Because of instruction set limitations it doesn't have
1131 any negative impact on performance. This was done mostly in order
1132 to make it possible to share assembler modules, such as bn_mul_mont
1133 implementations, between 32- and 64-bit builds without hassle.
1136 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1137 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1141 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1142 dedicated Montgomery multiplication procedure, is introduced.
1143 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1144 "64-bit" performance on certain 32-bit targets.
1147 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1148 in SSL structures. New SSL ctrl to set maximum send fragment size.
1149 Save memory by seeting the I/O buffer sizes dynamically instead of
1150 using the maximum available value.
1153 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1154 in addition to the text details.
1157 *) Very, very preliminary EXPERIMENTAL support for printing of general
1158 ASN1 structures. This currently produces rather ugly output and doesn't
1159 handle several customised structures at all.
1162 *) Integrated support for PVK file format and some related formats such
1163 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1164 these in the 'rsa' and 'dsa' utilities.
1167 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1170 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1171 place for the (very old) "NETSCAPE" format certificates which are now
1172 handled using new ASN1 code equivalents.
1175 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1176 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1177 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1180 *) Modify CRL distribution points extension code to print out previously
1181 unsupported fields. Enhance extension setting code to allow setting of
1185 *) Add print and set support for Issuing Distribution Point CRL extension.
1188 *) Change 'Configure' script to enable Camellia by default.
1191 Changes between 0.9.8r and 0.9.8s [xx XXX xxxx]
1193 *) Fix ssl_ciph.c set-up race.
1194 [Adam Langley (Google)]
1196 *) Fix spurious failures in ecdsatest.c.
1197 [Emilia Käsper (Google)]
1199 *) Fix the BIO_f_buffer() implementation (which was mixing different
1200 interpretations of the '..._len' fields).
1201 [Adam Langley (Google)]
1203 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1204 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1205 threads won't reuse the same blinding coefficients.
1207 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1208 lock to call BN_BLINDING_invert_ex, and avoids one use of
1209 BN_BLINDING_update for each BN_BLINDING structure (previously,
1210 the last update always remained unused).
1211 [Emilia Käsper (Google)]
1213 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1214 for multi-threaded use of ECDH.
1215 [Adam Langley (Google)]
1217 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1220 *) Add protection against ECDSA timing attacks as mentioned in the paper
1221 by Billy Bob Brumley and Nicola Tuveri, see:
1223 http://eprint.iacr.org/2011/232.pdf
1225 [Billy Bob Brumley and Nicola Tuveri]
1227 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1229 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1230 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1232 *) Fix bug in string printing code: if *any* escaping is enabled we must
1233 escape the escape character (backslash) or the resulting string is
1237 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1239 *) Disable code workaround for ancient and obsolete Netscape browsers
1240 and servers: an attacker can use it in a ciphersuite downgrade attack.
1241 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1244 *) Fixed J-PAKE implementation error, originally discovered by
1245 Sebastien Martini, further info and confirmation from Stefan
1246 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1249 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1251 *) Fix extension code to avoid race conditions which can result in a buffer
1252 overrun vulnerability: resumed sessions must not be modified as they can
1253 be shared by multiple threads. CVE-2010-3864
1256 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1259 *) Don't reencode certificate when calculating signature: cache and use
1260 the original encoding instead. This makes signature verification of
1261 some broken encodings work correctly.
1264 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1265 is also one of the inputs.
1266 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1268 *) Don't repeatedly append PBE algorithms to table if they already exist.
1269 Sort table on each new add. This effectively makes the table read only
1270 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1274 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1276 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1279 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1280 access or freeing data twice (CVE-2010-0742)
1281 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1283 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1284 common in certificates and some applications which only call
1285 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1289 Reduce copying into .apps and .test in makevms.com
1290 Don't try to use blank CA certificate in CA.com
1291 Allow use of C files from original directories in maketests.com
1292 [Steven M. Schweda" <sms@antinode.info>]
1294 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1296 *) When rejecting SSL/TLS records due to an incorrect version number, never
1297 update s->server with a new major version number. As of
1298 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1299 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1300 the previous behavior could result in a read attempt at NULL when
1301 receiving specific incorrect SSL/TLS records once record payload
1302 protection is active. (CVE-2010-0740)
1303 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1305 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1306 could be crashed if the relevant tables were not present (e.g. chrooted).
1307 [Tomas Hoger <thoger@redhat.com>]
1309 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1311 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1312 [Martin Olsson, Neel Mehta]
1314 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1315 accommodate for stack sorting, always a write lock!).
1318 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1319 excessive delays in the RAND_poll(): over a minute. As a workaround
1320 include a time check in the inner Heap32Next loop too.
1323 *) The code that handled flushing of data in SSL/TLS originally used the
1324 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1325 the problem outlined in PR#1949. The fix suggested there however can
1326 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1327 of Apache). So instead simplify the code to flush unconditionally.
1328 This should be fine since flushing with no data to flush is a no op.
1331 *) Handle TLS versions 2.0 and later properly and correctly use the
1332 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1333 off ancient servers have a habit of sticking around for a while...
1336 *) Modify compression code so it frees up structures without using the
1337 ex_data callbacks. This works around a problem where some applications
1338 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1339 restarting) then use compression (e.g. SSL with compression) later.
1340 This results in significant per-connection memory leaks and
1341 has caused some security issues including CVE-2008-1678 and
1345 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1346 change when encrypting or decrypting.
1349 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1350 connect and renegotiate with servers which do not support RI.
1351 Until RI is more widely deployed this option is enabled by default.
1354 *) Add "missing" ssl ctrls to clear options and mode.
1357 *) If client attempts to renegotiate and doesn't support RI respond with
1358 a no_renegotiation alert as required by RFC5746. Some renegotiating
1359 TLS clients will continue a connection gracefully when they receive
1360 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1361 waiting for a server hello which it will never receive. Now we treat a
1362 received no_renegotiation alert as a fatal error. This is because
1363 applications requesting a renegotiation might well expect it to succeed
1364 and would have no code in place to handle the server denying it so the
1365 only safe thing to do is to terminate the connection.
1368 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1369 peer supports secure renegotiation and 0 otherwise. Print out peer
1370 renegotiation support in s_client/s_server.
1373 *) Replace the highly broken and deprecated SPKAC certification method with
1374 the updated NID creation version. This should correctly handle UTF8.
1377 *) Implement RFC5746. Re-enable renegotiation but require the extension
1378 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1379 turns out to be a bad idea. It has been replaced by
1380 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1381 SSL_CTX_set_options(). This is really not recommended unless you
1382 know what you are doing.
1383 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1385 *) Fixes to stateless session resumption handling. Use initial_ctx when
1386 issuing and attempting to decrypt tickets in case it has changed during
1387 servername handling. Use a non-zero length session ID when attempting
1388 stateless session resumption: this makes it possible to determine if
1389 a resumption has occurred immediately after receiving server hello
1390 (several places in OpenSSL subtly assume this) instead of later in
1394 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1395 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1396 fixes for a few places where the return code is not checked
1398 [Julia Lawall <julia@diku.dk>]
1400 *) Add --strict-warnings option to Configure script to include devteam
1401 warnings in other configurations.
1404 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1405 makes it possible to install openssl libraries in locations which
1406 have names other than "lib", for example "/usr/lib64" which some
1408 [Steve Henson, based on patch from Jeremy Utley]
1410 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1411 X690 8.9.12 and can produce some misleading textual output of OIDs.
1412 [Steve Henson, reported by Dan Kaminsky]
1414 *) Delete MD2 from algorithm tables. This follows the recommendation in
1415 several standards that it is not used in new applications due to
1416 several cryptographic weaknesses. For binary compatibility reasons
1417 the MD2 API is still compiled in by default.
1420 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1424 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1425 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1427 [Guenter <lists@gknw.net>]
1429 *) Fix the server certificate chain building code to use X509_verify_cert(),
1430 it used to have an ad-hoc builder which was unable to cope with anything
1431 other than a simple chain.
1432 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1434 *) Don't check self signed certificate signatures in X509_verify_cert()
1435 by default (a flag can override this): it just wastes time without
1436 adding any security. As a useful side effect self signed root CAs
1437 with non-FIPS digests are now usable in FIPS mode.
1440 *) In dtls1_process_out_of_seq_message() the check if the current message
1441 is already buffered was missing. For every new message was memory
1442 allocated, allowing an attacker to perform an denial of service attack
1443 with sending out of seq handshake messages until there is no memory
1444 left. Additionally every future messege was buffered, even if the
1445 sequence number made no sense and would be part of another handshake.
1446 So only messages with sequence numbers less than 10 in advance will be
1447 buffered. (CVE-2009-1378)
1448 [Robin Seggelmann, discovered by Daniel Mentz]
1450 *) Records are buffered if they arrive with a future epoch to be
1451 processed after finishing the corresponding handshake. There is
1452 currently no limitation to this buffer allowing an attacker to perform
1453 a DOS attack with sending records with future epochs until there is no
1454 memory left. This patch adds the pqueue_size() function to detemine
1455 the size of a buffer and limits the record buffer to 100 entries.
1457 [Robin Seggelmann, discovered by Daniel Mentz]
1459 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1460 parent structure is freed. (CVE-2009-1379)
1463 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1464 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1467 [Ilya O. <vrghost@gmail.com>]
1469 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1471 *) Disable renegotiation completely - this fixes a severe security
1472 problem (CVE-2009-3555) at the cost of breaking all
1473 renegotiation. Renegotiation can be re-enabled by setting
1474 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1475 run-time. This is really not recommended unless you know what
1479 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1481 *) Don't set val to NULL when freeing up structures, it is freed up by
1482 underlying code. If sizeof(void *) > sizeof(long) this can result in
1483 zeroing past the valid field. (CVE-2009-0789)
1484 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1486 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1487 checked correctly. This would allow some invalid signed attributes to
1488 appear to verify correctly. (CVE-2009-0591)
1489 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1491 *) Reject UniversalString and BMPString types with invalid lengths. This
1492 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1493 a legal length. (CVE-2009-0590)
1496 *) Set S/MIME signing as the default purpose rather than setting it
1497 unconditionally. This allows applications to override it at the store
1501 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1502 to handle some structures.
1505 *) Improve efficiency of mem_gets: don't search whole buffer each time
1507 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1509 *) New -hex option for openssl rand.
1512 *) Print out UTF8String and NumericString when parsing ASN1.
1515 *) Support NumericString type for name components.
1518 *) Allow CC in the environment to override the automatically chosen
1519 compiler. Note that nothing is done to ensure flags work with the
1523 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1525 *) Properly check EVP_VerifyFinal() and similar return values
1527 [Ben Laurie, Bodo Moeller, Google Security Team]
1529 *) Enable TLS extensions by default.
1532 *) Allow the CHIL engine to be loaded, whether the application is
1533 multithreaded or not. (This does not release the developer from the
1534 obligation to set up the dynamic locking callbacks.)
1535 [Sander Temme <sander@temme.net>]
1537 *) Use correct exit code if there is an error in dgst command.
1538 [Steve Henson; problem pointed out by Roland Dirlewanger]
1540 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1541 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1544 *) Add experimental JPAKE support, including demo authentication in
1545 s_client and s_server.
1548 *) Set the comparison function in v3_addr_canonize().
1549 [Rob Austein <sra@hactrn.net>]
1551 *) Add support for XMPP STARTTLS in s_client.
1552 [Philip Paeps <philip@freebsd.org>]
1554 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1555 to ensure that even with this option, only ciphersuites in the
1556 server's preference list will be accepted. (Note that the option
1557 applies only when resuming a session, so the earlier behavior was
1558 just about the algorithm choice for symmetric cryptography.)
1561 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1563 *) Fix NULL pointer dereference if a DTLS server received
1564 ChangeCipherSpec as first record (CVE-2009-1386).
1567 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1568 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1571 *) The fix in 0.9.8c that supposedly got rid of unsafe
1572 double-checked locking was incomplete for RSA blinding,
1573 addressing just one layer of what turns out to have been
1574 doubly unsafe triple-checked locking.
1576 So now fix this for real by retiring the MONT_HELPER macro
1577 in crypto/rsa/rsa_eay.c.
1579 [Bodo Moeller; problem pointed out by Marius Schilder]
1581 *) Various precautionary measures:
1583 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1585 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1586 (NB: This would require knowledge of the secret session ticket key
1587 to exploit, in which case you'd be SOL either way.)
1589 - Change bn_nist.c so that it will properly handle input BIGNUMs
1590 outside the expected range.
1592 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1595 [Neel Mehta, Bodo Moeller]
1597 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1598 the load fails. Useful for distros.
1599 [Ben Laurie and the FreeBSD team]
1601 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1604 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1607 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1609 This work was sponsored by Logica.
1612 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1613 keystores. Support for SSL/TLS client authentication too.
1614 Not compiled unless enable-capieng specified to Configure.
1616 This work was sponsored by Logica.
1619 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1620 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1621 attribute creation routines such as certifcate requests and PKCS#12
1625 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1627 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1628 handshake which could lead to a cilent crash as found using the
1629 Codenomicon TLS test suite (CVE-2008-1672)
1630 [Steve Henson, Mark Cox]
1632 *) Fix double free in TLS server name extensions which could lead to
1633 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1636 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1638 Clear the error queue to ensure that error entries left from
1639 older function calls do not interfere with the correct operation.
1640 [Lutz Jaenicke, Erik de Castro Lopo]
1642 *) Remove root CA certificates of commercial CAs:
1644 The OpenSSL project does not recommend any specific CA and does not
1645 have any policy with respect to including or excluding any CA.
1646 Therefore it does not make any sense to ship an arbitrary selection
1647 of root CA certificates with the OpenSSL software.
1650 *) RSA OAEP patches to fix two separate invalid memory reads.
1651 The first one involves inputs when 'lzero' is greater than
1652 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1653 before the beginning of from). The second one involves inputs where
1654 the 'db' section contains nothing but zeroes (there is a one-byte
1655 invalid read after the end of 'db').
1656 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1658 *) Partial backport from 0.9.9-dev:
1660 Introduce bn_mul_mont (dedicated Montgomery multiplication
1661 procedure) as a candidate for BIGNUM assembler implementation.
1662 While 0.9.9-dev uses assembler for various architectures, only
1663 x86_64 is available by default here in the 0.9.8 branch, and
1664 32-bit x86 is available through a compile-time setting.
1666 To try the 32-bit x86 assembler implementation, use Configure
1667 option "enable-montasm" (which exists only for this backport).
1669 As "enable-montasm" for 32-bit x86 disclaims code stability
1670 anyway, in this constellation we activate additional code
1671 backported from 0.9.9-dev for further performance improvements,
1672 namely BN_from_montgomery_word. (To enable this otherwise,
1673 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1675 [Andy Polyakov (backport partially by Bodo Moeller)]
1677 *) Add TLS session ticket callback. This allows an application to set
1678 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1679 values. This is useful for key rollover for example where several key
1680 sets may exist with different names.
1683 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1684 This was broken until now in 0.9.8 releases, such that the only way
1685 a registered ENGINE could be used (assuming it initialises
1686 successfully on the host) was to explicitly set it as the default
1687 for the relevant algorithms. This is in contradiction with 0.9.7
1688 behaviour and the documentation. With this fix, when an ENGINE is
1689 registered into a given algorithm's table of implementations, the
1690 'uptodate' flag is reset so that auto-discovery will be used next
1691 time a new context for that algorithm attempts to select an
1693 [Ian Lister (tweaked by Geoff Thorpe)]
1695 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1696 implemention in the following ways:
1698 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1701 Lack of BER streaming support means one pass streaming processing is
1702 only supported if data is detached: setting the streaming flag is
1703 ignored for embedded content.
1705 CMS support is disabled by default and must be explicitly enabled
1706 with the enable-cms configuration option.
1709 *) Update the GMP engine glue to do direct copies between BIGNUM and
1710 mpz_t when openssl and GMP use the same limb size. Otherwise the
1711 existing "conversion via a text string export" trick is still used.
1712 [Paul Sheer <paulsheer@gmail.com>]
1714 *) Zlib compression BIO. This is a filter BIO which compressed and
1715 uncompresses any data passed through it.
1718 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1719 RFC3394 compatible AES key wrapping.
1722 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1723 sets string data without copying. X509_ALGOR_set0() and
1724 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1725 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1726 from an X509_ATTRIBUTE structure optionally checking it occurs only
1727 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1731 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1732 to get the expected BN_FLG_CONSTTIME behavior.
1733 [Bodo Moeller (Google)]
1737 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1738 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1739 - added some more tests to do_tests.pl
1740 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1741 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1742 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1743 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1744 - various changes to netware.pl to enable gcc-cross builds on Win32
1746 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1747 - various changes to fix missing prototype warnings
1748 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1749 - added AES, WHIRLPOOL and CPUID assembler code to build files
1750 - added missing AES assembler make rules to mk1mf.pl
1751 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1752 [Guenter Knauf <eflash@gmx.net>]
1754 *) Implement certificate status request TLS extension defined in RFC3546.
1755 A client can set the appropriate parameters and receive the encoded
1756 OCSP response via a callback. A server can query the supplied parameters
1757 and set the encoded OCSP response in the callback. Add simplified examples
1758 to s_client and s_server.
1761 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1763 *) Fix various bugs:
1764 + Binary incompatibility of ssl_ctx_st structure
1765 + DTLS interoperation with non-compliant servers
1766 + Don't call get_session_cb() without proposed session
1767 + Fix ia64 assembler code
1768 [Andy Polyakov, Steve Henson]
1770 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1772 *) DTLS Handshake overhaul. There were longstanding issues with
1773 OpenSSL DTLS implementation, which were making it impossible for
1774 RFC 4347 compliant client to communicate with OpenSSL server.
1775 Unfortunately just fixing these incompatibilities would "cut off"
1776 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1777 server keeps tolerating non RFC compliant syntax. The opposite is
1778 not true, 0.9.8f client can not communicate with earlier server.
1779 This update even addresses CVE-2007-4995.
1782 *) Changes to avoid need for function casts in OpenSSL: some compilers
1783 (gcc 4.2 and later) reject their use.
1784 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1787 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1788 RFC4507bis. The encrypted ticket format is an encrypted encoded
1789 SSL_SESSION structure, that way new session features are automatically
1792 If a client application caches session in an SSL_SESSION structure
1793 support is transparent because tickets are now stored in the encoded
1796 The SSL_CTX structure automatically generates keys for ticket
1797 protection in servers so again support should be possible
1798 with no application modification.
1800 If a client or server wishes to disable RFC4507 support then the option
1801 SSL_OP_NO_TICKET can be set.
1803 Add a TLS extension debugging callback to allow the contents of any client
1804 or server extensions to be examined.
1806 This work was sponsored by Google.
1809 *) Add initial support for TLS extensions, specifically for the server_name
1810 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1811 have new members for a host name. The SSL data structure has an
1812 additional member SSL_CTX *initial_ctx so that new sessions can be
1813 stored in that context to allow for session resumption, even after the
1814 SSL has been switched to a new SSL_CTX in reaction to a client's
1815 server_name extension.
1817 New functions (subject to change):
1819 SSL_get_servername()
1820 SSL_get_servername_type()
1823 New CTRL codes and macros (subject to change):
1825 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1826 - SSL_CTX_set_tlsext_servername_callback()
1827 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1828 - SSL_CTX_set_tlsext_servername_arg()
1829 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1831 openssl s_client has a new '-servername ...' option.
1833 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1834 '-key2 ...', '-servername_fatal' (subject to change). This allows
1835 testing the HostName extension for a specific single host name ('-cert'
1836 and '-key' remain fallbacks for handshakes without HostName
1837 negotiation). If the unrecogninzed_name alert has to be sent, this by
1838 default is a warning; it becomes fatal with the '-servername_fatal'
1841 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1843 *) Add AES and SSE2 assembly language support to VC++ build.
1846 *) Mitigate attack on final subtraction in Montgomery reduction.
1849 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1850 (which previously caused an internal error).
1853 *) Squeeze another 10% out of IGE mode when in != out.
1856 *) AES IGE mode speedup.
1857 [Dean Gaudet (Google)]
1859 *) Add the Korean symmetric 128-bit cipher SEED (see
1860 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1861 add SEED ciphersuites from RFC 4162:
1863 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1864 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1865 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1866 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1868 To minimize changes between patchlevels in the OpenSSL 0.9.8
1869 series, SEED remains excluded from compilation unless OpenSSL
1870 is configured with 'enable-seed'.
1871 [KISA, Bodo Moeller]
1873 *) Mitigate branch prediction attacks, which can be practical if a
1874 single processor is shared, allowing a spy process to extract
1875 information. For detailed background information, see
1876 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1877 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1878 and Necessary Software Countermeasures"). The core of the change
1879 are new versions BN_div_no_branch() and
1880 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1881 respectively, which are slower, but avoid the security-relevant
1882 conditional branches. These are automatically called by BN_div()
1883 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1884 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1885 remove a conditional branch.
1887 BN_FLG_CONSTTIME is the new name for the previous
1888 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1889 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1890 in the exponent causes BN_mod_exp_mont() to use the alternative
1891 implementation in BN_mod_exp_mont_consttime().) The old name
1892 remains as a deprecated alias.
1894 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1895 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1896 constant-time implementations for more than just exponentiation.
1897 Here too the old name is kept as a deprecated alias.
1899 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1900 the BN_BLINDING structure gets an independent copy of the
1901 modulus. This means that the previous "BIGNUM *m" argument to
1902 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1903 essentially becomes "const BIGNUM *m", although we can't actually
1904 change this in the header file before 0.9.9. It allows
1905 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1906 enable BN_FLG_CONSTTIME.
1908 [Matthew D Wood (Intel Corp)]
1910 *) In the SSL/TLS server implementation, be strict about session ID
1911 context matching (which matters if an application uses a single
1912 external cache for different purposes). Previously,
1913 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1914 set. This did ensure strict client verification, but meant that,
1915 with applications using a single external cache for quite
1916 different requirements, clients could circumvent ciphersuite
1917 restrictions for a given session ID context by starting a session
1918 in a different context.
1921 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1922 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1923 authentication-only ciphersuites.
1926 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1927 not complete and could lead to a possible single byte overflow
1928 (CVE-2007-5135) [Ben Laurie]
1930 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1932 *) Since AES128 and AES256 (and similarly Camellia128 and
1933 Camellia256) share a single mask bit in the logic of
1934 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1935 kludge to work properly if AES128 is available and AES256 isn't
1936 (or if Camellia128 is available and Camellia256 isn't).
1939 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1940 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1941 When a point or a seed is encoded in a BIT STRING, we need to
1942 prevent the removal of trailing zero bits to get the proper DER
1943 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1944 of a NamedBitList, for which trailing 0 bits need to be removed.)
1947 *) Have SSL/TLS server implementation tolerate "mismatched" record
1948 protocol version while receiving ClientHello even if the
1949 ClientHello is fragmented. (The server can't insist on the
1950 particular protocol version it has chosen before the ServerHello
1951 message has informed the client about his choice.)
1954 *) Add RFC 3779 support.
1955 [Rob Austein for ARIN, Ben Laurie]
1957 *) Load error codes if they are not already present instead of using a
1958 static variable. This allows them to be cleanly unloaded and reloaded.
1959 Improve header file function name parsing.
1962 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1963 or CAPABILITY handshake as required by RFCs.
1966 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1968 *) Introduce limits to prevent malicious keys being able to
1969 cause a denial of service. (CVE-2006-2940)
1970 [Steve Henson, Bodo Moeller]
1972 *) Fix ASN.1 parsing of certain invalid structures that can result
1973 in a denial of service. (CVE-2006-2937) [Steve Henson]
1975 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1976 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1978 *) Fix SSL client code which could crash if connecting to a
1979 malicious SSLv2 server. (CVE-2006-4343)
1980 [Tavis Ormandy and Will Drewry, Google Security Team]
1982 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1983 match only those. Before that, "AES256-SHA" would be interpreted
1984 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1985 the same strength classification in 0.9.7h) as we currently only
1986 have a single AES bit in the ciphersuite description bitmap.
1987 That change, however, also applied to ciphersuite strings such as
1988 "RC4-MD5" that intentionally matched multiple ciphersuites --
1989 namely, SSL 2.0 ciphersuites in addition to the more common ones
1990 from SSL 3.0/TLS 1.0.
1992 So we change the selection algorithm again: Naming an explicit
1993 ciphersuite selects this one ciphersuite, and any other similar
1994 ciphersuite (same bitmap) from *other* protocol versions.
1995 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1996 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1998 Since SSL 2.0 does not have any ciphersuites for which the
1999 128/256 bit distinction would be relevant, this works for now.
2000 The proper fix will be to use different bits for AES128 and
2001 AES256, which would have avoided the problems from the beginning;
2002 however, bits are scarce, so we can only do this in a new release
2003 (not just a patchlevel) when we can change the SSL_CIPHER
2004 definition to split the single 'unsigned long mask' bitmap into
2005 multiple values to extend the available space.
2009 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2011 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2012 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2014 *) Add AES IGE and biIGE modes.
2017 *) Change the Unix randomness entropy gathering to use poll() when
2018 possible instead of select(), since the latter has some
2019 undesirable limitations.
2020 [Darryl Miles via Richard Levitte and Bodo Moeller]
2022 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2023 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2024 cannot be implicitly activated as part of, e.g., the "AES" alias.
2025 However, please upgrade to OpenSSL 0.9.9[-dev] for
2026 non-experimental use of the ECC ciphersuites to get TLS extension
2027 support, which is required for curve and point format negotiation
2028 to avoid potential handshake problems.
2031 *) Disable rogue ciphersuites:
2033 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2034 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2035 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2037 The latter two were purportedly from
2038 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2041 Also deactivate the remaining ciphersuites from
2042 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2043 unofficial, and the ID has long expired.
2046 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2047 dual-core machines) and other potential thread-safety issues.
2050 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2051 versions), which is now available for royalty-free use
2052 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2053 Also, add Camellia TLS ciphersuites from RFC 4132.
2055 To minimize changes between patchlevels in the OpenSSL 0.9.8
2056 series, Camellia remains excluded from compilation unless OpenSSL
2057 is configured with 'enable-camellia'.
2060 *) Disable the padding bug check when compression is in use. The padding
2061 bug check assumes the first packet is of even length, this is not
2062 necessarily true if compresssion is enabled and can result in false
2063 positives causing handshake failure. The actual bug test is ancient
2064 code so it is hoped that implementations will either have fixed it by
2065 now or any which still have the bug do not support compression.
2068 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2070 *) When applying a cipher rule check to see if string match is an explicit
2071 cipher suite and only match that one cipher suite if it is.
2074 *) Link in manifests for VC++ if needed.
2075 [Austin Ziegler <halostatue@gmail.com>]
2077 *) Update support for ECC-based TLS ciphersuites according to
2078 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2079 TLS extensions, which are supported starting with the 0.9.9
2080 branch, not in the OpenSSL 0.9.8 branch).
2083 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2084 opaque EVP_CIPHER_CTX handling.
2087 *) Fixes and enhancements to zlib compression code. We now only use
2088 "zlib1.dll" and use the default __cdecl calling convention on Win32
2089 to conform with the standards mentioned here:
2090 http://www.zlib.net/DLL_FAQ.txt
2091 Static zlib linking now works on Windows and the new --with-zlib-include
2092 --with-zlib-lib options to Configure can be used to supply the location
2093 of the headers and library. Gracefully handle case where zlib library
2097 *) Several fixes and enhancements to the OID generation code. The old code
2098 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2099 handle numbers larger than ULONG_MAX, truncated printing and had a
2100 non standard OBJ_obj2txt() behaviour.
2103 *) Add support for building of engines under engine/ as shared libraries
2104 under VC++ build system.
2107 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2108 Hopefully, we will not see any false combination of paths any more.
2111 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2113 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2114 (part of SSL_OP_ALL). This option used to disable the
2115 countermeasure against man-in-the-middle protocol-version
2116 rollback in the SSL 2.0 server implementation, which is a bad
2117 idea. (CVE-2005-2969)
2119 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2120 for Information Security, National Institute of Advanced Industrial
2121 Science and Technology [AIST], Japan)]
2123 *) Add two function to clear and return the verify parameter flags.
2126 *) Keep cipherlists sorted in the source instead of sorting them at
2127 runtime, thus removing the need for a lock.
2130 *) Avoid some small subgroup attacks in Diffie-Hellman.
2131 [Nick Mathewson and Ben Laurie]
2133 *) Add functions for well-known primes.
2136 *) Extended Windows CE support.
2137 [Satoshi Nakamura and Andy Polyakov]
2139 *) Initialize SSL_METHOD structures at compile time instead of during
2140 runtime, thus removing the need for a lock.
2143 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2144 attempting to decrypt each encrypted key in turn. Add support to
2148 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2150 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2153 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2156 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2157 key into the same file any more.
2160 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2163 *) Add -utf8 command line and config file option to 'ca'.
2164 [Stefan <stf@udoma.org]
2166 *) Removed the macro des_crypt(), as it seems to conflict with some
2167 libraries. Use DES_crypt().
2170 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2171 involves renaming the source and generated shared-libs for
2172 both. The engines will accept the corrected or legacy ids
2173 ('ncipher' and '4758_cca' respectively) when binding. NB,
2174 this only applies when building 'shared'.
2175 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2177 *) Add attribute functions to EVP_PKEY structure. Modify
2178 PKCS12_create() to recognize a CSP name attribute and
2179 use it. Make -CSP option work again in pkcs12 utility.
2182 *) Add new functionality to the bn blinding code:
2183 - automatic re-creation of the BN_BLINDING parameters after
2184 a fixed number of uses (currently 32)
2185 - add new function for parameter creation
2186 - introduce flags to control the update behaviour of the
2187 BN_BLINDING parameters
2188 - hide BN_BLINDING structure
2189 Add a second BN_BLINDING slot to the RSA structure to improve
2190 performance when a single RSA object is shared among several
2194 *) Add support for DTLS.
2195 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2197 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2198 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2201 *) Remove buggy and incompletet DH cert support from
2202 ssl/ssl_rsa.c and ssl/s3_both.c
2205 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2206 the apps/openssl applications.
2209 *) Compile clean with "-Wall -Wmissing-prototypes
2210 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2211 DEBUG_SAFESTACK must also be set.
2214 *) Change ./Configure so that certain algorithms can be disabled by default.
2215 The new counterpiece to "no-xxx" is "enable-xxx".
2217 The patented RC5 and MDC2 algorithms will now be disabled unless
2218 "enable-rc5" and "enable-mdc2", respectively, are specified.
2220 (IDEA remains enabled despite being patented. This is because IDEA
2221 is frequently required for interoperability, and there is no license
2222 fee for non-commercial use. As before, "no-idea" can be used to
2223 avoid this algorithm.)
2227 *) Add processing of proxy certificates (see RFC 3820). This work was
2228 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2229 EGEE (Enabling Grids for E-science in Europe).
2232 *) RC4 performance overhaul on modern architectures/implementations, such
2233 as Intel P4, IA-64 and AMD64.
2236 *) New utility extract-section.pl. This can be used specify an alternative
2237 section number in a pod file instead of having to treat each file as
2238 a separate case in Makefile. This can be done by adding two lines to the
2241 =for comment openssl_section:XXX
2243 The blank line is mandatory.
2247 *) New arguments -certform, -keyform and -pass for s_client and s_server
2248 to allow alternative format key and certificate files and passphrase
2252 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2253 update associated structures and add various utility functions.
2255 Add new policy related verify parameters, include policy checking in
2256 standard verify code. Enhance 'smime' application with extra parameters
2257 to support policy checking and print out.
2260 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2261 Nehemiah processors. These extensions support AES encryption in hardware
2262 as well as RNG (though RNG support is currently disabled).
2263 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2265 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2268 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2269 [Andy Polyakov and a number of other people]
2271 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2272 implementation contributed by IBM.
2273 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2275 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2276 exponent rather than 'unsigned long'. There is a corresponding change to
2277 the new 'rsa_keygen' element of the RSA_METHOD structure.
2278 [Jelte Jansen, Geoff Thorpe]
2280 *) Functionality for creating the initial serial number file is now
2281 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2283 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2284 number file to 1, which is bound to cause problems. To avoid
2285 the problems while respecting compatibility between different 0.9.7
2286 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2287 CA.pl for serial number initialization. With the new release 0.9.8,
2288 we can fix the problem directly in the 'ca' utility.)
2291 *) Reduced header interdepencies by declaring more opaque objects in
2292 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2293 give fewer recursive includes, which could break lazy source code - so
2294 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2295 developers should define this symbol when building and using openssl to
2296 ensure they track the recommended behaviour, interfaces, [etc], but
2297 backwards-compatible behaviour prevails when this isn't defined.
2300 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2303 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2304 This will generate a random key of the appropriate length based on the
2305 cipher context. The EVP_CIPHER can provide its own random key generation
2306 routine to support keys of a specific form. This is used in the des and
2307 3des routines to generate a key of the correct parity. Update S/MIME
2308 code to use new functions and hence generate correct parity DES keys.
2309 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2310 valid (weak or incorrect parity).
2313 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2314 as looking them up. This is useful when the verified structure may contain
2315 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2316 present unless the new PKCS7_NO_CRL flag is asserted.
2319 *) Extend ASN1 oid configuration module. It now additionally accepts the
2322 shortName = some long name, 1.2.3.4
2325 *) Reimplemented the BN_CTX implementation. There is now no more static
2326 limitation on the number of variables it can handle nor the depth of the
2327 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2328 information can now expand as required, and rather than having a single
2329 static array of bignums, BN_CTX now uses a linked-list of such arrays
2330 allowing it to expand on demand whilst maintaining the usefulness of
2331 BN_CTX's "bundling".
2334 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2335 to allow all RSA operations to function using a single BN_CTX.
2338 *) Preliminary support for certificate policy evaluation and checking. This
2339 is initially intended to pass the tests outlined in "Conformance Testing
2340 of Relying Party Client Certificate Path Processing Logic" v1.07.
2343 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2344 remained unused and not that useful. A variety of other little bignum
2345 tweaks and fixes have also been made continuing on from the audit (see
2349 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2350 associated ASN1, EVP and SSL functions and old ASN1 macros.
2353 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2354 and this should never fail. So the return value from the use of
2355 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2356 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2359 *) BN_CTX_get() should return zero-valued bignums, providing the same
2360 initialised value as BN_new().
2361 [Geoff Thorpe, suggested by Ulf Möller]
2363 *) Support for inhibitAnyPolicy certificate extension.
2366 *) An audit of the BIGNUM code is underway, for which debugging code is
2367 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2368 is considered valid when processing BIGNUMs, and causes execution to
2369 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2370 further steps are taken to deliberately pollute unused data in BIGNUM
2371 structures to try and expose faulty code further on. For now, openssl will
2372 (in its default mode of operation) continue to tolerate the inconsistent
2373 forms that it has tolerated in the past, but authors and packagers should
2374 consider trying openssl and their own applications when compiled with
2375 these debugging symbols defined. It will help highlight potential bugs in
2376 their own code, and will improve the test coverage for OpenSSL itself. At
2377 some point, these tighter rules will become openssl's default to improve
2378 maintainability, though the assert()s and other overheads will remain only
2379 in debugging configurations. See bn.h for more details.
2380 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2382 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2383 that can only be obtained through BN_CTX_new() (which implicitly
2384 initialises it). The presence of this function only made it possible
2385 to overwrite an existing structure (and cause memory leaks).
2388 *) Because of the callback-based approach for implementing LHASH as a
2389 template type, lh_insert() adds opaque objects to hash-tables and
2390 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2391 to clean up those corresponding objects before destroying the hash table
2392 (and losing the object pointers). So some over-zealous constifications in
2393 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2394 objects as "const" and the lh_doall[_arg] callback wrappers are not
2395 prototyped to have "const" restrictions on the object pointers they are
2396 given (and so aren't required to cast them away any more).
2399 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2400 (speed) prefers to use its own implementation. The two implementations
2401 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2402 its object type properly exposed (MS_TM) instead of casting to/from "char
2403 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2404 aren't necessarily the greatest nomenclatures - but this is what was used
2405 internally to the implementation so I've used that for now.
2408 *) Ensure that deprecated functions do not get compiled when
2409 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2410 the self-tests were still using deprecated key-generation functions so
2411 these have been updated also.
2414 *) Reorganise PKCS#7 code to separate the digest location functionality
2415 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2416 New function PKCS7_set_digest() to set the digest type for PKCS#7
2417 digestedData type. Add additional code to correctly generate the
2418 digestedData type and add support for this type in PKCS7 initialization
2422 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2423 structure of type "other".
2426 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2427 sure the loop does correctly stop and breaking ("division by zero")
2428 modulus operations are not performed. The (pre-generated) prime
2429 table crypto/bn/bn_prime.h was already correct, but it could not be
2430 re-generated on some platforms because of the "division by zero"
2431 situation in the script.
2432 [Ralf S. Engelschall]
2434 *) Update support for ECC-based TLS ciphersuites according to
2435 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2436 SHA-1 now is only used for "small" curves (where the
2437 representation of a field element takes up to 24 bytes); for
2438 larger curves, the field element resulting from ECDH is directly
2439 used as premaster secret.
2440 [Douglas Stebila (Sun Microsystems Laboratories)]
2442 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2443 curve secp160r1 to the tests.
2444 [Douglas Stebila (Sun Microsystems Laboratories)]
2446 *) Add the possibility to load symbols globally with DSO.
2447 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2449 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2450 control of the error stack.
2453 *) Add support for STORE in ENGINE.
2456 *) Add the STORE type. The intention is to provide a common interface
2457 to certificate and key stores, be they simple file-based stores, or
2458 HSM-type store, or LDAP stores, or...
2459 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2462 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2463 pass a list of arguments to any function as well as provide a way
2464 for a function to pass data back to the caller.
2467 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2468 works like BUF_strdup() but can be used to duplicate a portion of
2469 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2473 *) Add the function sk_find_ex() which works like sk_find(), but will
2474 return an index to an element even if an exact match couldn't be
2475 found. The index is guaranteed to point at the element where the
2476 searched-for key would be inserted to preserve sorting order.
2479 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2480 takes an extra flags argument for optional functionality. Currently,
2481 the following flags are defined:
2483 OBJ_BSEARCH_VALUE_ON_NOMATCH
2484 This one gets OBJ_bsearch_ex() to return a pointer to the first
2485 element where the comparing function returns a negative or zero
2488 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2489 This one gets OBJ_bsearch_ex() to return a pointer to the first
2490 element where the comparing function returns zero. This is useful
2491 if there are more than one element where the comparing function
2495 *) Make it possible to create self-signed certificates with 'openssl ca'
2496 in such a way that the self-signed certificate becomes part of the
2497 CA database and uses the same mechanisms for serial number generation
2498 as all other certificate signing. The new flag '-selfsign' enables
2499 this functionality. Adapt CA.sh and CA.pl.in.
2502 *) Add functionality to check the public key of a certificate request
2503 against a given private. This is useful to check that a certificate
2504 request can be signed by that key (self-signing).
2507 *) Make it possible to have multiple active certificates with the same
2508 subject in the CA index file. This is done only if the keyword
2509 'unique_subject' is set to 'no' in the main CA section (default
2510 if 'CA_default') of the configuration file. The value is saved
2511 with the database itself in a separate index attribute file,
2512 named like the index file with '.attr' appended to the name.
2515 *) Generate muti valued AVAs using '+' notation in config files for
2519 *) Support for nameConstraints certificate extension.
2522 *) Support for policyConstraints certificate extension.
2525 *) Support for policyMappings certificate extension.
2528 *) Make sure the default DSA_METHOD implementation only uses its
2529 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2530 and change its own handlers to be NULL so as to remove unnecessary
2531 indirection. This lets alternative implementations fallback to the
2532 default implementation more easily.
2535 *) Support for directoryName in GeneralName related extensions
2539 *) Make it possible to link applications using Makefile.shared.
2540 Make that possible even when linking against static libraries!
2543 *) Support for single pass processing for S/MIME signing. This now
2544 means that S/MIME signing can be done from a pipe, in addition
2545 cleartext signing (multipart/signed type) is effectively streaming
2546 and the signed data does not need to be all held in memory.
2548 This is done with a new flag PKCS7_STREAM. When this flag is set
2549 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2550 is done after the data is output (and digests calculated) in
2551 SMIME_write_PKCS7().
2554 *) Add full support for -rpath/-R, both in shared libraries and
2555 applications, at least on the platforms where it's known how
2559 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2560 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2561 will now compute a table of multiples of the generator that
2562 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2563 faster (notably in the case of a single point multiplication,
2564 scalar * generator).
2565 [Nils Larsch, Bodo Moeller]
2567 *) IPv6 support for certificate extensions. The various extensions
2568 which use the IP:a.b.c.d can now take IPv6 addresses using the
2569 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2573 *) Added an ENGINE that implements RSA by performing private key
2574 exponentiations with the GMP library. The conversions to and from
2575 GMP's mpz_t format aren't optimised nor are any montgomery forms
2576 cached, and on x86 it appears OpenSSL's own performance has caught up.
2577 However there are likely to be other architectures where GMP could
2578 provide a boost. This ENGINE is not built in by default, but it can be
2579 specified at Configure time and should be accompanied by the necessary
2580 linker additions, eg;
2581 ./config -DOPENSSL_USE_GMP -lgmp
2584 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2585 testing availability of engines with "-t" - the old behaviour is
2586 produced by increasing the feature's verbosity with "-tt".
2589 *) ECDSA routines: under certain error conditions uninitialized BN objects
2590 could be freed. Solution: make sure initialization is performed early
2591 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2595 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2596 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2597 software implementations. For DSA and DH, parameter generation can
2598 also be overriden by providing the appropriate method callbacks.
2601 *) Change the "progress" mechanism used in key-generation and
2602 primality testing to functions that take a new BN_GENCB pointer in
2603 place of callback/argument pairs. The new API functions have "_ex"
2604 postfixes and the older functions are reimplemented as wrappers for
2605 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2606 declarations of the old functions to help (graceful) attempts to
2607 migrate to the new functions. Also, the new key-generation API
2608 functions operate on a caller-supplied key-structure and return
2609 success/failure rather than returning a key or NULL - this is to
2610 help make "keygen" another member function of RSA_METHOD etc.
2612 Example for using the new callback interface:
2614 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2618 BN_GENCB_set(&my_cb, my_callback, my_arg);
2620 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2621 /* For the meaning of a, b in calls to my_callback(), see the
2622 * documentation of the function that calls the callback.
2623 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2624 * my_callback should return 1 if it wants BN_is_prime_ex()
2625 * to continue, or 0 to stop.
2630 *) Change the ZLIB compression method to be stateful, and make it
2631 available to TLS with the number defined in
2632 draft-ietf-tls-compression-04.txt.
2635 *) Add the ASN.1 structures and functions for CertificatePair, which
2636 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2638 CertificatePair ::= SEQUENCE {
2639 forward [0] Certificate OPTIONAL,
2640 reverse [1] Certificate OPTIONAL,
2641 -- at least one of the pair shall be present -- }
2643 Also implement the PEM functions to read and write certificate
2644 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2646 This needed to be defined, mostly for the sake of the LDAP
2647 attribute crossCertificatePair, but may prove useful elsewhere as
2651 *) Make it possible to inhibit symlinking of shared libraries in
2652 Makefile.shared, for Cygwin's sake.
2655 *) Extend the BIGNUM API by creating a function
2656 void BN_set_negative(BIGNUM *a, int neg);
2657 and a macro that behave like
2658 int BN_is_negative(const BIGNUM *a);
2660 to avoid the need to access 'a->neg' directly in applications.
2663 *) Implement fast modular reduction for pseudo-Mersenne primes
2664 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2665 EC_GROUP_new_curve_GFp() will now automatically use this
2667 [Nils Larsch <nla@trustcenter.de>]
2669 *) Add new lock type (CRYPTO_LOCK_BN).
2672 *) Change the ENGINE framework to automatically load engines
2673 dynamically from specific directories unless they could be
2674 found to already be built in or loaded. Move all the
2675 current engines except for the cryptodev one to a new
2677 The engines in engines/ are built as shared libraries if
2678 the "shared" options was given to ./Configure or ./config.
2679 Otherwise, they are inserted in libcrypto.a.
2680 /usr/local/ssl/engines is the default directory for dynamic
2681 engines, but that can be overriden at configure time through
2682 the usual use of --prefix and/or --openssldir, and at run
2683 time with the environment variable OPENSSL_ENGINES.
2684 [Geoff Thorpe and Richard Levitte]
2686 *) Add Makefile.shared, a helper makefile to build shared
2687 libraries. Addapt Makefile.org.
2690 *) Add version info to Win32 DLLs.
2691 [Peter 'Luna' Runestig" <peter@runestig.com>]
2693 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2694 can be added using this API to created arbitrary PKCS#12
2695 files while avoiding the low level API.
2697 New options to PKCS12_create(), key or cert can be NULL and
2698 will then be omitted from the output file. The encryption
2699 algorithm NIDs can be set to -1 for no encryption, the mac
2700 iteration count can be set to 0 to omit the mac.
2702 Enhance pkcs12 utility by making the -nokeys and -nocerts
2703 options work when creating a PKCS#12 file. New option -nomac
2704 to omit the mac, NONE can be set for an encryption algorithm.
2705 New code is modified to use the enhanced PKCS12_create()
2706 instead of the low level API.
2709 *) Extend ASN1 encoder to support indefinite length constructed
2710 encoding. This can output sequences tags and octet strings in
2711 this form. Modify pk7_asn1.c to support indefinite length
2712 encoding. This is experimental and needs additional code to
2713 be useful, such as an ASN1 bio and some enhanced streaming
2716 Extend template encode functionality so that tagging is passed
2717 down to the template encoder.
2720 *) Let 'openssl req' fail if an argument to '-newkey' is not
2721 recognized instead of using RSA as a default.
2724 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2725 As these are not official, they are not included in "ALL";
2726 the "ECCdraft" ciphersuite group alias can be used to select them.
2727 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2729 *) Add ECDH engine support.
2730 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2732 *) Add ECDH in new directory crypto/ecdh/.
2733 [Douglas Stebila (Sun Microsystems Laboratories)]
2735 *) Let BN_rand_range() abort with an error after 100 iterations
2736 without success (which indicates a broken PRNG).
2739 *) Change BN_mod_sqrt() so that it verifies that the input value
2740 is really the square of the return value. (Previously,
2741 BN_mod_sqrt would show GIGO behaviour.)
2744 *) Add named elliptic curves over binary fields from X9.62, SECG,
2745 and WAP/WTLS; add OIDs that were still missing.
2747 [Sheueling Chang Shantz and Douglas Stebila
2748 (Sun Microsystems Laboratories)]
2750 *) Extend the EC library for elliptic curves over binary fields
2751 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2754 EC_GF2m_simple_method
2758 EC_GROUP_new_curve_GF2m
2759 EC_GROUP_set_curve_GF2m
2760 EC_GROUP_get_curve_GF2m
2761 EC_POINT_set_affine_coordinates_GF2m
2762 EC_POINT_get_affine_coordinates_GF2m
2763 EC_POINT_set_compressed_coordinates_GF2m
2765 Point compression for binary fields is disabled by default for
2766 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2769 As binary polynomials are represented as BIGNUMs, various members
2770 of the EC_GROUP and EC_POINT data structures can be shared
2771 between the implementations for prime fields and binary fields;
2772 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2773 are essentially identical to their ..._GFp counterparts.
2774 (For simplicity, the '..._GFp' prefix has been dropped from
2775 various internal method names.)
2777 An internal 'field_div' method (similar to 'field_mul' and
2778 'field_sqr') has been added; this is used only for binary fields.
2780 [Sheueling Chang Shantz and Douglas Stebila
2781 (Sun Microsystems Laboratories)]
2783 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2784 through methods ('mul', 'precompute_mult').
2786 The generic implementations (now internally called 'ec_wNAF_mul'
2787 and 'ec_wNAF_precomputed_mult') remain the default if these
2788 methods are undefined.
2790 [Sheueling Chang Shantz and Douglas Stebila
2791 (Sun Microsystems Laboratories)]
2793 *) New function EC_GROUP_get_degree, which is defined through
2794 EC_METHOD. For curves over prime fields, this returns the bit
2795 length of the modulus.
2797 [Sheueling Chang Shantz and Douglas Stebila
2798 (Sun Microsystems Laboratories)]
2800 *) New functions EC_GROUP_dup, EC_POINT_dup.
2801 (These simply call ..._new and ..._copy).
2803 [Sheueling Chang Shantz and Douglas Stebila
2804 (Sun Microsystems Laboratories)]
2806 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2807 Polynomials are represented as BIGNUMs (where the sign bit is not
2808 used) in the following functions [macros]:
2811 BN_GF2m_sub [= BN_GF2m_add]
2812 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2813 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2814 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2816 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2817 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2818 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2819 BN_GF2m_cmp [= BN_ucmp]
2821 (Note that only the 'mod' functions are actually for fields GF(2^m).
2822 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2824 For some functions, an the irreducible polynomial defining a
2825 field can be given as an 'unsigned int[]' with strictly
2826 decreasing elements giving the indices of those bits that are set;
2827 i.e., p[] represents the polynomial
2828 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2830 p[0] > p[1] > ... > p[k] = 0.
2831 This applies to the following functions:
2836 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2837 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2839 BN_GF2m_mod_sqrt_arr
2840 BN_GF2m_mod_solve_quad_arr
2844 Conversion can be performed by the following functions:
2849 bntest.c has additional tests for binary polynomial arithmetic.
2851 Two implementations for BN_GF2m_mod_div() are available.
2852 The default algorithm simply uses BN_GF2m_mod_inv() and
2853 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2854 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2855 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2857 [Sheueling Chang Shantz and Douglas Stebila
2858 (Sun Microsystems Laboratories)]
2860 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2861 functionality is disabled at compile-time.
2862 [Douglas Stebila <douglas.stebila@sun.com>]
2864 *) Change default behaviour of 'openssl asn1parse' so that more
2865 information is visible when viewing, e.g., a certificate:
2867 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2868 mode the content of non-printable OCTET STRINGs is output in a
2869 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2870 avoid the appearance of a printable string.
2871 [Nils Larsch <nla@trustcenter.de>]
2873 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2875 EC_GROUP_set_asn1_flag()
2876 EC_GROUP_get_asn1_flag()
2877 EC_GROUP_set_point_conversion_form()
2878 EC_GROUP_get_point_conversion_form()
2879 These control ASN1 encoding details:
2880 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2881 has been set to OPENSSL_EC_NAMED_CURVE.
2882 - Points are encoded in uncompressed form by default; options for
2883 asn1_for are as for point2oct, namely
2884 POINT_CONVERSION_COMPRESSED
2885 POINT_CONVERSION_UNCOMPRESSED
2886 POINT_CONVERSION_HYBRID
2888 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2891 EC_GROUP_get0_seed()
2892 EC_GROUP_get_seed_len()
2893 This is used only for ASN1 purposes (so far).
2894 [Nils Larsch <nla@trustcenter.de>]
2896 *) Add 'field_type' member to EC_METHOD, which holds the NID
2897 of the appropriate field type OID. The new function
2898 EC_METHOD_get_field_type() returns this value.
2899 [Nils Larsch <nla@trustcenter.de>]
2904 EC_POINT_point2hex()
2905 EC_POINT_hex2point()
2906 providing useful interfaces to EC_POINT_point2oct() and
2907 EC_POINT_oct2point().
2908 [Nils Larsch <nla@trustcenter.de>]
2910 *) Change internals of the EC library so that the functions
2911 EC_GROUP_set_generator()
2912 EC_GROUP_get_generator()
2913 EC_GROUP_get_order()
2914 EC_GROUP_get_cofactor()
2915 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2916 to methods, which would lead to unnecessary code duplication when
2917 adding different types of curves.
2918 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2920 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2921 arithmetic, and such that modified wNAFs are generated
2922 (which avoid length expansion in many cases).
2925 *) Add a function EC_GROUP_check_discriminant() (defined via
2926 EC_METHOD) that verifies that the curve discriminant is non-zero.
2928 Add a function EC_GROUP_check() that makes some sanity tests
2929 on a EC_GROUP, its generator and order. This includes
2930 EC_GROUP_check_discriminant().
2931 [Nils Larsch <nla@trustcenter.de>]
2933 *) Add ECDSA in new directory crypto/ecdsa/.
2935 Add applications 'openssl ecparam' and 'openssl ecdsa'
2936 (these are based on 'openssl dsaparam' and 'openssl dsa').
2938 ECDSA support is also included in various other files across the
2939 library. Most notably,
2940 - 'openssl req' now has a '-newkey ecdsa:file' option;
2941 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2942 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2943 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2944 them suitable for ECDSA where domain parameters must be
2945 extracted before the specific public key;
2946 - ECDSA engine support has been added.
2947 [Nils Larsch <nla@trustcenter.de>]
2949 *) Include some named elliptic curves, and add OIDs from X9.62,
2950 SECG, and WAP/WTLS. Each curve can be obtained from the new
2952 EC_GROUP_new_by_curve_name(),
2953 and the list of available named curves can be obtained with
2954 EC_get_builtin_curves().
2955 Also add a 'curve_name' member to EC_GROUP objects, which can be
2957 EC_GROUP_set_curve_name()
2958 EC_GROUP_get_curve_name()
2959 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2961 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2962 was actually never needed) and in BN_mul(). The removal in BN_mul()
2963 required a small change in bn_mul_part_recursive() and the addition
2964 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2965 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2966 bn_sub_words() and bn_add_words() except they take arrays with
2970 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2972 *) Cleanse PEM buffers before freeing them since they may contain
2974 [Benjamin Bennett <ben@psc.edu>]
2976 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2977 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2978 authentication-only ciphersuites.
2981 *) Since AES128 and AES256 share a single mask bit in the logic of
2982 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2983 kludge to work properly if AES128 is available and AES256 isn't.
2986 *) Expand security boundary to match 1.1.1 module.
2989 *) Remove redundant features: hash file source, editing of test vectors
2990 modify fipsld to use external fips_premain.c signature.
2993 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2994 run algorithm test programs.
2997 *) Make algorithm test programs more tolerant of whitespace.
3000 *) Have SSL/TLS server implementation tolerate "mismatched" record
3001 protocol version while receiving ClientHello even if the
3002 ClientHello is fragmented. (The server can't insist on the
3003 particular protocol version it has chosen before the ServerHello
3004 message has informed the client about his choice.)
3007 *) Load error codes if they are not already present instead of using a
3008 static variable. This allows them to be cleanly unloaded and reloaded.
3011 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3013 *) Introduce limits to prevent malicious keys being able to
3014 cause a denial of service. (CVE-2006-2940)
3015 [Steve Henson, Bodo Moeller]
3017 *) Fix ASN.1 parsing of certain invalid structures that can result
3018 in a denial of service. (CVE-2006-2937) [Steve Henson]
3020 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3021 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3023 *) Fix SSL client code which could crash if connecting to a
3024 malicious SSLv2 server. (CVE-2006-4343)
3025 [Tavis Ormandy and Will Drewry, Google Security Team]
3027 *) Change ciphersuite string processing so that an explicit
3028 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3029 will no longer include "AES128-SHA"), and any other similar
3030 ciphersuite (same bitmap) from *other* protocol versions (so that
3031 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3032 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3033 changes from 0.9.8b and 0.9.8d.
3036 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3038 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3039 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3041 *) Change the Unix randomness entropy gathering to use poll() when
3042 possible instead of select(), since the latter has some
3043 undesirable limitations.
3044 [Darryl Miles via Richard Levitte and Bodo Moeller]
3046 *) Disable rogue ciphersuites:
3048 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3049 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3050 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3052 The latter two were purportedly from
3053 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3056 Also deactive the remaining ciphersuites from
3057 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3058 unofficial, and the ID has long expired.
3061 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3062 dual-core machines) and other potential thread-safety issues.
3065 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3067 *) Adapt fipsld and the build system to link against the validated FIPS
3068 module in FIPS mode.
3071 *) Fixes for VC++ 2005 build under Windows.
3074 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3075 from a Windows bash shell such as MSYS. It is autodetected from the
3076 "config" script when run from a VC++ environment. Modify standard VC++
3077 build to use fipscanister.o from the GNU make build.
3080 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3082 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3083 The value now differs depending on if you build for FIPS or not.
3084 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3085 safely run with a non-FIPSed libcrypto, as it may crash because of
3086 the difference induced by this change.
3089 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3091 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3092 (part of SSL_OP_ALL). This option used to disable the
3093 countermeasure against man-in-the-middle protocol-version
3094 rollback in the SSL 2.0 server implementation, which is a bad
3095 idea. (CVE-2005-2969)
3097 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3098 for Information Security, National Institute of Advanced Industrial
3099 Science and Technology [AIST], Japan)]
3101 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3102 mainly for FIPS compliance and not fully integrated at this stage.
3105 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3106 the exponentiation using a fixed-length exponent. (Otherwise,
3107 the information leaked through timing could expose the secret key
3108 after many signatures; cf. Bleichenbacher's attack on DSA with
3112 *) Make a new fixed-window mod_exp implementation the default for
3113 RSA, DSA, and DH private-key operations so that the sequence of
3114 squares and multiplies and the memory access pattern are
3115 independent of the particular secret key. This will mitigate
3116 cache-timing and potential related attacks.
3118 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3119 and this is automatically used by BN_mod_exp_mont() if the new flag
3120 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3121 will use this BN flag for private exponents unless the flag
3122 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3123 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3125 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3127 *) Change the client implementation for SSLv23_method() and
3128 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3129 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3130 (Previously, the SSL 2.0 backwards compatible Client Hello
3131 message format would be used even with SSL_OP_NO_SSLv2.)
3134 *) Add support for smime-type MIME parameter in S/MIME messages which some
3138 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3139 a threadsafe manner. Modify rsa code to use new function and add calls
3140 to dsa and dh code (which had race conditions before).
3143 *) Include the fixed error library code in the C error file definitions
3144 instead of fixing them up at runtime. This keeps the error code
3145 structures constant.
3148 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3150 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3153 *) Fixes for newer kerberos headers. NB: the casts are needed because
3154 the 'length' field is signed on one version and unsigned on another
3155 with no (?) obvious way to tell the difference, without these VC++
3156 complains. Also the "definition" of FAR (blank) is no longer included
3157 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3158 some needed definitions.
3161 *) Undo Cygwin change.
3164 *) Added support for proxy certificates according to RFC 3820.
3165 Because they may be a security thread to unaware applications,
3166 they must be explicitely allowed in run-time. See
3167 docs/HOWTO/proxy_certificates.txt for further information.
3170 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3172 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3173 server and client random values. Previously
3174 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3175 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3177 This change has negligible security impact because:
3179 1. Server and client random values still have 24 bytes of pseudo random
3182 2. Server and client random values are sent in the clear in the initial
3185 3. The master secret is derived using the premaster secret (48 bytes in
3186 size for static RSA ciphersuites) as well as client server and random
3189 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3192 [Stephen Henson, reported by UK NISCC]
3194 *) Use Windows randomness collection on Cygwin.
3197 *) Fix hang in EGD/PRNGD query when communication socket is closed
3198 prematurely by EGD/PRNGD.
3199 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3201 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3204 *) Back-port of selected performance improvements from development
3205 branch, as well as improved support for PowerPC platforms.
3208 *) Add lots of checks for memory allocation failure, error codes to indicate
3209 failure and freeing up memory if a failure occurs.
3210 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3212 *) Add new -passin argument to dgst.
3215 *) Perform some character comparisons of different types in X509_NAME_cmp:
3216 this is needed for some certificates that reencode DNs into UTF8Strings
3217 (in violation of RFC3280) and can't or wont issue name rollover
3221 *) Make an explicit check during certificate validation to see that
3222 the CA setting in each certificate on the chain is correct. As a
3223 side effect always do the following basic checks on extensions,
3224 not just when there's an associated purpose to the check:
3226 - if there is an unhandled critical extension (unless the user
3227 has chosen to ignore this fault)
3228 - if the path length has been exceeded (if one is set at all)
3229 - that certain extensions fit the associated purpose (if one has
3233 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3235 *) Avoid a race condition when CRLs are checked in a multi threaded
3236 environment. This would happen due to the reordering of the revoked
3237 entries during signature checking and serial number lookup. Now the
3238 encoding is cached and the serial number sort performed under a lock.
3239 Add new STACK function sk_is_sorted().
3242 *) Add Delta CRL to the extension code.
3245 *) Various fixes to s3_pkt.c so alerts are sent properly.
3246 [David Holmes <d.holmes@f5.com>]
3248 *) Reduce the chances of duplicate issuer name and serial numbers (in
3249 violation of RFC3280) using the OpenSSL certificate creation utilities.
3250 This is done by creating a random 64 bit value for the initial serial
3251 number when a serial number file is created or when a self signed
3252 certificate is created using 'openssl req -x509'. The initial serial
3253 number file is created using 'openssl x509 -next_serial' in CA.pl
3254 rather than being initialized to 1.
3257 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3259 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3260 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3261 [Joe Orton, Steve Henson]
3263 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3265 [Joe Orton, Steve Henson]
3267 *) Make it possible to have multiple active certificates with the same
3268 subject in the CA index file. This is done only if the keyword
3269 'unique_subject' is set to 'no' in the main CA section (default
3270 if 'CA_default') of the configuration file. The value is saved
3271 with the database itself in a separate index attribute file,
3272 named like the index file with '.attr' appended to the name.
3275 *) X509 verify fixes. Disable broken certificate workarounds when
3276 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3277 keyUsage extension present. Don't accept CRLs with unhandled critical
3278 extensions: since verify currently doesn't process CRL extensions this
3279 rejects a CRL with *any* critical extensions. Add new verify error codes
3283 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3284 A clarification of RFC2560 will require the use of OCTET STRINGs and
3285 some implementations cannot handle the current raw format. Since OpenSSL
3286 copies and compares OCSP nonces as opaque blobs without any attempt at
3287 parsing them this should not create any compatibility issues.
3290 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3291 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3292 this HMAC (and other) operations are several times slower than OpenSSL
3296 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3297 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3299 *) Use the correct content when signing type "other".
3302 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3304 *) Fix various bugs revealed by running the NISCC test suite:
3306 Stop out of bounds reads in the ASN1 code when presented with
3307 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3309 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3311 If verify callback ignores invalid public key errors don't try to check
3312 certificate signature with the NULL public key.
3316 *) New -ignore_err option in ocsp application to stop the server
3317 exiting on the first error in a request.
3320 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3321 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3325 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3326 extra data after the compression methods not only for TLS 1.0
3327 but also for SSL 3.0 (as required by the specification).
3328 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3330 *) Change X509_certificate_type() to mark the key as exported/exportable
3331 when it's 512 *bits* long, not 512 bytes.
3334 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3335 blocks during encryption.
3338 *) Various fixes to base64 BIO and non blocking I/O. On write
3339 flushes were not handled properly if the BIO retried. On read
3340 data was not being buffered properly and had various logic bugs.
3341 This also affects blocking I/O when the data being decoded is a
3345 *) Various S/MIME bugfixes and compatibility changes:
3346 output correct application/pkcs7 MIME type if
3347 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3348 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3349 of files as .eml work). Correctly handle very long lines in MIME
3353 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3355 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3356 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3357 a protocol version number mismatch like a decryption error
3358 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3361 *) Turn on RSA blinding by default in the default implementation
3362 to avoid a timing attack. Applications that don't want it can call
3363 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3364 They would be ill-advised to do so in most cases.
3365 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3367 *) Change RSA blinding code so that it works when the PRNG is not
3368 seeded (in this case, the secret RSA exponent is abused as
3369 an unpredictable seed -- if it is not unpredictable, there
3370 is no point in blinding anyway). Make RSA blinding thread-safe
3371 by remembering the creator's thread ID in rsa->blinding and
3372 having all other threads use local one-time blinding factors
3373 (this requires more computation than sharing rsa->blinding, but
3374 avoids excessive locking; and if an RSA object is not shared
3375 between threads, blinding will still be very fast).
3378 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3379 ENGINE as defaults for all supported algorithms irrespective of
3380 the 'flags' parameter. 'flags' is now honoured, so applications
3381 should make sure they are passing it correctly.
3384 *) Target "mingw" now allows native Windows code to be generated in
3385 the Cygwin environment as well as with the MinGW compiler.
3388 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3390 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3391 via timing by performing a MAC computation even if incorrrect
3392 block cipher padding has been found. This is a countermeasure
3393 against active attacks where the attacker has to distinguish
3394 between bad padding and a MAC verification error. (CVE-2003-0078)
3396 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3397 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3398 Martin Vuagnoux (EPFL, Ilion)]
3400 *) Make the no-err option work as intended. The intention with no-err
3401 is not to have the whole error stack handling routines removed from
3402 libcrypto, it's only intended to remove all the function name and
3403 reason texts, thereby removing some of the footprint that may not
3404 be interesting if those errors aren't displayed anyway.
3406 NOTE: it's still possible for any application or module to have it's
3407 own set of error texts inserted. The routines are there, just not
3408 used by default when no-err is given.
3411 *) Add support for FreeBSD on IA64.
3412 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3414 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3415 Kerberos function mit_des_cbc_cksum(). Before this change,
3416 the value returned by DES_cbc_cksum() was like the one from
3417 mit_des_cbc_cksum(), except the bytes were swapped.
3418 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3420 *) Allow an application to disable the automatic SSL chain building.
3421 Before this a rather primitive chain build was always performed in
3422 ssl3_output_cert_chain(): an application had no way to send the
3423 correct chain if the automatic operation produced an incorrect result.
3425 Now the chain builder is disabled if either:
3427 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3429 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3431 The reasoning behind this is that an application would not want the
3432 auto chain building to take place if extra chain certificates are
3433 present and it might also want a means of sending no additional
3434 certificates (for example the chain has two certificates and the
3438 *) Add the possibility to build without the ENGINE framework.
3439 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3441 *) Under Win32 gmtime() can return NULL: check return value in
3442 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3445 *) DSA routines: under certain error conditions uninitialized BN objects
3446 could be freed. Solution: make sure initialization is performed early
3447 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3448 Nils Larsch <nla@trustcenter.de> via PR#459)
3451 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3452 checked on reconnect on the client side, therefore session resumption
3453 could still fail with a "ssl session id is different" error. This
3454 behaviour is masked when SSL_OP_ALL is used due to
3455 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3456 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3457 followup to PR #377.
3460 *) IA-32 assembler support enhancements: unified ELF targets, support
3461 for SCO/Caldera platforms, fix for Cygwin shared build.
3464 *) Add support for FreeBSD on sparc64. As a consequence, support for
3465 FreeBSD on non-x86 processors is separate from x86 processors on
3466 the config script, much like the NetBSD support.
3467 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3469 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3471 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3474 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3475 code (06) was taken as the first octet of the session ID and the last
3476 octet was ignored consequently. As a result SSLv2 client side session
3477 caching could not have worked due to the session ID mismatch between
3479 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3483 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3484 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3488 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3489 seems that in spite of existing for more than a year, many application
3490 author have done nothing to provide the necessary callbacks, which
3491 means that this particular engine will not work properly anywhere.
3492 This is a very unfortunate situation which forces us, in the name
3493 of usability, to give the hw_ncipher.c a static lock, which is part
3495 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3496 appear in 0.9.8 or later. We EXPECT application authors to have
3497 dealt properly with this when 0.9.8 is released (unless we actually
3498 make such changes in the libcrypto locking code that changes will
3499 have to be made anyway).
3502 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3503 octets have been read, EOF or an error occurs. Without this change
3504 some truncated ASN1 structures will not produce an error.
3507 *) Disable Heimdal support, since it hasn't been fully implemented.
3508 Still give the possibility to force the use of Heimdal, but with
3509 warnings and a request that patches get sent to openssl-dev.
3512 *) Add the VC-CE target, introduce the WINCE sysname, and add
3513 INSTALL.WCE and appropriate conditionals to make it build.
3514 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3516 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3517 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3518 edit numbers of the version.
3519 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3521 *) Introduce safe string copy and catenation functions
3522 (BUF_strlcpy() and BUF_strlcat()).
3523 [Ben Laurie (CHATS) and Richard Levitte]
3525 *) Avoid using fixed-size buffers for one-line DNs.
3526 [Ben Laurie (CHATS)]
3528 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3529 resizing buffers containing secrets, and use where appropriate.
3530 [Ben Laurie (CHATS)]
3532 *) Avoid using fixed size buffers for configuration file location.
3533 [Ben Laurie (CHATS)]
3535 *) Avoid filename truncation for various CA files.
3536 [Ben Laurie (CHATS)]
3538 *) Use sizeof in preference to magic numbers.
3539 [Ben Laurie (CHATS)]
3541 *) Avoid filename truncation in cert requests.
3542 [Ben Laurie (CHATS)]
3544 *) Add assertions to check for (supposedly impossible) buffer
3546 [Ben Laurie (CHATS)]
3548 *) Don't cache truncated DNS entries in the local cache (this could
3549 potentially lead to a spoofing attack).
3550 [Ben Laurie (CHATS)]
3552 *) Fix various buffers to be large enough for hex/decimal
3553 representations in a platform independent manner.
3554 [Ben Laurie (CHATS)]
3556 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3557 resizing buffers containing secrets, and use where appropriate.
3558 [Ben Laurie (CHATS)]
3560 *) Add BIO_indent() to avoid much slightly worrying code to do
3562 [Ben Laurie (CHATS)]
3564 *) Convert sprintf()/BIO_puts() to BIO_printf().
3565 [Ben Laurie (CHATS)]
3567 *) buffer_gets() could terminate with the buffer only half
3569 [Ben Laurie (CHATS)]
3571 *) Add assertions to prevent user-supplied crypto functions from
3572 overflowing internal buffers by having large block sizes, etc.
3573 [Ben Laurie (CHATS)]
3575 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3577 [Ben Laurie (CHATS)]
3579 *) Eliminate unused copy of key in RC4.
3580 [Ben Laurie (CHATS)]
3582 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3583 [Ben Laurie (CHATS)]
3585 *) Fix off-by-one error in EGD path.
3586 [Ben Laurie (CHATS)]
3588 *) If RANDFILE path is too long, ignore instead of truncating.
3589 [Ben Laurie (CHATS)]
3591 *) Eliminate unused and incorrectly sized X.509 structure
3593 [Ben Laurie (CHATS)]
3595 *) Eliminate unused and dangerous function knumber().
3596 [Ben Laurie (CHATS)]
3598 *) Eliminate unused and dangerous structure, KSSL_ERR.
3599 [Ben Laurie (CHATS)]
3601 *) Protect against overlong session ID context length in an encoded
3602 session object. Since these are local, this does not appear to be
3604 [Ben Laurie (CHATS)]
3606 *) Change from security patch (see 0.9.6e below) that did not affect
3607 the 0.9.6 release series:
3609 Remote buffer overflow in SSL3 protocol - an attacker could
3610 supply an oversized master key in Kerberos-enabled versions.
3612 [Ben Laurie (CHATS)]
3614 *) Change the SSL kerb5 codes to match RFC 2712.
3617 *) Make -nameopt work fully for req and add -reqopt switch.
3618 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3620 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3621 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3623 *) Make sure tests can be performed even if the corresponding algorithms
3624 have been removed entirely. This was also the last step to make
3625 OpenSSL compilable with DJGPP under all reasonable conditions.
3626 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3628 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3629 to allow version independent disabling of normally unselected ciphers,
3630 which may be activated as a side-effect of selecting a single cipher.
3632 (E.g., cipher list string "RSA" enables ciphersuites that are left
3633 out of "ALL" because they do not provide symmetric encryption.
3634 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3635 [Lutz Jaenicke, Bodo Moeller]
3637 *) Add appropriate support for separate platform-dependent build
3638 directories. The recommended way to make a platform-dependent
3639 build directory is the following (tested on Linux), maybe with
3642 # Place yourself outside of the OpenSSL source tree. In
3643 # this example, the environment variable OPENSSL_SOURCE
3644 # is assumed to contain the absolute OpenSSL source directory.
3645 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3646 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3647 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3648 mkdir -p `dirname $F`
3649 ln -s $OPENSSL_SOURCE/$F $F
3652 To be absolutely sure not to disturb the source tree, a "make clean"
3653 is a good thing. If it isn't successfull, don't worry about it,
3654 it probably means the source directory is very clean.
3657 *) Make sure any ENGINE control commands make local copies of string
3658 pointers passed to them whenever necessary. Otherwise it is possible
3659 the caller may have overwritten (or deallocated) the original string
3660 data when a later ENGINE operation tries to use the stored values.
3661 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3663 *) Improve diagnostics in file reading and command-line digests.
3664 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3666 *) Add AES modes CFB and OFB to the object database. Correct an
3667 error in AES-CFB decryption.
3670 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3671 allows existing EVP_CIPHER_CTX structures to be reused after
3672 calling EVP_*Final(). This behaviour is used by encryption
3673 BIOs and some applications. This has the side effect that
3674 applications must explicitly clean up cipher contexts with
3675 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3678 *) Check the values of dna and dnb in bn_mul_recursive before calling
3679 bn_mul_comba (a non zero value means the a or b arrays do not contain
3680 n2 elements) and fallback to bn_mul_normal if either is not zero.
3683 *) Fix escaping of non-ASCII characters when using the -subj option
3684 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3687 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3688 form for "surname", serialNumber has no short form.
3689 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3690 therefore remove "mail" short name for "internet 7".
3691 The OID for unique identifiers in X509 certificates is
3692 x500UniqueIdentifier, not uniqueIdentifier.
3693 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3696 *) Add an "init" command to the ENGINE config module and auto initialize
3697 ENGINEs. Without any "init" command the ENGINE will be initialized
3698 after all ctrl commands have been executed on it. If init=1 the
3699 ENGINE is initailized at that point (ctrls before that point are run
3700 on the uninitialized ENGINE and after on the initialized one). If
3701 init=0 then the ENGINE will not be iniatialized at all.
3704 *) Fix the 'app_verify_callback' interface so that the user-defined
3705 argument is actually passed to the callback: In the
3706 SSL_CTX_set_cert_verify_callback() prototype, the callback
3707 declaration has been changed from
3710 int (*cb)(X509_STORE_CTX *,void *);
3711 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3712 i=s->ctx->app_verify_callback(&ctx)
3713 has been changed into
3714 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3716 To update applications using SSL_CTX_set_cert_verify_callback(),
3717 a dummy argument can be added to their callback functions.
3718 [D. K. Smetters <smetters@parc.xerox.com>]
3720 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3721 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3723 *) Add and OPENSSL_LOAD_CONF define which will cause
3724 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3725 This allows older applications to transparently support certain
3726 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3727 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3728 load the config file and OPENSSL_add_all_algorithms_conf() which will
3729 always load it have also been added.
3732 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3733 Adjust NIDs and EVP layer.
3734 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3736 *) Config modules support in openssl utility.
3738 Most commands now load modules from the config file,
3739 though in a few (such as version) this isn't done
3740 because it couldn't be used for anything.
3742 In the case of ca and req the config file used is
3743 the same as the utility itself: that is the -config
3744 command line option can be used to specify an
3748 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3749 use "openssl_conf" if filename is NULL use default openssl config file.
3752 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3753 config section name. Add a new flag to tolerate a missing config file
3754 and move code to CONF_modules_load_file().
3757 *) Support for crypto accelerator cards from Accelerated Encryption
3758 Processing, www.aep.ie. (Use engine 'aep')
3759 The support was copied from 0.9.6c [engine] and adapted/corrected
3760 to work with the new engine framework.
3761 [AEP Inc. and Richard Levitte]
3763 *) Support for SureWare crypto accelerator cards from Baltimore
3764 Technologies. (Use engine 'sureware')
3765 The support was copied from 0.9.6c [engine] and adapted
3766 to work with the new engine framework.
3769 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3770 make the newer ENGINE framework commands for the CHIL engine work.
3771 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3773 *) Make it possible to produce shared libraries on ReliantUNIX.
3774 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3776 *) Add the configuration target debug-linux-ppro.
3777 Make 'openssl rsa' use the general key loading routines
3778 implemented in apps.c, and make those routines able to
3779 handle the key format FORMAT_NETSCAPE and the variant
3781 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3783 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3784 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3786 *) Add -keyform to rsautl, and document -engine.
3787 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3789 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3790 BIO_R_NO_SUCH_FILE error code rather than the generic
3791 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3794 *) Add new functions
3796 ERR_peek_last_error_line
3797 ERR_peek_last_error_line_data.
3798 These are similar to
3801 ERR_peek_error_line_data,
3802 but report on the latest error recorded rather than the first one
3803 still in the error queue.
3804 [Ben Laurie, Bodo Moeller]
3806 *) default_algorithms option in ENGINE config module. This allows things
3808 default_algorithms = ALL
3809 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3812 *) Prelminary ENGINE config module.
3815 *) New experimental application configuration code.
3818 *) Change the AES code to follow the same name structure as all other
3819 symmetric ciphers, and behave the same way. Move everything to
3820 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3821 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3823 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3824 [Ben Laurie and Theo de Raadt]
3826 *) Add option to output public keys in req command.
3827 [Massimiliano Pala madwolf@openca.org]
3829 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3830 (up to about 10% better than before for P-192 and P-224).
3833 *) New functions/macros
3835 SSL_CTX_set_msg_callback(ctx, cb)
3836 SSL_CTX_set_msg_callback_arg(ctx, arg)
3837 SSL_set_msg_callback(ssl, cb)
3838 SSL_set_msg_callback_arg(ssl, arg)
3840 to request calling a callback function
3842 void cb(int write_p, int version, int content_type,
3843 const void *buf, size_t len, SSL *ssl, void *arg)
3845 whenever a protocol message has been completely received
3846 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3847 protocol version according to which the SSL library interprets
3848 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3849 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3850 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3851 specification (change_cipher_spec(20), alert(21), handshake(22)).
3852 'buf' and 'len' point to the actual message, 'ssl' to the
3853 SSL object, and 'arg' is the application-defined value set by
3854 SSL[_CTX]_set_msg_callback_arg().
3856 'openssl s_client' and 'openssl s_server' have new '-msg' options
3857 to enable a callback that displays all protocol messages.
3860 *) Change the shared library support so shared libraries are built as
3861 soon as the corresponding static library is finished, and thereby get
3862 openssl and the test programs linked against the shared library.
3863 This still only happens when the keyword "shard" has been given to
3864 the configuration scripts.
3866 NOTE: shared library support is still an experimental thing, and
3867 backward binary compatibility is still not guaranteed.
3868 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3870 *) Add support for Subject Information Access extension.
3871 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3873 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3874 additional bytes when new memory had to be allocated, not just
3875 when reusing an existing buffer.
3878 *) New command line and configuration option 'utf8' for the req command.
3879 This allows field values to be specified as UTF8 strings.
3882 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3883 runs for the former and machine-readable output for the latter.
3886 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3887 of the e-mail address in the DN (i.e., it will go into a certificate
3888 extension only). The new configuration file option 'email_in_dn = no'
3889 has the same effect.
3890 [Massimiliano Pala madwolf@openca.org]
3892 *) Change all functions with names starting with des_ to be starting
3893 with DES_ instead. Add wrappers that are compatible with libdes,
3894 but are named _ossl_old_des_*. Finally, add macros that map the
3895 des_* symbols to the corresponding _ossl_old_des_* if libdes
3896 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3897 desired, the des_* symbols will be mapped to DES_*, with one
3900 Since we provide two compatibility mappings, the user needs to
3901 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3902 compatibility is desired. The default (i.e., when that macro
3903 isn't defined) is OpenSSL 0.9.6c compatibility.
3905 There are also macros that enable and disable the support of old
3906 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3907 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3908 are defined, the default will apply: to support the old des routines.
3910 In either case, one must include openssl/des.h to get the correct
3911 definitions. Do not try to just include openssl/des_old.h, that
3914 NOTE: This is a major break of an old API into a new one. Software
3915 authors are encouraged to switch to the DES_ style functions. Some
3916 time in the future, des_old.h and the libdes compatibility functions
3917 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3918 default), and then completely removed.
3921 *) Test for certificates which contain unsupported critical extensions.
3922 If such a certificate is found during a verify operation it is
3923 rejected by default: this behaviour can be overridden by either
3924 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3925 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3926 X509_supported_extension() has also been added which returns 1 if a
3927 particular extension is supported.
3930 *) Modify the behaviour of EVP cipher functions in similar way to digests
3931 to retain compatibility with existing code.
3934 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3935 compatibility with existing code. In particular the 'ctx' parameter does
3936 not have to be to be initialized before the call to EVP_DigestInit() and
3937 it is tidied up after a call to EVP_DigestFinal(). New function
3938 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3939 EVP_MD_CTX_copy() changed to not require the destination to be
3940 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3941 requires the destination to be valid.
3943 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3944 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3947 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3948 so that complete 'Handshake' protocol structures are kept in memory
3949 instead of overwriting 'msg_type' and 'length' with 'body' data.
3952 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3953 [Massimo Santin via Richard Levitte]
3955 *) Major restructuring to the underlying ENGINE code. This includes
3956 reduction of linker bloat, separation of pure "ENGINE" manipulation
3957 (initialisation, etc) from functionality dealing with implementations
3958 of specific crypto iterfaces. This change also introduces integrated
3959 support for symmetric ciphers and digest implementations - so ENGINEs
3960 can now accelerate these by providing EVP_CIPHER and EVP_MD
3961 implementations of their own. This is detailed in crypto/engine/README
3962 as it couldn't be adequately described here. However, there are a few
3963 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3964 were changed in the original introduction of ENGINE code have now
3965 reverted back - the hooking from this code to ENGINE is now a good
3966 deal more passive and at run-time, operations deal directly with
3967 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3968 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3969 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3970 they were not being used by the framework as there is no concept of a
3971 BIGNUM_METHOD and they could not be generalised to the new
3972 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3973 ENGINE_cpy() has been removed as it cannot be consistently defined in
3977 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3980 *) Change mkdef.pl to sort symbols that get the same entry number,
3981 and make sure the automatically generated functions ERR_load_*
3982 become part of libeay.num as well.
3985 *) New function SSL_renegotiate_pending(). This returns true once
3986 renegotiation has been requested (either SSL_renegotiate() call
3987 or HelloRequest/ClientHello receveived from the peer) and becomes
3988 false once a handshake has been completed.
3989 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3990 sends a HelloRequest, but does not ensure that a handshake takes
3991 place. SSL_renegotiate_pending() is useful for checking if the
3992 client has followed the request.)
3995 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3996 By default, clients may request session resumption even during
3997 renegotiation (if session ID contexts permit); with this option,
3998 session resumption is possible only in the first handshake.
4000 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4001 more bits available for options that should not be part of
4002 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4005 *) Add some demos for certificate and certificate request creation.
4008 *) Make maximum certificate chain size accepted from the peer application
4009 settable (SSL*_get/set_max_cert_list()), as proposed by
4010 "Douglas E. Engert" <deengert@anl.gov>.
4013 *) Add support for shared libraries for Unixware-7
4014 (Boyd Lynn Gerber <gerberb@zenez.com>).
4017 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4018 be done prior to destruction. Use this to unload error strings from
4019 ENGINEs that load their own error strings. NB: This adds two new API
4020 functions to "get" and "set" this destroy handler in an ENGINE.
4023 *) Alter all existing ENGINE implementations (except "openssl" and
4024 "openbsd") to dynamically instantiate their own error strings. This
4025 makes them more flexible to be built both as statically-linked ENGINEs
4026 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4027 Also, add stub code to each that makes building them as self-contained
4028 shared-libraries easier (see README.ENGINE).
4031 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4032 implementations into applications that are completely implemented in
4033 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4034 commands that can be used to configure what shared-library to load and
4035 to control aspects of the way it is handled. Also, made an update to
4036 the README.ENGINE file that brings its information up-to-date and
4037 provides some information and instructions on the "dynamic" ENGINE
4038 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4041 *) Make it possible to unload ranges of ERR strings with a new
4042 "ERR_unload_strings" function.
4045 *) Add a copy() function to EVP_MD.
4048 *) Make EVP_MD routines take a context pointer instead of just the
4049 md_data void pointer.
4052 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4053 that the digest can only process a single chunk of data
4054 (typically because it is provided by a piece of
4055 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4056 is only going to provide a single chunk of data, and hence the
4057 framework needn't accumulate the data for oneshot drivers.
4060 *) As with "ERR", make it possible to replace the underlying "ex_data"
4061 functions. This change also alters the storage and management of global
4062 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4063 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4064 index counters. The API functions that use this state have been changed
4065 to take a "class_index" rather than pointers to the class's local STACK
4066 and counter, and there is now an API function to dynamically create new
4067 classes. This centralisation allows us to (a) plug a lot of the
4068 thread-safety problems that existed, and (b) makes it possible to clean
4069 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4070 such data would previously have always leaked in application code and
4071 workarounds were in place to make the memory debugging turn a blind eye
4072 to it. Application code that doesn't use this new function will still
4073 leak as before, but their memory debugging output will announce it now
4074 rather than letting it slide.
4076 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4077 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4078 has a return value to indicate success or failure.
4081 *) Make it possible to replace the underlying "ERR" functions such that the
4082 global state (2 LHASH tables and 2 locks) is only used by the "default"
4083 implementation. This change also adds two functions to "get" and "set"
4084 the implementation prior to it being automatically set the first time
4085 any other ERR function takes place. Ie. an application can call "get",
4086 pass the return value to a module it has just loaded, and that module
4087 can call its own "set" function using that value. This means the
4088 module's "ERR" operations will use (and modify) the error state in the
4089 application and not in its own statically linked copy of OpenSSL code.
4092 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4093 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4094 the operation, and provides a more encapsulated way for external code
4095 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4096 to use these functions rather than manually incrementing the counts.
4098 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4101 *) Add EVP test program.
4104 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4107 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4108 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4109 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4110 These allow a CRL to be built without having to access X509_CRL fields
4111 directly. Modify 'ca' application to use new functions.
4114 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4115 bug workarounds. Rollback attack detection is a security feature.
4116 The problem will only arise on OpenSSL servers when TLSv1 is not
4117 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4118 Software authors not wanting to support TLSv1 will have special reasons
4119 for their choice and can explicitly enable this option.
4120 [Bodo Moeller, Lutz Jaenicke]
4122 *) Rationalise EVP so it can be extended: don't include a union of
4123 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4124 (similar to those existing for EVP_CIPHER_CTX).
4129 EVP_MD_CTX_init(&md); /* new function call */
4130 EVP_DigestInit(&md, EVP_sha1());
4131 EVP_DigestUpdate(&md, in, len);
4132 EVP_DigestFinal(&md, out, NULL);
4133 EVP_MD_CTX_cleanup(&md); /* new function call */
4137 *) Make DES key schedule conform to the usual scheme, as well as
4138 correcting its structure. This means that calls to DES functions
4139 now have to pass a pointer to a des_key_schedule instead of a
4140 plain des_key_schedule (which was actually always a pointer
4143 des_key_schedule ks;
4145 des_set_key_checked(..., &ks);
4146 des_ncbc_encrypt(..., &ks, ...);
4148 (Note that a later change renames 'des_...' into 'DES_...'.)
4151 *) Initial reduction of linker bloat: the use of some functions, such as
4152 PEM causes large amounts of unused functions to be linked in due to
4153 poor organisation. For example pem_all.c contains every PEM function
4154 which has a knock on effect of linking in large amounts of (unused)
4155 ASN1 code. Grouping together similar functions and splitting unrelated
4156 functions prevents this.
4159 *) Cleanup of EVP macros.
4162 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4163 correct _ecb suffix.
4166 *) Add initial OCSP responder support to ocsp application. The
4167 revocation information is handled using the text based index
4168 use by the ca application. The responder can either handle
4169 requests generated internally, supplied in files (for example
4170 via a CGI script) or using an internal minimal server.
4173 *) Add configuration choices to get zlib compression for TLS.
4176 *) Changes to Kerberos SSL for RFC 2712 compliance:
4177 1. Implemented real KerberosWrapper, instead of just using
4178 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4179 2. Implemented optional authenticator field of KerberosWrapper.
4181 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4182 and authenticator structs; see crypto/krb5/.
4184 Generalized Kerberos calls to support multiple Kerberos libraries.
4185 [Vern Staats <staatsvr@asc.hpc.mil>,
4186 Jeffrey Altman <jaltman@columbia.edu>
4187 via Richard Levitte]
4189 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4190 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4191 values for each of the key sizes rather than having just
4192 parameters (and 'speed' generating keys each time).
4195 *) Speed up EVP routines.
4198 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4199 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4200 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4201 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4203 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4204 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4205 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4208 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4210 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4213 *) Added the OS2-EMX target.
4214 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4216 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4217 to support NCONF routines in extension code. New function CONF_set_nconf()
4218 to allow functions which take an NCONF to also handle the old LHASH
4219 structure: this means that the old CONF compatible routines can be
4220 retained (in particular wrt extensions) without having to duplicate the
4221 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4224 *) Enhance the general user interface with mechanisms for inner control
4225 and with possibilities to have yes/no kind of prompts.
4228 *) Change all calls to low level digest routines in the library and
4229 applications to use EVP. Add missing calls to HMAC_cleanup() and
4230 don't assume HMAC_CTX can be copied using memcpy().
4231 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4233 *) Add the possibility to control engines through control names but with
4234 arbitrary arguments instead of just a string.
4235 Change the key loaders to take a UI_METHOD instead of a callback
4236 function pointer. NOTE: this breaks binary compatibility with earlier
4237 versions of OpenSSL [engine].
4238 Adapt the nCipher code for these new conditions and add a card insertion
4242 *) Enhance the general user interface with mechanisms to better support
4243 dialog box interfaces, application-defined prompts, the possibility
4244 to use defaults (for example default passwords from somewhere else)
4245 and interrupts/cancellations.
4248 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4249 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4252 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4253 tidy up some unnecessarily weird code in 'sk_new()').
4254 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4256 *) Change the key loading routines for ENGINEs to use the same kind
4257 callback (pem_password_cb) as all other routines that need this
4261 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4262 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4263 than this minimum value is recommended.
4266 *) New random seeder for OpenVMS, using the system process statistics
4267 that are easily reachable.
4270 *) Windows apparently can't transparently handle global
4271 variables defined in DLLs. Initialisations such as:
4273 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4275 wont compile. This is used by the any applications that need to
4276 declare their own ASN1 modules. This was fixed by adding the option
4277 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4278 needed for static libraries under Win32.
4281 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4282 setting of purpose and trust fields. New X509_STORE trust and
4283 purpose functions and tidy up setting in other SSL functions.
4286 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4287 structure. These are inherited by X509_STORE_CTX when it is
4288 initialised. This allows various defaults to be set in the
4289 X509_STORE structure (such as flags for CRL checking and custom
4290 purpose or trust settings) for functions which only use X509_STORE_CTX
4291 internally such as S/MIME.
4293 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4294 trust settings if they are not set in X509_STORE. This allows X509_STORE
4295 purposes and trust (in S/MIME for example) to override any set by default.
4297 Add command line options for CRL checking to smime, s_client and s_server
4301 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4302 are set then the CRL is looked up in the X509_STORE structure and
4303 its validity and signature checked, then if the certificate is found
4304 in the CRL the verify fails with a revoked error.
4306 Various new CRL related callbacks added to X509_STORE_CTX structure.
4308 Command line options added to 'verify' application to support this.
4310 This needs some additional work, such as being able to handle multiple
4311 CRLs with different times, extension based lookup (rather than just
4312 by subject name) and ultimately more complete V2 CRL extension
4316 *) Add a general user interface API (crypto/ui/). This is designed
4317 to replace things like des_read_password and friends (backward
4318 compatibility functions using this new API are provided).
4319 The purpose is to remove prompting functions from the DES code
4320 section as well as provide for prompting through dialog boxes in
4321 a window system and the like.
4324 *) Add "ex_data" support to ENGINE so implementations can add state at a
4325 per-structure level rather than having to store it globally.
4328 *) Make it possible for ENGINE structures to be copied when retrieved by
4329 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4330 This causes the "original" ENGINE structure to act like a template,
4331 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4332 operational state can be localised to each ENGINE structure, despite the
4333 fact they all share the same "methods". New ENGINE structures returned in
4334 this case have no functional references and the return value is the single
4335 structural reference. This matches the single structural reference returned
4336 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4340 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4341 needs to match any other type at all we need to manually clear the
4345 *) Changes to the "openssl engine" utility to include;
4346 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4347 about an ENGINE's available control commands.
4348 - executing control commands from command line arguments using the
4349 '-pre' and '-post' switches. '-post' is only used if '-t' is
4350 specified and the ENGINE is successfully initialised. The syntax for
4351 the individual commands are colon-separated, for example;
4352 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4355 *) New dynamic control command support for ENGINEs. ENGINEs can now
4356 declare their own commands (numbers), names (strings), descriptions,
4357 and input types for run-time discovery by calling applications. A
4358 subset of these commands are implicitly classed as "executable"
4359 depending on their input type, and only these can be invoked through
4360 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4361 can be based on user input, config files, etc). The distinction is
4362 that "executable" commands cannot return anything other than a boolean
4363 result and can only support numeric or string input, whereas some
4364 discoverable commands may only be for direct use through
4365 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4366 pointers, or other custom uses. The "executable" commands are to
4367 support parameterisations of ENGINE behaviour that can be
4368 unambiguously defined by ENGINEs and used consistently across any
4369 OpenSSL-based application. Commands have been added to all the
4370 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4371 control over shared-library paths without source code alterations.
4374 *) Changed all ENGINE implementations to dynamically allocate their
4375 ENGINEs rather than declaring them statically. Apart from this being
4376 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4377 this also allows the implementations to compile without using the
4378 internal engine_int.h header.
4381 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4382 'const' value. Any code that should be able to modify a RAND_METHOD
4383 should already have non-const pointers to it (ie. they should only
4384 modify their own ones).
4387 *) Made a variety of little tweaks to the ENGINE code.
4388 - "atalla" and "ubsec" string definitions were moved from header files
4389 to C code. "nuron" string definitions were placed in variables
4390 rather than hard-coded - allowing parameterisation of these values
4391 later on via ctrl() commands.
4392 - Removed unused "#if 0"'d code.
4393 - Fixed engine list iteration code so it uses ENGINE_free() to release
4394 structural references.
4395 - Constified the RAND_METHOD element of ENGINE structures.
4396 - Constified various get/set functions as appropriate and added
4397 missing functions (including a catch-all ENGINE_cpy that duplicates
4398 all ENGINE values onto a new ENGINE except reference counts/state).
4399 - Removed NULL parameter checks in get/set functions. Setting a method
4400 or function to NULL is a way of cancelling out a previously set
4401 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4402 and doesn't justify the extra error symbols and code.
4403 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4404 flags from engine_int.h to engine.h.
4405 - Changed prototypes for ENGINE handler functions (init(), finish(),
4406 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4409 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4410 to the algorithm using long division. The binary algorithm can be
4411 used only if the modulus is odd. On 32-bit systems, it is faster
4412 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4413 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4414 up to 450 bits. In 64-bit environments, the binary algorithm
4415 appears to be advantageous for much longer moduli; here we use it
4416 for moduli up to 2048 bits.
4419 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4420 could not support the combine flag in choice fields.
4423 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4424 extensions from a certificate request to the certificate.
4427 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4428 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4429 file: this allows the display of the certificate about to be
4430 signed to be customised, to allow certain fields to be included
4431 or excluded and extension details. The old system didn't display
4432 multicharacter strings properly, omitted fields not in the policy
4433 and couldn't display additional details such as extensions.
4436 *) Function EC_POINTs_mul for multiple scalar multiplication
4437 of an arbitrary number of elliptic curve points
4438 \sum scalars[i]*points[i],
4439 optionally including the generator defined for the EC_GROUP:
4440 scalar*generator + \sum scalars[i]*points[i].
4442 EC_POINT_mul is a simple wrapper function for the typical case
4443 that the point list has just one item (besides the optional
4447 *) First EC_METHODs for curves over GF(p):
4449 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4450 operations and provides various method functions that can also
4451 operate with faster implementations of modular arithmetic.
4453 EC_GFp_mont_method() reuses most functions that are part of
4454 EC_GFp_simple_method, but uses Montgomery arithmetic.
4456 [Bodo Moeller; point addition and point doubling
4457 implementation directly derived from source code provided by
4458 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4460 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4461 crypto/ec/ec_lib.c):
4463 Curves are EC_GROUP objects (with an optional group generator)
4464 based on EC_METHODs that are built into the library.
4466 Points are EC_POINT objects based on EC_GROUP objects.
4468 Most of the framework would be able to handle curves over arbitrary
4469 finite fields, but as there are no obvious types for fields other
4470 than GF(p), some functions are limited to that for now.
4473 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4474 that the file contains a complete HTTP response.
4477 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4478 change the def and num file printf format specifier from "%-40sXXX"
4479 to "%-39s XXX". The latter will always guarantee a space after the
4480 field while the former will cause them to run together if the field
4481 is 40 of more characters long.
4484 *) Constify the cipher and digest 'method' functions and structures
4485 and modify related functions to take constant EVP_MD and EVP_CIPHER
4489 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4490 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4493 *) Modify EVP_Digest*() routines so they now return values. Although the
4494 internal software routines can never fail additional hardware versions
4498 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4500 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4501 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4504 ERR_R_NESTED_ASN1_ERROR
4506 ERR_R_MISSING_ASN1_EOS
4507 were 4 .. 9, conflicting with
4508 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4510 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4511 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4513 Add new error code 'ERR_R_INTERNAL_ERROR'.
4516 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4520 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4521 sets the subject name for a new request or supersedes the
4522 subject name in a given request. Formats that can be parsed are
4523 'CN=Some Name, OU=myOU, C=IT'
4525 'CN=Some Name/OU=myOU/C=IT'.
4527 Add options '-batch' and '-verbose' to 'openssl req'.
4528 [Massimiliano Pala <madwolf@hackmasters.net>]
4530 *) Introduce the possibility to access global variables through
4531 functions on platform were that's the best way to handle exporting
4532 global variables in shared libraries. To enable this functionality,
4533 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4534 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4535 is normally done by Configure or something similar).
4537 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4538 in the source file (foo.c) like this:
4540 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4541 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4543 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4544 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4546 OPENSSL_DECLARE_GLOBAL(int,foo);
4547 #define foo OPENSSL_GLOBAL_REF(foo)
4548 OPENSSL_DECLARE_GLOBAL(double,bar);
4549 #define bar OPENSSL_GLOBAL_REF(bar)
4551 The #defines are very important, and therefore so is including the
4552 header file everywhere where the defined globals are used.
4554 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4555 of ASN.1 items, but that structure is a bit different.
4557 The largest change is in util/mkdef.pl which has been enhanced with
4558 better and easier to understand logic to choose which symbols should
4559 go into the Windows .def files as well as a number of fixes and code
4560 cleanup (among others, algorithm keywords are now sorted
4561 lexicographically to avoid constant rewrites).
4564 *) In BN_div() keep a copy of the sign of 'num' before writing the
4565 result to 'rm' because if rm==num the value will be overwritten
4566 and produce the wrong result if 'num' is negative: this caused
4567 problems with BN_mod() and BN_nnmod().
4570 *) Function OCSP_request_verify(). This checks the signature on an
4571 OCSP request and verifies the signer certificate. The signer
4572 certificate is just checked for a generic purpose and OCSP request
4576 *) Add OCSP_check_validity() function to check the validity of OCSP
4577 responses. OCSP responses are prepared in real time and may only
4578 be a few seconds old. Simply checking that the current time lies
4579 between thisUpdate and nextUpdate max reject otherwise valid responses
4580 caused by either OCSP responder or client clock inaccuracy. Instead
4581 we allow thisUpdate and nextUpdate to fall within a certain period of
4582 the current time. The age of the response can also optionally be
4583 checked. Two new options -validity_period and -status_age added to
4587 *) If signature or public key algorithm is unrecognized print out its
4588 OID rather that just UNKNOWN.
4591 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4592 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4593 ID to be generated from the issuer certificate alone which can then be
4594 passed to OCSP_id_issuer_cmp().
4597 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4598 ASN1 modules to export functions returning ASN1_ITEM pointers
4599 instead of the ASN1_ITEM structures themselves. This adds several
4600 new macros which allow the underlying ASN1 function/structure to
4601 be accessed transparently. As a result code should not use ASN1_ITEM
4602 references directly (such as &X509_it) but instead use the relevant
4603 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4604 use of the new ASN1 code on platforms where exporting structures
4605 is problematical (for example in shared libraries) but exporting
4606 functions returning pointers to structures is not.
4609 *) Add support for overriding the generation of SSL/TLS session IDs.
4610 These callbacks can be registered either in an SSL_CTX or per SSL.
4611 The purpose of this is to allow applications to control, if they wish,
4612 the arbitrary values chosen for use as session IDs, particularly as it
4613 can be useful for session caching in multiple-server environments. A
4614 command-line switch for testing this (and any client code that wishes
4615 to use such a feature) has been added to "s_server".
4616 [Geoff Thorpe, Lutz Jaenicke]
4618 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4619 of the form '#if defined(...) || defined(...) || ...' and
4620 '#if !defined(...) && !defined(...) && ...'. This also avoids
4621 the growing number of special cases it was previously handling.
4624 *) Make all configuration macros available for application by making
4625 sure they are available in opensslconf.h, by giving them names starting
4626 with "OPENSSL_" to avoid conflicts with other packages and by making
4627 sure e_os2.h will cover all platform-specific cases together with
4629 Additionally, it is now possible to define configuration/platform-
4630 specific names (called "system identities"). In the C code, these
4631 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4632 macro with the name beginning with "OPENSSL_SYS_", which is determined
4633 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4637 *) New option -set_serial to 'req' and 'x509' this allows the serial
4638 number to use to be specified on the command line. Previously self
4639 signed certificates were hard coded with serial number 0 and the
4640 CA options of 'x509' had to use a serial number in a file which was
4644 *) New options to 'ca' utility to support V2 CRL entry extensions.
4645 Currently CRL reason, invalidity date and hold instruction are
4646 supported. Add new CRL extensions to V3 code and some new objects.
4649 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4650 disable standard block padding (aka PKCS#5 padding) in the EVP
4651 API, which was previously mandatory. This means that the data is
4652 not padded in any way and so the total length much be a multiple
4653 of the block size, otherwise an error occurs.
4656 *) Initial (incomplete) OCSP SSL support.
4659 *) New function OCSP_parse_url(). This splits up a URL into its host,
4660 port and path components: primarily to parse OCSP URLs. New -url
4661 option to ocsp utility.
4664 *) New nonce behavior. The return value of OCSP_check_nonce() now
4665 reflects the various checks performed. Applications can decide
4666 whether to tolerate certain situations such as an absent nonce
4667 in a response when one was present in a request: the ocsp application
4668 just prints out a warning. New function OCSP_add1_basic_nonce()
4669 this is to allow responders to include a nonce in a response even if
4670 the request is nonce-less.
4673 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4674 skipped when using openssl x509 multiple times on a single input file,
4675 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4678 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4679 set string type: to handle setting ASN1_TIME structures. Fix ca
4680 utility to correctly initialize revocation date of CRLs.
4683 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4684 the clients preferred ciphersuites and rather use its own preferences.
4685 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4686 Internet Explorer by ensuring unchanged hash method during stepup.
4687 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4690 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4691 to aes and add a new 'exist' option to print out symbols that don't
4695 *) Additional options to ocsp utility to allow flags to be set and
4696 additional certificates supplied.
4699 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4700 OCSP client a number of certificate to only verify the response
4704 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4705 handle the new API. Currently only ECB, CBC modes supported. Add new
4708 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4709 Encryption Standard (AES) Ciphersuites for Transport Layer
4710 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4711 not enabled by default and were not part of the "ALL" ciphersuite
4712 alias because they were not yet official; they could be
4713 explicitly requested by specifying the "AESdraft" ciphersuite
4714 group alias. In the final release of OpenSSL 0.9.7, the group
4715 alias is called "AES" and is part of "ALL".)
4716 [Ben Laurie, Steve Henson, Bodo Moeller]
4718 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4719 request to response.
4722 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4723 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4724 extract information from a certificate request. OCSP_response_create()
4725 creates a response and optionally adds a basic response structure.
4726 OCSP_basic_add1_status() adds a complete single response to a basic
4727 response and returns the OCSP_SINGLERESP structure just added (to allow
4728 extensions to be included for example). OCSP_basic_add1_cert() adds a
4729 certificate to a basic response and OCSP_basic_sign() signs a basic
4730 response with various flags. New helper functions ASN1_TIME_check()
4731 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4732 (converts ASN1_TIME to GeneralizedTime).
4735 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4736 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4737 structure from a certificate. X509_pubkey_digest() digests the public_key
4738 contents: this is used in various key identifiers.
4741 *) Make sk_sort() tolerate a NULL argument.
4742 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4744 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4745 passed by the function are trusted implicitly. If any of them signed the
4746 response then it is assumed to be valid and is not verified.
4749 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4750 to data. This was previously part of the PKCS7 ASN1 code. This
4751 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4752 [Steve Henson, reported by Kenneth R. Robinette
4753 <support@securenetterm.com>]
4755 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4756 routines: without these tracing memory leaks is very painful.
4757 Fix leaks in PKCS12 and PKCS7 routines.
4760 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4761 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4762 effectively meant GeneralizedTime would never be used. Now it
4763 is initialised to -1 but X509_time_adj() now has to check the value
4764 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4765 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4766 [Steve Henson, reported by Kenneth R. Robinette
4767 <support@securenetterm.com>]
4769 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4770 result in a zero length in the ASN1_INTEGER structure which was
4771 not consistent with the structure when d2i_ASN1_INTEGER() was used
4772 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4773 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4774 where it did not print out a minus for negative ASN1_INTEGER.
4777 *) Add summary printout to ocsp utility. The various functions which
4778 convert status values to strings have been renamed to:
4779 OCSP_response_status_str(), OCSP_cert_status_str() and
4780 OCSP_crl_reason_str() and are no longer static. New options
4781 to verify nonce values and to disable verification. OCSP response
4782 printout format cleaned up.
4785 *) Add additional OCSP certificate checks. These are those specified
4786 in RFC2560. This consists of two separate checks: the CA of the
4787 certificate being checked must either be the OCSP signer certificate
4788 or the issuer of the OCSP signer certificate. In the latter case the
4789 OCSP signer certificate must contain the OCSP signing extended key
4790 usage. This check is performed by attempting to match the OCSP
4791 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4792 in the OCSP_CERTID structures of the response.
4795 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4796 and related routines. This uses the standard OpenSSL certificate
4797 verify routines to perform initial checks (just CA validity) and
4798 to obtain the certificate chain. Then additional checks will be
4799 performed on the chain. Currently the root CA is checked to see
4800 if it is explicitly trusted for OCSP signing. This is used to set
4801 a root CA as a global signing root: that is any certificate that
4802 chains to that CA is an acceptable OCSP signing certificate.
4805 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4806 extensions from a separate configuration file.
4807 As when reading extensions from the main configuration file,
4808 the '-extensions ...' option may be used for specifying the
4810 [Massimiliano Pala <madwolf@comune.modena.it>]
4812 *) New OCSP utility. Allows OCSP requests to be generated or
4813 read. The request can be sent to a responder and the output
4814 parsed, outputed or printed in text form. Not complete yet:
4815 still needs to check the OCSP response validity.
4818 *) New subcommands for 'openssl ca':
4819 'openssl ca -status <serial>' prints the status of the cert with
4820 the given serial number (according to the index file).
4821 'openssl ca -updatedb' updates the expiry status of certificates
4823 [Massimiliano Pala <madwolf@comune.modena.it>]
4825 *) New '-newreq-nodes' command option to CA.pl. This is like
4826 '-newreq', but calls 'openssl req' with the '-nodes' option
4827 so that the resulting key is not encrypted.
4828 [Damien Miller <djm@mindrot.org>]
4830 *) New configuration for the GNU Hurd.
4831 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4833 *) Initial code to implement OCSP basic response verify. This
4834 is currently incomplete. Currently just finds the signer's
4835 certificate and verifies the signature on the response.
4838 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4839 value of OPENSSLDIR. This is available via the new '-d' option
4840 to 'openssl version', and is also included in 'openssl version -a'.
4843 *) Allowing defining memory allocation callbacks that will be given
4844 file name and line number information in additional arguments
4845 (a const char* and an int). The basic functionality remains, as
4846 well as the original possibility to just replace malloc(),
4847 realloc() and free() by functions that do not know about these
4848 additional arguments. To register and find out the current
4849 settings for extended allocation functions, the following
4850 functions are provided:
4852 CRYPTO_set_mem_ex_functions
4853 CRYPTO_set_locked_mem_ex_functions
4854 CRYPTO_get_mem_ex_functions
4855 CRYPTO_get_locked_mem_ex_functions
4857 These work the same way as CRYPTO_set_mem_functions and friends.
4858 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4859 extended allocation function is enabled.
4860 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4861 a conventional allocation function is enabled.
4862 [Richard Levitte, Bodo Moeller]
4864 *) Finish off removing the remaining LHASH function pointer casts.
4865 There should no longer be any prototype-casting required when using
4866 the LHASH abstraction, and any casts that remain are "bugs". See
4867 the callback types and macros at the head of lhash.h for details
4868 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4871 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4872 If /dev/[u]random devices are not available or do not return enough
4873 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4875 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4876 /etc/entropy will be queried once each in this sequence, quering stops
4877 when enough entropy was collected without querying more sockets.
4880 *) Change the Unix RAND_poll() variant to be able to poll several
4881 random devices, as specified by DEVRANDOM, until a sufficient amount
4882 of data has been collected. We spend at most 10 ms on each file
4883 (select timeout) and read in non-blocking mode. DEVRANDOM now
4884 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4885 (previously it was just the string "/dev/urandom"), so on typical
4886 platforms the 10 ms delay will never occur.
4887 Also separate out the Unix variant to its own file, rand_unix.c.
4888 For VMS, there's a currently-empty rand_vms.c.
4891 *) Move OCSP client related routines to ocsp_cl.c. These
4892 provide utility functions which an application needing
4893 to issue a request to an OCSP responder and analyse the
4894 response will typically need: as opposed to those which an
4895 OCSP responder itself would need which will be added later.
4897 OCSP_request_sign() signs an OCSP request with an API similar
4898 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4899 response. OCSP_response_get1_basic() extracts basic response
4900 from response. OCSP_resp_find_status(): finds and extracts status
4901 information from an OCSP_CERTID structure (which will be created
4902 when the request structure is built). These are built from lower
4903 level functions which work on OCSP_SINGLERESP structures but
4904 wont normally be used unless the application wishes to examine
4905 extensions in the OCSP response for example.
4907 Replace nonce routines with a pair of functions.
4908 OCSP_request_add1_nonce() adds a nonce value and optionally
4909 generates a random value. OCSP_check_nonce() checks the
4910 validity of the nonce in an OCSP response.
4913 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4914 This doesn't copy the supplied OCSP_CERTID and avoids the
4915 need to free up the newly created id. Change return type
4916 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4917 This can then be used to add extensions to the request.
4918 Deleted OCSP_request_new(), since most of its functionality
4919 is now in OCSP_REQUEST_new() (and the case insensitive name
4920 clash) apart from the ability to set the request name which
4921 will be added elsewhere.
4924 *) Update OCSP API. Remove obsolete extensions argument from
4925 various functions. Extensions are now handled using the new
4926 OCSP extension code. New simple OCSP HTTP function which
4927 can be used to send requests and parse the response.
4930 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4931 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4932 uses the special reorder version of SET OF to sort the attributes
4933 and reorder them to match the encoded order. This resolves a long
4934 standing problem: a verify on a PKCS7 structure just after signing
4935 it used to fail because the attribute order did not match the
4936 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4937 it uses the received order. This is necessary to tolerate some broken
4938 software that does not order SET OF. This is handled by encoding
4939 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4940 to produce the required SET OF.
4943 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4944 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4945 files to get correct declarations of the ASN.1 item variables.
4948 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4949 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4950 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4951 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4952 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4953 ASN1_ITEM and no wrapper functions.
4956 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4957 replace the old function pointer based I/O routines. Change most of
4958 the *_d2i_bio() and *_d2i_fp() functions to use these.
4961 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4962 lines, recognice more "algorithms" that can be deselected, and make
4963 it complain about algorithm deselection that isn't recognised.
4966 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4967 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4968 to use new functions. Add NO_ASN1_OLD which can be set to remove
4969 some old style ASN1 functions: this can be used to determine if old
4970 code will still work when these eventually go away.
4973 *) New extension functions for OCSP structures, these follow the
4974 same conventions as certificates and CRLs.
4977 *) New function X509V3_add1_i2d(). This automatically encodes and
4978 adds an extension. Its behaviour can be customised with various
4979 flags to append, replace or delete. Various wrappers added for
4980 certifcates and CRLs.
4983 *) Fix to avoid calling the underlying ASN1 print routine when
4984 an extension cannot be parsed. Correct a typo in the
4985 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4988 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4989 entries for variables.
4992 *) Add functionality to apps/openssl.c for detecting locking
4993 problems: As the program is single-threaded, all we have
4994 to do is register a locking callback using an array for
4995 storing which locks are currently held by the program.
4998 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4999 SSL_get_ex_data_X509_STORE_idx(), which is used in
5000 ssl_verify_cert_chain() and thus can be called at any time
5001 during TLS/SSL handshakes so that thread-safety is essential.
5002 Unfortunately, the ex_data design is not at all suited
5003 for multi-threaded use, so it probably should be abolished.
5006 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5007 [Broadcom, tweaked and integrated by Geoff Thorpe]
5009 *) Move common extension printing code to new function
5010 X509V3_print_extensions(). Reorganise OCSP print routines and
5011 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5014 *) New function X509_signature_print() to remove duplication in some
5018 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5019 set (this was treated exactly the same as SET OF previously). This
5020 is used to reorder the STACK representing the structure to match the
5021 encoding. This will be used to get round a problem where a PKCS7
5022 structure which was signed could not be verified because the STACK
5023 order did not reflect the encoded order.
5026 *) Reimplement the OCSP ASN1 module using the new code.
5029 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5030 for its ASN1 operations. The old style function pointers still exist
5031 for now but they will eventually go away.
5034 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5035 completely replaces the old ASN1 functionality with a table driven
5036 encoder and decoder which interprets an ASN1_ITEM structure describing
5037 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5038 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5039 has also been converted to the new form.
5042 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5043 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5044 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5045 for negative moduli.
5048 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5049 of not touching the result's sign bit.
5052 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5056 *) Changed the LHASH code to use prototypes for callbacks, and created
5057 macros to declare and implement thin (optionally static) functions
5058 that provide type-safety and avoid function pointer casting for the
5059 type-specific callbacks.
5062 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5064 [Veers Staats <staatsvr@asc.hpc.mil>,
5065 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5067 *) Reformat the FAQ so the different questions and answers can be divided
5068 in sections depending on the subject.
5071 *) Have the zlib compression code load ZLIB.DLL dynamically under
5075 *) New function BN_mod_sqrt for computing square roots modulo a prime
5076 (using the probabilistic Tonelli-Shanks algorithm unless
5077 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5078 be handled deterministically).
5079 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5081 *) Make BN_mod_inverse faster by explicitly handling small quotients
5082 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5083 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5086 *) New function BN_kronecker.
5089 *) Fix BN_gcd so that it works on negative inputs; the result is
5090 positive unless both parameters are zero.
5091 Previously something reasonably close to an infinite loop was
5092 possible because numbers could be growing instead of shrinking
5093 in the implementation of Euclid's algorithm.
5096 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5097 sign of the number in question.
5099 Fix BN_is_word(a,w) to work correctly for w == 0.
5101 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5102 because its test if the absolute value of 'a' equals 'w'.
5103 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5104 it exists mostly for use in the implementations of BN_is_zero(),
5105 BN_is_one(), and BN_is_word().
5108 *) New function BN_swap.
5111 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5112 the exponentiation functions are more likely to produce reasonable
5113 results on negative inputs.
5116 *) Change BN_mod_mul so that the result is always non-negative.
5117 Previously, it could be negative if one of the factors was negative;
5118 I don't think anyone really wanted that behaviour.
5121 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5122 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5123 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5124 and add new functions:
5133 BN_mod_lshift1_quick
5137 These functions always generate non-negative results.
5139 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5140 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5142 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5143 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5144 be reduced modulo m.
5145 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5148 The following entry accidentily appeared in the CHANGES file
5149 distributed with OpenSSL 0.9.7. The modifications described in
5150 it do *not* apply to OpenSSL 0.9.7.
5152 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5153 was actually never needed) and in BN_mul(). The removal in BN_mul()
5154 required a small change in bn_mul_part_recursive() and the addition
5155 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5156 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5157 bn_sub_words() and bn_add_words() except they take arrays with
5162 *) In 'openssl passwd', verify passwords read from the terminal
5163 unless the '-salt' option is used (which usually means that
5164 verification would just waste user's time since the resulting
5165 hash is going to be compared with some given password hash)
5166 or the new '-noverify' option is used.
5168 This is an incompatible change, but it does not affect
5169 non-interactive use of 'openssl passwd' (passwords on the command
5170 line, '-stdin' option, '-in ...' option) and thus should not
5174 *) Remove all references to RSAref, since there's no more need for it.
5177 *) Make DSO load along a path given through an environment variable
5178 (SHLIB_PATH) with shl_load().
5181 *) Constify the ENGINE code as a result of BIGNUM constification.
5182 Also constify the RSA code and most things related to it. In a
5183 few places, most notable in the depth of the ASN.1 code, ugly
5184 casts back to non-const were required (to be solved at a later
5188 *) Make it so the openssl application has all engines loaded by default.
5191 *) Constify the BIGNUM routines a little more.
5194 *) Add the following functions:
5196 ENGINE_load_cswift()
5198 ENGINE_load_atalla()
5200 ENGINE_load_builtin_engines()
5202 That way, an application can itself choose if external engines that
5203 are built-in in OpenSSL shall ever be used or not. The benefit is
5204 that applications won't have to be linked with libdl or other dso
5205 libraries unless it's really needed.
5207 Changed 'openssl engine' to load all engines on demand.
5208 Changed the engine header files to avoid the duplication of some
5209 declarations (they differed!).
5212 *) 'openssl engine' can now list capabilities.
5215 *) Better error reporting in 'openssl engine'.
5218 *) Never call load_dh_param(NULL) in s_server.
5221 *) Add engine application. It can currently list engines by name and
5222 identity, and test if they are actually available.
5225 *) Improve RPM specification file by forcing symbolic linking and making
5226 sure the installed documentation is also owned by root.root.
5227 [Damien Miller <djm@mindrot.org>]
5229 *) Give the OpenSSL applications more possibilities to make use of
5230 keys (public as well as private) handled by engines.
5233 *) Add OCSP code that comes from CertCo.
5236 *) Add VMS support for the Rijndael code.
5239 *) Added untested support for Nuron crypto accelerator.
5242 *) Add support for external cryptographic devices. This code was
5243 previously distributed separately as the "engine" branch.
5244 [Geoff Thorpe, Richard Levitte]
5246 *) Rework the filename-translation in the DSO code. It is now possible to
5247 have far greater control over how a "name" is turned into a filename
5248 depending on the operating environment and any oddities about the
5249 different shared library filenames on each system.
5252 *) Support threads on FreeBSD-elf in Configure.
5255 *) Fix for SHA1 assembly problem with MASM: it produces
5256 warnings about corrupt line number information when assembling
5257 with debugging information. This is caused by the overlapping
5259 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5262 NCONF_get_number() has no error checking at all. As a replacement,
5263 NCONF_get_number_e() is defined (_e for "error checking") and is
5264 promoted strongly. The old NCONF_get_number is kept around for
5265 binary backward compatibility.
5266 Make it possible for methods to load from something other than a BIO,
5267 by providing a function pointer that is given a name instead of a BIO.
5268 For example, this could be used to load configuration data from an
5272 *) Fix for non blocking accept BIOs. Added new I/O special reason
5273 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5274 with non blocking I/O was not possible because no retry code was
5275 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5279 *) Added the beginnings of Rijndael support.
5282 *) Fix for bug in DirectoryString mask setting. Add support for
5283 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5284 to allow certificate printing to more controllable, additional
5285 'certopt' option to 'x509' to allow new printing options to be
5289 *) Clean old EAY MD5 hack from e_os.h.
5292 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5294 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5295 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5296 [Joe Orton, Steve Henson]
5298 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5300 *) Fix additional bug revealed by the NISCC test suite:
5302 Stop bug triggering large recursion when presented with
5303 certain ASN.1 tags (CVE-2003-0851)
5306 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5308 *) Fix various bugs revealed by running the NISCC test suite:
5310 Stop out of bounds reads in the ASN1 code when presented with
5311 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5313 If verify callback ignores invalid public key errors don't try to check
5314 certificate signature with the NULL public key.
5318 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5319 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5323 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5324 extra data after the compression methods not only for TLS 1.0
5325 but also for SSL 3.0 (as required by the specification).
5326 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5328 *) Change X509_certificate_type() to mark the key as exported/exportable
5329 when it's 512 *bits* long, not 512 bytes.
5332 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5334 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5335 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5336 a protocol version number mismatch like a decryption error
5337 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5340 *) Turn on RSA blinding by default in the default implementation
5341 to avoid a timing attack. Applications that don't want it can call
5342 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5343 They would be ill-advised to do so in most cases.
5344 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5346 *) Change RSA blinding code so that it works when the PRNG is not
5347 seeded (in this case, the secret RSA exponent is abused as
5348 an unpredictable seed -- if it is not unpredictable, there
5349 is no point in blinding anyway). Make RSA blinding thread-safe
5350 by remembering the creator's thread ID in rsa->blinding and
5351 having all other threads use local one-time blinding factors
5352 (this requires more computation than sharing rsa->blinding, but
5353 avoids excessive locking; and if an RSA object is not shared
5354 between threads, blinding will still be very fast).
5357 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5359 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5360 via timing by performing a MAC computation even if incorrrect
5361 block cipher padding has been found. This is a countermeasure
5362 against active attacks where the attacker has to distinguish
5363 between bad padding and a MAC verification error. (CVE-2003-0078)
5365 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5366 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5367 Martin Vuagnoux (EPFL, Ilion)]
5369 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5371 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5372 memory from it's contents. This is done with a counter that will
5373 place alternating values in each byte. This can be used to solve
5374 two issues: 1) the removal of calls to memset() by highly optimizing
5375 compilers, and 2) cleansing with other values than 0, since those can
5376 be read through on certain media, for example a swap space on disk.
5379 *) Bugfix: client side session caching did not work with external caching,
5380 because the session->cipher setting was not restored when reloading
5381 from the external cache. This problem was masked, when
5382 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5383 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5386 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5387 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5388 [Zeev Lieber <zeev-l@yahoo.com>]
5390 *) Undo an undocumented change introduced in 0.9.6e which caused
5391 repeated calls to OpenSSL_add_all_ciphers() and
5392 OpenSSL_add_all_digests() to be ignored, even after calling
5396 *) Change the default configuration reader to deal with last line not
5397 being properly terminated.
5400 *) Change X509_NAME_cmp() so it applies the special rules on handling
5401 DN values that are of type PrintableString, as well as RDNs of type
5402 emailAddress where the value has the type ia5String.
5403 [stefank@valicert.com via Richard Levitte]
5405 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5406 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5407 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5408 the bitwise-OR of the two for use by the majority of applications
5409 wanting this behaviour, and update the docs. The documented
5410 behaviour and actual behaviour were inconsistent and had been
5411 changing anyway, so this is more a bug-fix than a behavioural
5413 [Geoff Thorpe, diagnosed by Nadav Har'El]
5415 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5416 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5419 *) Fix initialization code race conditions in
5420 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5421 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5422 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5423 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5424 ssl2_get_cipher_by_char(),
5425 ssl3_get_cipher_by_char().
5426 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5428 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5429 the cached sessions are flushed, as the remove_cb() might use ex_data
5430 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5431 (see [openssl.org #212]).
5432 [Geoff Thorpe, Lutz Jaenicke]
5434 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5435 length, instead of the encoding length to d2i_ASN1_OBJECT.
5438 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5440 *) [In 0.9.6g-engine release:]
5441 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5442 [Lynn Gazis <lgazis@rainbow.com>]
5444 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5446 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5447 and get fix the header length calculation.
5448 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5449 Alon Kantor <alonk@checkpoint.com> (and others),
5452 *) Use proper error handling instead of 'assertions' in buffer
5453 overflow checks added in 0.9.6e. This prevents DoS (the
5454 assertions could call abort()).
5455 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5457 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5459 *) Add various sanity checks to asn1_get_length() to reject
5460 the ASN1 length bytes if they exceed sizeof(long), will appear
5461 negative or the content length exceeds the length of the
5463 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5465 *) Fix cipher selection routines: ciphers without encryption had no flags
5466 for the cipher strength set and where therefore not handled correctly
5467 by the selection routines (PR #130).
5470 *) Fix EVP_dsa_sha macro.
5474 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5475 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5476 that was added in OpenSSL 0.9.6d.
5478 As the countermeasure turned out to be incompatible with some
5479 broken SSL implementations, the new option is part of SSL_OP_ALL.
5480 SSL_OP_ALL is usually employed when compatibility with weird SSL
5481 implementations is desired (e.g. '-bugs' option to 's_client' and
5482 's_server'), so the new option is automatically set in many
5486 *) Changes in security patch:
5488 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5489 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5490 Air Force Materiel Command, USAF, under agreement number
5493 *) Add various sanity checks to asn1_get_length() to reject
5494 the ASN1 length bytes if they exceed sizeof(long), will appear
5495 negative or the content length exceeds the length of the
5496 supplied buffer. (CVE-2002-0659)
5497 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5499 *) Assertions for various potential buffer overflows, not known to
5501 [Ben Laurie (CHATS)]
5503 *) Various temporary buffers to hold ASCII versions of integers were
5504 too small for 64 bit platforms. (CVE-2002-0655)
5505 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5507 *) Remote buffer overflow in SSL3 protocol - an attacker could
5508 supply an oversized session ID to a client. (CVE-2002-0656)
5509 [Ben Laurie (CHATS)]
5511 *) Remote buffer overflow in SSL2 protocol - an attacker could
5512 supply an oversized client master key. (CVE-2002-0656)
5513 [Ben Laurie (CHATS)]
5515 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5517 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5518 encoded as NULL) with id-dsa-with-sha1.
5519 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5521 *) Check various X509_...() return values in apps/req.c.
5522 [Nils Larsch <nla@trustcenter.de>]
5524 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5525 an end-of-file condition would erronously be flagged, when the CRLF
5526 was just at the end of a processed block. The bug was discovered when
5527 processing data through a buffering memory BIO handing the data to a
5528 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5529 <ptsekov@syntrex.com> and Nedelcho Stanev.
5532 *) Implement a countermeasure against a vulnerability recently found
5533 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5534 before application data chunks to avoid the use of known IVs
5535 with data potentially chosen by the attacker.
5538 *) Fix length checks in ssl3_get_client_hello().
5541 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5542 to prevent ssl3_read_internal() from incorrectly assuming that
5543 ssl3_read_bytes() found application data while handshake
5544 processing was enabled when in fact s->s3->in_read_app_data was
5545 merely automatically cleared during the initial handshake.
5546 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5548 *) Fix object definitions for Private and Enterprise: they were not
5549 recognized in their shortname (=lowercase) representation. Extend
5550 obj_dat.pl to issue an error when using undefined keywords instead
5551 of silently ignoring the problem (Svenning Sorensen
5552 <sss@sss.dnsalias.net>).
5555 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5556 generators, i.e. generators other than 2 and 5. (Previously, the
5557 code did not properly initialise the 'add' and 'rem' values to
5558 BN_generate_prime().)
5560 In the new general case, we do not insist that 'generator' is
5561 actually a primitive root: This requirement is rather pointless;
5562 a generator of the order-q subgroup is just as good, if not
5566 *) Map new X509 verification errors to alerts. Discovered and submitted by
5567 Tom Wu <tom@arcot.com>.
5570 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5571 returning non-zero before the data has been completely received
5572 when using non-blocking I/O.
5573 [Bodo Moeller; problem pointed out by John Hughes]
5575 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5576 [Ben Laurie, Lutz Jaenicke]
5578 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5579 Yoram Zahavi <YoramZ@gilian.com>).
5582 *) Add information about CygWin 1.3 and on, and preserve proper
5583 configuration for the versions before that.
5584 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5586 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5587 check whether we deal with a copy of a session and do not delete from
5588 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5589 <izhar@checkpoint.com>.
5592 *) Do not store session data into the internal session cache, if it
5593 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5594 flag is set). Proposed by Aslam <aslam@funk.com>.
5597 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5601 *) [In 0.9.6d-engine release:]
5602 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5603 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5605 *) Add the configuration target linux-s390x.
5606 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5608 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5609 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5610 variable as an indication that a ClientHello message has been
5611 received. As the flag value will be lost between multiple
5612 invocations of ssl3_accept when using non-blocking I/O, the
5613 function may not be aware that a handshake has actually taken
5614 place, thus preventing a new session from being added to the
5617 To avoid this problem, we now set s->new_session to 2 instead of
5618 using a local variable.
5619 [Lutz Jaenicke, Bodo Moeller]
5621 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5622 if the SSL_R_LENGTH_MISMATCH error is detected.
5623 [Geoff Thorpe, Bodo Moeller]
5625 *) New 'shared_ldflag' column in Configure platform table.
5628 *) Fix EVP_CIPHER_mode macro.
5629 ["Dan S. Camper" <dan@bti.net>]
5631 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5632 type, we must throw them away by setting rr->length to 0.
5633 [D P Chang <dpc@qualys.com>]
5635 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5637 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5638 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5639 worked incorrectly for those cases where range = 10..._2 and
5640 3*range is two bits longer than range.)
5643 *) Only add signing time to PKCS7 structures if it is not already
5647 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5648 OBJ_ld_ce should be OBJ_id_ce.
5649 Also some ip-pda OIDs in crypto/objects/objects.txt were
5650 incorrect (cf. RFC 3039).
5651 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5653 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5654 returns early because it has nothing to do.
5655 [Andy Schneider <andy.schneider@bjss.co.uk>]
5657 *) [In 0.9.6c-engine release:]
5658 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5659 [Andy Schneider <andy.schneider@bjss.co.uk>]
5661 *) [In 0.9.6c-engine release:]
5662 Add support for Cryptographic Appliance's keyserver technology.
5663 (Use engine 'keyclient')
5664 [Cryptographic Appliances and Geoff Thorpe]
5666 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5667 is called via tools/c89.sh because arguments have to be
5668 rearranged (all '-L' options must appear before the first object
5670 [Richard Shapiro <rshapiro@abinitio.com>]
5672 *) [In 0.9.6c-engine release:]
5673 Add support for Broadcom crypto accelerator cards, backported
5675 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5677 *) [In 0.9.6c-engine release:]
5678 Add support for SureWare crypto accelerator cards from
5679 Baltimore Technologies. (Use engine 'sureware')
5680 [Baltimore Technologies and Mark Cox]
5682 *) [In 0.9.6c-engine release:]
5683 Add support for crypto accelerator cards from Accelerated
5684 Encryption Processing, www.aep.ie. (Use engine 'aep')
5685 [AEP Inc. and Mark Cox]
5687 *) Add a configuration entry for gcc on UnixWare.
5688 [Gary Benson <gbenson@redhat.com>]
5690 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5691 messages are stored in a single piece (fixed-length part and
5692 variable-length part combined) and fix various bugs found on the way.
5695 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5696 instead. BIO_gethostbyname() does not know what timeouts are
5697 appropriate, so entries would stay in cache even when they have
5699 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5701 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5702 faced with a pathologically small ClientHello fragment that does
5703 not contain client_version: Instead of aborting with an error,
5704 simply choose the highest available protocol version (i.e.,
5705 TLS 1.0 unless it is disabled). In practice, ClientHello
5706 messages are never sent like this, but this change gives us
5707 strictly correct behaviour at least for TLS.
5710 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5711 never resets s->method to s->ctx->method when called from within
5712 one of the SSL handshake functions.
5713 [Bodo Moeller; problem pointed out by Niko Baric]
5715 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5716 (sent using the client's version number) if client_version is
5717 smaller than the protocol version in use. Also change
5718 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5719 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5720 the client will at least see that alert.
5723 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5727 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5728 client receives HelloRequest while in a handshake.
5729 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5731 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5732 should end in 'break', not 'goto end' which circuments various
5733 cleanups done in state SSL_ST_OK. But session related stuff
5734 must be disabled for SSL_ST_OK in the case that we just sent a
5737 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5738 before just sending a HelloRequest.
5739 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5741 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5742 reveal whether illegal block cipher padding was found or a MAC
5743 verification error occured. (Neither SSLerr() codes nor alerts
5744 are directly visible to potential attackers, but the information
5745 may leak via logfiles.)
5747 Similar changes are not required for the SSL 2.0 implementation
5748 because the number of padding bytes is sent in clear for SSL 2.0,
5749 and the extra bytes are just ignored. However ssl/s2_pkt.c
5750 failed to verify that the purported number of padding bytes is in
5754 *) Add OpenUNIX-8 support including shared libraries
5755 (Boyd Lynn Gerber <gerberb@zenez.com>).
5758 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5759 'wristwatch attack' using huge encoding parameters (cf.
5760 James H. Manger's CRYPTO 2001 paper). Note that the
5761 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5762 encoding parameters and hence was not vulnerable.
5765 *) BN_sqr() bug fix.
5766 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5768 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5769 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5770 followed by modular reduction.
5771 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5773 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5774 equivalent based on BN_pseudo_rand() instead of BN_rand().
5777 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5778 This function was broken, as the check for a new client hello message
5779 to handle SGC did not allow these large messages.
5780 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5783 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5786 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5787 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5790 *) Rework the configuration and shared library support for Tru64 Unix.
5791 The configuration part makes use of modern compiler features and
5792 still retains old compiler behavior for those that run older versions
5793 of the OS. The shared library support part includes a variant that
5794 uses the RPATH feature, and is available through the special
5795 configuration target "alpha-cc-rpath", which will never be selected
5797 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5799 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5800 with the same message size as in ssl3_get_certificate_request().
5801 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5802 messages might inadvertently be reject as too long.
5803 [Petr Lampa <lampa@fee.vutbr.cz>]
5805 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5808 *) Modified SSL library such that the verify_callback that has been set
5809 specificly for an SSL object with SSL_set_verify() is actually being
5810 used. Before the change, a verify_callback set with this function was
5811 ignored and the verify_callback() set in the SSL_CTX at the time of
5812 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5813 to allow the necessary settings.
5816 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5817 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5818 done automatically (in contradiction to the requirements of the C
5819 standard). This made problems when used from OpenSSH.
5822 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5823 dh->length and always used
5825 BN_rand_range(priv_key, dh->p).
5827 BN_rand_range() is not necessary for Diffie-Hellman, and this
5828 specific range makes Diffie-Hellman unnecessarily inefficient if
5829 dh->length (recommended exponent length) is much smaller than the
5830 length of dh->p. We could use BN_rand_range() if the order of
5831 the subgroup was stored in the DH structure, but we only have
5836 BN_rand(priv_key, l, ...)
5838 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5844 RSA_eay_public_encrypt
5845 RSA_eay_private_decrypt
5846 RSA_eay_private_encrypt (signing)
5847 RSA_eay_public_decrypt (signature verification)
5849 (default implementations for RSA_public_encrypt,
5850 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5851 always reject numbers >= n.
5854 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5855 to synchronize access to 'locking_thread'. This is necessary on
5856 systems where access to 'locking_thread' (an 'unsigned long'
5857 variable) is not atomic.
5860 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5861 *before* setting the 'crypto_lock_rand' flag. The previous code had
5862 a race condition if 0 is a valid thread ID.
5863 [Travis Vitek <vitek@roguewave.com>]
5865 *) Add support for shared libraries under Irix.
5866 [Albert Chin-A-Young <china@thewrittenword.com>]
5868 *) Add configuration option to build on Linux on both big-endian and
5870 [Ralf Baechle <ralf@uni-koblenz.de>]
5872 *) Add the possibility to create shared libraries on HP-UX.
5875 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5877 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5878 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5879 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5880 PRNG state recovery was possible based on the output of
5881 one PRNG request appropriately sized to gain knowledge on
5882 'md' followed by enough consecutive 1-byte PRNG requests
5883 to traverse all of 'state'.
5885 1. When updating 'md_local' (the current thread's copy of 'md')
5886 during PRNG output generation, hash all of the previous
5887 'md_local' value, not just the half used for PRNG output.
5889 2. Make the number of bytes from 'state' included into the hash
5890 independent from the number of PRNG bytes requested.
5892 The first measure alone would be sufficient to avoid
5893 Markku-Juhani's attack. (Actually it had never occurred
5894 to me that the half of 'md_local' used for chaining was the
5895 half from which PRNG output bytes were taken -- I had always
5896 assumed that the secret half would be used.) The second
5897 measure makes sure that additional data from 'state' is never
5898 mixed into 'md_local' in small portions; this heuristically
5899 further strengthens the PRNG.
5902 *) Fix crypto/bn/asm/mips3.s.
5905 *) When only the key is given to "enc", the IV is undefined. Print out
5906 an error message in this case.
5909 *) Handle special case when X509_NAME is empty in X509 printing routines.
5912 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5913 positive and less than q.
5916 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5917 used: it isn't thread safe and the add_lock_callback should handle
5919 [Paul Rose <Paul.Rose@bridge.com>]
5921 *) Verify that incoming data obeys the block size in
5922 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5926 [Ulf Möller, Bodo Möller]
5928 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5929 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5930 when fixing the server behaviour for backwards-compatible 'client
5931 hello' messages. (Note that the attack is impractical against
5932 SSL 3.0 and TLS 1.0 anyway because length and version checking
5933 means that the probability of guessing a valid ciphertext is
5934 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5937 Before 0.9.5, the countermeasure (hide the error by generating a
5938 random 'decryption result') did not work properly because
5939 ERR_clear_error() was missing, meaning that SSL_get_error() would
5940 detect the supposedly ignored error.
5942 Both problems are now fixed.
5945 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5946 (previously it was 1024).
5949 *) Fix for compatibility mode trust settings: ignore trust settings
5950 unless some valid trust or reject settings are present.
5953 *) Fix for blowfish EVP: its a variable length cipher.
5956 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5957 parameters in DSA public key structures and return an error in the
5958 DSA routines if parameters are absent.
5961 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5962 in the current directory if neither $RANDFILE nor $HOME was set.
5963 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5964 caused some confusion to Windows users who haven't defined $HOME.
5965 Thus RAND_file_name() is changed again: e_os.h can define a
5966 DEFAULT_HOME, which will be used if $HOME is not set.
5967 For Windows, we use "C:"; on other platforms, we still require
5968 environment variables.
5970 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5971 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5972 having multiple threads call RAND_poll() concurrently.
5975 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5976 combination of a flag and a thread ID variable.
5977 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5978 flag), *other* threads can enter ssleay_add_bytes without obeying
5979 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5980 that they do not hold after the first thread unsets add_do_not_lock).
5983 *) Change bctest again: '-x' expressions are not available in all
5987 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5989 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5990 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5992 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5993 the default extension for executables, if any. Also, make the perl
5994 scripts that use symlink() to test if it really exists and use "cp"
5995 if it doesn't. All this made OpenSSL compilable and installable in
5999 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6000 If SEQUENCE is length is indefinite just set c->slen to the total
6001 amount of data available.
6002 [Steve Henson, reported by shige@FreeBSD.org]
6003 [This change does not apply to 0.9.7.]
6005 *) Change bctest to avoid here-documents inside command substitution
6006 (workaround for FreeBSD /bin/sh bug).
6007 For compatibility with Ultrix, avoid shell functions (introduced
6008 in the bctest version that searches along $PATH).
6011 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6012 with des_encrypt() defined on some operating systems, like Solaris
6016 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6017 On the Importance of Eliminating Errors in Cryptographic
6018 Computations, J. Cryptology 14 (2001) 2, 101-119,
6019 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6022 *) MIPS assembler BIGNUM division bug fix.
6025 *) Disabled incorrect Alpha assembler code.
6028 *) Fix PKCS#7 decode routines so they correctly update the length
6029 after reading an EOC for the EXPLICIT tag.
6031 [This change does not apply to 0.9.7.]
6033 *) Fix bug in PKCS#12 key generation routines. This was triggered
6034 if a 3DES key was generated with a 0 initial byte. Include
6035 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6036 (but broken) behaviour.
6039 *) Enhance bctest to search for a working bc along $PATH and print
6041 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6043 *) Fix memory leaks in err.c: free err_data string if necessary;
6044 don't write to the wrong index in ERR_set_error_data.
6047 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6051 *) Replace rdtsc with _emit statements for VC++ version 5.
6052 [Jeremy Cooper <jeremy@baymoo.org>]
6054 *) Make it possible to reuse SSLv2 sessions.
6057 *) In copy_email() check for >= 0 as a return value for
6058 X509_NAME_get_index_by_NID() since 0 is a valid index.
6059 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6061 *) Avoid coredump with unsupported or invalid public keys by checking if
6062 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6063 PKCS7_verify() fails with non detached data.
6066 *) Don't use getenv in library functions when run as setuid/setgid.
6067 New function OPENSSL_issetugid().
6070 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6071 due to incorrect handling of multi-threading:
6073 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6075 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6077 3. Count how many times MemCheck_off() has been called so that
6078 nested use can be treated correctly. This also avoids
6079 inband-signalling in the previous code (which relied on the
6080 assumption that thread ID 0 is impossible).
6083 *) Add "-rand" option also to s_client and s_server.
6086 *) Fix CPU detection on Irix 6.x.
6087 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6088 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6090 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6093 [This change does not apply to 0.9.7.]
6095 *) Use the cached encoding of an X509_NAME structure rather than
6096 copying it. This is apparently the reason for the libsafe "errors"
6097 but the code is actually correct.
6100 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6101 Bleichenbacher's DSA attack.
6102 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6103 to be set and top=0 forces the highest bit to be set; top=-1 is new
6104 and leaves the highest bit random.
6105 [Ulf Moeller, Bodo Moeller]
6107 *) In the NCONF_...-based implementations for CONF_... queries
6108 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6109 a temporary CONF structure with the data component set to NULL
6110 (which gives segmentation faults in lh_retrieve).
6111 Instead, use NULL for the CONF pointer in CONF_get_string and
6112 CONF_get_number (which may use environment variables) and directly
6113 return NULL from CONF_get_section.
6116 *) Fix potential buffer overrun for EBCDIC.
6119 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6120 keyUsage if basicConstraints absent for a CA.
6123 *) Make SMIME_write_PKCS7() write mail header values with a format that
6124 is more generally accepted (no spaces before the semicolon), since
6125 some programs can't parse those values properly otherwise. Also make
6126 sure BIO's that break lines after each write do not create invalid
6130 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6131 macros previously used would not encode an empty SEQUENCE OF
6132 and break the signature.
6134 [This change does not apply to 0.9.7.]
6136 *) Zero the premaster secret after deriving the master secret in
6140 *) Add some EVP_add_digest_alias registrations (as found in
6141 OpenSSL_add_all_digests()) to SSL_library_init()
6142 aka OpenSSL_add_ssl_algorithms(). This provides improved
6143 compatibility with peers using X.509 certificates
6144 with unconventional AlgorithmIdentifier OIDs.
6147 *) Fix for Irix with NO_ASM.
6148 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6150 *) ./config script fixes.
6151 [Ulf Moeller, Richard Levitte]
6153 *) Fix 'openssl passwd -1'.
6156 *) Change PKCS12_key_gen_asc() so it can cope with non null
6157 terminated strings whose length is passed in the passlen
6158 parameter, for example from PEM callbacks. This was done
6159 by adding an extra length parameter to asc2uni().
6160 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6162 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6163 call failed, free the DSA structure.
6166 *) Fix to uni2asc() to cope with zero length Unicode strings.
6167 These are present in some PKCS#12 files.
6170 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6171 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6172 when writing a 32767 byte record.
6173 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6175 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6176 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6178 (RSA objects have a reference count access to which is protected
6179 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6180 so they are meant to be shared between threads.)
6181 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6182 "Reddie, Steven" <Steven.Reddie@ca.com>]
6184 *) Fix a deadlock in CRYPTO_mem_leaks().
6187 *) Use better test patterns in bntest.
6190 *) rand_win.c fix for Borland C.
6193 *) BN_rshift bugfix for n == 0.
6196 *) Add a 'bctest' script that checks for some known 'bc' bugs
6197 so that 'make test' does not abort just because 'bc' is broken.
6200 *) Store verify_result within SSL_SESSION also for client side to
6201 avoid potential security hole. (Re-used sessions on the client side
6202 always resulted in verify_result==X509_V_OK, not using the original
6203 result of the server certificate verification.)
6206 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6207 SSL3_RT_APPLICATION_DATA, return 0.
6208 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6212 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6213 releases, have been re-implemented by renaming the previous
6214 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6215 and ssl3_read_internal, respectively, and adding 'peek' parameters
6216 to them. The new ssl[23]_{read,peek} functions are calls to
6217 ssl[23]_read_internal with the 'peek' flag set appropriately.
6218 A 'peek' parameter has also been added to ssl3_read_bytes, which
6219 does the actual work for ssl3_read_internal.
6222 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6223 the method-specific "init()" handler. Also clean up ex_data after
6224 calling the method-specific "finish()" handler. Previously, this was
6225 happening the other way round.
6228 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6229 The previous value, 12, was not always sufficient for BN_mod_exp().
6232 *) Make sure that shared libraries get the internal name engine with
6233 the full version number and not just 0. This should mark the
6234 shared libraries as not backward compatible. Of course, this should
6235 be changed again when we can guarantee backward binary compatibility.
6238 *) Fix typo in get_cert_by_subject() in by_dir.c
6239 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6241 *) Rework the system to generate shared libraries:
6243 - Make note of the expected extension for the shared libraries and
6244 if there is a need for symbolic links from for example libcrypto.so.0
6245 to libcrypto.so.0.9.7. There is extended info in Configure for
6248 - Make as few rebuilds of the shared libraries as possible.
6250 - Still avoid linking the OpenSSL programs with the shared libraries.
6252 - When installing, install the shared libraries separately from the
6256 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6258 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6259 and not in SSL_clear because the latter is also used by the
6260 accept/connect functions; previously, the settings made by
6261 SSL_set_read_ahead would be lost during the handshake.
6262 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6264 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6265 Previously, it would create entries for disableed algorithms no
6269 *) Added several new manual pages for SSL_* function.
6272 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6274 *) In ssl23_get_client_hello, generate an error message when faced
6275 with an initial SSL 3.0/TLS record that is too small to contain the
6276 first two bytes of the ClientHello message, i.e. client_version.
6277 (Note that this is a pathologic case that probably has never happened
6278 in real life.) The previous approach was to use the version number
6279 from the record header as a substitute; but our protocol choice
6280 should not depend on that one because it is not authenticated
6281 by the Finished messages.
6284 *) More robust randomness gathering functions for Windows.
6285 [Jeffrey Altman <jaltman@columbia.edu>]
6287 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6288 not set then we don't setup the error code for issuer check errors
6289 to avoid possibly overwriting other errors which the callback does
6290 handle. If an application does set the flag then we assume it knows
6291 what it is doing and can handle the new informational codes
6295 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6296 a general "ANY" type, as such it should be able to decode anything
6297 including tagged types. However it didn't check the class so it would
6298 wrongly interpret tagged types in the same way as their universal
6299 counterpart and unknown types were just rejected. Changed so that the
6300 tagged and unknown types are handled in the same way as a SEQUENCE:
6301 that is the encoding is stored intact. There is also a new type
6302 "V_ASN1_OTHER" which is used when the class is not universal, in this
6303 case we have no idea what the actual type is so we just lump them all
6307 *) On VMS, stdout may very well lead to a file that is written to
6308 in a record-oriented fashion. That means that every write() will
6309 write a separate record, which will be read separately by the
6310 programs trying to read from it. This can be very confusing.
6312 The solution is to put a BIO filter in the way that will buffer
6313 text until a linefeed is reached, and then write everything a
6314 line at a time, so every record written will be an actual line,
6315 not chunks of lines and not (usually doesn't happen, but I've
6316 seen it once) several lines in one record. BIO_f_linebuffer() is
6319 Currently, it's a VMS-only method, because that's where it has
6320 been tested well enough.
6323 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6324 it can return incorrect results.
6325 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6326 but it was in 0.9.6-beta[12].)
6329 *) Disable the check for content being present when verifying detached
6330 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6331 include zero length content when signing messages.
6334 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6335 BIO_ctrl (for BIO pairs).
6338 *) Add DSO method for VMS.
6341 *) Bug fix: Montgomery multiplication could produce results with the
6345 *) Add RPM specification openssl.spec and modify it to build three
6346 packages. The default package contains applications, application
6347 documentation and run-time libraries. The devel package contains
6348 include files, static libraries and function documentation. The
6349 doc package contains the contents of the doc directory. The original
6350 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6353 *) Add a large number of documentation files for many SSL routines.
6354 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6356 *) Add a configuration entry for Sony News 4.
6357 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6359 *) Don't set the two most significant bits to one when generating a
6360 random number < q in the DSA library.
6363 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6364 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6365 the underlying transport is blocking) if a handshake took place.
6366 (The default behaviour is needed by applications such as s_client
6367 and s_server that use select() to determine when to use SSL_read;
6368 but for applications that know in advance when to expect data, it
6369 just makes things more complicated.)
6372 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6376 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6377 work better on such systems.
6378 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6380 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6381 Update PKCS12_parse() so it copies the friendlyName and the
6382 keyid to the certificates aux info.
6385 *) Fix bug in PKCS7_verify() which caused an infinite loop
6386 if there was more than one signature.
6387 [Sven Uszpelkat <su@celocom.de>]
6389 *) Major change in util/mkdef.pl to include extra information
6390 about each symbol, as well as presentig variables as well
6391 as functions. This change means that there's n more need
6392 to rebuild the .num files when some algorithms are excluded.
6395 *) Allow the verify time to be set by an application,
6396 rather than always using the current time.
6399 *) Phase 2 verify code reorganisation. The certificate
6400 verify code now looks up an issuer certificate by a
6401 number of criteria: subject name, authority key id
6402 and key usage. It also verifies self signed certificates
6403 by the same criteria. The main comparison function is
6404 X509_check_issued() which performs these checks.
6406 Lot of changes were necessary in order to support this
6407 without completely rewriting the lookup code.
6409 Authority and subject key identifier are now cached.
6411 The LHASH 'certs' is X509_STORE has now been replaced
6412 by a STACK_OF(X509_OBJECT). This is mainly because an
6413 LHASH can't store or retrieve multiple objects with
6414 the same hash value.
6416 As a result various functions (which were all internal
6417 use only) have changed to handle the new X509_STORE
6418 structure. This will break anything that messed round
6419 with X509_STORE internally.
6421 The functions X509_STORE_add_cert() now checks for an
6422 exact match, rather than just subject name.
6424 The X509_STORE API doesn't directly support the retrieval
6425 of multiple certificates matching a given criteria, however
6426 this can be worked round by performing a lookup first
6427 (which will fill the cache with candidate certificates)
6428 and then examining the cache for matches. This is probably
6429 the best we can do without throwing out X509_LOOKUP
6430 entirely (maybe later...).
6432 The X509_VERIFY_CTX structure has been enhanced considerably.
6434 All certificate lookup operations now go via a get_issuer()
6435 callback. Although this currently uses an X509_STORE it
6436 can be replaced by custom lookups. This is a simple way
6437 to bypass the X509_STORE hackery necessary to make this
6438 work and makes it possible to use more efficient techniques
6439 in future. A very simple version which uses a simple
6440 STACK for its trusted certificate store is also provided
6441 using X509_STORE_CTX_trusted_stack().
6443 The verify_cb() and verify() callbacks now have equivalents
6444 in the X509_STORE_CTX structure.
6446 X509_STORE_CTX also has a 'flags' field which can be used
6447 to customise the verify behaviour.
6450 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6451 excludes S/MIME capabilities.
6454 *) When a certificate request is read in keep a copy of the
6455 original encoding of the signed data and use it when outputing
6456 again. Signatures then use the original encoding rather than
6457 a decoded, encoded version which may cause problems if the
6458 request is improperly encoded.
6461 *) For consistency with other BIO_puts implementations, call
6462 buffer_write(b, ...) directly in buffer_puts instead of calling
6465 In BIO_puts, increment b->num_write as in BIO_write.
6466 [Peter.Sylvester@EdelWeb.fr]
6468 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6469 BN_zero, we may not return a BIGNUM with an array consisting of
6473 *) Avoid calling abort() from within the library when problems are
6474 detected, except if preprocessor symbols have been defined
6475 (such as REF_CHECK, BN_DEBUG etc.).
6478 *) New openssl application 'rsautl'. This utility can be
6479 used for low level RSA operations. DER public key
6480 BIO/fp routines also added.
6483 *) New Configure entry and patches for compiling on QNX 4.
6484 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6486 *) A demo state-machine implementation was sponsored by
6487 Nuron (http://www.nuron.com/) and is now available in
6488 demos/state_machine.
6491 *) New options added to the 'dgst' utility for signature
6492 generation and verification.
6495 *) Unrecognized PKCS#7 content types are now handled via a
6496 catch all ASN1_TYPE structure. This allows unsupported
6497 types to be stored as a "blob" and an application can
6498 encode and decode it manually.
6501 *) Fix various signed/unsigned issues to make a_strex.c
6503 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6505 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6506 length if passed a buffer. ASN1_INTEGER_to_BN failed
6507 if passed a NULL BN and its argument was negative.
6508 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6510 *) Modification to PKCS#7 encoding routines to output definite
6511 length encoding. Since currently the whole structures are in
6512 memory there's not real point in using indefinite length
6513 constructed encoding. However if OpenSSL is compiled with
6514 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6517 *) Added BIO_vprintf() and BIO_vsnprintf().
6520 *) Added more prefixes to parse for in the the strings written
6521 through a logging bio, to cover all the levels that are available
6522 through syslog. The prefixes are now:
6524 PANIC, EMERG, EMR => LOG_EMERG
6525 ALERT, ALR => LOG_ALERT
6526 CRIT, CRI => LOG_CRIT
6527 ERROR, ERR => LOG_ERR
6528 WARNING, WARN, WAR => LOG_WARNING
6529 NOTICE, NOTE, NOT => LOG_NOTICE
6530 INFO, INF => LOG_INFO
6531 DEBUG, DBG => LOG_DEBUG
6533 and as before, if none of those prefixes are present at the
6534 beginning of the string, LOG_ERR is chosen.
6536 On Win32, the LOG_* levels are mapped according to this:
6538 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6539 LOG_WARNING => EVENTLOG_WARNING_TYPE
6540 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6544 *) Made it possible to reconfigure with just the configuration
6545 argument "reconf" or "reconfigure". The command line arguments
6546 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6547 and are retrieved from there when reconfiguring.
6551 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6553 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6556 *) The obj_dat.pl script was messing up the sorting of object
6557 names. The reason was that it compared the quoted version
6558 of strings as a result "OCSP" > "OCSP Signing" because
6559 " > SPACE. Changed script to store unquoted versions of
6560 names and add quotes on output. It was also omitting some
6561 names from the lookup table if they were given a default
6562 value (that is if SN is missing it is given the same
6563 value as LN and vice versa), these are now added on the
6564 grounds that if an object has a name we should be able to
6565 look it up. Finally added warning output when duplicate
6566 short or long names are found.
6569 *) Changes needed for Tandem NSK.
6570 [Scott Uroff <scott@xypro.com>]
6572 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6573 RSA_padding_check_SSLv23(), special padding was never detected
6574 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6575 version rollback attacks was not effective.
6577 In s23_clnt.c, don't use special rollback-attack detection padding
6578 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6579 client; similarly, in s23_srvr.c, don't do the rollback check if
6580 SSL 2.0 is the only protocol enabled in the server.
6583 *) Make it possible to get hexdumps of unprintable data with 'openssl
6584 asn1parse'. By implication, the functions ASN1_parse_dump() and
6585 BIO_dump_indent() are added.
6588 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6589 these print out strings and name structures based on various
6590 flags including RFC2253 support and proper handling of
6591 multibyte characters. Added options to the 'x509' utility
6592 to allow the various flags to be set.
6595 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6596 Also change the functions X509_cmp_current_time() and
6597 X509_gmtime_adj() work with an ASN1_TIME structure,
6598 this will enable certificates using GeneralizedTime in validity
6599 dates to be checked.
6602 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6603 negative public key encodings) on by default,
6604 NO_NEG_PUBKEY_BUG can be set to disable it.
6607 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6608 content octets. An i2c_ASN1_OBJECT is unnecessary because
6609 the encoding can be trivially obtained from the structure.
6612 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6613 not read locks (CRYPTO_r_[un]lock).
6616 *) A first attempt at creating official support for shared
6617 libraries through configuration. I've kept it so the
6618 default is static libraries only, and the OpenSSL programs
6619 are always statically linked for now, but there are
6620 preparations for dynamic linking in place.
6621 This has been tested on Linux and Tru64.
6624 *) Randomness polling function for Win9x, as described in:
6625 Peter Gutmann, Software Generation of Practically Strong
6629 *) Fix so PRNG is seeded in req if using an already existing
6633 *) New options to smime application. -inform and -outform
6634 allow alternative formats for the S/MIME message including
6635 PEM and DER. The -content option allows the content to be
6636 specified separately. This should allow things like Netscape
6637 form signing output easier to verify.
6640 *) Fix the ASN1 encoding of tags using the 'long form'.
6643 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6644 STRING types. These convert content octets to and from the
6645 underlying type. The actual tag and length octets are
6646 already assumed to have been read in and checked. These
6647 are needed because all other string types have virtually
6648 identical handling apart from the tag. By having versions
6649 of the ASN1 functions that just operate on content octets
6650 IMPLICIT tagging can be handled properly. It also allows
6651 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6652 and ASN1_INTEGER are identical apart from the tag.
6655 *) Change the handling of OID objects as follows:
6657 - New object identifiers are inserted in objects.txt, following
6658 the syntax given in objects.README.
6659 - objects.pl is used to process obj_mac.num and create a new
6661 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6664 This is currently kind of a hack, and the perl code in objects.pl
6665 isn't very elegant, but it works as I intended. The simplest way
6666 to check that it worked correctly is to look in obj_dat.h and
6667 check the array nid_objs and make sure the objects haven't moved
6668 around (this is important!). Additions are OK, as well as
6669 consistent name changes.
6672 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6675 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6676 The given file adds to whatever has already been seeded into the
6677 random pool through the RANDFILE configuration file option or
6678 environment variable, or the default random state file.
6681 *) mkstack.pl now sorts each macro group into lexical order.
6682 Previously the output order depended on the order the files
6683 appeared in the directory, resulting in needless rewriting
6687 *) Patches to make OpenSSL compile under Win32 again. Mostly
6688 work arounds for the VC++ problem that it treats func() as
6689 func(void). Also stripped out the parts of mkdef.pl that
6690 added extra typesafe functions: these no longer exist.
6693 *) Reorganisation of the stack code. The macros are now all
6694 collected in safestack.h . Each macro is defined in terms of
6695 a "stack macro" of the form SKM_<name>(type, a, b). The
6696 DEBUG_SAFESTACK is now handled in terms of function casts,
6697 this has the advantage of retaining type safety without the
6698 use of additional functions. If DEBUG_SAFESTACK is not defined
6699 then the non typesafe macros are used instead. Also modified the
6700 mkstack.pl script to handle the new form. Needs testing to see
6701 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6702 the default if no major problems. Similar behaviour for ASN1_SET_OF
6703 and PKCS12_STACK_OF.
6706 *) When some versions of IIS use the 'NET' form of private key the
6707 key derivation algorithm is different. Normally MD5(password) is
6708 used as a 128 bit RC4 key. In the modified case
6709 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6710 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6711 as the old Netscape_RSA functions except they have an additional
6712 'sgckey' parameter which uses the modified algorithm. Also added
6713 an -sgckey command line option to the rsa utility. Thanks to
6714 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6715 algorithm to openssl-dev.
6718 *) The evp_local.h macros were using 'c.##kname' which resulted in
6719 invalid expansion on some systems (SCO 5.0.5 for example).
6720 Corrected to 'c.kname'.
6721 [Phillip Porch <root@theporch.com>]
6723 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6724 a STACK of email addresses from a certificate or request, these look
6725 in the subject name and the subject alternative name extensions and
6726 omit any duplicate addresses.
6729 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6730 This makes DSA verification about 2 % faster.
6733 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6734 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6735 plus overhead for 1024 bit moduli).
6736 This makes exponentiations about 0.5 % faster for 1024 bit
6737 exponents (as measured by "openssl speed rsa2048").
6740 *) Rename memory handling macros to avoid conflicts with other
6742 Malloc => OPENSSL_malloc
6743 Malloc_locked => OPENSSL_malloc_locked
6744 Realloc => OPENSSL_realloc
6745 Free => OPENSSL_free
6748 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6749 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6752 *) CygWin32 support.
6753 [John Jarvie <jjarvie@newsguy.com>]
6755 *) The type-safe stack code has been rejigged. It is now only compiled
6756 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6757 by default all type-specific stack functions are "#define"d back to
6758 standard stack functions. This results in more streamlined output
6759 but retains the type-safety checking possibilities of the original
6763 *) The STACK code has been cleaned up, and certain type declarations
6764 that didn't make a lot of sense have been brought in line. This has
6765 also involved a cleanup of sorts in safestack.h to more correctly
6766 map type-safe stack functions onto their plain stack counterparts.
6767 This work has also resulted in a variety of "const"ifications of
6768 lots of the code, especially "_cmp" operations which should normally
6769 be prototyped with "const" parameters anyway.
6772 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6773 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6774 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6775 where all of 'md' is used each time the PRNG is used, but 'state'
6776 is used only indexed by a cyclic counter. As entropy may not be
6777 well distributed from the beginning, 'md' is important as a
6778 chaining variable. However, the output function chains only half
6779 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6780 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6781 in all of 'state' being rewritten, with the new values depending
6782 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6785 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6786 the handshake is continued after ssl_verify_cert_chain();
6787 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6788 can lead to 'unexplainable' connection aborts later.
6789 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6791 *) Major EVP API cipher revision.
6792 Add hooks for extra EVP features. This allows various cipher
6793 parameters to be set in the EVP interface. Support added for variable
6794 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6795 setting of RC2 and RC5 parameters.
6797 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6800 Remove lots of duplicated code from the EVP library. For example *every*
6801 cipher init() function handles the 'iv' in the same way according to the
6802 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6803 for CFB and OFB modes they zero ctx->num.
6805 New functionality allows removal of S/MIME code RC2 hack.
6807 Most of the routines have the same form and so can be declared in terms
6810 By shifting this to the top level EVP_CipherInit() it can be removed from
6811 all individual ciphers. If the cipher wants to handle IVs or keys
6812 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6815 Change lots of functions like EVP_EncryptUpdate() to now return a
6816 value: although software versions of the algorithms cannot fail
6817 any installed hardware versions can.
6820 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6821 this option is set, tolerate broken clients that send the negotiated
6822 protocol version number instead of the requested protocol version
6826 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6827 i.e. non-zero for export ciphersuites, zero otherwise.
6828 Previous versions had this flag inverted, inconsistent with
6829 rsa_tmp_cb (..._TMP_RSA_CB).
6830 [Bodo Moeller; problem reported by Amit Chopra]
6832 *) Add missing DSA library text string. Work around for some IIS
6833 key files with invalid SEQUENCE encoding.
6836 *) Add a document (doc/standards.txt) that list all kinds of standards
6837 and so on that are implemented in OpenSSL.
6840 *) Enhance c_rehash script. Old version would mishandle certificates
6841 with the same subject name hash and wouldn't handle CRLs at all.
6842 Added -fingerprint option to crl utility, to support new c_rehash
6846 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6849 *) Fix for SSL server purpose checking. Server checking was
6850 rejecting certificates which had extended key usage present
6851 but no ssl client purpose.
6852 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6854 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6855 is a little unclear about how a blank password is handled.
6856 Since the password in encoded as a BMPString with terminating
6857 double NULL a zero length password would end up as just the
6858 double NULL. However no password at all is different and is
6859 handled differently in the PKCS#12 key generation code. NS
6860 treats a blank password as zero length. MSIE treats it as no
6861 password on export: but it will try both on import. We now do
6862 the same: PKCS12_parse() tries zero length and no password if
6863 the password is set to "" or NULL (NULL is now a valid password:
6864 it wasn't before) as does the pkcs12 application.
6867 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6868 perror when PEM_read_bio_X509_REQ fails, the error message must
6869 be obtained from the error queue.
6872 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6873 it in ERR_remove_state if appropriate, and change ERR_get_state
6874 accordingly to avoid race conditions (this is necessary because
6875 thread_hash is no longer constant once set).
6878 *) Bugfix for linux-elf makefile.one.
6881 *) RSA_get_default_method() will now cause a default
6882 RSA_METHOD to be chosen if one doesn't exist already.
6883 Previously this was only set during a call to RSA_new()
6884 or RSA_new_method(NULL) meaning it was possible for
6885 RSA_get_default_method() to return NULL.
6888 *) Added native name translation to the existing DSO code
6889 that will convert (if the flag to do so is set) filenames
6890 that are sufficiently small and have no path information
6891 into a canonical native form. Eg. "blah" converted to
6892 "libblah.so" or "blah.dll" etc.
6895 *) New function ERR_error_string_n(e, buf, len) which is like
6896 ERR_error_string(e, buf), but writes at most 'len' bytes
6897 including the 0 terminator. For ERR_error_string_n, 'buf'
6899 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6901 *) CONF library reworked to become more general. A new CONF
6902 configuration file reader "class" is implemented as well as a
6903 new functions (NCONF_*, for "New CONF") to handle it. The now
6904 old CONF_* functions are still there, but are reimplemented to
6905 work in terms of the new functions. Also, a set of functions
6906 to handle the internal storage of the configuration data is
6907 provided to make it easier to write new configuration file
6908 reader "classes" (I can definitely see something reading a
6909 configuration file in XML format, for example), called _CONF_*,
6910 or "the configuration storage API"...
6912 The new configuration file reading functions are:
6914 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6915 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6917 NCONF_default, NCONF_WIN32
6919 NCONF_dump_fp, NCONF_dump_bio
6921 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6922 NCONF_new creates a new CONF object. This works in the same way
6923 as other interfaces in OpenSSL, like the BIO interface.
6924 NCONF_dump_* dump the internal storage of the configuration file,
6925 which is useful for debugging. All other functions take the same
6926 arguments as the old CONF_* functions wth the exception of the
6927 first that must be a `CONF *' instead of a `LHASH *'.
6929 To make it easer to use the new classes with the old CONF_* functions,
6930 the function CONF_set_default_method is provided.
6933 *) Add '-tls1' option to 'openssl ciphers', which was already
6934 mentioned in the documentation but had not been implemented.
6935 (This option is not yet really useful because even the additional
6936 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6939 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6940 OpenSSL-based applications) load shared libraries and bind to
6941 them in a portable way.
6942 [Geoff Thorpe, with contributions from Richard Levitte]
6944 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6946 *) Make sure _lrotl and _lrotr are only used with MSVC.
6948 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6949 (the default implementation of RAND_status).
6951 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6952 to '-clrext' (= clear extensions), as intended and documented.
6953 [Bodo Moeller; inconsistency pointed out by Michael Attili
6956 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6957 was larger than the MD block size.
6958 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6960 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6961 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6962 using the passed key: if the passed key was a private key the result
6963 of X509_print(), for example, would be to print out all the private key
6967 *) des_quad_cksum() byte order bug fix.
6968 [Ulf Möller, using the problem description in krb4-0.9.7, where
6969 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6971 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6973 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6975 *) For easily testing in shell scripts whether some command
6976 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6977 returns with exit code 0 iff no command of the given name is available.
6978 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6979 the output goes to stdout and nothing is printed to stderr.
6980 Additional arguments are always ignored.
6982 Since for each cipher there is a command of the same name,
6983 the 'no-cipher' compilation switches can be tested this way.
6985 ('openssl no-XXX' is not able to detect pseudo-commands such
6986 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6989 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6992 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6993 is set; it will be thrown away anyway because each handshake creates
6995 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6996 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6997 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6998 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7001 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7002 'Q' and 'R' lose their special meanings (quit/renegotiate).
7003 This is part of what -quiet does; unlike -quiet, -ign_eof
7004 does not suppress any output.
7007 *) Add compatibility options to the purpose and trust code. The
7008 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7009 accepts a certificate or CA, this was the previous behaviour,
7010 with all the associated security issues.
7012 X509_TRUST_COMPAT is the old trust behaviour: only and
7013 automatically trust self signed roots in certificate store. A
7014 new trust setting X509_TRUST_DEFAULT is used to specify that
7015 a purpose has no associated trust setting and it should instead
7016 use the value in the default purpose.
7019 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7020 and fix a memory leak.
7023 *) In util/mkerr.pl (which implements 'make errors'), preserve
7024 reason strings from the previous version of the .c file, as
7025 the default to have only downcase letters (and digits) in
7026 automatically generated reasons codes is not always appropriate.
7029 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7030 using strerror. Previously, ERR_reason_error_string() returned
7031 library names as reason strings for SYSerr; but SYSerr is a special
7032 case where small numbers are errno values, not library numbers.
7035 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7036 converts DSA parameters into DH parameters. (When creating parameters,
7037 DSA_generate_parameters is used.)
7040 *) Include 'length' (recommended exponent length) in C code generated
7041 by 'openssl dhparam -C'.
7044 *) The second argument to set_label in perlasm was already being used
7045 so couldn't be used as a "file scope" flag. Moved to third argument
7049 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7050 instead of RAND_bytes for encryption IVs and salts.
7053 *) Include RAND_status() into RAND_METHOD instead of implementing
7054 it only for md_rand.c Otherwise replacing the PRNG by calling
7055 RAND_set_rand_method would be impossible.
7058 *) Don't let DSA_generate_key() enter an infinite loop if the random
7059 number generation fails.
7062 *) New 'rand' application for creating pseudo-random output.
7065 *) Added configuration support for Linux/IA64
7066 [Rolf Haberrecker <rolf@suse.de>]
7068 *) Assembler module support for Mingw32.
7071 *) Shared library support for HPUX (in shlib/).
7072 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7074 *) Shared library support for Solaris gcc.
7075 [Lutz Behnke <behnke@trustcenter.de>]
7077 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7079 *) PKCS7_encrypt() was adding text MIME headers twice because they
7080 were added manually and by SMIME_crlf_copy().
7083 *) In bntest.c don't call BN_rand with zero bits argument.
7084 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7086 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7087 case was implemented. This caused BN_div_recp() to fail occasionally.
7090 *) Add an optional second argument to the set_label() in the perl
7091 assembly language builder. If this argument exists and is set
7092 to 1 it signals that the assembler should use a symbol whose
7093 scope is the entire file, not just the current function. This
7094 is needed with MASM which uses the format label:: for this scope.
7095 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7097 *) Change the ASN1 types so they are typedefs by default. Before
7098 almost all types were #define'd to ASN1_STRING which was causing
7099 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7103 *) Change names of new functions to the new get1/get0 naming
7104 convention: After 'get1', the caller owns a reference count
7105 and has to call ..._free; 'get0' returns a pointer to some
7106 data structure without incrementing reference counters.
7107 (Some of the existing 'get' functions increment a reference
7108 counter, some don't.)
7109 Similarly, 'set1' and 'add1' functions increase reference
7110 counters or duplicate objects.
7113 *) Allow for the possibility of temp RSA key generation failure:
7114 the code used to assume it always worked and crashed on failure.
7117 *) Fix potential buffer overrun problem in BIO_printf().
7118 [Ulf Möller, using public domain code by Patrick Powell; problem
7119 pointed out by David Sacerdote <das33@cornell.edu>]
7121 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7122 RAND_egd() and RAND_status(). In the command line application,
7123 the EGD socket can be specified like a seed file using RANDFILE
7127 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7128 Some CAs (e.g. Verisign) distribute certificates in this form.
7131 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7132 list to exclude them. This means that no special compilation option
7133 is needed to use anonymous DH: it just needs to be included in the
7137 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7138 EVP_MD_type. The old functionality is available in a new macro called
7139 EVP_MD_md(). Change code that uses it and update docs.
7142 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7143 where the 'void *' argument is replaced by a function pointer argument.
7144 Previously 'void *' was abused to point to functions, which works on
7145 many platforms, but is not correct. As these functions are usually
7146 called by macros defined in OpenSSL header files, most source code
7147 should work without changes.
7150 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7151 sections with information on -D... compiler switches used for
7152 compiling the library so that applications can see them. To enable
7153 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7154 must be defined. E.g.,
7155 #define OPENSSL_ALGORITHM_DEFINES
7156 #include <openssl/opensslconf.h>
7157 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7158 [Richard Levitte, Ulf and Bodo Möller]
7160 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7164 *) Change the 'other' type in certificate aux info to a STACK_OF
7165 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7166 the required ASN1 format: arbitrary types determined by an OID.
7169 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7170 argument to 'req'. This is not because the function is newer or
7171 better than others it just uses the work 'NEW' in the certificate
7172 request header lines. Some software needs this.
7175 *) Reorganise password command line arguments: now passwords can be
7176 obtained from various sources. Delete the PEM_cb function and make
7177 it the default behaviour: i.e. if the callback is NULL and the
7178 usrdata argument is not NULL interpret it as a null terminated pass
7179 phrase. If usrdata and the callback are NULL then the pass phrase
7180 is prompted for as usual.
7183 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7184 the support is automatically enabled. The resulting binaries will
7185 autodetect the card and use it if present.
7186 [Ben Laurie and Compaq Inc.]
7188 *) Work around for Netscape hang bug. This sends certificate request
7189 and server done in one record. Since this is perfectly legal in the
7190 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7191 the bugs/SSLv3 entry for more info.
7194 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7197 *) Add -rand argument to smime and pkcs12 applications and read/write
7201 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7204 *) Add command line password options to the remaining applications.
7207 *) Bug fix for BN_div_recp() for numerators with an even number of
7211 *) More tests in bntest.c, and changed test_bn output.
7214 *) ./config recognizes MacOS X now.
7217 *) Bug fix for BN_div() when the first words of num and divsor are
7218 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7221 *) Add support for various broken PKCS#8 formats, and command line
7222 options to produce them.
7225 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7226 get temporary BIGNUMs from a BN_CTX.
7229 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7233 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7234 include a #define from the old name to the new. The original intent
7235 was that statically linked binaries could for example just call
7236 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7237 link with digests. This never worked becayse SSLeay_add_all_digests()
7238 and SSLeay_add_all_ciphers() were in the same source file so calling
7239 one would link with the other. They are now in separate source files.
7242 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7245 *) Use a less unusual form of the Miller-Rabin primality test (it used
7246 a binary algorithm for exponentiation integrated into the Miller-Rabin
7247 loop, our standard modexp algorithms are faster).
7250 *) Support for the EBCDIC character set completed.
7251 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7253 *) Source code cleanups: use const where appropriate, eliminate casts,
7254 use void * instead of char * in lhash.
7257 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7258 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7259 this the server could overwrite ephemeral keys that the client
7263 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7264 using 50 iterations of the Rabin-Miller test.
7266 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7267 iterations of the Rabin-Miller test as required by the appendix
7268 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7269 As BN_is_prime_fasttest includes trial division, DSA parameter
7270 generation becomes much faster.
7272 This implies a change for the callback functions in DSA_is_prime
7273 and DSA_generate_parameters: The callback function is called once
7274 for each positive witness in the Rabin-Miller test, not just
7275 occasionally in the inner loop; and the parameters to the
7276 callback function now provide an iteration count for the outer
7277 loop rather than for the current invocation of the inner loop.
7278 DSA_generate_parameters additionally can call the callback
7279 function with an 'iteration count' of -1, meaning that a
7280 candidate has passed the trial division test (when q is generated
7281 from an application-provided seed, trial division is skipped).
7284 *) New function BN_is_prime_fasttest that optionally does trial
7285 division before starting the Rabin-Miller test and has
7286 an additional BN_CTX * argument (whereas BN_is_prime always
7287 has to allocate at least one BN_CTX).
7288 'callback(1, -1, cb_arg)' is called when a number has passed the
7289 trial division stage.
7292 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7296 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7299 *) New function BN_pseudo_rand().
7302 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7303 bignum version of BN_from_montgomery() with the working code from
7304 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7308 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7309 made it impossible to use the same SSL_SESSION data structure in
7310 SSL2 clients in multiple threads.
7313 *) The return value of RAND_load_file() no longer counts bytes obtained
7314 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7315 to seed the PRNG (previously an explicit byte count was required).
7316 [Ulf Möller, Bodo Möller]
7318 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7319 used (char *) instead of (void *) and had casts all over the place.
7322 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7325 *) Retain source code compatibility for BN_prime_checks macro:
7326 BN_is_prime(..., BN_prime_checks, ...) now uses
7327 BN_prime_checks_for_size to determine the appropriate number of
7328 Rabin-Miller iterations.
7331 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7332 DH_CHECK_P_NOT_SAFE_PRIME.
7333 (Check if this is true? OpenPGP calls them "strong".)
7336 *) Merge the functionality of "dh" and "gendh" programs into a new program
7337 "dhparam". The old programs are retained for now but will handle DH keys
7338 (instead of parameters) in future.
7341 *) Make the ciphers, s_server and s_client programs check the return values
7342 when a new cipher list is set.
7345 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7346 ciphers. Before when the 56bit ciphers were enabled the sorting was
7349 The syntax for the cipher sorting has been extended to support sorting by
7350 cipher-strength (using the strength_bits hard coded in the tables).
7351 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7353 Fix a bug in the cipher-command parser: when supplying a cipher command
7354 string with an "undefined" symbol (neither command nor alphanumeric
7355 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7356 an error is flagged.
7358 Due to the strength-sorting extension, the code of the
7359 ssl_create_cipher_list() function was completely rearranged. I hope that
7360 the readability was also increased :-)
7361 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7363 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7364 for the first serial number and places 2 in the serial number file. This
7365 avoids problems when the root CA is created with serial number zero and
7366 the first user certificate has the same issuer name and serial number
7370 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7371 the new code. Add documentation for this stuff.
7374 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7375 X509_*() to X509at_*() on the grounds that they don't handle X509
7376 structures and behave in an analagous way to the X509v3 functions:
7377 they shouldn't be called directly but wrapper functions should be used
7380 So we also now have some wrapper functions that call the X509at functions
7381 when passed certificate requests. (TO DO: similar things can be done with
7382 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7383 things. Some of these need some d2i or i2d and print functionality
7384 because they handle more complex structures.)
7387 *) Add missing #ifndefs that caused missing symbols when building libssl
7388 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7389 NO_RSA in ssl/s2*.c.
7390 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7392 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7393 has a return value which indicates the quality of the random data
7394 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7395 error queue. New function RAND_pseudo_bytes() generates output that is
7396 guaranteed to be unique but not unpredictable. RAND_add is like
7397 RAND_seed, but takes an extra argument for an entropy estimate
7398 (RAND_seed always assumes full entropy).
7401 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7402 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7403 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7404 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7405 false-positive rate of at most 2^-80 for random input.
7408 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7411 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7412 in the 0.9.5 release), this returns the chain
7413 from an X509_CTX structure with a dup of the stack and all
7414 the X509 reference counts upped: so the stack will exist
7415 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7418 Also make SSL_SESSION_print() print out the verify return
7422 *) Add manpage for the pkcs12 command. Also change the default
7423 behaviour so MAC iteration counts are used unless the new
7424 -nomaciter option is used. This improves file security and
7425 only older versions of MSIE (4.0 for example) need it.
7428 *) Honor the no-xxx Configure options when creating .DEF files.
7431 *) Add PKCS#10 attributes to field table: challengePassword,
7432 unstructuredName and unstructuredAddress. These are taken from
7433 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7434 international characters are used.
7436 More changes to X509_ATTRIBUTE code: allow the setting of types
7437 based on strings. Remove the 'loc' parameter when adding
7438 attributes because these will be a SET OF encoding which is sorted
7442 *) Initial changes to the 'req' utility to allow request generation
7443 automation. This will allow an application to just generate a template
7444 file containing all the field values and have req construct the
7447 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7448 used all over the place including certificate requests and PKCS#7
7449 structures. They are currently handled manually where necessary with
7450 some primitive wrappers for PKCS#7. The new functions behave in a
7451 manner analogous to the X509 extension functions: they allow
7452 attributes to be looked up by NID and added.
7454 Later something similar to the X509V3 code would be desirable to
7455 automatically handle the encoding, decoding and printing of the
7456 more complex types. The string types like challengePassword can
7457 be handled by the string table functions.
7459 Also modified the multi byte string table handling. Now there is
7460 a 'global mask' which masks out certain types. The table itself
7461 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7462 is useful when for example there is only one permissible type
7463 (as in countryName) and using the mask might result in no valid
7467 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7468 SSL_get_peer_finished to allow applications to obtain the latest
7469 Finished messages sent to the peer or expected from the peer,
7470 respectively. (SSL_get_peer_finished is usually the Finished message
7471 actually received from the peer, otherwise the protocol will be aborted.)
7473 As the Finished message are message digests of the complete handshake
7474 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7475 be used for external authentication procedures when the authentication
7476 provided by SSL/TLS is not desired or is not enough.
7479 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7480 the host supports BWX extension and if Compaq C is present on the
7481 $PATH. Just exploiting of the BWX extension results in 20-30%
7482 performance kick for some algorithms, e.g. DES and RC4 to mention
7483 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7487 *) Add support for MS "fast SGC". This is arguably a violation of the
7488 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7489 weak crypto and after checking the certificate is SGC a second one
7490 with strong crypto. MS SGC stops the first handshake after receiving
7491 the server certificate message and sends a second client hello. Since
7492 a server will typically do all the time consuming operations before
7493 expecting any further messages from the client (server key exchange
7494 is the most expensive) there is little difference between the two.
7496 To get OpenSSL to support MS SGC we have to permit a second client
7497 hello message after we have sent server done. In addition we have to
7498 reset the MAC if we do get this second client hello.
7501 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7502 if a DER encoded private key is RSA or DSA traditional format. Changed
7503 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7504 format DER encoded private key. Newer code should use PKCS#8 format which
7505 has the key type encoded in the ASN1 structure. Added DER private key
7506 support to pkcs8 application.
7509 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7510 ciphersuites has been selected (as required by the SSL 3/TLS 1
7511 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7512 is set, we interpret this as a request to violate the specification
7513 (the worst that can happen is a handshake failure, and 'correct'
7514 behaviour would result in a handshake failure anyway).
7517 *) In SSL_CTX_add_session, take into account that there might be multiple
7518 SSL_SESSION structures with the same session ID (e.g. when two threads
7519 concurrently obtain them from an external cache).
7520 The internal cache can handle only one SSL_SESSION with a given ID,
7521 so if there's a conflict, we now throw out the old one to achieve
7525 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7526 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7527 some routines that use cipher OIDs: some ciphers do not have OIDs
7528 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7532 *) Simplify the trust setting structure and code. Now we just have
7533 two sequences of OIDs for trusted and rejected settings. These will
7534 typically have values the same as the extended key usage extension
7535 and any application specific purposes.
7537 The trust checking code now has a default behaviour: it will just
7538 check for an object with the same NID as the passed id. Functions can
7539 be provided to override either the default behaviour or the behaviour
7540 for a given id. SSL client, server and email already have functions
7541 in place for compatibility: they check the NID and also return "trusted"
7542 if the certificate is self signed.
7545 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7546 traditional format into an EVP_PKEY structure.
7549 *) Add a password callback function PEM_cb() which either prompts for
7550 a password if usr_data is NULL or otherwise assumes it is a null
7551 terminated password. Allow passwords to be passed on command line
7552 environment or config files in a few more utilities.
7555 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7556 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7557 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7558 Update documentation.
7561 *) Support for ASN1 "NULL" type. This could be handled before by using
7562 ASN1_TYPE but there wasn't any function that would try to read a NULL
7563 and produce an error if it couldn't. For compatibility we also have
7564 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7565 don't allocate anything because they don't need to.
7568 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7570 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7572 *) Rebuild of the memory allocation routines used by OpenSSL code and
7573 possibly others as well. The purpose is to make an interface that
7574 provide hooks so anyone can build a separate set of allocation and
7575 deallocation routines to be used by OpenSSL, for example memory
7576 pool implementations, or something else, which was previously hard
7577 since Malloc(), Realloc() and Free() were defined as macros having
7578 the values malloc, realloc and free, respectively (except for Win32
7579 compilations). The same is provided for memory debugging code.
7580 OpenSSL already comes with functionality to find memory leaks, but
7581 this gives people a chance to debug other memory problems.
7583 With these changes, a new set of functions and macros have appeared:
7585 CRYPTO_set_mem_debug_functions() [F]
7586 CRYPTO_get_mem_debug_functions() [F]
7587 CRYPTO_dbg_set_options() [F]
7588 CRYPTO_dbg_get_options() [F]
7589 CRYPTO_malloc_debug_init() [M]
7591 The memory debug functions are NULL by default, unless the library
7592 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7593 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7594 gives the standard debugging functions that come with OpenSSL) or
7595 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7596 provided by the library user) must be used. When the standard
7597 debugging functions are used, CRYPTO_dbg_set_options can be used to
7598 request additional information:
7599 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7600 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7602 Also, things like CRYPTO_set_mem_functions will always give the
7603 expected result (the new set of functions is used for allocation
7604 and deallocation) at all times, regardless of platform and compiler
7607 To finish it up, some functions that were never use in any other
7608 way than through macros have a new API and new semantic:
7611 CRYPTO_dbg_realloc()
7614 All macros of value have retained their old syntax.
7615 [Richard Levitte and Bodo Moeller]
7617 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7618 ordering of SMIMECapabilities wasn't in "strength order" and there
7619 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7623 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7624 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7625 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7627 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7628 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7629 functionality to handle multipart/signed properly) and a utility
7630 called 'smime' to call all this stuff. This is based on code I
7631 originally wrote for Celo who have kindly allowed it to be
7632 included in OpenSSL.
7635 *) Add variants des_set_key_checked and des_set_key_unchecked of
7636 des_set_key (aka des_key_sched). Global variable des_check_key
7637 decides which of these is called by des_set_key; this way
7638 des_check_key behaves as it always did, but applications and
7639 the library itself, which was buggy for des_check_key == 1,
7640 have a cleaner way to pick the version they need.
7643 *) New function PKCS12_newpass() which changes the password of a
7647 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7648 dynamic mix. In both cases the ids can be used as an index into the
7649 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7650 functions so they accept a list of the field values and the
7651 application doesn't need to directly manipulate the X509_TRUST
7655 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7659 *) Modify the way the V3 extension code looks up extensions. This now
7660 works in a similar way to the object code: we have some "standard"
7661 extensions in a static table which is searched with OBJ_bsearch()
7662 and the application can add dynamic ones if needed. The file
7663 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7664 updated whenever a new extension is added to the core code and kept
7665 in ext_nid order. There is a simple program 'tabtest.c' which checks
7666 this. New extensions are not added too often so this file can readily
7667 be maintained manually.
7669 There are two big advantages in doing things this way. The extensions
7670 can be looked up immediately and no longer need to be "added" using
7671 X509V3_add_standard_extensions(): this function now does nothing.
7672 [Side note: I get *lots* of email saying the extension code doesn't
7673 work because people forget to call this function]
7674 Also no dynamic allocation is done unless new extensions are added:
7675 so if we don't add custom extensions there is no need to call
7676 X509V3_EXT_cleanup().
7679 *) Modify enc utility's salting as follows: make salting the default. Add a
7680 magic header, so unsalted files fail gracefully instead of just decrypting
7681 to garbage. This is because not salting is a big security hole, so people
7682 should be discouraged from doing it.
7685 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7686 digest to be passed on the command line but it only used this
7687 parameter when signing a certificate. Modified so all relevant
7688 operations are affected by the digest parameter including the
7689 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7690 DSA key was used because it didn't fix the digest.
7693 *) Initial certificate chain verify code. Currently tests the untrusted
7694 certificates for consistency with the verify purpose (which is set
7695 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7697 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7698 this is because it will reject chains with invalid extensions whereas
7699 every previous version of OpenSSL and SSLeay made no checks at all.
7701 Trust code: checks the root CA for the relevant trust settings. Trust
7702 settings have an initial value consistent with the verify purpose: e.g.
7703 if the verify purpose is for SSL client use it expects the CA to be
7704 trusted for SSL client use. However the default value can be changed to
7705 permit custom trust settings: one example of this would be to only trust
7706 certificates from a specific "secure" set of CAs.
7708 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7709 which should be used for version portability: especially since the
7710 verify structure is likely to change more often now.
7712 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7713 to set them. If not set then assume SSL clients will verify SSL servers
7716 Two new options to the verify program: -untrusted allows a set of
7717 untrusted certificates to be passed in and -purpose which sets the
7718 intended purpose of the certificate. If a purpose is set then the
7719 new chain verify code is used to check extension consistency.
7722 *) Support for the authority information access extension.
7725 *) Modify RSA and DSA PEM read routines to transparently handle
7726 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7727 public keys in a format compatible with certificate
7728 SubjectPublicKeyInfo structures. Unfortunately there were already
7729 functions called *_PublicKey_* which used various odd formats so
7730 these are retained for compatibility: however the DSA variants were
7731 never in a public release so they have been deleted. Changed dsa/rsa
7732 utilities to handle the new format: note no releases ever handled public
7733 keys so we should be OK.
7735 The primary motivation for this change is to avoid the same fiasco
7736 that dogs private keys: there are several incompatible private key
7737 formats some of which are standard and some OpenSSL specific and
7738 require various evil hacks to allow partial transparent handling and
7739 even then it doesn't work with DER formats. Given the option anything
7740 other than PKCS#8 should be dumped: but the other formats have to
7741 stay in the name of compatibility.
7743 With public keys and the benefit of hindsight one standard format
7744 is used which works with EVP_PKEY, RSA or DSA structures: though
7745 it clearly returns an error if you try to read the wrong kind of key.
7747 Added a -pubkey option to the 'x509' utility to output the public key.
7748 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7749 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7750 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7751 that do the same as the EVP_PKEY_assign_*() except they up the
7752 reference count of the added key (they don't "swallow" the
7756 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7757 CRLs would fail if the file contained no certificates or no CRLs:
7758 added a new function to read in both types and return the number
7759 read: this means that if none are read it will be an error. The
7760 DER versions of the certificate and CRL reader would always fail
7761 because it isn't possible to mix certificates and CRLs in DER format
7762 without choking one or the other routine. Changed this to just read
7763 a certificate: this is the best we can do. Also modified the code
7764 in apps/verify.c to take notice of return codes: it was previously
7765 attempting to read in certificates from NULL pointers and ignoring
7766 any errors: this is one reason why the cert and CRL reader seemed
7767 to work. It doesn't check return codes from the default certificate
7768 routines: these may well fail if the certificates aren't installed.
7771 *) Code to support otherName option in GeneralName.
7774 *) First update to verify code. Change the verify utility
7775 so it warns if it is passed a self signed certificate:
7776 for consistency with the normal behaviour. X509_verify
7777 has been modified to it will now verify a self signed
7778 certificate if *exactly* the same certificate appears
7779 in the store: it was previously impossible to trust a
7780 single self signed certificate. This means that:
7781 openssl verify ss.pem
7782 now gives a warning about a self signed certificate but
7783 openssl verify -CAfile ss.pem ss.pem
7787 *) For servers, store verify_result in SSL_SESSION data structure
7788 (and add it to external session representation).
7789 This is needed when client certificate verifications fails,
7790 but an application-provided verification callback (set by
7791 SSL_CTX_set_cert_verify_callback) allows accepting the session
7792 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7793 but returns 1): When the session is reused, we have to set
7794 ssl->verify_result to the appropriate error code to avoid
7796 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7798 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7799 case in PKCS7_dataInit() where the signed PKCS7 structure
7800 didn't contain any existing data because it was being created.
7801 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7803 *) Add a salt to the key derivation routines in enc.c. This
7804 forms the first 8 bytes of the encrypted file. Also add a
7805 -S option to allow a salt to be input on the command line.
7808 *) New function X509_cmp(). Oddly enough there wasn't a function
7809 to compare two certificates. We do this by working out the SHA1
7810 hash and comparing that. X509_cmp() will be needed by the trust
7814 *) SSL_get1_session() is like SSL_get_session(), but increments
7815 the reference count in the SSL_SESSION returned.
7816 [Geoff Thorpe <geoff@eu.c2.net>]
7818 *) Fix for 'req': it was adding a null to request attributes.
7819 Also change the X509_LOOKUP and X509_INFO code to handle
7820 certificate auxiliary information.
7823 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7827 *) Add the possibility to add extra information to the memory leak
7828 detecting output, to form tracebacks, showing from where each
7829 allocation was originated: CRYPTO_push_info("constant string") adds
7830 the string plus current file name and line number to a per-thread
7831 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7832 is like calling CYRPTO_pop_info() until the stack is empty.
7833 Also updated memory leak detection code to be multi-thread-safe.
7836 *) Add options -text and -noout to pkcs7 utility and delete the
7837 encryption options which never did anything. Update docs.
7840 *) Add options to some of the utilities to allow the pass phrase
7841 to be included on either the command line (not recommended on
7842 OSes like Unix) or read from the environment. Update the
7843 manpages and fix a few bugs.
7846 *) Add a few manpages for some of the openssl commands.
7849 *) Fix the -revoke option in ca. It was freeing up memory twice,
7850 leaking and not finding already revoked certificates.
7853 *) Extensive changes to support certificate auxiliary information.
7854 This involves the use of X509_CERT_AUX structure and X509_AUX
7855 functions. An X509_AUX function such as PEM_read_X509_AUX()
7856 can still read in a certificate file in the usual way but it
7857 will also read in any additional "auxiliary information". By
7858 doing things this way a fair degree of compatibility can be
7859 retained: existing certificates can have this information added
7860 using the new 'x509' options.
7862 Current auxiliary information includes an "alias" and some trust
7863 settings. The trust settings will ultimately be used in enhanced
7864 certificate chain verification routines: currently a certificate
7865 can only be trusted if it is self signed and then it is trusted
7869 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7870 The problem was that one of the replacement routines had not been working
7871 since SSLeay releases. For now the offending routine has been replaced
7872 with non-optimised assembler. Even so, this now gives around 95%
7873 performance improvement for 1024 bit RSA signs.
7876 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7877 handling. Most clients have the effective key size in bits equal to
7878 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7879 A few however don't do this and instead use the size of the decrypted key
7880 to determine the RC2 key length and the AlgorithmIdentifier to determine
7881 the effective key length. In this case the effective key length can still
7882 be 40 bits but the key length can be 168 bits for example. This is fixed
7883 by manually forcing an RC2 key into the EVP_PKEY structure because the
7884 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7885 the key length and effective key length are equal.
7888 *) Add a bunch of functions that should simplify the creation of
7889 X509_NAME structures. Now you should be able to do:
7890 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7891 and have it automatically work out the correct field type and fill in
7892 the structures. The more adventurous can try:
7893 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7894 and it will (hopefully) work out the correct multibyte encoding.
7897 *) Change the 'req' utility to use the new field handling and multibyte
7898 copy routines. Before the DN field creation was handled in an ad hoc
7899 way in req, ca, and x509 which was rather broken and didn't support
7900 BMPStrings or UTF8Strings. Since some software doesn't implement
7901 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7902 using the dirstring_type option. See the new comment in the default
7903 openssl.cnf for more info.
7906 *) Make crypto/rand/md_rand.c more robust:
7907 - Assure unique random numbers after fork().
7908 - Make sure that concurrent threads access the global counter and
7909 md serializably so that we never lose entropy in them
7910 or use exactly the same state in multiple threads.
7911 Access to the large state is not always serializable because
7912 the additional locking could be a performance killer, and
7913 md should be large enough anyway.
7916 *) New file apps/app_rand.c with commonly needed functionality
7917 for handling the random seed file.
7919 Use the random seed file in some applications that previously did not:
7921 dsaparam -genkey (which also ignored its '-rand' option),
7924 x509 (when signing).
7925 Except on systems with /dev/urandom, it is crucial to have a random
7926 seed file at least for key creation, DSA signing, and for DH exchanges;
7927 for RSA signatures we could do without one.
7929 gendh and gendsa (unlike genrsa) used to read only the first byte
7930 of each file listed in the '-rand' option. The function as previously
7931 found in genrsa is now in app_rand.c and is used by all programs
7932 that support '-rand'.
7935 *) In RAND_write_file, use mode 0600 for creating files;
7936 don't just chmod when it may be too late.
7939 *) Report an error from X509_STORE_load_locations
7940 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7943 *) New function ASN1_mbstring_copy() this copies a string in either
7944 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7945 into an ASN1_STRING type. A mask of permissible types is passed
7946 and it chooses the "minimal" type to use or an error if not type
7950 *) Add function equivalents to the various macros in asn1.h. The old
7951 macros are retained with an M_ prefix. Code inside the library can
7952 use the M_ macros. External code (including the openssl utility)
7953 should *NOT* in order to be "shared library friendly".
7956 *) Add various functions that can check a certificate's extensions
7957 to see if it usable for various purposes such as SSL client,
7958 server or S/MIME and CAs of these types. This is currently
7959 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7960 verification. Also added a -purpose flag to x509 utility to
7961 print out all the purposes.
7964 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7968 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7969 for, obtain and decode and extension and obtain its critical flag.
7970 This allows all the necessary extension code to be handled in a
7971 single function call.
7974 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7975 platforms. See crypto/rc4/rc4_enc.c for further details.
7978 *) New -noout option to asn1parse. This causes no output to be produced
7979 its main use is when combined with -strparse and -out to extract data
7980 from a file (which may not be in ASN.1 format).
7983 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7984 when producing the local key id.
7985 [Richard Levitte <levitte@stacken.kth.se>]
7987 *) New option -dhparam in s_server. This allows a DH parameter file to be
7988 stated explicitly. If it is not stated then it tries the first server
7989 certificate file. The previous behaviour hard coded the filename
7993 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7994 a public key to be input or output. For example:
7995 openssl rsa -in key.pem -pubout -out pubkey.pem
7996 Also added necessary DSA public key functions to handle this.
7999 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8000 in the message. This was handled by allowing
8001 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8002 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8004 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8005 to the end of the strings whereas this didn't. This would cause problems
8006 if strings read with d2i_ASN1_bytes() were later modified.
8007 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8009 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8010 data and it contains EOF it will end up returning an error. This is
8011 caused by input 46 bytes long. The cause is due to the way base64
8012 BIOs find the start of base64 encoded data. They do this by trying a
8013 trial decode on each line until they find one that works. When they
8014 do a flag is set and it starts again knowing it can pass all the
8015 data directly through the decoder. Unfortunately it doesn't reset
8016 the context it uses. This means that if EOF is reached an attempt
8017 is made to pass two EOFs through the context and this causes the
8018 resulting error. This can also cause other problems as well. As is
8019 usual with these problems it takes *ages* to find and the fix is
8020 trivial: move one line.
8021 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8023 *) Ugly workaround to get s_client and s_server working under Windows. The
8024 old code wouldn't work because it needed to select() on sockets and the
8025 tty (for keypresses and to see if data could be written). Win32 only
8026 supports select() on sockets so we select() with a 1s timeout on the
8027 sockets and then see if any characters are waiting to be read, if none
8028 are present then we retry, we also assume we can always write data to
8029 the tty. This isn't nice because the code then blocks until we've
8030 received a complete line of data and it is effectively polling the
8031 keyboard at 1s intervals: however it's quite a bit better than not
8032 working at all :-) A dedicated Windows application might handle this
8033 with an event loop for example.
8036 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8037 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8038 will be called when RSA_sign() and RSA_verify() are used. This is useful
8039 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8040 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8041 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8042 This necessitated the support of an extra signature type NID_md5_sha1
8043 for SSL signatures and modifications to the SSL library to use it instead
8044 of calling RSA_public_decrypt() and RSA_private_encrypt().
8047 *) Add new -verify -CAfile and -CApath options to the crl program, these
8048 will lookup a CRL issuers certificate and verify the signature in a
8049 similar way to the verify program. Tidy up the crl program so it
8050 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8051 less strict. It will now permit CRL extensions even if it is not
8052 a V2 CRL: this will allow it to tolerate some broken CRLs.
8055 *) Initialize all non-automatic variables each time one of the openssl
8056 sub-programs is started (this is necessary as they may be started
8057 multiple times from the "OpenSSL>" prompt).
8058 [Lennart Bang, Bodo Moeller]
8060 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8061 removing all other RSA functionality (this is what NO_RSA does). This
8062 is so (for example) those in the US can disable those operations covered
8063 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8067 *) Non-copying interface to BIO pairs.
8068 (still largely untested)
8071 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8072 ASCII string. This was handled independently in various places before.
8075 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8076 UTF8 strings a character at a time.
8079 *) Use client_version from client hello to select the protocol
8080 (s23_srvr.c) and for RSA client key exchange verification
8081 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8084 *) Add various utility functions to handle SPKACs, these were previously
8085 handled by poking round in the structure internals. Added new function
8086 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8087 print, verify and generate SPKACs. Based on an original idea from
8088 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8091 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8094 *) Allow the config file extension section to be overwritten on the
8095 command line. Based on an original idea from Massimiliano Pala
8096 <madwolf@comune.modena.it>. The new option is called -extensions
8097 and can be applied to ca, req and x509. Also -reqexts to override
8098 the request extensions in req and -crlexts to override the crl extensions
8102 *) Add new feature to the SPKAC handling in ca. Now you can include
8103 the same field multiple times by preceding it by "XXXX." for example:
8106 this is the same syntax as used in the req config file.
8109 *) Allow certificate extensions to be added to certificate requests. These
8110 are specified in a 'req_extensions' option of the req section of the
8111 config file. They can be printed out with the -text option to req but
8112 are otherwise ignored at present.
8115 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8116 data read consists of only the final block it would not decrypted because
8117 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8118 A misplaced 'break' also meant the decrypted final block might not be
8119 copied until the next read.
8122 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8123 a few extra parameters to the DH structure: these will be useful if
8124 for example we want the value of 'q' or implement X9.42 DH.
8127 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8128 provides hooks that allow the default DSA functions or functions on a
8129 "per key" basis to be replaced. This allows hardware acceleration and
8130 hardware key storage to be handled without major modification to the
8131 library. Also added low level modexp hooks and CRYPTO_EX structure and
8132 associated functions.
8135 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8136 as "read only": it can't be written to and the buffer it points to will
8137 not be freed. Reading from a read only BIO is much more efficient than
8138 a normal memory BIO. This was added because there are several times when
8139 an area of memory needs to be read from a BIO. The previous method was
8140 to create a memory BIO and write the data to it, this results in two
8141 copies of the data and an O(n^2) reading algorithm. There is a new
8142 function BIO_new_mem_buf() which creates a read only memory BIO from
8143 an area of memory. Also modified the PKCS#7 routines to use read only
8147 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8148 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8149 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8150 but a retry condition occured while trying to read the rest.
8153 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8154 NID_pkcs7_encrypted by default: this was wrong since this should almost
8155 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8156 the encrypted data type: this is a more sensible place to put it and it
8157 allows the PKCS#12 code to be tidied up that duplicated this
8161 *) Changed obj_dat.pl script so it takes its input and output files on
8162 the command line. This should avoid shell escape redirection problems
8166 *) Initial support for certificate extension requests, these are included
8167 in things like Xenroll certificate requests. Included functions to allow
8168 extensions to be obtained and added.
8171 *) -crlf option to s_client and s_server for sending newlines as
8172 CRLF (as required by many protocols).
8175 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8177 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8178 [Ralf S. Engelschall]
8180 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8181 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8183 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8187 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8188 DH parameters/keys (q is lost during that conversion, but the resulting
8189 DH parameters contain its length).
8191 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8192 much faster than DH_generate_parameters (which creates parameters
8193 where p = 2*q + 1), and also the smaller q makes DH computations
8194 much more efficient (160-bit exponentiation instead of 1024-bit
8195 exponentiation); so this provides a convenient way to support DHE
8196 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8197 utter importance to use
8198 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8200 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8201 when such DH parameters are used, because otherwise small subgroup
8202 attacks may become possible!
8205 *) Avoid memory leak in i2d_DHparams.
8208 *) Allow the -k option to be used more than once in the enc program:
8209 this allows the same encrypted message to be read by multiple recipients.
8212 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8213 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8214 it will always use the numerical form of the OID, even if it has a short
8218 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8219 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8220 otherwise bn_mod_exp was called. In the case of hardware keys for example
8221 no private key components need be present and it might store extra data
8222 in the RSA structure, which cannot be accessed from bn_mod_exp.
8223 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8224 private key operations.
8227 *) Added support for SPARC Linux.
8230 *) pem_password_cb function type incompatibly changed from
8231 typedef int pem_password_cb(char *buf, int size, int rwflag);
8233 ....(char *buf, int size, int rwflag, void *userdata);
8234 so that applications can pass data to their callbacks:
8235 The PEM[_ASN1]_{read,write}... functions and macros now take an
8236 additional void * argument, which is just handed through whenever
8237 the password callback is called.
8238 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8240 New function SSL_CTX_set_default_passwd_cb_userdata.
8242 Compatibility note: As many C implementations push function arguments
8243 onto the stack in reverse order, the new library version is likely to
8244 interoperate with programs that have been compiled with the old
8245 pem_password_cb definition (PEM_whatever takes some data that
8246 happens to be on the stack as its last argument, and the callback
8247 just ignores this garbage); but there is no guarantee whatsoever that
8250 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8251 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8252 problems not only on Windows, but also on some Unix platforms.
8253 To avoid problematic command lines, these definitions are now in an
8254 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8255 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8258 *) MIPS III/IV assembler module is reimplemented.
8261 *) More DES library cleanups: remove references to srand/rand and
8262 delete an unused file.
8265 *) Add support for the the free Netwide assembler (NASM) under Win32,
8266 since not many people have MASM (ml) and it can be hard to obtain.
8267 This is currently experimental but it seems to work OK and pass all
8268 the tests. Check out INSTALL.W32 for info.
8271 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8272 without temporary keys kept an extra copy of the server key,
8273 and connections with temporary keys did not free everything in case
8277 *) New function RSA_check_key and new openssl rsa option -check
8278 for verifying the consistency of RSA keys.
8279 [Ulf Moeller, Bodo Moeller]
8281 *) Various changes to make Win32 compile work:
8282 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8283 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8284 comparison" warnings.
8285 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8288 *) Add a debugging option to PKCS#5 v2 key generation function: when
8289 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8290 derived keys are printed to stderr.
8293 *) Copy the flags in ASN1_STRING_dup().
8294 [Roman E. Pavlov <pre@mo.msk.ru>]
8296 *) The x509 application mishandled signing requests containing DSA
8297 keys when the signing key was also DSA and the parameters didn't match.
8299 It was supposed to omit the parameters when they matched the signing key:
8300 the verifying software was then supposed to automatically use the CA's
8301 parameters if they were absent from the end user certificate.
8303 Omitting parameters is no longer recommended. The test was also
8304 the wrong way round! This was probably due to unusual behaviour in
8305 EVP_cmp_parameters() which returns 1 if the parameters match.
8306 This meant that parameters were omitted when they *didn't* match and
8307 the certificate was useless. Certificates signed with 'ca' didn't have
8309 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8311 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8312 The interface is as follows:
8313 Applications can use
8314 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8315 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8316 "off" is now the default.
8317 The library internally uses
8318 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8319 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8320 to disable memory-checking temporarily.
8322 Some inconsistent states that previously were possible (and were
8323 even the default) are now avoided.
8325 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8326 with each memory chunk allocated; this is occasionally more helpful
8327 than just having a counter.
8329 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8331 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8335 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8336 which largely parallels "options", but is for changing API behaviour,
8337 whereas "options" are about protocol behaviour.
8338 Initial "mode" flags are:
8340 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8341 a single record has been written.
8342 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8343 retries use the same buffer location.
8344 (But all of the contents must be
8348 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8351 *) Fix problems with no-hmac etc.
8352 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8354 *) New functions RSA_get_default_method(), RSA_set_method() and
8355 RSA_get_method(). These allows replacement of RSA_METHODs without having
8356 to mess around with the internals of an RSA structure.
8359 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8360 Also really enable memory leak checks in openssl.c and in some
8362 [Chad C. Mulligan, Bodo Moeller]
8364 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8365 up the length of negative integers. This has now been simplified to just
8366 store the length when it is first determined and use it later, rather
8367 than trying to keep track of where data is copied and updating it to
8369 [Steve Henson, reported by Brien Wheeler
8370 <bwheeler@authentica-security.com>]
8372 *) Add a new function PKCS7_signatureVerify. This allows the verification
8373 of a PKCS#7 signature but with the signing certificate passed to the
8374 function itself. This contrasts with PKCS7_dataVerify which assumes the
8375 certificate is present in the PKCS#7 structure. This isn't always the
8376 case: certificates can be omitted from a PKCS#7 structure and be
8377 distributed by "out of band" means (such as a certificate database).
8380 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8381 function prototypes in pem.h, also change util/mkdef.pl to add the
8382 necessary function names.
8385 *) mk1mf.pl (used by Windows builds) did not properly read the
8386 options set by Configure in the top level Makefile, and Configure
8387 was not even able to write more than one option correctly.
8388 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8391 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8392 file to be loaded from a BIO or FILE pointer. The BIO version will
8393 for example allow memory BIOs to contain config info.
8396 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8397 Whoever hopes to achieve shared-library compatibility across versions
8398 must use this, not the compile-time macro.
8399 (Exercise 0.9.4: Which is the minimum library version required by
8401 Note: All this applies only to multi-threaded programs, others don't
8405 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8406 through a BIO pair triggered the default case, i.e.
8407 SSLerr(...,SSL_R_UNKNOWN_STATE).
8410 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8411 can use the SSL library even if none of the specific BIOs is
8415 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8416 for the encoded length.
8417 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8419 *) Add initial documentation of the X509V3 functions.
8422 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8423 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8424 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8425 secure PKCS#8 private key format with a high iteration count.
8428 *) Fix determination of Perl interpreter: A perl or perl5
8429 _directory_ in $PATH was also accepted as the interpreter.
8430 [Ralf S. Engelschall]
8432 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8433 wrong with it but it was very old and did things like calling
8434 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8438 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8439 to use the new extension code.
8442 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8443 with macros. This should make it easier to change their form, add extra
8444 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8448 *) Add to configuration table a new entry that can specify an alternative
8449 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8450 according to Mark Crispin <MRC@Panda.COM>.
8454 *) DES CBC did not update the IV. Weird.
8457 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8458 Changing the behaviour of the former might break existing programs --
8459 where IV updating is needed, des_ncbc_encrypt can be used.
8462 *) When bntest is run from "make test" it drives bc to check its
8463 calculations, as well as internally checking them. If an internal check
8464 fails, it needs to cause bc to give a non-zero result or make test carries
8465 on without noticing the failure. Fixed.
8468 *) DES library cleanups.
8471 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8472 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8473 ciphers. NOTE: although the key derivation function has been verified
8474 against some published test vectors it has not been extensively tested
8475 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8479 *) Instead of "mkdir -p", which is not fully portable, use new
8480 Perl script "util/mkdir-p.pl".
8483 *) Rewrite the way password based encryption (PBE) is handled. It used to
8484 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8485 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8486 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8487 the 'parameter' field of the AlgorithmIdentifier is passed to the
8488 underlying key generation function so it must do its own ASN1 parsing.
8489 This has also changed the EVP_PBE_CipherInit() function which now has a
8490 'parameter' argument instead of literal salt and iteration count values
8491 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8494 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8495 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8496 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8497 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8498 value was just used as a "magic string" and not used directly its
8499 value doesn't matter.
8502 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8506 *) "linux-sparc64" configuration (ultrapenguin).
8507 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8508 "linux-sparc" configuration.
8509 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8511 *) config now generates no-xxx options for missing ciphers.
8514 *) Support the EBCDIC character set (work in progress).
8515 File ebcdic.c not yet included because it has a different license.
8516 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8518 *) Support BS2000/OSD-POSIX.
8519 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8521 *) Make callbacks for key generation use void * instead of char *.
8524 *) Make S/MIME samples compile (not yet tested).
8527 *) Additional typesafe stacks.
8530 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8534 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8536 *) New configuration variant "sco5-gcc".
8538 *) Updated some demos.
8539 [Sean O Riordain, Wade Scholine]
8541 *) Add missing BIO_free at exit of pkcs12 application.
8544 *) Fix memory leak in conf.c.
8547 *) Updates for Win32 to assembler version of MD5.
8550 *) Set #! path to perl in apps/der_chop to where we found it
8551 instead of using a fixed path.
8554 *) SHA library changes for irix64-mips4-cc.
8557 *) Improvements for VMS support.
8561 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8563 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8564 This also avoids the problems with SC4.2 and unpatched SC5.
8565 [Andy Polyakov <appro@fy.chalmers.se>]
8567 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8568 These are required because of the typesafe stack would otherwise break
8569 existing code. If old code used a structure member which used to be STACK
8570 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8571 sk_num or sk_value it would produce an error because the num, data members
8572 are not present in STACK_OF. Now it just produces a warning. sk_set
8573 replaces the old method of assigning a value to sk_value
8574 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8575 that does this will no longer work (and should use sk_set instead) but
8576 this could be regarded as a "questionable" behaviour anyway.
8579 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8580 correctly handle encrypted S/MIME data.
8583 *) Change type of various DES function arguments from des_cblock
8584 (which means, in function argument declarations, pointer to char)
8585 to des_cblock * (meaning pointer to array with 8 char elements),
8586 which allows the compiler to do more typechecking; it was like
8587 that back in SSLeay, but with lots of ugly casts.
8589 Introduce new type const_des_cblock.
8592 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8593 problems: find RecipientInfo structure that matches recipient certificate
8594 and initialise the ASN1 structures properly based on passed cipher.
8597 *) Belatedly make the BN tests actually check the results.
8600 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8601 to and from BNs: it was completely broken. New compilation option
8602 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8603 key elements as negative integers.
8606 *) Reorganize and speed up MD5.
8607 [Andy Polyakov <appro@fy.chalmers.se>]
8610 [Richard Levitte <richard@levitte.org>]
8612 *) New option -out to asn1parse to allow the parsed structure to be
8613 output to a file. This is most useful when combined with the -strparse
8614 option to examine the output of things like OCTET STRINGS.
8617 *) Make SSL library a little more fool-proof by not requiring any longer
8618 that SSL_set_{accept,connect}_state be called before
8619 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8620 in many applications because usually everything *appeared* to work as
8621 intended anyway -- now it really works as intended).
8624 *) Move openssl.cnf out of lib/.
8627 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8628 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8629 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8630 [Ralf S. Engelschall]
8632 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8633 handle PKCS#7 enveloped data properly.
8634 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8636 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8637 copying pointers. The cert_st handling is changed by this in
8638 various ways (and thus what used to be known as ctx->default_cert
8639 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8640 any longer when s->cert does not give us what we need).
8641 ssl_cert_instantiate becomes obsolete by this change.
8642 As soon as we've got the new code right (possibly it already is?),
8643 we have solved a couple of bugs of the earlier code where s->cert
8644 was used as if it could not have been shared with other SSL structures.
8646 Note that using the SSL API in certain dirty ways now will result
8647 in different behaviour than observed with earlier library versions:
8648 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8649 does not influence s as it used to.
8651 In order to clean up things more thoroughly, inside SSL_SESSION
8652 we don't use CERT any longer, but a new structure SESS_CERT
8653 that holds per-session data (if available); currently, this is
8654 the peer's certificate chain and, for clients, the server's certificate
8655 and temporary key. CERT holds only those values that can have
8656 meaningful defaults in an SSL_CTX.
8659 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8660 from the internal representation. Various PKCS#7 fixes: remove some
8661 evil casts and set the enc_dig_alg field properly based on the signing
8665 *) Allow PKCS#12 password to be set from the command line or the
8666 environment. Let 'ca' get its config file name from the environment
8667 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8671 *) Allow certificate policies extension to use an IA5STRING for the
8672 organization field. This is contrary to the PKIX definition but
8673 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8677 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8678 without disallowing inline assembler and the like for non-pedantic builds.
8681 *) Support Borland C++ builder.
8682 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8687 *) SHA-1 cleanups and performance enhancements.
8688 [Andy Polyakov <appro@fy.chalmers.se>]
8690 *) Sparc v8plus assembler for the bignum library.
8691 [Andy Polyakov <appro@fy.chalmers.se>]
8693 *) Accept any -xxx and +xxx compiler options in Configure.
8696 *) Update HPUX configuration.
8699 *) Add missing sk_<type>_unshift() function to safestack.h
8700 [Ralf S. Engelschall]
8702 *) New function SSL_CTX_use_certificate_chain_file that sets the
8703 "extra_cert"s in addition to the certificate. (This makes sense
8704 only for "PEM" format files, as chains as a whole are not
8708 *) Support verify_depth from the SSL API.
8709 x509_vfy.c had what can be considered an off-by-one-error:
8710 Its depth (which was not part of the external interface)
8711 was actually counting the number of certificates in a chain;
8712 now it really counts the depth.
8715 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8716 instead of X509err, which often resulted in confusing error
8717 messages since the error codes are not globally unique
8718 (e.g. an alleged error in ssl3_accept when a certificate
8719 didn't match the private key).
8721 *) New function SSL_CTX_set_session_id_context that allows to set a default
8722 value (so that you don't need SSL_set_session_id_context for each
8723 connection using the SSL_CTX).
8726 *) OAEP decoding bug fix.
8729 *) Support INSTALL_PREFIX for package builders, as proposed by
8733 *) New Configure options "threads" and "no-threads". For systems
8734 where the proper compiler options are known (currently Solaris
8735 and Linux), "threads" is the default.
8738 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8741 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8742 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8743 such as /usr/local/bin.
8746 *) "make linux-shared" to build shared libraries.
8747 [Niels Poppe <niels@netbox.org>]
8749 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8752 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8753 extension adding in x509 utility.
8756 *) Remove NOPROTO sections and error code comments.
8759 *) Partial rewrite of the DEF file generator to now parse the ANSI
8763 *) New Configure options --prefix=DIR and --openssldir=DIR.
8766 *) Complete rewrite of the error code script(s). It is all now handled
8767 by one script at the top level which handles error code gathering,
8768 header rewriting and C source file generation. It should be much better
8769 than the old method: it now uses a modified version of Ulf's parser to
8770 read the ANSI prototypes in all header files (thus the old K&R definitions
8771 aren't needed for error creation any more) and do a better job of
8772 translating function codes into names. The old 'ASN1 error code imbedded
8773 in a comment' is no longer necessary and it doesn't use .err files which
8774 have now been deleted. Also the error code call doesn't have to appear all
8775 on one line (which resulted in some large lines...).
8778 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8781 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8782 0 (which usually indicates a closed connection), but continue reading.
8785 *) Fix some race conditions.
8788 *) Add support for CRL distribution points extension. Add Certificate
8789 Policies and CRL distribution points documentation.
8792 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8795 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8796 8 of keying material. Merlin has also confirmed interop with this fix
8797 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8798 [Merlin Hughes <merlin@baltimore.ie>]
8800 *) Fix lots of warnings.
8801 [Richard Levitte <levitte@stacken.kth.se>]
8803 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8804 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8805 [Richard Levitte <levitte@stacken.kth.se>]
8807 *) Fix problems with sizeof(long) == 8.
8808 [Andy Polyakov <appro@fy.chalmers.se>]
8810 *) Change functions to ANSI C.
8813 *) Fix typos in error codes.
8814 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8816 *) Remove defunct assembler files from Configure.
8819 *) SPARC v8 assembler BIGNUM implementation.
8820 [Andy Polyakov <appro@fy.chalmers.se>]
8822 *) Support for Certificate Policies extension: both print and set.
8823 Various additions to support the r2i method this uses.
8826 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8827 return a const string when you are expecting an allocated buffer.
8830 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8831 types DirectoryString and DisplayText.
8834 *) Add code to allow r2i extensions to access the configuration database,
8835 add an LHASH database driver and add several ctx helper functions.
8838 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8839 fail when they extended the size of a BIGNUM.
8842 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8843 support typesafe stack.
8846 *) Fix typo in SSL_[gs]et_options().
8847 [Nils Frostberg <nils@medcom.se>]
8849 *) Delete various functions and files that belonged to the (now obsolete)
8850 old X509V3 handling code.
8853 *) New Configure option "rsaref".
8856 *) Don't auto-generate pem.h.
8859 *) Introduce type-safe ASN.1 SETs.
8862 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8863 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8865 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8866 that links with OpenSSL (well at least cause lots of warnings), but fear
8867 not: the conversion is trivial, and it eliminates loads of evil casts. A
8868 few STACKed things have been converted already. Feel free to convert more.
8869 In the fullness of time, I'll do away with the STACK type altogether.
8872 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8873 specified in <certfile> by updating the entry in the index.txt file.
8874 This way one no longer has to edit the index.txt file manually for
8875 revoking a certificate. The -revoke option does the gory details now.
8876 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8878 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8879 `-text' option at all and this way the `-noout -text' combination was
8880 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8881 [Ralf S. Engelschall]
8883 *) Make sure a corresponding plain text error message exists for the
8884 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8885 verify callback function determined that a certificate was revoked.
8886 [Ralf S. Engelschall]
8888 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8889 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8890 all available cipers including rc5, which was forgotten until now.
8891 In order to let the testing shell script know which algorithms
8892 are available, a new (up to now undocumented) command
8893 "openssl list-cipher-commands" is used.
8896 *) Bugfix: s_client occasionally would sleep in select() when
8897 it should have checked SSL_pending() first.
8900 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8901 the raw DSA values prior to ASN.1 encoding.
8904 *) Tweaks to Configure
8905 [Niels Poppe <niels@netbox.org>]
8907 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8911 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8914 *) New config option to avoid instructions that are illegal on the 80386.
8915 The default code is faster, but requires at least a 486.
8918 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8919 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8920 same as SSL2_VERSION anyway.
8923 *) New "-showcerts" option for s_client.
8926 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8927 application. Various cleanups and fixes.
8930 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8931 modify error routines to work internally. Add error codes and PBE init
8932 to library startup routines.
8935 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8936 packing functions to asn1 and evp. Changed function names and error
8937 codes along the way.
8940 *) PKCS12 integration: and so it begins... First of several patches to
8941 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8942 objects to objects.h
8945 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8946 and display support for Thawte strong extranet extension.
8949 *) Add LinuxPPC support.
8950 [Jeff Dubrule <igor@pobox.org>]
8952 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8953 bn_div_words in alpha.s.
8954 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8956 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8957 OAEP isn't supported when OpenSSL is built with RSAref.
8958 [Ulf Moeller <ulf@fitug.de>]
8960 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8961 so they no longer are missing under -DNOPROTO.
8962 [Soren S. Jorvang <soren@t.dk>]
8965 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8967 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8968 doesn't work when the session is reused. Coming soon!
8971 *) Fix a security hole, that allows sessions to be reused in the wrong
8972 context thus bypassing client cert protection! All software that uses
8973 client certs and session caches in multiple contexts NEEDS PATCHING to
8974 allow session reuse! A fuller solution is in the works.
8975 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8977 *) Some more source tree cleanups (removed obsolete files
8978 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8979 permission on "config" script to be executable) and a fix for the INSTALL
8981 [Ulf Moeller <ulf@fitug.de>]
8983 *) Remove some legacy and erroneous uses of malloc, free instead of
8985 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8987 *) Make rsa_oaep_test return non-zero on error.
8988 [Ulf Moeller <ulf@fitug.de>]
8990 *) Add support for native Solaris shared libraries. Configure
8991 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8992 if someone would make that last step automatic.
8993 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8995 *) ctx_size was not built with the right compiler during "make links". Fixed.
8998 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8999 except NULL ciphers". This means the default cipher list will no longer
9000 enable NULL ciphers. They need to be specifically enabled e.g. with
9001 the string "DEFAULT:eNULL".
9004 *) Fix to RSA private encryption routines: if p < q then it would
9005 occasionally produce an invalid result. This will only happen with
9006 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9009 *) Be less restrictive and allow also `perl util/perlpath.pl
9010 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9011 because this way one can also use an interpreter named `perl5' (which is
9012 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9013 installed as `perl').
9014 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9016 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9017 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9019 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9020 advapi32.lib to Win32 build and change the pem test comparision
9021 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9022 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9023 and crypto/des/ede_cbcm_enc.c.
9026 *) DES quad checksum was broken on big-endian architectures. Fixed.
9029 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9030 Win32 test batch file so it (might) work again. The Win32 test batch file
9031 is horrible: I feel ill....
9034 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9035 in e_os.h. Audit of header files to check ANSI and non ANSI
9036 sections: 10 functions were absent from non ANSI section and not exported
9037 from Windows DLLs. Fixed up libeay.num for new functions.
9040 *) Make `openssl version' output lines consistent.
9041 [Ralf S. Engelschall]
9043 *) Fix Win32 symbol export lists for BIO functions: Added
9044 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9045 to ms/libeay{16,32}.def.
9046 [Ralf S. Engelschall]
9048 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9049 fine under Unix and passes some trivial tests I've now added. But the
9050 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9051 added to make sure no one expects that this stuff really works in the
9052 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9053 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9055 [Ralf S. Engelschall]
9057 *) Fix the generation of two part addresses in perl.
9058 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9060 *) Add config entry for Linux on MIPS.
9061 [John Tobey <jtobey@channel1.com>]
9063 *) Make links whenever Configure is run, unless we are on Windoze.
9066 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9067 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9071 *) Add a useful kludge to allow package maintainers to specify compiler and
9072 other platforms details on the command line without having to patch the
9073 Configure script everytime: One now can use ``perl Configure
9074 <id>:<details>'', i.e. platform ids are allowed to have details appended
9075 to them (seperated by colons). This is treated as there would be a static
9076 pre-configured entry in Configure's %table under key <id> with value
9077 <details> and ``perl Configure <id>'' is called. So, when you want to
9078 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9079 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9080 now, which overrides the FreeBSD-elf entry on-the-fly.
9081 [Ralf S. Engelschall]
9083 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9086 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9087 on the `perl Configure ...' command line. This way one can compile
9088 OpenSSL libraries with Position Independent Code (PIC) which is needed
9089 for linking it into DSOs.
9090 [Ralf S. Engelschall]
9092 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9096 *) Cleaned up the LICENSE document: The official contact for any license
9097 questions now is the OpenSSL core team under openssl-core@openssl.org.
9098 And add a paragraph about the dual-license situation to make sure people
9099 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9100 to the OpenSSL toolkit.
9101 [Ralf S. Engelschall]
9103 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9104 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9105 Additonally cleaned up the `make links' target: Remove unnecessary
9106 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9107 to speed processing and no longer clutter the display with confusing
9108 stuff. Instead only the actually done links are displayed.
9109 [Ralf S. Engelschall]
9111 *) Permit null encryption ciphersuites, used for authentication only. It used
9112 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9113 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9117 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9118 signed attributes when verifying signatures (this would break them),
9119 the detached data encoding was wrong and public keys obtained using
9120 X509_get_pubkey() weren't freed.
9123 *) Add text documentation for the BUFFER functions. Also added a work around
9124 to a Win95 console bug. This was triggered by the password read stuff: the
9125 last character typed gets carried over to the next fread(). If you were
9126 generating a new cert request using 'req' for example then the last
9127 character of the passphrase would be CR which would then enter the first
9131 *) Added the new `Includes OpenSSL Cryptography Software' button as
9132 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9133 button and can be used by applications based on OpenSSL to show the
9134 relationship to the OpenSSL project.
9135 [Ralf S. Engelschall]
9137 *) Remove confusing variables in function signatures in files
9138 ssl/ssl_lib.c and ssl/ssl.h.
9139 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9141 *) Don't install bss_file.c under PREFIX/include/
9142 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9144 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9145 functions that return function pointers and has support for NT specific
9146 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9147 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9148 unsigned to signed types: this was killing the Win32 compile.
9151 *) Add new certificate file to stack functions,
9152 SSL_add_dir_cert_subjects_to_stack() and
9153 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9154 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9155 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9156 This means that Apache-SSL and similar packages don't have to mess around
9157 to add as many CAs as they want to the preferred list.
9160 *) Experiment with doxygen documentation. Currently only partially applied to
9162 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9163 openssl.doxy as the configuration file.
9166 *) Get rid of remaining C++-style comments which strict C compilers hate.
9167 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9169 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9170 compiled in by default: it has problems with large keys.
9173 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9174 DH private keys and/or callback functions which directly correspond to
9175 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9176 is needed for applications which have to configure certificates on a
9177 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9179 For the RSA certificate situation is makes no difference, but
9180 for the DSA certificate situation this fixes the "no shared cipher"
9181 problem where the OpenSSL cipher selection procedure failed because the
9182 temporary keys were not overtaken from the context and the API provided
9183 no way to reconfigure them.
9184 The new functions now let applications reconfigure the stuff and they
9185 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9186 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9187 non-public-API function ssl_cert_instantiate() is used as a helper
9188 function and also to reduce code redundancy inside ssl_rsa.c.
9189 [Ralf S. Engelschall]
9191 *) Move s_server -dcert and -dkey options out of the undocumented feature
9192 area because they are useful for the DSA situation and should be
9193 recognized by the users.
9194 [Ralf S. Engelschall]
9196 *) Fix the cipher decision scheme for export ciphers: the export bits are
9197 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9198 SSL_EXP_MASK. So, the original variable has to be used instead of the
9199 already masked variable.
9200 [Richard Levitte <levitte@stacken.kth.se>]
9202 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9203 [Richard Levitte <levitte@stacken.kth.se>]
9205 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9206 from `int' to `unsigned int' because it's a length and initialized by
9207 EVP_DigestFinal() which expects an `unsigned int *'.
9208 [Richard Levitte <levitte@stacken.kth.se>]
9210 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9211 script. Instead use the usual Shell->Perl transition trick.
9212 [Ralf S. Engelschall]
9214 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9215 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9216 -noout -modulus' as it's already the case for `openssl rsa -noout
9217 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9218 currently the public key is printed (a decision which was already done by
9219 `openssl dsa -modulus' in the past) which serves a similar purpose.
9220 Additionally the NO_RSA no longer completely removes the whole -modulus
9221 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9223 [Ralf S. Engelschall]
9225 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9226 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9227 [Arne Ansper <arne@ats.cyber.ee>]
9229 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9230 to be added. Now both 'req' and 'ca' can use new objects defined in the
9234 *) Add cool BIO that does syslog (or event log on NT).
9235 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9237 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9238 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9239 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9240 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9243 *) Add preliminary config info for new extension code.
9246 *) Make RSA_NO_PADDING really use no padding.
9247 [Ulf Moeller <ulf@fitug.de>]
9249 *) Generate errors when private/public key check is done.
9252 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9253 for some CRL extensions and new objects added.
9256 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9257 key usage extension and fuller support for authority key id.
9260 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9261 padding method for RSA, which is recommended for new applications in PKCS
9262 #1 v2.0 (RFC 2437, October 1998).
9263 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9264 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9265 against Bleichbacher's attack on RSA.
9266 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9269 *) Updates to the new SSL compression code
9270 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9272 *) Fix so that the version number in the master secret, when passed
9273 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9274 (because the server will not accept higher), that the version number
9275 is 0x03,0x01, not 0x03,0x00
9276 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9278 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9279 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9280 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9283 *) Support for RAW extensions where an arbitrary extension can be
9284 created by including its DER encoding. See apps/openssl.cnf for
9288 *) Make sure latest Perl versions don't interpret some generated C array
9289 code as Perl array code in the crypto/err/err_genc.pl script.
9290 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9292 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9293 not many people have the assembler. Various Win32 compilation fixes and
9294 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9298 *) Modify configure script 'Configure' to automatically create crypto/date.h
9299 file under Win32 and also build pem.h from pem.org. New script
9300 util/mkfiles.pl to create the MINFO file on environments that can't do a
9301 'make files': perl util/mkfiles.pl >MINFO should work.
9304 *) Major rework of DES function declarations, in the pursuit of correctness
9305 and purity. As a result, many evil casts evaporated, and some weirdness,
9306 too. You may find this causes warnings in your code. Zapping your evil
9307 casts will probably fix them. Mostly.
9310 *) Fix for a typo in asn1.h. Bug fix to object creation script
9311 obj_dat.pl. It considered a zero in an object definition to mean
9312 "end of object": none of the objects in objects.h have any zeros
9313 so it wasn't spotted.
9314 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9316 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9317 Masking (CBCM). In the absence of test vectors, the best I have been able
9318 to do is check that the decrypt undoes the encrypt, so far. Send me test
9319 vectors if you have them.
9322 *) Correct calculation of key length for export ciphers (too much space was
9323 allocated for null ciphers). This has not been tested!
9326 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9327 message is now correct (it understands "crypto" and "ssl" on its
9328 command line). There is also now an "update" option. This will update
9329 the util/ssleay.num and util/libeay.num files with any new functions.
9331 perl util/mkdef.pl crypto ssl update
9332 it will update them.
9335 *) Overhauled the Perl interface (perl/*):
9336 - ported BN stuff to OpenSSL's different BN library
9337 - made the perl/ source tree CVS-aware
9338 - renamed the package from SSLeay to OpenSSL (the files still contain
9339 their history because I've copied them in the repository)
9340 - removed obsolete files (the test scripts will be replaced
9341 by better Test::Harness variants in the future)
9342 [Ralf S. Engelschall]
9344 *) First cut for a very conservative source tree cleanup:
9345 1. merge various obsolete readme texts into doc/ssleay.txt
9346 where we collect the old documents and readme texts.
9347 2. remove the first part of files where I'm already sure that we no
9348 longer need them because of three reasons: either they are just temporary
9349 files which were left by Eric or they are preserved original files where
9350 I've verified that the diff is also available in the CVS via "cvs diff
9351 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9352 the crypto/md/ stuff).
9353 [Ralf S. Engelschall]
9355 *) More extension code. Incomplete support for subject and issuer alt
9356 name, issuer and authority key id. Change the i2v function parameters
9357 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9358 what that's for :-) Fix to ASN1 macro which messed up
9359 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9362 *) Preliminary support for ENUMERATED type. This is largely copied from the
9366 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9367 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9369 *) Make sure `make rehash' target really finds the `openssl' program.
9370 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9372 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9373 like to hear about it if this slows down other processors.
9376 *) Add CygWin32 platform information to Configure script.
9377 [Alan Batie <batie@aahz.jf.intel.com>]
9379 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9380 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9382 *) New program nseq to manipulate netscape certificate sequences
9385 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9389 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9390 but the BN code had some problems that would cause failures when
9391 doing certificate verification and some other functions.
9392 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9394 *) Add ASN1 and PEM code to support netscape certificate sequences.
9397 *) Add ASN1 and PEM code to support netscape certificate sequences.
9400 *) Add several PKIX and private extended key usage OIDs.
9403 *) Modify the 'ca' program to handle the new extension code. Modify
9404 openssl.cnf for new extension format, add comments.
9407 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9408 and add a sample to openssl.cnf so req -x509 now adds appropriate
9412 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9413 error code, add initial support to X509_print() and x509 application.
9416 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9417 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9418 stuff is currently isolated and isn't even compiled yet.
9421 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9422 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9423 Removed the versions check from X509 routines when loading extensions:
9424 this allows certain broken certificates that don't set the version
9425 properly to be processed.
9428 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9429 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9430 can still be regenerated with "make depend".
9433 *) Spelling mistake in C version of CAST-128.
9434 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9436 *) Changes to the error generation code. The perl script err-code.pl
9437 now reads in the old error codes and retains the old numbers, only
9438 adding new ones if necessary. It also only changes the .err files if new
9439 codes are added. The makefiles have been modified to only insert errors
9440 when needed (to avoid needlessly modifying header files). This is done
9441 by only inserting errors if the .err file is newer than the auto generated
9442 C file. To rebuild all the error codes from scratch (the old behaviour)
9443 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9444 or delete all the .err files.
9447 *) CAST-128 was incorrectly implemented for short keys. The C version has
9448 been fixed, but is untested. The assembler versions are also fixed, but
9449 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9450 to regenerate it if needed.
9451 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9452 Hagino <itojun@kame.net>]
9454 *) File was opened incorrectly in randfile.c.
9455 [Ulf Möller <ulf@fitug.de>]
9457 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9458 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9459 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9460 al: it's just almost always a UTCTime. Note this patch adds new error
9461 codes so do a "make errors" if there are problems.
9464 *) Correct Linux 1 recognition in config.
9465 [Ulf Möller <ulf@fitug.de>]
9467 *) Remove pointless MD5 hash when using DSA keys in ca.
9468 [Anonymous <nobody@replay.com>]
9470 *) Generate an error if given an empty string as a cert directory. Also
9471 generate an error if handed NULL (previously returned 0 to indicate an
9472 error, but didn't set one).
9473 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9475 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9478 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9479 parameters. This was causing a warning which killed off the Win32 compile.
9482 *) Remove C++ style comments from crypto/bn/bn_local.h.
9483 [Neil Costigan <neil.costigan@celocom.com>]
9485 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9486 based on a text string, looking up short and long names and finally
9487 "dot" format. The "dot" format stuff didn't work. Added new function
9488 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9489 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9490 OID is not part of the table.
9493 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9494 X509_LOOKUP_by_alias().
9497 *) Sort openssl functions by name.
9500 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9501 encryption from sample DSA keys (in case anyone is interested the password
9505 *) Make _all_ *_free functions accept a NULL pointer.
9506 [Frans Heymans <fheymans@isaserver.be>]
9508 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9510 [Anonymous <nobody@replay.com>]
9512 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9513 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9515 *) Don't blow it for numeric -newkey arguments to apps/req.
9516 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9518 *) Temp key "for export" tests were wrong in s3_srvr.c.
9519 [Anonymous <nobody@replay.com>]
9521 *) Add prototype for temp key callback functions
9522 SSL_CTX_set_tmp_{rsa,dh}_callback().
9525 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9526 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9529 *) X509_name_add_entry() freed the wrong thing after an error.
9530 [Arne Ansper <arne@ats.cyber.ee>]
9532 *) rsa_eay.c would attempt to free a NULL context.
9533 [Arne Ansper <arne@ats.cyber.ee>]
9535 *) BIO_s_socket() had a broken should_retry() on Windoze.
9536 [Arne Ansper <arne@ats.cyber.ee>]
9538 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9539 [Arne Ansper <arne@ats.cyber.ee>]
9541 *) Make sure the already existing X509_STORE->depth variable is initialized
9542 in X509_STORE_new(), but document the fact that this variable is still
9543 unused in the certificate verification process.
9544 [Ralf S. Engelschall]
9546 *) Fix the various library and apps files to free up pkeys obtained from
9547 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9550 *) Fix reference counting in X509_PUBKEY_get(). This makes
9551 demos/maurice/example2.c work, amongst others, probably.
9552 [Steve Henson and Ben Laurie]
9554 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9555 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9556 are no longer created. This way we have a single and consistent command
9557 line interface `openssl <command>', similar to `cvs <command>'.
9558 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9560 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9561 BIT STRING wrapper always have zero unused bits.
9564 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9567 *) Make the top-level INSTALL documentation easier to understand.
9570 *) Makefiles updated to exit if an error occurs in a sub-directory
9571 make (including if user presses ^C) [Paul Sutton]
9573 *) Make Montgomery context stuff explicit in RSA data structure.
9576 *) Fix build order of pem and err to allow for generated pem.h.
9579 *) Fix renumbering bug in X509_NAME_delete_entry().
9582 *) Enhanced the err-ins.pl script so it makes the error library number
9583 global and can add a library name. This is needed for external ASN1 and
9584 other error libraries.
9587 *) Fixed sk_insert which never worked properly.
9590 *) Fix ASN1 macros so they can handle indefinite length construted
9591 EXPLICIT tags. Some non standard certificates use these: they can now
9595 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9596 into a single doc/ssleay.txt bundle. This way the information is still
9597 preserved but no longer messes up this directory. Now it's new room for
9598 the new set of documenation files.
9599 [Ralf S. Engelschall]
9601 *) SETs were incorrectly DER encoded. This was a major pain, because they
9602 shared code with SEQUENCEs, which aren't coded the same. This means that
9603 almost everything to do with SETs or SEQUENCEs has either changed name or
9604 number of arguments.
9605 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9607 *) Fix test data to work with the above.
9610 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9611 was already fixed by Eric for 0.9.1 it seems.
9612 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9614 *) Autodetect FreeBSD3.
9617 *) Fix various bugs in Configure. This affects the following platforms:
9621 unixware-2.0-pentium
9625 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9626 before they are needed.
9629 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9633 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9635 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9636 changed SSLeay to OpenSSL in version strings.
9637 [Ralf S. Engelschall]
9639 *) Some fixups to the top-level documents.
9642 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9643 because the symlink to include/ was missing.
9644 [Ralf S. Engelschall]
9646 *) Incorporated the popular no-RSA/DSA-only patches
9647 which allow to compile a RSA-free SSLeay.
9648 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9650 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9651 when "ssleay" is still not found.
9652 [Ralf S. Engelschall]
9654 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9655 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9657 *) Updated the README file.
9658 [Ralf S. Engelschall]
9660 *) Added various .cvsignore files in the CVS repository subdirs
9661 to make a "cvs update" really silent.
9662 [Ralf S. Engelschall]
9664 *) Recompiled the error-definition header files and added
9665 missing symbols to the Win32 linker tables.
9666 [Ralf S. Engelschall]
9668 *) Cleaned up the top-level documents;
9669 o new files: CHANGES and LICENSE
9670 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9671 o merged COPYRIGHT into LICENSE
9672 o removed obsolete TODO file
9673 o renamed MICROSOFT to INSTALL.W32
9674 [Ralf S. Engelschall]
9676 *) Removed dummy files from the 0.9.1b source tree:
9677 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9678 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9679 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9680 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9681 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9682 [Ralf S. Engelschall]
9684 *) Added various platform portability fixes.
9687 *) The Genesis of the OpenSSL rpject:
9688 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9689 Young and Tim J. Hudson created while they were working for C2Net until
9691 [The OpenSSL Project]
9694 Changes between 0.9.0b and 0.9.1b [not released]
9696 *) Updated a few CA certificates under certs/
9699 *) Changed some BIGNUM api stuff.
9702 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9703 DGUX x86, Linux Alpha, etc.
9706 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9707 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9711 *) Add -strparse option to asn1pars program which parses nested
9713 [Dr Stephen Henson <shenson@bigfoot.com>]
9715 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9718 *) DSA fix for "ca" program.
9721 *) Added "-genkey" option to "dsaparam" program.
9724 *) Added RIPE MD160 (rmd160) message digest.
9727 *) Added -a (all) option to "ssleay version" command.
9730 *) Added PLATFORM define which is the id given to Configure.
9733 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9736 *) Extended the ASN.1 parser routines.
9739 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9742 *) Added a BN_CTX to the BN library.
9745 *) Fixed the weak key values in DES library
9748 *) Changed API in EVP library for cipher aliases.
9751 *) Added support for RC2/64bit cipher.
9754 *) Converted the lhash library to the crypto/mem.c functions.
9757 *) Added more recognized ASN.1 object ids.
9760 *) Added more RSA padding checks for SSL/TLS.
9763 *) Added BIO proxy/filter functionality.
9766 *) Added extra_certs to SSL_CTX which can be used
9767 send extra CA certificates to the client in the CA cert chain sending
9768 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9771 *) Now Fortezza is denied in the authentication phase because
9772 this is key exchange mechanism is not supported by SSLeay at all.
9775 *) Additional PKCS1 checks.
9778 *) Support the string "TLSv1" for all TLS v1 ciphers.
9781 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9782 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9785 *) Fixed a few memory leaks.
9788 *) Fixed various code and comment typos.
9791 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9792 bytes sent in the client random.
9793 [Edward Bishop <ebishop@spyglass.com>]