5 Changes between 1.0.1 and 1.1.0 [xx XXX xxxx]
7 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
8 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
9 performs algorithm blocking for unapproved PRNG types. Also do not
10 set PRNG type in FIPS_mode_set(): leave this to the application.
11 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
12 the standard OpenSSL PRNG: set additional data to a date time vector.
15 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
16 This shouldn't present any incompatibility problems because applications
17 shouldn't be using these directly and any that are will need to rethink
18 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
21 *) Extensive self tests and health checking required by SP800-90 DRBG.
22 Remove strength parameter from FIPS_drbg_instantiate and always
23 instantiate at maximum supported strength.
27 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
29 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
32 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
35 *) New function DH_compute_key_padded() to compute a DH key and pad with
36 leading zeroes if needed: this complies with SP800-56A et al.
39 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
40 anything, incomplete, subject to change and largely untested at present.
43 *) Modify fipscanisteronly build option to only build the necessary object
44 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
47 *) Add experimental option FIPSSYMS to give all symbols in
48 fipscanister.o and FIPS or fips prefix. This will avoid
49 conflicts with future versions of OpenSSL. Add perl script
50 util/fipsas.pl to preprocess assembly language source files
51 and rename any affected symbols.
54 *) Add selftest checks and algorithm block of non-fips algorithms in
55 FIPS mode. Remove DES2 from selftests.
58 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
59 return internal method without any ENGINE dependencies. Add new
60 tiny fips sign and verify functions.
63 *) New build option no-ec2m to disable characteristic 2 code.
66 *) New build option "fipscanisteronly". This only builds fipscanister.o
67 and (currently) associated fips utilities. Uses the file Makefile.fips
68 instead of Makefile.org as the prototype.
71 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
72 Update fips_gcmtest to use IV generator.
75 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
76 setting output buffer to NULL. The *Final function must be
77 called although it will not retrieve any additional data. The tag
78 can be set or retrieved with a ctrl. The IV length is by default 12
79 bytes (96 bits) but can be set to an alternative value. If the IV
80 length exceeds the maximum IV length (currently 16 bytes) it cannot be
84 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
85 underlying do_cipher function handles all cipher semantics itself
86 including padding and finalisation. This is useful if (for example)
87 an ENGINE cipher handles block padding itself. The behaviour of
88 do_cipher is subtly changed if this flag is set: the return value
89 is the number of characters written to the output buffer (zero is
90 no longer an error code) or a negative error code. Also if the
91 input buffer is NULL and length 0 finalisation should be performed.
94 *) If a candidate issuer certificate is already part of the constructed
95 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
98 *) Improve forward-security support: add functions
100 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
101 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
103 for use by SSL/TLS servers; the callback function will be called whenever a
104 new session is created, and gets to decide whether the session may be
105 cached to make it resumable (return 0) or not (return 1). (As by the
106 SSL/TLS protocol specifications, the session_id sent by the server will be
107 empty to indicate that the session is not resumable; also, the server will
108 not generate RFC 4507 (RFC 5077) session tickets.)
110 A simple reasonable callback implementation is to return is_forward_secure.
111 This parameter will be set to 1 or 0 depending on the ciphersuite selected
112 by the SSL/TLS server library, indicating whether it can provide forward
114 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
116 *) Add Next Protocol Negotiation,
117 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
118 disabled with a no-npn flag to config or Configure. Code donated
120 [Adam Langley <agl@google.com> and Ben Laurie]
122 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
123 all platforms. Move ssize_t definition from e_os.h to the public
124 header file e_os2.h as it now appears in public header file cms.h
127 *) New function OPENSSL_gmtime_diff to find the difference in days
128 and seconds between two tm structures. This will be used to provide
129 additional functionality for ASN1_TIME.
132 *) New -sigopt option to the ca, req and x509 utilities. Additional
133 signature parameters can be passed using this option and in
137 *) Add RSA PSS signing function. This will generate and set the
138 appropriate AlgorithmIdentifiers for PSS based on those in the
139 corresponding EVP_MD_CTX structure. No application support yet.
142 *) Support for companion algorithm specific ASN1 signing routines.
143 New function ASN1_item_sign_ctx() signs a pre-initialised
144 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
145 the appropriate parameters.
148 *) Add new algorithm specific ASN1 verification initialisation function
149 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
150 handling will be the same no matter what EVP_PKEY_METHOD is used.
151 Add a PSS handler to support verification of PSS signatures: checked
152 against a number of sample certificates.
155 *) Add signature printing for PSS. Add PSS OIDs.
156 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
158 *) Add algorithm specific signature printing. An individual ASN1 method
159 can now print out signatures instead of the standard hex dump.
161 More complex signatures (e.g. PSS) can print out more meaningful
162 information. Include DSA version that prints out the signature
166 *) Add -trusted_first option which attempts to find certificates in the
167 trusted store even if an untrusted chain is also supplied.
170 *) Initial experimental support for explicitly trusted non-root CAs.
171 OpenSSL still tries to build a complete chain to a root but if an
172 intermediate CA has a trust setting included that is used. The first
173 setting is used: whether to trust or reject.
176 *) New -verify_name option in command line utilities to set verification
180 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
181 Add CMAC pkey methods.
184 *) Experiemental regnegotiation in s_server -www mode. If the client
185 browses /reneg connection is renegotiated. If /renegcert it is
186 renegotiated requesting a certificate.
189 *) Add an "external" session cache for debugging purposes to s_server. This
190 should help trace issues which normally are only apparent in deployed
191 multi-process servers.
194 *) Experiemental password based recipient info support for CMS library:
195 implementing RFC3211.
198 *) Split password based encryption into PBES2 and PBKDF2 functions. This
199 neatly separates the code into cipher and PBE sections and is required
200 for some algorithms that split PBES2 into separate pieces (such as
204 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
205 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
206 BIO_set_cipher() and some obscure PEM functions were changed so they
207 can now return an error. The RAND changes required a change to the
208 RAND_METHOD structure.
211 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
212 a gcc attribute to warn if the result of a function is ignored. This
213 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
214 whose return value is often ignored.
217 Changes between 1.0.0d and 1.0.1 [xx XXX xxxx]
219 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
222 *) Add EC_GFp_nistp224_method(), a 64-bit optimized implementation for
223 elliptic curve NIST-P224 with constant-time single point multiplication on
224 typical inputs. EC_GROUP_new_by_curve_name() will automatically use this
225 (while EC_GROUP_new_curve_GFp() currently won't and prefers the more
226 flexible implementations).
228 The implementation requires support for the nonstandard type __uint128_t,
229 and so is disabled by default. To include this in your build of OpenSSL,
230 use -DEC_NISTP224_64_GCC_128 on the Configure (or config) command line,
231 and run "make depend" (or "make update").
232 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
234 *) Permit abbreviated handshakes when renegotiating using the function
235 SSL_renegotiate_abbreviated().
236 [Robin Seggelmann <seggelmann@fh-muenster.de>]
238 *) Add call to ENGINE_register_all_complete() to
239 ENGINE_load_builtin_engines(), so some implementations get used
240 automatically instead of needing explicit application support.
243 *) Add support for TLS key exporter as described in RFC5705.
244 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
246 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
247 a few changes are required:
249 Add SSL_OP_NO_TLSv1_1 flag.
251 Update version checking logic to handle version 1.1.
252 Add explicit IV handling (ported from DTLS code).
253 Add command line options to s_client/s_server.
256 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
258 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
259 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
261 *) Fix bug in string printing code: if *any* escaping is enabled we must
262 escape the escape character (backslash) or the resulting string is
266 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
268 *) Disable code workaround for ancient and obsolete Netscape browsers
269 and servers: an attacker can use it in a ciphersuite downgrade attack.
270 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
273 *) Fixed J-PAKE implementation error, originally discovered by
274 Sebastien Martini, further info and confirmation from Stefan
275 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
278 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
280 *) Fix extension code to avoid race conditions which can result in a buffer
281 overrun vulnerability: resumed sessions must not be modified as they can
282 be shared by multiple threads. CVE-2010-3864
285 *) Fix WIN32 build system to correctly link an ENGINE directory into
289 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
291 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
293 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
295 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
297 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
298 context. The operation can be customised via the ctrl mechanism in
299 case ENGINEs want to include additional functionality.
302 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
305 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
306 output hashes compatible with older versions of OpenSSL.
307 [Willy Weisz <weisz@vcpc.univie.ac.at>]
309 *) Fix compression algorithm handling: if resuming a session use the
310 compression algorithm of the resumed session instead of determining
311 it from client hello again. Don't allow server to change algorithm.
314 *) Add load_crls() function to apps tidying load_certs() too. Add option
315 to verify utility to allow additional CRLs to be included.
318 *) Update OCSP request code to permit adding custom headers to the request:
319 some responders need this.
322 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
324 [Julia Lawall <julia@diku.dk>]
326 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
327 needlessly dereferenced structures, used obsolete functions and
328 didn't handle all updated verify codes correctly.
331 *) Disable MD2 in the default configuration.
334 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
335 indicate the initial BIO being pushed or popped. This makes it possible
336 to determine whether the BIO is the one explicitly called or as a result
337 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
338 it handles reference counts correctly and doesn't zero out the I/O bio
339 when it is not being explicitly popped. WARNING: applications which
340 included workarounds for the old buggy behaviour will need to be modified
341 or they could free up already freed BIOs.
344 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
345 renaming to all platforms (within the 0.9.8 branch, this was
346 done conditionally on Netware platforms to avoid a name clash).
347 [Guenter <lists@gknw.net>]
349 *) Add ECDHE and PSK support to DTLS.
350 [Michael Tuexen <tuexen@fh-muenster.de>]
352 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
356 *) Add "missing" function EVP_MD_flags() (without this the only way to
357 retrieve a digest flags is by accessing the structure directly. Update
358 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
359 or cipher is registered as in the "from" argument. Print out all
360 registered digests in the dgst usage message instead of manually
361 attempting to work them out.
364 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
365 this allows the use of compression and extensions. Change default cipher
366 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
367 by default unless an application cipher string requests it.
370 *) Alter match criteria in PKCS12_parse(). It used to try to use local
371 key ids to find matching certificates and keys but some PKCS#12 files
372 don't follow the (somewhat unwritten) rules and this strategy fails.
373 Now just gather all certificates together and the first private key
374 then look for the first certificate that matches the key.
377 *) Support use of registered digest and cipher names for dgst and cipher
378 commands instead of having to add each one as a special case. So now
385 openssl dgst -sha256 foo
387 and this works for ENGINE based algorithms too.
391 *) Update Gost ENGINE to support parameter files.
392 [Victor B. Wagner <vitus@cryptocom.ru>]
394 *) Support GeneralizedTime in ca utility.
395 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
397 *) Enhance the hash format used for certificate directory links. The new
398 form uses the canonical encoding (meaning equivalent names will work
399 even if they aren't identical) and uses SHA1 instead of MD5. This form
400 is incompatible with the older format and as a result c_rehash should
401 be used to rebuild symbolic links.
404 *) Make PKCS#8 the default write format for private keys, replacing the
405 traditional format. This form is standardised, more secure and doesn't
406 include an implicit MD5 dependency.
409 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
410 committed to OpenSSL should pass this lot as a minimum.
413 *) Add session ticket override functionality for use by EAP-FAST.
414 [Jouni Malinen <j@w1.fi>]
416 *) Modify HMAC functions to return a value. Since these can be implemented
417 in an ENGINE errors can occur.
420 *) Type-checked OBJ_bsearch_ex.
423 *) Type-checked OBJ_bsearch. Also some constification necessitated
424 by type-checking. Still to come: TXT_DB, bsearch(?),
425 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
429 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
430 seconds to a tm structure directly, instead of going through OS
431 specific date routines. This avoids any issues with OS routines such
432 as the year 2038 bug. New *_adj() functions for ASN1 time structures
433 and X509_time_adj_ex() to cover the extended range. The existing
434 X509_time_adj() is still usable and will no longer have any date issues.
437 *) Delta CRL support. New use deltas option which will attempt to locate
438 and search any appropriate delta CRLs available.
440 This work was sponsored by Google.
443 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
444 code and add additional score elements. Validate alternate CRL paths
445 as part of the CRL checking and indicate a new error "CRL path validation
446 error" in this case. Applications wanting additional details can use
447 the verify callback and check the new "parent" field. If this is not
448 NULL CRL path validation is taking place. Existing applications wont
449 see this because it requires extended CRL support which is off by
452 This work was sponsored by Google.
455 *) Support for freshest CRL extension.
457 This work was sponsored by Google.
460 *) Initial indirect CRL support. Currently only supported in the CRLs
461 passed directly and not via lookup. Process certificate issuer
462 CRL entry extension and lookup CRL entries by bother issuer name
463 and serial number. Check and process CRL issuer entry in IDP extension.
465 This work was sponsored by Google.
468 *) Add support for distinct certificate and CRL paths. The CRL issuer
469 certificate is validated separately in this case. Only enabled if
470 an extended CRL support flag is set: this flag will enable additional
471 CRL functionality in future.
473 This work was sponsored by Google.
476 *) Add support for policy mappings extension.
478 This work was sponsored by Google.
481 *) Fixes to pathlength constraint, self issued certificate handling,
482 policy processing to align with RFC3280 and PKITS tests.
484 This work was sponsored by Google.
487 *) Support for name constraints certificate extension. DN, email, DNS
488 and URI types are currently supported.
490 This work was sponsored by Google.
493 *) To cater for systems that provide a pointer-based thread ID rather
494 than numeric, deprecate the current numeric thread ID mechanism and
495 replace it with a structure and associated callback type. This
496 mechanism allows a numeric "hash" to be extracted from a thread ID in
497 either case, and on platforms where pointers are larger than 'long',
498 mixing is done to help ensure the numeric 'hash' is usable even if it
499 can't be guaranteed unique. The default mechanism is to use "&errno"
500 as a pointer-based thread ID to distinguish between threads.
502 Applications that want to provide their own thread IDs should now use
503 CRYPTO_THREADID_set_callback() to register a callback that will call
504 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
506 Note that ERR_remove_state() is now deprecated, because it is tied
507 to the assumption that thread IDs are numeric. ERR_remove_state(0)
508 to free the current thread's error state should be replaced by
509 ERR_remove_thread_state(NULL).
511 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
512 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
513 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
514 application was previously providing a numeric thread callback that
515 was inappropriate for distinguishing threads, then uniqueness might
516 have been obtained with &errno that happened immediately in the
517 intermediate development versions of OpenSSL; this is no longer the
518 case, the numeric thread callback will now override the automatic use
520 [Geoff Thorpe, with help from Bodo Moeller]
522 *) Initial support for different CRL issuing certificates. This covers a
523 simple case where the self issued certificates in the chain exist and
524 the real CRL issuer is higher in the existing chain.
526 This work was sponsored by Google.
529 *) Removed effectively defunct crypto/store from the build.
532 *) Revamp of STACK to provide stronger type-checking. Still to come:
533 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
534 ASN1_STRING, CONF_VALUE.
537 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
538 RAM on SSL connections. This option can save about 34k per idle SSL.
541 *) Revamp of LHASH to provide stronger type-checking. Still to come:
542 STACK, TXT_DB, bsearch, qsort.
545 *) Initial support for Cryptographic Message Syntax (aka CMS) based
546 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
547 support for data, signedData, compressedData, digestedData and
548 encryptedData, envelopedData types included. Scripts to check against
549 RFC4134 examples draft and interop and consistency checks of many
550 content types and variants.
553 *) Add options to enc utility to support use of zlib compression BIO.
556 *) Extend mk1mf to support importing of options and assembly language
557 files from Configure script, currently only included in VC-WIN32.
558 The assembly language rules can now optionally generate the source
559 files from the associated perl scripts.
562 *) Implement remaining functionality needed to support GOST ciphersuites.
563 Interop testing has been performed using CryptoPro implementations.
564 [Victor B. Wagner <vitus@cryptocom.ru>]
566 *) s390x assembler pack.
569 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
573 *) Implement Opaque PRF Input TLS extension as specified in
574 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
575 official specification yet and no extension type assignment by
576 IANA exists, this extension (for now) will have to be explicitly
577 enabled when building OpenSSL by providing the extension number
578 to use. For example, specify an option
580 -DTLSEXT_TYPE_opaque_prf_input=0x9527
582 to the "config" or "Configure" script to enable the extension,
583 assuming extension number 0x9527 (which is a completely arbitrary
584 and unofficial assignment based on the MD5 hash of the Internet
585 Draft). Note that by doing so, you potentially lose
586 interoperability with other TLS implementations since these might
587 be using the same extension number for other purposes.
589 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
590 opaque PRF input value to use in the handshake. This will create
591 an interal copy of the length-'len' string at 'src', and will
592 return non-zero for success.
594 To get more control and flexibility, provide a callback function
597 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
598 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
602 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
605 Callback function 'cb' will be called in handshakes, and is
606 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
607 Argument 'arg' is for application purposes (the value as given to
608 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
609 be provided to the callback function). The callback function
610 has to return non-zero to report success: usually 1 to use opaque
611 PRF input just if possible, or 2 to enforce use of the opaque PRF
612 input. In the latter case, the library will abort the handshake
613 if opaque PRF input is not successfully negotiated.
615 Arguments 'peerinput' and 'len' given to the callback function
616 will always be NULL and 0 in the case of a client. A server will
617 see the client's opaque PRF input through these variables if
618 available (NULL and 0 otherwise). Note that if the server
619 provides an opaque PRF input, the length must be the same as the
620 length of the client's opaque PRF input.
622 Note that the callback function will only be called when creating
623 a new session (session resumption can resume whatever was
624 previously negotiated), and will not be called in SSL 2.0
625 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
626 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
627 for applications that need to enforce opaque PRF input.
631 *) Update ssl code to support digests other than SHA1+MD5 for handshake
634 [Victor B. Wagner <vitus@cryptocom.ru>]
636 *) Add RFC4507 support to OpenSSL. This includes the corrections in
637 RFC4507bis. The encrypted ticket format is an encrypted encoded
638 SSL_SESSION structure, that way new session features are automatically
641 If a client application caches session in an SSL_SESSION structure
642 support is transparent because tickets are now stored in the encoded
645 The SSL_CTX structure automatically generates keys for ticket
646 protection in servers so again support should be possible
647 with no application modification.
649 If a client or server wishes to disable RFC4507 support then the option
650 SSL_OP_NO_TICKET can be set.
652 Add a TLS extension debugging callback to allow the contents of any client
653 or server extensions to be examined.
655 This work was sponsored by Google.
658 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
659 OpenSSL should now compile cleanly on gcc 4.2
660 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
662 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
663 support including streaming MAC support: this is required for GOST
665 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
667 *) Add option -stream to use PKCS#7 streaming in smime utility. New
668 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
669 to output in BER and PEM format.
672 *) Experimental support for use of HMAC via EVP_PKEY interface. This
673 allows HMAC to be handled via the EVP_DigestSign*() interface. The
674 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
675 ENGINE support for HMAC keys which are unextractable. New -mac and
676 -macopt options to dgst utility.
679 *) New option -sigopt to dgst utility. Update dgst to use
680 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
681 alternative signing paramaters such as X9.31 or PSS in the dgst
685 *) Change ssl_cipher_apply_rule(), the internal function that does
686 the work each time a ciphersuite string requests enabling
687 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
688 removing ("!foo+bar") a class of ciphersuites: Now it maintains
689 the order of disabled ciphersuites such that those ciphersuites
690 that most recently went from enabled to disabled not only stay
691 in order with respect to each other, but also have higher priority
692 than other disabled ciphersuites the next time ciphersuites are
695 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
696 the same ciphersuites as with "HIGH" alone, but in a specific
697 order where the PSK ciphersuites come first (since they are the
698 most recently disabled ciphersuites when "HIGH" is parsed).
700 Also, change ssl_create_cipher_list() (using this new
701 funcionality) such that between otherwise identical
702 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
706 *) Change ssl_create_cipher_list() so that it automatically
707 arranges the ciphersuites in reasonable order before starting
708 to process the rule string. Thus, the definition for "DEFAULT"
709 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
710 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
711 This makes it much easier to arrive at a reasonable default order
712 in applications for which anonymous ciphers are OK (meaning
713 that you can't actually use DEFAULT).
714 [Bodo Moeller; suggested by Victor Duchovni]
716 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
717 processing) into multiple integers instead of setting
718 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
719 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
720 (These masks as well as the individual bit definitions are hidden
721 away into the non-exported interface ssl/ssl_locl.h, so this
722 change to the definition of the SSL_CIPHER structure shouldn't
723 affect applications.) This give us more bits for each of these
724 categories, so there is no longer a need to coagulate AES128 and
725 AES256 into a single algorithm bit, and to coagulate Camellia128
726 and Camellia256 into a single algorithm bit, which has led to all
729 Thus, among other things, the kludge introduced in 0.9.7m and
730 0.9.8e for masking out AES256 independently of AES128 or masking
731 out Camellia256 independently of AES256 is not needed here in 0.9.9.
733 With the change, we also introduce new ciphersuite aliases that
734 so far were missing: "AES128", "AES256", "CAMELLIA128", and
738 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
739 Use the leftmost N bytes of the signature input if the input is
740 larger than the prime q (with N being the size in bytes of q).
743 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
744 it yet and it is largely untested.
747 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
750 *) Initial incomplete changes to avoid need for function casts in OpenSSL
751 some compilers (gcc 4.2 and later) reject their use. Safestack is
752 reimplemented. Update ASN1 to avoid use of legacy functions.
755 *) Win32/64 targets are linked with Winsock2.
758 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
759 to external functions. This can be used to increase CRL handling
760 efficiency especially when CRLs are very large by (for example) storing
761 the CRL revoked certificates in a database.
764 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
765 new CRLs added to a directory can be used. New command line option
766 -verify_return_error to s_client and s_server. This causes real errors
767 to be returned by the verify callback instead of carrying on no matter
768 what. This reflects the way a "real world" verify callback would behave.
771 *) GOST engine, supporting several GOST algorithms and public key formats.
772 Kindly donated by Cryptocom.
775 *) Partial support for Issuing Distribution Point CRL extension. CRLs
776 partitioned by DP are handled but no indirect CRL or reason partitioning
777 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
778 selected via a scoring technique which handles IDP and AKID in CRLs.
781 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
782 will ultimately be used for all verify operations: this will remove the
783 X509_STORE dependency on certificate verification and allow alternative
784 lookup methods. X509_STORE based implementations of these two callbacks.
787 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
788 Modify get_crl() to find a valid (unexpired) CRL if possible.
791 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
792 this would be called X509_CRL_cmp() but that name is already used by
793 a function that just compares CRL issuer names. Cache several CRL
794 extensions in X509_CRL structure and cache CRLDP in X509.
797 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
798 this maps equivalent X509_NAME structures into a consistent structure.
799 Name comparison can then be performed rapidly using memcmp().
802 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
806 *) Allow digests to supply their own micalg string for S/MIME type using
807 the ctrl EVP_MD_CTRL_MICALG.
810 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
811 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
812 ctrl. It can then customise the structure before and/or after signing
816 *) New function OBJ_add_sigid() to allow application defined signature OIDs
817 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
818 to free up any added signature OIDs.
821 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
822 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
823 digest and cipher tables. New options added to openssl utility:
824 list-message-digest-algorithms and list-cipher-algorithms.
827 *) Change the array representation of binary polynomials: the list
828 of degrees of non-zero coefficients is now terminated with -1.
829 Previously it was terminated with 0, which was also part of the
830 value; thus, the array representation was not applicable to
831 polynomials where t^0 has coefficient zero. This change makes
832 the array representation useful in a more general context.
835 *) Various modifications and fixes to SSL/TLS cipher string
836 handling. For ECC, the code now distinguishes between fixed ECDH
837 with RSA certificates on the one hand and with ECDSA certificates
838 on the other hand, since these are separate ciphersuites. The
839 unused code for Fortezza ciphersuites has been removed.
841 For consistency with EDH, ephemeral ECDH is now called "EECDH"
842 (not "ECDHE"). For consistency with the code for DH
843 certificates, use of ECDH certificates is now considered ECDH
844 authentication, not RSA or ECDSA authentication (the latter is
845 merely the CA's signing algorithm and not actively used in the
848 The temporary ciphersuite alias "ECCdraft" is no longer
849 available, and ECC ciphersuites are no longer excluded from "ALL"
850 and "DEFAULT". The following aliases now exist for RFC 4492
851 ciphersuites, most of these by analogy with the DH case:
853 kECDHr - ECDH cert, signed with RSA
854 kECDHe - ECDH cert, signed with ECDSA
855 kECDH - ECDH cert (signed with either RSA or ECDSA)
856 kEECDH - ephemeral ECDH
857 ECDH - ECDH cert or ephemeral ECDH
863 AECDH - anonymous ECDH
864 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
868 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
869 Use correct micalg parameters depending on digest(s) in signed message.
872 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
873 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
876 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
877 an engine to register a method. Add ENGINE lookups for methods and
878 functional reference processing.
881 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
882 EVP_{Sign,Verify}* which allow an application to customise the signature
886 *) New -resign option to smime utility. This adds one or more signers
887 to an existing PKCS#7 signedData structure. Also -md option to use an
888 alternative message digest algorithm for signing.
891 *) Tidy up PKCS#7 routines and add new functions to make it easier to
892 create PKCS7 structures containing multiple signers. Update smime
893 application to support multiple signers.
896 *) New -macalg option to pkcs12 utility to allow setting of an alternative
900 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
901 Reorganize PBE internals to lookup from a static table using NIDs,
902 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
903 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
904 PRF which will be automatically used with PBES2.
907 *) Replace the algorithm specific calls to generate keys in "req" with the
911 *) Update PKCS#7 enveloped data routines to use new API. This is now
912 supported by any public key method supporting the encrypt operation. A
913 ctrl is added to allow the public key algorithm to examine or modify
914 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
918 *) Add a ctrl to asn1 method to allow a public key algorithm to express
919 a default digest type to use. In most cases this will be SHA1 but some
920 algorithms (such as GOST) need to specify an alternative digest. The
921 return value indicates how strong the prefernce is 1 means optional and
922 2 is mandatory (that is it is the only supported type). Modify
923 ASN1_item_sign() to accept a NULL digest argument to indicate it should
924 use the default md. Update openssl utilities to use the default digest
925 type for signing if it is not explicitly indicated.
928 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
929 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
930 signing method from the key type. This effectively removes the link
931 between digests and public key types.
934 *) Add an OID cross reference table and utility functions. Its purpose is to
935 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
936 rsaEncryption. This will allow some of the algorithm specific hackery
937 needed to use the correct OID to be removed.
940 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
941 structures for PKCS7_sign(). They are now set up by the relevant public
945 *) Add provisional EC pkey method with support for ECDSA and ECDH.
948 *) Add support for key derivation (agreement) in the API, DH method and
952 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
953 public and private key formats. As a side effect these add additional
954 command line functionality not previously available: DSA signatures can be
955 generated and verified using pkeyutl and DH key support and generation in
960 [Oliver Tappe <zooey@hirschkaefer.de>]
962 *) New make target "install_html_docs" installs HTML renditions of the
964 [Oliver Tappe <zooey@hirschkaefer.de>]
966 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
967 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
968 support key and parameter generation and add initial key generation
969 functionality for RSA.
972 *) Add functions for main EVP_PKEY_method operations. The undocumented
973 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
974 EVP_PKEY_{encrypt,decrypt}_old.
977 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
978 key API, doesn't do much yet.
981 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
982 public key algorithms. New option to openssl utility:
983 "list-public-key-algorithms" to print out info.
986 *) Implement the Supported Elliptic Curves Extension for
987 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
990 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
991 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
994 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
995 utilities such as rsa, dsa, dsaparam etc except they process any key
999 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1000 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1001 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1005 *) Initial support for pluggable public key ASN1.
1006 De-spaghettify the public key ASN1 handling. Move public and private
1007 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1008 algorithm specific handling to a single module within the relevant
1009 algorithm directory. Add functions to allow (near) opaque processing
1010 of public and private key structures.
1013 *) Implement the Supported Point Formats Extension for
1014 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1017 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1018 for the psk identity [hint] and the psk callback functions to the
1019 SSL_SESSION, SSL and SSL_CTX structure.
1022 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1026 SSL_CTX_use_psk_identity_hint
1027 SSL_get_psk_identity_hint
1028 SSL_get_psk_identity
1029 SSL_use_psk_identity_hint
1031 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1033 *) Add RFC 3161 compliant time stamp request creation, response generation
1034 and response verification functionality.
1035 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1037 *) Add initial support for TLS extensions, specifically for the server_name
1038 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1039 have new members for a host name. The SSL data structure has an
1040 additional member SSL_CTX *initial_ctx so that new sessions can be
1041 stored in that context to allow for session resumption, even after the
1042 SSL has been switched to a new SSL_CTX in reaction to a client's
1043 server_name extension.
1045 New functions (subject to change):
1047 SSL_get_servername()
1048 SSL_get_servername_type()
1051 New CTRL codes and macros (subject to change):
1053 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1054 - SSL_CTX_set_tlsext_servername_callback()
1055 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1056 - SSL_CTX_set_tlsext_servername_arg()
1057 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1059 openssl s_client has a new '-servername ...' option.
1061 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1062 '-key2 ...', '-servername_fatal' (subject to change). This allows
1063 testing the HostName extension for a specific single host name ('-cert'
1064 and '-key' remain fallbacks for handshakes without HostName
1065 negotiation). If the unrecogninzed_name alert has to be sent, this by
1066 default is a warning; it becomes fatal with the '-servername_fatal'
1069 [Peter Sylvester, Remy Allais, Christophe Renou]
1071 *) Whirlpool hash implementation is added.
1074 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1075 bn(64,32). Because of instruction set limitations it doesn't have
1076 any negative impact on performance. This was done mostly in order
1077 to make it possible to share assembler modules, such as bn_mul_mont
1078 implementations, between 32- and 64-bit builds without hassle.
1081 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1082 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1086 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1087 dedicated Montgomery multiplication procedure, is introduced.
1088 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1089 "64-bit" performance on certain 32-bit targets.
1092 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1093 in SSL structures. New SSL ctrl to set maximum send fragment size.
1094 Save memory by seeting the I/O buffer sizes dynamically instead of
1095 using the maximum available value.
1098 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1099 in addition to the text details.
1102 *) Very, very preliminary EXPERIMENTAL support for printing of general
1103 ASN1 structures. This currently produces rather ugly output and doesn't
1104 handle several customised structures at all.
1107 *) Integrated support for PVK file format and some related formats such
1108 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1109 these in the 'rsa' and 'dsa' utilities.
1112 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1115 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1116 place for the (very old) "NETSCAPE" format certificates which are now
1117 handled using new ASN1 code equivalents.
1120 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1121 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1122 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1125 *) Modify CRL distribution points extension code to print out previously
1126 unsupported fields. Enhance extension setting code to allow setting of
1130 *) Add print and set support for Issuing Distribution Point CRL extension.
1133 *) Change 'Configure' script to enable Camellia by default.
1136 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1138 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1139 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1141 *) Fix bug in string printing code: if *any* escaping is enabled we must
1142 escape the escape character (backslash) or the resulting string is
1146 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1148 *) Disable code workaround for ancient and obsolete Netscape browsers
1149 and servers: an attacker can use it in a ciphersuite downgrade attack.
1150 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1153 *) Fixed J-PAKE implementation error, originally discovered by
1154 Sebastien Martini, further info and confirmation from Stefan
1155 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1158 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1160 *) Fix extension code to avoid race conditions which can result in a buffer
1161 overrun vulnerability: resumed sessions must not be modified as they can
1162 be shared by multiple threads. CVE-2010-3864
1165 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1168 *) Don't reencode certificate when calculating signature: cache and use
1169 the original encoding instead. This makes signature verification of
1170 some broken encodings work correctly.
1173 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1174 is also one of the inputs.
1175 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1177 *) Don't repeatedly append PBE algorithms to table if they already exist.
1178 Sort table on each new add. This effectively makes the table read only
1179 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1183 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1185 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1188 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1189 access or freeing data twice (CVE-2010-0742)
1190 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1192 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1193 common in certificates and some applications which only call
1194 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1198 Reduce copying into .apps and .test in makevms.com
1199 Don't try to use blank CA certificate in CA.com
1200 Allow use of C files from original directories in maketests.com
1201 [Steven M. Schweda" <sms@antinode.info>]
1203 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1205 *) When rejecting SSL/TLS records due to an incorrect version number, never
1206 update s->server with a new major version number. As of
1207 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1208 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1209 the previous behavior could result in a read attempt at NULL when
1210 receiving specific incorrect SSL/TLS records once record payload
1211 protection is active. (CVE-2010-0740)
1212 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1214 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1215 could be crashed if the relevant tables were not present (e.g. chrooted).
1216 [Tomas Hoger <thoger@redhat.com>]
1218 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1220 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1221 [Martin Olsson, Neel Mehta]
1223 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1224 accommodate for stack sorting, always a write lock!).
1227 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1228 excessive delays in the RAND_poll(): over a minute. As a workaround
1229 include a time check in the inner Heap32Next loop too.
1232 *) The code that handled flushing of data in SSL/TLS originally used the
1233 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1234 the problem outlined in PR#1949. The fix suggested there however can
1235 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1236 of Apache). So instead simplify the code to flush unconditionally.
1237 This should be fine since flushing with no data to flush is a no op.
1240 *) Handle TLS versions 2.0 and later properly and correctly use the
1241 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1242 off ancient servers have a habit of sticking around for a while...
1245 *) Modify compression code so it frees up structures without using the
1246 ex_data callbacks. This works around a problem where some applications
1247 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1248 restarting) then use compression (e.g. SSL with compression) later.
1249 This results in significant per-connection memory leaks and
1250 has caused some security issues including CVE-2008-1678 and
1254 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1255 change when encrypting or decrypting.
1258 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1259 connect and renegotiate with servers which do not support RI.
1260 Until RI is more widely deployed this option is enabled by default.
1263 *) Add "missing" ssl ctrls to clear options and mode.
1266 *) If client attempts to renegotiate and doesn't support RI respond with
1267 a no_renegotiation alert as required by RFC5746. Some renegotiating
1268 TLS clients will continue a connection gracefully when they receive
1269 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1270 waiting for a server hello which it will never receive. Now we treat a
1271 received no_renegotiation alert as a fatal error. This is because
1272 applications requesting a renegotiation might well expect it to succeed
1273 and would have no code in place to handle the server denying it so the
1274 only safe thing to do is to terminate the connection.
1277 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1278 peer supports secure renegotiation and 0 otherwise. Print out peer
1279 renegotiation support in s_client/s_server.
1282 *) Replace the highly broken and deprecated SPKAC certification method with
1283 the updated NID creation version. This should correctly handle UTF8.
1286 *) Implement RFC5746. Re-enable renegotiation but require the extension
1287 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1288 turns out to be a bad idea. It has been replaced by
1289 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1290 SSL_CTX_set_options(). This is really not recommended unless you
1291 know what you are doing.
1292 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1294 *) Fixes to stateless session resumption handling. Use initial_ctx when
1295 issuing and attempting to decrypt tickets in case it has changed during
1296 servername handling. Use a non-zero length session ID when attempting
1297 stateless session resumption: this makes it possible to determine if
1298 a resumption has occurred immediately after receiving server hello
1299 (several places in OpenSSL subtly assume this) instead of later in
1303 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1304 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1305 fixes for a few places where the return code is not checked
1307 [Julia Lawall <julia@diku.dk>]
1309 *) Add --strict-warnings option to Configure script to include devteam
1310 warnings in other configurations.
1313 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1314 makes it possible to install openssl libraries in locations which
1315 have names other than "lib", for example "/usr/lib64" which some
1317 [Steve Henson, based on patch from Jeremy Utley]
1319 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1320 X690 8.9.12 and can produce some misleading textual output of OIDs.
1321 [Steve Henson, reported by Dan Kaminsky]
1323 *) Delete MD2 from algorithm tables. This follows the recommendation in
1324 several standards that it is not used in new applications due to
1325 several cryptographic weaknesses. For binary compatibility reasons
1326 the MD2 API is still compiled in by default.
1329 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1333 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1334 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1336 [Guenter <lists@gknw.net>]
1338 *) Fix the server certificate chain building code to use X509_verify_cert(),
1339 it used to have an ad-hoc builder which was unable to cope with anything
1340 other than a simple chain.
1341 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1343 *) Don't check self signed certificate signatures in X509_verify_cert()
1344 by default (a flag can override this): it just wastes time without
1345 adding any security. As a useful side effect self signed root CAs
1346 with non-FIPS digests are now usable in FIPS mode.
1349 *) In dtls1_process_out_of_seq_message() the check if the current message
1350 is already buffered was missing. For every new message was memory
1351 allocated, allowing an attacker to perform an denial of service attack
1352 with sending out of seq handshake messages until there is no memory
1353 left. Additionally every future messege was buffered, even if the
1354 sequence number made no sense and would be part of another handshake.
1355 So only messages with sequence numbers less than 10 in advance will be
1356 buffered. (CVE-2009-1378)
1357 [Robin Seggelmann, discovered by Daniel Mentz]
1359 *) Records are buffered if they arrive with a future epoch to be
1360 processed after finishing the corresponding handshake. There is
1361 currently no limitation to this buffer allowing an attacker to perform
1362 a DOS attack with sending records with future epochs until there is no
1363 memory left. This patch adds the pqueue_size() function to detemine
1364 the size of a buffer and limits the record buffer to 100 entries.
1366 [Robin Seggelmann, discovered by Daniel Mentz]
1368 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1369 parent structure is freed. (CVE-2009-1379)
1372 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1373 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1376 [Ilya O. <vrghost@gmail.com>]
1378 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1380 *) Disable renegotiation completely - this fixes a severe security
1381 problem (CVE-2009-3555) at the cost of breaking all
1382 renegotiation. Renegotiation can be re-enabled by setting
1383 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1384 run-time. This is really not recommended unless you know what
1388 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1390 *) Don't set val to NULL when freeing up structures, it is freed up by
1391 underlying code. If sizeof(void *) > sizeof(long) this can result in
1392 zeroing past the valid field. (CVE-2009-0789)
1393 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1395 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1396 checked correctly. This would allow some invalid signed attributes to
1397 appear to verify correctly. (CVE-2009-0591)
1398 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1400 *) Reject UniversalString and BMPString types with invalid lengths. This
1401 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1402 a legal length. (CVE-2009-0590)
1405 *) Set S/MIME signing as the default purpose rather than setting it
1406 unconditionally. This allows applications to override it at the store
1410 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1411 to handle some structures.
1414 *) Improve efficiency of mem_gets: don't search whole buffer each time
1416 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1418 *) New -hex option for openssl rand.
1421 *) Print out UTF8String and NumericString when parsing ASN1.
1424 *) Support NumericString type for name components.
1427 *) Allow CC in the environment to override the automatically chosen
1428 compiler. Note that nothing is done to ensure flags work with the
1432 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1434 *) Properly check EVP_VerifyFinal() and similar return values
1436 [Ben Laurie, Bodo Moeller, Google Security Team]
1438 *) Enable TLS extensions by default.
1441 *) Allow the CHIL engine to be loaded, whether the application is
1442 multithreaded or not. (This does not release the developer from the
1443 obligation to set up the dynamic locking callbacks.)
1444 [Sander Temme <sander@temme.net>]
1446 *) Use correct exit code if there is an error in dgst command.
1447 [Steve Henson; problem pointed out by Roland Dirlewanger]
1449 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1450 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1453 *) Add experimental JPAKE support, including demo authentication in
1454 s_client and s_server.
1457 *) Set the comparison function in v3_addr_canonize().
1458 [Rob Austein <sra@hactrn.net>]
1460 *) Add support for XMPP STARTTLS in s_client.
1461 [Philip Paeps <philip@freebsd.org>]
1463 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1464 to ensure that even with this option, only ciphersuites in the
1465 server's preference list will be accepted. (Note that the option
1466 applies only when resuming a session, so the earlier behavior was
1467 just about the algorithm choice for symmetric cryptography.)
1470 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1472 *) Fix NULL pointer dereference if a DTLS server received
1473 ChangeCipherSpec as first record (CVE-2009-1386).
1476 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1477 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1480 *) The fix in 0.9.8c that supposedly got rid of unsafe
1481 double-checked locking was incomplete for RSA blinding,
1482 addressing just one layer of what turns out to have been
1483 doubly unsafe triple-checked locking.
1485 So now fix this for real by retiring the MONT_HELPER macro
1486 in crypto/rsa/rsa_eay.c.
1488 [Bodo Moeller; problem pointed out by Marius Schilder]
1490 *) Various precautionary measures:
1492 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1494 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1495 (NB: This would require knowledge of the secret session ticket key
1496 to exploit, in which case you'd be SOL either way.)
1498 - Change bn_nist.c so that it will properly handle input BIGNUMs
1499 outside the expected range.
1501 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1504 [Neel Mehta, Bodo Moeller]
1506 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1507 the load fails. Useful for distros.
1508 [Ben Laurie and the FreeBSD team]
1510 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1513 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1516 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1518 This work was sponsored by Logica.
1521 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1522 keystores. Support for SSL/TLS client authentication too.
1523 Not compiled unless enable-capieng specified to Configure.
1525 This work was sponsored by Logica.
1528 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1529 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1530 attribute creation routines such as certifcate requests and PKCS#12
1534 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1536 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1537 handshake which could lead to a cilent crash as found using the
1538 Codenomicon TLS test suite (CVE-2008-1672)
1539 [Steve Henson, Mark Cox]
1541 *) Fix double free in TLS server name extensions which could lead to
1542 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1545 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1547 Clear the error queue to ensure that error entries left from
1548 older function calls do not interfere with the correct operation.
1549 [Lutz Jaenicke, Erik de Castro Lopo]
1551 *) Remove root CA certificates of commercial CAs:
1553 The OpenSSL project does not recommend any specific CA and does not
1554 have any policy with respect to including or excluding any CA.
1555 Therefore it does not make any sense to ship an arbitrary selection
1556 of root CA certificates with the OpenSSL software.
1559 *) RSA OAEP patches to fix two separate invalid memory reads.
1560 The first one involves inputs when 'lzero' is greater than
1561 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1562 before the beginning of from). The second one involves inputs where
1563 the 'db' section contains nothing but zeroes (there is a one-byte
1564 invalid read after the end of 'db').
1565 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1567 *) Partial backport from 0.9.9-dev:
1569 Introduce bn_mul_mont (dedicated Montgomery multiplication
1570 procedure) as a candidate for BIGNUM assembler implementation.
1571 While 0.9.9-dev uses assembler for various architectures, only
1572 x86_64 is available by default here in the 0.9.8 branch, and
1573 32-bit x86 is available through a compile-time setting.
1575 To try the 32-bit x86 assembler implementation, use Configure
1576 option "enable-montasm" (which exists only for this backport).
1578 As "enable-montasm" for 32-bit x86 disclaims code stability
1579 anyway, in this constellation we activate additional code
1580 backported from 0.9.9-dev for further performance improvements,
1581 namely BN_from_montgomery_word. (To enable this otherwise,
1582 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1584 [Andy Polyakov (backport partially by Bodo Moeller)]
1586 *) Add TLS session ticket callback. This allows an application to set
1587 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1588 values. This is useful for key rollover for example where several key
1589 sets may exist with different names.
1592 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1593 This was broken until now in 0.9.8 releases, such that the only way
1594 a registered ENGINE could be used (assuming it initialises
1595 successfully on the host) was to explicitly set it as the default
1596 for the relevant algorithms. This is in contradiction with 0.9.7
1597 behaviour and the documentation. With this fix, when an ENGINE is
1598 registered into a given algorithm's table of implementations, the
1599 'uptodate' flag is reset so that auto-discovery will be used next
1600 time a new context for that algorithm attempts to select an
1602 [Ian Lister (tweaked by Geoff Thorpe)]
1604 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1605 implemention in the following ways:
1607 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1610 Lack of BER streaming support means one pass streaming processing is
1611 only supported if data is detached: setting the streaming flag is
1612 ignored for embedded content.
1614 CMS support is disabled by default and must be explicitly enabled
1615 with the enable-cms configuration option.
1618 *) Update the GMP engine glue to do direct copies between BIGNUM and
1619 mpz_t when openssl and GMP use the same limb size. Otherwise the
1620 existing "conversion via a text string export" trick is still used.
1621 [Paul Sheer <paulsheer@gmail.com>]
1623 *) Zlib compression BIO. This is a filter BIO which compressed and
1624 uncompresses any data passed through it.
1627 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1628 RFC3394 compatible AES key wrapping.
1631 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1632 sets string data without copying. X509_ALGOR_set0() and
1633 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1634 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1635 from an X509_ATTRIBUTE structure optionally checking it occurs only
1636 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1640 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1641 to get the expected BN_FLG_CONSTTIME behavior.
1642 [Bodo Moeller (Google)]
1646 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1647 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1648 - added some more tests to do_tests.pl
1649 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1650 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1651 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1652 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1653 - various changes to netware.pl to enable gcc-cross builds on Win32
1655 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1656 - various changes to fix missing prototype warnings
1657 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1658 - added AES, WHIRLPOOL and CPUID assembler code to build files
1659 - added missing AES assembler make rules to mk1mf.pl
1660 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1661 [Guenter Knauf <eflash@gmx.net>]
1663 *) Implement certificate status request TLS extension defined in RFC3546.
1664 A client can set the appropriate parameters and receive the encoded
1665 OCSP response via a callback. A server can query the supplied parameters
1666 and set the encoded OCSP response in the callback. Add simplified examples
1667 to s_client and s_server.
1670 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1672 *) Fix various bugs:
1673 + Binary incompatibility of ssl_ctx_st structure
1674 + DTLS interoperation with non-compliant servers
1675 + Don't call get_session_cb() without proposed session
1676 + Fix ia64 assembler code
1677 [Andy Polyakov, Steve Henson]
1679 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1681 *) DTLS Handshake overhaul. There were longstanding issues with
1682 OpenSSL DTLS implementation, which were making it impossible for
1683 RFC 4347 compliant client to communicate with OpenSSL server.
1684 Unfortunately just fixing these incompatibilities would "cut off"
1685 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1686 server keeps tolerating non RFC compliant syntax. The opposite is
1687 not true, 0.9.8f client can not communicate with earlier server.
1688 This update even addresses CVE-2007-4995.
1691 *) Changes to avoid need for function casts in OpenSSL: some compilers
1692 (gcc 4.2 and later) reject their use.
1693 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1696 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1697 RFC4507bis. The encrypted ticket format is an encrypted encoded
1698 SSL_SESSION structure, that way new session features are automatically
1701 If a client application caches session in an SSL_SESSION structure
1702 support is transparent because tickets are now stored in the encoded
1705 The SSL_CTX structure automatically generates keys for ticket
1706 protection in servers so again support should be possible
1707 with no application modification.
1709 If a client or server wishes to disable RFC4507 support then the option
1710 SSL_OP_NO_TICKET can be set.
1712 Add a TLS extension debugging callback to allow the contents of any client
1713 or server extensions to be examined.
1715 This work was sponsored by Google.
1718 *) Add initial support for TLS extensions, specifically for the server_name
1719 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1720 have new members for a host name. The SSL data structure has an
1721 additional member SSL_CTX *initial_ctx so that new sessions can be
1722 stored in that context to allow for session resumption, even after the
1723 SSL has been switched to a new SSL_CTX in reaction to a client's
1724 server_name extension.
1726 New functions (subject to change):
1728 SSL_get_servername()
1729 SSL_get_servername_type()
1732 New CTRL codes and macros (subject to change):
1734 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1735 - SSL_CTX_set_tlsext_servername_callback()
1736 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1737 - SSL_CTX_set_tlsext_servername_arg()
1738 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1740 openssl s_client has a new '-servername ...' option.
1742 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1743 '-key2 ...', '-servername_fatal' (subject to change). This allows
1744 testing the HostName extension for a specific single host name ('-cert'
1745 and '-key' remain fallbacks for handshakes without HostName
1746 negotiation). If the unrecogninzed_name alert has to be sent, this by
1747 default is a warning; it becomes fatal with the '-servername_fatal'
1750 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1752 *) Add AES and SSE2 assembly language support to VC++ build.
1755 *) Mitigate attack on final subtraction in Montgomery reduction.
1758 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1759 (which previously caused an internal error).
1762 *) Squeeze another 10% out of IGE mode when in != out.
1765 *) AES IGE mode speedup.
1766 [Dean Gaudet (Google)]
1768 *) Add the Korean symmetric 128-bit cipher SEED (see
1769 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1770 add SEED ciphersuites from RFC 4162:
1772 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1773 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1774 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1775 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1777 To minimize changes between patchlevels in the OpenSSL 0.9.8
1778 series, SEED remains excluded from compilation unless OpenSSL
1779 is configured with 'enable-seed'.
1780 [KISA, Bodo Moeller]
1782 *) Mitigate branch prediction attacks, which can be practical if a
1783 single processor is shared, allowing a spy process to extract
1784 information. For detailed background information, see
1785 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1786 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1787 and Necessary Software Countermeasures"). The core of the change
1788 are new versions BN_div_no_branch() and
1789 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1790 respectively, which are slower, but avoid the security-relevant
1791 conditional branches. These are automatically called by BN_div()
1792 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1793 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1794 remove a conditional branch.
1796 BN_FLG_CONSTTIME is the new name for the previous
1797 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1798 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1799 in the exponent causes BN_mod_exp_mont() to use the alternative
1800 implementation in BN_mod_exp_mont_consttime().) The old name
1801 remains as a deprecated alias.
1803 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1804 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1805 constant-time implementations for more than just exponentiation.
1806 Here too the old name is kept as a deprecated alias.
1808 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1809 the BN_BLINDING structure gets an independent copy of the
1810 modulus. This means that the previous "BIGNUM *m" argument to
1811 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1812 essentially becomes "const BIGNUM *m", although we can't actually
1813 change this in the header file before 0.9.9. It allows
1814 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1815 enable BN_FLG_CONSTTIME.
1817 [Matthew D Wood (Intel Corp)]
1819 *) In the SSL/TLS server implementation, be strict about session ID
1820 context matching (which matters if an application uses a single
1821 external cache for different purposes). Previously,
1822 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1823 set. This did ensure strict client verification, but meant that,
1824 with applications using a single external cache for quite
1825 different requirements, clients could circumvent ciphersuite
1826 restrictions for a given session ID context by starting a session
1827 in a different context.
1830 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1831 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1832 authentication-only ciphersuites.
1835 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1836 not complete and could lead to a possible single byte overflow
1837 (CVE-2007-5135) [Ben Laurie]
1839 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1841 *) Since AES128 and AES256 (and similarly Camellia128 and
1842 Camellia256) share a single mask bit in the logic of
1843 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1844 kludge to work properly if AES128 is available and AES256 isn't
1845 (or if Camellia128 is available and Camellia256 isn't).
1848 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1849 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1850 When a point or a seed is encoded in a BIT STRING, we need to
1851 prevent the removal of trailing zero bits to get the proper DER
1852 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1853 of a NamedBitList, for which trailing 0 bits need to be removed.)
1856 *) Have SSL/TLS server implementation tolerate "mismatched" record
1857 protocol version while receiving ClientHello even if the
1858 ClientHello is fragmented. (The server can't insist on the
1859 particular protocol version it has chosen before the ServerHello
1860 message has informed the client about his choice.)
1863 *) Add RFC 3779 support.
1864 [Rob Austein for ARIN, Ben Laurie]
1866 *) Load error codes if they are not already present instead of using a
1867 static variable. This allows them to be cleanly unloaded and reloaded.
1868 Improve header file function name parsing.
1871 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1872 or CAPABILITY handshake as required by RFCs.
1875 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1877 *) Introduce limits to prevent malicious keys being able to
1878 cause a denial of service. (CVE-2006-2940)
1879 [Steve Henson, Bodo Moeller]
1881 *) Fix ASN.1 parsing of certain invalid structures that can result
1882 in a denial of service. (CVE-2006-2937) [Steve Henson]
1884 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1885 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1887 *) Fix SSL client code which could crash if connecting to a
1888 malicious SSLv2 server. (CVE-2006-4343)
1889 [Tavis Ormandy and Will Drewry, Google Security Team]
1891 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1892 match only those. Before that, "AES256-SHA" would be interpreted
1893 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1894 the same strength classification in 0.9.7h) as we currently only
1895 have a single AES bit in the ciphersuite description bitmap.
1896 That change, however, also applied to ciphersuite strings such as
1897 "RC4-MD5" that intentionally matched multiple ciphersuites --
1898 namely, SSL 2.0 ciphersuites in addition to the more common ones
1899 from SSL 3.0/TLS 1.0.
1901 So we change the selection algorithm again: Naming an explicit
1902 ciphersuite selects this one ciphersuite, and any other similar
1903 ciphersuite (same bitmap) from *other* protocol versions.
1904 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1905 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1907 Since SSL 2.0 does not have any ciphersuites for which the
1908 128/256 bit distinction would be relevant, this works for now.
1909 The proper fix will be to use different bits for AES128 and
1910 AES256, which would have avoided the problems from the beginning;
1911 however, bits are scarce, so we can only do this in a new release
1912 (not just a patchlevel) when we can change the SSL_CIPHER
1913 definition to split the single 'unsigned long mask' bitmap into
1914 multiple values to extend the available space.
1918 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1920 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1921 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1923 *) Add AES IGE and biIGE modes.
1926 *) Change the Unix randomness entropy gathering to use poll() when
1927 possible instead of select(), since the latter has some
1928 undesirable limitations.
1929 [Darryl Miles via Richard Levitte and Bodo Moeller]
1931 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1932 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1933 cannot be implicitly activated as part of, e.g., the "AES" alias.
1934 However, please upgrade to OpenSSL 0.9.9[-dev] for
1935 non-experimental use of the ECC ciphersuites to get TLS extension
1936 support, which is required for curve and point format negotiation
1937 to avoid potential handshake problems.
1940 *) Disable rogue ciphersuites:
1942 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1943 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1944 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1946 The latter two were purportedly from
1947 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1950 Also deactivate the remaining ciphersuites from
1951 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1952 unofficial, and the ID has long expired.
1955 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1956 dual-core machines) and other potential thread-safety issues.
1959 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1960 versions), which is now available for royalty-free use
1961 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1962 Also, add Camellia TLS ciphersuites from RFC 4132.
1964 To minimize changes between patchlevels in the OpenSSL 0.9.8
1965 series, Camellia remains excluded from compilation unless OpenSSL
1966 is configured with 'enable-camellia'.
1969 *) Disable the padding bug check when compression is in use. The padding
1970 bug check assumes the first packet is of even length, this is not
1971 necessarily true if compresssion is enabled and can result in false
1972 positives causing handshake failure. The actual bug test is ancient
1973 code so it is hoped that implementations will either have fixed it by
1974 now or any which still have the bug do not support compression.
1977 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1979 *) When applying a cipher rule check to see if string match is an explicit
1980 cipher suite and only match that one cipher suite if it is.
1983 *) Link in manifests for VC++ if needed.
1984 [Austin Ziegler <halostatue@gmail.com>]
1986 *) Update support for ECC-based TLS ciphersuites according to
1987 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1988 TLS extensions, which are supported starting with the 0.9.9
1989 branch, not in the OpenSSL 0.9.8 branch).
1992 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1993 opaque EVP_CIPHER_CTX handling.
1996 *) Fixes and enhancements to zlib compression code. We now only use
1997 "zlib1.dll" and use the default __cdecl calling convention on Win32
1998 to conform with the standards mentioned here:
1999 http://www.zlib.net/DLL_FAQ.txt
2000 Static zlib linking now works on Windows and the new --with-zlib-include
2001 --with-zlib-lib options to Configure can be used to supply the location
2002 of the headers and library. Gracefully handle case where zlib library
2006 *) Several fixes and enhancements to the OID generation code. The old code
2007 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2008 handle numbers larger than ULONG_MAX, truncated printing and had a
2009 non standard OBJ_obj2txt() behaviour.
2012 *) Add support for building of engines under engine/ as shared libraries
2013 under VC++ build system.
2016 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2017 Hopefully, we will not see any false combination of paths any more.
2020 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2022 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2023 (part of SSL_OP_ALL). This option used to disable the
2024 countermeasure against man-in-the-middle protocol-version
2025 rollback in the SSL 2.0 server implementation, which is a bad
2026 idea. (CVE-2005-2969)
2028 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2029 for Information Security, National Institute of Advanced Industrial
2030 Science and Technology [AIST], Japan)]
2032 *) Add two function to clear and return the verify parameter flags.
2035 *) Keep cipherlists sorted in the source instead of sorting them at
2036 runtime, thus removing the need for a lock.
2039 *) Avoid some small subgroup attacks in Diffie-Hellman.
2040 [Nick Mathewson and Ben Laurie]
2042 *) Add functions for well-known primes.
2045 *) Extended Windows CE support.
2046 [Satoshi Nakamura and Andy Polyakov]
2048 *) Initialize SSL_METHOD structures at compile time instead of during
2049 runtime, thus removing the need for a lock.
2052 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2053 attempting to decrypt each encrypted key in turn. Add support to
2057 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2059 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2062 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2065 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2066 key into the same file any more.
2069 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2072 *) Add -utf8 command line and config file option to 'ca'.
2073 [Stefan <stf@udoma.org]
2075 *) Removed the macro des_crypt(), as it seems to conflict with some
2076 libraries. Use DES_crypt().
2079 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2080 involves renaming the source and generated shared-libs for
2081 both. The engines will accept the corrected or legacy ids
2082 ('ncipher' and '4758_cca' respectively) when binding. NB,
2083 this only applies when building 'shared'.
2084 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2086 *) Add attribute functions to EVP_PKEY structure. Modify
2087 PKCS12_create() to recognize a CSP name attribute and
2088 use it. Make -CSP option work again in pkcs12 utility.
2091 *) Add new functionality to the bn blinding code:
2092 - automatic re-creation of the BN_BLINDING parameters after
2093 a fixed number of uses (currently 32)
2094 - add new function for parameter creation
2095 - introduce flags to control the update behaviour of the
2096 BN_BLINDING parameters
2097 - hide BN_BLINDING structure
2098 Add a second BN_BLINDING slot to the RSA structure to improve
2099 performance when a single RSA object is shared among several
2103 *) Add support for DTLS.
2104 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2106 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2107 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2110 *) Remove buggy and incompletet DH cert support from
2111 ssl/ssl_rsa.c and ssl/s3_both.c
2114 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2115 the apps/openssl applications.
2118 *) Compile clean with "-Wall -Wmissing-prototypes
2119 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2120 DEBUG_SAFESTACK must also be set.
2123 *) Change ./Configure so that certain algorithms can be disabled by default.
2124 The new counterpiece to "no-xxx" is "enable-xxx".
2126 The patented RC5 and MDC2 algorithms will now be disabled unless
2127 "enable-rc5" and "enable-mdc2", respectively, are specified.
2129 (IDEA remains enabled despite being patented. This is because IDEA
2130 is frequently required for interoperability, and there is no license
2131 fee for non-commercial use. As before, "no-idea" can be used to
2132 avoid this algorithm.)
2136 *) Add processing of proxy certificates (see RFC 3820). This work was
2137 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2138 EGEE (Enabling Grids for E-science in Europe).
2141 *) RC4 performance overhaul on modern architectures/implementations, such
2142 as Intel P4, IA-64 and AMD64.
2145 *) New utility extract-section.pl. This can be used specify an alternative
2146 section number in a pod file instead of having to treat each file as
2147 a separate case in Makefile. This can be done by adding two lines to the
2150 =for comment openssl_section:XXX
2152 The blank line is mandatory.
2156 *) New arguments -certform, -keyform and -pass for s_client and s_server
2157 to allow alternative format key and certificate files and passphrase
2161 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2162 update associated structures and add various utility functions.
2164 Add new policy related verify parameters, include policy checking in
2165 standard verify code. Enhance 'smime' application with extra parameters
2166 to support policy checking and print out.
2169 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2170 Nehemiah processors. These extensions support AES encryption in hardware
2171 as well as RNG (though RNG support is currently disabled).
2172 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2174 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2177 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2178 [Andy Polyakov and a number of other people]
2180 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2181 implementation contributed by IBM.
2182 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2184 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2185 exponent rather than 'unsigned long'. There is a corresponding change to
2186 the new 'rsa_keygen' element of the RSA_METHOD structure.
2187 [Jelte Jansen, Geoff Thorpe]
2189 *) Functionality for creating the initial serial number file is now
2190 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2192 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2193 number file to 1, which is bound to cause problems. To avoid
2194 the problems while respecting compatibility between different 0.9.7
2195 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2196 CA.pl for serial number initialization. With the new release 0.9.8,
2197 we can fix the problem directly in the 'ca' utility.)
2200 *) Reduced header interdepencies by declaring more opaque objects in
2201 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2202 give fewer recursive includes, which could break lazy source code - so
2203 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2204 developers should define this symbol when building and using openssl to
2205 ensure they track the recommended behaviour, interfaces, [etc], but
2206 backwards-compatible behaviour prevails when this isn't defined.
2209 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2212 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2213 This will generate a random key of the appropriate length based on the
2214 cipher context. The EVP_CIPHER can provide its own random key generation
2215 routine to support keys of a specific form. This is used in the des and
2216 3des routines to generate a key of the correct parity. Update S/MIME
2217 code to use new functions and hence generate correct parity DES keys.
2218 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2219 valid (weak or incorrect parity).
2222 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2223 as looking them up. This is useful when the verified structure may contain
2224 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2225 present unless the new PKCS7_NO_CRL flag is asserted.
2228 *) Extend ASN1 oid configuration module. It now additionally accepts the
2231 shortName = some long name, 1.2.3.4
2234 *) Reimplemented the BN_CTX implementation. There is now no more static
2235 limitation on the number of variables it can handle nor the depth of the
2236 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2237 information can now expand as required, and rather than having a single
2238 static array of bignums, BN_CTX now uses a linked-list of such arrays
2239 allowing it to expand on demand whilst maintaining the usefulness of
2240 BN_CTX's "bundling".
2243 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2244 to allow all RSA operations to function using a single BN_CTX.
2247 *) Preliminary support for certificate policy evaluation and checking. This
2248 is initially intended to pass the tests outlined in "Conformance Testing
2249 of Relying Party Client Certificate Path Processing Logic" v1.07.
2252 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2253 remained unused and not that useful. A variety of other little bignum
2254 tweaks and fixes have also been made continuing on from the audit (see
2258 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2259 associated ASN1, EVP and SSL functions and old ASN1 macros.
2262 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2263 and this should never fail. So the return value from the use of
2264 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2265 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2268 *) BN_CTX_get() should return zero-valued bignums, providing the same
2269 initialised value as BN_new().
2270 [Geoff Thorpe, suggested by Ulf Möller]
2272 *) Support for inhibitAnyPolicy certificate extension.
2275 *) An audit of the BIGNUM code is underway, for which debugging code is
2276 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2277 is considered valid when processing BIGNUMs, and causes execution to
2278 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2279 further steps are taken to deliberately pollute unused data in BIGNUM
2280 structures to try and expose faulty code further on. For now, openssl will
2281 (in its default mode of operation) continue to tolerate the inconsistent
2282 forms that it has tolerated in the past, but authors and packagers should
2283 consider trying openssl and their own applications when compiled with
2284 these debugging symbols defined. It will help highlight potential bugs in
2285 their own code, and will improve the test coverage for OpenSSL itself. At
2286 some point, these tighter rules will become openssl's default to improve
2287 maintainability, though the assert()s and other overheads will remain only
2288 in debugging configurations. See bn.h for more details.
2289 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2291 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2292 that can only be obtained through BN_CTX_new() (which implicitly
2293 initialises it). The presence of this function only made it possible
2294 to overwrite an existing structure (and cause memory leaks).
2297 *) Because of the callback-based approach for implementing LHASH as a
2298 template type, lh_insert() adds opaque objects to hash-tables and
2299 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2300 to clean up those corresponding objects before destroying the hash table
2301 (and losing the object pointers). So some over-zealous constifications in
2302 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2303 objects as "const" and the lh_doall[_arg] callback wrappers are not
2304 prototyped to have "const" restrictions on the object pointers they are
2305 given (and so aren't required to cast them away any more).
2308 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2309 (speed) prefers to use its own implementation. The two implementations
2310 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2311 its object type properly exposed (MS_TM) instead of casting to/from "char
2312 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2313 aren't necessarily the greatest nomenclatures - but this is what was used
2314 internally to the implementation so I've used that for now.
2317 *) Ensure that deprecated functions do not get compiled when
2318 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2319 the self-tests were still using deprecated key-generation functions so
2320 these have been updated also.
2323 *) Reorganise PKCS#7 code to separate the digest location functionality
2324 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2325 New function PKCS7_set_digest() to set the digest type for PKCS#7
2326 digestedData type. Add additional code to correctly generate the
2327 digestedData type and add support for this type in PKCS7 initialization
2331 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2332 structure of type "other".
2335 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2336 sure the loop does correctly stop and breaking ("division by zero")
2337 modulus operations are not performed. The (pre-generated) prime
2338 table crypto/bn/bn_prime.h was already correct, but it could not be
2339 re-generated on some platforms because of the "division by zero"
2340 situation in the script.
2341 [Ralf S. Engelschall]
2343 *) Update support for ECC-based TLS ciphersuites according to
2344 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2345 SHA-1 now is only used for "small" curves (where the
2346 representation of a field element takes up to 24 bytes); for
2347 larger curves, the field element resulting from ECDH is directly
2348 used as premaster secret.
2349 [Douglas Stebila (Sun Microsystems Laboratories)]
2351 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2352 curve secp160r1 to the tests.
2353 [Douglas Stebila (Sun Microsystems Laboratories)]
2355 *) Add the possibility to load symbols globally with DSO.
2356 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2358 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2359 control of the error stack.
2362 *) Add support for STORE in ENGINE.
2365 *) Add the STORE type. The intention is to provide a common interface
2366 to certificate and key stores, be they simple file-based stores, or
2367 HSM-type store, or LDAP stores, or...
2368 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2371 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2372 pass a list of arguments to any function as well as provide a way
2373 for a function to pass data back to the caller.
2376 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2377 works like BUF_strdup() but can be used to duplicate a portion of
2378 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2382 *) Add the function sk_find_ex() which works like sk_find(), but will
2383 return an index to an element even if an exact match couldn't be
2384 found. The index is guaranteed to point at the element where the
2385 searched-for key would be inserted to preserve sorting order.
2388 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2389 takes an extra flags argument for optional functionality. Currently,
2390 the following flags are defined:
2392 OBJ_BSEARCH_VALUE_ON_NOMATCH
2393 This one gets OBJ_bsearch_ex() to return a pointer to the first
2394 element where the comparing function returns a negative or zero
2397 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2398 This one gets OBJ_bsearch_ex() to return a pointer to the first
2399 element where the comparing function returns zero. This is useful
2400 if there are more than one element where the comparing function
2404 *) Make it possible to create self-signed certificates with 'openssl ca'
2405 in such a way that the self-signed certificate becomes part of the
2406 CA database and uses the same mechanisms for serial number generation
2407 as all other certificate signing. The new flag '-selfsign' enables
2408 this functionality. Adapt CA.sh and CA.pl.in.
2411 *) Add functionality to check the public key of a certificate request
2412 against a given private. This is useful to check that a certificate
2413 request can be signed by that key (self-signing).
2416 *) Make it possible to have multiple active certificates with the same
2417 subject in the CA index file. This is done only if the keyword
2418 'unique_subject' is set to 'no' in the main CA section (default
2419 if 'CA_default') of the configuration file. The value is saved
2420 with the database itself in a separate index attribute file,
2421 named like the index file with '.attr' appended to the name.
2424 *) Generate muti valued AVAs using '+' notation in config files for
2428 *) Support for nameConstraints certificate extension.
2431 *) Support for policyConstraints certificate extension.
2434 *) Support for policyMappings certificate extension.
2437 *) Make sure the default DSA_METHOD implementation only uses its
2438 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2439 and change its own handlers to be NULL so as to remove unnecessary
2440 indirection. This lets alternative implementations fallback to the
2441 default implementation more easily.
2444 *) Support for directoryName in GeneralName related extensions
2448 *) Make it possible to link applications using Makefile.shared.
2449 Make that possible even when linking against static libraries!
2452 *) Support for single pass processing for S/MIME signing. This now
2453 means that S/MIME signing can be done from a pipe, in addition
2454 cleartext signing (multipart/signed type) is effectively streaming
2455 and the signed data does not need to be all held in memory.
2457 This is done with a new flag PKCS7_STREAM. When this flag is set
2458 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2459 is done after the data is output (and digests calculated) in
2460 SMIME_write_PKCS7().
2463 *) Add full support for -rpath/-R, both in shared libraries and
2464 applications, at least on the platforms where it's known how
2468 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2469 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2470 will now compute a table of multiples of the generator that
2471 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2472 faster (notably in the case of a single point multiplication,
2473 scalar * generator).
2474 [Nils Larsch, Bodo Moeller]
2476 *) IPv6 support for certificate extensions. The various extensions
2477 which use the IP:a.b.c.d can now take IPv6 addresses using the
2478 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2482 *) Added an ENGINE that implements RSA by performing private key
2483 exponentiations with the GMP library. The conversions to and from
2484 GMP's mpz_t format aren't optimised nor are any montgomery forms
2485 cached, and on x86 it appears OpenSSL's own performance has caught up.
2486 However there are likely to be other architectures where GMP could
2487 provide a boost. This ENGINE is not built in by default, but it can be
2488 specified at Configure time and should be accompanied by the necessary
2489 linker additions, eg;
2490 ./config -DOPENSSL_USE_GMP -lgmp
2493 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2494 testing availability of engines with "-t" - the old behaviour is
2495 produced by increasing the feature's verbosity with "-tt".
2498 *) ECDSA routines: under certain error conditions uninitialized BN objects
2499 could be freed. Solution: make sure initialization is performed early
2500 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2504 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2505 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2506 software implementations. For DSA and DH, parameter generation can
2507 also be overriden by providing the appropriate method callbacks.
2510 *) Change the "progress" mechanism used in key-generation and
2511 primality testing to functions that take a new BN_GENCB pointer in
2512 place of callback/argument pairs. The new API functions have "_ex"
2513 postfixes and the older functions are reimplemented as wrappers for
2514 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2515 declarations of the old functions to help (graceful) attempts to
2516 migrate to the new functions. Also, the new key-generation API
2517 functions operate on a caller-supplied key-structure and return
2518 success/failure rather than returning a key or NULL - this is to
2519 help make "keygen" another member function of RSA_METHOD etc.
2521 Example for using the new callback interface:
2523 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2527 BN_GENCB_set(&my_cb, my_callback, my_arg);
2529 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2530 /* For the meaning of a, b in calls to my_callback(), see the
2531 * documentation of the function that calls the callback.
2532 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2533 * my_callback should return 1 if it wants BN_is_prime_ex()
2534 * to continue, or 0 to stop.
2539 *) Change the ZLIB compression method to be stateful, and make it
2540 available to TLS with the number defined in
2541 draft-ietf-tls-compression-04.txt.
2544 *) Add the ASN.1 structures and functions for CertificatePair, which
2545 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2547 CertificatePair ::= SEQUENCE {
2548 forward [0] Certificate OPTIONAL,
2549 reverse [1] Certificate OPTIONAL,
2550 -- at least one of the pair shall be present -- }
2552 Also implement the PEM functions to read and write certificate
2553 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2555 This needed to be defined, mostly for the sake of the LDAP
2556 attribute crossCertificatePair, but may prove useful elsewhere as
2560 *) Make it possible to inhibit symlinking of shared libraries in
2561 Makefile.shared, for Cygwin's sake.
2564 *) Extend the BIGNUM API by creating a function
2565 void BN_set_negative(BIGNUM *a, int neg);
2566 and a macro that behave like
2567 int BN_is_negative(const BIGNUM *a);
2569 to avoid the need to access 'a->neg' directly in applications.
2572 *) Implement fast modular reduction for pseudo-Mersenne primes
2573 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2574 EC_GROUP_new_curve_GFp() will now automatically use this
2576 [Nils Larsch <nla@trustcenter.de>]
2578 *) Add new lock type (CRYPTO_LOCK_BN).
2581 *) Change the ENGINE framework to automatically load engines
2582 dynamically from specific directories unless they could be
2583 found to already be built in or loaded. Move all the
2584 current engines except for the cryptodev one to a new
2586 The engines in engines/ are built as shared libraries if
2587 the "shared" options was given to ./Configure or ./config.
2588 Otherwise, they are inserted in libcrypto.a.
2589 /usr/local/ssl/engines is the default directory for dynamic
2590 engines, but that can be overriden at configure time through
2591 the usual use of --prefix and/or --openssldir, and at run
2592 time with the environment variable OPENSSL_ENGINES.
2593 [Geoff Thorpe and Richard Levitte]
2595 *) Add Makefile.shared, a helper makefile to build shared
2596 libraries. Addapt Makefile.org.
2599 *) Add version info to Win32 DLLs.
2600 [Peter 'Luna' Runestig" <peter@runestig.com>]
2602 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2603 can be added using this API to created arbitrary PKCS#12
2604 files while avoiding the low level API.
2606 New options to PKCS12_create(), key or cert can be NULL and
2607 will then be omitted from the output file. The encryption
2608 algorithm NIDs can be set to -1 for no encryption, the mac
2609 iteration count can be set to 0 to omit the mac.
2611 Enhance pkcs12 utility by making the -nokeys and -nocerts
2612 options work when creating a PKCS#12 file. New option -nomac
2613 to omit the mac, NONE can be set for an encryption algorithm.
2614 New code is modified to use the enhanced PKCS12_create()
2615 instead of the low level API.
2618 *) Extend ASN1 encoder to support indefinite length constructed
2619 encoding. This can output sequences tags and octet strings in
2620 this form. Modify pk7_asn1.c to support indefinite length
2621 encoding. This is experimental and needs additional code to
2622 be useful, such as an ASN1 bio and some enhanced streaming
2625 Extend template encode functionality so that tagging is passed
2626 down to the template encoder.
2629 *) Let 'openssl req' fail if an argument to '-newkey' is not
2630 recognized instead of using RSA as a default.
2633 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2634 As these are not official, they are not included in "ALL";
2635 the "ECCdraft" ciphersuite group alias can be used to select them.
2636 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2638 *) Add ECDH engine support.
2639 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2641 *) Add ECDH in new directory crypto/ecdh/.
2642 [Douglas Stebila (Sun Microsystems Laboratories)]
2644 *) Let BN_rand_range() abort with an error after 100 iterations
2645 without success (which indicates a broken PRNG).
2648 *) Change BN_mod_sqrt() so that it verifies that the input value
2649 is really the square of the return value. (Previously,
2650 BN_mod_sqrt would show GIGO behaviour.)
2653 *) Add named elliptic curves over binary fields from X9.62, SECG,
2654 and WAP/WTLS; add OIDs that were still missing.
2656 [Sheueling Chang Shantz and Douglas Stebila
2657 (Sun Microsystems Laboratories)]
2659 *) Extend the EC library for elliptic curves over binary fields
2660 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2663 EC_GF2m_simple_method
2667 EC_GROUP_new_curve_GF2m
2668 EC_GROUP_set_curve_GF2m
2669 EC_GROUP_get_curve_GF2m
2670 EC_POINT_set_affine_coordinates_GF2m
2671 EC_POINT_get_affine_coordinates_GF2m
2672 EC_POINT_set_compressed_coordinates_GF2m
2674 Point compression for binary fields is disabled by default for
2675 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2678 As binary polynomials are represented as BIGNUMs, various members
2679 of the EC_GROUP and EC_POINT data structures can be shared
2680 between the implementations for prime fields and binary fields;
2681 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2682 are essentially identical to their ..._GFp counterparts.
2683 (For simplicity, the '..._GFp' prefix has been dropped from
2684 various internal method names.)
2686 An internal 'field_div' method (similar to 'field_mul' and
2687 'field_sqr') has been added; this is used only for binary fields.
2689 [Sheueling Chang Shantz and Douglas Stebila
2690 (Sun Microsystems Laboratories)]
2692 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2693 through methods ('mul', 'precompute_mult').
2695 The generic implementations (now internally called 'ec_wNAF_mul'
2696 and 'ec_wNAF_precomputed_mult') remain the default if these
2697 methods are undefined.
2699 [Sheueling Chang Shantz and Douglas Stebila
2700 (Sun Microsystems Laboratories)]
2702 *) New function EC_GROUP_get_degree, which is defined through
2703 EC_METHOD. For curves over prime fields, this returns the bit
2704 length of the modulus.
2706 [Sheueling Chang Shantz and Douglas Stebila
2707 (Sun Microsystems Laboratories)]
2709 *) New functions EC_GROUP_dup, EC_POINT_dup.
2710 (These simply call ..._new and ..._copy).
2712 [Sheueling Chang Shantz and Douglas Stebila
2713 (Sun Microsystems Laboratories)]
2715 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2716 Polynomials are represented as BIGNUMs (where the sign bit is not
2717 used) in the following functions [macros]:
2720 BN_GF2m_sub [= BN_GF2m_add]
2721 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2722 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2723 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2725 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2726 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2727 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2728 BN_GF2m_cmp [= BN_ucmp]
2730 (Note that only the 'mod' functions are actually for fields GF(2^m).
2731 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2733 For some functions, an the irreducible polynomial defining a
2734 field can be given as an 'unsigned int[]' with strictly
2735 decreasing elements giving the indices of those bits that are set;
2736 i.e., p[] represents the polynomial
2737 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2739 p[0] > p[1] > ... > p[k] = 0.
2740 This applies to the following functions:
2745 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2746 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2748 BN_GF2m_mod_sqrt_arr
2749 BN_GF2m_mod_solve_quad_arr
2753 Conversion can be performed by the following functions:
2758 bntest.c has additional tests for binary polynomial arithmetic.
2760 Two implementations for BN_GF2m_mod_div() are available.
2761 The default algorithm simply uses BN_GF2m_mod_inv() and
2762 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2763 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2764 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2766 [Sheueling Chang Shantz and Douglas Stebila
2767 (Sun Microsystems Laboratories)]
2769 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2770 functionality is disabled at compile-time.
2771 [Douglas Stebila <douglas.stebila@sun.com>]
2773 *) Change default behaviour of 'openssl asn1parse' so that more
2774 information is visible when viewing, e.g., a certificate:
2776 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2777 mode the content of non-printable OCTET STRINGs is output in a
2778 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2779 avoid the appearance of a printable string.
2780 [Nils Larsch <nla@trustcenter.de>]
2782 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2784 EC_GROUP_set_asn1_flag()
2785 EC_GROUP_get_asn1_flag()
2786 EC_GROUP_set_point_conversion_form()
2787 EC_GROUP_get_point_conversion_form()
2788 These control ASN1 encoding details:
2789 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2790 has been set to OPENSSL_EC_NAMED_CURVE.
2791 - Points are encoded in uncompressed form by default; options for
2792 asn1_for are as for point2oct, namely
2793 POINT_CONVERSION_COMPRESSED
2794 POINT_CONVERSION_UNCOMPRESSED
2795 POINT_CONVERSION_HYBRID
2797 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2800 EC_GROUP_get0_seed()
2801 EC_GROUP_get_seed_len()
2802 This is used only for ASN1 purposes (so far).
2803 [Nils Larsch <nla@trustcenter.de>]
2805 *) Add 'field_type' member to EC_METHOD, which holds the NID
2806 of the appropriate field type OID. The new function
2807 EC_METHOD_get_field_type() returns this value.
2808 [Nils Larsch <nla@trustcenter.de>]
2813 EC_POINT_point2hex()
2814 EC_POINT_hex2point()
2815 providing useful interfaces to EC_POINT_point2oct() and
2816 EC_POINT_oct2point().
2817 [Nils Larsch <nla@trustcenter.de>]
2819 *) Change internals of the EC library so that the functions
2820 EC_GROUP_set_generator()
2821 EC_GROUP_get_generator()
2822 EC_GROUP_get_order()
2823 EC_GROUP_get_cofactor()
2824 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2825 to methods, which would lead to unnecessary code duplication when
2826 adding different types of curves.
2827 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2829 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2830 arithmetic, and such that modified wNAFs are generated
2831 (which avoid length expansion in many cases).
2834 *) Add a function EC_GROUP_check_discriminant() (defined via
2835 EC_METHOD) that verifies that the curve discriminant is non-zero.
2837 Add a function EC_GROUP_check() that makes some sanity tests
2838 on a EC_GROUP, its generator and order. This includes
2839 EC_GROUP_check_discriminant().
2840 [Nils Larsch <nla@trustcenter.de>]
2842 *) Add ECDSA in new directory crypto/ecdsa/.
2844 Add applications 'openssl ecparam' and 'openssl ecdsa'
2845 (these are based on 'openssl dsaparam' and 'openssl dsa').
2847 ECDSA support is also included in various other files across the
2848 library. Most notably,
2849 - 'openssl req' now has a '-newkey ecdsa:file' option;
2850 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2851 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2852 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2853 them suitable for ECDSA where domain parameters must be
2854 extracted before the specific public key;
2855 - ECDSA engine support has been added.
2856 [Nils Larsch <nla@trustcenter.de>]
2858 *) Include some named elliptic curves, and add OIDs from X9.62,
2859 SECG, and WAP/WTLS. Each curve can be obtained from the new
2861 EC_GROUP_new_by_curve_name(),
2862 and the list of available named curves can be obtained with
2863 EC_get_builtin_curves().
2864 Also add a 'curve_name' member to EC_GROUP objects, which can be
2866 EC_GROUP_set_curve_name()
2867 EC_GROUP_get_curve_name()
2868 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2870 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2871 was actually never needed) and in BN_mul(). The removal in BN_mul()
2872 required a small change in bn_mul_part_recursive() and the addition
2873 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2874 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2875 bn_sub_words() and bn_add_words() except they take arrays with
2879 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2881 *) Cleanse PEM buffers before freeing them since they may contain
2883 [Benjamin Bennett <ben@psc.edu>]
2885 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2886 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2887 authentication-only ciphersuites.
2890 *) Since AES128 and AES256 share a single mask bit in the logic of
2891 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2892 kludge to work properly if AES128 is available and AES256 isn't.
2895 *) Expand security boundary to match 1.1.1 module.
2898 *) Remove redundant features: hash file source, editing of test vectors
2899 modify fipsld to use external fips_premain.c signature.
2902 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2903 run algorithm test programs.
2906 *) Make algorithm test programs more tolerant of whitespace.
2909 *) Have SSL/TLS server implementation tolerate "mismatched" record
2910 protocol version while receiving ClientHello even if the
2911 ClientHello is fragmented. (The server can't insist on the
2912 particular protocol version it has chosen before the ServerHello
2913 message has informed the client about his choice.)
2916 *) Load error codes if they are not already present instead of using a
2917 static variable. This allows them to be cleanly unloaded and reloaded.
2920 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2922 *) Introduce limits to prevent malicious keys being able to
2923 cause a denial of service. (CVE-2006-2940)
2924 [Steve Henson, Bodo Moeller]
2926 *) Fix ASN.1 parsing of certain invalid structures that can result
2927 in a denial of service. (CVE-2006-2937) [Steve Henson]
2929 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2930 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2932 *) Fix SSL client code which could crash if connecting to a
2933 malicious SSLv2 server. (CVE-2006-4343)
2934 [Tavis Ormandy and Will Drewry, Google Security Team]
2936 *) Change ciphersuite string processing so that an explicit
2937 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2938 will no longer include "AES128-SHA"), and any other similar
2939 ciphersuite (same bitmap) from *other* protocol versions (so that
2940 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2941 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2942 changes from 0.9.8b and 0.9.8d.
2945 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2947 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2948 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2950 *) Change the Unix randomness entropy gathering to use poll() when
2951 possible instead of select(), since the latter has some
2952 undesirable limitations.
2953 [Darryl Miles via Richard Levitte and Bodo Moeller]
2955 *) Disable rogue ciphersuites:
2957 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2958 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2959 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2961 The latter two were purportedly from
2962 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2965 Also deactive the remaining ciphersuites from
2966 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2967 unofficial, and the ID has long expired.
2970 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2971 dual-core machines) and other potential thread-safety issues.
2974 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2976 *) Adapt fipsld and the build system to link against the validated FIPS
2977 module in FIPS mode.
2980 *) Fixes for VC++ 2005 build under Windows.
2983 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2984 from a Windows bash shell such as MSYS. It is autodetected from the
2985 "config" script when run from a VC++ environment. Modify standard VC++
2986 build to use fipscanister.o from the GNU make build.
2989 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2991 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2992 The value now differs depending on if you build for FIPS or not.
2993 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2994 safely run with a non-FIPSed libcrypto, as it may crash because of
2995 the difference induced by this change.
2998 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3000 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3001 (part of SSL_OP_ALL). This option used to disable the
3002 countermeasure against man-in-the-middle protocol-version
3003 rollback in the SSL 2.0 server implementation, which is a bad
3004 idea. (CVE-2005-2969)
3006 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3007 for Information Security, National Institute of Advanced Industrial
3008 Science and Technology [AIST], Japan)]
3010 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3011 mainly for FIPS compliance and not fully integrated at this stage.
3014 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3015 the exponentiation using a fixed-length exponent. (Otherwise,
3016 the information leaked through timing could expose the secret key
3017 after many signatures; cf. Bleichenbacher's attack on DSA with
3021 *) Make a new fixed-window mod_exp implementation the default for
3022 RSA, DSA, and DH private-key operations so that the sequence of
3023 squares and multiplies and the memory access pattern are
3024 independent of the particular secret key. This will mitigate
3025 cache-timing and potential related attacks.
3027 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3028 and this is automatically used by BN_mod_exp_mont() if the new flag
3029 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3030 will use this BN flag for private exponents unless the flag
3031 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3032 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3034 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3036 *) Change the client implementation for SSLv23_method() and
3037 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3038 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3039 (Previously, the SSL 2.0 backwards compatible Client Hello
3040 message format would be used even with SSL_OP_NO_SSLv2.)
3043 *) Add support for smime-type MIME parameter in S/MIME messages which some
3047 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3048 a threadsafe manner. Modify rsa code to use new function and add calls
3049 to dsa and dh code (which had race conditions before).
3052 *) Include the fixed error library code in the C error file definitions
3053 instead of fixing them up at runtime. This keeps the error code
3054 structures constant.
3057 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3059 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3062 *) Fixes for newer kerberos headers. NB: the casts are needed because
3063 the 'length' field is signed on one version and unsigned on another
3064 with no (?) obvious way to tell the difference, without these VC++
3065 complains. Also the "definition" of FAR (blank) is no longer included
3066 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3067 some needed definitions.
3070 *) Undo Cygwin change.
3073 *) Added support for proxy certificates according to RFC 3820.
3074 Because they may be a security thread to unaware applications,
3075 they must be explicitely allowed in run-time. See
3076 docs/HOWTO/proxy_certificates.txt for further information.
3079 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3081 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3082 server and client random values. Previously
3083 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3084 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3086 This change has negligible security impact because:
3088 1. Server and client random values still have 24 bytes of pseudo random
3091 2. Server and client random values are sent in the clear in the initial
3094 3. The master secret is derived using the premaster secret (48 bytes in
3095 size for static RSA ciphersuites) as well as client server and random
3098 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3101 [Stephen Henson, reported by UK NISCC]
3103 *) Use Windows randomness collection on Cygwin.
3106 *) Fix hang in EGD/PRNGD query when communication socket is closed
3107 prematurely by EGD/PRNGD.
3108 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3110 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3113 *) Back-port of selected performance improvements from development
3114 branch, as well as improved support for PowerPC platforms.
3117 *) Add lots of checks for memory allocation failure, error codes to indicate
3118 failure and freeing up memory if a failure occurs.
3119 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3121 *) Add new -passin argument to dgst.
3124 *) Perform some character comparisons of different types in X509_NAME_cmp:
3125 this is needed for some certificates that reencode DNs into UTF8Strings
3126 (in violation of RFC3280) and can't or wont issue name rollover
3130 *) Make an explicit check during certificate validation to see that
3131 the CA setting in each certificate on the chain is correct. As a
3132 side effect always do the following basic checks on extensions,
3133 not just when there's an associated purpose to the check:
3135 - if there is an unhandled critical extension (unless the user
3136 has chosen to ignore this fault)
3137 - if the path length has been exceeded (if one is set at all)
3138 - that certain extensions fit the associated purpose (if one has
3142 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3144 *) Avoid a race condition when CRLs are checked in a multi threaded
3145 environment. This would happen due to the reordering of the revoked
3146 entries during signature checking and serial number lookup. Now the
3147 encoding is cached and the serial number sort performed under a lock.
3148 Add new STACK function sk_is_sorted().
3151 *) Add Delta CRL to the extension code.
3154 *) Various fixes to s3_pkt.c so alerts are sent properly.
3155 [David Holmes <d.holmes@f5.com>]
3157 *) Reduce the chances of duplicate issuer name and serial numbers (in
3158 violation of RFC3280) using the OpenSSL certificate creation utilities.
3159 This is done by creating a random 64 bit value for the initial serial
3160 number when a serial number file is created or when a self signed
3161 certificate is created using 'openssl req -x509'. The initial serial
3162 number file is created using 'openssl x509 -next_serial' in CA.pl
3163 rather than being initialized to 1.
3166 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3168 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3169 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3170 [Joe Orton, Steve Henson]
3172 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3174 [Joe Orton, Steve Henson]
3176 *) Make it possible to have multiple active certificates with the same
3177 subject in the CA index file. This is done only if the keyword
3178 'unique_subject' is set to 'no' in the main CA section (default
3179 if 'CA_default') of the configuration file. The value is saved
3180 with the database itself in a separate index attribute file,
3181 named like the index file with '.attr' appended to the name.
3184 *) X509 verify fixes. Disable broken certificate workarounds when
3185 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3186 keyUsage extension present. Don't accept CRLs with unhandled critical
3187 extensions: since verify currently doesn't process CRL extensions this
3188 rejects a CRL with *any* critical extensions. Add new verify error codes
3192 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3193 A clarification of RFC2560 will require the use of OCTET STRINGs and
3194 some implementations cannot handle the current raw format. Since OpenSSL
3195 copies and compares OCSP nonces as opaque blobs without any attempt at
3196 parsing them this should not create any compatibility issues.
3199 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3200 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3201 this HMAC (and other) operations are several times slower than OpenSSL
3205 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3206 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3208 *) Use the correct content when signing type "other".
3211 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3213 *) Fix various bugs revealed by running the NISCC test suite:
3215 Stop out of bounds reads in the ASN1 code when presented with
3216 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3218 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3220 If verify callback ignores invalid public key errors don't try to check
3221 certificate signature with the NULL public key.
3225 *) New -ignore_err option in ocsp application to stop the server
3226 exiting on the first error in a request.
3229 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3230 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3234 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3235 extra data after the compression methods not only for TLS 1.0
3236 but also for SSL 3.0 (as required by the specification).
3237 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3239 *) Change X509_certificate_type() to mark the key as exported/exportable
3240 when it's 512 *bits* long, not 512 bytes.
3243 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3244 blocks during encryption.
3247 *) Various fixes to base64 BIO and non blocking I/O. On write
3248 flushes were not handled properly if the BIO retried. On read
3249 data was not being buffered properly and had various logic bugs.
3250 This also affects blocking I/O when the data being decoded is a
3254 *) Various S/MIME bugfixes and compatibility changes:
3255 output correct application/pkcs7 MIME type if
3256 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3257 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3258 of files as .eml work). Correctly handle very long lines in MIME
3262 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3264 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3265 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3266 a protocol version number mismatch like a decryption error
3267 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3270 *) Turn on RSA blinding by default in the default implementation
3271 to avoid a timing attack. Applications that don't want it can call
3272 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3273 They would be ill-advised to do so in most cases.
3274 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3276 *) Change RSA blinding code so that it works when the PRNG is not
3277 seeded (in this case, the secret RSA exponent is abused as
3278 an unpredictable seed -- if it is not unpredictable, there
3279 is no point in blinding anyway). Make RSA blinding thread-safe
3280 by remembering the creator's thread ID in rsa->blinding and
3281 having all other threads use local one-time blinding factors
3282 (this requires more computation than sharing rsa->blinding, but
3283 avoids excessive locking; and if an RSA object is not shared
3284 between threads, blinding will still be very fast).
3287 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3288 ENGINE as defaults for all supported algorithms irrespective of
3289 the 'flags' parameter. 'flags' is now honoured, so applications
3290 should make sure they are passing it correctly.
3293 *) Target "mingw" now allows native Windows code to be generated in
3294 the Cygwin environment as well as with the MinGW compiler.
3297 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3299 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3300 via timing by performing a MAC computation even if incorrrect
3301 block cipher padding has been found. This is a countermeasure
3302 against active attacks where the attacker has to distinguish
3303 between bad padding and a MAC verification error. (CVE-2003-0078)
3305 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3306 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3307 Martin Vuagnoux (EPFL, Ilion)]
3309 *) Make the no-err option work as intended. The intention with no-err
3310 is not to have the whole error stack handling routines removed from
3311 libcrypto, it's only intended to remove all the function name and
3312 reason texts, thereby removing some of the footprint that may not
3313 be interesting if those errors aren't displayed anyway.
3315 NOTE: it's still possible for any application or module to have it's
3316 own set of error texts inserted. The routines are there, just not
3317 used by default when no-err is given.
3320 *) Add support for FreeBSD on IA64.
3321 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3323 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3324 Kerberos function mit_des_cbc_cksum(). Before this change,
3325 the value returned by DES_cbc_cksum() was like the one from
3326 mit_des_cbc_cksum(), except the bytes were swapped.
3327 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3329 *) Allow an application to disable the automatic SSL chain building.
3330 Before this a rather primitive chain build was always performed in
3331 ssl3_output_cert_chain(): an application had no way to send the
3332 correct chain if the automatic operation produced an incorrect result.
3334 Now the chain builder is disabled if either:
3336 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3338 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3340 The reasoning behind this is that an application would not want the
3341 auto chain building to take place if extra chain certificates are
3342 present and it might also want a means of sending no additional
3343 certificates (for example the chain has two certificates and the
3347 *) Add the possibility to build without the ENGINE framework.
3348 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3350 *) Under Win32 gmtime() can return NULL: check return value in
3351 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3354 *) DSA routines: under certain error conditions uninitialized BN objects
3355 could be freed. Solution: make sure initialization is performed early
3356 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3357 Nils Larsch <nla@trustcenter.de> via PR#459)
3360 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3361 checked on reconnect on the client side, therefore session resumption
3362 could still fail with a "ssl session id is different" error. This
3363 behaviour is masked when SSL_OP_ALL is used due to
3364 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3365 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3366 followup to PR #377.
3369 *) IA-32 assembler support enhancements: unified ELF targets, support
3370 for SCO/Caldera platforms, fix for Cygwin shared build.
3373 *) Add support for FreeBSD on sparc64. As a consequence, support for
3374 FreeBSD on non-x86 processors is separate from x86 processors on
3375 the config script, much like the NetBSD support.
3376 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3378 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3380 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3383 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3384 code (06) was taken as the first octet of the session ID and the last
3385 octet was ignored consequently. As a result SSLv2 client side session
3386 caching could not have worked due to the session ID mismatch between
3388 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3392 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3393 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3397 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3398 seems that in spite of existing for more than a year, many application
3399 author have done nothing to provide the necessary callbacks, which
3400 means that this particular engine will not work properly anywhere.
3401 This is a very unfortunate situation which forces us, in the name
3402 of usability, to give the hw_ncipher.c a static lock, which is part
3404 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3405 appear in 0.9.8 or later. We EXPECT application authors to have
3406 dealt properly with this when 0.9.8 is released (unless we actually
3407 make such changes in the libcrypto locking code that changes will
3408 have to be made anyway).
3411 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3412 octets have been read, EOF or an error occurs. Without this change
3413 some truncated ASN1 structures will not produce an error.
3416 *) Disable Heimdal support, since it hasn't been fully implemented.
3417 Still give the possibility to force the use of Heimdal, but with
3418 warnings and a request that patches get sent to openssl-dev.
3421 *) Add the VC-CE target, introduce the WINCE sysname, and add
3422 INSTALL.WCE and appropriate conditionals to make it build.
3423 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3425 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3426 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3427 edit numbers of the version.
3428 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3430 *) Introduce safe string copy and catenation functions
3431 (BUF_strlcpy() and BUF_strlcat()).
3432 [Ben Laurie (CHATS) and Richard Levitte]
3434 *) Avoid using fixed-size buffers for one-line DNs.
3435 [Ben Laurie (CHATS)]
3437 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3438 resizing buffers containing secrets, and use where appropriate.
3439 [Ben Laurie (CHATS)]
3441 *) Avoid using fixed size buffers for configuration file location.
3442 [Ben Laurie (CHATS)]
3444 *) Avoid filename truncation for various CA files.
3445 [Ben Laurie (CHATS)]
3447 *) Use sizeof in preference to magic numbers.
3448 [Ben Laurie (CHATS)]
3450 *) Avoid filename truncation in cert requests.
3451 [Ben Laurie (CHATS)]
3453 *) Add assertions to check for (supposedly impossible) buffer
3455 [Ben Laurie (CHATS)]
3457 *) Don't cache truncated DNS entries in the local cache (this could
3458 potentially lead to a spoofing attack).
3459 [Ben Laurie (CHATS)]
3461 *) Fix various buffers to be large enough for hex/decimal
3462 representations in a platform independent manner.
3463 [Ben Laurie (CHATS)]
3465 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3466 resizing buffers containing secrets, and use where appropriate.
3467 [Ben Laurie (CHATS)]
3469 *) Add BIO_indent() to avoid much slightly worrying code to do
3471 [Ben Laurie (CHATS)]
3473 *) Convert sprintf()/BIO_puts() to BIO_printf().
3474 [Ben Laurie (CHATS)]
3476 *) buffer_gets() could terminate with the buffer only half
3478 [Ben Laurie (CHATS)]
3480 *) Add assertions to prevent user-supplied crypto functions from
3481 overflowing internal buffers by having large block sizes, etc.
3482 [Ben Laurie (CHATS)]
3484 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3486 [Ben Laurie (CHATS)]
3488 *) Eliminate unused copy of key in RC4.
3489 [Ben Laurie (CHATS)]
3491 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3492 [Ben Laurie (CHATS)]
3494 *) Fix off-by-one error in EGD path.
3495 [Ben Laurie (CHATS)]
3497 *) If RANDFILE path is too long, ignore instead of truncating.
3498 [Ben Laurie (CHATS)]
3500 *) Eliminate unused and incorrectly sized X.509 structure
3502 [Ben Laurie (CHATS)]
3504 *) Eliminate unused and dangerous function knumber().
3505 [Ben Laurie (CHATS)]
3507 *) Eliminate unused and dangerous structure, KSSL_ERR.
3508 [Ben Laurie (CHATS)]
3510 *) Protect against overlong session ID context length in an encoded
3511 session object. Since these are local, this does not appear to be
3513 [Ben Laurie (CHATS)]
3515 *) Change from security patch (see 0.9.6e below) that did not affect
3516 the 0.9.6 release series:
3518 Remote buffer overflow in SSL3 protocol - an attacker could
3519 supply an oversized master key in Kerberos-enabled versions.
3521 [Ben Laurie (CHATS)]
3523 *) Change the SSL kerb5 codes to match RFC 2712.
3526 *) Make -nameopt work fully for req and add -reqopt switch.
3527 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3529 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3530 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3532 *) Make sure tests can be performed even if the corresponding algorithms
3533 have been removed entirely. This was also the last step to make
3534 OpenSSL compilable with DJGPP under all reasonable conditions.
3535 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3537 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3538 to allow version independent disabling of normally unselected ciphers,
3539 which may be activated as a side-effect of selecting a single cipher.
3541 (E.g., cipher list string "RSA" enables ciphersuites that are left
3542 out of "ALL" because they do not provide symmetric encryption.
3543 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3544 [Lutz Jaenicke, Bodo Moeller]
3546 *) Add appropriate support for separate platform-dependent build
3547 directories. The recommended way to make a platform-dependent
3548 build directory is the following (tested on Linux), maybe with
3551 # Place yourself outside of the OpenSSL source tree. In
3552 # this example, the environment variable OPENSSL_SOURCE
3553 # is assumed to contain the absolute OpenSSL source directory.
3554 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3555 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3556 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3557 mkdir -p `dirname $F`
3558 ln -s $OPENSSL_SOURCE/$F $F
3561 To be absolutely sure not to disturb the source tree, a "make clean"
3562 is a good thing. If it isn't successfull, don't worry about it,
3563 it probably means the source directory is very clean.
3566 *) Make sure any ENGINE control commands make local copies of string
3567 pointers passed to them whenever necessary. Otherwise it is possible
3568 the caller may have overwritten (or deallocated) the original string
3569 data when a later ENGINE operation tries to use the stored values.
3570 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3572 *) Improve diagnostics in file reading and command-line digests.
3573 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3575 *) Add AES modes CFB and OFB to the object database. Correct an
3576 error in AES-CFB decryption.
3579 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3580 allows existing EVP_CIPHER_CTX structures to be reused after
3581 calling EVP_*Final(). This behaviour is used by encryption
3582 BIOs and some applications. This has the side effect that
3583 applications must explicitly clean up cipher contexts with
3584 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3587 *) Check the values of dna and dnb in bn_mul_recursive before calling
3588 bn_mul_comba (a non zero value means the a or b arrays do not contain
3589 n2 elements) and fallback to bn_mul_normal if either is not zero.
3592 *) Fix escaping of non-ASCII characters when using the -subj option
3593 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3596 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3597 form for "surname", serialNumber has no short form.
3598 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3599 therefore remove "mail" short name for "internet 7".
3600 The OID for unique identifiers in X509 certificates is
3601 x500UniqueIdentifier, not uniqueIdentifier.
3602 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3605 *) Add an "init" command to the ENGINE config module and auto initialize
3606 ENGINEs. Without any "init" command the ENGINE will be initialized
3607 after all ctrl commands have been executed on it. If init=1 the
3608 ENGINE is initailized at that point (ctrls before that point are run
3609 on the uninitialized ENGINE and after on the initialized one). If
3610 init=0 then the ENGINE will not be iniatialized at all.
3613 *) Fix the 'app_verify_callback' interface so that the user-defined
3614 argument is actually passed to the callback: In the
3615 SSL_CTX_set_cert_verify_callback() prototype, the callback
3616 declaration has been changed from
3619 int (*cb)(X509_STORE_CTX *,void *);
3620 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3621 i=s->ctx->app_verify_callback(&ctx)
3622 has been changed into
3623 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3625 To update applications using SSL_CTX_set_cert_verify_callback(),
3626 a dummy argument can be added to their callback functions.
3627 [D. K. Smetters <smetters@parc.xerox.com>]
3629 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3630 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3632 *) Add and OPENSSL_LOAD_CONF define which will cause
3633 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3634 This allows older applications to transparently support certain
3635 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3636 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3637 load the config file and OPENSSL_add_all_algorithms_conf() which will
3638 always load it have also been added.
3641 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3642 Adjust NIDs and EVP layer.
3643 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3645 *) Config modules support in openssl utility.
3647 Most commands now load modules from the config file,
3648 though in a few (such as version) this isn't done
3649 because it couldn't be used for anything.
3651 In the case of ca and req the config file used is
3652 the same as the utility itself: that is the -config
3653 command line option can be used to specify an
3657 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3658 use "openssl_conf" if filename is NULL use default openssl config file.
3661 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3662 config section name. Add a new flag to tolerate a missing config file
3663 and move code to CONF_modules_load_file().
3666 *) Support for crypto accelerator cards from Accelerated Encryption
3667 Processing, www.aep.ie. (Use engine 'aep')
3668 The support was copied from 0.9.6c [engine] and adapted/corrected
3669 to work with the new engine framework.
3670 [AEP Inc. and Richard Levitte]
3672 *) Support for SureWare crypto accelerator cards from Baltimore
3673 Technologies. (Use engine 'sureware')
3674 The support was copied from 0.9.6c [engine] and adapted
3675 to work with the new engine framework.
3678 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3679 make the newer ENGINE framework commands for the CHIL engine work.
3680 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3682 *) Make it possible to produce shared libraries on ReliantUNIX.
3683 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3685 *) Add the configuration target debug-linux-ppro.
3686 Make 'openssl rsa' use the general key loading routines
3687 implemented in apps.c, and make those routines able to
3688 handle the key format FORMAT_NETSCAPE and the variant
3690 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3692 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3693 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3695 *) Add -keyform to rsautl, and document -engine.
3696 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3698 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3699 BIO_R_NO_SUCH_FILE error code rather than the generic
3700 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3703 *) Add new functions
3705 ERR_peek_last_error_line
3706 ERR_peek_last_error_line_data.
3707 These are similar to
3710 ERR_peek_error_line_data,
3711 but report on the latest error recorded rather than the first one
3712 still in the error queue.
3713 [Ben Laurie, Bodo Moeller]
3715 *) default_algorithms option in ENGINE config module. This allows things
3717 default_algorithms = ALL
3718 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3721 *) Prelminary ENGINE config module.
3724 *) New experimental application configuration code.
3727 *) Change the AES code to follow the same name structure as all other
3728 symmetric ciphers, and behave the same way. Move everything to
3729 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3730 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3732 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3733 [Ben Laurie and Theo de Raadt]
3735 *) Add option to output public keys in req command.
3736 [Massimiliano Pala madwolf@openca.org]
3738 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3739 (up to about 10% better than before for P-192 and P-224).
3742 *) New functions/macros
3744 SSL_CTX_set_msg_callback(ctx, cb)
3745 SSL_CTX_set_msg_callback_arg(ctx, arg)
3746 SSL_set_msg_callback(ssl, cb)
3747 SSL_set_msg_callback_arg(ssl, arg)
3749 to request calling a callback function
3751 void cb(int write_p, int version, int content_type,
3752 const void *buf, size_t len, SSL *ssl, void *arg)
3754 whenever a protocol message has been completely received
3755 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3756 protocol version according to which the SSL library interprets
3757 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3758 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3759 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3760 specification (change_cipher_spec(20), alert(21), handshake(22)).
3761 'buf' and 'len' point to the actual message, 'ssl' to the
3762 SSL object, and 'arg' is the application-defined value set by
3763 SSL[_CTX]_set_msg_callback_arg().
3765 'openssl s_client' and 'openssl s_server' have new '-msg' options
3766 to enable a callback that displays all protocol messages.
3769 *) Change the shared library support so shared libraries are built as
3770 soon as the corresponding static library is finished, and thereby get
3771 openssl and the test programs linked against the shared library.
3772 This still only happens when the keyword "shard" has been given to
3773 the configuration scripts.
3775 NOTE: shared library support is still an experimental thing, and
3776 backward binary compatibility is still not guaranteed.
3777 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3779 *) Add support for Subject Information Access extension.
3780 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3782 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3783 additional bytes when new memory had to be allocated, not just
3784 when reusing an existing buffer.
3787 *) New command line and configuration option 'utf8' for the req command.
3788 This allows field values to be specified as UTF8 strings.
3791 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3792 runs for the former and machine-readable output for the latter.
3795 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3796 of the e-mail address in the DN (i.e., it will go into a certificate
3797 extension only). The new configuration file option 'email_in_dn = no'
3798 has the same effect.
3799 [Massimiliano Pala madwolf@openca.org]
3801 *) Change all functions with names starting with des_ to be starting
3802 with DES_ instead. Add wrappers that are compatible with libdes,
3803 but are named _ossl_old_des_*. Finally, add macros that map the
3804 des_* symbols to the corresponding _ossl_old_des_* if libdes
3805 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3806 desired, the des_* symbols will be mapped to DES_*, with one
3809 Since we provide two compatibility mappings, the user needs to
3810 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3811 compatibility is desired. The default (i.e., when that macro
3812 isn't defined) is OpenSSL 0.9.6c compatibility.
3814 There are also macros that enable and disable the support of old
3815 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3816 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3817 are defined, the default will apply: to support the old des routines.
3819 In either case, one must include openssl/des.h to get the correct
3820 definitions. Do not try to just include openssl/des_old.h, that
3823 NOTE: This is a major break of an old API into a new one. Software
3824 authors are encouraged to switch to the DES_ style functions. Some
3825 time in the future, des_old.h and the libdes compatibility functions
3826 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3827 default), and then completely removed.
3830 *) Test for certificates which contain unsupported critical extensions.
3831 If such a certificate is found during a verify operation it is
3832 rejected by default: this behaviour can be overridden by either
3833 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3834 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3835 X509_supported_extension() has also been added which returns 1 if a
3836 particular extension is supported.
3839 *) Modify the behaviour of EVP cipher functions in similar way to digests
3840 to retain compatibility with existing code.
3843 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3844 compatibility with existing code. In particular the 'ctx' parameter does
3845 not have to be to be initialized before the call to EVP_DigestInit() and
3846 it is tidied up after a call to EVP_DigestFinal(). New function
3847 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3848 EVP_MD_CTX_copy() changed to not require the destination to be
3849 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3850 requires the destination to be valid.
3852 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3853 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3856 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3857 so that complete 'Handshake' protocol structures are kept in memory
3858 instead of overwriting 'msg_type' and 'length' with 'body' data.
3861 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3862 [Massimo Santin via Richard Levitte]
3864 *) Major restructuring to the underlying ENGINE code. This includes
3865 reduction of linker bloat, separation of pure "ENGINE" manipulation
3866 (initialisation, etc) from functionality dealing with implementations
3867 of specific crypto iterfaces. This change also introduces integrated
3868 support for symmetric ciphers and digest implementations - so ENGINEs
3869 can now accelerate these by providing EVP_CIPHER and EVP_MD
3870 implementations of their own. This is detailed in crypto/engine/README
3871 as it couldn't be adequately described here. However, there are a few
3872 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3873 were changed in the original introduction of ENGINE code have now
3874 reverted back - the hooking from this code to ENGINE is now a good
3875 deal more passive and at run-time, operations deal directly with
3876 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3877 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3878 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3879 they were not being used by the framework as there is no concept of a
3880 BIGNUM_METHOD and they could not be generalised to the new
3881 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3882 ENGINE_cpy() has been removed as it cannot be consistently defined in
3886 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3889 *) Change mkdef.pl to sort symbols that get the same entry number,
3890 and make sure the automatically generated functions ERR_load_*
3891 become part of libeay.num as well.
3894 *) New function SSL_renegotiate_pending(). This returns true once
3895 renegotiation has been requested (either SSL_renegotiate() call
3896 or HelloRequest/ClientHello receveived from the peer) and becomes
3897 false once a handshake has been completed.
3898 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3899 sends a HelloRequest, but does not ensure that a handshake takes
3900 place. SSL_renegotiate_pending() is useful for checking if the
3901 client has followed the request.)
3904 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3905 By default, clients may request session resumption even during
3906 renegotiation (if session ID contexts permit); with this option,
3907 session resumption is possible only in the first handshake.
3909 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3910 more bits available for options that should not be part of
3911 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3914 *) Add some demos for certificate and certificate request creation.
3917 *) Make maximum certificate chain size accepted from the peer application
3918 settable (SSL*_get/set_max_cert_list()), as proposed by
3919 "Douglas E. Engert" <deengert@anl.gov>.
3922 *) Add support for shared libraries for Unixware-7
3923 (Boyd Lynn Gerber <gerberb@zenez.com>).
3926 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3927 be done prior to destruction. Use this to unload error strings from
3928 ENGINEs that load their own error strings. NB: This adds two new API
3929 functions to "get" and "set" this destroy handler in an ENGINE.
3932 *) Alter all existing ENGINE implementations (except "openssl" and
3933 "openbsd") to dynamically instantiate their own error strings. This
3934 makes them more flexible to be built both as statically-linked ENGINEs
3935 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3936 Also, add stub code to each that makes building them as self-contained
3937 shared-libraries easier (see README.ENGINE).
3940 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3941 implementations into applications that are completely implemented in
3942 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3943 commands that can be used to configure what shared-library to load and
3944 to control aspects of the way it is handled. Also, made an update to
3945 the README.ENGINE file that brings its information up-to-date and
3946 provides some information and instructions on the "dynamic" ENGINE
3947 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3950 *) Make it possible to unload ranges of ERR strings with a new
3951 "ERR_unload_strings" function.
3954 *) Add a copy() function to EVP_MD.
3957 *) Make EVP_MD routines take a context pointer instead of just the
3958 md_data void pointer.
3961 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3962 that the digest can only process a single chunk of data
3963 (typically because it is provided by a piece of
3964 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3965 is only going to provide a single chunk of data, and hence the
3966 framework needn't accumulate the data for oneshot drivers.
3969 *) As with "ERR", make it possible to replace the underlying "ex_data"
3970 functions. This change also alters the storage and management of global
3971 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3972 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3973 index counters. The API functions that use this state have been changed
3974 to take a "class_index" rather than pointers to the class's local STACK
3975 and counter, and there is now an API function to dynamically create new
3976 classes. This centralisation allows us to (a) plug a lot of the
3977 thread-safety problems that existed, and (b) makes it possible to clean
3978 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3979 such data would previously have always leaked in application code and
3980 workarounds were in place to make the memory debugging turn a blind eye
3981 to it. Application code that doesn't use this new function will still
3982 leak as before, but their memory debugging output will announce it now
3983 rather than letting it slide.
3985 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3986 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3987 has a return value to indicate success or failure.
3990 *) Make it possible to replace the underlying "ERR" functions such that the
3991 global state (2 LHASH tables and 2 locks) is only used by the "default"
3992 implementation. This change also adds two functions to "get" and "set"
3993 the implementation prior to it being automatically set the first time
3994 any other ERR function takes place. Ie. an application can call "get",
3995 pass the return value to a module it has just loaded, and that module
3996 can call its own "set" function using that value. This means the
3997 module's "ERR" operations will use (and modify) the error state in the
3998 application and not in its own statically linked copy of OpenSSL code.
4001 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4002 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4003 the operation, and provides a more encapsulated way for external code
4004 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4005 to use these functions rather than manually incrementing the counts.
4007 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4010 *) Add EVP test program.
4013 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4016 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4017 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4018 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4019 These allow a CRL to be built without having to access X509_CRL fields
4020 directly. Modify 'ca' application to use new functions.
4023 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4024 bug workarounds. Rollback attack detection is a security feature.
4025 The problem will only arise on OpenSSL servers when TLSv1 is not
4026 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4027 Software authors not wanting to support TLSv1 will have special reasons
4028 for their choice and can explicitly enable this option.
4029 [Bodo Moeller, Lutz Jaenicke]
4031 *) Rationalise EVP so it can be extended: don't include a union of
4032 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4033 (similar to those existing for EVP_CIPHER_CTX).
4038 EVP_MD_CTX_init(&md); /* new function call */
4039 EVP_DigestInit(&md, EVP_sha1());
4040 EVP_DigestUpdate(&md, in, len);
4041 EVP_DigestFinal(&md, out, NULL);
4042 EVP_MD_CTX_cleanup(&md); /* new function call */
4046 *) Make DES key schedule conform to the usual scheme, as well as
4047 correcting its structure. This means that calls to DES functions
4048 now have to pass a pointer to a des_key_schedule instead of a
4049 plain des_key_schedule (which was actually always a pointer
4052 des_key_schedule ks;
4054 des_set_key_checked(..., &ks);
4055 des_ncbc_encrypt(..., &ks, ...);
4057 (Note that a later change renames 'des_...' into 'DES_...'.)
4060 *) Initial reduction of linker bloat: the use of some functions, such as
4061 PEM causes large amounts of unused functions to be linked in due to
4062 poor organisation. For example pem_all.c contains every PEM function
4063 which has a knock on effect of linking in large amounts of (unused)
4064 ASN1 code. Grouping together similar functions and splitting unrelated
4065 functions prevents this.
4068 *) Cleanup of EVP macros.
4071 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4072 correct _ecb suffix.
4075 *) Add initial OCSP responder support to ocsp application. The
4076 revocation information is handled using the text based index
4077 use by the ca application. The responder can either handle
4078 requests generated internally, supplied in files (for example
4079 via a CGI script) or using an internal minimal server.
4082 *) Add configuration choices to get zlib compression for TLS.
4085 *) Changes to Kerberos SSL for RFC 2712 compliance:
4086 1. Implemented real KerberosWrapper, instead of just using
4087 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4088 2. Implemented optional authenticator field of KerberosWrapper.
4090 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4091 and authenticator structs; see crypto/krb5/.
4093 Generalized Kerberos calls to support multiple Kerberos libraries.
4094 [Vern Staats <staatsvr@asc.hpc.mil>,
4095 Jeffrey Altman <jaltman@columbia.edu>
4096 via Richard Levitte]
4098 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4099 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4100 values for each of the key sizes rather than having just
4101 parameters (and 'speed' generating keys each time).
4104 *) Speed up EVP routines.
4107 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4108 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4109 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4110 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4112 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4113 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4114 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4117 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4119 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4122 *) Added the OS2-EMX target.
4123 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4125 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4126 to support NCONF routines in extension code. New function CONF_set_nconf()
4127 to allow functions which take an NCONF to also handle the old LHASH
4128 structure: this means that the old CONF compatible routines can be
4129 retained (in particular wrt extensions) without having to duplicate the
4130 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4133 *) Enhance the general user interface with mechanisms for inner control
4134 and with possibilities to have yes/no kind of prompts.
4137 *) Change all calls to low level digest routines in the library and
4138 applications to use EVP. Add missing calls to HMAC_cleanup() and
4139 don't assume HMAC_CTX can be copied using memcpy().
4140 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4142 *) Add the possibility to control engines through control names but with
4143 arbitrary arguments instead of just a string.
4144 Change the key loaders to take a UI_METHOD instead of a callback
4145 function pointer. NOTE: this breaks binary compatibility with earlier
4146 versions of OpenSSL [engine].
4147 Adapt the nCipher code for these new conditions and add a card insertion
4151 *) Enhance the general user interface with mechanisms to better support
4152 dialog box interfaces, application-defined prompts, the possibility
4153 to use defaults (for example default passwords from somewhere else)
4154 and interrupts/cancellations.
4157 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4158 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4161 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4162 tidy up some unnecessarily weird code in 'sk_new()').
4163 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4165 *) Change the key loading routines for ENGINEs to use the same kind
4166 callback (pem_password_cb) as all other routines that need this
4170 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4171 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4172 than this minimum value is recommended.
4175 *) New random seeder for OpenVMS, using the system process statistics
4176 that are easily reachable.
4179 *) Windows apparently can't transparently handle global
4180 variables defined in DLLs. Initialisations such as:
4182 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4184 wont compile. This is used by the any applications that need to
4185 declare their own ASN1 modules. This was fixed by adding the option
4186 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4187 needed for static libraries under Win32.
4190 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4191 setting of purpose and trust fields. New X509_STORE trust and
4192 purpose functions and tidy up setting in other SSL functions.
4195 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4196 structure. These are inherited by X509_STORE_CTX when it is
4197 initialised. This allows various defaults to be set in the
4198 X509_STORE structure (such as flags for CRL checking and custom
4199 purpose or trust settings) for functions which only use X509_STORE_CTX
4200 internally such as S/MIME.
4202 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4203 trust settings if they are not set in X509_STORE. This allows X509_STORE
4204 purposes and trust (in S/MIME for example) to override any set by default.
4206 Add command line options for CRL checking to smime, s_client and s_server
4210 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4211 are set then the CRL is looked up in the X509_STORE structure and
4212 its validity and signature checked, then if the certificate is found
4213 in the CRL the verify fails with a revoked error.
4215 Various new CRL related callbacks added to X509_STORE_CTX structure.
4217 Command line options added to 'verify' application to support this.
4219 This needs some additional work, such as being able to handle multiple
4220 CRLs with different times, extension based lookup (rather than just
4221 by subject name) and ultimately more complete V2 CRL extension
4225 *) Add a general user interface API (crypto/ui/). This is designed
4226 to replace things like des_read_password and friends (backward
4227 compatibility functions using this new API are provided).
4228 The purpose is to remove prompting functions from the DES code
4229 section as well as provide for prompting through dialog boxes in
4230 a window system and the like.
4233 *) Add "ex_data" support to ENGINE so implementations can add state at a
4234 per-structure level rather than having to store it globally.
4237 *) Make it possible for ENGINE structures to be copied when retrieved by
4238 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4239 This causes the "original" ENGINE structure to act like a template,
4240 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4241 operational state can be localised to each ENGINE structure, despite the
4242 fact they all share the same "methods". New ENGINE structures returned in
4243 this case have no functional references and the return value is the single
4244 structural reference. This matches the single structural reference returned
4245 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4249 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4250 needs to match any other type at all we need to manually clear the
4254 *) Changes to the "openssl engine" utility to include;
4255 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4256 about an ENGINE's available control commands.
4257 - executing control commands from command line arguments using the
4258 '-pre' and '-post' switches. '-post' is only used if '-t' is
4259 specified and the ENGINE is successfully initialised. The syntax for
4260 the individual commands are colon-separated, for example;
4261 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4264 *) New dynamic control command support for ENGINEs. ENGINEs can now
4265 declare their own commands (numbers), names (strings), descriptions,
4266 and input types for run-time discovery by calling applications. A
4267 subset of these commands are implicitly classed as "executable"
4268 depending on their input type, and only these can be invoked through
4269 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4270 can be based on user input, config files, etc). The distinction is
4271 that "executable" commands cannot return anything other than a boolean
4272 result and can only support numeric or string input, whereas some
4273 discoverable commands may only be for direct use through
4274 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4275 pointers, or other custom uses. The "executable" commands are to
4276 support parameterisations of ENGINE behaviour that can be
4277 unambiguously defined by ENGINEs and used consistently across any
4278 OpenSSL-based application. Commands have been added to all the
4279 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4280 control over shared-library paths without source code alterations.
4283 *) Changed all ENGINE implementations to dynamically allocate their
4284 ENGINEs rather than declaring them statically. Apart from this being
4285 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4286 this also allows the implementations to compile without using the
4287 internal engine_int.h header.
4290 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4291 'const' value. Any code that should be able to modify a RAND_METHOD
4292 should already have non-const pointers to it (ie. they should only
4293 modify their own ones).
4296 *) Made a variety of little tweaks to the ENGINE code.
4297 - "atalla" and "ubsec" string definitions were moved from header files
4298 to C code. "nuron" string definitions were placed in variables
4299 rather than hard-coded - allowing parameterisation of these values
4300 later on via ctrl() commands.
4301 - Removed unused "#if 0"'d code.
4302 - Fixed engine list iteration code so it uses ENGINE_free() to release
4303 structural references.
4304 - Constified the RAND_METHOD element of ENGINE structures.
4305 - Constified various get/set functions as appropriate and added
4306 missing functions (including a catch-all ENGINE_cpy that duplicates
4307 all ENGINE values onto a new ENGINE except reference counts/state).
4308 - Removed NULL parameter checks in get/set functions. Setting a method
4309 or function to NULL is a way of cancelling out a previously set
4310 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4311 and doesn't justify the extra error symbols and code.
4312 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4313 flags from engine_int.h to engine.h.
4314 - Changed prototypes for ENGINE handler functions (init(), finish(),
4315 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4318 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4319 to the algorithm using long division. The binary algorithm can be
4320 used only if the modulus is odd. On 32-bit systems, it is faster
4321 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4322 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4323 up to 450 bits. In 64-bit environments, the binary algorithm
4324 appears to be advantageous for much longer moduli; here we use it
4325 for moduli up to 2048 bits.
4328 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4329 could not support the combine flag in choice fields.
4332 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4333 extensions from a certificate request to the certificate.
4336 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4337 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4338 file: this allows the display of the certificate about to be
4339 signed to be customised, to allow certain fields to be included
4340 or excluded and extension details. The old system didn't display
4341 multicharacter strings properly, omitted fields not in the policy
4342 and couldn't display additional details such as extensions.
4345 *) Function EC_POINTs_mul for multiple scalar multiplication
4346 of an arbitrary number of elliptic curve points
4347 \sum scalars[i]*points[i],
4348 optionally including the generator defined for the EC_GROUP:
4349 scalar*generator + \sum scalars[i]*points[i].
4351 EC_POINT_mul is a simple wrapper function for the typical case
4352 that the point list has just one item (besides the optional
4356 *) First EC_METHODs for curves over GF(p):
4358 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4359 operations and provides various method functions that can also
4360 operate with faster implementations of modular arithmetic.
4362 EC_GFp_mont_method() reuses most functions that are part of
4363 EC_GFp_simple_method, but uses Montgomery arithmetic.
4365 [Bodo Moeller; point addition and point doubling
4366 implementation directly derived from source code provided by
4367 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4369 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4370 crypto/ec/ec_lib.c):
4372 Curves are EC_GROUP objects (with an optional group generator)
4373 based on EC_METHODs that are built into the library.
4375 Points are EC_POINT objects based on EC_GROUP objects.
4377 Most of the framework would be able to handle curves over arbitrary
4378 finite fields, but as there are no obvious types for fields other
4379 than GF(p), some functions are limited to that for now.
4382 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4383 that the file contains a complete HTTP response.
4386 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4387 change the def and num file printf format specifier from "%-40sXXX"
4388 to "%-39s XXX". The latter will always guarantee a space after the
4389 field while the former will cause them to run together if the field
4390 is 40 of more characters long.
4393 *) Constify the cipher and digest 'method' functions and structures
4394 and modify related functions to take constant EVP_MD and EVP_CIPHER
4398 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4399 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4402 *) Modify EVP_Digest*() routines so they now return values. Although the
4403 internal software routines can never fail additional hardware versions
4407 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4409 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4410 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4413 ERR_R_NESTED_ASN1_ERROR
4415 ERR_R_MISSING_ASN1_EOS
4416 were 4 .. 9, conflicting with
4417 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4419 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4420 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4422 Add new error code 'ERR_R_INTERNAL_ERROR'.
4425 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4429 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4430 sets the subject name for a new request or supersedes the
4431 subject name in a given request. Formats that can be parsed are
4432 'CN=Some Name, OU=myOU, C=IT'
4434 'CN=Some Name/OU=myOU/C=IT'.
4436 Add options '-batch' and '-verbose' to 'openssl req'.
4437 [Massimiliano Pala <madwolf@hackmasters.net>]
4439 *) Introduce the possibility to access global variables through
4440 functions on platform were that's the best way to handle exporting
4441 global variables in shared libraries. To enable this functionality,
4442 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4443 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4444 is normally done by Configure or something similar).
4446 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4447 in the source file (foo.c) like this:
4449 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4450 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4452 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4453 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4455 OPENSSL_DECLARE_GLOBAL(int,foo);
4456 #define foo OPENSSL_GLOBAL_REF(foo)
4457 OPENSSL_DECLARE_GLOBAL(double,bar);
4458 #define bar OPENSSL_GLOBAL_REF(bar)
4460 The #defines are very important, and therefore so is including the
4461 header file everywhere where the defined globals are used.
4463 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4464 of ASN.1 items, but that structure is a bit different.
4466 The largest change is in util/mkdef.pl which has been enhanced with
4467 better and easier to understand logic to choose which symbols should
4468 go into the Windows .def files as well as a number of fixes and code
4469 cleanup (among others, algorithm keywords are now sorted
4470 lexicographically to avoid constant rewrites).
4473 *) In BN_div() keep a copy of the sign of 'num' before writing the
4474 result to 'rm' because if rm==num the value will be overwritten
4475 and produce the wrong result if 'num' is negative: this caused
4476 problems with BN_mod() and BN_nnmod().
4479 *) Function OCSP_request_verify(). This checks the signature on an
4480 OCSP request and verifies the signer certificate. The signer
4481 certificate is just checked for a generic purpose and OCSP request
4485 *) Add OCSP_check_validity() function to check the validity of OCSP
4486 responses. OCSP responses are prepared in real time and may only
4487 be a few seconds old. Simply checking that the current time lies
4488 between thisUpdate and nextUpdate max reject otherwise valid responses
4489 caused by either OCSP responder or client clock inaccuracy. Instead
4490 we allow thisUpdate and nextUpdate to fall within a certain period of
4491 the current time. The age of the response can also optionally be
4492 checked. Two new options -validity_period and -status_age added to
4496 *) If signature or public key algorithm is unrecognized print out its
4497 OID rather that just UNKNOWN.
4500 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4501 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4502 ID to be generated from the issuer certificate alone which can then be
4503 passed to OCSP_id_issuer_cmp().
4506 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4507 ASN1 modules to export functions returning ASN1_ITEM pointers
4508 instead of the ASN1_ITEM structures themselves. This adds several
4509 new macros which allow the underlying ASN1 function/structure to
4510 be accessed transparently. As a result code should not use ASN1_ITEM
4511 references directly (such as &X509_it) but instead use the relevant
4512 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4513 use of the new ASN1 code on platforms where exporting structures
4514 is problematical (for example in shared libraries) but exporting
4515 functions returning pointers to structures is not.
4518 *) Add support for overriding the generation of SSL/TLS session IDs.
4519 These callbacks can be registered either in an SSL_CTX or per SSL.
4520 The purpose of this is to allow applications to control, if they wish,
4521 the arbitrary values chosen for use as session IDs, particularly as it
4522 can be useful for session caching in multiple-server environments. A
4523 command-line switch for testing this (and any client code that wishes
4524 to use such a feature) has been added to "s_server".
4525 [Geoff Thorpe, Lutz Jaenicke]
4527 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4528 of the form '#if defined(...) || defined(...) || ...' and
4529 '#if !defined(...) && !defined(...) && ...'. This also avoids
4530 the growing number of special cases it was previously handling.
4533 *) Make all configuration macros available for application by making
4534 sure they are available in opensslconf.h, by giving them names starting
4535 with "OPENSSL_" to avoid conflicts with other packages and by making
4536 sure e_os2.h will cover all platform-specific cases together with
4538 Additionally, it is now possible to define configuration/platform-
4539 specific names (called "system identities"). In the C code, these
4540 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4541 macro with the name beginning with "OPENSSL_SYS_", which is determined
4542 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4546 *) New option -set_serial to 'req' and 'x509' this allows the serial
4547 number to use to be specified on the command line. Previously self
4548 signed certificates were hard coded with serial number 0 and the
4549 CA options of 'x509' had to use a serial number in a file which was
4553 *) New options to 'ca' utility to support V2 CRL entry extensions.
4554 Currently CRL reason, invalidity date and hold instruction are
4555 supported. Add new CRL extensions to V3 code and some new objects.
4558 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4559 disable standard block padding (aka PKCS#5 padding) in the EVP
4560 API, which was previously mandatory. This means that the data is
4561 not padded in any way and so the total length much be a multiple
4562 of the block size, otherwise an error occurs.
4565 *) Initial (incomplete) OCSP SSL support.
4568 *) New function OCSP_parse_url(). This splits up a URL into its host,
4569 port and path components: primarily to parse OCSP URLs. New -url
4570 option to ocsp utility.
4573 *) New nonce behavior. The return value of OCSP_check_nonce() now
4574 reflects the various checks performed. Applications can decide
4575 whether to tolerate certain situations such as an absent nonce
4576 in a response when one was present in a request: the ocsp application
4577 just prints out a warning. New function OCSP_add1_basic_nonce()
4578 this is to allow responders to include a nonce in a response even if
4579 the request is nonce-less.
4582 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4583 skipped when using openssl x509 multiple times on a single input file,
4584 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4587 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4588 set string type: to handle setting ASN1_TIME structures. Fix ca
4589 utility to correctly initialize revocation date of CRLs.
4592 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4593 the clients preferred ciphersuites and rather use its own preferences.
4594 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4595 Internet Explorer by ensuring unchanged hash method during stepup.
4596 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4599 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4600 to aes and add a new 'exist' option to print out symbols that don't
4604 *) Additional options to ocsp utility to allow flags to be set and
4605 additional certificates supplied.
4608 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4609 OCSP client a number of certificate to only verify the response
4613 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4614 handle the new API. Currently only ECB, CBC modes supported. Add new
4617 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4618 Encryption Standard (AES) Ciphersuites for Transport Layer
4619 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4620 not enabled by default and were not part of the "ALL" ciphersuite
4621 alias because they were not yet official; they could be
4622 explicitly requested by specifying the "AESdraft" ciphersuite
4623 group alias. In the final release of OpenSSL 0.9.7, the group
4624 alias is called "AES" and is part of "ALL".)
4625 [Ben Laurie, Steve Henson, Bodo Moeller]
4627 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4628 request to response.
4631 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4632 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4633 extract information from a certificate request. OCSP_response_create()
4634 creates a response and optionally adds a basic response structure.
4635 OCSP_basic_add1_status() adds a complete single response to a basic
4636 response and returns the OCSP_SINGLERESP structure just added (to allow
4637 extensions to be included for example). OCSP_basic_add1_cert() adds a
4638 certificate to a basic response and OCSP_basic_sign() signs a basic
4639 response with various flags. New helper functions ASN1_TIME_check()
4640 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4641 (converts ASN1_TIME to GeneralizedTime).
4644 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4645 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4646 structure from a certificate. X509_pubkey_digest() digests the public_key
4647 contents: this is used in various key identifiers.
4650 *) Make sk_sort() tolerate a NULL argument.
4651 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4653 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4654 passed by the function are trusted implicitly. If any of them signed the
4655 response then it is assumed to be valid and is not verified.
4658 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4659 to data. This was previously part of the PKCS7 ASN1 code. This
4660 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4661 [Steve Henson, reported by Kenneth R. Robinette
4662 <support@securenetterm.com>]
4664 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4665 routines: without these tracing memory leaks is very painful.
4666 Fix leaks in PKCS12 and PKCS7 routines.
4669 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4670 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4671 effectively meant GeneralizedTime would never be used. Now it
4672 is initialised to -1 but X509_time_adj() now has to check the value
4673 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4674 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4675 [Steve Henson, reported by Kenneth R. Robinette
4676 <support@securenetterm.com>]
4678 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4679 result in a zero length in the ASN1_INTEGER structure which was
4680 not consistent with the structure when d2i_ASN1_INTEGER() was used
4681 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4682 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4683 where it did not print out a minus for negative ASN1_INTEGER.
4686 *) Add summary printout to ocsp utility. The various functions which
4687 convert status values to strings have been renamed to:
4688 OCSP_response_status_str(), OCSP_cert_status_str() and
4689 OCSP_crl_reason_str() and are no longer static. New options
4690 to verify nonce values and to disable verification. OCSP response
4691 printout format cleaned up.
4694 *) Add additional OCSP certificate checks. These are those specified
4695 in RFC2560. This consists of two separate checks: the CA of the
4696 certificate being checked must either be the OCSP signer certificate
4697 or the issuer of the OCSP signer certificate. In the latter case the
4698 OCSP signer certificate must contain the OCSP signing extended key
4699 usage. This check is performed by attempting to match the OCSP
4700 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4701 in the OCSP_CERTID structures of the response.
4704 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4705 and related routines. This uses the standard OpenSSL certificate
4706 verify routines to perform initial checks (just CA validity) and
4707 to obtain the certificate chain. Then additional checks will be
4708 performed on the chain. Currently the root CA is checked to see
4709 if it is explicitly trusted for OCSP signing. This is used to set
4710 a root CA as a global signing root: that is any certificate that
4711 chains to that CA is an acceptable OCSP signing certificate.
4714 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4715 extensions from a separate configuration file.
4716 As when reading extensions from the main configuration file,
4717 the '-extensions ...' option may be used for specifying the
4719 [Massimiliano Pala <madwolf@comune.modena.it>]
4721 *) New OCSP utility. Allows OCSP requests to be generated or
4722 read. The request can be sent to a responder and the output
4723 parsed, outputed or printed in text form. Not complete yet:
4724 still needs to check the OCSP response validity.
4727 *) New subcommands for 'openssl ca':
4728 'openssl ca -status <serial>' prints the status of the cert with
4729 the given serial number (according to the index file).
4730 'openssl ca -updatedb' updates the expiry status of certificates
4732 [Massimiliano Pala <madwolf@comune.modena.it>]
4734 *) New '-newreq-nodes' command option to CA.pl. This is like
4735 '-newreq', but calls 'openssl req' with the '-nodes' option
4736 so that the resulting key is not encrypted.
4737 [Damien Miller <djm@mindrot.org>]
4739 *) New configuration for the GNU Hurd.
4740 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4742 *) Initial code to implement OCSP basic response verify. This
4743 is currently incomplete. Currently just finds the signer's
4744 certificate and verifies the signature on the response.
4747 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4748 value of OPENSSLDIR. This is available via the new '-d' option
4749 to 'openssl version', and is also included in 'openssl version -a'.
4752 *) Allowing defining memory allocation callbacks that will be given
4753 file name and line number information in additional arguments
4754 (a const char* and an int). The basic functionality remains, as
4755 well as the original possibility to just replace malloc(),
4756 realloc() and free() by functions that do not know about these
4757 additional arguments. To register and find out the current
4758 settings for extended allocation functions, the following
4759 functions are provided:
4761 CRYPTO_set_mem_ex_functions
4762 CRYPTO_set_locked_mem_ex_functions
4763 CRYPTO_get_mem_ex_functions
4764 CRYPTO_get_locked_mem_ex_functions
4766 These work the same way as CRYPTO_set_mem_functions and friends.
4767 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4768 extended allocation function is enabled.
4769 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4770 a conventional allocation function is enabled.
4771 [Richard Levitte, Bodo Moeller]
4773 *) Finish off removing the remaining LHASH function pointer casts.
4774 There should no longer be any prototype-casting required when using
4775 the LHASH abstraction, and any casts that remain are "bugs". See
4776 the callback types and macros at the head of lhash.h for details
4777 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4780 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4781 If /dev/[u]random devices are not available or do not return enough
4782 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4784 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4785 /etc/entropy will be queried once each in this sequence, quering stops
4786 when enough entropy was collected without querying more sockets.
4789 *) Change the Unix RAND_poll() variant to be able to poll several
4790 random devices, as specified by DEVRANDOM, until a sufficient amount
4791 of data has been collected. We spend at most 10 ms on each file
4792 (select timeout) and read in non-blocking mode. DEVRANDOM now
4793 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4794 (previously it was just the string "/dev/urandom"), so on typical
4795 platforms the 10 ms delay will never occur.
4796 Also separate out the Unix variant to its own file, rand_unix.c.
4797 For VMS, there's a currently-empty rand_vms.c.
4800 *) Move OCSP client related routines to ocsp_cl.c. These
4801 provide utility functions which an application needing
4802 to issue a request to an OCSP responder and analyse the
4803 response will typically need: as opposed to those which an
4804 OCSP responder itself would need which will be added later.
4806 OCSP_request_sign() signs an OCSP request with an API similar
4807 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4808 response. OCSP_response_get1_basic() extracts basic response
4809 from response. OCSP_resp_find_status(): finds and extracts status
4810 information from an OCSP_CERTID structure (which will be created
4811 when the request structure is built). These are built from lower
4812 level functions which work on OCSP_SINGLERESP structures but
4813 wont normally be used unless the application wishes to examine
4814 extensions in the OCSP response for example.
4816 Replace nonce routines with a pair of functions.
4817 OCSP_request_add1_nonce() adds a nonce value and optionally
4818 generates a random value. OCSP_check_nonce() checks the
4819 validity of the nonce in an OCSP response.
4822 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4823 This doesn't copy the supplied OCSP_CERTID and avoids the
4824 need to free up the newly created id. Change return type
4825 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4826 This can then be used to add extensions to the request.
4827 Deleted OCSP_request_new(), since most of its functionality
4828 is now in OCSP_REQUEST_new() (and the case insensitive name
4829 clash) apart from the ability to set the request name which
4830 will be added elsewhere.
4833 *) Update OCSP API. Remove obsolete extensions argument from
4834 various functions. Extensions are now handled using the new
4835 OCSP extension code. New simple OCSP HTTP function which
4836 can be used to send requests and parse the response.
4839 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4840 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4841 uses the special reorder version of SET OF to sort the attributes
4842 and reorder them to match the encoded order. This resolves a long
4843 standing problem: a verify on a PKCS7 structure just after signing
4844 it used to fail because the attribute order did not match the
4845 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4846 it uses the received order. This is necessary to tolerate some broken
4847 software that does not order SET OF. This is handled by encoding
4848 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4849 to produce the required SET OF.
4852 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4853 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4854 files to get correct declarations of the ASN.1 item variables.
4857 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4858 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4859 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4860 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4861 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4862 ASN1_ITEM and no wrapper functions.
4865 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4866 replace the old function pointer based I/O routines. Change most of
4867 the *_d2i_bio() and *_d2i_fp() functions to use these.
4870 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4871 lines, recognice more "algorithms" that can be deselected, and make
4872 it complain about algorithm deselection that isn't recognised.
4875 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4876 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4877 to use new functions. Add NO_ASN1_OLD which can be set to remove
4878 some old style ASN1 functions: this can be used to determine if old
4879 code will still work when these eventually go away.
4882 *) New extension functions for OCSP structures, these follow the
4883 same conventions as certificates and CRLs.
4886 *) New function X509V3_add1_i2d(). This automatically encodes and
4887 adds an extension. Its behaviour can be customised with various
4888 flags to append, replace or delete. Various wrappers added for
4889 certifcates and CRLs.
4892 *) Fix to avoid calling the underlying ASN1 print routine when
4893 an extension cannot be parsed. Correct a typo in the
4894 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4897 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4898 entries for variables.
4901 *) Add functionality to apps/openssl.c for detecting locking
4902 problems: As the program is single-threaded, all we have
4903 to do is register a locking callback using an array for
4904 storing which locks are currently held by the program.
4907 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4908 SSL_get_ex_data_X509_STORE_idx(), which is used in
4909 ssl_verify_cert_chain() and thus can be called at any time
4910 during TLS/SSL handshakes so that thread-safety is essential.
4911 Unfortunately, the ex_data design is not at all suited
4912 for multi-threaded use, so it probably should be abolished.
4915 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4916 [Broadcom, tweaked and integrated by Geoff Thorpe]
4918 *) Move common extension printing code to new function
4919 X509V3_print_extensions(). Reorganise OCSP print routines and
4920 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4923 *) New function X509_signature_print() to remove duplication in some
4927 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4928 set (this was treated exactly the same as SET OF previously). This
4929 is used to reorder the STACK representing the structure to match the
4930 encoding. This will be used to get round a problem where a PKCS7
4931 structure which was signed could not be verified because the STACK
4932 order did not reflect the encoded order.
4935 *) Reimplement the OCSP ASN1 module using the new code.
4938 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4939 for its ASN1 operations. The old style function pointers still exist
4940 for now but they will eventually go away.
4943 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4944 completely replaces the old ASN1 functionality with a table driven
4945 encoder and decoder which interprets an ASN1_ITEM structure describing
4946 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4947 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4948 has also been converted to the new form.
4951 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4952 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4953 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4954 for negative moduli.
4957 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4958 of not touching the result's sign bit.
4961 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4965 *) Changed the LHASH code to use prototypes for callbacks, and created
4966 macros to declare and implement thin (optionally static) functions
4967 that provide type-safety and avoid function pointer casting for the
4968 type-specific callbacks.
4971 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4973 [Veers Staats <staatsvr@asc.hpc.mil>,
4974 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4976 *) Reformat the FAQ so the different questions and answers can be divided
4977 in sections depending on the subject.
4980 *) Have the zlib compression code load ZLIB.DLL dynamically under
4984 *) New function BN_mod_sqrt for computing square roots modulo a prime
4985 (using the probabilistic Tonelli-Shanks algorithm unless
4986 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4987 be handled deterministically).
4988 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4990 *) Make BN_mod_inverse faster by explicitly handling small quotients
4991 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4992 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4995 *) New function BN_kronecker.
4998 *) Fix BN_gcd so that it works on negative inputs; the result is
4999 positive unless both parameters are zero.
5000 Previously something reasonably close to an infinite loop was
5001 possible because numbers could be growing instead of shrinking
5002 in the implementation of Euclid's algorithm.
5005 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5006 sign of the number in question.
5008 Fix BN_is_word(a,w) to work correctly for w == 0.
5010 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5011 because its test if the absolute value of 'a' equals 'w'.
5012 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5013 it exists mostly for use in the implementations of BN_is_zero(),
5014 BN_is_one(), and BN_is_word().
5017 *) New function BN_swap.
5020 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5021 the exponentiation functions are more likely to produce reasonable
5022 results on negative inputs.
5025 *) Change BN_mod_mul so that the result is always non-negative.
5026 Previously, it could be negative if one of the factors was negative;
5027 I don't think anyone really wanted that behaviour.
5030 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5031 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5032 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5033 and add new functions:
5042 BN_mod_lshift1_quick
5046 These functions always generate non-negative results.
5048 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5049 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5051 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5052 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5053 be reduced modulo m.
5054 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5057 The following entry accidentily appeared in the CHANGES file
5058 distributed with OpenSSL 0.9.7. The modifications described in
5059 it do *not* apply to OpenSSL 0.9.7.
5061 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5062 was actually never needed) and in BN_mul(). The removal in BN_mul()
5063 required a small change in bn_mul_part_recursive() and the addition
5064 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5065 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5066 bn_sub_words() and bn_add_words() except they take arrays with
5071 *) In 'openssl passwd', verify passwords read from the terminal
5072 unless the '-salt' option is used (which usually means that
5073 verification would just waste user's time since the resulting
5074 hash is going to be compared with some given password hash)
5075 or the new '-noverify' option is used.
5077 This is an incompatible change, but it does not affect
5078 non-interactive use of 'openssl passwd' (passwords on the command
5079 line, '-stdin' option, '-in ...' option) and thus should not
5083 *) Remove all references to RSAref, since there's no more need for it.
5086 *) Make DSO load along a path given through an environment variable
5087 (SHLIB_PATH) with shl_load().
5090 *) Constify the ENGINE code as a result of BIGNUM constification.
5091 Also constify the RSA code and most things related to it. In a
5092 few places, most notable in the depth of the ASN.1 code, ugly
5093 casts back to non-const were required (to be solved at a later
5097 *) Make it so the openssl application has all engines loaded by default.
5100 *) Constify the BIGNUM routines a little more.
5103 *) Add the following functions:
5105 ENGINE_load_cswift()
5107 ENGINE_load_atalla()
5109 ENGINE_load_builtin_engines()
5111 That way, an application can itself choose if external engines that
5112 are built-in in OpenSSL shall ever be used or not. The benefit is
5113 that applications won't have to be linked with libdl or other dso
5114 libraries unless it's really needed.
5116 Changed 'openssl engine' to load all engines on demand.
5117 Changed the engine header files to avoid the duplication of some
5118 declarations (they differed!).
5121 *) 'openssl engine' can now list capabilities.
5124 *) Better error reporting in 'openssl engine'.
5127 *) Never call load_dh_param(NULL) in s_server.
5130 *) Add engine application. It can currently list engines by name and
5131 identity, and test if they are actually available.
5134 *) Improve RPM specification file by forcing symbolic linking and making
5135 sure the installed documentation is also owned by root.root.
5136 [Damien Miller <djm@mindrot.org>]
5138 *) Give the OpenSSL applications more possibilities to make use of
5139 keys (public as well as private) handled by engines.
5142 *) Add OCSP code that comes from CertCo.
5145 *) Add VMS support for the Rijndael code.
5148 *) Added untested support for Nuron crypto accelerator.
5151 *) Add support for external cryptographic devices. This code was
5152 previously distributed separately as the "engine" branch.
5153 [Geoff Thorpe, Richard Levitte]
5155 *) Rework the filename-translation in the DSO code. It is now possible to
5156 have far greater control over how a "name" is turned into a filename
5157 depending on the operating environment and any oddities about the
5158 different shared library filenames on each system.
5161 *) Support threads on FreeBSD-elf in Configure.
5164 *) Fix for SHA1 assembly problem with MASM: it produces
5165 warnings about corrupt line number information when assembling
5166 with debugging information. This is caused by the overlapping
5168 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5171 NCONF_get_number() has no error checking at all. As a replacement,
5172 NCONF_get_number_e() is defined (_e for "error checking") and is
5173 promoted strongly. The old NCONF_get_number is kept around for
5174 binary backward compatibility.
5175 Make it possible for methods to load from something other than a BIO,
5176 by providing a function pointer that is given a name instead of a BIO.
5177 For example, this could be used to load configuration data from an
5181 *) Fix for non blocking accept BIOs. Added new I/O special reason
5182 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5183 with non blocking I/O was not possible because no retry code was
5184 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5188 *) Added the beginnings of Rijndael support.
5191 *) Fix for bug in DirectoryString mask setting. Add support for
5192 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5193 to allow certificate printing to more controllable, additional
5194 'certopt' option to 'x509' to allow new printing options to be
5198 *) Clean old EAY MD5 hack from e_os.h.
5201 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5203 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5204 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5205 [Joe Orton, Steve Henson]
5207 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5209 *) Fix additional bug revealed by the NISCC test suite:
5211 Stop bug triggering large recursion when presented with
5212 certain ASN.1 tags (CVE-2003-0851)
5215 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5217 *) Fix various bugs revealed by running the NISCC test suite:
5219 Stop out of bounds reads in the ASN1 code when presented with
5220 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5222 If verify callback ignores invalid public key errors don't try to check
5223 certificate signature with the NULL public key.
5227 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5228 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5232 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5233 extra data after the compression methods not only for TLS 1.0
5234 but also for SSL 3.0 (as required by the specification).
5235 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5237 *) Change X509_certificate_type() to mark the key as exported/exportable
5238 when it's 512 *bits* long, not 512 bytes.
5241 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5243 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5244 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5245 a protocol version number mismatch like a decryption error
5246 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5249 *) Turn on RSA blinding by default in the default implementation
5250 to avoid a timing attack. Applications that don't want it can call
5251 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5252 They would be ill-advised to do so in most cases.
5253 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5255 *) Change RSA blinding code so that it works when the PRNG is not
5256 seeded (in this case, the secret RSA exponent is abused as
5257 an unpredictable seed -- if it is not unpredictable, there
5258 is no point in blinding anyway). Make RSA blinding thread-safe
5259 by remembering the creator's thread ID in rsa->blinding and
5260 having all other threads use local one-time blinding factors
5261 (this requires more computation than sharing rsa->blinding, but
5262 avoids excessive locking; and if an RSA object is not shared
5263 between threads, blinding will still be very fast).
5266 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5268 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5269 via timing by performing a MAC computation even if incorrrect
5270 block cipher padding has been found. This is a countermeasure
5271 against active attacks where the attacker has to distinguish
5272 between bad padding and a MAC verification error. (CVE-2003-0078)
5274 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5275 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5276 Martin Vuagnoux (EPFL, Ilion)]
5278 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5280 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5281 memory from it's contents. This is done with a counter that will
5282 place alternating values in each byte. This can be used to solve
5283 two issues: 1) the removal of calls to memset() by highly optimizing
5284 compilers, and 2) cleansing with other values than 0, since those can
5285 be read through on certain media, for example a swap space on disk.
5288 *) Bugfix: client side session caching did not work with external caching,
5289 because the session->cipher setting was not restored when reloading
5290 from the external cache. This problem was masked, when
5291 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5292 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5295 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5296 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5297 [Zeev Lieber <zeev-l@yahoo.com>]
5299 *) Undo an undocumented change introduced in 0.9.6e which caused
5300 repeated calls to OpenSSL_add_all_ciphers() and
5301 OpenSSL_add_all_digests() to be ignored, even after calling
5305 *) Change the default configuration reader to deal with last line not
5306 being properly terminated.
5309 *) Change X509_NAME_cmp() so it applies the special rules on handling
5310 DN values that are of type PrintableString, as well as RDNs of type
5311 emailAddress where the value has the type ia5String.
5312 [stefank@valicert.com via Richard Levitte]
5314 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5315 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5316 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5317 the bitwise-OR of the two for use by the majority of applications
5318 wanting this behaviour, and update the docs. The documented
5319 behaviour and actual behaviour were inconsistent and had been
5320 changing anyway, so this is more a bug-fix than a behavioural
5322 [Geoff Thorpe, diagnosed by Nadav Har'El]
5324 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5325 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5328 *) Fix initialization code race conditions in
5329 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5330 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5331 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5332 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5333 ssl2_get_cipher_by_char(),
5334 ssl3_get_cipher_by_char().
5335 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5337 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5338 the cached sessions are flushed, as the remove_cb() might use ex_data
5339 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5340 (see [openssl.org #212]).
5341 [Geoff Thorpe, Lutz Jaenicke]
5343 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5344 length, instead of the encoding length to d2i_ASN1_OBJECT.
5347 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5349 *) [In 0.9.6g-engine release:]
5350 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5351 [Lynn Gazis <lgazis@rainbow.com>]
5353 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5355 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5356 and get fix the header length calculation.
5357 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5358 Alon Kantor <alonk@checkpoint.com> (and others),
5361 *) Use proper error handling instead of 'assertions' in buffer
5362 overflow checks added in 0.9.6e. This prevents DoS (the
5363 assertions could call abort()).
5364 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5366 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5368 *) Add various sanity checks to asn1_get_length() to reject
5369 the ASN1 length bytes if they exceed sizeof(long), will appear
5370 negative or the content length exceeds the length of the
5372 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5374 *) Fix cipher selection routines: ciphers without encryption had no flags
5375 for the cipher strength set and where therefore not handled correctly
5376 by the selection routines (PR #130).
5379 *) Fix EVP_dsa_sha macro.
5383 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5384 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5385 that was added in OpenSSL 0.9.6d.
5387 As the countermeasure turned out to be incompatible with some
5388 broken SSL implementations, the new option is part of SSL_OP_ALL.
5389 SSL_OP_ALL is usually employed when compatibility with weird SSL
5390 implementations is desired (e.g. '-bugs' option to 's_client' and
5391 's_server'), so the new option is automatically set in many
5395 *) Changes in security patch:
5397 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5398 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5399 Air Force Materiel Command, USAF, under agreement number
5402 *) Add various sanity checks to asn1_get_length() to reject
5403 the ASN1 length bytes if they exceed sizeof(long), will appear
5404 negative or the content length exceeds the length of the
5405 supplied buffer. (CVE-2002-0659)
5406 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5408 *) Assertions for various potential buffer overflows, not known to
5410 [Ben Laurie (CHATS)]
5412 *) Various temporary buffers to hold ASCII versions of integers were
5413 too small for 64 bit platforms. (CVE-2002-0655)
5414 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5416 *) Remote buffer overflow in SSL3 protocol - an attacker could
5417 supply an oversized session ID to a client. (CVE-2002-0656)
5418 [Ben Laurie (CHATS)]
5420 *) Remote buffer overflow in SSL2 protocol - an attacker could
5421 supply an oversized client master key. (CVE-2002-0656)
5422 [Ben Laurie (CHATS)]
5424 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5426 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5427 encoded as NULL) with id-dsa-with-sha1.
5428 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5430 *) Check various X509_...() return values in apps/req.c.
5431 [Nils Larsch <nla@trustcenter.de>]
5433 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5434 an end-of-file condition would erronously be flagged, when the CRLF
5435 was just at the end of a processed block. The bug was discovered when
5436 processing data through a buffering memory BIO handing the data to a
5437 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5438 <ptsekov@syntrex.com> and Nedelcho Stanev.
5441 *) Implement a countermeasure against a vulnerability recently found
5442 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5443 before application data chunks to avoid the use of known IVs
5444 with data potentially chosen by the attacker.
5447 *) Fix length checks in ssl3_get_client_hello().
5450 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5451 to prevent ssl3_read_internal() from incorrectly assuming that
5452 ssl3_read_bytes() found application data while handshake
5453 processing was enabled when in fact s->s3->in_read_app_data was
5454 merely automatically cleared during the initial handshake.
5455 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5457 *) Fix object definitions for Private and Enterprise: they were not
5458 recognized in their shortname (=lowercase) representation. Extend
5459 obj_dat.pl to issue an error when using undefined keywords instead
5460 of silently ignoring the problem (Svenning Sorensen
5461 <sss@sss.dnsalias.net>).
5464 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5465 generators, i.e. generators other than 2 and 5. (Previously, the
5466 code did not properly initialise the 'add' and 'rem' values to
5467 BN_generate_prime().)
5469 In the new general case, we do not insist that 'generator' is
5470 actually a primitive root: This requirement is rather pointless;
5471 a generator of the order-q subgroup is just as good, if not
5475 *) Map new X509 verification errors to alerts. Discovered and submitted by
5476 Tom Wu <tom@arcot.com>.
5479 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5480 returning non-zero before the data has been completely received
5481 when using non-blocking I/O.
5482 [Bodo Moeller; problem pointed out by John Hughes]
5484 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5485 [Ben Laurie, Lutz Jaenicke]
5487 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5488 Yoram Zahavi <YoramZ@gilian.com>).
5491 *) Add information about CygWin 1.3 and on, and preserve proper
5492 configuration for the versions before that.
5493 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5495 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5496 check whether we deal with a copy of a session and do not delete from
5497 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5498 <izhar@checkpoint.com>.
5501 *) Do not store session data into the internal session cache, if it
5502 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5503 flag is set). Proposed by Aslam <aslam@funk.com>.
5506 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5510 *) [In 0.9.6d-engine release:]
5511 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5512 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5514 *) Add the configuration target linux-s390x.
5515 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5517 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5518 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5519 variable as an indication that a ClientHello message has been
5520 received. As the flag value will be lost between multiple
5521 invocations of ssl3_accept when using non-blocking I/O, the
5522 function may not be aware that a handshake has actually taken
5523 place, thus preventing a new session from being added to the
5526 To avoid this problem, we now set s->new_session to 2 instead of
5527 using a local variable.
5528 [Lutz Jaenicke, Bodo Moeller]
5530 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5531 if the SSL_R_LENGTH_MISMATCH error is detected.
5532 [Geoff Thorpe, Bodo Moeller]
5534 *) New 'shared_ldflag' column in Configure platform table.
5537 *) Fix EVP_CIPHER_mode macro.
5538 ["Dan S. Camper" <dan@bti.net>]
5540 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5541 type, we must throw them away by setting rr->length to 0.
5542 [D P Chang <dpc@qualys.com>]
5544 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5546 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5547 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5548 worked incorrectly for those cases where range = 10..._2 and
5549 3*range is two bits longer than range.)
5552 *) Only add signing time to PKCS7 structures if it is not already
5556 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5557 OBJ_ld_ce should be OBJ_id_ce.
5558 Also some ip-pda OIDs in crypto/objects/objects.txt were
5559 incorrect (cf. RFC 3039).
5560 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5562 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5563 returns early because it has nothing to do.
5564 [Andy Schneider <andy.schneider@bjss.co.uk>]
5566 *) [In 0.9.6c-engine release:]
5567 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5568 [Andy Schneider <andy.schneider@bjss.co.uk>]
5570 *) [In 0.9.6c-engine release:]
5571 Add support for Cryptographic Appliance's keyserver technology.
5572 (Use engine 'keyclient')
5573 [Cryptographic Appliances and Geoff Thorpe]
5575 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5576 is called via tools/c89.sh because arguments have to be
5577 rearranged (all '-L' options must appear before the first object
5579 [Richard Shapiro <rshapiro@abinitio.com>]
5581 *) [In 0.9.6c-engine release:]
5582 Add support for Broadcom crypto accelerator cards, backported
5584 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5586 *) [In 0.9.6c-engine release:]
5587 Add support for SureWare crypto accelerator cards from
5588 Baltimore Technologies. (Use engine 'sureware')
5589 [Baltimore Technologies and Mark Cox]
5591 *) [In 0.9.6c-engine release:]
5592 Add support for crypto accelerator cards from Accelerated
5593 Encryption Processing, www.aep.ie. (Use engine 'aep')
5594 [AEP Inc. and Mark Cox]
5596 *) Add a configuration entry for gcc on UnixWare.
5597 [Gary Benson <gbenson@redhat.com>]
5599 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5600 messages are stored in a single piece (fixed-length part and
5601 variable-length part combined) and fix various bugs found on the way.
5604 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5605 instead. BIO_gethostbyname() does not know what timeouts are
5606 appropriate, so entries would stay in cache even when they have
5608 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5610 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5611 faced with a pathologically small ClientHello fragment that does
5612 not contain client_version: Instead of aborting with an error,
5613 simply choose the highest available protocol version (i.e.,
5614 TLS 1.0 unless it is disabled). In practice, ClientHello
5615 messages are never sent like this, but this change gives us
5616 strictly correct behaviour at least for TLS.
5619 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5620 never resets s->method to s->ctx->method when called from within
5621 one of the SSL handshake functions.
5622 [Bodo Moeller; problem pointed out by Niko Baric]
5624 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5625 (sent using the client's version number) if client_version is
5626 smaller than the protocol version in use. Also change
5627 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5628 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5629 the client will at least see that alert.
5632 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5636 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5637 client receives HelloRequest while in a handshake.
5638 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5640 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5641 should end in 'break', not 'goto end' which circuments various
5642 cleanups done in state SSL_ST_OK. But session related stuff
5643 must be disabled for SSL_ST_OK in the case that we just sent a
5646 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5647 before just sending a HelloRequest.
5648 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5650 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5651 reveal whether illegal block cipher padding was found or a MAC
5652 verification error occured. (Neither SSLerr() codes nor alerts
5653 are directly visible to potential attackers, but the information
5654 may leak via logfiles.)
5656 Similar changes are not required for the SSL 2.0 implementation
5657 because the number of padding bytes is sent in clear for SSL 2.0,
5658 and the extra bytes are just ignored. However ssl/s2_pkt.c
5659 failed to verify that the purported number of padding bytes is in
5663 *) Add OpenUNIX-8 support including shared libraries
5664 (Boyd Lynn Gerber <gerberb@zenez.com>).
5667 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5668 'wristwatch attack' using huge encoding parameters (cf.
5669 James H. Manger's CRYPTO 2001 paper). Note that the
5670 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5671 encoding parameters and hence was not vulnerable.
5674 *) BN_sqr() bug fix.
5675 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5677 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5678 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5679 followed by modular reduction.
5680 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5682 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5683 equivalent based on BN_pseudo_rand() instead of BN_rand().
5686 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5687 This function was broken, as the check for a new client hello message
5688 to handle SGC did not allow these large messages.
5689 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5692 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5695 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5696 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5699 *) Rework the configuration and shared library support for Tru64 Unix.
5700 The configuration part makes use of modern compiler features and
5701 still retains old compiler behavior for those that run older versions
5702 of the OS. The shared library support part includes a variant that
5703 uses the RPATH feature, and is available through the special
5704 configuration target "alpha-cc-rpath", which will never be selected
5706 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5708 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5709 with the same message size as in ssl3_get_certificate_request().
5710 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5711 messages might inadvertently be reject as too long.
5712 [Petr Lampa <lampa@fee.vutbr.cz>]
5714 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5717 *) Modified SSL library such that the verify_callback that has been set
5718 specificly for an SSL object with SSL_set_verify() is actually being
5719 used. Before the change, a verify_callback set with this function was
5720 ignored and the verify_callback() set in the SSL_CTX at the time of
5721 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5722 to allow the necessary settings.
5725 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5726 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5727 done automatically (in contradiction to the requirements of the C
5728 standard). This made problems when used from OpenSSH.
5731 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5732 dh->length and always used
5734 BN_rand_range(priv_key, dh->p).
5736 BN_rand_range() is not necessary for Diffie-Hellman, and this
5737 specific range makes Diffie-Hellman unnecessarily inefficient if
5738 dh->length (recommended exponent length) is much smaller than the
5739 length of dh->p. We could use BN_rand_range() if the order of
5740 the subgroup was stored in the DH structure, but we only have
5745 BN_rand(priv_key, l, ...)
5747 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5753 RSA_eay_public_encrypt
5754 RSA_eay_private_decrypt
5755 RSA_eay_private_encrypt (signing)
5756 RSA_eay_public_decrypt (signature verification)
5758 (default implementations for RSA_public_encrypt,
5759 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5760 always reject numbers >= n.
5763 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5764 to synchronize access to 'locking_thread'. This is necessary on
5765 systems where access to 'locking_thread' (an 'unsigned long'
5766 variable) is not atomic.
5769 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5770 *before* setting the 'crypto_lock_rand' flag. The previous code had
5771 a race condition if 0 is a valid thread ID.
5772 [Travis Vitek <vitek@roguewave.com>]
5774 *) Add support for shared libraries under Irix.
5775 [Albert Chin-A-Young <china@thewrittenword.com>]
5777 *) Add configuration option to build on Linux on both big-endian and
5779 [Ralf Baechle <ralf@uni-koblenz.de>]
5781 *) Add the possibility to create shared libraries on HP-UX.
5784 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5786 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5787 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5788 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5789 PRNG state recovery was possible based on the output of
5790 one PRNG request appropriately sized to gain knowledge on
5791 'md' followed by enough consecutive 1-byte PRNG requests
5792 to traverse all of 'state'.
5794 1. When updating 'md_local' (the current thread's copy of 'md')
5795 during PRNG output generation, hash all of the previous
5796 'md_local' value, not just the half used for PRNG output.
5798 2. Make the number of bytes from 'state' included into the hash
5799 independent from the number of PRNG bytes requested.
5801 The first measure alone would be sufficient to avoid
5802 Markku-Juhani's attack. (Actually it had never occurred
5803 to me that the half of 'md_local' used for chaining was the
5804 half from which PRNG output bytes were taken -- I had always
5805 assumed that the secret half would be used.) The second
5806 measure makes sure that additional data from 'state' is never
5807 mixed into 'md_local' in small portions; this heuristically
5808 further strengthens the PRNG.
5811 *) Fix crypto/bn/asm/mips3.s.
5814 *) When only the key is given to "enc", the IV is undefined. Print out
5815 an error message in this case.
5818 *) Handle special case when X509_NAME is empty in X509 printing routines.
5821 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5822 positive and less than q.
5825 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5826 used: it isn't thread safe and the add_lock_callback should handle
5828 [Paul Rose <Paul.Rose@bridge.com>]
5830 *) Verify that incoming data obeys the block size in
5831 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5835 [Ulf Möller, Bodo Möller]
5837 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5838 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5839 when fixing the server behaviour for backwards-compatible 'client
5840 hello' messages. (Note that the attack is impractical against
5841 SSL 3.0 and TLS 1.0 anyway because length and version checking
5842 means that the probability of guessing a valid ciphertext is
5843 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5846 Before 0.9.5, the countermeasure (hide the error by generating a
5847 random 'decryption result') did not work properly because
5848 ERR_clear_error() was missing, meaning that SSL_get_error() would
5849 detect the supposedly ignored error.
5851 Both problems are now fixed.
5854 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5855 (previously it was 1024).
5858 *) Fix for compatibility mode trust settings: ignore trust settings
5859 unless some valid trust or reject settings are present.
5862 *) Fix for blowfish EVP: its a variable length cipher.
5865 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5866 parameters in DSA public key structures and return an error in the
5867 DSA routines if parameters are absent.
5870 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5871 in the current directory if neither $RANDFILE nor $HOME was set.
5872 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5873 caused some confusion to Windows users who haven't defined $HOME.
5874 Thus RAND_file_name() is changed again: e_os.h can define a
5875 DEFAULT_HOME, which will be used if $HOME is not set.
5876 For Windows, we use "C:"; on other platforms, we still require
5877 environment variables.
5879 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5880 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5881 having multiple threads call RAND_poll() concurrently.
5884 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5885 combination of a flag and a thread ID variable.
5886 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5887 flag), *other* threads can enter ssleay_add_bytes without obeying
5888 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5889 that they do not hold after the first thread unsets add_do_not_lock).
5892 *) Change bctest again: '-x' expressions are not available in all
5896 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5898 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5899 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5901 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5902 the default extension for executables, if any. Also, make the perl
5903 scripts that use symlink() to test if it really exists and use "cp"
5904 if it doesn't. All this made OpenSSL compilable and installable in
5908 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5909 If SEQUENCE is length is indefinite just set c->slen to the total
5910 amount of data available.
5911 [Steve Henson, reported by shige@FreeBSD.org]
5912 [This change does not apply to 0.9.7.]
5914 *) Change bctest to avoid here-documents inside command substitution
5915 (workaround for FreeBSD /bin/sh bug).
5916 For compatibility with Ultrix, avoid shell functions (introduced
5917 in the bctest version that searches along $PATH).
5920 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5921 with des_encrypt() defined on some operating systems, like Solaris
5925 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5926 On the Importance of Eliminating Errors in Cryptographic
5927 Computations, J. Cryptology 14 (2001) 2, 101-119,
5928 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5931 *) MIPS assembler BIGNUM division bug fix.
5934 *) Disabled incorrect Alpha assembler code.
5937 *) Fix PKCS#7 decode routines so they correctly update the length
5938 after reading an EOC for the EXPLICIT tag.
5940 [This change does not apply to 0.9.7.]
5942 *) Fix bug in PKCS#12 key generation routines. This was triggered
5943 if a 3DES key was generated with a 0 initial byte. Include
5944 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5945 (but broken) behaviour.
5948 *) Enhance bctest to search for a working bc along $PATH and print
5950 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5952 *) Fix memory leaks in err.c: free err_data string if necessary;
5953 don't write to the wrong index in ERR_set_error_data.
5956 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5960 *) Replace rdtsc with _emit statements for VC++ version 5.
5961 [Jeremy Cooper <jeremy@baymoo.org>]
5963 *) Make it possible to reuse SSLv2 sessions.
5966 *) In copy_email() check for >= 0 as a return value for
5967 X509_NAME_get_index_by_NID() since 0 is a valid index.
5968 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5970 *) Avoid coredump with unsupported or invalid public keys by checking if
5971 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5972 PKCS7_verify() fails with non detached data.
5975 *) Don't use getenv in library functions when run as setuid/setgid.
5976 New function OPENSSL_issetugid().
5979 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5980 due to incorrect handling of multi-threading:
5982 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5984 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5986 3. Count how many times MemCheck_off() has been called so that
5987 nested use can be treated correctly. This also avoids
5988 inband-signalling in the previous code (which relied on the
5989 assumption that thread ID 0 is impossible).
5992 *) Add "-rand" option also to s_client and s_server.
5995 *) Fix CPU detection on Irix 6.x.
5996 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5997 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5999 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6002 [This change does not apply to 0.9.7.]
6004 *) Use the cached encoding of an X509_NAME structure rather than
6005 copying it. This is apparently the reason for the libsafe "errors"
6006 but the code is actually correct.
6009 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6010 Bleichenbacher's DSA attack.
6011 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6012 to be set and top=0 forces the highest bit to be set; top=-1 is new
6013 and leaves the highest bit random.
6014 [Ulf Moeller, Bodo Moeller]
6016 *) In the NCONF_...-based implementations for CONF_... queries
6017 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6018 a temporary CONF structure with the data component set to NULL
6019 (which gives segmentation faults in lh_retrieve).
6020 Instead, use NULL for the CONF pointer in CONF_get_string and
6021 CONF_get_number (which may use environment variables) and directly
6022 return NULL from CONF_get_section.
6025 *) Fix potential buffer overrun for EBCDIC.
6028 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6029 keyUsage if basicConstraints absent for a CA.
6032 *) Make SMIME_write_PKCS7() write mail header values with a format that
6033 is more generally accepted (no spaces before the semicolon), since
6034 some programs can't parse those values properly otherwise. Also make
6035 sure BIO's that break lines after each write do not create invalid
6039 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6040 macros previously used would not encode an empty SEQUENCE OF
6041 and break the signature.
6043 [This change does not apply to 0.9.7.]
6045 *) Zero the premaster secret after deriving the master secret in
6049 *) Add some EVP_add_digest_alias registrations (as found in
6050 OpenSSL_add_all_digests()) to SSL_library_init()
6051 aka OpenSSL_add_ssl_algorithms(). This provides improved
6052 compatibility with peers using X.509 certificates
6053 with unconventional AlgorithmIdentifier OIDs.
6056 *) Fix for Irix with NO_ASM.
6057 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6059 *) ./config script fixes.
6060 [Ulf Moeller, Richard Levitte]
6062 *) Fix 'openssl passwd -1'.
6065 *) Change PKCS12_key_gen_asc() so it can cope with non null
6066 terminated strings whose length is passed in the passlen
6067 parameter, for example from PEM callbacks. This was done
6068 by adding an extra length parameter to asc2uni().
6069 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6071 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6072 call failed, free the DSA structure.
6075 *) Fix to uni2asc() to cope with zero length Unicode strings.
6076 These are present in some PKCS#12 files.
6079 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6080 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6081 when writing a 32767 byte record.
6082 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6084 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6085 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6087 (RSA objects have a reference count access to which is protected
6088 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6089 so they are meant to be shared between threads.)
6090 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6091 "Reddie, Steven" <Steven.Reddie@ca.com>]
6093 *) Fix a deadlock in CRYPTO_mem_leaks().
6096 *) Use better test patterns in bntest.
6099 *) rand_win.c fix for Borland C.
6102 *) BN_rshift bugfix for n == 0.
6105 *) Add a 'bctest' script that checks for some known 'bc' bugs
6106 so that 'make test' does not abort just because 'bc' is broken.
6109 *) Store verify_result within SSL_SESSION also for client side to
6110 avoid potential security hole. (Re-used sessions on the client side
6111 always resulted in verify_result==X509_V_OK, not using the original
6112 result of the server certificate verification.)
6115 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6116 SSL3_RT_APPLICATION_DATA, return 0.
6117 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6121 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6122 releases, have been re-implemented by renaming the previous
6123 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6124 and ssl3_read_internal, respectively, and adding 'peek' parameters
6125 to them. The new ssl[23]_{read,peek} functions are calls to
6126 ssl[23]_read_internal with the 'peek' flag set appropriately.
6127 A 'peek' parameter has also been added to ssl3_read_bytes, which
6128 does the actual work for ssl3_read_internal.
6131 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6132 the method-specific "init()" handler. Also clean up ex_data after
6133 calling the method-specific "finish()" handler. Previously, this was
6134 happening the other way round.
6137 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6138 The previous value, 12, was not always sufficient for BN_mod_exp().
6141 *) Make sure that shared libraries get the internal name engine with
6142 the full version number and not just 0. This should mark the
6143 shared libraries as not backward compatible. Of course, this should
6144 be changed again when we can guarantee backward binary compatibility.
6147 *) Fix typo in get_cert_by_subject() in by_dir.c
6148 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6150 *) Rework the system to generate shared libraries:
6152 - Make note of the expected extension for the shared libraries and
6153 if there is a need for symbolic links from for example libcrypto.so.0
6154 to libcrypto.so.0.9.7. There is extended info in Configure for
6157 - Make as few rebuilds of the shared libraries as possible.
6159 - Still avoid linking the OpenSSL programs with the shared libraries.
6161 - When installing, install the shared libraries separately from the
6165 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6167 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6168 and not in SSL_clear because the latter is also used by the
6169 accept/connect functions; previously, the settings made by
6170 SSL_set_read_ahead would be lost during the handshake.
6171 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6173 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6174 Previously, it would create entries for disableed algorithms no
6178 *) Added several new manual pages for SSL_* function.
6181 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6183 *) In ssl23_get_client_hello, generate an error message when faced
6184 with an initial SSL 3.0/TLS record that is too small to contain the
6185 first two bytes of the ClientHello message, i.e. client_version.
6186 (Note that this is a pathologic case that probably has never happened
6187 in real life.) The previous approach was to use the version number
6188 from the record header as a substitute; but our protocol choice
6189 should not depend on that one because it is not authenticated
6190 by the Finished messages.
6193 *) More robust randomness gathering functions for Windows.
6194 [Jeffrey Altman <jaltman@columbia.edu>]
6196 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6197 not set then we don't setup the error code for issuer check errors
6198 to avoid possibly overwriting other errors which the callback does
6199 handle. If an application does set the flag then we assume it knows
6200 what it is doing and can handle the new informational codes
6204 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6205 a general "ANY" type, as such it should be able to decode anything
6206 including tagged types. However it didn't check the class so it would
6207 wrongly interpret tagged types in the same way as their universal
6208 counterpart and unknown types were just rejected. Changed so that the
6209 tagged and unknown types are handled in the same way as a SEQUENCE:
6210 that is the encoding is stored intact. There is also a new type
6211 "V_ASN1_OTHER" which is used when the class is not universal, in this
6212 case we have no idea what the actual type is so we just lump them all
6216 *) On VMS, stdout may very well lead to a file that is written to
6217 in a record-oriented fashion. That means that every write() will
6218 write a separate record, which will be read separately by the
6219 programs trying to read from it. This can be very confusing.
6221 The solution is to put a BIO filter in the way that will buffer
6222 text until a linefeed is reached, and then write everything a
6223 line at a time, so every record written will be an actual line,
6224 not chunks of lines and not (usually doesn't happen, but I've
6225 seen it once) several lines in one record. BIO_f_linebuffer() is
6228 Currently, it's a VMS-only method, because that's where it has
6229 been tested well enough.
6232 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6233 it can return incorrect results.
6234 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6235 but it was in 0.9.6-beta[12].)
6238 *) Disable the check for content being present when verifying detached
6239 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6240 include zero length content when signing messages.
6243 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6244 BIO_ctrl (for BIO pairs).
6247 *) Add DSO method for VMS.
6250 *) Bug fix: Montgomery multiplication could produce results with the
6254 *) Add RPM specification openssl.spec and modify it to build three
6255 packages. The default package contains applications, application
6256 documentation and run-time libraries. The devel package contains
6257 include files, static libraries and function documentation. The
6258 doc package contains the contents of the doc directory. The original
6259 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6262 *) Add a large number of documentation files for many SSL routines.
6263 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6265 *) Add a configuration entry for Sony News 4.
6266 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6268 *) Don't set the two most significant bits to one when generating a
6269 random number < q in the DSA library.
6272 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6273 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6274 the underlying transport is blocking) if a handshake took place.
6275 (The default behaviour is needed by applications such as s_client
6276 and s_server that use select() to determine when to use SSL_read;
6277 but for applications that know in advance when to expect data, it
6278 just makes things more complicated.)
6281 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6285 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6286 work better on such systems.
6287 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6289 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6290 Update PKCS12_parse() so it copies the friendlyName and the
6291 keyid to the certificates aux info.
6294 *) Fix bug in PKCS7_verify() which caused an infinite loop
6295 if there was more than one signature.
6296 [Sven Uszpelkat <su@celocom.de>]
6298 *) Major change in util/mkdef.pl to include extra information
6299 about each symbol, as well as presentig variables as well
6300 as functions. This change means that there's n more need
6301 to rebuild the .num files when some algorithms are excluded.
6304 *) Allow the verify time to be set by an application,
6305 rather than always using the current time.
6308 *) Phase 2 verify code reorganisation. The certificate
6309 verify code now looks up an issuer certificate by a
6310 number of criteria: subject name, authority key id
6311 and key usage. It also verifies self signed certificates
6312 by the same criteria. The main comparison function is
6313 X509_check_issued() which performs these checks.
6315 Lot of changes were necessary in order to support this
6316 without completely rewriting the lookup code.
6318 Authority and subject key identifier are now cached.
6320 The LHASH 'certs' is X509_STORE has now been replaced
6321 by a STACK_OF(X509_OBJECT). This is mainly because an
6322 LHASH can't store or retrieve multiple objects with
6323 the same hash value.
6325 As a result various functions (which were all internal
6326 use only) have changed to handle the new X509_STORE
6327 structure. This will break anything that messed round
6328 with X509_STORE internally.
6330 The functions X509_STORE_add_cert() now checks for an
6331 exact match, rather than just subject name.
6333 The X509_STORE API doesn't directly support the retrieval
6334 of multiple certificates matching a given criteria, however
6335 this can be worked round by performing a lookup first
6336 (which will fill the cache with candidate certificates)
6337 and then examining the cache for matches. This is probably
6338 the best we can do without throwing out X509_LOOKUP
6339 entirely (maybe later...).
6341 The X509_VERIFY_CTX structure has been enhanced considerably.
6343 All certificate lookup operations now go via a get_issuer()
6344 callback. Although this currently uses an X509_STORE it
6345 can be replaced by custom lookups. This is a simple way
6346 to bypass the X509_STORE hackery necessary to make this
6347 work and makes it possible to use more efficient techniques
6348 in future. A very simple version which uses a simple
6349 STACK for its trusted certificate store is also provided
6350 using X509_STORE_CTX_trusted_stack().
6352 The verify_cb() and verify() callbacks now have equivalents
6353 in the X509_STORE_CTX structure.
6355 X509_STORE_CTX also has a 'flags' field which can be used
6356 to customise the verify behaviour.
6359 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6360 excludes S/MIME capabilities.
6363 *) When a certificate request is read in keep a copy of the
6364 original encoding of the signed data and use it when outputing
6365 again. Signatures then use the original encoding rather than
6366 a decoded, encoded version which may cause problems if the
6367 request is improperly encoded.
6370 *) For consistency with other BIO_puts implementations, call
6371 buffer_write(b, ...) directly in buffer_puts instead of calling
6374 In BIO_puts, increment b->num_write as in BIO_write.
6375 [Peter.Sylvester@EdelWeb.fr]
6377 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6378 BN_zero, we may not return a BIGNUM with an array consisting of
6382 *) Avoid calling abort() from within the library when problems are
6383 detected, except if preprocessor symbols have been defined
6384 (such as REF_CHECK, BN_DEBUG etc.).
6387 *) New openssl application 'rsautl'. This utility can be
6388 used for low level RSA operations. DER public key
6389 BIO/fp routines also added.
6392 *) New Configure entry and patches for compiling on QNX 4.
6393 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6395 *) A demo state-machine implementation was sponsored by
6396 Nuron (http://www.nuron.com/) and is now available in
6397 demos/state_machine.
6400 *) New options added to the 'dgst' utility for signature
6401 generation and verification.
6404 *) Unrecognized PKCS#7 content types are now handled via a
6405 catch all ASN1_TYPE structure. This allows unsupported
6406 types to be stored as a "blob" and an application can
6407 encode and decode it manually.
6410 *) Fix various signed/unsigned issues to make a_strex.c
6412 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6414 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6415 length if passed a buffer. ASN1_INTEGER_to_BN failed
6416 if passed a NULL BN and its argument was negative.
6417 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6419 *) Modification to PKCS#7 encoding routines to output definite
6420 length encoding. Since currently the whole structures are in
6421 memory there's not real point in using indefinite length
6422 constructed encoding. However if OpenSSL is compiled with
6423 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6426 *) Added BIO_vprintf() and BIO_vsnprintf().
6429 *) Added more prefixes to parse for in the the strings written
6430 through a logging bio, to cover all the levels that are available
6431 through syslog. The prefixes are now:
6433 PANIC, EMERG, EMR => LOG_EMERG
6434 ALERT, ALR => LOG_ALERT
6435 CRIT, CRI => LOG_CRIT
6436 ERROR, ERR => LOG_ERR
6437 WARNING, WARN, WAR => LOG_WARNING
6438 NOTICE, NOTE, NOT => LOG_NOTICE
6439 INFO, INF => LOG_INFO
6440 DEBUG, DBG => LOG_DEBUG
6442 and as before, if none of those prefixes are present at the
6443 beginning of the string, LOG_ERR is chosen.
6445 On Win32, the LOG_* levels are mapped according to this:
6447 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6448 LOG_WARNING => EVENTLOG_WARNING_TYPE
6449 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6453 *) Made it possible to reconfigure with just the configuration
6454 argument "reconf" or "reconfigure". The command line arguments
6455 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6456 and are retrieved from there when reconfiguring.
6460 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6462 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6465 *) The obj_dat.pl script was messing up the sorting of object
6466 names. The reason was that it compared the quoted version
6467 of strings as a result "OCSP" > "OCSP Signing" because
6468 " > SPACE. Changed script to store unquoted versions of
6469 names and add quotes on output. It was also omitting some
6470 names from the lookup table if they were given a default
6471 value (that is if SN is missing it is given the same
6472 value as LN and vice versa), these are now added on the
6473 grounds that if an object has a name we should be able to
6474 look it up. Finally added warning output when duplicate
6475 short or long names are found.
6478 *) Changes needed for Tandem NSK.
6479 [Scott Uroff <scott@xypro.com>]
6481 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6482 RSA_padding_check_SSLv23(), special padding was never detected
6483 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6484 version rollback attacks was not effective.
6486 In s23_clnt.c, don't use special rollback-attack detection padding
6487 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6488 client; similarly, in s23_srvr.c, don't do the rollback check if
6489 SSL 2.0 is the only protocol enabled in the server.
6492 *) Make it possible to get hexdumps of unprintable data with 'openssl
6493 asn1parse'. By implication, the functions ASN1_parse_dump() and
6494 BIO_dump_indent() are added.
6497 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6498 these print out strings and name structures based on various
6499 flags including RFC2253 support and proper handling of
6500 multibyte characters. Added options to the 'x509' utility
6501 to allow the various flags to be set.
6504 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6505 Also change the functions X509_cmp_current_time() and
6506 X509_gmtime_adj() work with an ASN1_TIME structure,
6507 this will enable certificates using GeneralizedTime in validity
6508 dates to be checked.
6511 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6512 negative public key encodings) on by default,
6513 NO_NEG_PUBKEY_BUG can be set to disable it.
6516 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6517 content octets. An i2c_ASN1_OBJECT is unnecessary because
6518 the encoding can be trivially obtained from the structure.
6521 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6522 not read locks (CRYPTO_r_[un]lock).
6525 *) A first attempt at creating official support for shared
6526 libraries through configuration. I've kept it so the
6527 default is static libraries only, and the OpenSSL programs
6528 are always statically linked for now, but there are
6529 preparations for dynamic linking in place.
6530 This has been tested on Linux and Tru64.
6533 *) Randomness polling function for Win9x, as described in:
6534 Peter Gutmann, Software Generation of Practically Strong
6538 *) Fix so PRNG is seeded in req if using an already existing
6542 *) New options to smime application. -inform and -outform
6543 allow alternative formats for the S/MIME message including
6544 PEM and DER. The -content option allows the content to be
6545 specified separately. This should allow things like Netscape
6546 form signing output easier to verify.
6549 *) Fix the ASN1 encoding of tags using the 'long form'.
6552 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6553 STRING types. These convert content octets to and from the
6554 underlying type. The actual tag and length octets are
6555 already assumed to have been read in and checked. These
6556 are needed because all other string types have virtually
6557 identical handling apart from the tag. By having versions
6558 of the ASN1 functions that just operate on content octets
6559 IMPLICIT tagging can be handled properly. It also allows
6560 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6561 and ASN1_INTEGER are identical apart from the tag.
6564 *) Change the handling of OID objects as follows:
6566 - New object identifiers are inserted in objects.txt, following
6567 the syntax given in objects.README.
6568 - objects.pl is used to process obj_mac.num and create a new
6570 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6573 This is currently kind of a hack, and the perl code in objects.pl
6574 isn't very elegant, but it works as I intended. The simplest way
6575 to check that it worked correctly is to look in obj_dat.h and
6576 check the array nid_objs and make sure the objects haven't moved
6577 around (this is important!). Additions are OK, as well as
6578 consistent name changes.
6581 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6584 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6585 The given file adds to whatever has already been seeded into the
6586 random pool through the RANDFILE configuration file option or
6587 environment variable, or the default random state file.
6590 *) mkstack.pl now sorts each macro group into lexical order.
6591 Previously the output order depended on the order the files
6592 appeared in the directory, resulting in needless rewriting
6596 *) Patches to make OpenSSL compile under Win32 again. Mostly
6597 work arounds for the VC++ problem that it treats func() as
6598 func(void). Also stripped out the parts of mkdef.pl that
6599 added extra typesafe functions: these no longer exist.
6602 *) Reorganisation of the stack code. The macros are now all
6603 collected in safestack.h . Each macro is defined in terms of
6604 a "stack macro" of the form SKM_<name>(type, a, b). The
6605 DEBUG_SAFESTACK is now handled in terms of function casts,
6606 this has the advantage of retaining type safety without the
6607 use of additional functions. If DEBUG_SAFESTACK is not defined
6608 then the non typesafe macros are used instead. Also modified the
6609 mkstack.pl script to handle the new form. Needs testing to see
6610 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6611 the default if no major problems. Similar behaviour for ASN1_SET_OF
6612 and PKCS12_STACK_OF.
6615 *) When some versions of IIS use the 'NET' form of private key the
6616 key derivation algorithm is different. Normally MD5(password) is
6617 used as a 128 bit RC4 key. In the modified case
6618 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6619 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6620 as the old Netscape_RSA functions except they have an additional
6621 'sgckey' parameter which uses the modified algorithm. Also added
6622 an -sgckey command line option to the rsa utility. Thanks to
6623 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6624 algorithm to openssl-dev.
6627 *) The evp_local.h macros were using 'c.##kname' which resulted in
6628 invalid expansion on some systems (SCO 5.0.5 for example).
6629 Corrected to 'c.kname'.
6630 [Phillip Porch <root@theporch.com>]
6632 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6633 a STACK of email addresses from a certificate or request, these look
6634 in the subject name and the subject alternative name extensions and
6635 omit any duplicate addresses.
6638 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6639 This makes DSA verification about 2 % faster.
6642 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6643 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6644 plus overhead for 1024 bit moduli).
6645 This makes exponentiations about 0.5 % faster for 1024 bit
6646 exponents (as measured by "openssl speed rsa2048").
6649 *) Rename memory handling macros to avoid conflicts with other
6651 Malloc => OPENSSL_malloc
6652 Malloc_locked => OPENSSL_malloc_locked
6653 Realloc => OPENSSL_realloc
6654 Free => OPENSSL_free
6657 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6658 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6661 *) CygWin32 support.
6662 [John Jarvie <jjarvie@newsguy.com>]
6664 *) The type-safe stack code has been rejigged. It is now only compiled
6665 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6666 by default all type-specific stack functions are "#define"d back to
6667 standard stack functions. This results in more streamlined output
6668 but retains the type-safety checking possibilities of the original
6672 *) The STACK code has been cleaned up, and certain type declarations
6673 that didn't make a lot of sense have been brought in line. This has
6674 also involved a cleanup of sorts in safestack.h to more correctly
6675 map type-safe stack functions onto their plain stack counterparts.
6676 This work has also resulted in a variety of "const"ifications of
6677 lots of the code, especially "_cmp" operations which should normally
6678 be prototyped with "const" parameters anyway.
6681 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6682 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6683 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6684 where all of 'md' is used each time the PRNG is used, but 'state'
6685 is used only indexed by a cyclic counter. As entropy may not be
6686 well distributed from the beginning, 'md' is important as a
6687 chaining variable. However, the output function chains only half
6688 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6689 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6690 in all of 'state' being rewritten, with the new values depending
6691 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6694 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6695 the handshake is continued after ssl_verify_cert_chain();
6696 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6697 can lead to 'unexplainable' connection aborts later.
6698 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6700 *) Major EVP API cipher revision.
6701 Add hooks for extra EVP features. This allows various cipher
6702 parameters to be set in the EVP interface. Support added for variable
6703 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6704 setting of RC2 and RC5 parameters.
6706 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6709 Remove lots of duplicated code from the EVP library. For example *every*
6710 cipher init() function handles the 'iv' in the same way according to the
6711 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6712 for CFB and OFB modes they zero ctx->num.
6714 New functionality allows removal of S/MIME code RC2 hack.
6716 Most of the routines have the same form and so can be declared in terms
6719 By shifting this to the top level EVP_CipherInit() it can be removed from
6720 all individual ciphers. If the cipher wants to handle IVs or keys
6721 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6724 Change lots of functions like EVP_EncryptUpdate() to now return a
6725 value: although software versions of the algorithms cannot fail
6726 any installed hardware versions can.
6729 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6730 this option is set, tolerate broken clients that send the negotiated
6731 protocol version number instead of the requested protocol version
6735 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6736 i.e. non-zero for export ciphersuites, zero otherwise.
6737 Previous versions had this flag inverted, inconsistent with
6738 rsa_tmp_cb (..._TMP_RSA_CB).
6739 [Bodo Moeller; problem reported by Amit Chopra]
6741 *) Add missing DSA library text string. Work around for some IIS
6742 key files with invalid SEQUENCE encoding.
6745 *) Add a document (doc/standards.txt) that list all kinds of standards
6746 and so on that are implemented in OpenSSL.
6749 *) Enhance c_rehash script. Old version would mishandle certificates
6750 with the same subject name hash and wouldn't handle CRLs at all.
6751 Added -fingerprint option to crl utility, to support new c_rehash
6755 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6758 *) Fix for SSL server purpose checking. Server checking was
6759 rejecting certificates which had extended key usage present
6760 but no ssl client purpose.
6761 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6763 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6764 is a little unclear about how a blank password is handled.
6765 Since the password in encoded as a BMPString with terminating
6766 double NULL a zero length password would end up as just the
6767 double NULL. However no password at all is different and is
6768 handled differently in the PKCS#12 key generation code. NS
6769 treats a blank password as zero length. MSIE treats it as no
6770 password on export: but it will try both on import. We now do
6771 the same: PKCS12_parse() tries zero length and no password if
6772 the password is set to "" or NULL (NULL is now a valid password:
6773 it wasn't before) as does the pkcs12 application.
6776 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6777 perror when PEM_read_bio_X509_REQ fails, the error message must
6778 be obtained from the error queue.
6781 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6782 it in ERR_remove_state if appropriate, and change ERR_get_state
6783 accordingly to avoid race conditions (this is necessary because
6784 thread_hash is no longer constant once set).
6787 *) Bugfix for linux-elf makefile.one.
6790 *) RSA_get_default_method() will now cause a default
6791 RSA_METHOD to be chosen if one doesn't exist already.
6792 Previously this was only set during a call to RSA_new()
6793 or RSA_new_method(NULL) meaning it was possible for
6794 RSA_get_default_method() to return NULL.
6797 *) Added native name translation to the existing DSO code
6798 that will convert (if the flag to do so is set) filenames
6799 that are sufficiently small and have no path information
6800 into a canonical native form. Eg. "blah" converted to
6801 "libblah.so" or "blah.dll" etc.
6804 *) New function ERR_error_string_n(e, buf, len) which is like
6805 ERR_error_string(e, buf), but writes at most 'len' bytes
6806 including the 0 terminator. For ERR_error_string_n, 'buf'
6808 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6810 *) CONF library reworked to become more general. A new CONF
6811 configuration file reader "class" is implemented as well as a
6812 new functions (NCONF_*, for "New CONF") to handle it. The now
6813 old CONF_* functions are still there, but are reimplemented to
6814 work in terms of the new functions. Also, a set of functions
6815 to handle the internal storage of the configuration data is
6816 provided to make it easier to write new configuration file
6817 reader "classes" (I can definitely see something reading a
6818 configuration file in XML format, for example), called _CONF_*,
6819 or "the configuration storage API"...
6821 The new configuration file reading functions are:
6823 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6824 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6826 NCONF_default, NCONF_WIN32
6828 NCONF_dump_fp, NCONF_dump_bio
6830 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6831 NCONF_new creates a new CONF object. This works in the same way
6832 as other interfaces in OpenSSL, like the BIO interface.
6833 NCONF_dump_* dump the internal storage of the configuration file,
6834 which is useful for debugging. All other functions take the same
6835 arguments as the old CONF_* functions wth the exception of the
6836 first that must be a `CONF *' instead of a `LHASH *'.
6838 To make it easer to use the new classes with the old CONF_* functions,
6839 the function CONF_set_default_method is provided.
6842 *) Add '-tls1' option to 'openssl ciphers', which was already
6843 mentioned in the documentation but had not been implemented.
6844 (This option is not yet really useful because even the additional
6845 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6848 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6849 OpenSSL-based applications) load shared libraries and bind to
6850 them in a portable way.
6851 [Geoff Thorpe, with contributions from Richard Levitte]
6853 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6855 *) Make sure _lrotl and _lrotr are only used with MSVC.
6857 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6858 (the default implementation of RAND_status).
6860 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6861 to '-clrext' (= clear extensions), as intended and documented.
6862 [Bodo Moeller; inconsistency pointed out by Michael Attili
6865 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6866 was larger than the MD block size.
6867 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6869 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6870 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6871 using the passed key: if the passed key was a private key the result
6872 of X509_print(), for example, would be to print out all the private key
6876 *) des_quad_cksum() byte order bug fix.
6877 [Ulf Möller, using the problem description in krb4-0.9.7, where
6878 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6880 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6882 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6884 *) For easily testing in shell scripts whether some command
6885 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6886 returns with exit code 0 iff no command of the given name is available.
6887 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6888 the output goes to stdout and nothing is printed to stderr.
6889 Additional arguments are always ignored.
6891 Since for each cipher there is a command of the same name,
6892 the 'no-cipher' compilation switches can be tested this way.
6894 ('openssl no-XXX' is not able to detect pseudo-commands such
6895 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6898 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6901 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6902 is set; it will be thrown away anyway because each handshake creates
6904 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6905 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6906 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6907 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6910 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6911 'Q' and 'R' lose their special meanings (quit/renegotiate).
6912 This is part of what -quiet does; unlike -quiet, -ign_eof
6913 does not suppress any output.
6916 *) Add compatibility options to the purpose and trust code. The
6917 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6918 accepts a certificate or CA, this was the previous behaviour,
6919 with all the associated security issues.
6921 X509_TRUST_COMPAT is the old trust behaviour: only and
6922 automatically trust self signed roots in certificate store. A
6923 new trust setting X509_TRUST_DEFAULT is used to specify that
6924 a purpose has no associated trust setting and it should instead
6925 use the value in the default purpose.
6928 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6929 and fix a memory leak.
6932 *) In util/mkerr.pl (which implements 'make errors'), preserve
6933 reason strings from the previous version of the .c file, as
6934 the default to have only downcase letters (and digits) in
6935 automatically generated reasons codes is not always appropriate.
6938 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6939 using strerror. Previously, ERR_reason_error_string() returned
6940 library names as reason strings for SYSerr; but SYSerr is a special
6941 case where small numbers are errno values, not library numbers.
6944 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6945 converts DSA parameters into DH parameters. (When creating parameters,
6946 DSA_generate_parameters is used.)
6949 *) Include 'length' (recommended exponent length) in C code generated
6950 by 'openssl dhparam -C'.
6953 *) The second argument to set_label in perlasm was already being used
6954 so couldn't be used as a "file scope" flag. Moved to third argument
6958 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6959 instead of RAND_bytes for encryption IVs and salts.
6962 *) Include RAND_status() into RAND_METHOD instead of implementing
6963 it only for md_rand.c Otherwise replacing the PRNG by calling
6964 RAND_set_rand_method would be impossible.
6967 *) Don't let DSA_generate_key() enter an infinite loop if the random
6968 number generation fails.
6971 *) New 'rand' application for creating pseudo-random output.
6974 *) Added configuration support for Linux/IA64
6975 [Rolf Haberrecker <rolf@suse.de>]
6977 *) Assembler module support for Mingw32.
6980 *) Shared library support for HPUX (in shlib/).
6981 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6983 *) Shared library support for Solaris gcc.
6984 [Lutz Behnke <behnke@trustcenter.de>]
6986 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6988 *) PKCS7_encrypt() was adding text MIME headers twice because they
6989 were added manually and by SMIME_crlf_copy().
6992 *) In bntest.c don't call BN_rand with zero bits argument.
6993 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6995 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6996 case was implemented. This caused BN_div_recp() to fail occasionally.
6999 *) Add an optional second argument to the set_label() in the perl
7000 assembly language builder. If this argument exists and is set
7001 to 1 it signals that the assembler should use a symbol whose
7002 scope is the entire file, not just the current function. This
7003 is needed with MASM which uses the format label:: for this scope.
7004 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7006 *) Change the ASN1 types so they are typedefs by default. Before
7007 almost all types were #define'd to ASN1_STRING which was causing
7008 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7012 *) Change names of new functions to the new get1/get0 naming
7013 convention: After 'get1', the caller owns a reference count
7014 and has to call ..._free; 'get0' returns a pointer to some
7015 data structure without incrementing reference counters.
7016 (Some of the existing 'get' functions increment a reference
7017 counter, some don't.)
7018 Similarly, 'set1' and 'add1' functions increase reference
7019 counters or duplicate objects.
7022 *) Allow for the possibility of temp RSA key generation failure:
7023 the code used to assume it always worked and crashed on failure.
7026 *) Fix potential buffer overrun problem in BIO_printf().
7027 [Ulf Möller, using public domain code by Patrick Powell; problem
7028 pointed out by David Sacerdote <das33@cornell.edu>]
7030 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7031 RAND_egd() and RAND_status(). In the command line application,
7032 the EGD socket can be specified like a seed file using RANDFILE
7036 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7037 Some CAs (e.g. Verisign) distribute certificates in this form.
7040 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7041 list to exclude them. This means that no special compilation option
7042 is needed to use anonymous DH: it just needs to be included in the
7046 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7047 EVP_MD_type. The old functionality is available in a new macro called
7048 EVP_MD_md(). Change code that uses it and update docs.
7051 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7052 where the 'void *' argument is replaced by a function pointer argument.
7053 Previously 'void *' was abused to point to functions, which works on
7054 many platforms, but is not correct. As these functions are usually
7055 called by macros defined in OpenSSL header files, most source code
7056 should work without changes.
7059 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7060 sections with information on -D... compiler switches used for
7061 compiling the library so that applications can see them. To enable
7062 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7063 must be defined. E.g.,
7064 #define OPENSSL_ALGORITHM_DEFINES
7065 #include <openssl/opensslconf.h>
7066 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7067 [Richard Levitte, Ulf and Bodo Möller]
7069 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7073 *) Change the 'other' type in certificate aux info to a STACK_OF
7074 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7075 the required ASN1 format: arbitrary types determined by an OID.
7078 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7079 argument to 'req'. This is not because the function is newer or
7080 better than others it just uses the work 'NEW' in the certificate
7081 request header lines. Some software needs this.
7084 *) Reorganise password command line arguments: now passwords can be
7085 obtained from various sources. Delete the PEM_cb function and make
7086 it the default behaviour: i.e. if the callback is NULL and the
7087 usrdata argument is not NULL interpret it as a null terminated pass
7088 phrase. If usrdata and the callback are NULL then the pass phrase
7089 is prompted for as usual.
7092 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7093 the support is automatically enabled. The resulting binaries will
7094 autodetect the card and use it if present.
7095 [Ben Laurie and Compaq Inc.]
7097 *) Work around for Netscape hang bug. This sends certificate request
7098 and server done in one record. Since this is perfectly legal in the
7099 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7100 the bugs/SSLv3 entry for more info.
7103 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7106 *) Add -rand argument to smime and pkcs12 applications and read/write
7110 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7113 *) Add command line password options to the remaining applications.
7116 *) Bug fix for BN_div_recp() for numerators with an even number of
7120 *) More tests in bntest.c, and changed test_bn output.
7123 *) ./config recognizes MacOS X now.
7126 *) Bug fix for BN_div() when the first words of num and divsor are
7127 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7130 *) Add support for various broken PKCS#8 formats, and command line
7131 options to produce them.
7134 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7135 get temporary BIGNUMs from a BN_CTX.
7138 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7142 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7143 include a #define from the old name to the new. The original intent
7144 was that statically linked binaries could for example just call
7145 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7146 link with digests. This never worked becayse SSLeay_add_all_digests()
7147 and SSLeay_add_all_ciphers() were in the same source file so calling
7148 one would link with the other. They are now in separate source files.
7151 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7154 *) Use a less unusual form of the Miller-Rabin primality test (it used
7155 a binary algorithm for exponentiation integrated into the Miller-Rabin
7156 loop, our standard modexp algorithms are faster).
7159 *) Support for the EBCDIC character set completed.
7160 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7162 *) Source code cleanups: use const where appropriate, eliminate casts,
7163 use void * instead of char * in lhash.
7166 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7167 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7168 this the server could overwrite ephemeral keys that the client
7172 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7173 using 50 iterations of the Rabin-Miller test.
7175 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7176 iterations of the Rabin-Miller test as required by the appendix
7177 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7178 As BN_is_prime_fasttest includes trial division, DSA parameter
7179 generation becomes much faster.
7181 This implies a change for the callback functions in DSA_is_prime
7182 and DSA_generate_parameters: The callback function is called once
7183 for each positive witness in the Rabin-Miller test, not just
7184 occasionally in the inner loop; and the parameters to the
7185 callback function now provide an iteration count for the outer
7186 loop rather than for the current invocation of the inner loop.
7187 DSA_generate_parameters additionally can call the callback
7188 function with an 'iteration count' of -1, meaning that a
7189 candidate has passed the trial division test (when q is generated
7190 from an application-provided seed, trial division is skipped).
7193 *) New function BN_is_prime_fasttest that optionally does trial
7194 division before starting the Rabin-Miller test and has
7195 an additional BN_CTX * argument (whereas BN_is_prime always
7196 has to allocate at least one BN_CTX).
7197 'callback(1, -1, cb_arg)' is called when a number has passed the
7198 trial division stage.
7201 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7205 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7208 *) New function BN_pseudo_rand().
7211 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7212 bignum version of BN_from_montgomery() with the working code from
7213 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7217 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7218 made it impossible to use the same SSL_SESSION data structure in
7219 SSL2 clients in multiple threads.
7222 *) The return value of RAND_load_file() no longer counts bytes obtained
7223 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7224 to seed the PRNG (previously an explicit byte count was required).
7225 [Ulf Möller, Bodo Möller]
7227 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7228 used (char *) instead of (void *) and had casts all over the place.
7231 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7234 *) Retain source code compatibility for BN_prime_checks macro:
7235 BN_is_prime(..., BN_prime_checks, ...) now uses
7236 BN_prime_checks_for_size to determine the appropriate number of
7237 Rabin-Miller iterations.
7240 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7241 DH_CHECK_P_NOT_SAFE_PRIME.
7242 (Check if this is true? OpenPGP calls them "strong".)
7245 *) Merge the functionality of "dh" and "gendh" programs into a new program
7246 "dhparam". The old programs are retained for now but will handle DH keys
7247 (instead of parameters) in future.
7250 *) Make the ciphers, s_server and s_client programs check the return values
7251 when a new cipher list is set.
7254 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7255 ciphers. Before when the 56bit ciphers were enabled the sorting was
7258 The syntax for the cipher sorting has been extended to support sorting by
7259 cipher-strength (using the strength_bits hard coded in the tables).
7260 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7262 Fix a bug in the cipher-command parser: when supplying a cipher command
7263 string with an "undefined" symbol (neither command nor alphanumeric
7264 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7265 an error is flagged.
7267 Due to the strength-sorting extension, the code of the
7268 ssl_create_cipher_list() function was completely rearranged. I hope that
7269 the readability was also increased :-)
7270 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7272 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7273 for the first serial number and places 2 in the serial number file. This
7274 avoids problems when the root CA is created with serial number zero and
7275 the first user certificate has the same issuer name and serial number
7279 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7280 the new code. Add documentation for this stuff.
7283 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7284 X509_*() to X509at_*() on the grounds that they don't handle X509
7285 structures and behave in an analagous way to the X509v3 functions:
7286 they shouldn't be called directly but wrapper functions should be used
7289 So we also now have some wrapper functions that call the X509at functions
7290 when passed certificate requests. (TO DO: similar things can be done with
7291 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7292 things. Some of these need some d2i or i2d and print functionality
7293 because they handle more complex structures.)
7296 *) Add missing #ifndefs that caused missing symbols when building libssl
7297 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7298 NO_RSA in ssl/s2*.c.
7299 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7301 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7302 has a return value which indicates the quality of the random data
7303 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7304 error queue. New function RAND_pseudo_bytes() generates output that is
7305 guaranteed to be unique but not unpredictable. RAND_add is like
7306 RAND_seed, but takes an extra argument for an entropy estimate
7307 (RAND_seed always assumes full entropy).
7310 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7311 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7312 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7313 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7314 false-positive rate of at most 2^-80 for random input.
7317 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7320 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7321 in the 0.9.5 release), this returns the chain
7322 from an X509_CTX structure with a dup of the stack and all
7323 the X509 reference counts upped: so the stack will exist
7324 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7327 Also make SSL_SESSION_print() print out the verify return
7331 *) Add manpage for the pkcs12 command. Also change the default
7332 behaviour so MAC iteration counts are used unless the new
7333 -nomaciter option is used. This improves file security and
7334 only older versions of MSIE (4.0 for example) need it.
7337 *) Honor the no-xxx Configure options when creating .DEF files.
7340 *) Add PKCS#10 attributes to field table: challengePassword,
7341 unstructuredName and unstructuredAddress. These are taken from
7342 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7343 international characters are used.
7345 More changes to X509_ATTRIBUTE code: allow the setting of types
7346 based on strings. Remove the 'loc' parameter when adding
7347 attributes because these will be a SET OF encoding which is sorted
7351 *) Initial changes to the 'req' utility to allow request generation
7352 automation. This will allow an application to just generate a template
7353 file containing all the field values and have req construct the
7356 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7357 used all over the place including certificate requests and PKCS#7
7358 structures. They are currently handled manually where necessary with
7359 some primitive wrappers for PKCS#7. The new functions behave in a
7360 manner analogous to the X509 extension functions: they allow
7361 attributes to be looked up by NID and added.
7363 Later something similar to the X509V3 code would be desirable to
7364 automatically handle the encoding, decoding and printing of the
7365 more complex types. The string types like challengePassword can
7366 be handled by the string table functions.
7368 Also modified the multi byte string table handling. Now there is
7369 a 'global mask' which masks out certain types. The table itself
7370 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7371 is useful when for example there is only one permissible type
7372 (as in countryName) and using the mask might result in no valid
7376 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7377 SSL_get_peer_finished to allow applications to obtain the latest
7378 Finished messages sent to the peer or expected from the peer,
7379 respectively. (SSL_get_peer_finished is usually the Finished message
7380 actually received from the peer, otherwise the protocol will be aborted.)
7382 As the Finished message are message digests of the complete handshake
7383 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7384 be used for external authentication procedures when the authentication
7385 provided by SSL/TLS is not desired or is not enough.
7388 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7389 the host supports BWX extension and if Compaq C is present on the
7390 $PATH. Just exploiting of the BWX extension results in 20-30%
7391 performance kick for some algorithms, e.g. DES and RC4 to mention
7392 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7396 *) Add support for MS "fast SGC". This is arguably a violation of the
7397 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7398 weak crypto and after checking the certificate is SGC a second one
7399 with strong crypto. MS SGC stops the first handshake after receiving
7400 the server certificate message and sends a second client hello. Since
7401 a server will typically do all the time consuming operations before
7402 expecting any further messages from the client (server key exchange
7403 is the most expensive) there is little difference between the two.
7405 To get OpenSSL to support MS SGC we have to permit a second client
7406 hello message after we have sent server done. In addition we have to
7407 reset the MAC if we do get this second client hello.
7410 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7411 if a DER encoded private key is RSA or DSA traditional format. Changed
7412 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7413 format DER encoded private key. Newer code should use PKCS#8 format which
7414 has the key type encoded in the ASN1 structure. Added DER private key
7415 support to pkcs8 application.
7418 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7419 ciphersuites has been selected (as required by the SSL 3/TLS 1
7420 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7421 is set, we interpret this as a request to violate the specification
7422 (the worst that can happen is a handshake failure, and 'correct'
7423 behaviour would result in a handshake failure anyway).
7426 *) In SSL_CTX_add_session, take into account that there might be multiple
7427 SSL_SESSION structures with the same session ID (e.g. when two threads
7428 concurrently obtain them from an external cache).
7429 The internal cache can handle only one SSL_SESSION with a given ID,
7430 so if there's a conflict, we now throw out the old one to achieve
7434 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7435 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7436 some routines that use cipher OIDs: some ciphers do not have OIDs
7437 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7441 *) Simplify the trust setting structure and code. Now we just have
7442 two sequences of OIDs for trusted and rejected settings. These will
7443 typically have values the same as the extended key usage extension
7444 and any application specific purposes.
7446 The trust checking code now has a default behaviour: it will just
7447 check for an object with the same NID as the passed id. Functions can
7448 be provided to override either the default behaviour or the behaviour
7449 for a given id. SSL client, server and email already have functions
7450 in place for compatibility: they check the NID and also return "trusted"
7451 if the certificate is self signed.
7454 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7455 traditional format into an EVP_PKEY structure.
7458 *) Add a password callback function PEM_cb() which either prompts for
7459 a password if usr_data is NULL or otherwise assumes it is a null
7460 terminated password. Allow passwords to be passed on command line
7461 environment or config files in a few more utilities.
7464 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7465 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7466 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7467 Update documentation.
7470 *) Support for ASN1 "NULL" type. This could be handled before by using
7471 ASN1_TYPE but there wasn't any function that would try to read a NULL
7472 and produce an error if it couldn't. For compatibility we also have
7473 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7474 don't allocate anything because they don't need to.
7477 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7479 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7481 *) Rebuild of the memory allocation routines used by OpenSSL code and
7482 possibly others as well. The purpose is to make an interface that
7483 provide hooks so anyone can build a separate set of allocation and
7484 deallocation routines to be used by OpenSSL, for example memory
7485 pool implementations, or something else, which was previously hard
7486 since Malloc(), Realloc() and Free() were defined as macros having
7487 the values malloc, realloc and free, respectively (except for Win32
7488 compilations). The same is provided for memory debugging code.
7489 OpenSSL already comes with functionality to find memory leaks, but
7490 this gives people a chance to debug other memory problems.
7492 With these changes, a new set of functions and macros have appeared:
7494 CRYPTO_set_mem_debug_functions() [F]
7495 CRYPTO_get_mem_debug_functions() [F]
7496 CRYPTO_dbg_set_options() [F]
7497 CRYPTO_dbg_get_options() [F]
7498 CRYPTO_malloc_debug_init() [M]
7500 The memory debug functions are NULL by default, unless the library
7501 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7502 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7503 gives the standard debugging functions that come with OpenSSL) or
7504 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7505 provided by the library user) must be used. When the standard
7506 debugging functions are used, CRYPTO_dbg_set_options can be used to
7507 request additional information:
7508 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7509 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7511 Also, things like CRYPTO_set_mem_functions will always give the
7512 expected result (the new set of functions is used for allocation
7513 and deallocation) at all times, regardless of platform and compiler
7516 To finish it up, some functions that were never use in any other
7517 way than through macros have a new API and new semantic:
7520 CRYPTO_dbg_realloc()
7523 All macros of value have retained their old syntax.
7524 [Richard Levitte and Bodo Moeller]
7526 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7527 ordering of SMIMECapabilities wasn't in "strength order" and there
7528 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7532 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7533 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7534 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7536 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7537 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7538 functionality to handle multipart/signed properly) and a utility
7539 called 'smime' to call all this stuff. This is based on code I
7540 originally wrote for Celo who have kindly allowed it to be
7541 included in OpenSSL.
7544 *) Add variants des_set_key_checked and des_set_key_unchecked of
7545 des_set_key (aka des_key_sched). Global variable des_check_key
7546 decides which of these is called by des_set_key; this way
7547 des_check_key behaves as it always did, but applications and
7548 the library itself, which was buggy for des_check_key == 1,
7549 have a cleaner way to pick the version they need.
7552 *) New function PKCS12_newpass() which changes the password of a
7556 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7557 dynamic mix. In both cases the ids can be used as an index into the
7558 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7559 functions so they accept a list of the field values and the
7560 application doesn't need to directly manipulate the X509_TRUST
7564 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7568 *) Modify the way the V3 extension code looks up extensions. This now
7569 works in a similar way to the object code: we have some "standard"
7570 extensions in a static table which is searched with OBJ_bsearch()
7571 and the application can add dynamic ones if needed. The file
7572 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7573 updated whenever a new extension is added to the core code and kept
7574 in ext_nid order. There is a simple program 'tabtest.c' which checks
7575 this. New extensions are not added too often so this file can readily
7576 be maintained manually.
7578 There are two big advantages in doing things this way. The extensions
7579 can be looked up immediately and no longer need to be "added" using
7580 X509V3_add_standard_extensions(): this function now does nothing.
7581 [Side note: I get *lots* of email saying the extension code doesn't
7582 work because people forget to call this function]
7583 Also no dynamic allocation is done unless new extensions are added:
7584 so if we don't add custom extensions there is no need to call
7585 X509V3_EXT_cleanup().
7588 *) Modify enc utility's salting as follows: make salting the default. Add a
7589 magic header, so unsalted files fail gracefully instead of just decrypting
7590 to garbage. This is because not salting is a big security hole, so people
7591 should be discouraged from doing it.
7594 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7595 digest to be passed on the command line but it only used this
7596 parameter when signing a certificate. Modified so all relevant
7597 operations are affected by the digest parameter including the
7598 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7599 DSA key was used because it didn't fix the digest.
7602 *) Initial certificate chain verify code. Currently tests the untrusted
7603 certificates for consistency with the verify purpose (which is set
7604 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7606 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7607 this is because it will reject chains with invalid extensions whereas
7608 every previous version of OpenSSL and SSLeay made no checks at all.
7610 Trust code: checks the root CA for the relevant trust settings. Trust
7611 settings have an initial value consistent with the verify purpose: e.g.
7612 if the verify purpose is for SSL client use it expects the CA to be
7613 trusted for SSL client use. However the default value can be changed to
7614 permit custom trust settings: one example of this would be to only trust
7615 certificates from a specific "secure" set of CAs.
7617 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7618 which should be used for version portability: especially since the
7619 verify structure is likely to change more often now.
7621 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7622 to set them. If not set then assume SSL clients will verify SSL servers
7625 Two new options to the verify program: -untrusted allows a set of
7626 untrusted certificates to be passed in and -purpose which sets the
7627 intended purpose of the certificate. If a purpose is set then the
7628 new chain verify code is used to check extension consistency.
7631 *) Support for the authority information access extension.
7634 *) Modify RSA and DSA PEM read routines to transparently handle
7635 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7636 public keys in a format compatible with certificate
7637 SubjectPublicKeyInfo structures. Unfortunately there were already
7638 functions called *_PublicKey_* which used various odd formats so
7639 these are retained for compatibility: however the DSA variants were
7640 never in a public release so they have been deleted. Changed dsa/rsa
7641 utilities to handle the new format: note no releases ever handled public
7642 keys so we should be OK.
7644 The primary motivation for this change is to avoid the same fiasco
7645 that dogs private keys: there are several incompatible private key
7646 formats some of which are standard and some OpenSSL specific and
7647 require various evil hacks to allow partial transparent handling and
7648 even then it doesn't work with DER formats. Given the option anything
7649 other than PKCS#8 should be dumped: but the other formats have to
7650 stay in the name of compatibility.
7652 With public keys and the benefit of hindsight one standard format
7653 is used which works with EVP_PKEY, RSA or DSA structures: though
7654 it clearly returns an error if you try to read the wrong kind of key.
7656 Added a -pubkey option to the 'x509' utility to output the public key.
7657 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7658 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7659 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7660 that do the same as the EVP_PKEY_assign_*() except they up the
7661 reference count of the added key (they don't "swallow" the
7665 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7666 CRLs would fail if the file contained no certificates or no CRLs:
7667 added a new function to read in both types and return the number
7668 read: this means that if none are read it will be an error. The
7669 DER versions of the certificate and CRL reader would always fail
7670 because it isn't possible to mix certificates and CRLs in DER format
7671 without choking one or the other routine. Changed this to just read
7672 a certificate: this is the best we can do. Also modified the code
7673 in apps/verify.c to take notice of return codes: it was previously
7674 attempting to read in certificates from NULL pointers and ignoring
7675 any errors: this is one reason why the cert and CRL reader seemed
7676 to work. It doesn't check return codes from the default certificate
7677 routines: these may well fail if the certificates aren't installed.
7680 *) Code to support otherName option in GeneralName.
7683 *) First update to verify code. Change the verify utility
7684 so it warns if it is passed a self signed certificate:
7685 for consistency with the normal behaviour. X509_verify
7686 has been modified to it will now verify a self signed
7687 certificate if *exactly* the same certificate appears
7688 in the store: it was previously impossible to trust a
7689 single self signed certificate. This means that:
7690 openssl verify ss.pem
7691 now gives a warning about a self signed certificate but
7692 openssl verify -CAfile ss.pem ss.pem
7696 *) For servers, store verify_result in SSL_SESSION data structure
7697 (and add it to external session representation).
7698 This is needed when client certificate verifications fails,
7699 but an application-provided verification callback (set by
7700 SSL_CTX_set_cert_verify_callback) allows accepting the session
7701 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7702 but returns 1): When the session is reused, we have to set
7703 ssl->verify_result to the appropriate error code to avoid
7705 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7707 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7708 case in PKCS7_dataInit() where the signed PKCS7 structure
7709 didn't contain any existing data because it was being created.
7710 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7712 *) Add a salt to the key derivation routines in enc.c. This
7713 forms the first 8 bytes of the encrypted file. Also add a
7714 -S option to allow a salt to be input on the command line.
7717 *) New function X509_cmp(). Oddly enough there wasn't a function
7718 to compare two certificates. We do this by working out the SHA1
7719 hash and comparing that. X509_cmp() will be needed by the trust
7723 *) SSL_get1_session() is like SSL_get_session(), but increments
7724 the reference count in the SSL_SESSION returned.
7725 [Geoff Thorpe <geoff@eu.c2.net>]
7727 *) Fix for 'req': it was adding a null to request attributes.
7728 Also change the X509_LOOKUP and X509_INFO code to handle
7729 certificate auxiliary information.
7732 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7736 *) Add the possibility to add extra information to the memory leak
7737 detecting output, to form tracebacks, showing from where each
7738 allocation was originated: CRYPTO_push_info("constant string") adds
7739 the string plus current file name and line number to a per-thread
7740 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7741 is like calling CYRPTO_pop_info() until the stack is empty.
7742 Also updated memory leak detection code to be multi-thread-safe.
7745 *) Add options -text and -noout to pkcs7 utility and delete the
7746 encryption options which never did anything. Update docs.
7749 *) Add options to some of the utilities to allow the pass phrase
7750 to be included on either the command line (not recommended on
7751 OSes like Unix) or read from the environment. Update the
7752 manpages and fix a few bugs.
7755 *) Add a few manpages for some of the openssl commands.
7758 *) Fix the -revoke option in ca. It was freeing up memory twice,
7759 leaking and not finding already revoked certificates.
7762 *) Extensive changes to support certificate auxiliary information.
7763 This involves the use of X509_CERT_AUX structure and X509_AUX
7764 functions. An X509_AUX function such as PEM_read_X509_AUX()
7765 can still read in a certificate file in the usual way but it
7766 will also read in any additional "auxiliary information". By
7767 doing things this way a fair degree of compatibility can be
7768 retained: existing certificates can have this information added
7769 using the new 'x509' options.
7771 Current auxiliary information includes an "alias" and some trust
7772 settings. The trust settings will ultimately be used in enhanced
7773 certificate chain verification routines: currently a certificate
7774 can only be trusted if it is self signed and then it is trusted
7778 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7779 The problem was that one of the replacement routines had not been working
7780 since SSLeay releases. For now the offending routine has been replaced
7781 with non-optimised assembler. Even so, this now gives around 95%
7782 performance improvement for 1024 bit RSA signs.
7785 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7786 handling. Most clients have the effective key size in bits equal to
7787 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7788 A few however don't do this and instead use the size of the decrypted key
7789 to determine the RC2 key length and the AlgorithmIdentifier to determine
7790 the effective key length. In this case the effective key length can still
7791 be 40 bits but the key length can be 168 bits for example. This is fixed
7792 by manually forcing an RC2 key into the EVP_PKEY structure because the
7793 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7794 the key length and effective key length are equal.
7797 *) Add a bunch of functions that should simplify the creation of
7798 X509_NAME structures. Now you should be able to do:
7799 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7800 and have it automatically work out the correct field type and fill in
7801 the structures. The more adventurous can try:
7802 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7803 and it will (hopefully) work out the correct multibyte encoding.
7806 *) Change the 'req' utility to use the new field handling and multibyte
7807 copy routines. Before the DN field creation was handled in an ad hoc
7808 way in req, ca, and x509 which was rather broken and didn't support
7809 BMPStrings or UTF8Strings. Since some software doesn't implement
7810 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7811 using the dirstring_type option. See the new comment in the default
7812 openssl.cnf for more info.
7815 *) Make crypto/rand/md_rand.c more robust:
7816 - Assure unique random numbers after fork().
7817 - Make sure that concurrent threads access the global counter and
7818 md serializably so that we never lose entropy in them
7819 or use exactly the same state in multiple threads.
7820 Access to the large state is not always serializable because
7821 the additional locking could be a performance killer, and
7822 md should be large enough anyway.
7825 *) New file apps/app_rand.c with commonly needed functionality
7826 for handling the random seed file.
7828 Use the random seed file in some applications that previously did not:
7830 dsaparam -genkey (which also ignored its '-rand' option),
7833 x509 (when signing).
7834 Except on systems with /dev/urandom, it is crucial to have a random
7835 seed file at least for key creation, DSA signing, and for DH exchanges;
7836 for RSA signatures we could do without one.
7838 gendh and gendsa (unlike genrsa) used to read only the first byte
7839 of each file listed in the '-rand' option. The function as previously
7840 found in genrsa is now in app_rand.c and is used by all programs
7841 that support '-rand'.
7844 *) In RAND_write_file, use mode 0600 for creating files;
7845 don't just chmod when it may be too late.
7848 *) Report an error from X509_STORE_load_locations
7849 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7852 *) New function ASN1_mbstring_copy() this copies a string in either
7853 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7854 into an ASN1_STRING type. A mask of permissible types is passed
7855 and it chooses the "minimal" type to use or an error if not type
7859 *) Add function equivalents to the various macros in asn1.h. The old
7860 macros are retained with an M_ prefix. Code inside the library can
7861 use the M_ macros. External code (including the openssl utility)
7862 should *NOT* in order to be "shared library friendly".
7865 *) Add various functions that can check a certificate's extensions
7866 to see if it usable for various purposes such as SSL client,
7867 server or S/MIME and CAs of these types. This is currently
7868 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7869 verification. Also added a -purpose flag to x509 utility to
7870 print out all the purposes.
7873 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7877 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7878 for, obtain and decode and extension and obtain its critical flag.
7879 This allows all the necessary extension code to be handled in a
7880 single function call.
7883 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7884 platforms. See crypto/rc4/rc4_enc.c for further details.
7887 *) New -noout option to asn1parse. This causes no output to be produced
7888 its main use is when combined with -strparse and -out to extract data
7889 from a file (which may not be in ASN.1 format).
7892 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7893 when producing the local key id.
7894 [Richard Levitte <levitte@stacken.kth.se>]
7896 *) New option -dhparam in s_server. This allows a DH parameter file to be
7897 stated explicitly. If it is not stated then it tries the first server
7898 certificate file. The previous behaviour hard coded the filename
7902 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7903 a public key to be input or output. For example:
7904 openssl rsa -in key.pem -pubout -out pubkey.pem
7905 Also added necessary DSA public key functions to handle this.
7908 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7909 in the message. This was handled by allowing
7910 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7911 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7913 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7914 to the end of the strings whereas this didn't. This would cause problems
7915 if strings read with d2i_ASN1_bytes() were later modified.
7916 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7918 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7919 data and it contains EOF it will end up returning an error. This is
7920 caused by input 46 bytes long. The cause is due to the way base64
7921 BIOs find the start of base64 encoded data. They do this by trying a
7922 trial decode on each line until they find one that works. When they
7923 do a flag is set and it starts again knowing it can pass all the
7924 data directly through the decoder. Unfortunately it doesn't reset
7925 the context it uses. This means that if EOF is reached an attempt
7926 is made to pass two EOFs through the context and this causes the
7927 resulting error. This can also cause other problems as well. As is
7928 usual with these problems it takes *ages* to find and the fix is
7929 trivial: move one line.
7930 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7932 *) Ugly workaround to get s_client and s_server working under Windows. The
7933 old code wouldn't work because it needed to select() on sockets and the
7934 tty (for keypresses and to see if data could be written). Win32 only
7935 supports select() on sockets so we select() with a 1s timeout on the
7936 sockets and then see if any characters are waiting to be read, if none
7937 are present then we retry, we also assume we can always write data to
7938 the tty. This isn't nice because the code then blocks until we've
7939 received a complete line of data and it is effectively polling the
7940 keyboard at 1s intervals: however it's quite a bit better than not
7941 working at all :-) A dedicated Windows application might handle this
7942 with an event loop for example.
7945 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7946 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7947 will be called when RSA_sign() and RSA_verify() are used. This is useful
7948 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7949 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7950 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7951 This necessitated the support of an extra signature type NID_md5_sha1
7952 for SSL signatures and modifications to the SSL library to use it instead
7953 of calling RSA_public_decrypt() and RSA_private_encrypt().
7956 *) Add new -verify -CAfile and -CApath options to the crl program, these
7957 will lookup a CRL issuers certificate and verify the signature in a
7958 similar way to the verify program. Tidy up the crl program so it
7959 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7960 less strict. It will now permit CRL extensions even if it is not
7961 a V2 CRL: this will allow it to tolerate some broken CRLs.
7964 *) Initialize all non-automatic variables each time one of the openssl
7965 sub-programs is started (this is necessary as they may be started
7966 multiple times from the "OpenSSL>" prompt).
7967 [Lennart Bang, Bodo Moeller]
7969 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7970 removing all other RSA functionality (this is what NO_RSA does). This
7971 is so (for example) those in the US can disable those operations covered
7972 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7976 *) Non-copying interface to BIO pairs.
7977 (still largely untested)
7980 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7981 ASCII string. This was handled independently in various places before.
7984 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7985 UTF8 strings a character at a time.
7988 *) Use client_version from client hello to select the protocol
7989 (s23_srvr.c) and for RSA client key exchange verification
7990 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7993 *) Add various utility functions to handle SPKACs, these were previously
7994 handled by poking round in the structure internals. Added new function
7995 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7996 print, verify and generate SPKACs. Based on an original idea from
7997 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8000 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8003 *) Allow the config file extension section to be overwritten on the
8004 command line. Based on an original idea from Massimiliano Pala
8005 <madwolf@comune.modena.it>. The new option is called -extensions
8006 and can be applied to ca, req and x509. Also -reqexts to override
8007 the request extensions in req and -crlexts to override the crl extensions
8011 *) Add new feature to the SPKAC handling in ca. Now you can include
8012 the same field multiple times by preceding it by "XXXX." for example:
8015 this is the same syntax as used in the req config file.
8018 *) Allow certificate extensions to be added to certificate requests. These
8019 are specified in a 'req_extensions' option of the req section of the
8020 config file. They can be printed out with the -text option to req but
8021 are otherwise ignored at present.
8024 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8025 data read consists of only the final block it would not decrypted because
8026 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8027 A misplaced 'break' also meant the decrypted final block might not be
8028 copied until the next read.
8031 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8032 a few extra parameters to the DH structure: these will be useful if
8033 for example we want the value of 'q' or implement X9.42 DH.
8036 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8037 provides hooks that allow the default DSA functions or functions on a
8038 "per key" basis to be replaced. This allows hardware acceleration and
8039 hardware key storage to be handled without major modification to the
8040 library. Also added low level modexp hooks and CRYPTO_EX structure and
8041 associated functions.
8044 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8045 as "read only": it can't be written to and the buffer it points to will
8046 not be freed. Reading from a read only BIO is much more efficient than
8047 a normal memory BIO. This was added because there are several times when
8048 an area of memory needs to be read from a BIO. The previous method was
8049 to create a memory BIO and write the data to it, this results in two
8050 copies of the data and an O(n^2) reading algorithm. There is a new
8051 function BIO_new_mem_buf() which creates a read only memory BIO from
8052 an area of memory. Also modified the PKCS#7 routines to use read only
8056 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8057 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8058 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8059 but a retry condition occured while trying to read the rest.
8062 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8063 NID_pkcs7_encrypted by default: this was wrong since this should almost
8064 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8065 the encrypted data type: this is a more sensible place to put it and it
8066 allows the PKCS#12 code to be tidied up that duplicated this
8070 *) Changed obj_dat.pl script so it takes its input and output files on
8071 the command line. This should avoid shell escape redirection problems
8075 *) Initial support for certificate extension requests, these are included
8076 in things like Xenroll certificate requests. Included functions to allow
8077 extensions to be obtained and added.
8080 *) -crlf option to s_client and s_server for sending newlines as
8081 CRLF (as required by many protocols).
8084 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8086 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8087 [Ralf S. Engelschall]
8089 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8090 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8092 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8096 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8097 DH parameters/keys (q is lost during that conversion, but the resulting
8098 DH parameters contain its length).
8100 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8101 much faster than DH_generate_parameters (which creates parameters
8102 where p = 2*q + 1), and also the smaller q makes DH computations
8103 much more efficient (160-bit exponentiation instead of 1024-bit
8104 exponentiation); so this provides a convenient way to support DHE
8105 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8106 utter importance to use
8107 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8109 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8110 when such DH parameters are used, because otherwise small subgroup
8111 attacks may become possible!
8114 *) Avoid memory leak in i2d_DHparams.
8117 *) Allow the -k option to be used more than once in the enc program:
8118 this allows the same encrypted message to be read by multiple recipients.
8121 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8122 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8123 it will always use the numerical form of the OID, even if it has a short
8127 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8128 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8129 otherwise bn_mod_exp was called. In the case of hardware keys for example
8130 no private key components need be present and it might store extra data
8131 in the RSA structure, which cannot be accessed from bn_mod_exp.
8132 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8133 private key operations.
8136 *) Added support for SPARC Linux.
8139 *) pem_password_cb function type incompatibly changed from
8140 typedef int pem_password_cb(char *buf, int size, int rwflag);
8142 ....(char *buf, int size, int rwflag, void *userdata);
8143 so that applications can pass data to their callbacks:
8144 The PEM[_ASN1]_{read,write}... functions and macros now take an
8145 additional void * argument, which is just handed through whenever
8146 the password callback is called.
8147 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8149 New function SSL_CTX_set_default_passwd_cb_userdata.
8151 Compatibility note: As many C implementations push function arguments
8152 onto the stack in reverse order, the new library version is likely to
8153 interoperate with programs that have been compiled with the old
8154 pem_password_cb definition (PEM_whatever takes some data that
8155 happens to be on the stack as its last argument, and the callback
8156 just ignores this garbage); but there is no guarantee whatsoever that
8159 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8160 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8161 problems not only on Windows, but also on some Unix platforms.
8162 To avoid problematic command lines, these definitions are now in an
8163 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8164 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8167 *) MIPS III/IV assembler module is reimplemented.
8170 *) More DES library cleanups: remove references to srand/rand and
8171 delete an unused file.
8174 *) Add support for the the free Netwide assembler (NASM) under Win32,
8175 since not many people have MASM (ml) and it can be hard to obtain.
8176 This is currently experimental but it seems to work OK and pass all
8177 the tests. Check out INSTALL.W32 for info.
8180 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8181 without temporary keys kept an extra copy of the server key,
8182 and connections with temporary keys did not free everything in case
8186 *) New function RSA_check_key and new openssl rsa option -check
8187 for verifying the consistency of RSA keys.
8188 [Ulf Moeller, Bodo Moeller]
8190 *) Various changes to make Win32 compile work:
8191 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8192 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8193 comparison" warnings.
8194 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8197 *) Add a debugging option to PKCS#5 v2 key generation function: when
8198 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8199 derived keys are printed to stderr.
8202 *) Copy the flags in ASN1_STRING_dup().
8203 [Roman E. Pavlov <pre@mo.msk.ru>]
8205 *) The x509 application mishandled signing requests containing DSA
8206 keys when the signing key was also DSA and the parameters didn't match.
8208 It was supposed to omit the parameters when they matched the signing key:
8209 the verifying software was then supposed to automatically use the CA's
8210 parameters if they were absent from the end user certificate.
8212 Omitting parameters is no longer recommended. The test was also
8213 the wrong way round! This was probably due to unusual behaviour in
8214 EVP_cmp_parameters() which returns 1 if the parameters match.
8215 This meant that parameters were omitted when they *didn't* match and
8216 the certificate was useless. Certificates signed with 'ca' didn't have
8218 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8220 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8221 The interface is as follows:
8222 Applications can use
8223 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8224 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8225 "off" is now the default.
8226 The library internally uses
8227 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8228 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8229 to disable memory-checking temporarily.
8231 Some inconsistent states that previously were possible (and were
8232 even the default) are now avoided.
8234 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8235 with each memory chunk allocated; this is occasionally more helpful
8236 than just having a counter.
8238 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8240 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8244 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8245 which largely parallels "options", but is for changing API behaviour,
8246 whereas "options" are about protocol behaviour.
8247 Initial "mode" flags are:
8249 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8250 a single record has been written.
8251 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8252 retries use the same buffer location.
8253 (But all of the contents must be
8257 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8260 *) Fix problems with no-hmac etc.
8261 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8263 *) New functions RSA_get_default_method(), RSA_set_method() and
8264 RSA_get_method(). These allows replacement of RSA_METHODs without having
8265 to mess around with the internals of an RSA structure.
8268 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8269 Also really enable memory leak checks in openssl.c and in some
8271 [Chad C. Mulligan, Bodo Moeller]
8273 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8274 up the length of negative integers. This has now been simplified to just
8275 store the length when it is first determined and use it later, rather
8276 than trying to keep track of where data is copied and updating it to
8278 [Steve Henson, reported by Brien Wheeler
8279 <bwheeler@authentica-security.com>]
8281 *) Add a new function PKCS7_signatureVerify. This allows the verification
8282 of a PKCS#7 signature but with the signing certificate passed to the
8283 function itself. This contrasts with PKCS7_dataVerify which assumes the
8284 certificate is present in the PKCS#7 structure. This isn't always the
8285 case: certificates can be omitted from a PKCS#7 structure and be
8286 distributed by "out of band" means (such as a certificate database).
8289 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8290 function prototypes in pem.h, also change util/mkdef.pl to add the
8291 necessary function names.
8294 *) mk1mf.pl (used by Windows builds) did not properly read the
8295 options set by Configure in the top level Makefile, and Configure
8296 was not even able to write more than one option correctly.
8297 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8300 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8301 file to be loaded from a BIO or FILE pointer. The BIO version will
8302 for example allow memory BIOs to contain config info.
8305 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8306 Whoever hopes to achieve shared-library compatibility across versions
8307 must use this, not the compile-time macro.
8308 (Exercise 0.9.4: Which is the minimum library version required by
8310 Note: All this applies only to multi-threaded programs, others don't
8314 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8315 through a BIO pair triggered the default case, i.e.
8316 SSLerr(...,SSL_R_UNKNOWN_STATE).
8319 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8320 can use the SSL library even if none of the specific BIOs is
8324 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8325 for the encoded length.
8326 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8328 *) Add initial documentation of the X509V3 functions.
8331 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8332 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8333 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8334 secure PKCS#8 private key format with a high iteration count.
8337 *) Fix determination of Perl interpreter: A perl or perl5
8338 _directory_ in $PATH was also accepted as the interpreter.
8339 [Ralf S. Engelschall]
8341 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8342 wrong with it but it was very old and did things like calling
8343 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8347 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8348 to use the new extension code.
8351 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8352 with macros. This should make it easier to change their form, add extra
8353 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8357 *) Add to configuration table a new entry that can specify an alternative
8358 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8359 according to Mark Crispin <MRC@Panda.COM>.
8363 *) DES CBC did not update the IV. Weird.
8366 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8367 Changing the behaviour of the former might break existing programs --
8368 where IV updating is needed, des_ncbc_encrypt can be used.
8371 *) When bntest is run from "make test" it drives bc to check its
8372 calculations, as well as internally checking them. If an internal check
8373 fails, it needs to cause bc to give a non-zero result or make test carries
8374 on without noticing the failure. Fixed.
8377 *) DES library cleanups.
8380 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8381 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8382 ciphers. NOTE: although the key derivation function has been verified
8383 against some published test vectors it has not been extensively tested
8384 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8388 *) Instead of "mkdir -p", which is not fully portable, use new
8389 Perl script "util/mkdir-p.pl".
8392 *) Rewrite the way password based encryption (PBE) is handled. It used to
8393 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8394 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8395 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8396 the 'parameter' field of the AlgorithmIdentifier is passed to the
8397 underlying key generation function so it must do its own ASN1 parsing.
8398 This has also changed the EVP_PBE_CipherInit() function which now has a
8399 'parameter' argument instead of literal salt and iteration count values
8400 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8403 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8404 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8405 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8406 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8407 value was just used as a "magic string" and not used directly its
8408 value doesn't matter.
8411 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8415 *) "linux-sparc64" configuration (ultrapenguin).
8416 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8417 "linux-sparc" configuration.
8418 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8420 *) config now generates no-xxx options for missing ciphers.
8423 *) Support the EBCDIC character set (work in progress).
8424 File ebcdic.c not yet included because it has a different license.
8425 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8427 *) Support BS2000/OSD-POSIX.
8428 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8430 *) Make callbacks for key generation use void * instead of char *.
8433 *) Make S/MIME samples compile (not yet tested).
8436 *) Additional typesafe stacks.
8439 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8443 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8445 *) New configuration variant "sco5-gcc".
8447 *) Updated some demos.
8448 [Sean O Riordain, Wade Scholine]
8450 *) Add missing BIO_free at exit of pkcs12 application.
8453 *) Fix memory leak in conf.c.
8456 *) Updates for Win32 to assembler version of MD5.
8459 *) Set #! path to perl in apps/der_chop to where we found it
8460 instead of using a fixed path.
8463 *) SHA library changes for irix64-mips4-cc.
8466 *) Improvements for VMS support.
8470 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8472 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8473 This also avoids the problems with SC4.2 and unpatched SC5.
8474 [Andy Polyakov <appro@fy.chalmers.se>]
8476 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8477 These are required because of the typesafe stack would otherwise break
8478 existing code. If old code used a structure member which used to be STACK
8479 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8480 sk_num or sk_value it would produce an error because the num, data members
8481 are not present in STACK_OF. Now it just produces a warning. sk_set
8482 replaces the old method of assigning a value to sk_value
8483 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8484 that does this will no longer work (and should use sk_set instead) but
8485 this could be regarded as a "questionable" behaviour anyway.
8488 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8489 correctly handle encrypted S/MIME data.
8492 *) Change type of various DES function arguments from des_cblock
8493 (which means, in function argument declarations, pointer to char)
8494 to des_cblock * (meaning pointer to array with 8 char elements),
8495 which allows the compiler to do more typechecking; it was like
8496 that back in SSLeay, but with lots of ugly casts.
8498 Introduce new type const_des_cblock.
8501 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8502 problems: find RecipientInfo structure that matches recipient certificate
8503 and initialise the ASN1 structures properly based on passed cipher.
8506 *) Belatedly make the BN tests actually check the results.
8509 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8510 to and from BNs: it was completely broken. New compilation option
8511 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8512 key elements as negative integers.
8515 *) Reorganize and speed up MD5.
8516 [Andy Polyakov <appro@fy.chalmers.se>]
8519 [Richard Levitte <richard@levitte.org>]
8521 *) New option -out to asn1parse to allow the parsed structure to be
8522 output to a file. This is most useful when combined with the -strparse
8523 option to examine the output of things like OCTET STRINGS.
8526 *) Make SSL library a little more fool-proof by not requiring any longer
8527 that SSL_set_{accept,connect}_state be called before
8528 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8529 in many applications because usually everything *appeared* to work as
8530 intended anyway -- now it really works as intended).
8533 *) Move openssl.cnf out of lib/.
8536 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8537 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8538 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8539 [Ralf S. Engelschall]
8541 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8542 handle PKCS#7 enveloped data properly.
8543 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8545 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8546 copying pointers. The cert_st handling is changed by this in
8547 various ways (and thus what used to be known as ctx->default_cert
8548 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8549 any longer when s->cert does not give us what we need).
8550 ssl_cert_instantiate becomes obsolete by this change.
8551 As soon as we've got the new code right (possibly it already is?),
8552 we have solved a couple of bugs of the earlier code where s->cert
8553 was used as if it could not have been shared with other SSL structures.
8555 Note that using the SSL API in certain dirty ways now will result
8556 in different behaviour than observed with earlier library versions:
8557 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8558 does not influence s as it used to.
8560 In order to clean up things more thoroughly, inside SSL_SESSION
8561 we don't use CERT any longer, but a new structure SESS_CERT
8562 that holds per-session data (if available); currently, this is
8563 the peer's certificate chain and, for clients, the server's certificate
8564 and temporary key. CERT holds only those values that can have
8565 meaningful defaults in an SSL_CTX.
8568 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8569 from the internal representation. Various PKCS#7 fixes: remove some
8570 evil casts and set the enc_dig_alg field properly based on the signing
8574 *) Allow PKCS#12 password to be set from the command line or the
8575 environment. Let 'ca' get its config file name from the environment
8576 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8580 *) Allow certificate policies extension to use an IA5STRING for the
8581 organization field. This is contrary to the PKIX definition but
8582 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8586 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8587 without disallowing inline assembler and the like for non-pedantic builds.
8590 *) Support Borland C++ builder.
8591 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8596 *) SHA-1 cleanups and performance enhancements.
8597 [Andy Polyakov <appro@fy.chalmers.se>]
8599 *) Sparc v8plus assembler for the bignum library.
8600 [Andy Polyakov <appro@fy.chalmers.se>]
8602 *) Accept any -xxx and +xxx compiler options in Configure.
8605 *) Update HPUX configuration.
8608 *) Add missing sk_<type>_unshift() function to safestack.h
8609 [Ralf S. Engelschall]
8611 *) New function SSL_CTX_use_certificate_chain_file that sets the
8612 "extra_cert"s in addition to the certificate. (This makes sense
8613 only for "PEM" format files, as chains as a whole are not
8617 *) Support verify_depth from the SSL API.
8618 x509_vfy.c had what can be considered an off-by-one-error:
8619 Its depth (which was not part of the external interface)
8620 was actually counting the number of certificates in a chain;
8621 now it really counts the depth.
8624 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8625 instead of X509err, which often resulted in confusing error
8626 messages since the error codes are not globally unique
8627 (e.g. an alleged error in ssl3_accept when a certificate
8628 didn't match the private key).
8630 *) New function SSL_CTX_set_session_id_context that allows to set a default
8631 value (so that you don't need SSL_set_session_id_context for each
8632 connection using the SSL_CTX).
8635 *) OAEP decoding bug fix.
8638 *) Support INSTALL_PREFIX for package builders, as proposed by
8642 *) New Configure options "threads" and "no-threads". For systems
8643 where the proper compiler options are known (currently Solaris
8644 and Linux), "threads" is the default.
8647 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8650 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8651 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8652 such as /usr/local/bin.
8655 *) "make linux-shared" to build shared libraries.
8656 [Niels Poppe <niels@netbox.org>]
8658 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8661 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8662 extension adding in x509 utility.
8665 *) Remove NOPROTO sections and error code comments.
8668 *) Partial rewrite of the DEF file generator to now parse the ANSI
8672 *) New Configure options --prefix=DIR and --openssldir=DIR.
8675 *) Complete rewrite of the error code script(s). It is all now handled
8676 by one script at the top level which handles error code gathering,
8677 header rewriting and C source file generation. It should be much better
8678 than the old method: it now uses a modified version of Ulf's parser to
8679 read the ANSI prototypes in all header files (thus the old K&R definitions
8680 aren't needed for error creation any more) and do a better job of
8681 translating function codes into names. The old 'ASN1 error code imbedded
8682 in a comment' is no longer necessary and it doesn't use .err files which
8683 have now been deleted. Also the error code call doesn't have to appear all
8684 on one line (which resulted in some large lines...).
8687 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8690 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8691 0 (which usually indicates a closed connection), but continue reading.
8694 *) Fix some race conditions.
8697 *) Add support for CRL distribution points extension. Add Certificate
8698 Policies and CRL distribution points documentation.
8701 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8704 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8705 8 of keying material. Merlin has also confirmed interop with this fix
8706 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8707 [Merlin Hughes <merlin@baltimore.ie>]
8709 *) Fix lots of warnings.
8710 [Richard Levitte <levitte@stacken.kth.se>]
8712 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8713 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8714 [Richard Levitte <levitte@stacken.kth.se>]
8716 *) Fix problems with sizeof(long) == 8.
8717 [Andy Polyakov <appro@fy.chalmers.se>]
8719 *) Change functions to ANSI C.
8722 *) Fix typos in error codes.
8723 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8725 *) Remove defunct assembler files from Configure.
8728 *) SPARC v8 assembler BIGNUM implementation.
8729 [Andy Polyakov <appro@fy.chalmers.se>]
8731 *) Support for Certificate Policies extension: both print and set.
8732 Various additions to support the r2i method this uses.
8735 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8736 return a const string when you are expecting an allocated buffer.
8739 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8740 types DirectoryString and DisplayText.
8743 *) Add code to allow r2i extensions to access the configuration database,
8744 add an LHASH database driver and add several ctx helper functions.
8747 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8748 fail when they extended the size of a BIGNUM.
8751 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8752 support typesafe stack.
8755 *) Fix typo in SSL_[gs]et_options().
8756 [Nils Frostberg <nils@medcom.se>]
8758 *) Delete various functions and files that belonged to the (now obsolete)
8759 old X509V3 handling code.
8762 *) New Configure option "rsaref".
8765 *) Don't auto-generate pem.h.
8768 *) Introduce type-safe ASN.1 SETs.
8771 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8772 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8774 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8775 that links with OpenSSL (well at least cause lots of warnings), but fear
8776 not: the conversion is trivial, and it eliminates loads of evil casts. A
8777 few STACKed things have been converted already. Feel free to convert more.
8778 In the fullness of time, I'll do away with the STACK type altogether.
8781 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8782 specified in <certfile> by updating the entry in the index.txt file.
8783 This way one no longer has to edit the index.txt file manually for
8784 revoking a certificate. The -revoke option does the gory details now.
8785 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8787 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8788 `-text' option at all and this way the `-noout -text' combination was
8789 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8790 [Ralf S. Engelschall]
8792 *) Make sure a corresponding plain text error message exists for the
8793 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8794 verify callback function determined that a certificate was revoked.
8795 [Ralf S. Engelschall]
8797 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8798 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8799 all available cipers including rc5, which was forgotten until now.
8800 In order to let the testing shell script know which algorithms
8801 are available, a new (up to now undocumented) command
8802 "openssl list-cipher-commands" is used.
8805 *) Bugfix: s_client occasionally would sleep in select() when
8806 it should have checked SSL_pending() first.
8809 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8810 the raw DSA values prior to ASN.1 encoding.
8813 *) Tweaks to Configure
8814 [Niels Poppe <niels@netbox.org>]
8816 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8820 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8823 *) New config option to avoid instructions that are illegal on the 80386.
8824 The default code is faster, but requires at least a 486.
8827 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8828 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8829 same as SSL2_VERSION anyway.
8832 *) New "-showcerts" option for s_client.
8835 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8836 application. Various cleanups and fixes.
8839 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8840 modify error routines to work internally. Add error codes and PBE init
8841 to library startup routines.
8844 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8845 packing functions to asn1 and evp. Changed function names and error
8846 codes along the way.
8849 *) PKCS12 integration: and so it begins... First of several patches to
8850 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8851 objects to objects.h
8854 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8855 and display support for Thawte strong extranet extension.
8858 *) Add LinuxPPC support.
8859 [Jeff Dubrule <igor@pobox.org>]
8861 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8862 bn_div_words in alpha.s.
8863 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8865 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8866 OAEP isn't supported when OpenSSL is built with RSAref.
8867 [Ulf Moeller <ulf@fitug.de>]
8869 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8870 so they no longer are missing under -DNOPROTO.
8871 [Soren S. Jorvang <soren@t.dk>]
8874 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8876 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8877 doesn't work when the session is reused. Coming soon!
8880 *) Fix a security hole, that allows sessions to be reused in the wrong
8881 context thus bypassing client cert protection! All software that uses
8882 client certs and session caches in multiple contexts NEEDS PATCHING to
8883 allow session reuse! A fuller solution is in the works.
8884 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8886 *) Some more source tree cleanups (removed obsolete files
8887 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8888 permission on "config" script to be executable) and a fix for the INSTALL
8890 [Ulf Moeller <ulf@fitug.de>]
8892 *) Remove some legacy and erroneous uses of malloc, free instead of
8894 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8896 *) Make rsa_oaep_test return non-zero on error.
8897 [Ulf Moeller <ulf@fitug.de>]
8899 *) Add support for native Solaris shared libraries. Configure
8900 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8901 if someone would make that last step automatic.
8902 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8904 *) ctx_size was not built with the right compiler during "make links". Fixed.
8907 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8908 except NULL ciphers". This means the default cipher list will no longer
8909 enable NULL ciphers. They need to be specifically enabled e.g. with
8910 the string "DEFAULT:eNULL".
8913 *) Fix to RSA private encryption routines: if p < q then it would
8914 occasionally produce an invalid result. This will only happen with
8915 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8918 *) Be less restrictive and allow also `perl util/perlpath.pl
8919 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8920 because this way one can also use an interpreter named `perl5' (which is
8921 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8922 installed as `perl').
8923 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8925 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8926 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8928 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8929 advapi32.lib to Win32 build and change the pem test comparision
8930 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8931 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8932 and crypto/des/ede_cbcm_enc.c.
8935 *) DES quad checksum was broken on big-endian architectures. Fixed.
8938 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8939 Win32 test batch file so it (might) work again. The Win32 test batch file
8940 is horrible: I feel ill....
8943 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8944 in e_os.h. Audit of header files to check ANSI and non ANSI
8945 sections: 10 functions were absent from non ANSI section and not exported
8946 from Windows DLLs. Fixed up libeay.num for new functions.
8949 *) Make `openssl version' output lines consistent.
8950 [Ralf S. Engelschall]
8952 *) Fix Win32 symbol export lists for BIO functions: Added
8953 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8954 to ms/libeay{16,32}.def.
8955 [Ralf S. Engelschall]
8957 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8958 fine under Unix and passes some trivial tests I've now added. But the
8959 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8960 added to make sure no one expects that this stuff really works in the
8961 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8962 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8964 [Ralf S. Engelschall]
8966 *) Fix the generation of two part addresses in perl.
8967 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8969 *) Add config entry for Linux on MIPS.
8970 [John Tobey <jtobey@channel1.com>]
8972 *) Make links whenever Configure is run, unless we are on Windoze.
8975 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8976 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8980 *) Add a useful kludge to allow package maintainers to specify compiler and
8981 other platforms details on the command line without having to patch the
8982 Configure script everytime: One now can use ``perl Configure
8983 <id>:<details>'', i.e. platform ids are allowed to have details appended
8984 to them (seperated by colons). This is treated as there would be a static
8985 pre-configured entry in Configure's %table under key <id> with value
8986 <details> and ``perl Configure <id>'' is called. So, when you want to
8987 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8988 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8989 now, which overrides the FreeBSD-elf entry on-the-fly.
8990 [Ralf S. Engelschall]
8992 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8995 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8996 on the `perl Configure ...' command line. This way one can compile
8997 OpenSSL libraries with Position Independent Code (PIC) which is needed
8998 for linking it into DSOs.
8999 [Ralf S. Engelschall]
9001 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9005 *) Cleaned up the LICENSE document: The official contact for any license
9006 questions now is the OpenSSL core team under openssl-core@openssl.org.
9007 And add a paragraph about the dual-license situation to make sure people
9008 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9009 to the OpenSSL toolkit.
9010 [Ralf S. Engelschall]
9012 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9013 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9014 Additonally cleaned up the `make links' target: Remove unnecessary
9015 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9016 to speed processing and no longer clutter the display with confusing
9017 stuff. Instead only the actually done links are displayed.
9018 [Ralf S. Engelschall]
9020 *) Permit null encryption ciphersuites, used for authentication only. It used
9021 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9022 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9026 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9027 signed attributes when verifying signatures (this would break them),
9028 the detached data encoding was wrong and public keys obtained using
9029 X509_get_pubkey() weren't freed.
9032 *) Add text documentation for the BUFFER functions. Also added a work around
9033 to a Win95 console bug. This was triggered by the password read stuff: the
9034 last character typed gets carried over to the next fread(). If you were
9035 generating a new cert request using 'req' for example then the last
9036 character of the passphrase would be CR which would then enter the first
9040 *) Added the new `Includes OpenSSL Cryptography Software' button as
9041 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9042 button and can be used by applications based on OpenSSL to show the
9043 relationship to the OpenSSL project.
9044 [Ralf S. Engelschall]
9046 *) Remove confusing variables in function signatures in files
9047 ssl/ssl_lib.c and ssl/ssl.h.
9048 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9050 *) Don't install bss_file.c under PREFIX/include/
9051 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9053 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9054 functions that return function pointers and has support for NT specific
9055 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9056 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9057 unsigned to signed types: this was killing the Win32 compile.
9060 *) Add new certificate file to stack functions,
9061 SSL_add_dir_cert_subjects_to_stack() and
9062 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9063 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9064 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9065 This means that Apache-SSL and similar packages don't have to mess around
9066 to add as many CAs as they want to the preferred list.
9069 *) Experiment with doxygen documentation. Currently only partially applied to
9071 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9072 openssl.doxy as the configuration file.
9075 *) Get rid of remaining C++-style comments which strict C compilers hate.
9076 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9078 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9079 compiled in by default: it has problems with large keys.
9082 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9083 DH private keys and/or callback functions which directly correspond to
9084 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9085 is needed for applications which have to configure certificates on a
9086 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9088 For the RSA certificate situation is makes no difference, but
9089 for the DSA certificate situation this fixes the "no shared cipher"
9090 problem where the OpenSSL cipher selection procedure failed because the
9091 temporary keys were not overtaken from the context and the API provided
9092 no way to reconfigure them.
9093 The new functions now let applications reconfigure the stuff and they
9094 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9095 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9096 non-public-API function ssl_cert_instantiate() is used as a helper
9097 function and also to reduce code redundancy inside ssl_rsa.c.
9098 [Ralf S. Engelschall]
9100 *) Move s_server -dcert and -dkey options out of the undocumented feature
9101 area because they are useful for the DSA situation and should be
9102 recognized by the users.
9103 [Ralf S. Engelschall]
9105 *) Fix the cipher decision scheme for export ciphers: the export bits are
9106 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9107 SSL_EXP_MASK. So, the original variable has to be used instead of the
9108 already masked variable.
9109 [Richard Levitte <levitte@stacken.kth.se>]
9111 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9112 [Richard Levitte <levitte@stacken.kth.se>]
9114 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9115 from `int' to `unsigned int' because it's a length and initialized by
9116 EVP_DigestFinal() which expects an `unsigned int *'.
9117 [Richard Levitte <levitte@stacken.kth.se>]
9119 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9120 script. Instead use the usual Shell->Perl transition trick.
9121 [Ralf S. Engelschall]
9123 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9124 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9125 -noout -modulus' as it's already the case for `openssl rsa -noout
9126 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9127 currently the public key is printed (a decision which was already done by
9128 `openssl dsa -modulus' in the past) which serves a similar purpose.
9129 Additionally the NO_RSA no longer completely removes the whole -modulus
9130 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9132 [Ralf S. Engelschall]
9134 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9135 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9136 [Arne Ansper <arne@ats.cyber.ee>]
9138 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9139 to be added. Now both 'req' and 'ca' can use new objects defined in the
9143 *) Add cool BIO that does syslog (or event log on NT).
9144 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9146 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9147 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9148 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9149 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9152 *) Add preliminary config info for new extension code.
9155 *) Make RSA_NO_PADDING really use no padding.
9156 [Ulf Moeller <ulf@fitug.de>]
9158 *) Generate errors when private/public key check is done.
9161 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9162 for some CRL extensions and new objects added.
9165 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9166 key usage extension and fuller support for authority key id.
9169 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9170 padding method for RSA, which is recommended for new applications in PKCS
9171 #1 v2.0 (RFC 2437, October 1998).
9172 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9173 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9174 against Bleichbacher's attack on RSA.
9175 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9178 *) Updates to the new SSL compression code
9179 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9181 *) Fix so that the version number in the master secret, when passed
9182 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9183 (because the server will not accept higher), that the version number
9184 is 0x03,0x01, not 0x03,0x00
9185 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9187 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9188 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9189 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9192 *) Support for RAW extensions where an arbitrary extension can be
9193 created by including its DER encoding. See apps/openssl.cnf for
9197 *) Make sure latest Perl versions don't interpret some generated C array
9198 code as Perl array code in the crypto/err/err_genc.pl script.
9199 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9201 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9202 not many people have the assembler. Various Win32 compilation fixes and
9203 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9207 *) Modify configure script 'Configure' to automatically create crypto/date.h
9208 file under Win32 and also build pem.h from pem.org. New script
9209 util/mkfiles.pl to create the MINFO file on environments that can't do a
9210 'make files': perl util/mkfiles.pl >MINFO should work.
9213 *) Major rework of DES function declarations, in the pursuit of correctness
9214 and purity. As a result, many evil casts evaporated, and some weirdness,
9215 too. You may find this causes warnings in your code. Zapping your evil
9216 casts will probably fix them. Mostly.
9219 *) Fix for a typo in asn1.h. Bug fix to object creation script
9220 obj_dat.pl. It considered a zero in an object definition to mean
9221 "end of object": none of the objects in objects.h have any zeros
9222 so it wasn't spotted.
9223 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9225 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9226 Masking (CBCM). In the absence of test vectors, the best I have been able
9227 to do is check that the decrypt undoes the encrypt, so far. Send me test
9228 vectors if you have them.
9231 *) Correct calculation of key length for export ciphers (too much space was
9232 allocated for null ciphers). This has not been tested!
9235 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9236 message is now correct (it understands "crypto" and "ssl" on its
9237 command line). There is also now an "update" option. This will update
9238 the util/ssleay.num and util/libeay.num files with any new functions.
9240 perl util/mkdef.pl crypto ssl update
9241 it will update them.
9244 *) Overhauled the Perl interface (perl/*):
9245 - ported BN stuff to OpenSSL's different BN library
9246 - made the perl/ source tree CVS-aware
9247 - renamed the package from SSLeay to OpenSSL (the files still contain
9248 their history because I've copied them in the repository)
9249 - removed obsolete files (the test scripts will be replaced
9250 by better Test::Harness variants in the future)
9251 [Ralf S. Engelschall]
9253 *) First cut for a very conservative source tree cleanup:
9254 1. merge various obsolete readme texts into doc/ssleay.txt
9255 where we collect the old documents and readme texts.
9256 2. remove the first part of files where I'm already sure that we no
9257 longer need them because of three reasons: either they are just temporary
9258 files which were left by Eric or they are preserved original files where
9259 I've verified that the diff is also available in the CVS via "cvs diff
9260 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9261 the crypto/md/ stuff).
9262 [Ralf S. Engelschall]
9264 *) More extension code. Incomplete support for subject and issuer alt
9265 name, issuer and authority key id. Change the i2v function parameters
9266 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9267 what that's for :-) Fix to ASN1 macro which messed up
9268 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9271 *) Preliminary support for ENUMERATED type. This is largely copied from the
9275 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9276 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9278 *) Make sure `make rehash' target really finds the `openssl' program.
9279 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9281 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9282 like to hear about it if this slows down other processors.
9285 *) Add CygWin32 platform information to Configure script.
9286 [Alan Batie <batie@aahz.jf.intel.com>]
9288 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9289 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9291 *) New program nseq to manipulate netscape certificate sequences
9294 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9298 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9299 but the BN code had some problems that would cause failures when
9300 doing certificate verification and some other functions.
9301 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9303 *) Add ASN1 and PEM code to support netscape certificate sequences.
9306 *) Add ASN1 and PEM code to support netscape certificate sequences.
9309 *) Add several PKIX and private extended key usage OIDs.
9312 *) Modify the 'ca' program to handle the new extension code. Modify
9313 openssl.cnf for new extension format, add comments.
9316 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9317 and add a sample to openssl.cnf so req -x509 now adds appropriate
9321 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9322 error code, add initial support to X509_print() and x509 application.
9325 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9326 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9327 stuff is currently isolated and isn't even compiled yet.
9330 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9331 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9332 Removed the versions check from X509 routines when loading extensions:
9333 this allows certain broken certificates that don't set the version
9334 properly to be processed.
9337 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9338 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9339 can still be regenerated with "make depend".
9342 *) Spelling mistake in C version of CAST-128.
9343 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9345 *) Changes to the error generation code. The perl script err-code.pl
9346 now reads in the old error codes and retains the old numbers, only
9347 adding new ones if necessary. It also only changes the .err files if new
9348 codes are added. The makefiles have been modified to only insert errors
9349 when needed (to avoid needlessly modifying header files). This is done
9350 by only inserting errors if the .err file is newer than the auto generated
9351 C file. To rebuild all the error codes from scratch (the old behaviour)
9352 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9353 or delete all the .err files.
9356 *) CAST-128 was incorrectly implemented for short keys. The C version has
9357 been fixed, but is untested. The assembler versions are also fixed, but
9358 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9359 to regenerate it if needed.
9360 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9361 Hagino <itojun@kame.net>]
9363 *) File was opened incorrectly in randfile.c.
9364 [Ulf Möller <ulf@fitug.de>]
9366 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9367 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9368 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9369 al: it's just almost always a UTCTime. Note this patch adds new error
9370 codes so do a "make errors" if there are problems.
9373 *) Correct Linux 1 recognition in config.
9374 [Ulf Möller <ulf@fitug.de>]
9376 *) Remove pointless MD5 hash when using DSA keys in ca.
9377 [Anonymous <nobody@replay.com>]
9379 *) Generate an error if given an empty string as a cert directory. Also
9380 generate an error if handed NULL (previously returned 0 to indicate an
9381 error, but didn't set one).
9382 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9384 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9387 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9388 parameters. This was causing a warning which killed off the Win32 compile.
9391 *) Remove C++ style comments from crypto/bn/bn_local.h.
9392 [Neil Costigan <neil.costigan@celocom.com>]
9394 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9395 based on a text string, looking up short and long names and finally
9396 "dot" format. The "dot" format stuff didn't work. Added new function
9397 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9398 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9399 OID is not part of the table.
9402 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9403 X509_LOOKUP_by_alias().
9406 *) Sort openssl functions by name.
9409 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9410 encryption from sample DSA keys (in case anyone is interested the password
9414 *) Make _all_ *_free functions accept a NULL pointer.
9415 [Frans Heymans <fheymans@isaserver.be>]
9417 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9419 [Anonymous <nobody@replay.com>]
9421 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9422 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9424 *) Don't blow it for numeric -newkey arguments to apps/req.
9425 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9427 *) Temp key "for export" tests were wrong in s3_srvr.c.
9428 [Anonymous <nobody@replay.com>]
9430 *) Add prototype for temp key callback functions
9431 SSL_CTX_set_tmp_{rsa,dh}_callback().
9434 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9435 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9438 *) X509_name_add_entry() freed the wrong thing after an error.
9439 [Arne Ansper <arne@ats.cyber.ee>]
9441 *) rsa_eay.c would attempt to free a NULL context.
9442 [Arne Ansper <arne@ats.cyber.ee>]
9444 *) BIO_s_socket() had a broken should_retry() on Windoze.
9445 [Arne Ansper <arne@ats.cyber.ee>]
9447 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9448 [Arne Ansper <arne@ats.cyber.ee>]
9450 *) Make sure the already existing X509_STORE->depth variable is initialized
9451 in X509_STORE_new(), but document the fact that this variable is still
9452 unused in the certificate verification process.
9453 [Ralf S. Engelschall]
9455 *) Fix the various library and apps files to free up pkeys obtained from
9456 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9459 *) Fix reference counting in X509_PUBKEY_get(). This makes
9460 demos/maurice/example2.c work, amongst others, probably.
9461 [Steve Henson and Ben Laurie]
9463 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9464 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9465 are no longer created. This way we have a single and consistent command
9466 line interface `openssl <command>', similar to `cvs <command>'.
9467 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9469 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9470 BIT STRING wrapper always have zero unused bits.
9473 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9476 *) Make the top-level INSTALL documentation easier to understand.
9479 *) Makefiles updated to exit if an error occurs in a sub-directory
9480 make (including if user presses ^C) [Paul Sutton]
9482 *) Make Montgomery context stuff explicit in RSA data structure.
9485 *) Fix build order of pem and err to allow for generated pem.h.
9488 *) Fix renumbering bug in X509_NAME_delete_entry().
9491 *) Enhanced the err-ins.pl script so it makes the error library number
9492 global and can add a library name. This is needed for external ASN1 and
9493 other error libraries.
9496 *) Fixed sk_insert which never worked properly.
9499 *) Fix ASN1 macros so they can handle indefinite length construted
9500 EXPLICIT tags. Some non standard certificates use these: they can now
9504 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9505 into a single doc/ssleay.txt bundle. This way the information is still
9506 preserved but no longer messes up this directory. Now it's new room for
9507 the new set of documenation files.
9508 [Ralf S. Engelschall]
9510 *) SETs were incorrectly DER encoded. This was a major pain, because they
9511 shared code with SEQUENCEs, which aren't coded the same. This means that
9512 almost everything to do with SETs or SEQUENCEs has either changed name or
9513 number of arguments.
9514 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9516 *) Fix test data to work with the above.
9519 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9520 was already fixed by Eric for 0.9.1 it seems.
9521 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9523 *) Autodetect FreeBSD3.
9526 *) Fix various bugs in Configure. This affects the following platforms:
9530 unixware-2.0-pentium
9534 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9535 before they are needed.
9538 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9542 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9544 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9545 changed SSLeay to OpenSSL in version strings.
9546 [Ralf S. Engelschall]
9548 *) Some fixups to the top-level documents.
9551 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9552 because the symlink to include/ was missing.
9553 [Ralf S. Engelschall]
9555 *) Incorporated the popular no-RSA/DSA-only patches
9556 which allow to compile a RSA-free SSLeay.
9557 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9559 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9560 when "ssleay" is still not found.
9561 [Ralf S. Engelschall]
9563 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9564 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9566 *) Updated the README file.
9567 [Ralf S. Engelschall]
9569 *) Added various .cvsignore files in the CVS repository subdirs
9570 to make a "cvs update" really silent.
9571 [Ralf S. Engelschall]
9573 *) Recompiled the error-definition header files and added
9574 missing symbols to the Win32 linker tables.
9575 [Ralf S. Engelschall]
9577 *) Cleaned up the top-level documents;
9578 o new files: CHANGES and LICENSE
9579 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9580 o merged COPYRIGHT into LICENSE
9581 o removed obsolete TODO file
9582 o renamed MICROSOFT to INSTALL.W32
9583 [Ralf S. Engelschall]
9585 *) Removed dummy files from the 0.9.1b source tree:
9586 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9587 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9588 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9589 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9590 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9591 [Ralf S. Engelschall]
9593 *) Added various platform portability fixes.
9596 *) The Genesis of the OpenSSL rpject:
9597 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9598 Young and Tim J. Hudson created while they were working for C2Net until
9600 [The OpenSSL Project]
9603 Changes between 0.9.0b and 0.9.1b [not released]
9605 *) Updated a few CA certificates under certs/
9608 *) Changed some BIGNUM api stuff.
9611 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9612 DGUX x86, Linux Alpha, etc.
9615 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9616 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9620 *) Add -strparse option to asn1pars program which parses nested
9622 [Dr Stephen Henson <shenson@bigfoot.com>]
9624 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9627 *) DSA fix for "ca" program.
9630 *) Added "-genkey" option to "dsaparam" program.
9633 *) Added RIPE MD160 (rmd160) message digest.
9636 *) Added -a (all) option to "ssleay version" command.
9639 *) Added PLATFORM define which is the id given to Configure.
9642 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9645 *) Extended the ASN.1 parser routines.
9648 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9651 *) Added a BN_CTX to the BN library.
9654 *) Fixed the weak key values in DES library
9657 *) Changed API in EVP library for cipher aliases.
9660 *) Added support for RC2/64bit cipher.
9663 *) Converted the lhash library to the crypto/mem.c functions.
9666 *) Added more recognized ASN.1 object ids.
9669 *) Added more RSA padding checks for SSL/TLS.
9672 *) Added BIO proxy/filter functionality.
9675 *) Added extra_certs to SSL_CTX which can be used
9676 send extra CA certificates to the client in the CA cert chain sending
9677 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9680 *) Now Fortezza is denied in the authentication phase because
9681 this is key exchange mechanism is not supported by SSLeay at all.
9684 *) Additional PKCS1 checks.
9687 *) Support the string "TLSv1" for all TLS v1 ciphers.
9690 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9691 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9694 *) Fixed a few memory leaks.
9697 *) Fixed various code and comment typos.
9700 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9701 bytes sent in the client random.
9702 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob