5 Changes between 1.0.0 and 1.1.0 [xx XXX xxxx]
7 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8 a few changes are required:
10 Add SSL_OP_NO_TLSv1_1 flag.
12 Update version checking logic to handle version 1.1.
13 Add explicit IV handling (ported from DTLS code).
14 Add command line options to s_client/s_server.
17 *) Experiemental password based recipient info support for CMS library:
21 *) Split password based encryption into PBES2 and PBKDF2 functions. This
22 neatly separates the code into cipher and PBE sections and is required
23 for some algorithms that split PBES2 into separate pieces (such as
27 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
28 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
29 BIO_set_cipher() and some obscure PEM functions were changed so they
30 can now return an error. The RAND changes required a change to the
31 RAND_METHOD structure.
34 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
35 a gcc attribute to warn if the result of a function is ignored. This
36 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
37 whose return value is often ignored.
40 Changes between 0.9.8m (?) and 1.0.0 [xx XXX xxxx]
42 *) Add load_crls() function to apps tidying load_certs() too. Add option
43 to verify utility to allow additional CRLs to be included.
46 *) Update OCSP request code to permit adding custom headers to the request:
47 some responders need this.
50 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
52 [Julia Lawall <julia@diku.dk>]
54 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
55 needlessly dereferenced structures, used obsolete functions and
56 didn't handle all updated verify codes correctly.
59 *) Disable MD2 in the default configuration.
62 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
63 indicate the initial BIO being pushed or popped. This makes it possible
64 to determine whether the BIO is the one explicitly called or as a result
65 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
66 it handles reference counts correctly and doesn't zero out the I/O bio
67 when it is not being explicitly popped. WARNING: applications which
68 included workarounds for the old buggy behaviour will need to be modified
69 or they could free up already freed BIOs.
72 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
73 renaming to all platforms (within the 0.9.8 branch, this was
74 done conditionally on Netware platforms to avoid a name clash).
75 [Guenter <lists@gknw.net>]
77 *) Add ECDHE and PSK support to DTLS.
78 [Michael Tuexen <tuexen@fh-muenster.de>]
80 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
84 *) Add "missing" function EVP_MD_flags() (without this the only way to
85 retrieve a digest flags is by accessing the structure directly. Update
86 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
87 or cipher is registered as in the "from" argument. Print out all
88 registered digests in the dgst usage message instead of manually
89 attempting to work them out.
92 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
93 this allows the use of compression and extensions. Change default cipher
94 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
95 by default unless an application cipher string requests it.
98 *) Alter match criteria in PKCS12_parse(). It used to try to use local
99 key ids to find matching certificates and keys but some PKCS#12 files
100 don't follow the (somewhat unwritten) rules and this strategy fails.
101 Now just gather all certificates together and the first private key
102 then look for the first certificate that matches the key.
105 *) Support use of registered digest and cipher names for dgst and cipher
106 commands instead of having to add each one as a special case. So now
113 openssl dgst -sha256 foo
115 and this works for ENGINE based algorithms too.
119 *) Update Gost ENGINE to support parameter files.
120 [Victor B. Wagner <vitus@cryptocom.ru>]
122 *) Support GeneralizedTime in ca utility.
123 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
125 *) Enhance the hash format used for certificate directory links. The new
126 form uses the canonical encoding (meaning equivalent names will work
127 even if they aren't identical) and uses SHA1 instead of MD5. This form
128 is incompatible with the older format and as a result c_rehash should
129 be used to rebuild symbolic links.
132 *) Make PKCS#8 the default write format for private keys, replacing the
133 traditional format. This form is standardised, more secure and doesn't
134 include an implicit MD5 dependency.
137 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
138 committed to OpenSSL should pass this lot as a minimum.
141 *) Add session ticket override functionality for use by EAP-FAST.
142 [Jouni Malinen <j@w1.fi>]
144 *) Modify HMAC functions to return a value. Since these can be implemented
145 in an ENGINE errors can occur.
148 *) Type-checked OBJ_bsearch_ex.
151 *) Type-checked OBJ_bsearch. Also some constification necessitated
152 by type-checking. Still to come: TXT_DB, bsearch(?),
153 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
157 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
158 seconds to a tm structure directly, instead of going through OS
159 specific date routines. This avoids any issues with OS routines such
160 as the year 2038 bug. New *_adj() functions for ASN1 time structures
161 and X509_time_adj_ex() to cover the extended range. The existing
162 X509_time_adj() is still usable and will no longer have any date issues.
165 *) Delta CRL support. New use deltas option which will attempt to locate
166 and search any appropriate delta CRLs available.
168 This work was sponsored by Google.
171 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
172 code and add additional score elements. Validate alternate CRL paths
173 as part of the CRL checking and indicate a new error "CRL path validation
174 error" in this case. Applications wanting additional details can use
175 the verify callback and check the new "parent" field. If this is not
176 NULL CRL path validation is taking place. Existing applications wont
177 see this because it requires extended CRL support which is off by
180 This work was sponsored by Google.
183 *) Support for freshest CRL extension.
185 This work was sponsored by Google.
188 *) Initial indirect CRL support. Currently only supported in the CRLs
189 passed directly and not via lookup. Process certificate issuer
190 CRL entry extension and lookup CRL entries by bother issuer name
191 and serial number. Check and process CRL issuer entry in IDP extension.
193 This work was sponsored by Google.
196 *) Add support for distinct certificate and CRL paths. The CRL issuer
197 certificate is validated separately in this case. Only enabled if
198 an extended CRL support flag is set: this flag will enable additional
199 CRL functionality in future.
201 This work was sponsored by Google.
204 *) Add support for policy mappings extension.
206 This work was sponsored by Google.
209 *) Fixes to pathlength constraint, self issued certificate handling,
210 policy processing to align with RFC3280 and PKITS tests.
212 This work was sponsored by Google.
215 *) Support for name constraints certificate extension. DN, email, DNS
216 and URI types are currently supported.
218 This work was sponsored by Google.
221 *) To cater for systems that provide a pointer-based thread ID rather
222 than numeric, deprecate the current numeric thread ID mechanism and
223 replace it with a structure and associated callback type. This
224 mechanism allows a numeric "hash" to be extracted from a thread ID in
225 either case, and on platforms where pointers are larger than 'long',
226 mixing is done to help ensure the numeric 'hash' is usable even if it
227 can't be guaranteed unique. The default mechanism is to use "&errno"
228 as a pointer-based thread ID to distinguish between threads.
230 Applications that want to provide their own thread IDs should now use
231 CRYPTO_THREADID_set_callback() to register a callback that will call
232 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
234 Note that ERR_remove_state() is now deprecated, because it is tied
235 to the assumption that thread IDs are numeric. ERR_remove_state(0)
236 to free the current thread's error state should be replaced by
237 ERR_remove_thread_state(NULL).
239 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
240 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
241 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
242 application was previously providing a numeric thread callback that
243 was inappropriate for distinguishing threads, then uniqueness might
244 have been obtained with &errno that happened immediately in the
245 intermediate development versions of OpenSSL; this is no longer the
246 case, the numeric thread callback will now override the automatic use
248 [Geoff Thorpe, with help from Bodo Moeller]
250 *) Initial support for different CRL issuing certificates. This covers a
251 simple case where the self issued certificates in the chain exist and
252 the real CRL issuer is higher in the existing chain.
254 This work was sponsored by Google.
257 *) Removed effectively defunct crypto/store from the build.
260 *) Revamp of STACK to provide stronger type-checking. Still to come:
261 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
262 ASN1_STRING, CONF_VALUE.
265 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
266 RAM on SSL connections. This option can save about 34k per idle SSL.
269 *) Revamp of LHASH to provide stronger type-checking. Still to come:
270 STACK, TXT_DB, bsearch, qsort.
273 *) Initial support for Cryptographic Message Syntax (aka CMS) based
274 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
275 support for data, signedData, compressedData, digestedData and
276 encryptedData, envelopedData types included. Scripts to check against
277 RFC4134 examples draft and interop and consistency checks of many
278 content types and variants.
281 *) Add options to enc utility to support use of zlib compression BIO.
284 *) Extend mk1mf to support importing of options and assembly language
285 files from Configure script, currently only included in VC-WIN32.
286 The assembly language rules can now optionally generate the source
287 files from the associated perl scripts.
290 *) Implement remaining functionality needed to support GOST ciphersuites.
291 Interop testing has been performed using CryptoPro implementations.
292 [Victor B. Wagner <vitus@cryptocom.ru>]
294 *) s390x assembler pack.
297 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
301 *) Implement Opaque PRF Input TLS extension as specified in
302 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
303 official specification yet and no extension type assignment by
304 IANA exists, this extension (for now) will have to be explicitly
305 enabled when building OpenSSL by providing the extension number
306 to use. For example, specify an option
308 -DTLSEXT_TYPE_opaque_prf_input=0x9527
310 to the "config" or "Configure" script to enable the extension,
311 assuming extension number 0x9527 (which is a completely arbitrary
312 and unofficial assignment based on the MD5 hash of the Internet
313 Draft). Note that by doing so, you potentially lose
314 interoperability with other TLS implementations since these might
315 be using the same extension number for other purposes.
317 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
318 opaque PRF input value to use in the handshake. This will create
319 an interal copy of the length-'len' string at 'src', and will
320 return non-zero for success.
322 To get more control and flexibility, provide a callback function
325 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
326 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
330 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
333 Callback function 'cb' will be called in handshakes, and is
334 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
335 Argument 'arg' is for application purposes (the value as given to
336 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
337 be provided to the callback function). The callback function
338 has to return non-zero to report success: usually 1 to use opaque
339 PRF input just if possible, or 2 to enforce use of the opaque PRF
340 input. In the latter case, the library will abort the handshake
341 if opaque PRF input is not successfully negotiated.
343 Arguments 'peerinput' and 'len' given to the callback function
344 will always be NULL and 0 in the case of a client. A server will
345 see the client's opaque PRF input through these variables if
346 available (NULL and 0 otherwise). Note that if the server
347 provides an opaque PRF input, the length must be the same as the
348 length of the client's opaque PRF input.
350 Note that the callback function will only be called when creating
351 a new session (session resumption can resume whatever was
352 previously negotiated), and will not be called in SSL 2.0
353 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
354 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
355 for applications that need to enforce opaque PRF input.
359 *) Update ssl code to support digests other than SHA1+MD5 for handshake
362 [Victor B. Wagner <vitus@cryptocom.ru>]
364 *) Add RFC4507 support to OpenSSL. This includes the corrections in
365 RFC4507bis. The encrypted ticket format is an encrypted encoded
366 SSL_SESSION structure, that way new session features are automatically
369 If a client application caches session in an SSL_SESSION structure
370 support is transparent because tickets are now stored in the encoded
373 The SSL_CTX structure automatically generates keys for ticket
374 protection in servers so again support should be possible
375 with no application modification.
377 If a client or server wishes to disable RFC4507 support then the option
378 SSL_OP_NO_TICKET can be set.
380 Add a TLS extension debugging callback to allow the contents of any client
381 or server extensions to be examined.
383 This work was sponsored by Google.
386 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
387 OpenSSL should now compile cleanly on gcc 4.2
388 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
390 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
391 support including streaming MAC support: this is required for GOST
393 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
395 *) Add option -stream to use PKCS#7 streaming in smime utility. New
396 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
397 to output in BER and PEM format.
400 *) Experimental support for use of HMAC via EVP_PKEY interface. This
401 allows HMAC to be handled via the EVP_DigestSign*() interface. The
402 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
403 ENGINE support for HMAC keys which are unextractable. New -mac and
404 -macopt options to dgst utility.
407 *) New option -sigopt to dgst utility. Update dgst to use
408 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
409 alternative signing paramaters such as X9.31 or PSS in the dgst
413 *) Change ssl_cipher_apply_rule(), the internal function that does
414 the work each time a ciphersuite string requests enabling
415 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
416 removing ("!foo+bar") a class of ciphersuites: Now it maintains
417 the order of disabled ciphersuites such that those ciphersuites
418 that most recently went from enabled to disabled not only stay
419 in order with respect to each other, but also have higher priority
420 than other disabled ciphersuites the next time ciphersuites are
423 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
424 the same ciphersuites as with "HIGH" alone, but in a specific
425 order where the PSK ciphersuites come first (since they are the
426 most recently disabled ciphersuites when "HIGH" is parsed).
428 Also, change ssl_create_cipher_list() (using this new
429 funcionality) such that between otherwise identical
430 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
434 *) Change ssl_create_cipher_list() so that it automatically
435 arranges the ciphersuites in reasonable order before starting
436 to process the rule string. Thus, the definition for "DEFAULT"
437 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
438 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
439 This makes it much easier to arrive at a reasonable default order
440 in applications for which anonymous ciphers are OK (meaning
441 that you can't actually use DEFAULT).
442 [Bodo Moeller; suggested by Victor Duchovni]
444 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
445 processing) into multiple integers instead of setting
446 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
447 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
448 (These masks as well as the individual bit definitions are hidden
449 away into the non-exported interface ssl/ssl_locl.h, so this
450 change to the definition of the SSL_CIPHER structure shouldn't
451 affect applications.) This give us more bits for each of these
452 categories, so there is no longer a need to coagulate AES128 and
453 AES256 into a single algorithm bit, and to coagulate Camellia128
454 and Camellia256 into a single algorithm bit, which has led to all
457 Thus, among other things, the kludge introduced in 0.9.7m and
458 0.9.8e for masking out AES256 independently of AES128 or masking
459 out Camellia256 independently of AES256 is not needed here in 0.9.9.
461 With the change, we also introduce new ciphersuite aliases that
462 so far were missing: "AES128", "AES256", "CAMELLIA128", and
466 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
467 Use the leftmost N bytes of the signature input if the input is
468 larger than the prime q (with N being the size in bytes of q).
471 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
472 it yet and it is largely untested.
475 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
478 *) Initial incomplete changes to avoid need for function casts in OpenSSL
479 some compilers (gcc 4.2 and later) reject their use. Safestack is
480 reimplemented. Update ASN1 to avoid use of legacy functions.
483 *) Win32/64 targets are linked with Winsock2.
486 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
487 to external functions. This can be used to increase CRL handling
488 efficiency especially when CRLs are very large by (for example) storing
489 the CRL revoked certificates in a database.
492 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
493 new CRLs added to a directory can be used. New command line option
494 -verify_return_error to s_client and s_server. This causes real errors
495 to be returned by the verify callback instead of carrying on no matter
496 what. This reflects the way a "real world" verify callback would behave.
499 *) GOST engine, supporting several GOST algorithms and public key formats.
500 Kindly donated by Cryptocom.
503 *) Partial support for Issuing Distribution Point CRL extension. CRLs
504 partitioned by DP are handled but no indirect CRL or reason partitioning
505 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
506 selected via a scoring technique which handles IDP and AKID in CRLs.
509 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
510 will ultimately be used for all verify operations: this will remove the
511 X509_STORE dependency on certificate verification and allow alternative
512 lookup methods. X509_STORE based implementations of these two callbacks.
515 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
516 Modify get_crl() to find a valid (unexpired) CRL if possible.
519 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
520 this would be called X509_CRL_cmp() but that name is already used by
521 a function that just compares CRL issuer names. Cache several CRL
522 extensions in X509_CRL structure and cache CRLDP in X509.
525 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
526 this maps equivalent X509_NAME structures into a consistent structure.
527 Name comparison can then be performed rapidly using memcmp().
530 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
534 *) Allow digests to supply their own micalg string for S/MIME type using
535 the ctrl EVP_MD_CTRL_MICALG.
538 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
539 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
540 ctrl. It can then customise the structure before and/or after signing
544 *) New function OBJ_add_sigid() to allow application defined signature OIDs
545 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
546 to free up any added signature OIDs.
549 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
550 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
551 digest and cipher tables. New options added to openssl utility:
552 list-message-digest-algorithms and list-cipher-algorithms.
555 *) Change the array representation of binary polynomials: the list
556 of degrees of non-zero coefficients is now terminated with -1.
557 Previously it was terminated with 0, which was also part of the
558 value; thus, the array representation was not applicable to
559 polynomials where t^0 has coefficient zero. This change makes
560 the array representation useful in a more general context.
563 *) Various modifications and fixes to SSL/TLS cipher string
564 handling. For ECC, the code now distinguishes between fixed ECDH
565 with RSA certificates on the one hand and with ECDSA certificates
566 on the other hand, since these are separate ciphersuites. The
567 unused code for Fortezza ciphersuites has been removed.
569 For consistency with EDH, ephemeral ECDH is now called "EECDH"
570 (not "ECDHE"). For consistency with the code for DH
571 certificates, use of ECDH certificates is now considered ECDH
572 authentication, not RSA or ECDSA authentication (the latter is
573 merely the CA's signing algorithm and not actively used in the
576 The temporary ciphersuite alias "ECCdraft" is no longer
577 available, and ECC ciphersuites are no longer excluded from "ALL"
578 and "DEFAULT". The following aliases now exist for RFC 4492
579 ciphersuites, most of these by analogy with the DH case:
581 kECDHr - ECDH cert, signed with RSA
582 kECDHe - ECDH cert, signed with ECDSA
583 kECDH - ECDH cert (signed with either RSA or ECDSA)
584 kEECDH - ephemeral ECDH
585 ECDH - ECDH cert or ephemeral ECDH
591 AECDH - anonymous ECDH
592 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
596 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
597 Use correct micalg parameters depending on digest(s) in signed message.
600 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
601 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
604 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
605 an engine to register a method. Add ENGINE lookups for methods and
606 functional reference processing.
609 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
610 EVP_{Sign,Verify}* which allow an application to customise the signature
614 *) New -resign option to smime utility. This adds one or more signers
615 to an existing PKCS#7 signedData structure. Also -md option to use an
616 alternative message digest algorithm for signing.
619 *) Tidy up PKCS#7 routines and add new functions to make it easier to
620 create PKCS7 structures containing multiple signers. Update smime
621 application to support multiple signers.
624 *) New -macalg option to pkcs12 utility to allow setting of an alternative
628 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
629 Reorganize PBE internals to lookup from a static table using NIDs,
630 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
631 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
632 PRF which will be automatically used with PBES2.
635 *) Replace the algorithm specific calls to generate keys in "req" with the
639 *) Update PKCS#7 enveloped data routines to use new API. This is now
640 supported by any public key method supporting the encrypt operation. A
641 ctrl is added to allow the public key algorithm to examine or modify
642 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
646 *) Add a ctrl to asn1 method to allow a public key algorithm to express
647 a default digest type to use. In most cases this will be SHA1 but some
648 algorithms (such as GOST) need to specify an alternative digest. The
649 return value indicates how strong the prefernce is 1 means optional and
650 2 is mandatory (that is it is the only supported type). Modify
651 ASN1_item_sign() to accept a NULL digest argument to indicate it should
652 use the default md. Update openssl utilities to use the default digest
653 type for signing if it is not explicitly indicated.
656 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
657 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
658 signing method from the key type. This effectively removes the link
659 between digests and public key types.
662 *) Add an OID cross reference table and utility functions. Its purpose is to
663 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
664 rsaEncryption. This will allow some of the algorithm specific hackery
665 needed to use the correct OID to be removed.
668 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
669 structures for PKCS7_sign(). They are now set up by the relevant public
673 *) Add provisional EC pkey method with support for ECDSA and ECDH.
676 *) Add support for key derivation (agreement) in the API, DH method and
680 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
681 public and private key formats. As a side effect these add additional
682 command line functionality not previously available: DSA signatures can be
683 generated and verified using pkeyutl and DH key support and generation in
688 [Oliver Tappe <zooey@hirschkaefer.de>]
690 *) New make target "install_html_docs" installs HTML renditions of the
692 [Oliver Tappe <zooey@hirschkaefer.de>]
694 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
695 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
696 support key and parameter generation and add initial key generation
697 functionality for RSA.
700 *) Add functions for main EVP_PKEY_method operations. The undocumented
701 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
702 EVP_PKEY_{encrypt,decrypt}_old.
705 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
706 key API, doesn't do much yet.
709 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
710 public key algorithms. New option to openssl utility:
711 "list-public-key-algorithms" to print out info.
714 *) Implement the Supported Elliptic Curves Extension for
715 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
718 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
719 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
722 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
723 utilities such as rsa, dsa, dsaparam etc except they process any key
727 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
728 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
729 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
733 *) Initial support for pluggable public key ASN1.
734 De-spaghettify the public key ASN1 handling. Move public and private
735 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
736 algorithm specific handling to a single module within the relevant
737 algorithm directory. Add functions to allow (near) opaque processing
738 of public and private key structures.
741 *) Implement the Supported Point Formats Extension for
742 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
745 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
746 for the psk identity [hint] and the psk callback functions to the
747 SSL_SESSION, SSL and SSL_CTX structure.
750 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
754 SSL_CTX_use_psk_identity_hint
755 SSL_get_psk_identity_hint
757 SSL_use_psk_identity_hint
759 [Mika Kousa and Pasi Eronen of Nokia Corporation]
761 *) Add RFC 3161 compliant time stamp request creation, response generation
762 and response verification functionality.
763 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
765 *) Add initial support for TLS extensions, specifically for the server_name
766 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
767 have new members for a host name. The SSL data structure has an
768 additional member SSL_CTX *initial_ctx so that new sessions can be
769 stored in that context to allow for session resumption, even after the
770 SSL has been switched to a new SSL_CTX in reaction to a client's
771 server_name extension.
773 New functions (subject to change):
776 SSL_get_servername_type()
779 New CTRL codes and macros (subject to change):
781 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
782 - SSL_CTX_set_tlsext_servername_callback()
783 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
784 - SSL_CTX_set_tlsext_servername_arg()
785 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
787 openssl s_client has a new '-servername ...' option.
789 openssl s_server has new options '-servername_host ...', '-cert2 ...',
790 '-key2 ...', '-servername_fatal' (subject to change). This allows
791 testing the HostName extension for a specific single host name ('-cert'
792 and '-key' remain fallbacks for handshakes without HostName
793 negotiation). If the unrecogninzed_name alert has to be sent, this by
794 default is a warning; it becomes fatal with the '-servername_fatal'
797 [Peter Sylvester, Remy Allais, Christophe Renou]
799 *) Whirlpool hash implementation is added.
802 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
803 bn(64,32). Because of instruction set limitations it doesn't have
804 any negative impact on performance. This was done mostly in order
805 to make it possible to share assembler modules, such as bn_mul_mont
806 implementations, between 32- and 64-bit builds without hassle.
809 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
810 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
814 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
815 dedicated Montgomery multiplication procedure, is introduced.
816 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
817 "64-bit" performance on certain 32-bit targets.
820 *) New option SSL_OP_NO_COMP to disable use of compression selectively
821 in SSL structures. New SSL ctrl to set maximum send fragment size.
822 Save memory by seeting the I/O buffer sizes dynamically instead of
823 using the maximum available value.
826 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
827 in addition to the text details.
830 *) Very, very preliminary EXPERIMENTAL support for printing of general
831 ASN1 structures. This currently produces rather ugly output and doesn't
832 handle several customised structures at all.
835 *) Integrated support for PVK file format and some related formats such
836 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
837 these in the 'rsa' and 'dsa' utilities.
840 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
843 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
844 place for the (very old) "NETSCAPE" format certificates which are now
845 handled using new ASN1 code equivalents.
848 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
849 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
850 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
853 *) Modify CRL distribution points extension code to print out previously
854 unsupported fields. Enhance extension setting code to allow setting of
858 *) Add print and set support for Issuing Distribution Point CRL extension.
861 *) Change 'Configure' script to enable Camellia by default.
864 Changes between 0.9.8l (?) and 0.9.8m (?) [xx XXX xxxx]
866 *) If client attempts to renegotiate and doesn't support RI respond with
867 a no_renegotiation alert as required by draft-ietf-tls-renegotiation.
868 Some renegotiating TLS clients will continue a connection gracefully
869 when they receive the alert. Unfortunately OpenSSL mishandled
870 this alert and would hang waiting for a server hello which it will never
871 receive. Now we treat a received no_renegotiation alert as a fatal
872 error. This is because applications requesting a renegotiation might well
873 expect it to succeed and would have no code in place to handle the server
874 denying it so the only safe thing to do is to terminate the connection.
877 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
878 peer supports secure renegotiation and 0 otherwise. Print out peer
879 renegotiation support in s_client/s_server.
882 *) Replace the highly broken and deprecated SPKAC certification method with
883 the updated NID creation version. This should correctly handle UTF8.
886 *) Implement draft-ietf-tls-renegotiation. Re-enable
887 renegotiation but require the extension as needed. Unfortunately,
888 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION turns out to be a
889 bad idea. It has been replaced by
890 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
891 SSL_CTX_set_options(). This is really not recommended unless you
892 know what you are doing.
893 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
895 *) Fixes to stateless session resumption handling. Use initial_ctx when
896 issuing and attempting to decrypt tickets in case it has changed during
897 servername handling. Use a non-zero length session ID when attempting
898 stateless session resumption: this makes it possible to determine if
899 a resumption has occurred immediately after receiving server hello
900 (several places in OpenSSL subtly assume this) instead of later in
904 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
905 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
906 fixes for a few places where the return code is not checked
908 [Julia Lawall <julia@diku.dk>]
910 *) Add --strict-warnings option to Configure script to include devteam
911 warnings in other configurations.
914 *) Add support for --libdir option and LIBDIR variable in makefiles. This
915 makes it possible to install openssl libraries in locations which
916 have names other than "lib", for example "/usr/lib64" which some
918 [Steve Henson, based on patch from Jeremy Utley]
920 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
921 X690 8.9.12 and can produce some misleading textual output of OIDs.
922 [Steve Henson, reported by Dan Kaminsky]
924 *) Delete MD2 from algorithm tables. This follows the recommendation in
925 several standards that it is not used in new applications due to
926 several cryptographic weaknesses. For binary compatibility reasons
927 the MD2 API is still compiled in by default.
930 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
934 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
935 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
937 [Guenter <lists@gknw.net>]
939 *) Fix the server certificate chain building code to use X509_verify_cert(),
940 it used to have an ad-hoc builder which was unable to cope with anything
941 other than a simple chain.
942 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
944 *) Don't check self signed certificate signatures in X509_verify_cert()
945 by default (a flag can override this): it just wastes time without
946 adding any security. As a useful side effect self signed root CAs
947 with non-FIPS digests are now usable in FIPS mode.
950 *) In dtls1_process_out_of_seq_message() the check if the current message
951 is already buffered was missing. For every new message was memory
952 allocated, allowing an attacker to perform an denial of service attack
953 with sending out of seq handshake messages until there is no memory
954 left. Additionally every future messege was buffered, even if the
955 sequence number made no sense and would be part of another handshake.
956 So only messages with sequence numbers less than 10 in advance will be
957 buffered. (CVE-2009-1378)
958 [Robin Seggelmann, discovered by Daniel Mentz]
960 *) Records are buffered if they arrive with a future epoch to be
961 processed after finishing the corresponding handshake. There is
962 currently no limitation to this buffer allowing an attacker to perform
963 a DOS attack with sending records with future epochs until there is no
964 memory left. This patch adds the pqueue_size() function to detemine
965 the size of a buffer and limits the record buffer to 100 entries.
967 [Robin Seggelmann, discovered by Daniel Mentz]
969 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
970 parent structure is freed. (CVE-2009-1379)
973 *) Handle non-blocking I/O properly in SSL_shutdown() call.
974 [Darryl Miles <darryl-mailinglists@netbauds.net>]
977 [Ilya O. <vrghost@gmail.com>]
979 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
981 *) Disable renegotiation completely - this fixes a severe security
982 problem (CVE-2009-3555) at the cost of breaking all
983 renegotiation. Renegotiation can be re-enabled by setting
984 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
985 run-time. This is really not recommended unless you know what
989 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
991 *) Don't set val to NULL when freeing up structures, it is freed up by
992 underlying code. If sizeof(void *) > sizeof(long) this can result in
993 zeroing past the valid field. (CVE-2009-0789)
994 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
996 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
997 checked correctly. This would allow some invalid signed attributes to
998 appear to verify correctly. (CVE-2009-0591)
999 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1001 *) Reject UniversalString and BMPString types with invalid lengths. This
1002 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1003 a legal length. (CVE-2009-0590)
1006 *) Set S/MIME signing as the default purpose rather than setting it
1007 unconditionally. This allows applications to override it at the store
1011 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1012 to handle some structures.
1015 *) Improve efficiency of mem_gets: don't search whole buffer each time
1017 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1019 *) New -hex option for openssl rand.
1022 *) Print out UTF8String and NumericString when parsing ASN1.
1025 *) Support NumericString type for name components.
1028 *) Allow CC in the environment to override the automatically chosen
1029 compiler. Note that nothing is done to ensure flags work with the
1033 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1035 *) Properly check EVP_VerifyFinal() and similar return values
1037 [Ben Laurie, Bodo Moeller, Google Security Team]
1039 *) Enable TLS extensions by default.
1042 *) Allow the CHIL engine to be loaded, whether the application is
1043 multithreaded or not. (This does not release the developer from the
1044 obligation to set up the dynamic locking callbacks.)
1045 [Sander Temme <sander@temme.net>]
1047 *) Use correct exit code if there is an error in dgst command.
1048 [Steve Henson; problem pointed out by Roland Dirlewanger]
1050 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1051 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1054 *) Add experimental JPAKE support, including demo authentication in
1055 s_client and s_server.
1058 *) Set the comparison function in v3_addr_canonize().
1059 [Rob Austein <sra@hactrn.net>]
1061 *) Add support for XMPP STARTTLS in s_client.
1062 [Philip Paeps <philip@freebsd.org>]
1064 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1065 to ensure that even with this option, only ciphersuites in the
1066 server's preference list will be accepted. (Note that the option
1067 applies only when resuming a session, so the earlier behavior was
1068 just about the algorithm choice for symmetric cryptography.)
1071 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1073 *) Fix NULL pointer dereference if a DTLS server received
1074 ChangeCipherSpec as first record (CVE-2009-1386).
1077 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1078 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1081 *) The fix in 0.9.8c that supposedly got rid of unsafe
1082 double-checked locking was incomplete for RSA blinding,
1083 addressing just one layer of what turns out to have been
1084 doubly unsafe triple-checked locking.
1086 So now fix this for real by retiring the MONT_HELPER macro
1087 in crypto/rsa/rsa_eay.c.
1089 [Bodo Moeller; problem pointed out by Marius Schilder]
1091 *) Various precautionary measures:
1093 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1095 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1096 (NB: This would require knowledge of the secret session ticket key
1097 to exploit, in which case you'd be SOL either way.)
1099 - Change bn_nist.c so that it will properly handle input BIGNUMs
1100 outside the expected range.
1102 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1105 [Neel Mehta, Bodo Moeller]
1107 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1108 the load fails. Useful for distros.
1109 [Ben Laurie and the FreeBSD team]
1111 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1114 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1117 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1119 This work was sponsored by Logica.
1122 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1123 keystores. Support for SSL/TLS client authentication too.
1124 Not compiled unless enable-capieng specified to Configure.
1126 This work was sponsored by Logica.
1129 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1130 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1131 attribute creation routines such as certifcate requests and PKCS#12
1135 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1137 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1138 handshake which could lead to a cilent crash as found using the
1139 Codenomicon TLS test suite (CVE-2008-1672)
1140 [Steve Henson, Mark Cox]
1142 *) Fix double free in TLS server name extensions which could lead to
1143 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1146 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1148 Clear the error queue to ensure that error entries left from
1149 older function calls do not interfere with the correct operation.
1150 [Lutz Jaenicke, Erik de Castro Lopo]
1152 *) Remove root CA certificates of commercial CAs:
1154 The OpenSSL project does not recommend any specific CA and does not
1155 have any policy with respect to including or excluding any CA.
1156 Therefore it does not make any sense to ship an arbitrary selection
1157 of root CA certificates with the OpenSSL software.
1160 *) RSA OAEP patches to fix two separate invalid memory reads.
1161 The first one involves inputs when 'lzero' is greater than
1162 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1163 before the beginning of from). The second one involves inputs where
1164 the 'db' section contains nothing but zeroes (there is a one-byte
1165 invalid read after the end of 'db').
1166 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1168 *) Partial backport from 0.9.9-dev:
1170 Introduce bn_mul_mont (dedicated Montgomery multiplication
1171 procedure) as a candidate for BIGNUM assembler implementation.
1172 While 0.9.9-dev uses assembler for various architectures, only
1173 x86_64 is available by default here in the 0.9.8 branch, and
1174 32-bit x86 is available through a compile-time setting.
1176 To try the 32-bit x86 assembler implementation, use Configure
1177 option "enable-montasm" (which exists only for this backport).
1179 As "enable-montasm" for 32-bit x86 disclaims code stability
1180 anyway, in this constellation we activate additional code
1181 backported from 0.9.9-dev for further performance improvements,
1182 namely BN_from_montgomery_word. (To enable this otherwise,
1183 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1185 [Andy Polyakov (backport partially by Bodo Moeller)]
1187 *) Add TLS session ticket callback. This allows an application to set
1188 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1189 values. This is useful for key rollover for example where several key
1190 sets may exist with different names.
1193 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1194 This was broken until now in 0.9.8 releases, such that the only way
1195 a registered ENGINE could be used (assuming it initialises
1196 successfully on the host) was to explicitly set it as the default
1197 for the relevant algorithms. This is in contradiction with 0.9.7
1198 behaviour and the documentation. With this fix, when an ENGINE is
1199 registered into a given algorithm's table of implementations, the
1200 'uptodate' flag is reset so that auto-discovery will be used next
1201 time a new context for that algorithm attempts to select an
1203 [Ian Lister (tweaked by Geoff Thorpe)]
1205 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1206 implemention in the following ways:
1208 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1211 Lack of BER streaming support means one pass streaming processing is
1212 only supported if data is detached: setting the streaming flag is
1213 ignored for embedded content.
1215 CMS support is disabled by default and must be explicitly enabled
1216 with the enable-cms configuration option.
1219 *) Update the GMP engine glue to do direct copies between BIGNUM and
1220 mpz_t when openssl and GMP use the same limb size. Otherwise the
1221 existing "conversion via a text string export" trick is still used.
1222 [Paul Sheer <paulsheer@gmail.com>]
1224 *) Zlib compression BIO. This is a filter BIO which compressed and
1225 uncompresses any data passed through it.
1228 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1229 RFC3394 compatible AES key wrapping.
1232 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1233 sets string data without copying. X509_ALGOR_set0() and
1234 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1235 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1236 from an X509_ATTRIBUTE structure optionally checking it occurs only
1237 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1241 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1242 to get the expected BN_FLG_CONSTTIME behavior.
1243 [Bodo Moeller (Google)]
1247 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1248 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1249 - added some more tests to do_tests.pl
1250 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1251 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1252 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1253 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1254 - various changes to netware.pl to enable gcc-cross builds on Win32
1256 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1257 - various changes to fix missing prototype warnings
1258 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1259 - added AES, WHIRLPOOL and CPUID assembler code to build files
1260 - added missing AES assembler make rules to mk1mf.pl
1261 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1262 [Guenter Knauf <eflash@gmx.net>]
1264 *) Implement certificate status request TLS extension defined in RFC3546.
1265 A client can set the appropriate parameters and receive the encoded
1266 OCSP response via a callback. A server can query the supplied parameters
1267 and set the encoded OCSP response in the callback. Add simplified examples
1268 to s_client and s_server.
1271 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1273 *) Fix various bugs:
1274 + Binary incompatibility of ssl_ctx_st structure
1275 + DTLS interoperation with non-compliant servers
1276 + Don't call get_session_cb() without proposed session
1277 + Fix ia64 assembler code
1278 [Andy Polyakov, Steve Henson]
1280 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1282 *) DTLS Handshake overhaul. There were longstanding issues with
1283 OpenSSL DTLS implementation, which were making it impossible for
1284 RFC 4347 compliant client to communicate with OpenSSL server.
1285 Unfortunately just fixing these incompatibilities would "cut off"
1286 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1287 server keeps tolerating non RFC compliant syntax. The opposite is
1288 not true, 0.9.8f client can not communicate with earlier server.
1289 This update even addresses CVE-2007-4995.
1292 *) Changes to avoid need for function casts in OpenSSL: some compilers
1293 (gcc 4.2 and later) reject their use.
1294 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1297 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1298 RFC4507bis. The encrypted ticket format is an encrypted encoded
1299 SSL_SESSION structure, that way new session features are automatically
1302 If a client application caches session in an SSL_SESSION structure
1303 support is transparent because tickets are now stored in the encoded
1306 The SSL_CTX structure automatically generates keys for ticket
1307 protection in servers so again support should be possible
1308 with no application modification.
1310 If a client or server wishes to disable RFC4507 support then the option
1311 SSL_OP_NO_TICKET can be set.
1313 Add a TLS extension debugging callback to allow the contents of any client
1314 or server extensions to be examined.
1316 This work was sponsored by Google.
1319 *) Add initial support for TLS extensions, specifically for the server_name
1320 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1321 have new members for a host name. The SSL data structure has an
1322 additional member SSL_CTX *initial_ctx so that new sessions can be
1323 stored in that context to allow for session resumption, even after the
1324 SSL has been switched to a new SSL_CTX in reaction to a client's
1325 server_name extension.
1327 New functions (subject to change):
1329 SSL_get_servername()
1330 SSL_get_servername_type()
1333 New CTRL codes and macros (subject to change):
1335 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1336 - SSL_CTX_set_tlsext_servername_callback()
1337 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1338 - SSL_CTX_set_tlsext_servername_arg()
1339 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1341 openssl s_client has a new '-servername ...' option.
1343 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1344 '-key2 ...', '-servername_fatal' (subject to change). This allows
1345 testing the HostName extension for a specific single host name ('-cert'
1346 and '-key' remain fallbacks for handshakes without HostName
1347 negotiation). If the unrecogninzed_name alert has to be sent, this by
1348 default is a warning; it becomes fatal with the '-servername_fatal'
1351 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1353 *) Add AES and SSE2 assembly language support to VC++ build.
1356 *) Mitigate attack on final subtraction in Montgomery reduction.
1359 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1360 (which previously caused an internal error).
1363 *) Squeeze another 10% out of IGE mode when in != out.
1366 *) AES IGE mode speedup.
1367 [Dean Gaudet (Google)]
1369 *) Add the Korean symmetric 128-bit cipher SEED (see
1370 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1371 add SEED ciphersuites from RFC 4162:
1373 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1374 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1375 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1376 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1378 To minimize changes between patchlevels in the OpenSSL 0.9.8
1379 series, SEED remains excluded from compilation unless OpenSSL
1380 is configured with 'enable-seed'.
1381 [KISA, Bodo Moeller]
1383 *) Mitigate branch prediction attacks, which can be practical if a
1384 single processor is shared, allowing a spy process to extract
1385 information. For detailed background information, see
1386 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1387 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1388 and Necessary Software Countermeasures"). The core of the change
1389 are new versions BN_div_no_branch() and
1390 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1391 respectively, which are slower, but avoid the security-relevant
1392 conditional branches. These are automatically called by BN_div()
1393 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1394 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1395 remove a conditional branch.
1397 BN_FLG_CONSTTIME is the new name for the previous
1398 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1399 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1400 in the exponent causes BN_mod_exp_mont() to use the alternative
1401 implementation in BN_mod_exp_mont_consttime().) The old name
1402 remains as a deprecated alias.
1404 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1405 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1406 constant-time implementations for more than just exponentiation.
1407 Here too the old name is kept as a deprecated alias.
1409 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1410 the BN_BLINDING structure gets an independent copy of the
1411 modulus. This means that the previous "BIGNUM *m" argument to
1412 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1413 essentially becomes "const BIGNUM *m", although we can't actually
1414 change this in the header file before 0.9.9. It allows
1415 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1416 enable BN_FLG_CONSTTIME.
1418 [Matthew D Wood (Intel Corp)]
1420 *) In the SSL/TLS server implementation, be strict about session ID
1421 context matching (which matters if an application uses a single
1422 external cache for different purposes). Previously,
1423 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1424 set. This did ensure strict client verification, but meant that,
1425 with applications using a single external cache for quite
1426 different requirements, clients could circumvent ciphersuite
1427 restrictions for a given session ID context by starting a session
1428 in a different context.
1431 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1432 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1433 authentication-only ciphersuites.
1436 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1437 not complete and could lead to a possible single byte overflow
1438 (CVE-2007-5135) [Ben Laurie]
1440 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1442 *) Since AES128 and AES256 (and similarly Camellia128 and
1443 Camellia256) share a single mask bit in the logic of
1444 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1445 kludge to work properly if AES128 is available and AES256 isn't
1446 (or if Camellia128 is available and Camellia256 isn't).
1449 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1450 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1451 When a point or a seed is encoded in a BIT STRING, we need to
1452 prevent the removal of trailing zero bits to get the proper DER
1453 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1454 of a NamedBitList, for which trailing 0 bits need to be removed.)
1457 *) Have SSL/TLS server implementation tolerate "mismatched" record
1458 protocol version while receiving ClientHello even if the
1459 ClientHello is fragmented. (The server can't insist on the
1460 particular protocol version it has chosen before the ServerHello
1461 message has informed the client about his choice.)
1464 *) Add RFC 3779 support.
1465 [Rob Austein for ARIN, Ben Laurie]
1467 *) Load error codes if they are not already present instead of using a
1468 static variable. This allows them to be cleanly unloaded and reloaded.
1469 Improve header file function name parsing.
1472 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1473 or CAPABILITY handshake as required by RFCs.
1476 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1478 *) Introduce limits to prevent malicious keys being able to
1479 cause a denial of service. (CVE-2006-2940)
1480 [Steve Henson, Bodo Moeller]
1482 *) Fix ASN.1 parsing of certain invalid structures that can result
1483 in a denial of service. (CVE-2006-2937) [Steve Henson]
1485 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1486 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1488 *) Fix SSL client code which could crash if connecting to a
1489 malicious SSLv2 server. (CVE-2006-4343)
1490 [Tavis Ormandy and Will Drewry, Google Security Team]
1492 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1493 match only those. Before that, "AES256-SHA" would be interpreted
1494 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1495 the same strength classification in 0.9.7h) as we currently only
1496 have a single AES bit in the ciphersuite description bitmap.
1497 That change, however, also applied to ciphersuite strings such as
1498 "RC4-MD5" that intentionally matched multiple ciphersuites --
1499 namely, SSL 2.0 ciphersuites in addition to the more common ones
1500 from SSL 3.0/TLS 1.0.
1502 So we change the selection algorithm again: Naming an explicit
1503 ciphersuite selects this one ciphersuite, and any other similar
1504 ciphersuite (same bitmap) from *other* protocol versions.
1505 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1506 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1508 Since SSL 2.0 does not have any ciphersuites for which the
1509 128/256 bit distinction would be relevant, this works for now.
1510 The proper fix will be to use different bits for AES128 and
1511 AES256, which would have avoided the problems from the beginning;
1512 however, bits are scarce, so we can only do this in a new release
1513 (not just a patchlevel) when we can change the SSL_CIPHER
1514 definition to split the single 'unsigned long mask' bitmap into
1515 multiple values to extend the available space.
1519 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1521 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1522 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1524 *) Add AES IGE and biIGE modes.
1527 *) Change the Unix randomness entropy gathering to use poll() when
1528 possible instead of select(), since the latter has some
1529 undesirable limitations.
1530 [Darryl Miles via Richard Levitte and Bodo Moeller]
1532 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1533 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1534 cannot be implicitly activated as part of, e.g., the "AES" alias.
1535 However, please upgrade to OpenSSL 0.9.9[-dev] for
1536 non-experimental use of the ECC ciphersuites to get TLS extension
1537 support, which is required for curve and point format negotiation
1538 to avoid potential handshake problems.
1541 *) Disable rogue ciphersuites:
1543 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1544 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1545 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1547 The latter two were purportedly from
1548 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1551 Also deactivate the remaining ciphersuites from
1552 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1553 unofficial, and the ID has long expired.
1556 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1557 dual-core machines) and other potential thread-safety issues.
1560 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1561 versions), which is now available for royalty-free use
1562 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1563 Also, add Camellia TLS ciphersuites from RFC 4132.
1565 To minimize changes between patchlevels in the OpenSSL 0.9.8
1566 series, Camellia remains excluded from compilation unless OpenSSL
1567 is configured with 'enable-camellia'.
1570 *) Disable the padding bug check when compression is in use. The padding
1571 bug check assumes the first packet is of even length, this is not
1572 necessarily true if compresssion is enabled and can result in false
1573 positives causing handshake failure. The actual bug test is ancient
1574 code so it is hoped that implementations will either have fixed it by
1575 now or any which still have the bug do not support compression.
1578 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1580 *) When applying a cipher rule check to see if string match is an explicit
1581 cipher suite and only match that one cipher suite if it is.
1584 *) Link in manifests for VC++ if needed.
1585 [Austin Ziegler <halostatue@gmail.com>]
1587 *) Update support for ECC-based TLS ciphersuites according to
1588 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1589 TLS extensions, which are supported starting with the 0.9.9
1590 branch, not in the OpenSSL 0.9.8 branch).
1593 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1594 opaque EVP_CIPHER_CTX handling.
1597 *) Fixes and enhancements to zlib compression code. We now only use
1598 "zlib1.dll" and use the default __cdecl calling convention on Win32
1599 to conform with the standards mentioned here:
1600 http://www.zlib.net/DLL_FAQ.txt
1601 Static zlib linking now works on Windows and the new --with-zlib-include
1602 --with-zlib-lib options to Configure can be used to supply the location
1603 of the headers and library. Gracefully handle case where zlib library
1607 *) Several fixes and enhancements to the OID generation code. The old code
1608 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1609 handle numbers larger than ULONG_MAX, truncated printing and had a
1610 non standard OBJ_obj2txt() behaviour.
1613 *) Add support for building of engines under engine/ as shared libraries
1614 under VC++ build system.
1617 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1618 Hopefully, we will not see any false combination of paths any more.
1621 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1623 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1624 (part of SSL_OP_ALL). This option used to disable the
1625 countermeasure against man-in-the-middle protocol-version
1626 rollback in the SSL 2.0 server implementation, which is a bad
1627 idea. (CVE-2005-2969)
1629 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1630 for Information Security, National Institute of Advanced Industrial
1631 Science and Technology [AIST], Japan)]
1633 *) Add two function to clear and return the verify parameter flags.
1636 *) Keep cipherlists sorted in the source instead of sorting them at
1637 runtime, thus removing the need for a lock.
1640 *) Avoid some small subgroup attacks in Diffie-Hellman.
1641 [Nick Mathewson and Ben Laurie]
1643 *) Add functions for well-known primes.
1646 *) Extended Windows CE support.
1647 [Satoshi Nakamura and Andy Polyakov]
1649 *) Initialize SSL_METHOD structures at compile time instead of during
1650 runtime, thus removing the need for a lock.
1653 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1654 attempting to decrypt each encrypted key in turn. Add support to
1658 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1660 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1663 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1666 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1667 key into the same file any more.
1670 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1673 *) Add -utf8 command line and config file option to 'ca'.
1674 [Stefan <stf@udoma.org]
1676 *) Removed the macro des_crypt(), as it seems to conflict with some
1677 libraries. Use DES_crypt().
1680 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1681 involves renaming the source and generated shared-libs for
1682 both. The engines will accept the corrected or legacy ids
1683 ('ncipher' and '4758_cca' respectively) when binding. NB,
1684 this only applies when building 'shared'.
1685 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1687 *) Add attribute functions to EVP_PKEY structure. Modify
1688 PKCS12_create() to recognize a CSP name attribute and
1689 use it. Make -CSP option work again in pkcs12 utility.
1692 *) Add new functionality to the bn blinding code:
1693 - automatic re-creation of the BN_BLINDING parameters after
1694 a fixed number of uses (currently 32)
1695 - add new function for parameter creation
1696 - introduce flags to control the update behaviour of the
1697 BN_BLINDING parameters
1698 - hide BN_BLINDING structure
1699 Add a second BN_BLINDING slot to the RSA structure to improve
1700 performance when a single RSA object is shared among several
1704 *) Add support for DTLS.
1705 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1707 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1708 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1711 *) Remove buggy and incompletet DH cert support from
1712 ssl/ssl_rsa.c and ssl/s3_both.c
1715 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1716 the apps/openssl applications.
1719 *) Compile clean with "-Wall -Wmissing-prototypes
1720 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1721 DEBUG_SAFESTACK must also be set.
1724 *) Change ./Configure so that certain algorithms can be disabled by default.
1725 The new counterpiece to "no-xxx" is "enable-xxx".
1727 The patented RC5 and MDC2 algorithms will now be disabled unless
1728 "enable-rc5" and "enable-mdc2", respectively, are specified.
1730 (IDEA remains enabled despite being patented. This is because IDEA
1731 is frequently required for interoperability, and there is no license
1732 fee for non-commercial use. As before, "no-idea" can be used to
1733 avoid this algorithm.)
1737 *) Add processing of proxy certificates (see RFC 3820). This work was
1738 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1739 EGEE (Enabling Grids for E-science in Europe).
1742 *) RC4 performance overhaul on modern architectures/implementations, such
1743 as Intel P4, IA-64 and AMD64.
1746 *) New utility extract-section.pl. This can be used specify an alternative
1747 section number in a pod file instead of having to treat each file as
1748 a separate case in Makefile. This can be done by adding two lines to the
1751 =for comment openssl_section:XXX
1753 The blank line is mandatory.
1757 *) New arguments -certform, -keyform and -pass for s_client and s_server
1758 to allow alternative format key and certificate files and passphrase
1762 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1763 update associated structures and add various utility functions.
1765 Add new policy related verify parameters, include policy checking in
1766 standard verify code. Enhance 'smime' application with extra parameters
1767 to support policy checking and print out.
1770 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1771 Nehemiah processors. These extensions support AES encryption in hardware
1772 as well as RNG (though RNG support is currently disabled).
1773 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1775 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1778 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1779 [Andy Polyakov and a number of other people]
1781 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1782 implementation contributed by IBM.
1783 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1785 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1786 exponent rather than 'unsigned long'. There is a corresponding change to
1787 the new 'rsa_keygen' element of the RSA_METHOD structure.
1788 [Jelte Jansen, Geoff Thorpe]
1790 *) Functionality for creating the initial serial number file is now
1791 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1793 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1794 number file to 1, which is bound to cause problems. To avoid
1795 the problems while respecting compatibility between different 0.9.7
1796 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1797 CA.pl for serial number initialization. With the new release 0.9.8,
1798 we can fix the problem directly in the 'ca' utility.)
1801 *) Reduced header interdepencies by declaring more opaque objects in
1802 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1803 give fewer recursive includes, which could break lazy source code - so
1804 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1805 developers should define this symbol when building and using openssl to
1806 ensure they track the recommended behaviour, interfaces, [etc], but
1807 backwards-compatible behaviour prevails when this isn't defined.
1810 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1813 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1814 This will generate a random key of the appropriate length based on the
1815 cipher context. The EVP_CIPHER can provide its own random key generation
1816 routine to support keys of a specific form. This is used in the des and
1817 3des routines to generate a key of the correct parity. Update S/MIME
1818 code to use new functions and hence generate correct parity DES keys.
1819 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1820 valid (weak or incorrect parity).
1823 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1824 as looking them up. This is useful when the verified structure may contain
1825 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1826 present unless the new PKCS7_NO_CRL flag is asserted.
1829 *) Extend ASN1 oid configuration module. It now additionally accepts the
1832 shortName = some long name, 1.2.3.4
1835 *) Reimplemented the BN_CTX implementation. There is now no more static
1836 limitation on the number of variables it can handle nor the depth of the
1837 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1838 information can now expand as required, and rather than having a single
1839 static array of bignums, BN_CTX now uses a linked-list of such arrays
1840 allowing it to expand on demand whilst maintaining the usefulness of
1841 BN_CTX's "bundling".
1844 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1845 to allow all RSA operations to function using a single BN_CTX.
1848 *) Preliminary support for certificate policy evaluation and checking. This
1849 is initially intended to pass the tests outlined in "Conformance Testing
1850 of Relying Party Client Certificate Path Processing Logic" v1.07.
1853 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1854 remained unused and not that useful. A variety of other little bignum
1855 tweaks and fixes have also been made continuing on from the audit (see
1859 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1860 associated ASN1, EVP and SSL functions and old ASN1 macros.
1863 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1864 and this should never fail. So the return value from the use of
1865 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1866 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1869 *) BN_CTX_get() should return zero-valued bignums, providing the same
1870 initialised value as BN_new().
1871 [Geoff Thorpe, suggested by Ulf Möller]
1873 *) Support for inhibitAnyPolicy certificate extension.
1876 *) An audit of the BIGNUM code is underway, for which debugging code is
1877 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1878 is considered valid when processing BIGNUMs, and causes execution to
1879 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1880 further steps are taken to deliberately pollute unused data in BIGNUM
1881 structures to try and expose faulty code further on. For now, openssl will
1882 (in its default mode of operation) continue to tolerate the inconsistent
1883 forms that it has tolerated in the past, but authors and packagers should
1884 consider trying openssl and their own applications when compiled with
1885 these debugging symbols defined. It will help highlight potential bugs in
1886 their own code, and will improve the test coverage for OpenSSL itself. At
1887 some point, these tighter rules will become openssl's default to improve
1888 maintainability, though the assert()s and other overheads will remain only
1889 in debugging configurations. See bn.h for more details.
1890 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1892 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1893 that can only be obtained through BN_CTX_new() (which implicitly
1894 initialises it). The presence of this function only made it possible
1895 to overwrite an existing structure (and cause memory leaks).
1898 *) Because of the callback-based approach for implementing LHASH as a
1899 template type, lh_insert() adds opaque objects to hash-tables and
1900 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1901 to clean up those corresponding objects before destroying the hash table
1902 (and losing the object pointers). So some over-zealous constifications in
1903 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1904 objects as "const" and the lh_doall[_arg] callback wrappers are not
1905 prototyped to have "const" restrictions on the object pointers they are
1906 given (and so aren't required to cast them away any more).
1909 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1910 (speed) prefers to use its own implementation. The two implementations
1911 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1912 its object type properly exposed (MS_TM) instead of casting to/from "char
1913 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1914 aren't necessarily the greatest nomenclatures - but this is what was used
1915 internally to the implementation so I've used that for now.
1918 *) Ensure that deprecated functions do not get compiled when
1919 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1920 the self-tests were still using deprecated key-generation functions so
1921 these have been updated also.
1924 *) Reorganise PKCS#7 code to separate the digest location functionality
1925 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1926 New function PKCS7_set_digest() to set the digest type for PKCS#7
1927 digestedData type. Add additional code to correctly generate the
1928 digestedData type and add support for this type in PKCS7 initialization
1932 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1933 structure of type "other".
1936 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1937 sure the loop does correctly stop and breaking ("division by zero")
1938 modulus operations are not performed. The (pre-generated) prime
1939 table crypto/bn/bn_prime.h was already correct, but it could not be
1940 re-generated on some platforms because of the "division by zero"
1941 situation in the script.
1942 [Ralf S. Engelschall]
1944 *) Update support for ECC-based TLS ciphersuites according to
1945 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1946 SHA-1 now is only used for "small" curves (where the
1947 representation of a field element takes up to 24 bytes); for
1948 larger curves, the field element resulting from ECDH is directly
1949 used as premaster secret.
1950 [Douglas Stebila (Sun Microsystems Laboratories)]
1952 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1953 curve secp160r1 to the tests.
1954 [Douglas Stebila (Sun Microsystems Laboratories)]
1956 *) Add the possibility to load symbols globally with DSO.
1957 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1959 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1960 control of the error stack.
1963 *) Add support for STORE in ENGINE.
1966 *) Add the STORE type. The intention is to provide a common interface
1967 to certificate and key stores, be they simple file-based stores, or
1968 HSM-type store, or LDAP stores, or...
1969 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1972 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1973 pass a list of arguments to any function as well as provide a way
1974 for a function to pass data back to the caller.
1977 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1978 works like BUF_strdup() but can be used to duplicate a portion of
1979 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1983 *) Add the function sk_find_ex() which works like sk_find(), but will
1984 return an index to an element even if an exact match couldn't be
1985 found. The index is guaranteed to point at the element where the
1986 searched-for key would be inserted to preserve sorting order.
1989 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1990 takes an extra flags argument for optional functionality. Currently,
1991 the following flags are defined:
1993 OBJ_BSEARCH_VALUE_ON_NOMATCH
1994 This one gets OBJ_bsearch_ex() to return a pointer to the first
1995 element where the comparing function returns a negative or zero
1998 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1999 This one gets OBJ_bsearch_ex() to return a pointer to the first
2000 element where the comparing function returns zero. This is useful
2001 if there are more than one element where the comparing function
2005 *) Make it possible to create self-signed certificates with 'openssl ca'
2006 in such a way that the self-signed certificate becomes part of the
2007 CA database and uses the same mechanisms for serial number generation
2008 as all other certificate signing. The new flag '-selfsign' enables
2009 this functionality. Adapt CA.sh and CA.pl.in.
2012 *) Add functionality to check the public key of a certificate request
2013 against a given private. This is useful to check that a certificate
2014 request can be signed by that key (self-signing).
2017 *) Make it possible to have multiple active certificates with the same
2018 subject in the CA index file. This is done only if the keyword
2019 'unique_subject' is set to 'no' in the main CA section (default
2020 if 'CA_default') of the configuration file. The value is saved
2021 with the database itself in a separate index attribute file,
2022 named like the index file with '.attr' appended to the name.
2025 *) Generate muti valued AVAs using '+' notation in config files for
2029 *) Support for nameConstraints certificate extension.
2032 *) Support for policyConstraints certificate extension.
2035 *) Support for policyMappings certificate extension.
2038 *) Make sure the default DSA_METHOD implementation only uses its
2039 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2040 and change its own handlers to be NULL so as to remove unnecessary
2041 indirection. This lets alternative implementations fallback to the
2042 default implementation more easily.
2045 *) Support for directoryName in GeneralName related extensions
2049 *) Make it possible to link applications using Makefile.shared.
2050 Make that possible even when linking against static libraries!
2053 *) Support for single pass processing for S/MIME signing. This now
2054 means that S/MIME signing can be done from a pipe, in addition
2055 cleartext signing (multipart/signed type) is effectively streaming
2056 and the signed data does not need to be all held in memory.
2058 This is done with a new flag PKCS7_STREAM. When this flag is set
2059 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2060 is done after the data is output (and digests calculated) in
2061 SMIME_write_PKCS7().
2064 *) Add full support for -rpath/-R, both in shared libraries and
2065 applications, at least on the platforms where it's known how
2069 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2070 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2071 will now compute a table of multiples of the generator that
2072 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2073 faster (notably in the case of a single point multiplication,
2074 scalar * generator).
2075 [Nils Larsch, Bodo Moeller]
2077 *) IPv6 support for certificate extensions. The various extensions
2078 which use the IP:a.b.c.d can now take IPv6 addresses using the
2079 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2083 *) Added an ENGINE that implements RSA by performing private key
2084 exponentiations with the GMP library. The conversions to and from
2085 GMP's mpz_t format aren't optimised nor are any montgomery forms
2086 cached, and on x86 it appears OpenSSL's own performance has caught up.
2087 However there are likely to be other architectures where GMP could
2088 provide a boost. This ENGINE is not built in by default, but it can be
2089 specified at Configure time and should be accompanied by the necessary
2090 linker additions, eg;
2091 ./config -DOPENSSL_USE_GMP -lgmp
2094 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2095 testing availability of engines with "-t" - the old behaviour is
2096 produced by increasing the feature's verbosity with "-tt".
2099 *) ECDSA routines: under certain error conditions uninitialized BN objects
2100 could be freed. Solution: make sure initialization is performed early
2101 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2105 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2106 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2107 software implementations. For DSA and DH, parameter generation can
2108 also be overriden by providing the appropriate method callbacks.
2111 *) Change the "progress" mechanism used in key-generation and
2112 primality testing to functions that take a new BN_GENCB pointer in
2113 place of callback/argument pairs. The new API functions have "_ex"
2114 postfixes and the older functions are reimplemented as wrappers for
2115 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2116 declarations of the old functions to help (graceful) attempts to
2117 migrate to the new functions. Also, the new key-generation API
2118 functions operate on a caller-supplied key-structure and return
2119 success/failure rather than returning a key or NULL - this is to
2120 help make "keygen" another member function of RSA_METHOD etc.
2122 Example for using the new callback interface:
2124 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2128 BN_GENCB_set(&my_cb, my_callback, my_arg);
2130 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2131 /* For the meaning of a, b in calls to my_callback(), see the
2132 * documentation of the function that calls the callback.
2133 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2134 * my_callback should return 1 if it wants BN_is_prime_ex()
2135 * to continue, or 0 to stop.
2140 *) Change the ZLIB compression method to be stateful, and make it
2141 available to TLS with the number defined in
2142 draft-ietf-tls-compression-04.txt.
2145 *) Add the ASN.1 structures and functions for CertificatePair, which
2146 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2148 CertificatePair ::= SEQUENCE {
2149 forward [0] Certificate OPTIONAL,
2150 reverse [1] Certificate OPTIONAL,
2151 -- at least one of the pair shall be present -- }
2153 Also implement the PEM functions to read and write certificate
2154 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2156 This needed to be defined, mostly for the sake of the LDAP
2157 attribute crossCertificatePair, but may prove useful elsewhere as
2161 *) Make it possible to inhibit symlinking of shared libraries in
2162 Makefile.shared, for Cygwin's sake.
2165 *) Extend the BIGNUM API by creating a function
2166 void BN_set_negative(BIGNUM *a, int neg);
2167 and a macro that behave like
2168 int BN_is_negative(const BIGNUM *a);
2170 to avoid the need to access 'a->neg' directly in applications.
2173 *) Implement fast modular reduction for pseudo-Mersenne primes
2174 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2175 EC_GROUP_new_curve_GFp() will now automatically use this
2177 [Nils Larsch <nla@trustcenter.de>]
2179 *) Add new lock type (CRYPTO_LOCK_BN).
2182 *) Change the ENGINE framework to automatically load engines
2183 dynamically from specific directories unless they could be
2184 found to already be built in or loaded. Move all the
2185 current engines except for the cryptodev one to a new
2187 The engines in engines/ are built as shared libraries if
2188 the "shared" options was given to ./Configure or ./config.
2189 Otherwise, they are inserted in libcrypto.a.
2190 /usr/local/ssl/engines is the default directory for dynamic
2191 engines, but that can be overriden at configure time through
2192 the usual use of --prefix and/or --openssldir, and at run
2193 time with the environment variable OPENSSL_ENGINES.
2194 [Geoff Thorpe and Richard Levitte]
2196 *) Add Makefile.shared, a helper makefile to build shared
2197 libraries. Addapt Makefile.org.
2200 *) Add version info to Win32 DLLs.
2201 [Peter 'Luna' Runestig" <peter@runestig.com>]
2203 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2204 can be added using this API to created arbitrary PKCS#12
2205 files while avoiding the low level API.
2207 New options to PKCS12_create(), key or cert can be NULL and
2208 will then be omitted from the output file. The encryption
2209 algorithm NIDs can be set to -1 for no encryption, the mac
2210 iteration count can be set to 0 to omit the mac.
2212 Enhance pkcs12 utility by making the -nokeys and -nocerts
2213 options work when creating a PKCS#12 file. New option -nomac
2214 to omit the mac, NONE can be set for an encryption algorithm.
2215 New code is modified to use the enhanced PKCS12_create()
2216 instead of the low level API.
2219 *) Extend ASN1 encoder to support indefinite length constructed
2220 encoding. This can output sequences tags and octet strings in
2221 this form. Modify pk7_asn1.c to support indefinite length
2222 encoding. This is experimental and needs additional code to
2223 be useful, such as an ASN1 bio and some enhanced streaming
2226 Extend template encode functionality so that tagging is passed
2227 down to the template encoder.
2230 *) Let 'openssl req' fail if an argument to '-newkey' is not
2231 recognized instead of using RSA as a default.
2234 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2235 As these are not official, they are not included in "ALL";
2236 the "ECCdraft" ciphersuite group alias can be used to select them.
2237 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2239 *) Add ECDH engine support.
2240 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2242 *) Add ECDH in new directory crypto/ecdh/.
2243 [Douglas Stebila (Sun Microsystems Laboratories)]
2245 *) Let BN_rand_range() abort with an error after 100 iterations
2246 without success (which indicates a broken PRNG).
2249 *) Change BN_mod_sqrt() so that it verifies that the input value
2250 is really the square of the return value. (Previously,
2251 BN_mod_sqrt would show GIGO behaviour.)
2254 *) Add named elliptic curves over binary fields from X9.62, SECG,
2255 and WAP/WTLS; add OIDs that were still missing.
2257 [Sheueling Chang Shantz and Douglas Stebila
2258 (Sun Microsystems Laboratories)]
2260 *) Extend the EC library for elliptic curves over binary fields
2261 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2264 EC_GF2m_simple_method
2268 EC_GROUP_new_curve_GF2m
2269 EC_GROUP_set_curve_GF2m
2270 EC_GROUP_get_curve_GF2m
2271 EC_POINT_set_affine_coordinates_GF2m
2272 EC_POINT_get_affine_coordinates_GF2m
2273 EC_POINT_set_compressed_coordinates_GF2m
2275 Point compression for binary fields is disabled by default for
2276 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2279 As binary polynomials are represented as BIGNUMs, various members
2280 of the EC_GROUP and EC_POINT data structures can be shared
2281 between the implementations for prime fields and binary fields;
2282 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2283 are essentially identical to their ..._GFp counterparts.
2284 (For simplicity, the '..._GFp' prefix has been dropped from
2285 various internal method names.)
2287 An internal 'field_div' method (similar to 'field_mul' and
2288 'field_sqr') has been added; this is used only for binary fields.
2290 [Sheueling Chang Shantz and Douglas Stebila
2291 (Sun Microsystems Laboratories)]
2293 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2294 through methods ('mul', 'precompute_mult').
2296 The generic implementations (now internally called 'ec_wNAF_mul'
2297 and 'ec_wNAF_precomputed_mult') remain the default if these
2298 methods are undefined.
2300 [Sheueling Chang Shantz and Douglas Stebila
2301 (Sun Microsystems Laboratories)]
2303 *) New function EC_GROUP_get_degree, which is defined through
2304 EC_METHOD. For curves over prime fields, this returns the bit
2305 length of the modulus.
2307 [Sheueling Chang Shantz and Douglas Stebila
2308 (Sun Microsystems Laboratories)]
2310 *) New functions EC_GROUP_dup, EC_POINT_dup.
2311 (These simply call ..._new and ..._copy).
2313 [Sheueling Chang Shantz and Douglas Stebila
2314 (Sun Microsystems Laboratories)]
2316 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2317 Polynomials are represented as BIGNUMs (where the sign bit is not
2318 used) in the following functions [macros]:
2321 BN_GF2m_sub [= BN_GF2m_add]
2322 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2323 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2324 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2326 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2327 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2328 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2329 BN_GF2m_cmp [= BN_ucmp]
2331 (Note that only the 'mod' functions are actually for fields GF(2^m).
2332 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2334 For some functions, an the irreducible polynomial defining a
2335 field can be given as an 'unsigned int[]' with strictly
2336 decreasing elements giving the indices of those bits that are set;
2337 i.e., p[] represents the polynomial
2338 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2340 p[0] > p[1] > ... > p[k] = 0.
2341 This applies to the following functions:
2346 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2347 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2349 BN_GF2m_mod_sqrt_arr
2350 BN_GF2m_mod_solve_quad_arr
2354 Conversion can be performed by the following functions:
2359 bntest.c has additional tests for binary polynomial arithmetic.
2361 Two implementations for BN_GF2m_mod_div() are available.
2362 The default algorithm simply uses BN_GF2m_mod_inv() and
2363 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2364 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2365 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2367 [Sheueling Chang Shantz and Douglas Stebila
2368 (Sun Microsystems Laboratories)]
2370 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2371 functionality is disabled at compile-time.
2372 [Douglas Stebila <douglas.stebila@sun.com>]
2374 *) Change default behaviour of 'openssl asn1parse' so that more
2375 information is visible when viewing, e.g., a certificate:
2377 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2378 mode the content of non-printable OCTET STRINGs is output in a
2379 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2380 avoid the appearance of a printable string.
2381 [Nils Larsch <nla@trustcenter.de>]
2383 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2385 EC_GROUP_set_asn1_flag()
2386 EC_GROUP_get_asn1_flag()
2387 EC_GROUP_set_point_conversion_form()
2388 EC_GROUP_get_point_conversion_form()
2389 These control ASN1 encoding details:
2390 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2391 has been set to OPENSSL_EC_NAMED_CURVE.
2392 - Points are encoded in uncompressed form by default; options for
2393 asn1_for are as for point2oct, namely
2394 POINT_CONVERSION_COMPRESSED
2395 POINT_CONVERSION_UNCOMPRESSED
2396 POINT_CONVERSION_HYBRID
2398 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2401 EC_GROUP_get0_seed()
2402 EC_GROUP_get_seed_len()
2403 This is used only for ASN1 purposes (so far).
2404 [Nils Larsch <nla@trustcenter.de>]
2406 *) Add 'field_type' member to EC_METHOD, which holds the NID
2407 of the appropriate field type OID. The new function
2408 EC_METHOD_get_field_type() returns this value.
2409 [Nils Larsch <nla@trustcenter.de>]
2414 EC_POINT_point2hex()
2415 EC_POINT_hex2point()
2416 providing useful interfaces to EC_POINT_point2oct() and
2417 EC_POINT_oct2point().
2418 [Nils Larsch <nla@trustcenter.de>]
2420 *) Change internals of the EC library so that the functions
2421 EC_GROUP_set_generator()
2422 EC_GROUP_get_generator()
2423 EC_GROUP_get_order()
2424 EC_GROUP_get_cofactor()
2425 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2426 to methods, which would lead to unnecessary code duplication when
2427 adding different types of curves.
2428 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2430 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2431 arithmetic, and such that modified wNAFs are generated
2432 (which avoid length expansion in many cases).
2435 *) Add a function EC_GROUP_check_discriminant() (defined via
2436 EC_METHOD) that verifies that the curve discriminant is non-zero.
2438 Add a function EC_GROUP_check() that makes some sanity tests
2439 on a EC_GROUP, its generator and order. This includes
2440 EC_GROUP_check_discriminant().
2441 [Nils Larsch <nla@trustcenter.de>]
2443 *) Add ECDSA in new directory crypto/ecdsa/.
2445 Add applications 'openssl ecparam' and 'openssl ecdsa'
2446 (these are based on 'openssl dsaparam' and 'openssl dsa').
2448 ECDSA support is also included in various other files across the
2449 library. Most notably,
2450 - 'openssl req' now has a '-newkey ecdsa:file' option;
2451 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2452 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2453 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2454 them suitable for ECDSA where domain parameters must be
2455 extracted before the specific public key;
2456 - ECDSA engine support has been added.
2457 [Nils Larsch <nla@trustcenter.de>]
2459 *) Include some named elliptic curves, and add OIDs from X9.62,
2460 SECG, and WAP/WTLS. Each curve can be obtained from the new
2462 EC_GROUP_new_by_curve_name(),
2463 and the list of available named curves can be obtained with
2464 EC_get_builtin_curves().
2465 Also add a 'curve_name' member to EC_GROUP objects, which can be
2467 EC_GROUP_set_curve_name()
2468 EC_GROUP_get_curve_name()
2469 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2471 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2472 was actually never needed) and in BN_mul(). The removal in BN_mul()
2473 required a small change in bn_mul_part_recursive() and the addition
2474 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2475 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2476 bn_sub_words() and bn_add_words() except they take arrays with
2480 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2482 *) Cleanse PEM buffers before freeing them since they may contain
2484 [Benjamin Bennett <ben@psc.edu>]
2486 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2487 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2488 authentication-only ciphersuites.
2491 *) Since AES128 and AES256 share a single mask bit in the logic of
2492 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2493 kludge to work properly if AES128 is available and AES256 isn't.
2496 *) Expand security boundary to match 1.1.1 module.
2499 *) Remove redundant features: hash file source, editing of test vectors
2500 modify fipsld to use external fips_premain.c signature.
2503 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2504 run algorithm test programs.
2507 *) Make algorithm test programs more tolerant of whitespace.
2510 *) Have SSL/TLS server implementation tolerate "mismatched" record
2511 protocol version while receiving ClientHello even if the
2512 ClientHello is fragmented. (The server can't insist on the
2513 particular protocol version it has chosen before the ServerHello
2514 message has informed the client about his choice.)
2517 *) Load error codes if they are not already present instead of using a
2518 static variable. This allows them to be cleanly unloaded and reloaded.
2521 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2523 *) Introduce limits to prevent malicious keys being able to
2524 cause a denial of service. (CVE-2006-2940)
2525 [Steve Henson, Bodo Moeller]
2527 *) Fix ASN.1 parsing of certain invalid structures that can result
2528 in a denial of service. (CVE-2006-2937) [Steve Henson]
2530 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2531 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2533 *) Fix SSL client code which could crash if connecting to a
2534 malicious SSLv2 server. (CVE-2006-4343)
2535 [Tavis Ormandy and Will Drewry, Google Security Team]
2537 *) Change ciphersuite string processing so that an explicit
2538 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2539 will no longer include "AES128-SHA"), and any other similar
2540 ciphersuite (same bitmap) from *other* protocol versions (so that
2541 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2542 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2543 changes from 0.9.8b and 0.9.8d.
2546 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2548 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2549 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2551 *) Change the Unix randomness entropy gathering to use poll() when
2552 possible instead of select(), since the latter has some
2553 undesirable limitations.
2554 [Darryl Miles via Richard Levitte and Bodo Moeller]
2556 *) Disable rogue ciphersuites:
2558 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2559 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2560 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2562 The latter two were purportedly from
2563 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2566 Also deactive the remaining ciphersuites from
2567 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2568 unofficial, and the ID has long expired.
2571 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2572 dual-core machines) and other potential thread-safety issues.
2575 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2577 *) Adapt fipsld and the build system to link against the validated FIPS
2578 module in FIPS mode.
2581 *) Fixes for VC++ 2005 build under Windows.
2584 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2585 from a Windows bash shell such as MSYS. It is autodetected from the
2586 "config" script when run from a VC++ environment. Modify standard VC++
2587 build to use fipscanister.o from the GNU make build.
2590 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2592 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2593 The value now differs depending on if you build for FIPS or not.
2594 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2595 safely run with a non-FIPSed libcrypto, as it may crash because of
2596 the difference induced by this change.
2599 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2601 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2602 (part of SSL_OP_ALL). This option used to disable the
2603 countermeasure against man-in-the-middle protocol-version
2604 rollback in the SSL 2.0 server implementation, which is a bad
2605 idea. (CVE-2005-2969)
2607 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2608 for Information Security, National Institute of Advanced Industrial
2609 Science and Technology [AIST], Japan)]
2611 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2612 mainly for FIPS compliance and not fully integrated at this stage.
2615 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2616 the exponentiation using a fixed-length exponent. (Otherwise,
2617 the information leaked through timing could expose the secret key
2618 after many signatures; cf. Bleichenbacher's attack on DSA with
2622 *) Make a new fixed-window mod_exp implementation the default for
2623 RSA, DSA, and DH private-key operations so that the sequence of
2624 squares and multiplies and the memory access pattern are
2625 independent of the particular secret key. This will mitigate
2626 cache-timing and potential related attacks.
2628 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2629 and this is automatically used by BN_mod_exp_mont() if the new flag
2630 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2631 will use this BN flag for private exponents unless the flag
2632 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2633 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2635 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2637 *) Change the client implementation for SSLv23_method() and
2638 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2639 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2640 (Previously, the SSL 2.0 backwards compatible Client Hello
2641 message format would be used even with SSL_OP_NO_SSLv2.)
2644 *) Add support for smime-type MIME parameter in S/MIME messages which some
2648 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2649 a threadsafe manner. Modify rsa code to use new function and add calls
2650 to dsa and dh code (which had race conditions before).
2653 *) Include the fixed error library code in the C error file definitions
2654 instead of fixing them up at runtime. This keeps the error code
2655 structures constant.
2658 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2660 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2663 *) Fixes for newer kerberos headers. NB: the casts are needed because
2664 the 'length' field is signed on one version and unsigned on another
2665 with no (?) obvious way to tell the difference, without these VC++
2666 complains. Also the "definition" of FAR (blank) is no longer included
2667 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2668 some needed definitions.
2671 *) Undo Cygwin change.
2674 *) Added support for proxy certificates according to RFC 3820.
2675 Because they may be a security thread to unaware applications,
2676 they must be explicitely allowed in run-time. See
2677 docs/HOWTO/proxy_certificates.txt for further information.
2680 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2682 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2683 server and client random values. Previously
2684 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2685 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2687 This change has negligible security impact because:
2689 1. Server and client random values still have 24 bytes of pseudo random
2692 2. Server and client random values are sent in the clear in the initial
2695 3. The master secret is derived using the premaster secret (48 bytes in
2696 size for static RSA ciphersuites) as well as client server and random
2699 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2702 [Stephen Henson, reported by UK NISCC]
2704 *) Use Windows randomness collection on Cygwin.
2707 *) Fix hang in EGD/PRNGD query when communication socket is closed
2708 prematurely by EGD/PRNGD.
2709 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2711 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2714 *) Back-port of selected performance improvements from development
2715 branch, as well as improved support for PowerPC platforms.
2718 *) Add lots of checks for memory allocation failure, error codes to indicate
2719 failure and freeing up memory if a failure occurs.
2720 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2722 *) Add new -passin argument to dgst.
2725 *) Perform some character comparisons of different types in X509_NAME_cmp:
2726 this is needed for some certificates that reencode DNs into UTF8Strings
2727 (in violation of RFC3280) and can't or wont issue name rollover
2731 *) Make an explicit check during certificate validation to see that
2732 the CA setting in each certificate on the chain is correct. As a
2733 side effect always do the following basic checks on extensions,
2734 not just when there's an associated purpose to the check:
2736 - if there is an unhandled critical extension (unless the user
2737 has chosen to ignore this fault)
2738 - if the path length has been exceeded (if one is set at all)
2739 - that certain extensions fit the associated purpose (if one has
2743 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2745 *) Avoid a race condition when CRLs are checked in a multi threaded
2746 environment. This would happen due to the reordering of the revoked
2747 entries during signature checking and serial number lookup. Now the
2748 encoding is cached and the serial number sort performed under a lock.
2749 Add new STACK function sk_is_sorted().
2752 *) Add Delta CRL to the extension code.
2755 *) Various fixes to s3_pkt.c so alerts are sent properly.
2756 [David Holmes <d.holmes@f5.com>]
2758 *) Reduce the chances of duplicate issuer name and serial numbers (in
2759 violation of RFC3280) using the OpenSSL certificate creation utilities.
2760 This is done by creating a random 64 bit value for the initial serial
2761 number when a serial number file is created or when a self signed
2762 certificate is created using 'openssl req -x509'. The initial serial
2763 number file is created using 'openssl x509 -next_serial' in CA.pl
2764 rather than being initialized to 1.
2767 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2769 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2770 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2771 [Joe Orton, Steve Henson]
2773 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2775 [Joe Orton, Steve Henson]
2777 *) Make it possible to have multiple active certificates with the same
2778 subject in the CA index file. This is done only if the keyword
2779 'unique_subject' is set to 'no' in the main CA section (default
2780 if 'CA_default') of the configuration file. The value is saved
2781 with the database itself in a separate index attribute file,
2782 named like the index file with '.attr' appended to the name.
2785 *) X509 verify fixes. Disable broken certificate workarounds when
2786 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2787 keyUsage extension present. Don't accept CRLs with unhandled critical
2788 extensions: since verify currently doesn't process CRL extensions this
2789 rejects a CRL with *any* critical extensions. Add new verify error codes
2793 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2794 A clarification of RFC2560 will require the use of OCTET STRINGs and
2795 some implementations cannot handle the current raw format. Since OpenSSL
2796 copies and compares OCSP nonces as opaque blobs without any attempt at
2797 parsing them this should not create any compatibility issues.
2800 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2801 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2802 this HMAC (and other) operations are several times slower than OpenSSL
2806 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2807 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2809 *) Use the correct content when signing type "other".
2812 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2814 *) Fix various bugs revealed by running the NISCC test suite:
2816 Stop out of bounds reads in the ASN1 code when presented with
2817 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2819 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2821 If verify callback ignores invalid public key errors don't try to check
2822 certificate signature with the NULL public key.
2826 *) New -ignore_err option in ocsp application to stop the server
2827 exiting on the first error in a request.
2830 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2831 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2835 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2836 extra data after the compression methods not only for TLS 1.0
2837 but also for SSL 3.0 (as required by the specification).
2838 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2840 *) Change X509_certificate_type() to mark the key as exported/exportable
2841 when it's 512 *bits* long, not 512 bytes.
2844 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2845 blocks during encryption.
2848 *) Various fixes to base64 BIO and non blocking I/O. On write
2849 flushes were not handled properly if the BIO retried. On read
2850 data was not being buffered properly and had various logic bugs.
2851 This also affects blocking I/O when the data being decoded is a
2855 *) Various S/MIME bugfixes and compatibility changes:
2856 output correct application/pkcs7 MIME type if
2857 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2858 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2859 of files as .eml work). Correctly handle very long lines in MIME
2863 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2865 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2866 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2867 a protocol version number mismatch like a decryption error
2868 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2871 *) Turn on RSA blinding by default in the default implementation
2872 to avoid a timing attack. Applications that don't want it can call
2873 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2874 They would be ill-advised to do so in most cases.
2875 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2877 *) Change RSA blinding code so that it works when the PRNG is not
2878 seeded (in this case, the secret RSA exponent is abused as
2879 an unpredictable seed -- if it is not unpredictable, there
2880 is no point in blinding anyway). Make RSA blinding thread-safe
2881 by remembering the creator's thread ID in rsa->blinding and
2882 having all other threads use local one-time blinding factors
2883 (this requires more computation than sharing rsa->blinding, but
2884 avoids excessive locking; and if an RSA object is not shared
2885 between threads, blinding will still be very fast).
2888 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2889 ENGINE as defaults for all supported algorithms irrespective of
2890 the 'flags' parameter. 'flags' is now honoured, so applications
2891 should make sure they are passing it correctly.
2894 *) Target "mingw" now allows native Windows code to be generated in
2895 the Cygwin environment as well as with the MinGW compiler.
2898 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2900 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2901 via timing by performing a MAC computation even if incorrrect
2902 block cipher padding has been found. This is a countermeasure
2903 against active attacks where the attacker has to distinguish
2904 between bad padding and a MAC verification error. (CVE-2003-0078)
2906 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2907 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2908 Martin Vuagnoux (EPFL, Ilion)]
2910 *) Make the no-err option work as intended. The intention with no-err
2911 is not to have the whole error stack handling routines removed from
2912 libcrypto, it's only intended to remove all the function name and
2913 reason texts, thereby removing some of the footprint that may not
2914 be interesting if those errors aren't displayed anyway.
2916 NOTE: it's still possible for any application or module to have it's
2917 own set of error texts inserted. The routines are there, just not
2918 used by default when no-err is given.
2921 *) Add support for FreeBSD on IA64.
2922 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2924 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2925 Kerberos function mit_des_cbc_cksum(). Before this change,
2926 the value returned by DES_cbc_cksum() was like the one from
2927 mit_des_cbc_cksum(), except the bytes were swapped.
2928 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2930 *) Allow an application to disable the automatic SSL chain building.
2931 Before this a rather primitive chain build was always performed in
2932 ssl3_output_cert_chain(): an application had no way to send the
2933 correct chain if the automatic operation produced an incorrect result.
2935 Now the chain builder is disabled if either:
2937 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2939 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2941 The reasoning behind this is that an application would not want the
2942 auto chain building to take place if extra chain certificates are
2943 present and it might also want a means of sending no additional
2944 certificates (for example the chain has two certificates and the
2948 *) Add the possibility to build without the ENGINE framework.
2949 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2951 *) Under Win32 gmtime() can return NULL: check return value in
2952 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2955 *) DSA routines: under certain error conditions uninitialized BN objects
2956 could be freed. Solution: make sure initialization is performed early
2957 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2958 Nils Larsch <nla@trustcenter.de> via PR#459)
2961 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2962 checked on reconnect on the client side, therefore session resumption
2963 could still fail with a "ssl session id is different" error. This
2964 behaviour is masked when SSL_OP_ALL is used due to
2965 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2966 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2967 followup to PR #377.
2970 *) IA-32 assembler support enhancements: unified ELF targets, support
2971 for SCO/Caldera platforms, fix for Cygwin shared build.
2974 *) Add support for FreeBSD on sparc64. As a consequence, support for
2975 FreeBSD on non-x86 processors is separate from x86 processors on
2976 the config script, much like the NetBSD support.
2977 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2979 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2981 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2984 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2985 code (06) was taken as the first octet of the session ID and the last
2986 octet was ignored consequently. As a result SSLv2 client side session
2987 caching could not have worked due to the session ID mismatch between
2989 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2993 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2994 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2998 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2999 seems that in spite of existing for more than a year, many application
3000 author have done nothing to provide the necessary callbacks, which
3001 means that this particular engine will not work properly anywhere.
3002 This is a very unfortunate situation which forces us, in the name
3003 of usability, to give the hw_ncipher.c a static lock, which is part
3005 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3006 appear in 0.9.8 or later. We EXPECT application authors to have
3007 dealt properly with this when 0.9.8 is released (unless we actually
3008 make such changes in the libcrypto locking code that changes will
3009 have to be made anyway).
3012 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3013 octets have been read, EOF or an error occurs. Without this change
3014 some truncated ASN1 structures will not produce an error.
3017 *) Disable Heimdal support, since it hasn't been fully implemented.
3018 Still give the possibility to force the use of Heimdal, but with
3019 warnings and a request that patches get sent to openssl-dev.
3022 *) Add the VC-CE target, introduce the WINCE sysname, and add
3023 INSTALL.WCE and appropriate conditionals to make it build.
3024 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3026 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3027 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3028 edit numbers of the version.
3029 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3031 *) Introduce safe string copy and catenation functions
3032 (BUF_strlcpy() and BUF_strlcat()).
3033 [Ben Laurie (CHATS) and Richard Levitte]
3035 *) Avoid using fixed-size buffers for one-line DNs.
3036 [Ben Laurie (CHATS)]
3038 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3039 resizing buffers containing secrets, and use where appropriate.
3040 [Ben Laurie (CHATS)]
3042 *) Avoid using fixed size buffers for configuration file location.
3043 [Ben Laurie (CHATS)]
3045 *) Avoid filename truncation for various CA files.
3046 [Ben Laurie (CHATS)]
3048 *) Use sizeof in preference to magic numbers.
3049 [Ben Laurie (CHATS)]
3051 *) Avoid filename truncation in cert requests.
3052 [Ben Laurie (CHATS)]
3054 *) Add assertions to check for (supposedly impossible) buffer
3056 [Ben Laurie (CHATS)]
3058 *) Don't cache truncated DNS entries in the local cache (this could
3059 potentially lead to a spoofing attack).
3060 [Ben Laurie (CHATS)]
3062 *) Fix various buffers to be large enough for hex/decimal
3063 representations in a platform independent manner.
3064 [Ben Laurie (CHATS)]
3066 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3067 resizing buffers containing secrets, and use where appropriate.
3068 [Ben Laurie (CHATS)]
3070 *) Add BIO_indent() to avoid much slightly worrying code to do
3072 [Ben Laurie (CHATS)]
3074 *) Convert sprintf()/BIO_puts() to BIO_printf().
3075 [Ben Laurie (CHATS)]
3077 *) buffer_gets() could terminate with the buffer only half
3079 [Ben Laurie (CHATS)]
3081 *) Add assertions to prevent user-supplied crypto functions from
3082 overflowing internal buffers by having large block sizes, etc.
3083 [Ben Laurie (CHATS)]
3085 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3087 [Ben Laurie (CHATS)]
3089 *) Eliminate unused copy of key in RC4.
3090 [Ben Laurie (CHATS)]
3092 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3093 [Ben Laurie (CHATS)]
3095 *) Fix off-by-one error in EGD path.
3096 [Ben Laurie (CHATS)]
3098 *) If RANDFILE path is too long, ignore instead of truncating.
3099 [Ben Laurie (CHATS)]
3101 *) Eliminate unused and incorrectly sized X.509 structure
3103 [Ben Laurie (CHATS)]
3105 *) Eliminate unused and dangerous function knumber().
3106 [Ben Laurie (CHATS)]
3108 *) Eliminate unused and dangerous structure, KSSL_ERR.
3109 [Ben Laurie (CHATS)]
3111 *) Protect against overlong session ID context length in an encoded
3112 session object. Since these are local, this does not appear to be
3114 [Ben Laurie (CHATS)]
3116 *) Change from security patch (see 0.9.6e below) that did not affect
3117 the 0.9.6 release series:
3119 Remote buffer overflow in SSL3 protocol - an attacker could
3120 supply an oversized master key in Kerberos-enabled versions.
3122 [Ben Laurie (CHATS)]
3124 *) Change the SSL kerb5 codes to match RFC 2712.
3127 *) Make -nameopt work fully for req and add -reqopt switch.
3128 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3130 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3131 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3133 *) Make sure tests can be performed even if the corresponding algorithms
3134 have been removed entirely. This was also the last step to make
3135 OpenSSL compilable with DJGPP under all reasonable conditions.
3136 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3138 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3139 to allow version independent disabling of normally unselected ciphers,
3140 which may be activated as a side-effect of selecting a single cipher.
3142 (E.g., cipher list string "RSA" enables ciphersuites that are left
3143 out of "ALL" because they do not provide symmetric encryption.
3144 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3145 [Lutz Jaenicke, Bodo Moeller]
3147 *) Add appropriate support for separate platform-dependent build
3148 directories. The recommended way to make a platform-dependent
3149 build directory is the following (tested on Linux), maybe with
3152 # Place yourself outside of the OpenSSL source tree. In
3153 # this example, the environment variable OPENSSL_SOURCE
3154 # is assumed to contain the absolute OpenSSL source directory.
3155 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3156 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3157 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3158 mkdir -p `dirname $F`
3159 ln -s $OPENSSL_SOURCE/$F $F
3162 To be absolutely sure not to disturb the source tree, a "make clean"
3163 is a good thing. If it isn't successfull, don't worry about it,
3164 it probably means the source directory is very clean.
3167 *) Make sure any ENGINE control commands make local copies of string
3168 pointers passed to them whenever necessary. Otherwise it is possible
3169 the caller may have overwritten (or deallocated) the original string
3170 data when a later ENGINE operation tries to use the stored values.
3171 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3173 *) Improve diagnostics in file reading and command-line digests.
3174 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3176 *) Add AES modes CFB and OFB to the object database. Correct an
3177 error in AES-CFB decryption.
3180 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3181 allows existing EVP_CIPHER_CTX structures to be reused after
3182 calling EVP_*Final(). This behaviour is used by encryption
3183 BIOs and some applications. This has the side effect that
3184 applications must explicitly clean up cipher contexts with
3185 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3188 *) Check the values of dna and dnb in bn_mul_recursive before calling
3189 bn_mul_comba (a non zero value means the a or b arrays do not contain
3190 n2 elements) and fallback to bn_mul_normal if either is not zero.
3193 *) Fix escaping of non-ASCII characters when using the -subj option
3194 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3197 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3198 form for "surname", serialNumber has no short form.
3199 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3200 therefore remove "mail" short name for "internet 7".
3201 The OID for unique identifiers in X509 certificates is
3202 x500UniqueIdentifier, not uniqueIdentifier.
3203 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3206 *) Add an "init" command to the ENGINE config module and auto initialize
3207 ENGINEs. Without any "init" command the ENGINE will be initialized
3208 after all ctrl commands have been executed on it. If init=1 the
3209 ENGINE is initailized at that point (ctrls before that point are run
3210 on the uninitialized ENGINE and after on the initialized one). If
3211 init=0 then the ENGINE will not be iniatialized at all.
3214 *) Fix the 'app_verify_callback' interface so that the user-defined
3215 argument is actually passed to the callback: In the
3216 SSL_CTX_set_cert_verify_callback() prototype, the callback
3217 declaration has been changed from
3220 int (*cb)(X509_STORE_CTX *,void *);
3221 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3222 i=s->ctx->app_verify_callback(&ctx)
3223 has been changed into
3224 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3226 To update applications using SSL_CTX_set_cert_verify_callback(),
3227 a dummy argument can be added to their callback functions.
3228 [D. K. Smetters <smetters@parc.xerox.com>]
3230 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3231 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3233 *) Add and OPENSSL_LOAD_CONF define which will cause
3234 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3235 This allows older applications to transparently support certain
3236 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3237 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3238 load the config file and OPENSSL_add_all_algorithms_conf() which will
3239 always load it have also been added.
3242 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3243 Adjust NIDs and EVP layer.
3244 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3246 *) Config modules support in openssl utility.
3248 Most commands now load modules from the config file,
3249 though in a few (such as version) this isn't done
3250 because it couldn't be used for anything.
3252 In the case of ca and req the config file used is
3253 the same as the utility itself: that is the -config
3254 command line option can be used to specify an
3258 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3259 use "openssl_conf" if filename is NULL use default openssl config file.
3262 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3263 config section name. Add a new flag to tolerate a missing config file
3264 and move code to CONF_modules_load_file().
3267 *) Support for crypto accelerator cards from Accelerated Encryption
3268 Processing, www.aep.ie. (Use engine 'aep')
3269 The support was copied from 0.9.6c [engine] and adapted/corrected
3270 to work with the new engine framework.
3271 [AEP Inc. and Richard Levitte]
3273 *) Support for SureWare crypto accelerator cards from Baltimore
3274 Technologies. (Use engine 'sureware')
3275 The support was copied from 0.9.6c [engine] and adapted
3276 to work with the new engine framework.
3279 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3280 make the newer ENGINE framework commands for the CHIL engine work.
3281 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3283 *) Make it possible to produce shared libraries on ReliantUNIX.
3284 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3286 *) Add the configuration target debug-linux-ppro.
3287 Make 'openssl rsa' use the general key loading routines
3288 implemented in apps.c, and make those routines able to
3289 handle the key format FORMAT_NETSCAPE and the variant
3291 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3293 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3294 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3296 *) Add -keyform to rsautl, and document -engine.
3297 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3299 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3300 BIO_R_NO_SUCH_FILE error code rather than the generic
3301 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3304 *) Add new functions
3306 ERR_peek_last_error_line
3307 ERR_peek_last_error_line_data.
3308 These are similar to
3311 ERR_peek_error_line_data,
3312 but report on the latest error recorded rather than the first one
3313 still in the error queue.
3314 [Ben Laurie, Bodo Moeller]
3316 *) default_algorithms option in ENGINE config module. This allows things
3318 default_algorithms = ALL
3319 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3322 *) Prelminary ENGINE config module.
3325 *) New experimental application configuration code.
3328 *) Change the AES code to follow the same name structure as all other
3329 symmetric ciphers, and behave the same way. Move everything to
3330 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3331 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3333 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3334 [Ben Laurie and Theo de Raadt]
3336 *) Add option to output public keys in req command.
3337 [Massimiliano Pala madwolf@openca.org]
3339 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3340 (up to about 10% better than before for P-192 and P-224).
3343 *) New functions/macros
3345 SSL_CTX_set_msg_callback(ctx, cb)
3346 SSL_CTX_set_msg_callback_arg(ctx, arg)
3347 SSL_set_msg_callback(ssl, cb)
3348 SSL_set_msg_callback_arg(ssl, arg)
3350 to request calling a callback function
3352 void cb(int write_p, int version, int content_type,
3353 const void *buf, size_t len, SSL *ssl, void *arg)
3355 whenever a protocol message has been completely received
3356 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3357 protocol version according to which the SSL library interprets
3358 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3359 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3360 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3361 specification (change_cipher_spec(20), alert(21), handshake(22)).
3362 'buf' and 'len' point to the actual message, 'ssl' to the
3363 SSL object, and 'arg' is the application-defined value set by
3364 SSL[_CTX]_set_msg_callback_arg().
3366 'openssl s_client' and 'openssl s_server' have new '-msg' options
3367 to enable a callback that displays all protocol messages.
3370 *) Change the shared library support so shared libraries are built as
3371 soon as the corresponding static library is finished, and thereby get
3372 openssl and the test programs linked against the shared library.
3373 This still only happens when the keyword "shard" has been given to
3374 the configuration scripts.
3376 NOTE: shared library support is still an experimental thing, and
3377 backward binary compatibility is still not guaranteed.
3378 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3380 *) Add support for Subject Information Access extension.
3381 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3383 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3384 additional bytes when new memory had to be allocated, not just
3385 when reusing an existing buffer.
3388 *) New command line and configuration option 'utf8' for the req command.
3389 This allows field values to be specified as UTF8 strings.
3392 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3393 runs for the former and machine-readable output for the latter.
3396 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3397 of the e-mail address in the DN (i.e., it will go into a certificate
3398 extension only). The new configuration file option 'email_in_dn = no'
3399 has the same effect.
3400 [Massimiliano Pala madwolf@openca.org]
3402 *) Change all functions with names starting with des_ to be starting
3403 with DES_ instead. Add wrappers that are compatible with libdes,
3404 but are named _ossl_old_des_*. Finally, add macros that map the
3405 des_* symbols to the corresponding _ossl_old_des_* if libdes
3406 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3407 desired, the des_* symbols will be mapped to DES_*, with one
3410 Since we provide two compatibility mappings, the user needs to
3411 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3412 compatibility is desired. The default (i.e., when that macro
3413 isn't defined) is OpenSSL 0.9.6c compatibility.
3415 There are also macros that enable and disable the support of old
3416 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3417 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3418 are defined, the default will apply: to support the old des routines.
3420 In either case, one must include openssl/des.h to get the correct
3421 definitions. Do not try to just include openssl/des_old.h, that
3424 NOTE: This is a major break of an old API into a new one. Software
3425 authors are encouraged to switch to the DES_ style functions. Some
3426 time in the future, des_old.h and the libdes compatibility functions
3427 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3428 default), and then completely removed.
3431 *) Test for certificates which contain unsupported critical extensions.
3432 If such a certificate is found during a verify operation it is
3433 rejected by default: this behaviour can be overridden by either
3434 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3435 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3436 X509_supported_extension() has also been added which returns 1 if a
3437 particular extension is supported.
3440 *) Modify the behaviour of EVP cipher functions in similar way to digests
3441 to retain compatibility with existing code.
3444 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3445 compatibility with existing code. In particular the 'ctx' parameter does
3446 not have to be to be initialized before the call to EVP_DigestInit() and
3447 it is tidied up after a call to EVP_DigestFinal(). New function
3448 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3449 EVP_MD_CTX_copy() changed to not require the destination to be
3450 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3451 requires the destination to be valid.
3453 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3454 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3457 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3458 so that complete 'Handshake' protocol structures are kept in memory
3459 instead of overwriting 'msg_type' and 'length' with 'body' data.
3462 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3463 [Massimo Santin via Richard Levitte]
3465 *) Major restructuring to the underlying ENGINE code. This includes
3466 reduction of linker bloat, separation of pure "ENGINE" manipulation
3467 (initialisation, etc) from functionality dealing with implementations
3468 of specific crypto iterfaces. This change also introduces integrated
3469 support for symmetric ciphers and digest implementations - so ENGINEs
3470 can now accelerate these by providing EVP_CIPHER and EVP_MD
3471 implementations of their own. This is detailed in crypto/engine/README
3472 as it couldn't be adequately described here. However, there are a few
3473 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3474 were changed in the original introduction of ENGINE code have now
3475 reverted back - the hooking from this code to ENGINE is now a good
3476 deal more passive and at run-time, operations deal directly with
3477 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3478 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3479 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3480 they were not being used by the framework as there is no concept of a
3481 BIGNUM_METHOD and they could not be generalised to the new
3482 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3483 ENGINE_cpy() has been removed as it cannot be consistently defined in
3487 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3490 *) Change mkdef.pl to sort symbols that get the same entry number,
3491 and make sure the automatically generated functions ERR_load_*
3492 become part of libeay.num as well.
3495 *) New function SSL_renegotiate_pending(). This returns true once
3496 renegotiation has been requested (either SSL_renegotiate() call
3497 or HelloRequest/ClientHello receveived from the peer) and becomes
3498 false once a handshake has been completed.
3499 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3500 sends a HelloRequest, but does not ensure that a handshake takes
3501 place. SSL_renegotiate_pending() is useful for checking if the
3502 client has followed the request.)
3505 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3506 By default, clients may request session resumption even during
3507 renegotiation (if session ID contexts permit); with this option,
3508 session resumption is possible only in the first handshake.
3510 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3511 more bits available for options that should not be part of
3512 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3515 *) Add some demos for certificate and certificate request creation.
3518 *) Make maximum certificate chain size accepted from the peer application
3519 settable (SSL*_get/set_max_cert_list()), as proposed by
3520 "Douglas E. Engert" <deengert@anl.gov>.
3523 *) Add support for shared libraries for Unixware-7
3524 (Boyd Lynn Gerber <gerberb@zenez.com>).
3527 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3528 be done prior to destruction. Use this to unload error strings from
3529 ENGINEs that load their own error strings. NB: This adds two new API
3530 functions to "get" and "set" this destroy handler in an ENGINE.
3533 *) Alter all existing ENGINE implementations (except "openssl" and
3534 "openbsd") to dynamically instantiate their own error strings. This
3535 makes them more flexible to be built both as statically-linked ENGINEs
3536 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3537 Also, add stub code to each that makes building them as self-contained
3538 shared-libraries easier (see README.ENGINE).
3541 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3542 implementations into applications that are completely implemented in
3543 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3544 commands that can be used to configure what shared-library to load and
3545 to control aspects of the way it is handled. Also, made an update to
3546 the README.ENGINE file that brings its information up-to-date and
3547 provides some information and instructions on the "dynamic" ENGINE
3548 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3551 *) Make it possible to unload ranges of ERR strings with a new
3552 "ERR_unload_strings" function.
3555 *) Add a copy() function to EVP_MD.
3558 *) Make EVP_MD routines take a context pointer instead of just the
3559 md_data void pointer.
3562 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3563 that the digest can only process a single chunk of data
3564 (typically because it is provided by a piece of
3565 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3566 is only going to provide a single chunk of data, and hence the
3567 framework needn't accumulate the data for oneshot drivers.
3570 *) As with "ERR", make it possible to replace the underlying "ex_data"
3571 functions. This change also alters the storage and management of global
3572 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3573 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3574 index counters. The API functions that use this state have been changed
3575 to take a "class_index" rather than pointers to the class's local STACK
3576 and counter, and there is now an API function to dynamically create new
3577 classes. This centralisation allows us to (a) plug a lot of the
3578 thread-safety problems that existed, and (b) makes it possible to clean
3579 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3580 such data would previously have always leaked in application code and
3581 workarounds were in place to make the memory debugging turn a blind eye
3582 to it. Application code that doesn't use this new function will still
3583 leak as before, but their memory debugging output will announce it now
3584 rather than letting it slide.
3586 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3587 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3588 has a return value to indicate success or failure.
3591 *) Make it possible to replace the underlying "ERR" functions such that the
3592 global state (2 LHASH tables and 2 locks) is only used by the "default"
3593 implementation. This change also adds two functions to "get" and "set"
3594 the implementation prior to it being automatically set the first time
3595 any other ERR function takes place. Ie. an application can call "get",
3596 pass the return value to a module it has just loaded, and that module
3597 can call its own "set" function using that value. This means the
3598 module's "ERR" operations will use (and modify) the error state in the
3599 application and not in its own statically linked copy of OpenSSL code.
3602 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3603 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3604 the operation, and provides a more encapsulated way for external code
3605 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3606 to use these functions rather than manually incrementing the counts.
3608 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3611 *) Add EVP test program.
3614 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3617 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3618 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3619 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3620 These allow a CRL to be built without having to access X509_CRL fields
3621 directly. Modify 'ca' application to use new functions.
3624 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3625 bug workarounds. Rollback attack detection is a security feature.
3626 The problem will only arise on OpenSSL servers when TLSv1 is not
3627 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3628 Software authors not wanting to support TLSv1 will have special reasons
3629 for their choice and can explicitly enable this option.
3630 [Bodo Moeller, Lutz Jaenicke]
3632 *) Rationalise EVP so it can be extended: don't include a union of
3633 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3634 (similar to those existing for EVP_CIPHER_CTX).
3639 EVP_MD_CTX_init(&md); /* new function call */
3640 EVP_DigestInit(&md, EVP_sha1());
3641 EVP_DigestUpdate(&md, in, len);
3642 EVP_DigestFinal(&md, out, NULL);
3643 EVP_MD_CTX_cleanup(&md); /* new function call */
3647 *) Make DES key schedule conform to the usual scheme, as well as
3648 correcting its structure. This means that calls to DES functions
3649 now have to pass a pointer to a des_key_schedule instead of a
3650 plain des_key_schedule (which was actually always a pointer
3653 des_key_schedule ks;
3655 des_set_key_checked(..., &ks);
3656 des_ncbc_encrypt(..., &ks, ...);
3658 (Note that a later change renames 'des_...' into 'DES_...'.)
3661 *) Initial reduction of linker bloat: the use of some functions, such as
3662 PEM causes large amounts of unused functions to be linked in due to
3663 poor organisation. For example pem_all.c contains every PEM function
3664 which has a knock on effect of linking in large amounts of (unused)
3665 ASN1 code. Grouping together similar functions and splitting unrelated
3666 functions prevents this.
3669 *) Cleanup of EVP macros.
3672 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3673 correct _ecb suffix.
3676 *) Add initial OCSP responder support to ocsp application. The
3677 revocation information is handled using the text based index
3678 use by the ca application. The responder can either handle
3679 requests generated internally, supplied in files (for example
3680 via a CGI script) or using an internal minimal server.
3683 *) Add configuration choices to get zlib compression for TLS.
3686 *) Changes to Kerberos SSL for RFC 2712 compliance:
3687 1. Implemented real KerberosWrapper, instead of just using
3688 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3689 2. Implemented optional authenticator field of KerberosWrapper.
3691 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3692 and authenticator structs; see crypto/krb5/.
3694 Generalized Kerberos calls to support multiple Kerberos libraries.
3695 [Vern Staats <staatsvr@asc.hpc.mil>,
3696 Jeffrey Altman <jaltman@columbia.edu>
3697 via Richard Levitte]
3699 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3700 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3701 values for each of the key sizes rather than having just
3702 parameters (and 'speed' generating keys each time).
3705 *) Speed up EVP routines.
3708 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3709 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3710 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3711 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3713 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3714 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3715 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3718 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3720 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3723 *) Added the OS2-EMX target.
3724 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3726 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3727 to support NCONF routines in extension code. New function CONF_set_nconf()
3728 to allow functions which take an NCONF to also handle the old LHASH
3729 structure: this means that the old CONF compatible routines can be
3730 retained (in particular wrt extensions) without having to duplicate the
3731 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3734 *) Enhance the general user interface with mechanisms for inner control
3735 and with possibilities to have yes/no kind of prompts.
3738 *) Change all calls to low level digest routines in the library and
3739 applications to use EVP. Add missing calls to HMAC_cleanup() and
3740 don't assume HMAC_CTX can be copied using memcpy().
3741 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3743 *) Add the possibility to control engines through control names but with
3744 arbitrary arguments instead of just a string.
3745 Change the key loaders to take a UI_METHOD instead of a callback
3746 function pointer. NOTE: this breaks binary compatibility with earlier
3747 versions of OpenSSL [engine].
3748 Adapt the nCipher code for these new conditions and add a card insertion
3752 *) Enhance the general user interface with mechanisms to better support
3753 dialog box interfaces, application-defined prompts, the possibility
3754 to use defaults (for example default passwords from somewhere else)
3755 and interrupts/cancellations.
3758 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3759 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3762 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3763 tidy up some unnecessarily weird code in 'sk_new()').
3764 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3766 *) Change the key loading routines for ENGINEs to use the same kind
3767 callback (pem_password_cb) as all other routines that need this
3771 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3772 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3773 than this minimum value is recommended.
3776 *) New random seeder for OpenVMS, using the system process statistics
3777 that are easily reachable.
3780 *) Windows apparently can't transparently handle global
3781 variables defined in DLLs. Initialisations such as:
3783 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3785 wont compile. This is used by the any applications that need to
3786 declare their own ASN1 modules. This was fixed by adding the option
3787 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3788 needed for static libraries under Win32.
3791 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3792 setting of purpose and trust fields. New X509_STORE trust and
3793 purpose functions and tidy up setting in other SSL functions.
3796 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3797 structure. These are inherited by X509_STORE_CTX when it is
3798 initialised. This allows various defaults to be set in the
3799 X509_STORE structure (such as flags for CRL checking and custom
3800 purpose or trust settings) for functions which only use X509_STORE_CTX
3801 internally such as S/MIME.
3803 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3804 trust settings if they are not set in X509_STORE. This allows X509_STORE
3805 purposes and trust (in S/MIME for example) to override any set by default.
3807 Add command line options for CRL checking to smime, s_client and s_server
3811 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3812 are set then the CRL is looked up in the X509_STORE structure and
3813 its validity and signature checked, then if the certificate is found
3814 in the CRL the verify fails with a revoked error.
3816 Various new CRL related callbacks added to X509_STORE_CTX structure.
3818 Command line options added to 'verify' application to support this.
3820 This needs some additional work, such as being able to handle multiple
3821 CRLs with different times, extension based lookup (rather than just
3822 by subject name) and ultimately more complete V2 CRL extension
3826 *) Add a general user interface API (crypto/ui/). This is designed
3827 to replace things like des_read_password and friends (backward
3828 compatibility functions using this new API are provided).
3829 The purpose is to remove prompting functions from the DES code
3830 section as well as provide for prompting through dialog boxes in
3831 a window system and the like.
3834 *) Add "ex_data" support to ENGINE so implementations can add state at a
3835 per-structure level rather than having to store it globally.
3838 *) Make it possible for ENGINE structures to be copied when retrieved by
3839 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3840 This causes the "original" ENGINE structure to act like a template,
3841 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3842 operational state can be localised to each ENGINE structure, despite the
3843 fact they all share the same "methods". New ENGINE structures returned in
3844 this case have no functional references and the return value is the single
3845 structural reference. This matches the single structural reference returned
3846 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3850 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3851 needs to match any other type at all we need to manually clear the
3855 *) Changes to the "openssl engine" utility to include;
3856 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3857 about an ENGINE's available control commands.
3858 - executing control commands from command line arguments using the
3859 '-pre' and '-post' switches. '-post' is only used if '-t' is
3860 specified and the ENGINE is successfully initialised. The syntax for
3861 the individual commands are colon-separated, for example;
3862 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3865 *) New dynamic control command support for ENGINEs. ENGINEs can now
3866 declare their own commands (numbers), names (strings), descriptions,
3867 and input types for run-time discovery by calling applications. A
3868 subset of these commands are implicitly classed as "executable"
3869 depending on their input type, and only these can be invoked through
3870 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3871 can be based on user input, config files, etc). The distinction is
3872 that "executable" commands cannot return anything other than a boolean
3873 result and can only support numeric or string input, whereas some
3874 discoverable commands may only be for direct use through
3875 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3876 pointers, or other custom uses. The "executable" commands are to
3877 support parameterisations of ENGINE behaviour that can be
3878 unambiguously defined by ENGINEs and used consistently across any
3879 OpenSSL-based application. Commands have been added to all the
3880 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3881 control over shared-library paths without source code alterations.
3884 *) Changed all ENGINE implementations to dynamically allocate their
3885 ENGINEs rather than declaring them statically. Apart from this being
3886 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3887 this also allows the implementations to compile without using the
3888 internal engine_int.h header.
3891 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3892 'const' value. Any code that should be able to modify a RAND_METHOD
3893 should already have non-const pointers to it (ie. they should only
3894 modify their own ones).
3897 *) Made a variety of little tweaks to the ENGINE code.
3898 - "atalla" and "ubsec" string definitions were moved from header files
3899 to C code. "nuron" string definitions were placed in variables
3900 rather than hard-coded - allowing parameterisation of these values
3901 later on via ctrl() commands.
3902 - Removed unused "#if 0"'d code.
3903 - Fixed engine list iteration code so it uses ENGINE_free() to release
3904 structural references.
3905 - Constified the RAND_METHOD element of ENGINE structures.
3906 - Constified various get/set functions as appropriate and added
3907 missing functions (including a catch-all ENGINE_cpy that duplicates
3908 all ENGINE values onto a new ENGINE except reference counts/state).
3909 - Removed NULL parameter checks in get/set functions. Setting a method
3910 or function to NULL is a way of cancelling out a previously set
3911 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3912 and doesn't justify the extra error symbols and code.
3913 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3914 flags from engine_int.h to engine.h.
3915 - Changed prototypes for ENGINE handler functions (init(), finish(),
3916 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3919 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3920 to the algorithm using long division. The binary algorithm can be
3921 used only if the modulus is odd. On 32-bit systems, it is faster
3922 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3923 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3924 up to 450 bits. In 64-bit environments, the binary algorithm
3925 appears to be advantageous for much longer moduli; here we use it
3926 for moduli up to 2048 bits.
3929 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3930 could not support the combine flag in choice fields.
3933 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3934 extensions from a certificate request to the certificate.
3937 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3938 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3939 file: this allows the display of the certificate about to be
3940 signed to be customised, to allow certain fields to be included
3941 or excluded and extension details. The old system didn't display
3942 multicharacter strings properly, omitted fields not in the policy
3943 and couldn't display additional details such as extensions.
3946 *) Function EC_POINTs_mul for multiple scalar multiplication
3947 of an arbitrary number of elliptic curve points
3948 \sum scalars[i]*points[i],
3949 optionally including the generator defined for the EC_GROUP:
3950 scalar*generator + \sum scalars[i]*points[i].
3952 EC_POINT_mul is a simple wrapper function for the typical case
3953 that the point list has just one item (besides the optional
3957 *) First EC_METHODs for curves over GF(p):
3959 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3960 operations and provides various method functions that can also
3961 operate with faster implementations of modular arithmetic.
3963 EC_GFp_mont_method() reuses most functions that are part of
3964 EC_GFp_simple_method, but uses Montgomery arithmetic.
3966 [Bodo Moeller; point addition and point doubling
3967 implementation directly derived from source code provided by
3968 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3970 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3971 crypto/ec/ec_lib.c):
3973 Curves are EC_GROUP objects (with an optional group generator)
3974 based on EC_METHODs that are built into the library.
3976 Points are EC_POINT objects based on EC_GROUP objects.
3978 Most of the framework would be able to handle curves over arbitrary
3979 finite fields, but as there are no obvious types for fields other
3980 than GF(p), some functions are limited to that for now.
3983 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3984 that the file contains a complete HTTP response.
3987 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3988 change the def and num file printf format specifier from "%-40sXXX"
3989 to "%-39s XXX". The latter will always guarantee a space after the
3990 field while the former will cause them to run together if the field
3991 is 40 of more characters long.
3994 *) Constify the cipher and digest 'method' functions and structures
3995 and modify related functions to take constant EVP_MD and EVP_CIPHER
3999 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4000 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4003 *) Modify EVP_Digest*() routines so they now return values. Although the
4004 internal software routines can never fail additional hardware versions
4008 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4010 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4011 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4014 ERR_R_NESTED_ASN1_ERROR
4016 ERR_R_MISSING_ASN1_EOS
4017 were 4 .. 9, conflicting with
4018 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4020 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4021 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4023 Add new error code 'ERR_R_INTERNAL_ERROR'.
4026 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4030 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4031 sets the subject name for a new request or supersedes the
4032 subject name in a given request. Formats that can be parsed are
4033 'CN=Some Name, OU=myOU, C=IT'
4035 'CN=Some Name/OU=myOU/C=IT'.
4037 Add options '-batch' and '-verbose' to 'openssl req'.
4038 [Massimiliano Pala <madwolf@hackmasters.net>]
4040 *) Introduce the possibility to access global variables through
4041 functions on platform were that's the best way to handle exporting
4042 global variables in shared libraries. To enable this functionality,
4043 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4044 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4045 is normally done by Configure or something similar).
4047 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4048 in the source file (foo.c) like this:
4050 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4051 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4053 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4054 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4056 OPENSSL_DECLARE_GLOBAL(int,foo);
4057 #define foo OPENSSL_GLOBAL_REF(foo)
4058 OPENSSL_DECLARE_GLOBAL(double,bar);
4059 #define bar OPENSSL_GLOBAL_REF(bar)
4061 The #defines are very important, and therefore so is including the
4062 header file everywhere where the defined globals are used.
4064 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4065 of ASN.1 items, but that structure is a bit different.
4067 The largest change is in util/mkdef.pl which has been enhanced with
4068 better and easier to understand logic to choose which symbols should
4069 go into the Windows .def files as well as a number of fixes and code
4070 cleanup (among others, algorithm keywords are now sorted
4071 lexicographically to avoid constant rewrites).
4074 *) In BN_div() keep a copy of the sign of 'num' before writing the
4075 result to 'rm' because if rm==num the value will be overwritten
4076 and produce the wrong result if 'num' is negative: this caused
4077 problems with BN_mod() and BN_nnmod().
4080 *) Function OCSP_request_verify(). This checks the signature on an
4081 OCSP request and verifies the signer certificate. The signer
4082 certificate is just checked for a generic purpose and OCSP request
4086 *) Add OCSP_check_validity() function to check the validity of OCSP
4087 responses. OCSP responses are prepared in real time and may only
4088 be a few seconds old. Simply checking that the current time lies
4089 between thisUpdate and nextUpdate max reject otherwise valid responses
4090 caused by either OCSP responder or client clock inaccuracy. Instead
4091 we allow thisUpdate and nextUpdate to fall within a certain period of
4092 the current time. The age of the response can also optionally be
4093 checked. Two new options -validity_period and -status_age added to
4097 *) If signature or public key algorithm is unrecognized print out its
4098 OID rather that just UNKNOWN.
4101 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4102 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4103 ID to be generated from the issuer certificate alone which can then be
4104 passed to OCSP_id_issuer_cmp().
4107 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4108 ASN1 modules to export functions returning ASN1_ITEM pointers
4109 instead of the ASN1_ITEM structures themselves. This adds several
4110 new macros which allow the underlying ASN1 function/structure to
4111 be accessed transparently. As a result code should not use ASN1_ITEM
4112 references directly (such as &X509_it) but instead use the relevant
4113 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4114 use of the new ASN1 code on platforms where exporting structures
4115 is problematical (for example in shared libraries) but exporting
4116 functions returning pointers to structures is not.
4119 *) Add support for overriding the generation of SSL/TLS session IDs.
4120 These callbacks can be registered either in an SSL_CTX or per SSL.
4121 The purpose of this is to allow applications to control, if they wish,
4122 the arbitrary values chosen for use as session IDs, particularly as it
4123 can be useful for session caching in multiple-server environments. A
4124 command-line switch for testing this (and any client code that wishes
4125 to use such a feature) has been added to "s_server".
4126 [Geoff Thorpe, Lutz Jaenicke]
4128 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4129 of the form '#if defined(...) || defined(...) || ...' and
4130 '#if !defined(...) && !defined(...) && ...'. This also avoids
4131 the growing number of special cases it was previously handling.
4134 *) Make all configuration macros available for application by making
4135 sure they are available in opensslconf.h, by giving them names starting
4136 with "OPENSSL_" to avoid conflicts with other packages and by making
4137 sure e_os2.h will cover all platform-specific cases together with
4139 Additionally, it is now possible to define configuration/platform-
4140 specific names (called "system identities"). In the C code, these
4141 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4142 macro with the name beginning with "OPENSSL_SYS_", which is determined
4143 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4147 *) New option -set_serial to 'req' and 'x509' this allows the serial
4148 number to use to be specified on the command line. Previously self
4149 signed certificates were hard coded with serial number 0 and the
4150 CA options of 'x509' had to use a serial number in a file which was
4154 *) New options to 'ca' utility to support V2 CRL entry extensions.
4155 Currently CRL reason, invalidity date and hold instruction are
4156 supported. Add new CRL extensions to V3 code and some new objects.
4159 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4160 disable standard block padding (aka PKCS#5 padding) in the EVP
4161 API, which was previously mandatory. This means that the data is
4162 not padded in any way and so the total length much be a multiple
4163 of the block size, otherwise an error occurs.
4166 *) Initial (incomplete) OCSP SSL support.
4169 *) New function OCSP_parse_url(). This splits up a URL into its host,
4170 port and path components: primarily to parse OCSP URLs. New -url
4171 option to ocsp utility.
4174 *) New nonce behavior. The return value of OCSP_check_nonce() now
4175 reflects the various checks performed. Applications can decide
4176 whether to tolerate certain situations such as an absent nonce
4177 in a response when one was present in a request: the ocsp application
4178 just prints out a warning. New function OCSP_add1_basic_nonce()
4179 this is to allow responders to include a nonce in a response even if
4180 the request is nonce-less.
4183 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4184 skipped when using openssl x509 multiple times on a single input file,
4185 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4188 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4189 set string type: to handle setting ASN1_TIME structures. Fix ca
4190 utility to correctly initialize revocation date of CRLs.
4193 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4194 the clients preferred ciphersuites and rather use its own preferences.
4195 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4196 Internet Explorer by ensuring unchanged hash method during stepup.
4197 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4200 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4201 to aes and add a new 'exist' option to print out symbols that don't
4205 *) Additional options to ocsp utility to allow flags to be set and
4206 additional certificates supplied.
4209 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4210 OCSP client a number of certificate to only verify the response
4214 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4215 handle the new API. Currently only ECB, CBC modes supported. Add new
4218 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4219 Encryption Standard (AES) Ciphersuites for Transport Layer
4220 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4221 not enabled by default and were not part of the "ALL" ciphersuite
4222 alias because they were not yet official; they could be
4223 explicitly requested by specifying the "AESdraft" ciphersuite
4224 group alias. In the final release of OpenSSL 0.9.7, the group
4225 alias is called "AES" and is part of "ALL".)
4226 [Ben Laurie, Steve Henson, Bodo Moeller]
4228 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4229 request to response.
4232 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4233 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4234 extract information from a certificate request. OCSP_response_create()
4235 creates a response and optionally adds a basic response structure.
4236 OCSP_basic_add1_status() adds a complete single response to a basic
4237 response and returns the OCSP_SINGLERESP structure just added (to allow
4238 extensions to be included for example). OCSP_basic_add1_cert() adds a
4239 certificate to a basic response and OCSP_basic_sign() signs a basic
4240 response with various flags. New helper functions ASN1_TIME_check()
4241 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4242 (converts ASN1_TIME to GeneralizedTime).
4245 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4246 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4247 structure from a certificate. X509_pubkey_digest() digests the public_key
4248 contents: this is used in various key identifiers.
4251 *) Make sk_sort() tolerate a NULL argument.
4252 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4254 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4255 passed by the function are trusted implicitly. If any of them signed the
4256 response then it is assumed to be valid and is not verified.
4259 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4260 to data. This was previously part of the PKCS7 ASN1 code. This
4261 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4262 [Steve Henson, reported by Kenneth R. Robinette
4263 <support@securenetterm.com>]
4265 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4266 routines: without these tracing memory leaks is very painful.
4267 Fix leaks in PKCS12 and PKCS7 routines.
4270 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4271 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4272 effectively meant GeneralizedTime would never be used. Now it
4273 is initialised to -1 but X509_time_adj() now has to check the value
4274 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4275 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4276 [Steve Henson, reported by Kenneth R. Robinette
4277 <support@securenetterm.com>]
4279 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4280 result in a zero length in the ASN1_INTEGER structure which was
4281 not consistent with the structure when d2i_ASN1_INTEGER() was used
4282 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4283 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4284 where it did not print out a minus for negative ASN1_INTEGER.
4287 *) Add summary printout to ocsp utility. The various functions which
4288 convert status values to strings have been renamed to:
4289 OCSP_response_status_str(), OCSP_cert_status_str() and
4290 OCSP_crl_reason_str() and are no longer static. New options
4291 to verify nonce values and to disable verification. OCSP response
4292 printout format cleaned up.
4295 *) Add additional OCSP certificate checks. These are those specified
4296 in RFC2560. This consists of two separate checks: the CA of the
4297 certificate being checked must either be the OCSP signer certificate
4298 or the issuer of the OCSP signer certificate. In the latter case the
4299 OCSP signer certificate must contain the OCSP signing extended key
4300 usage. This check is performed by attempting to match the OCSP
4301 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4302 in the OCSP_CERTID structures of the response.
4305 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4306 and related routines. This uses the standard OpenSSL certificate
4307 verify routines to perform initial checks (just CA validity) and
4308 to obtain the certificate chain. Then additional checks will be
4309 performed on the chain. Currently the root CA is checked to see
4310 if it is explicitly trusted for OCSP signing. This is used to set
4311 a root CA as a global signing root: that is any certificate that
4312 chains to that CA is an acceptable OCSP signing certificate.
4315 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4316 extensions from a separate configuration file.
4317 As when reading extensions from the main configuration file,
4318 the '-extensions ...' option may be used for specifying the
4320 [Massimiliano Pala <madwolf@comune.modena.it>]
4322 *) New OCSP utility. Allows OCSP requests to be generated or
4323 read. The request can be sent to a responder and the output
4324 parsed, outputed or printed in text form. Not complete yet:
4325 still needs to check the OCSP response validity.
4328 *) New subcommands for 'openssl ca':
4329 'openssl ca -status <serial>' prints the status of the cert with
4330 the given serial number (according to the index file).
4331 'openssl ca -updatedb' updates the expiry status of certificates
4333 [Massimiliano Pala <madwolf@comune.modena.it>]
4335 *) New '-newreq-nodes' command option to CA.pl. This is like
4336 '-newreq', but calls 'openssl req' with the '-nodes' option
4337 so that the resulting key is not encrypted.
4338 [Damien Miller <djm@mindrot.org>]
4340 *) New configuration for the GNU Hurd.
4341 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4343 *) Initial code to implement OCSP basic response verify. This
4344 is currently incomplete. Currently just finds the signer's
4345 certificate and verifies the signature on the response.
4348 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4349 value of OPENSSLDIR. This is available via the new '-d' option
4350 to 'openssl version', and is also included in 'openssl version -a'.
4353 *) Allowing defining memory allocation callbacks that will be given
4354 file name and line number information in additional arguments
4355 (a const char* and an int). The basic functionality remains, as
4356 well as the original possibility to just replace malloc(),
4357 realloc() and free() by functions that do not know about these
4358 additional arguments. To register and find out the current
4359 settings for extended allocation functions, the following
4360 functions are provided:
4362 CRYPTO_set_mem_ex_functions
4363 CRYPTO_set_locked_mem_ex_functions
4364 CRYPTO_get_mem_ex_functions
4365 CRYPTO_get_locked_mem_ex_functions
4367 These work the same way as CRYPTO_set_mem_functions and friends.
4368 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4369 extended allocation function is enabled.
4370 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4371 a conventional allocation function is enabled.
4372 [Richard Levitte, Bodo Moeller]
4374 *) Finish off removing the remaining LHASH function pointer casts.
4375 There should no longer be any prototype-casting required when using
4376 the LHASH abstraction, and any casts that remain are "bugs". See
4377 the callback types and macros at the head of lhash.h for details
4378 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4381 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4382 If /dev/[u]random devices are not available or do not return enough
4383 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4385 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4386 /etc/entropy will be queried once each in this sequence, quering stops
4387 when enough entropy was collected without querying more sockets.
4390 *) Change the Unix RAND_poll() variant to be able to poll several
4391 random devices, as specified by DEVRANDOM, until a sufficient amount
4392 of data has been collected. We spend at most 10 ms on each file
4393 (select timeout) and read in non-blocking mode. DEVRANDOM now
4394 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4395 (previously it was just the string "/dev/urandom"), so on typical
4396 platforms the 10 ms delay will never occur.
4397 Also separate out the Unix variant to its own file, rand_unix.c.
4398 For VMS, there's a currently-empty rand_vms.c.
4401 *) Move OCSP client related routines to ocsp_cl.c. These
4402 provide utility functions which an application needing
4403 to issue a request to an OCSP responder and analyse the
4404 response will typically need: as opposed to those which an
4405 OCSP responder itself would need which will be added later.
4407 OCSP_request_sign() signs an OCSP request with an API similar
4408 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4409 response. OCSP_response_get1_basic() extracts basic response
4410 from response. OCSP_resp_find_status(): finds and extracts status
4411 information from an OCSP_CERTID structure (which will be created
4412 when the request structure is built). These are built from lower
4413 level functions which work on OCSP_SINGLERESP structures but
4414 wont normally be used unless the application wishes to examine
4415 extensions in the OCSP response for example.
4417 Replace nonce routines with a pair of functions.
4418 OCSP_request_add1_nonce() adds a nonce value and optionally
4419 generates a random value. OCSP_check_nonce() checks the
4420 validity of the nonce in an OCSP response.
4423 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4424 This doesn't copy the supplied OCSP_CERTID and avoids the
4425 need to free up the newly created id. Change return type
4426 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4427 This can then be used to add extensions to the request.
4428 Deleted OCSP_request_new(), since most of its functionality
4429 is now in OCSP_REQUEST_new() (and the case insensitive name
4430 clash) apart from the ability to set the request name which
4431 will be added elsewhere.
4434 *) Update OCSP API. Remove obsolete extensions argument from
4435 various functions. Extensions are now handled using the new
4436 OCSP extension code. New simple OCSP HTTP function which
4437 can be used to send requests and parse the response.
4440 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4441 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4442 uses the special reorder version of SET OF to sort the attributes
4443 and reorder them to match the encoded order. This resolves a long
4444 standing problem: a verify on a PKCS7 structure just after signing
4445 it used to fail because the attribute order did not match the
4446 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4447 it uses the received order. This is necessary to tolerate some broken
4448 software that does not order SET OF. This is handled by encoding
4449 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4450 to produce the required SET OF.
4453 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4454 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4455 files to get correct declarations of the ASN.1 item variables.
4458 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4459 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4460 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4461 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4462 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4463 ASN1_ITEM and no wrapper functions.
4466 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4467 replace the old function pointer based I/O routines. Change most of
4468 the *_d2i_bio() and *_d2i_fp() functions to use these.
4471 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4472 lines, recognice more "algorithms" that can be deselected, and make
4473 it complain about algorithm deselection that isn't recognised.
4476 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4477 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4478 to use new functions. Add NO_ASN1_OLD which can be set to remove
4479 some old style ASN1 functions: this can be used to determine if old
4480 code will still work when these eventually go away.
4483 *) New extension functions for OCSP structures, these follow the
4484 same conventions as certificates and CRLs.
4487 *) New function X509V3_add1_i2d(). This automatically encodes and
4488 adds an extension. Its behaviour can be customised with various
4489 flags to append, replace or delete. Various wrappers added for
4490 certifcates and CRLs.
4493 *) Fix to avoid calling the underlying ASN1 print routine when
4494 an extension cannot be parsed. Correct a typo in the
4495 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4498 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4499 entries for variables.
4502 *) Add functionality to apps/openssl.c for detecting locking
4503 problems: As the program is single-threaded, all we have
4504 to do is register a locking callback using an array for
4505 storing which locks are currently held by the program.
4508 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4509 SSL_get_ex_data_X509_STORE_idx(), which is used in
4510 ssl_verify_cert_chain() and thus can be called at any time
4511 during TLS/SSL handshakes so that thread-safety is essential.
4512 Unfortunately, the ex_data design is not at all suited
4513 for multi-threaded use, so it probably should be abolished.
4516 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4517 [Broadcom, tweaked and integrated by Geoff Thorpe]
4519 *) Move common extension printing code to new function
4520 X509V3_print_extensions(). Reorganise OCSP print routines and
4521 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4524 *) New function X509_signature_print() to remove duplication in some
4528 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4529 set (this was treated exactly the same as SET OF previously). This
4530 is used to reorder the STACK representing the structure to match the
4531 encoding. This will be used to get round a problem where a PKCS7
4532 structure which was signed could not be verified because the STACK
4533 order did not reflect the encoded order.
4536 *) Reimplement the OCSP ASN1 module using the new code.
4539 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4540 for its ASN1 operations. The old style function pointers still exist
4541 for now but they will eventually go away.
4544 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4545 completely replaces the old ASN1 functionality with a table driven
4546 encoder and decoder which interprets an ASN1_ITEM structure describing
4547 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4548 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4549 has also been converted to the new form.
4552 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4553 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4554 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4555 for negative moduli.
4558 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4559 of not touching the result's sign bit.
4562 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4566 *) Changed the LHASH code to use prototypes for callbacks, and created
4567 macros to declare and implement thin (optionally static) functions
4568 that provide type-safety and avoid function pointer casting for the
4569 type-specific callbacks.
4572 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4574 [Veers Staats <staatsvr@asc.hpc.mil>,
4575 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4577 *) Reformat the FAQ so the different questions and answers can be divided
4578 in sections depending on the subject.
4581 *) Have the zlib compression code load ZLIB.DLL dynamically under
4585 *) New function BN_mod_sqrt for computing square roots modulo a prime
4586 (using the probabilistic Tonelli-Shanks algorithm unless
4587 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4588 be handled deterministically).
4589 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4591 *) Make BN_mod_inverse faster by explicitly handling small quotients
4592 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4593 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4596 *) New function BN_kronecker.
4599 *) Fix BN_gcd so that it works on negative inputs; the result is
4600 positive unless both parameters are zero.
4601 Previously something reasonably close to an infinite loop was
4602 possible because numbers could be growing instead of shrinking
4603 in the implementation of Euclid's algorithm.
4606 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4607 sign of the number in question.
4609 Fix BN_is_word(a,w) to work correctly for w == 0.
4611 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4612 because its test if the absolute value of 'a' equals 'w'.
4613 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4614 it exists mostly for use in the implementations of BN_is_zero(),
4615 BN_is_one(), and BN_is_word().
4618 *) New function BN_swap.
4621 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4622 the exponentiation functions are more likely to produce reasonable
4623 results on negative inputs.
4626 *) Change BN_mod_mul so that the result is always non-negative.
4627 Previously, it could be negative if one of the factors was negative;
4628 I don't think anyone really wanted that behaviour.
4631 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4632 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4633 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4634 and add new functions:
4643 BN_mod_lshift1_quick
4647 These functions always generate non-negative results.
4649 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4650 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4652 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4653 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4654 be reduced modulo m.
4655 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4658 The following entry accidentily appeared in the CHANGES file
4659 distributed with OpenSSL 0.9.7. The modifications described in
4660 it do *not* apply to OpenSSL 0.9.7.
4662 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4663 was actually never needed) and in BN_mul(). The removal in BN_mul()
4664 required a small change in bn_mul_part_recursive() and the addition
4665 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4666 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4667 bn_sub_words() and bn_add_words() except they take arrays with
4672 *) In 'openssl passwd', verify passwords read from the terminal
4673 unless the '-salt' option is used (which usually means that
4674 verification would just waste user's time since the resulting
4675 hash is going to be compared with some given password hash)
4676 or the new '-noverify' option is used.
4678 This is an incompatible change, but it does not affect
4679 non-interactive use of 'openssl passwd' (passwords on the command
4680 line, '-stdin' option, '-in ...' option) and thus should not
4684 *) Remove all references to RSAref, since there's no more need for it.
4687 *) Make DSO load along a path given through an environment variable
4688 (SHLIB_PATH) with shl_load().
4691 *) Constify the ENGINE code as a result of BIGNUM constification.
4692 Also constify the RSA code and most things related to it. In a
4693 few places, most notable in the depth of the ASN.1 code, ugly
4694 casts back to non-const were required (to be solved at a later
4698 *) Make it so the openssl application has all engines loaded by default.
4701 *) Constify the BIGNUM routines a little more.
4704 *) Add the following functions:
4706 ENGINE_load_cswift()
4708 ENGINE_load_atalla()
4710 ENGINE_load_builtin_engines()
4712 That way, an application can itself choose if external engines that
4713 are built-in in OpenSSL shall ever be used or not. The benefit is
4714 that applications won't have to be linked with libdl or other dso
4715 libraries unless it's really needed.
4717 Changed 'openssl engine' to load all engines on demand.
4718 Changed the engine header files to avoid the duplication of some
4719 declarations (they differed!).
4722 *) 'openssl engine' can now list capabilities.
4725 *) Better error reporting in 'openssl engine'.
4728 *) Never call load_dh_param(NULL) in s_server.
4731 *) Add engine application. It can currently list engines by name and
4732 identity, and test if they are actually available.
4735 *) Improve RPM specification file by forcing symbolic linking and making
4736 sure the installed documentation is also owned by root.root.
4737 [Damien Miller <djm@mindrot.org>]
4739 *) Give the OpenSSL applications more possibilities to make use of
4740 keys (public as well as private) handled by engines.
4743 *) Add OCSP code that comes from CertCo.
4746 *) Add VMS support for the Rijndael code.
4749 *) Added untested support for Nuron crypto accelerator.
4752 *) Add support for external cryptographic devices. This code was
4753 previously distributed separately as the "engine" branch.
4754 [Geoff Thorpe, Richard Levitte]
4756 *) Rework the filename-translation in the DSO code. It is now possible to
4757 have far greater control over how a "name" is turned into a filename
4758 depending on the operating environment and any oddities about the
4759 different shared library filenames on each system.
4762 *) Support threads on FreeBSD-elf in Configure.
4765 *) Fix for SHA1 assembly problem with MASM: it produces
4766 warnings about corrupt line number information when assembling
4767 with debugging information. This is caused by the overlapping
4769 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4772 NCONF_get_number() has no error checking at all. As a replacement,
4773 NCONF_get_number_e() is defined (_e for "error checking") and is
4774 promoted strongly. The old NCONF_get_number is kept around for
4775 binary backward compatibility.
4776 Make it possible for methods to load from something other than a BIO,
4777 by providing a function pointer that is given a name instead of a BIO.
4778 For example, this could be used to load configuration data from an
4782 *) Fix for non blocking accept BIOs. Added new I/O special reason
4783 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4784 with non blocking I/O was not possible because no retry code was
4785 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4789 *) Added the beginnings of Rijndael support.
4792 *) Fix for bug in DirectoryString mask setting. Add support for
4793 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4794 to allow certificate printing to more controllable, additional
4795 'certopt' option to 'x509' to allow new printing options to be
4799 *) Clean old EAY MD5 hack from e_os.h.
4802 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4804 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4805 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4806 [Joe Orton, Steve Henson]
4808 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4810 *) Fix additional bug revealed by the NISCC test suite:
4812 Stop bug triggering large recursion when presented with
4813 certain ASN.1 tags (CVE-2003-0851)
4816 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4818 *) Fix various bugs revealed by running the NISCC test suite:
4820 Stop out of bounds reads in the ASN1 code when presented with
4821 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4823 If verify callback ignores invalid public key errors don't try to check
4824 certificate signature with the NULL public key.
4828 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4829 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4833 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4834 extra data after the compression methods not only for TLS 1.0
4835 but also for SSL 3.0 (as required by the specification).
4836 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4838 *) Change X509_certificate_type() to mark the key as exported/exportable
4839 when it's 512 *bits* long, not 512 bytes.
4842 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4844 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4845 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4846 a protocol version number mismatch like a decryption error
4847 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4850 *) Turn on RSA blinding by default in the default implementation
4851 to avoid a timing attack. Applications that don't want it can call
4852 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4853 They would be ill-advised to do so in most cases.
4854 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4856 *) Change RSA blinding code so that it works when the PRNG is not
4857 seeded (in this case, the secret RSA exponent is abused as
4858 an unpredictable seed -- if it is not unpredictable, there
4859 is no point in blinding anyway). Make RSA blinding thread-safe
4860 by remembering the creator's thread ID in rsa->blinding and
4861 having all other threads use local one-time blinding factors
4862 (this requires more computation than sharing rsa->blinding, but
4863 avoids excessive locking; and if an RSA object is not shared
4864 between threads, blinding will still be very fast).
4867 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4869 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4870 via timing by performing a MAC computation even if incorrrect
4871 block cipher padding has been found. This is a countermeasure
4872 against active attacks where the attacker has to distinguish
4873 between bad padding and a MAC verification error. (CVE-2003-0078)
4875 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4876 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4877 Martin Vuagnoux (EPFL, Ilion)]
4879 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4881 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4882 memory from it's contents. This is done with a counter that will
4883 place alternating values in each byte. This can be used to solve
4884 two issues: 1) the removal of calls to memset() by highly optimizing
4885 compilers, and 2) cleansing with other values than 0, since those can
4886 be read through on certain media, for example a swap space on disk.
4889 *) Bugfix: client side session caching did not work with external caching,
4890 because the session->cipher setting was not restored when reloading
4891 from the external cache. This problem was masked, when
4892 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4893 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4896 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4897 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4898 [Zeev Lieber <zeev-l@yahoo.com>]
4900 *) Undo an undocumented change introduced in 0.9.6e which caused
4901 repeated calls to OpenSSL_add_all_ciphers() and
4902 OpenSSL_add_all_digests() to be ignored, even after calling
4906 *) Change the default configuration reader to deal with last line not
4907 being properly terminated.
4910 *) Change X509_NAME_cmp() so it applies the special rules on handling
4911 DN values that are of type PrintableString, as well as RDNs of type
4912 emailAddress where the value has the type ia5String.
4913 [stefank@valicert.com via Richard Levitte]
4915 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4916 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4917 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4918 the bitwise-OR of the two for use by the majority of applications
4919 wanting this behaviour, and update the docs. The documented
4920 behaviour and actual behaviour were inconsistent and had been
4921 changing anyway, so this is more a bug-fix than a behavioural
4923 [Geoff Thorpe, diagnosed by Nadav Har'El]
4925 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4926 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4929 *) Fix initialization code race conditions in
4930 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4931 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4932 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4933 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4934 ssl2_get_cipher_by_char(),
4935 ssl3_get_cipher_by_char().
4936 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4938 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4939 the cached sessions are flushed, as the remove_cb() might use ex_data
4940 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4941 (see [openssl.org #212]).
4942 [Geoff Thorpe, Lutz Jaenicke]
4944 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4945 length, instead of the encoding length to d2i_ASN1_OBJECT.
4948 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4950 *) [In 0.9.6g-engine release:]
4951 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4952 [Lynn Gazis <lgazis@rainbow.com>]
4954 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4956 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4957 and get fix the header length calculation.
4958 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4959 Alon Kantor <alonk@checkpoint.com> (and others),
4962 *) Use proper error handling instead of 'assertions' in buffer
4963 overflow checks added in 0.9.6e. This prevents DoS (the
4964 assertions could call abort()).
4965 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4967 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4969 *) Add various sanity checks to asn1_get_length() to reject
4970 the ASN1 length bytes if they exceed sizeof(long), will appear
4971 negative or the content length exceeds the length of the
4973 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4975 *) Fix cipher selection routines: ciphers without encryption had no flags
4976 for the cipher strength set and where therefore not handled correctly
4977 by the selection routines (PR #130).
4980 *) Fix EVP_dsa_sha macro.
4984 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4985 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4986 that was added in OpenSSL 0.9.6d.
4988 As the countermeasure turned out to be incompatible with some
4989 broken SSL implementations, the new option is part of SSL_OP_ALL.
4990 SSL_OP_ALL is usually employed when compatibility with weird SSL
4991 implementations is desired (e.g. '-bugs' option to 's_client' and
4992 's_server'), so the new option is automatically set in many
4996 *) Changes in security patch:
4998 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4999 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5000 Air Force Materiel Command, USAF, under agreement number
5003 *) Add various sanity checks to asn1_get_length() to reject
5004 the ASN1 length bytes if they exceed sizeof(long), will appear
5005 negative or the content length exceeds the length of the
5006 supplied buffer. (CVE-2002-0659)
5007 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5009 *) Assertions for various potential buffer overflows, not known to
5011 [Ben Laurie (CHATS)]
5013 *) Various temporary buffers to hold ASCII versions of integers were
5014 too small for 64 bit platforms. (CVE-2002-0655)
5015 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5017 *) Remote buffer overflow in SSL3 protocol - an attacker could
5018 supply an oversized session ID to a client. (CVE-2002-0656)
5019 [Ben Laurie (CHATS)]
5021 *) Remote buffer overflow in SSL2 protocol - an attacker could
5022 supply an oversized client master key. (CVE-2002-0656)
5023 [Ben Laurie (CHATS)]
5025 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5027 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5028 encoded as NULL) with id-dsa-with-sha1.
5029 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5031 *) Check various X509_...() return values in apps/req.c.
5032 [Nils Larsch <nla@trustcenter.de>]
5034 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5035 an end-of-file condition would erronously be flagged, when the CRLF
5036 was just at the end of a processed block. The bug was discovered when
5037 processing data through a buffering memory BIO handing the data to a
5038 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5039 <ptsekov@syntrex.com> and Nedelcho Stanev.
5042 *) Implement a countermeasure against a vulnerability recently found
5043 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5044 before application data chunks to avoid the use of known IVs
5045 with data potentially chosen by the attacker.
5048 *) Fix length checks in ssl3_get_client_hello().
5051 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5052 to prevent ssl3_read_internal() from incorrectly assuming that
5053 ssl3_read_bytes() found application data while handshake
5054 processing was enabled when in fact s->s3->in_read_app_data was
5055 merely automatically cleared during the initial handshake.
5056 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5058 *) Fix object definitions for Private and Enterprise: they were not
5059 recognized in their shortname (=lowercase) representation. Extend
5060 obj_dat.pl to issue an error when using undefined keywords instead
5061 of silently ignoring the problem (Svenning Sorensen
5062 <sss@sss.dnsalias.net>).
5065 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5066 generators, i.e. generators other than 2 and 5. (Previously, the
5067 code did not properly initialise the 'add' and 'rem' values to
5068 BN_generate_prime().)
5070 In the new general case, we do not insist that 'generator' is
5071 actually a primitive root: This requirement is rather pointless;
5072 a generator of the order-q subgroup is just as good, if not
5076 *) Map new X509 verification errors to alerts. Discovered and submitted by
5077 Tom Wu <tom@arcot.com>.
5080 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5081 returning non-zero before the data has been completely received
5082 when using non-blocking I/O.
5083 [Bodo Moeller; problem pointed out by John Hughes]
5085 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5086 [Ben Laurie, Lutz Jaenicke]
5088 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5089 Yoram Zahavi <YoramZ@gilian.com>).
5092 *) Add information about CygWin 1.3 and on, and preserve proper
5093 configuration for the versions before that.
5094 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5096 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5097 check whether we deal with a copy of a session and do not delete from
5098 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5099 <izhar@checkpoint.com>.
5102 *) Do not store session data into the internal session cache, if it
5103 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5104 flag is set). Proposed by Aslam <aslam@funk.com>.
5107 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5111 *) [In 0.9.6d-engine release:]
5112 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5113 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5115 *) Add the configuration target linux-s390x.
5116 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5118 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5119 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5120 variable as an indication that a ClientHello message has been
5121 received. As the flag value will be lost between multiple
5122 invocations of ssl3_accept when using non-blocking I/O, the
5123 function may not be aware that a handshake has actually taken
5124 place, thus preventing a new session from being added to the
5127 To avoid this problem, we now set s->new_session to 2 instead of
5128 using a local variable.
5129 [Lutz Jaenicke, Bodo Moeller]
5131 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5132 if the SSL_R_LENGTH_MISMATCH error is detected.
5133 [Geoff Thorpe, Bodo Moeller]
5135 *) New 'shared_ldflag' column in Configure platform table.
5138 *) Fix EVP_CIPHER_mode macro.
5139 ["Dan S. Camper" <dan@bti.net>]
5141 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5142 type, we must throw them away by setting rr->length to 0.
5143 [D P Chang <dpc@qualys.com>]
5145 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5147 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5148 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5149 worked incorrectly for those cases where range = 10..._2 and
5150 3*range is two bits longer than range.)
5153 *) Only add signing time to PKCS7 structures if it is not already
5157 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5158 OBJ_ld_ce should be OBJ_id_ce.
5159 Also some ip-pda OIDs in crypto/objects/objects.txt were
5160 incorrect (cf. RFC 3039).
5161 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5163 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5164 returns early because it has nothing to do.
5165 [Andy Schneider <andy.schneider@bjss.co.uk>]
5167 *) [In 0.9.6c-engine release:]
5168 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5169 [Andy Schneider <andy.schneider@bjss.co.uk>]
5171 *) [In 0.9.6c-engine release:]
5172 Add support for Cryptographic Appliance's keyserver technology.
5173 (Use engine 'keyclient')
5174 [Cryptographic Appliances and Geoff Thorpe]
5176 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5177 is called via tools/c89.sh because arguments have to be
5178 rearranged (all '-L' options must appear before the first object
5180 [Richard Shapiro <rshapiro@abinitio.com>]
5182 *) [In 0.9.6c-engine release:]
5183 Add support for Broadcom crypto accelerator cards, backported
5185 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5187 *) [In 0.9.6c-engine release:]
5188 Add support for SureWare crypto accelerator cards from
5189 Baltimore Technologies. (Use engine 'sureware')
5190 [Baltimore Technologies and Mark Cox]
5192 *) [In 0.9.6c-engine release:]
5193 Add support for crypto accelerator cards from Accelerated
5194 Encryption Processing, www.aep.ie. (Use engine 'aep')
5195 [AEP Inc. and Mark Cox]
5197 *) Add a configuration entry for gcc on UnixWare.
5198 [Gary Benson <gbenson@redhat.com>]
5200 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5201 messages are stored in a single piece (fixed-length part and
5202 variable-length part combined) and fix various bugs found on the way.
5205 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5206 instead. BIO_gethostbyname() does not know what timeouts are
5207 appropriate, so entries would stay in cache even when they have
5209 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5211 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5212 faced with a pathologically small ClientHello fragment that does
5213 not contain client_version: Instead of aborting with an error,
5214 simply choose the highest available protocol version (i.e.,
5215 TLS 1.0 unless it is disabled). In practice, ClientHello
5216 messages are never sent like this, but this change gives us
5217 strictly correct behaviour at least for TLS.
5220 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5221 never resets s->method to s->ctx->method when called from within
5222 one of the SSL handshake functions.
5223 [Bodo Moeller; problem pointed out by Niko Baric]
5225 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5226 (sent using the client's version number) if client_version is
5227 smaller than the protocol version in use. Also change
5228 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5229 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5230 the client will at least see that alert.
5233 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5237 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5238 client receives HelloRequest while in a handshake.
5239 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5241 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5242 should end in 'break', not 'goto end' which circuments various
5243 cleanups done in state SSL_ST_OK. But session related stuff
5244 must be disabled for SSL_ST_OK in the case that we just sent a
5247 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5248 before just sending a HelloRequest.
5249 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5251 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5252 reveal whether illegal block cipher padding was found or a MAC
5253 verification error occured. (Neither SSLerr() codes nor alerts
5254 are directly visible to potential attackers, but the information
5255 may leak via logfiles.)
5257 Similar changes are not required for the SSL 2.0 implementation
5258 because the number of padding bytes is sent in clear for SSL 2.0,
5259 and the extra bytes are just ignored. However ssl/s2_pkt.c
5260 failed to verify that the purported number of padding bytes is in
5264 *) Add OpenUNIX-8 support including shared libraries
5265 (Boyd Lynn Gerber <gerberb@zenez.com>).
5268 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5269 'wristwatch attack' using huge encoding parameters (cf.
5270 James H. Manger's CRYPTO 2001 paper). Note that the
5271 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5272 encoding parameters and hence was not vulnerable.
5275 *) BN_sqr() bug fix.
5276 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5278 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5279 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5280 followed by modular reduction.
5281 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5283 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5284 equivalent based on BN_pseudo_rand() instead of BN_rand().
5287 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5288 This function was broken, as the check for a new client hello message
5289 to handle SGC did not allow these large messages.
5290 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5293 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5296 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5297 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5300 *) Rework the configuration and shared library support for Tru64 Unix.
5301 The configuration part makes use of modern compiler features and
5302 still retains old compiler behavior for those that run older versions
5303 of the OS. The shared library support part includes a variant that
5304 uses the RPATH feature, and is available through the special
5305 configuration target "alpha-cc-rpath", which will never be selected
5307 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5309 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5310 with the same message size as in ssl3_get_certificate_request().
5311 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5312 messages might inadvertently be reject as too long.
5313 [Petr Lampa <lampa@fee.vutbr.cz>]
5315 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5318 *) Modified SSL library such that the verify_callback that has been set
5319 specificly for an SSL object with SSL_set_verify() is actually being
5320 used. Before the change, a verify_callback set with this function was
5321 ignored and the verify_callback() set in the SSL_CTX at the time of
5322 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5323 to allow the necessary settings.
5326 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5327 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5328 done automatically (in contradiction to the requirements of the C
5329 standard). This made problems when used from OpenSSH.
5332 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5333 dh->length and always used
5335 BN_rand_range(priv_key, dh->p).
5337 BN_rand_range() is not necessary for Diffie-Hellman, and this
5338 specific range makes Diffie-Hellman unnecessarily inefficient if
5339 dh->length (recommended exponent length) is much smaller than the
5340 length of dh->p. We could use BN_rand_range() if the order of
5341 the subgroup was stored in the DH structure, but we only have
5346 BN_rand(priv_key, l, ...)
5348 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5354 RSA_eay_public_encrypt
5355 RSA_eay_private_decrypt
5356 RSA_eay_private_encrypt (signing)
5357 RSA_eay_public_decrypt (signature verification)
5359 (default implementations for RSA_public_encrypt,
5360 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5361 always reject numbers >= n.
5364 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5365 to synchronize access to 'locking_thread'. This is necessary on
5366 systems where access to 'locking_thread' (an 'unsigned long'
5367 variable) is not atomic.
5370 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5371 *before* setting the 'crypto_lock_rand' flag. The previous code had
5372 a race condition if 0 is a valid thread ID.
5373 [Travis Vitek <vitek@roguewave.com>]
5375 *) Add support for shared libraries under Irix.
5376 [Albert Chin-A-Young <china@thewrittenword.com>]
5378 *) Add configuration option to build on Linux on both big-endian and
5380 [Ralf Baechle <ralf@uni-koblenz.de>]
5382 *) Add the possibility to create shared libraries on HP-UX.
5385 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5387 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5388 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5389 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5390 PRNG state recovery was possible based on the output of
5391 one PRNG request appropriately sized to gain knowledge on
5392 'md' followed by enough consecutive 1-byte PRNG requests
5393 to traverse all of 'state'.
5395 1. When updating 'md_local' (the current thread's copy of 'md')
5396 during PRNG output generation, hash all of the previous
5397 'md_local' value, not just the half used for PRNG output.
5399 2. Make the number of bytes from 'state' included into the hash
5400 independent from the number of PRNG bytes requested.
5402 The first measure alone would be sufficient to avoid
5403 Markku-Juhani's attack. (Actually it had never occurred
5404 to me that the half of 'md_local' used for chaining was the
5405 half from which PRNG output bytes were taken -- I had always
5406 assumed that the secret half would be used.) The second
5407 measure makes sure that additional data from 'state' is never
5408 mixed into 'md_local' in small portions; this heuristically
5409 further strengthens the PRNG.
5412 *) Fix crypto/bn/asm/mips3.s.
5415 *) When only the key is given to "enc", the IV is undefined. Print out
5416 an error message in this case.
5419 *) Handle special case when X509_NAME is empty in X509 printing routines.
5422 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5423 positive and less than q.
5426 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5427 used: it isn't thread safe and the add_lock_callback should handle
5429 [Paul Rose <Paul.Rose@bridge.com>]
5431 *) Verify that incoming data obeys the block size in
5432 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5436 [Ulf Möller, Bodo Möller]
5438 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5439 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5440 when fixing the server behaviour for backwards-compatible 'client
5441 hello' messages. (Note that the attack is impractical against
5442 SSL 3.0 and TLS 1.0 anyway because length and version checking
5443 means that the probability of guessing a valid ciphertext is
5444 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5447 Before 0.9.5, the countermeasure (hide the error by generating a
5448 random 'decryption result') did not work properly because
5449 ERR_clear_error() was missing, meaning that SSL_get_error() would
5450 detect the supposedly ignored error.
5452 Both problems are now fixed.
5455 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5456 (previously it was 1024).
5459 *) Fix for compatibility mode trust settings: ignore trust settings
5460 unless some valid trust or reject settings are present.
5463 *) Fix for blowfish EVP: its a variable length cipher.
5466 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5467 parameters in DSA public key structures and return an error in the
5468 DSA routines if parameters are absent.
5471 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5472 in the current directory if neither $RANDFILE nor $HOME was set.
5473 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5474 caused some confusion to Windows users who haven't defined $HOME.
5475 Thus RAND_file_name() is changed again: e_os.h can define a
5476 DEFAULT_HOME, which will be used if $HOME is not set.
5477 For Windows, we use "C:"; on other platforms, we still require
5478 environment variables.
5480 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5481 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5482 having multiple threads call RAND_poll() concurrently.
5485 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5486 combination of a flag and a thread ID variable.
5487 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5488 flag), *other* threads can enter ssleay_add_bytes without obeying
5489 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5490 that they do not hold after the first thread unsets add_do_not_lock).
5493 *) Change bctest again: '-x' expressions are not available in all
5497 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5499 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5500 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5502 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5503 the default extension for executables, if any. Also, make the perl
5504 scripts that use symlink() to test if it really exists and use "cp"
5505 if it doesn't. All this made OpenSSL compilable and installable in
5509 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5510 If SEQUENCE is length is indefinite just set c->slen to the total
5511 amount of data available.
5512 [Steve Henson, reported by shige@FreeBSD.org]
5513 [This change does not apply to 0.9.7.]
5515 *) Change bctest to avoid here-documents inside command substitution
5516 (workaround for FreeBSD /bin/sh bug).
5517 For compatibility with Ultrix, avoid shell functions (introduced
5518 in the bctest version that searches along $PATH).
5521 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5522 with des_encrypt() defined on some operating systems, like Solaris
5526 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5527 On the Importance of Eliminating Errors in Cryptographic
5528 Computations, J. Cryptology 14 (2001) 2, 101-119,
5529 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5532 *) MIPS assembler BIGNUM division bug fix.
5535 *) Disabled incorrect Alpha assembler code.
5538 *) Fix PKCS#7 decode routines so they correctly update the length
5539 after reading an EOC for the EXPLICIT tag.
5541 [This change does not apply to 0.9.7.]
5543 *) Fix bug in PKCS#12 key generation routines. This was triggered
5544 if a 3DES key was generated with a 0 initial byte. Include
5545 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5546 (but broken) behaviour.
5549 *) Enhance bctest to search for a working bc along $PATH and print
5551 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5553 *) Fix memory leaks in err.c: free err_data string if necessary;
5554 don't write to the wrong index in ERR_set_error_data.
5557 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5561 *) Replace rdtsc with _emit statements for VC++ version 5.
5562 [Jeremy Cooper <jeremy@baymoo.org>]
5564 *) Make it possible to reuse SSLv2 sessions.
5567 *) In copy_email() check for >= 0 as a return value for
5568 X509_NAME_get_index_by_NID() since 0 is a valid index.
5569 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5571 *) Avoid coredump with unsupported or invalid public keys by checking if
5572 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5573 PKCS7_verify() fails with non detached data.
5576 *) Don't use getenv in library functions when run as setuid/setgid.
5577 New function OPENSSL_issetugid().
5580 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5581 due to incorrect handling of multi-threading:
5583 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5585 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5587 3. Count how many times MemCheck_off() has been called so that
5588 nested use can be treated correctly. This also avoids
5589 inband-signalling in the previous code (which relied on the
5590 assumption that thread ID 0 is impossible).
5593 *) Add "-rand" option also to s_client and s_server.
5596 *) Fix CPU detection on Irix 6.x.
5597 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5598 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5600 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5603 [This change does not apply to 0.9.7.]
5605 *) Use the cached encoding of an X509_NAME structure rather than
5606 copying it. This is apparently the reason for the libsafe "errors"
5607 but the code is actually correct.
5610 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5611 Bleichenbacher's DSA attack.
5612 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5613 to be set and top=0 forces the highest bit to be set; top=-1 is new
5614 and leaves the highest bit random.
5615 [Ulf Moeller, Bodo Moeller]
5617 *) In the NCONF_...-based implementations for CONF_... queries
5618 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5619 a temporary CONF structure with the data component set to NULL
5620 (which gives segmentation faults in lh_retrieve).
5621 Instead, use NULL for the CONF pointer in CONF_get_string and
5622 CONF_get_number (which may use environment variables) and directly
5623 return NULL from CONF_get_section.
5626 *) Fix potential buffer overrun for EBCDIC.
5629 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5630 keyUsage if basicConstraints absent for a CA.
5633 *) Make SMIME_write_PKCS7() write mail header values with a format that
5634 is more generally accepted (no spaces before the semicolon), since
5635 some programs can't parse those values properly otherwise. Also make
5636 sure BIO's that break lines after each write do not create invalid
5640 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5641 macros previously used would not encode an empty SEQUENCE OF
5642 and break the signature.
5644 [This change does not apply to 0.9.7.]
5646 *) Zero the premaster secret after deriving the master secret in
5650 *) Add some EVP_add_digest_alias registrations (as found in
5651 OpenSSL_add_all_digests()) to SSL_library_init()
5652 aka OpenSSL_add_ssl_algorithms(). This provides improved
5653 compatibility with peers using X.509 certificates
5654 with unconventional AlgorithmIdentifier OIDs.
5657 *) Fix for Irix with NO_ASM.
5658 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5660 *) ./config script fixes.
5661 [Ulf Moeller, Richard Levitte]
5663 *) Fix 'openssl passwd -1'.
5666 *) Change PKCS12_key_gen_asc() so it can cope with non null
5667 terminated strings whose length is passed in the passlen
5668 parameter, for example from PEM callbacks. This was done
5669 by adding an extra length parameter to asc2uni().
5670 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5672 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5673 call failed, free the DSA structure.
5676 *) Fix to uni2asc() to cope with zero length Unicode strings.
5677 These are present in some PKCS#12 files.
5680 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5681 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5682 when writing a 32767 byte record.
5683 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5685 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5686 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5688 (RSA objects have a reference count access to which is protected
5689 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5690 so they are meant to be shared between threads.)
5691 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5692 "Reddie, Steven" <Steven.Reddie@ca.com>]
5694 *) Fix a deadlock in CRYPTO_mem_leaks().
5697 *) Use better test patterns in bntest.
5700 *) rand_win.c fix for Borland C.
5703 *) BN_rshift bugfix for n == 0.
5706 *) Add a 'bctest' script that checks for some known 'bc' bugs
5707 so that 'make test' does not abort just because 'bc' is broken.
5710 *) Store verify_result within SSL_SESSION also for client side to
5711 avoid potential security hole. (Re-used sessions on the client side
5712 always resulted in verify_result==X509_V_OK, not using the original
5713 result of the server certificate verification.)
5716 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5717 SSL3_RT_APPLICATION_DATA, return 0.
5718 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5722 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5723 releases, have been re-implemented by renaming the previous
5724 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5725 and ssl3_read_internal, respectively, and adding 'peek' parameters
5726 to them. The new ssl[23]_{read,peek} functions are calls to
5727 ssl[23]_read_internal with the 'peek' flag set appropriately.
5728 A 'peek' parameter has also been added to ssl3_read_bytes, which
5729 does the actual work for ssl3_read_internal.
5732 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5733 the method-specific "init()" handler. Also clean up ex_data after
5734 calling the method-specific "finish()" handler. Previously, this was
5735 happening the other way round.
5738 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5739 The previous value, 12, was not always sufficient for BN_mod_exp().
5742 *) Make sure that shared libraries get the internal name engine with
5743 the full version number and not just 0. This should mark the
5744 shared libraries as not backward compatible. Of course, this should
5745 be changed again when we can guarantee backward binary compatibility.
5748 *) Fix typo in get_cert_by_subject() in by_dir.c
5749 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5751 *) Rework the system to generate shared libraries:
5753 - Make note of the expected extension for the shared libraries and
5754 if there is a need for symbolic links from for example libcrypto.so.0
5755 to libcrypto.so.0.9.7. There is extended info in Configure for
5758 - Make as few rebuilds of the shared libraries as possible.
5760 - Still avoid linking the OpenSSL programs with the shared libraries.
5762 - When installing, install the shared libraries separately from the
5766 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5768 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5769 and not in SSL_clear because the latter is also used by the
5770 accept/connect functions; previously, the settings made by
5771 SSL_set_read_ahead would be lost during the handshake.
5772 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5774 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5775 Previously, it would create entries for disableed algorithms no
5779 *) Added several new manual pages for SSL_* function.
5782 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5784 *) In ssl23_get_client_hello, generate an error message when faced
5785 with an initial SSL 3.0/TLS record that is too small to contain the
5786 first two bytes of the ClientHello message, i.e. client_version.
5787 (Note that this is a pathologic case that probably has never happened
5788 in real life.) The previous approach was to use the version number
5789 from the record header as a substitute; but our protocol choice
5790 should not depend on that one because it is not authenticated
5791 by the Finished messages.
5794 *) More robust randomness gathering functions for Windows.
5795 [Jeffrey Altman <jaltman@columbia.edu>]
5797 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5798 not set then we don't setup the error code for issuer check errors
5799 to avoid possibly overwriting other errors which the callback does
5800 handle. If an application does set the flag then we assume it knows
5801 what it is doing and can handle the new informational codes
5805 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5806 a general "ANY" type, as such it should be able to decode anything
5807 including tagged types. However it didn't check the class so it would
5808 wrongly interpret tagged types in the same way as their universal
5809 counterpart and unknown types were just rejected. Changed so that the
5810 tagged and unknown types are handled in the same way as a SEQUENCE:
5811 that is the encoding is stored intact. There is also a new type
5812 "V_ASN1_OTHER" which is used when the class is not universal, in this
5813 case we have no idea what the actual type is so we just lump them all
5817 *) On VMS, stdout may very well lead to a file that is written to
5818 in a record-oriented fashion. That means that every write() will
5819 write a separate record, which will be read separately by the
5820 programs trying to read from it. This can be very confusing.
5822 The solution is to put a BIO filter in the way that will buffer
5823 text until a linefeed is reached, and then write everything a
5824 line at a time, so every record written will be an actual line,
5825 not chunks of lines and not (usually doesn't happen, but I've
5826 seen it once) several lines in one record. BIO_f_linebuffer() is
5829 Currently, it's a VMS-only method, because that's where it has
5830 been tested well enough.
5833 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5834 it can return incorrect results.
5835 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5836 but it was in 0.9.6-beta[12].)
5839 *) Disable the check for content being present when verifying detached
5840 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5841 include zero length content when signing messages.
5844 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5845 BIO_ctrl (for BIO pairs).
5848 *) Add DSO method for VMS.
5851 *) Bug fix: Montgomery multiplication could produce results with the
5855 *) Add RPM specification openssl.spec and modify it to build three
5856 packages. The default package contains applications, application
5857 documentation and run-time libraries. The devel package contains
5858 include files, static libraries and function documentation. The
5859 doc package contains the contents of the doc directory. The original
5860 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5863 *) Add a large number of documentation files for many SSL routines.
5864 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5866 *) Add a configuration entry for Sony News 4.
5867 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5869 *) Don't set the two most significant bits to one when generating a
5870 random number < q in the DSA library.
5873 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5874 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5875 the underlying transport is blocking) if a handshake took place.
5876 (The default behaviour is needed by applications such as s_client
5877 and s_server that use select() to determine when to use SSL_read;
5878 but for applications that know in advance when to expect data, it
5879 just makes things more complicated.)
5882 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5886 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5887 work better on such systems.
5888 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5890 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5891 Update PKCS12_parse() so it copies the friendlyName and the
5892 keyid to the certificates aux info.
5895 *) Fix bug in PKCS7_verify() which caused an infinite loop
5896 if there was more than one signature.
5897 [Sven Uszpelkat <su@celocom.de>]
5899 *) Major change in util/mkdef.pl to include extra information
5900 about each symbol, as well as presentig variables as well
5901 as functions. This change means that there's n more need
5902 to rebuild the .num files when some algorithms are excluded.
5905 *) Allow the verify time to be set by an application,
5906 rather than always using the current time.
5909 *) Phase 2 verify code reorganisation. The certificate
5910 verify code now looks up an issuer certificate by a
5911 number of criteria: subject name, authority key id
5912 and key usage. It also verifies self signed certificates
5913 by the same criteria. The main comparison function is
5914 X509_check_issued() which performs these checks.
5916 Lot of changes were necessary in order to support this
5917 without completely rewriting the lookup code.
5919 Authority and subject key identifier are now cached.
5921 The LHASH 'certs' is X509_STORE has now been replaced
5922 by a STACK_OF(X509_OBJECT). This is mainly because an
5923 LHASH can't store or retrieve multiple objects with
5924 the same hash value.
5926 As a result various functions (which were all internal
5927 use only) have changed to handle the new X509_STORE
5928 structure. This will break anything that messed round
5929 with X509_STORE internally.
5931 The functions X509_STORE_add_cert() now checks for an
5932 exact match, rather than just subject name.
5934 The X509_STORE API doesn't directly support the retrieval
5935 of multiple certificates matching a given criteria, however
5936 this can be worked round by performing a lookup first
5937 (which will fill the cache with candidate certificates)
5938 and then examining the cache for matches. This is probably
5939 the best we can do without throwing out X509_LOOKUP
5940 entirely (maybe later...).
5942 The X509_VERIFY_CTX structure has been enhanced considerably.
5944 All certificate lookup operations now go via a get_issuer()
5945 callback. Although this currently uses an X509_STORE it
5946 can be replaced by custom lookups. This is a simple way
5947 to bypass the X509_STORE hackery necessary to make this
5948 work and makes it possible to use more efficient techniques
5949 in future. A very simple version which uses a simple
5950 STACK for its trusted certificate store is also provided
5951 using X509_STORE_CTX_trusted_stack().
5953 The verify_cb() and verify() callbacks now have equivalents
5954 in the X509_STORE_CTX structure.
5956 X509_STORE_CTX also has a 'flags' field which can be used
5957 to customise the verify behaviour.
5960 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5961 excludes S/MIME capabilities.
5964 *) When a certificate request is read in keep a copy of the
5965 original encoding of the signed data and use it when outputing
5966 again. Signatures then use the original encoding rather than
5967 a decoded, encoded version which may cause problems if the
5968 request is improperly encoded.
5971 *) For consistency with other BIO_puts implementations, call
5972 buffer_write(b, ...) directly in buffer_puts instead of calling
5975 In BIO_puts, increment b->num_write as in BIO_write.
5976 [Peter.Sylvester@EdelWeb.fr]
5978 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5979 BN_zero, we may not return a BIGNUM with an array consisting of
5983 *) Avoid calling abort() from within the library when problems are
5984 detected, except if preprocessor symbols have been defined
5985 (such as REF_CHECK, BN_DEBUG etc.).
5988 *) New openssl application 'rsautl'. This utility can be
5989 used for low level RSA operations. DER public key
5990 BIO/fp routines also added.
5993 *) New Configure entry and patches for compiling on QNX 4.
5994 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5996 *) A demo state-machine implementation was sponsored by
5997 Nuron (http://www.nuron.com/) and is now available in
5998 demos/state_machine.
6001 *) New options added to the 'dgst' utility for signature
6002 generation and verification.
6005 *) Unrecognized PKCS#7 content types are now handled via a
6006 catch all ASN1_TYPE structure. This allows unsupported
6007 types to be stored as a "blob" and an application can
6008 encode and decode it manually.
6011 *) Fix various signed/unsigned issues to make a_strex.c
6013 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6015 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6016 length if passed a buffer. ASN1_INTEGER_to_BN failed
6017 if passed a NULL BN and its argument was negative.
6018 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6020 *) Modification to PKCS#7 encoding routines to output definite
6021 length encoding. Since currently the whole structures are in
6022 memory there's not real point in using indefinite length
6023 constructed encoding. However if OpenSSL is compiled with
6024 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6027 *) Added BIO_vprintf() and BIO_vsnprintf().
6030 *) Added more prefixes to parse for in the the strings written
6031 through a logging bio, to cover all the levels that are available
6032 through syslog. The prefixes are now:
6034 PANIC, EMERG, EMR => LOG_EMERG
6035 ALERT, ALR => LOG_ALERT
6036 CRIT, CRI => LOG_CRIT
6037 ERROR, ERR => LOG_ERR
6038 WARNING, WARN, WAR => LOG_WARNING
6039 NOTICE, NOTE, NOT => LOG_NOTICE
6040 INFO, INF => LOG_INFO
6041 DEBUG, DBG => LOG_DEBUG
6043 and as before, if none of those prefixes are present at the
6044 beginning of the string, LOG_ERR is chosen.
6046 On Win32, the LOG_* levels are mapped according to this:
6048 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6049 LOG_WARNING => EVENTLOG_WARNING_TYPE
6050 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6054 *) Made it possible to reconfigure with just the configuration
6055 argument "reconf" or "reconfigure". The command line arguments
6056 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6057 and are retrieved from there when reconfiguring.
6061 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6063 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6066 *) The obj_dat.pl script was messing up the sorting of object
6067 names. The reason was that it compared the quoted version
6068 of strings as a result "OCSP" > "OCSP Signing" because
6069 " > SPACE. Changed script to store unquoted versions of
6070 names and add quotes on output. It was also omitting some
6071 names from the lookup table if they were given a default
6072 value (that is if SN is missing it is given the same
6073 value as LN and vice versa), these are now added on the
6074 grounds that if an object has a name we should be able to
6075 look it up. Finally added warning output when duplicate
6076 short or long names are found.
6079 *) Changes needed for Tandem NSK.
6080 [Scott Uroff <scott@xypro.com>]
6082 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6083 RSA_padding_check_SSLv23(), special padding was never detected
6084 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6085 version rollback attacks was not effective.
6087 In s23_clnt.c, don't use special rollback-attack detection padding
6088 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6089 client; similarly, in s23_srvr.c, don't do the rollback check if
6090 SSL 2.0 is the only protocol enabled in the server.
6093 *) Make it possible to get hexdumps of unprintable data with 'openssl
6094 asn1parse'. By implication, the functions ASN1_parse_dump() and
6095 BIO_dump_indent() are added.
6098 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6099 these print out strings and name structures based on various
6100 flags including RFC2253 support and proper handling of
6101 multibyte characters. Added options to the 'x509' utility
6102 to allow the various flags to be set.
6105 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6106 Also change the functions X509_cmp_current_time() and
6107 X509_gmtime_adj() work with an ASN1_TIME structure,
6108 this will enable certificates using GeneralizedTime in validity
6109 dates to be checked.
6112 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6113 negative public key encodings) on by default,
6114 NO_NEG_PUBKEY_BUG can be set to disable it.
6117 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6118 content octets. An i2c_ASN1_OBJECT is unnecessary because
6119 the encoding can be trivially obtained from the structure.
6122 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6123 not read locks (CRYPTO_r_[un]lock).
6126 *) A first attempt at creating official support for shared
6127 libraries through configuration. I've kept it so the
6128 default is static libraries only, and the OpenSSL programs
6129 are always statically linked for now, but there are
6130 preparations for dynamic linking in place.
6131 This has been tested on Linux and Tru64.
6134 *) Randomness polling function for Win9x, as described in:
6135 Peter Gutmann, Software Generation of Practically Strong
6139 *) Fix so PRNG is seeded in req if using an already existing
6143 *) New options to smime application. -inform and -outform
6144 allow alternative formats for the S/MIME message including
6145 PEM and DER. The -content option allows the content to be
6146 specified separately. This should allow things like Netscape
6147 form signing output easier to verify.
6150 *) Fix the ASN1 encoding of tags using the 'long form'.
6153 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6154 STRING types. These convert content octets to and from the
6155 underlying type. The actual tag and length octets are
6156 already assumed to have been read in and checked. These
6157 are needed because all other string types have virtually
6158 identical handling apart from the tag. By having versions
6159 of the ASN1 functions that just operate on content octets
6160 IMPLICIT tagging can be handled properly. It also allows
6161 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6162 and ASN1_INTEGER are identical apart from the tag.
6165 *) Change the handling of OID objects as follows:
6167 - New object identifiers are inserted in objects.txt, following
6168 the syntax given in objects.README.
6169 - objects.pl is used to process obj_mac.num and create a new
6171 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6174 This is currently kind of a hack, and the perl code in objects.pl
6175 isn't very elegant, but it works as I intended. The simplest way
6176 to check that it worked correctly is to look in obj_dat.h and
6177 check the array nid_objs and make sure the objects haven't moved
6178 around (this is important!). Additions are OK, as well as
6179 consistent name changes.
6182 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6185 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6186 The given file adds to whatever has already been seeded into the
6187 random pool through the RANDFILE configuration file option or
6188 environment variable, or the default random state file.
6191 *) mkstack.pl now sorts each macro group into lexical order.
6192 Previously the output order depended on the order the files
6193 appeared in the directory, resulting in needless rewriting
6197 *) Patches to make OpenSSL compile under Win32 again. Mostly
6198 work arounds for the VC++ problem that it treats func() as
6199 func(void). Also stripped out the parts of mkdef.pl that
6200 added extra typesafe functions: these no longer exist.
6203 *) Reorganisation of the stack code. The macros are now all
6204 collected in safestack.h . Each macro is defined in terms of
6205 a "stack macro" of the form SKM_<name>(type, a, b). The
6206 DEBUG_SAFESTACK is now handled in terms of function casts,
6207 this has the advantage of retaining type safety without the
6208 use of additional functions. If DEBUG_SAFESTACK is not defined
6209 then the non typesafe macros are used instead. Also modified the
6210 mkstack.pl script to handle the new form. Needs testing to see
6211 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6212 the default if no major problems. Similar behaviour for ASN1_SET_OF
6213 and PKCS12_STACK_OF.
6216 *) When some versions of IIS use the 'NET' form of private key the
6217 key derivation algorithm is different. Normally MD5(password) is
6218 used as a 128 bit RC4 key. In the modified case
6219 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6220 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6221 as the old Netscape_RSA functions except they have an additional
6222 'sgckey' parameter which uses the modified algorithm. Also added
6223 an -sgckey command line option to the rsa utility. Thanks to
6224 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6225 algorithm to openssl-dev.
6228 *) The evp_local.h macros were using 'c.##kname' which resulted in
6229 invalid expansion on some systems (SCO 5.0.5 for example).
6230 Corrected to 'c.kname'.
6231 [Phillip Porch <root@theporch.com>]
6233 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6234 a STACK of email addresses from a certificate or request, these look
6235 in the subject name and the subject alternative name extensions and
6236 omit any duplicate addresses.
6239 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6240 This makes DSA verification about 2 % faster.
6243 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6244 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6245 plus overhead for 1024 bit moduli).
6246 This makes exponentiations about 0.5 % faster for 1024 bit
6247 exponents (as measured by "openssl speed rsa2048").
6250 *) Rename memory handling macros to avoid conflicts with other
6252 Malloc => OPENSSL_malloc
6253 Malloc_locked => OPENSSL_malloc_locked
6254 Realloc => OPENSSL_realloc
6255 Free => OPENSSL_free
6258 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6259 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6262 *) CygWin32 support.
6263 [John Jarvie <jjarvie@newsguy.com>]
6265 *) The type-safe stack code has been rejigged. It is now only compiled
6266 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6267 by default all type-specific stack functions are "#define"d back to
6268 standard stack functions. This results in more streamlined output
6269 but retains the type-safety checking possibilities of the original
6273 *) The STACK code has been cleaned up, and certain type declarations
6274 that didn't make a lot of sense have been brought in line. This has
6275 also involved a cleanup of sorts in safestack.h to more correctly
6276 map type-safe stack functions onto their plain stack counterparts.
6277 This work has also resulted in a variety of "const"ifications of
6278 lots of the code, especially "_cmp" operations which should normally
6279 be prototyped with "const" parameters anyway.
6282 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6283 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6284 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6285 where all of 'md' is used each time the PRNG is used, but 'state'
6286 is used only indexed by a cyclic counter. As entropy may not be
6287 well distributed from the beginning, 'md' is important as a
6288 chaining variable. However, the output function chains only half
6289 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6290 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6291 in all of 'state' being rewritten, with the new values depending
6292 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6295 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6296 the handshake is continued after ssl_verify_cert_chain();
6297 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6298 can lead to 'unexplainable' connection aborts later.
6299 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6301 *) Major EVP API cipher revision.
6302 Add hooks for extra EVP features. This allows various cipher
6303 parameters to be set in the EVP interface. Support added for variable
6304 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6305 setting of RC2 and RC5 parameters.
6307 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6310 Remove lots of duplicated code from the EVP library. For example *every*
6311 cipher init() function handles the 'iv' in the same way according to the
6312 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6313 for CFB and OFB modes they zero ctx->num.
6315 New functionality allows removal of S/MIME code RC2 hack.
6317 Most of the routines have the same form and so can be declared in terms
6320 By shifting this to the top level EVP_CipherInit() it can be removed from
6321 all individual ciphers. If the cipher wants to handle IVs or keys
6322 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6325 Change lots of functions like EVP_EncryptUpdate() to now return a
6326 value: although software versions of the algorithms cannot fail
6327 any installed hardware versions can.
6330 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6331 this option is set, tolerate broken clients that send the negotiated
6332 protocol version number instead of the requested protocol version
6336 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6337 i.e. non-zero for export ciphersuites, zero otherwise.
6338 Previous versions had this flag inverted, inconsistent with
6339 rsa_tmp_cb (..._TMP_RSA_CB).
6340 [Bodo Moeller; problem reported by Amit Chopra]
6342 *) Add missing DSA library text string. Work around for some IIS
6343 key files with invalid SEQUENCE encoding.
6346 *) Add a document (doc/standards.txt) that list all kinds of standards
6347 and so on that are implemented in OpenSSL.
6350 *) Enhance c_rehash script. Old version would mishandle certificates
6351 with the same subject name hash and wouldn't handle CRLs at all.
6352 Added -fingerprint option to crl utility, to support new c_rehash
6356 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6359 *) Fix for SSL server purpose checking. Server checking was
6360 rejecting certificates which had extended key usage present
6361 but no ssl client purpose.
6362 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6364 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6365 is a little unclear about how a blank password is handled.
6366 Since the password in encoded as a BMPString with terminating
6367 double NULL a zero length password would end up as just the
6368 double NULL. However no password at all is different and is
6369 handled differently in the PKCS#12 key generation code. NS
6370 treats a blank password as zero length. MSIE treats it as no
6371 password on export: but it will try both on import. We now do
6372 the same: PKCS12_parse() tries zero length and no password if
6373 the password is set to "" or NULL (NULL is now a valid password:
6374 it wasn't before) as does the pkcs12 application.
6377 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6378 perror when PEM_read_bio_X509_REQ fails, the error message must
6379 be obtained from the error queue.
6382 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6383 it in ERR_remove_state if appropriate, and change ERR_get_state
6384 accordingly to avoid race conditions (this is necessary because
6385 thread_hash is no longer constant once set).
6388 *) Bugfix for linux-elf makefile.one.
6391 *) RSA_get_default_method() will now cause a default
6392 RSA_METHOD to be chosen if one doesn't exist already.
6393 Previously this was only set during a call to RSA_new()
6394 or RSA_new_method(NULL) meaning it was possible for
6395 RSA_get_default_method() to return NULL.
6398 *) Added native name translation to the existing DSO code
6399 that will convert (if the flag to do so is set) filenames
6400 that are sufficiently small and have no path information
6401 into a canonical native form. Eg. "blah" converted to
6402 "libblah.so" or "blah.dll" etc.
6405 *) New function ERR_error_string_n(e, buf, len) which is like
6406 ERR_error_string(e, buf), but writes at most 'len' bytes
6407 including the 0 terminator. For ERR_error_string_n, 'buf'
6409 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6411 *) CONF library reworked to become more general. A new CONF
6412 configuration file reader "class" is implemented as well as a
6413 new functions (NCONF_*, for "New CONF") to handle it. The now
6414 old CONF_* functions are still there, but are reimplemented to
6415 work in terms of the new functions. Also, a set of functions
6416 to handle the internal storage of the configuration data is
6417 provided to make it easier to write new configuration file
6418 reader "classes" (I can definitely see something reading a
6419 configuration file in XML format, for example), called _CONF_*,
6420 or "the configuration storage API"...
6422 The new configuration file reading functions are:
6424 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6425 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6427 NCONF_default, NCONF_WIN32
6429 NCONF_dump_fp, NCONF_dump_bio
6431 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6432 NCONF_new creates a new CONF object. This works in the same way
6433 as other interfaces in OpenSSL, like the BIO interface.
6434 NCONF_dump_* dump the internal storage of the configuration file,
6435 which is useful for debugging. All other functions take the same
6436 arguments as the old CONF_* functions wth the exception of the
6437 first that must be a `CONF *' instead of a `LHASH *'.
6439 To make it easer to use the new classes with the old CONF_* functions,
6440 the function CONF_set_default_method is provided.
6443 *) Add '-tls1' option to 'openssl ciphers', which was already
6444 mentioned in the documentation but had not been implemented.
6445 (This option is not yet really useful because even the additional
6446 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6449 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6450 OpenSSL-based applications) load shared libraries and bind to
6451 them in a portable way.
6452 [Geoff Thorpe, with contributions from Richard Levitte]
6454 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6456 *) Make sure _lrotl and _lrotr are only used with MSVC.
6458 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6459 (the default implementation of RAND_status).
6461 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6462 to '-clrext' (= clear extensions), as intended and documented.
6463 [Bodo Moeller; inconsistency pointed out by Michael Attili
6466 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6467 was larger than the MD block size.
6468 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6470 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6471 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6472 using the passed key: if the passed key was a private key the result
6473 of X509_print(), for example, would be to print out all the private key
6477 *) des_quad_cksum() byte order bug fix.
6478 [Ulf Möller, using the problem description in krb4-0.9.7, where
6479 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6481 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6483 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6485 *) For easily testing in shell scripts whether some command
6486 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6487 returns with exit code 0 iff no command of the given name is available.
6488 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6489 the output goes to stdout and nothing is printed to stderr.
6490 Additional arguments are always ignored.
6492 Since for each cipher there is a command of the same name,
6493 the 'no-cipher' compilation switches can be tested this way.
6495 ('openssl no-XXX' is not able to detect pseudo-commands such
6496 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6499 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6502 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6503 is set; it will be thrown away anyway because each handshake creates
6505 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6506 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6507 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6508 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6511 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6512 'Q' and 'R' lose their special meanings (quit/renegotiate).
6513 This is part of what -quiet does; unlike -quiet, -ign_eof
6514 does not suppress any output.
6517 *) Add compatibility options to the purpose and trust code. The
6518 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6519 accepts a certificate or CA, this was the previous behaviour,
6520 with all the associated security issues.
6522 X509_TRUST_COMPAT is the old trust behaviour: only and
6523 automatically trust self signed roots in certificate store. A
6524 new trust setting X509_TRUST_DEFAULT is used to specify that
6525 a purpose has no associated trust setting and it should instead
6526 use the value in the default purpose.
6529 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6530 and fix a memory leak.
6533 *) In util/mkerr.pl (which implements 'make errors'), preserve
6534 reason strings from the previous version of the .c file, as
6535 the default to have only downcase letters (and digits) in
6536 automatically generated reasons codes is not always appropriate.
6539 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6540 using strerror. Previously, ERR_reason_error_string() returned
6541 library names as reason strings for SYSerr; but SYSerr is a special
6542 case where small numbers are errno values, not library numbers.
6545 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6546 converts DSA parameters into DH parameters. (When creating parameters,
6547 DSA_generate_parameters is used.)
6550 *) Include 'length' (recommended exponent length) in C code generated
6551 by 'openssl dhparam -C'.
6554 *) The second argument to set_label in perlasm was already being used
6555 so couldn't be used as a "file scope" flag. Moved to third argument
6559 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6560 instead of RAND_bytes for encryption IVs and salts.
6563 *) Include RAND_status() into RAND_METHOD instead of implementing
6564 it only for md_rand.c Otherwise replacing the PRNG by calling
6565 RAND_set_rand_method would be impossible.
6568 *) Don't let DSA_generate_key() enter an infinite loop if the random
6569 number generation fails.
6572 *) New 'rand' application for creating pseudo-random output.
6575 *) Added configuration support for Linux/IA64
6576 [Rolf Haberrecker <rolf@suse.de>]
6578 *) Assembler module support for Mingw32.
6581 *) Shared library support for HPUX (in shlib/).
6582 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6584 *) Shared library support for Solaris gcc.
6585 [Lutz Behnke <behnke@trustcenter.de>]
6587 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6589 *) PKCS7_encrypt() was adding text MIME headers twice because they
6590 were added manually and by SMIME_crlf_copy().
6593 *) In bntest.c don't call BN_rand with zero bits argument.
6594 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6596 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6597 case was implemented. This caused BN_div_recp() to fail occasionally.
6600 *) Add an optional second argument to the set_label() in the perl
6601 assembly language builder. If this argument exists and is set
6602 to 1 it signals that the assembler should use a symbol whose
6603 scope is the entire file, not just the current function. This
6604 is needed with MASM which uses the format label:: for this scope.
6605 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6607 *) Change the ASN1 types so they are typedefs by default. Before
6608 almost all types were #define'd to ASN1_STRING which was causing
6609 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6613 *) Change names of new functions to the new get1/get0 naming
6614 convention: After 'get1', the caller owns a reference count
6615 and has to call ..._free; 'get0' returns a pointer to some
6616 data structure without incrementing reference counters.
6617 (Some of the existing 'get' functions increment a reference
6618 counter, some don't.)
6619 Similarly, 'set1' and 'add1' functions increase reference
6620 counters or duplicate objects.
6623 *) Allow for the possibility of temp RSA key generation failure:
6624 the code used to assume it always worked and crashed on failure.
6627 *) Fix potential buffer overrun problem in BIO_printf().
6628 [Ulf Möller, using public domain code by Patrick Powell; problem
6629 pointed out by David Sacerdote <das33@cornell.edu>]
6631 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6632 RAND_egd() and RAND_status(). In the command line application,
6633 the EGD socket can be specified like a seed file using RANDFILE
6637 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6638 Some CAs (e.g. Verisign) distribute certificates in this form.
6641 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6642 list to exclude them. This means that no special compilation option
6643 is needed to use anonymous DH: it just needs to be included in the
6647 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6648 EVP_MD_type. The old functionality is available in a new macro called
6649 EVP_MD_md(). Change code that uses it and update docs.
6652 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6653 where the 'void *' argument is replaced by a function pointer argument.
6654 Previously 'void *' was abused to point to functions, which works on
6655 many platforms, but is not correct. As these functions are usually
6656 called by macros defined in OpenSSL header files, most source code
6657 should work without changes.
6660 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6661 sections with information on -D... compiler switches used for
6662 compiling the library so that applications can see them. To enable
6663 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6664 must be defined. E.g.,
6665 #define OPENSSL_ALGORITHM_DEFINES
6666 #include <openssl/opensslconf.h>
6667 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6668 [Richard Levitte, Ulf and Bodo Möller]
6670 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6674 *) Change the 'other' type in certificate aux info to a STACK_OF
6675 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6676 the required ASN1 format: arbitrary types determined by an OID.
6679 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6680 argument to 'req'. This is not because the function is newer or
6681 better than others it just uses the work 'NEW' in the certificate
6682 request header lines. Some software needs this.
6685 *) Reorganise password command line arguments: now passwords can be
6686 obtained from various sources. Delete the PEM_cb function and make
6687 it the default behaviour: i.e. if the callback is NULL and the
6688 usrdata argument is not NULL interpret it as a null terminated pass
6689 phrase. If usrdata and the callback are NULL then the pass phrase
6690 is prompted for as usual.
6693 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6694 the support is automatically enabled. The resulting binaries will
6695 autodetect the card and use it if present.
6696 [Ben Laurie and Compaq Inc.]
6698 *) Work around for Netscape hang bug. This sends certificate request
6699 and server done in one record. Since this is perfectly legal in the
6700 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6701 the bugs/SSLv3 entry for more info.
6704 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6707 *) Add -rand argument to smime and pkcs12 applications and read/write
6711 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6714 *) Add command line password options to the remaining applications.
6717 *) Bug fix for BN_div_recp() for numerators with an even number of
6721 *) More tests in bntest.c, and changed test_bn output.
6724 *) ./config recognizes MacOS X now.
6727 *) Bug fix for BN_div() when the first words of num and divsor are
6728 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6731 *) Add support for various broken PKCS#8 formats, and command line
6732 options to produce them.
6735 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6736 get temporary BIGNUMs from a BN_CTX.
6739 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6743 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6744 include a #define from the old name to the new. The original intent
6745 was that statically linked binaries could for example just call
6746 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6747 link with digests. This never worked becayse SSLeay_add_all_digests()
6748 and SSLeay_add_all_ciphers() were in the same source file so calling
6749 one would link with the other. They are now in separate source files.
6752 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6755 *) Use a less unusual form of the Miller-Rabin primality test (it used
6756 a binary algorithm for exponentiation integrated into the Miller-Rabin
6757 loop, our standard modexp algorithms are faster).
6760 *) Support for the EBCDIC character set completed.
6761 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6763 *) Source code cleanups: use const where appropriate, eliminate casts,
6764 use void * instead of char * in lhash.
6767 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6768 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6769 this the server could overwrite ephemeral keys that the client
6773 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6774 using 50 iterations of the Rabin-Miller test.
6776 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6777 iterations of the Rabin-Miller test as required by the appendix
6778 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6779 As BN_is_prime_fasttest includes trial division, DSA parameter
6780 generation becomes much faster.
6782 This implies a change for the callback functions in DSA_is_prime
6783 and DSA_generate_parameters: The callback function is called once
6784 for each positive witness in the Rabin-Miller test, not just
6785 occasionally in the inner loop; and the parameters to the
6786 callback function now provide an iteration count for the outer
6787 loop rather than for the current invocation of the inner loop.
6788 DSA_generate_parameters additionally can call the callback
6789 function with an 'iteration count' of -1, meaning that a
6790 candidate has passed the trial division test (when q is generated
6791 from an application-provided seed, trial division is skipped).
6794 *) New function BN_is_prime_fasttest that optionally does trial
6795 division before starting the Rabin-Miller test and has
6796 an additional BN_CTX * argument (whereas BN_is_prime always
6797 has to allocate at least one BN_CTX).
6798 'callback(1, -1, cb_arg)' is called when a number has passed the
6799 trial division stage.
6802 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6806 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6809 *) New function BN_pseudo_rand().
6812 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6813 bignum version of BN_from_montgomery() with the working code from
6814 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6818 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6819 made it impossible to use the same SSL_SESSION data structure in
6820 SSL2 clients in multiple threads.
6823 *) The return value of RAND_load_file() no longer counts bytes obtained
6824 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6825 to seed the PRNG (previously an explicit byte count was required).
6826 [Ulf Möller, Bodo Möller]
6828 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6829 used (char *) instead of (void *) and had casts all over the place.
6832 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6835 *) Retain source code compatibility for BN_prime_checks macro:
6836 BN_is_prime(..., BN_prime_checks, ...) now uses
6837 BN_prime_checks_for_size to determine the appropriate number of
6838 Rabin-Miller iterations.
6841 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6842 DH_CHECK_P_NOT_SAFE_PRIME.
6843 (Check if this is true? OpenPGP calls them "strong".)
6846 *) Merge the functionality of "dh" and "gendh" programs into a new program
6847 "dhparam". The old programs are retained for now but will handle DH keys
6848 (instead of parameters) in future.
6851 *) Make the ciphers, s_server and s_client programs check the return values
6852 when a new cipher list is set.
6855 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6856 ciphers. Before when the 56bit ciphers were enabled the sorting was
6859 The syntax for the cipher sorting has been extended to support sorting by
6860 cipher-strength (using the strength_bits hard coded in the tables).
6861 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6863 Fix a bug in the cipher-command parser: when supplying a cipher command
6864 string with an "undefined" symbol (neither command nor alphanumeric
6865 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6866 an error is flagged.
6868 Due to the strength-sorting extension, the code of the
6869 ssl_create_cipher_list() function was completely rearranged. I hope that
6870 the readability was also increased :-)
6871 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6873 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6874 for the first serial number and places 2 in the serial number file. This
6875 avoids problems when the root CA is created with serial number zero and
6876 the first user certificate has the same issuer name and serial number
6880 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6881 the new code. Add documentation for this stuff.
6884 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6885 X509_*() to X509at_*() on the grounds that they don't handle X509
6886 structures and behave in an analagous way to the X509v3 functions:
6887 they shouldn't be called directly but wrapper functions should be used
6890 So we also now have some wrapper functions that call the X509at functions
6891 when passed certificate requests. (TO DO: similar things can be done with
6892 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6893 things. Some of these need some d2i or i2d and print functionality
6894 because they handle more complex structures.)
6897 *) Add missing #ifndefs that caused missing symbols when building libssl
6898 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6899 NO_RSA in ssl/s2*.c.
6900 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6902 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6903 has a return value which indicates the quality of the random data
6904 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6905 error queue. New function RAND_pseudo_bytes() generates output that is
6906 guaranteed to be unique but not unpredictable. RAND_add is like
6907 RAND_seed, but takes an extra argument for an entropy estimate
6908 (RAND_seed always assumes full entropy).
6911 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6912 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6913 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6914 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6915 false-positive rate of at most 2^-80 for random input.
6918 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6921 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6922 in the 0.9.5 release), this returns the chain
6923 from an X509_CTX structure with a dup of the stack and all
6924 the X509 reference counts upped: so the stack will exist
6925 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6928 Also make SSL_SESSION_print() print out the verify return
6932 *) Add manpage for the pkcs12 command. Also change the default
6933 behaviour so MAC iteration counts are used unless the new
6934 -nomaciter option is used. This improves file security and
6935 only older versions of MSIE (4.0 for example) need it.
6938 *) Honor the no-xxx Configure options when creating .DEF files.
6941 *) Add PKCS#10 attributes to field table: challengePassword,
6942 unstructuredName and unstructuredAddress. These are taken from
6943 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6944 international characters are used.
6946 More changes to X509_ATTRIBUTE code: allow the setting of types
6947 based on strings. Remove the 'loc' parameter when adding
6948 attributes because these will be a SET OF encoding which is sorted
6952 *) Initial changes to the 'req' utility to allow request generation
6953 automation. This will allow an application to just generate a template
6954 file containing all the field values and have req construct the
6957 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6958 used all over the place including certificate requests and PKCS#7
6959 structures. They are currently handled manually where necessary with
6960 some primitive wrappers for PKCS#7. The new functions behave in a
6961 manner analogous to the X509 extension functions: they allow
6962 attributes to be looked up by NID and added.
6964 Later something similar to the X509V3 code would be desirable to
6965 automatically handle the encoding, decoding and printing of the
6966 more complex types. The string types like challengePassword can
6967 be handled by the string table functions.
6969 Also modified the multi byte string table handling. Now there is
6970 a 'global mask' which masks out certain types. The table itself
6971 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6972 is useful when for example there is only one permissible type
6973 (as in countryName) and using the mask might result in no valid
6977 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6978 SSL_get_peer_finished to allow applications to obtain the latest
6979 Finished messages sent to the peer or expected from the peer,
6980 respectively. (SSL_get_peer_finished is usually the Finished message
6981 actually received from the peer, otherwise the protocol will be aborted.)
6983 As the Finished message are message digests of the complete handshake
6984 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6985 be used for external authentication procedures when the authentication
6986 provided by SSL/TLS is not desired or is not enough.
6989 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6990 the host supports BWX extension and if Compaq C is present on the
6991 $PATH. Just exploiting of the BWX extension results in 20-30%
6992 performance kick for some algorithms, e.g. DES and RC4 to mention
6993 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6997 *) Add support for MS "fast SGC". This is arguably a violation of the
6998 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6999 weak crypto and after checking the certificate is SGC a second one
7000 with strong crypto. MS SGC stops the first handshake after receiving
7001 the server certificate message and sends a second client hello. Since
7002 a server will typically do all the time consuming operations before
7003 expecting any further messages from the client (server key exchange
7004 is the most expensive) there is little difference between the two.
7006 To get OpenSSL to support MS SGC we have to permit a second client
7007 hello message after we have sent server done. In addition we have to
7008 reset the MAC if we do get this second client hello.
7011 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7012 if a DER encoded private key is RSA or DSA traditional format. Changed
7013 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7014 format DER encoded private key. Newer code should use PKCS#8 format which
7015 has the key type encoded in the ASN1 structure. Added DER private key
7016 support to pkcs8 application.
7019 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7020 ciphersuites has been selected (as required by the SSL 3/TLS 1
7021 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7022 is set, we interpret this as a request to violate the specification
7023 (the worst that can happen is a handshake failure, and 'correct'
7024 behaviour would result in a handshake failure anyway).
7027 *) In SSL_CTX_add_session, take into account that there might be multiple
7028 SSL_SESSION structures with the same session ID (e.g. when two threads
7029 concurrently obtain them from an external cache).
7030 The internal cache can handle only one SSL_SESSION with a given ID,
7031 so if there's a conflict, we now throw out the old one to achieve
7035 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7036 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7037 some routines that use cipher OIDs: some ciphers do not have OIDs
7038 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7042 *) Simplify the trust setting structure and code. Now we just have
7043 two sequences of OIDs for trusted and rejected settings. These will
7044 typically have values the same as the extended key usage extension
7045 and any application specific purposes.
7047 The trust checking code now has a default behaviour: it will just
7048 check for an object with the same NID as the passed id. Functions can
7049 be provided to override either the default behaviour or the behaviour
7050 for a given id. SSL client, server and email already have functions
7051 in place for compatibility: they check the NID and also return "trusted"
7052 if the certificate is self signed.
7055 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7056 traditional format into an EVP_PKEY structure.
7059 *) Add a password callback function PEM_cb() which either prompts for
7060 a password if usr_data is NULL or otherwise assumes it is a null
7061 terminated password. Allow passwords to be passed on command line
7062 environment or config files in a few more utilities.
7065 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7066 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7067 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7068 Update documentation.
7071 *) Support for ASN1 "NULL" type. This could be handled before by using
7072 ASN1_TYPE but there wasn't any function that would try to read a NULL
7073 and produce an error if it couldn't. For compatibility we also have
7074 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7075 don't allocate anything because they don't need to.
7078 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7080 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7082 *) Rebuild of the memory allocation routines used by OpenSSL code and
7083 possibly others as well. The purpose is to make an interface that
7084 provide hooks so anyone can build a separate set of allocation and
7085 deallocation routines to be used by OpenSSL, for example memory
7086 pool implementations, or something else, which was previously hard
7087 since Malloc(), Realloc() and Free() were defined as macros having
7088 the values malloc, realloc and free, respectively (except for Win32
7089 compilations). The same is provided for memory debugging code.
7090 OpenSSL already comes with functionality to find memory leaks, but
7091 this gives people a chance to debug other memory problems.
7093 With these changes, a new set of functions and macros have appeared:
7095 CRYPTO_set_mem_debug_functions() [F]
7096 CRYPTO_get_mem_debug_functions() [F]
7097 CRYPTO_dbg_set_options() [F]
7098 CRYPTO_dbg_get_options() [F]
7099 CRYPTO_malloc_debug_init() [M]
7101 The memory debug functions are NULL by default, unless the library
7102 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7103 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7104 gives the standard debugging functions that come with OpenSSL) or
7105 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7106 provided by the library user) must be used. When the standard
7107 debugging functions are used, CRYPTO_dbg_set_options can be used to
7108 request additional information:
7109 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7110 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7112 Also, things like CRYPTO_set_mem_functions will always give the
7113 expected result (the new set of functions is used for allocation
7114 and deallocation) at all times, regardless of platform and compiler
7117 To finish it up, some functions that were never use in any other
7118 way than through macros have a new API and new semantic:
7121 CRYPTO_dbg_realloc()
7124 All macros of value have retained their old syntax.
7125 [Richard Levitte and Bodo Moeller]
7127 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7128 ordering of SMIMECapabilities wasn't in "strength order" and there
7129 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7133 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7134 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7135 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7137 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7138 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7139 functionality to handle multipart/signed properly) and a utility
7140 called 'smime' to call all this stuff. This is based on code I
7141 originally wrote for Celo who have kindly allowed it to be
7142 included in OpenSSL.
7145 *) Add variants des_set_key_checked and des_set_key_unchecked of
7146 des_set_key (aka des_key_sched). Global variable des_check_key
7147 decides which of these is called by des_set_key; this way
7148 des_check_key behaves as it always did, but applications and
7149 the library itself, which was buggy for des_check_key == 1,
7150 have a cleaner way to pick the version they need.
7153 *) New function PKCS12_newpass() which changes the password of a
7157 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7158 dynamic mix. In both cases the ids can be used as an index into the
7159 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7160 functions so they accept a list of the field values and the
7161 application doesn't need to directly manipulate the X509_TRUST
7165 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7169 *) Modify the way the V3 extension code looks up extensions. This now
7170 works in a similar way to the object code: we have some "standard"
7171 extensions in a static table which is searched with OBJ_bsearch()
7172 and the application can add dynamic ones if needed. The file
7173 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7174 updated whenever a new extension is added to the core code and kept
7175 in ext_nid order. There is a simple program 'tabtest.c' which checks
7176 this. New extensions are not added too often so this file can readily
7177 be maintained manually.
7179 There are two big advantages in doing things this way. The extensions
7180 can be looked up immediately and no longer need to be "added" using
7181 X509V3_add_standard_extensions(): this function now does nothing.
7182 [Side note: I get *lots* of email saying the extension code doesn't
7183 work because people forget to call this function]
7184 Also no dynamic allocation is done unless new extensions are added:
7185 so if we don't add custom extensions there is no need to call
7186 X509V3_EXT_cleanup().
7189 *) Modify enc utility's salting as follows: make salting the default. Add a
7190 magic header, so unsalted files fail gracefully instead of just decrypting
7191 to garbage. This is because not salting is a big security hole, so people
7192 should be discouraged from doing it.
7195 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7196 digest to be passed on the command line but it only used this
7197 parameter when signing a certificate. Modified so all relevant
7198 operations are affected by the digest parameter including the
7199 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7200 DSA key was used because it didn't fix the digest.
7203 *) Initial certificate chain verify code. Currently tests the untrusted
7204 certificates for consistency with the verify purpose (which is set
7205 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7207 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7208 this is because it will reject chains with invalid extensions whereas
7209 every previous version of OpenSSL and SSLeay made no checks at all.
7211 Trust code: checks the root CA for the relevant trust settings. Trust
7212 settings have an initial value consistent with the verify purpose: e.g.
7213 if the verify purpose is for SSL client use it expects the CA to be
7214 trusted for SSL client use. However the default value can be changed to
7215 permit custom trust settings: one example of this would be to only trust
7216 certificates from a specific "secure" set of CAs.
7218 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7219 which should be used for version portability: especially since the
7220 verify structure is likely to change more often now.
7222 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7223 to set them. If not set then assume SSL clients will verify SSL servers
7226 Two new options to the verify program: -untrusted allows a set of
7227 untrusted certificates to be passed in and -purpose which sets the
7228 intended purpose of the certificate. If a purpose is set then the
7229 new chain verify code is used to check extension consistency.
7232 *) Support for the authority information access extension.
7235 *) Modify RSA and DSA PEM read routines to transparently handle
7236 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7237 public keys in a format compatible with certificate
7238 SubjectPublicKeyInfo structures. Unfortunately there were already
7239 functions called *_PublicKey_* which used various odd formats so
7240 these are retained for compatibility: however the DSA variants were
7241 never in a public release so they have been deleted. Changed dsa/rsa
7242 utilities to handle the new format: note no releases ever handled public
7243 keys so we should be OK.
7245 The primary motivation for this change is to avoid the same fiasco
7246 that dogs private keys: there are several incompatible private key
7247 formats some of which are standard and some OpenSSL specific and
7248 require various evil hacks to allow partial transparent handling and
7249 even then it doesn't work with DER formats. Given the option anything
7250 other than PKCS#8 should be dumped: but the other formats have to
7251 stay in the name of compatibility.
7253 With public keys and the benefit of hindsight one standard format
7254 is used which works with EVP_PKEY, RSA or DSA structures: though
7255 it clearly returns an error if you try to read the wrong kind of key.
7257 Added a -pubkey option to the 'x509' utility to output the public key.
7258 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7259 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7260 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7261 that do the same as the EVP_PKEY_assign_*() except they up the
7262 reference count of the added key (they don't "swallow" the
7266 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7267 CRLs would fail if the file contained no certificates or no CRLs:
7268 added a new function to read in both types and return the number
7269 read: this means that if none are read it will be an error. The
7270 DER versions of the certificate and CRL reader would always fail
7271 because it isn't possible to mix certificates and CRLs in DER format
7272 without choking one or the other routine. Changed this to just read
7273 a certificate: this is the best we can do. Also modified the code
7274 in apps/verify.c to take notice of return codes: it was previously
7275 attempting to read in certificates from NULL pointers and ignoring
7276 any errors: this is one reason why the cert and CRL reader seemed
7277 to work. It doesn't check return codes from the default certificate
7278 routines: these may well fail if the certificates aren't installed.
7281 *) Code to support otherName option in GeneralName.
7284 *) First update to verify code. Change the verify utility
7285 so it warns if it is passed a self signed certificate:
7286 for consistency with the normal behaviour. X509_verify
7287 has been modified to it will now verify a self signed
7288 certificate if *exactly* the same certificate appears
7289 in the store: it was previously impossible to trust a
7290 single self signed certificate. This means that:
7291 openssl verify ss.pem
7292 now gives a warning about a self signed certificate but
7293 openssl verify -CAfile ss.pem ss.pem
7297 *) For servers, store verify_result in SSL_SESSION data structure
7298 (and add it to external session representation).
7299 This is needed when client certificate verifications fails,
7300 but an application-provided verification callback (set by
7301 SSL_CTX_set_cert_verify_callback) allows accepting the session
7302 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7303 but returns 1): When the session is reused, we have to set
7304 ssl->verify_result to the appropriate error code to avoid
7306 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7308 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7309 case in PKCS7_dataInit() where the signed PKCS7 structure
7310 didn't contain any existing data because it was being created.
7311 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7313 *) Add a salt to the key derivation routines in enc.c. This
7314 forms the first 8 bytes of the encrypted file. Also add a
7315 -S option to allow a salt to be input on the command line.
7318 *) New function X509_cmp(). Oddly enough there wasn't a function
7319 to compare two certificates. We do this by working out the SHA1
7320 hash and comparing that. X509_cmp() will be needed by the trust
7324 *) SSL_get1_session() is like SSL_get_session(), but increments
7325 the reference count in the SSL_SESSION returned.
7326 [Geoff Thorpe <geoff@eu.c2.net>]
7328 *) Fix for 'req': it was adding a null to request attributes.
7329 Also change the X509_LOOKUP and X509_INFO code to handle
7330 certificate auxiliary information.
7333 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7337 *) Add the possibility to add extra information to the memory leak
7338 detecting output, to form tracebacks, showing from where each
7339 allocation was originated: CRYPTO_push_info("constant string") adds
7340 the string plus current file name and line number to a per-thread
7341 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7342 is like calling CYRPTO_pop_info() until the stack is empty.
7343 Also updated memory leak detection code to be multi-thread-safe.
7346 *) Add options -text and -noout to pkcs7 utility and delete the
7347 encryption options which never did anything. Update docs.
7350 *) Add options to some of the utilities to allow the pass phrase
7351 to be included on either the command line (not recommended on
7352 OSes like Unix) or read from the environment. Update the
7353 manpages and fix a few bugs.
7356 *) Add a few manpages for some of the openssl commands.
7359 *) Fix the -revoke option in ca. It was freeing up memory twice,
7360 leaking and not finding already revoked certificates.
7363 *) Extensive changes to support certificate auxiliary information.
7364 This involves the use of X509_CERT_AUX structure and X509_AUX
7365 functions. An X509_AUX function such as PEM_read_X509_AUX()
7366 can still read in a certificate file in the usual way but it
7367 will also read in any additional "auxiliary information". By
7368 doing things this way a fair degree of compatibility can be
7369 retained: existing certificates can have this information added
7370 using the new 'x509' options.
7372 Current auxiliary information includes an "alias" and some trust
7373 settings. The trust settings will ultimately be used in enhanced
7374 certificate chain verification routines: currently a certificate
7375 can only be trusted if it is self signed and then it is trusted
7379 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7380 The problem was that one of the replacement routines had not been working
7381 since SSLeay releases. For now the offending routine has been replaced
7382 with non-optimised assembler. Even so, this now gives around 95%
7383 performance improvement for 1024 bit RSA signs.
7386 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7387 handling. Most clients have the effective key size in bits equal to
7388 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7389 A few however don't do this and instead use the size of the decrypted key
7390 to determine the RC2 key length and the AlgorithmIdentifier to determine
7391 the effective key length. In this case the effective key length can still
7392 be 40 bits but the key length can be 168 bits for example. This is fixed
7393 by manually forcing an RC2 key into the EVP_PKEY structure because the
7394 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7395 the key length and effective key length are equal.
7398 *) Add a bunch of functions that should simplify the creation of
7399 X509_NAME structures. Now you should be able to do:
7400 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7401 and have it automatically work out the correct field type and fill in
7402 the structures. The more adventurous can try:
7403 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7404 and it will (hopefully) work out the correct multibyte encoding.
7407 *) Change the 'req' utility to use the new field handling and multibyte
7408 copy routines. Before the DN field creation was handled in an ad hoc
7409 way in req, ca, and x509 which was rather broken and didn't support
7410 BMPStrings or UTF8Strings. Since some software doesn't implement
7411 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7412 using the dirstring_type option. See the new comment in the default
7413 openssl.cnf for more info.
7416 *) Make crypto/rand/md_rand.c more robust:
7417 - Assure unique random numbers after fork().
7418 - Make sure that concurrent threads access the global counter and
7419 md serializably so that we never lose entropy in them
7420 or use exactly the same state in multiple threads.
7421 Access to the large state is not always serializable because
7422 the additional locking could be a performance killer, and
7423 md should be large enough anyway.
7426 *) New file apps/app_rand.c with commonly needed functionality
7427 for handling the random seed file.
7429 Use the random seed file in some applications that previously did not:
7431 dsaparam -genkey (which also ignored its '-rand' option),
7434 x509 (when signing).
7435 Except on systems with /dev/urandom, it is crucial to have a random
7436 seed file at least for key creation, DSA signing, and for DH exchanges;
7437 for RSA signatures we could do without one.
7439 gendh and gendsa (unlike genrsa) used to read only the first byte
7440 of each file listed in the '-rand' option. The function as previously
7441 found in genrsa is now in app_rand.c and is used by all programs
7442 that support '-rand'.
7445 *) In RAND_write_file, use mode 0600 for creating files;
7446 don't just chmod when it may be too late.
7449 *) Report an error from X509_STORE_load_locations
7450 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7453 *) New function ASN1_mbstring_copy() this copies a string in either
7454 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7455 into an ASN1_STRING type. A mask of permissible types is passed
7456 and it chooses the "minimal" type to use or an error if not type
7460 *) Add function equivalents to the various macros in asn1.h. The old
7461 macros are retained with an M_ prefix. Code inside the library can
7462 use the M_ macros. External code (including the openssl utility)
7463 should *NOT* in order to be "shared library friendly".
7466 *) Add various functions that can check a certificate's extensions
7467 to see if it usable for various purposes such as SSL client,
7468 server or S/MIME and CAs of these types. This is currently
7469 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7470 verification. Also added a -purpose flag to x509 utility to
7471 print out all the purposes.
7474 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7478 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7479 for, obtain and decode and extension and obtain its critical flag.
7480 This allows all the necessary extension code to be handled in a
7481 single function call.
7484 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7485 platforms. See crypto/rc4/rc4_enc.c for further details.
7488 *) New -noout option to asn1parse. This causes no output to be produced
7489 its main use is when combined with -strparse and -out to extract data
7490 from a file (which may not be in ASN.1 format).
7493 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7494 when producing the local key id.
7495 [Richard Levitte <levitte@stacken.kth.se>]
7497 *) New option -dhparam in s_server. This allows a DH parameter file to be
7498 stated explicitly. If it is not stated then it tries the first server
7499 certificate file. The previous behaviour hard coded the filename
7503 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7504 a public key to be input or output. For example:
7505 openssl rsa -in key.pem -pubout -out pubkey.pem
7506 Also added necessary DSA public key functions to handle this.
7509 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7510 in the message. This was handled by allowing
7511 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7512 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7514 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7515 to the end of the strings whereas this didn't. This would cause problems
7516 if strings read with d2i_ASN1_bytes() were later modified.
7517 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7519 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7520 data and it contains EOF it will end up returning an error. This is
7521 caused by input 46 bytes long. The cause is due to the way base64
7522 BIOs find the start of base64 encoded data. They do this by trying a
7523 trial decode on each line until they find one that works. When they
7524 do a flag is set and it starts again knowing it can pass all the
7525 data directly through the decoder. Unfortunately it doesn't reset
7526 the context it uses. This means that if EOF is reached an attempt
7527 is made to pass two EOFs through the context and this causes the
7528 resulting error. This can also cause other problems as well. As is
7529 usual with these problems it takes *ages* to find and the fix is
7530 trivial: move one line.
7531 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7533 *) Ugly workaround to get s_client and s_server working under Windows. The
7534 old code wouldn't work because it needed to select() on sockets and the
7535 tty (for keypresses and to see if data could be written). Win32 only
7536 supports select() on sockets so we select() with a 1s timeout on the
7537 sockets and then see if any characters are waiting to be read, if none
7538 are present then we retry, we also assume we can always write data to
7539 the tty. This isn't nice because the code then blocks until we've
7540 received a complete line of data and it is effectively polling the
7541 keyboard at 1s intervals: however it's quite a bit better than not
7542 working at all :-) A dedicated Windows application might handle this
7543 with an event loop for example.
7546 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7547 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7548 will be called when RSA_sign() and RSA_verify() are used. This is useful
7549 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7550 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7551 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7552 This necessitated the support of an extra signature type NID_md5_sha1
7553 for SSL signatures and modifications to the SSL library to use it instead
7554 of calling RSA_public_decrypt() and RSA_private_encrypt().
7557 *) Add new -verify -CAfile and -CApath options to the crl program, these
7558 will lookup a CRL issuers certificate and verify the signature in a
7559 similar way to the verify program. Tidy up the crl program so it
7560 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7561 less strict. It will now permit CRL extensions even if it is not
7562 a V2 CRL: this will allow it to tolerate some broken CRLs.
7565 *) Initialize all non-automatic variables each time one of the openssl
7566 sub-programs is started (this is necessary as they may be started
7567 multiple times from the "OpenSSL>" prompt).
7568 [Lennart Bang, Bodo Moeller]
7570 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7571 removing all other RSA functionality (this is what NO_RSA does). This
7572 is so (for example) those in the US can disable those operations covered
7573 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7577 *) Non-copying interface to BIO pairs.
7578 (still largely untested)
7581 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7582 ASCII string. This was handled independently in various places before.
7585 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7586 UTF8 strings a character at a time.
7589 *) Use client_version from client hello to select the protocol
7590 (s23_srvr.c) and for RSA client key exchange verification
7591 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7594 *) Add various utility functions to handle SPKACs, these were previously
7595 handled by poking round in the structure internals. Added new function
7596 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7597 print, verify and generate SPKACs. Based on an original idea from
7598 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7601 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7604 *) Allow the config file extension section to be overwritten on the
7605 command line. Based on an original idea from Massimiliano Pala
7606 <madwolf@comune.modena.it>. The new option is called -extensions
7607 and can be applied to ca, req and x509. Also -reqexts to override
7608 the request extensions in req and -crlexts to override the crl extensions
7612 *) Add new feature to the SPKAC handling in ca. Now you can include
7613 the same field multiple times by preceding it by "XXXX." for example:
7616 this is the same syntax as used in the req config file.
7619 *) Allow certificate extensions to be added to certificate requests. These
7620 are specified in a 'req_extensions' option of the req section of the
7621 config file. They can be printed out with the -text option to req but
7622 are otherwise ignored at present.
7625 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7626 data read consists of only the final block it would not decrypted because
7627 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7628 A misplaced 'break' also meant the decrypted final block might not be
7629 copied until the next read.
7632 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7633 a few extra parameters to the DH structure: these will be useful if
7634 for example we want the value of 'q' or implement X9.42 DH.
7637 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7638 provides hooks that allow the default DSA functions or functions on a
7639 "per key" basis to be replaced. This allows hardware acceleration and
7640 hardware key storage to be handled without major modification to the
7641 library. Also added low level modexp hooks and CRYPTO_EX structure and
7642 associated functions.
7645 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7646 as "read only": it can't be written to and the buffer it points to will
7647 not be freed. Reading from a read only BIO is much more efficient than
7648 a normal memory BIO. This was added because there are several times when
7649 an area of memory needs to be read from a BIO. The previous method was
7650 to create a memory BIO and write the data to it, this results in two
7651 copies of the data and an O(n^2) reading algorithm. There is a new
7652 function BIO_new_mem_buf() which creates a read only memory BIO from
7653 an area of memory. Also modified the PKCS#7 routines to use read only
7657 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7658 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7659 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7660 but a retry condition occured while trying to read the rest.
7663 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7664 NID_pkcs7_encrypted by default: this was wrong since this should almost
7665 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7666 the encrypted data type: this is a more sensible place to put it and it
7667 allows the PKCS#12 code to be tidied up that duplicated this
7671 *) Changed obj_dat.pl script so it takes its input and output files on
7672 the command line. This should avoid shell escape redirection problems
7676 *) Initial support for certificate extension requests, these are included
7677 in things like Xenroll certificate requests. Included functions to allow
7678 extensions to be obtained and added.
7681 *) -crlf option to s_client and s_server for sending newlines as
7682 CRLF (as required by many protocols).
7685 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7687 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7688 [Ralf S. Engelschall]
7690 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7691 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7693 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7697 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7698 DH parameters/keys (q is lost during that conversion, but the resulting
7699 DH parameters contain its length).
7701 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7702 much faster than DH_generate_parameters (which creates parameters
7703 where p = 2*q + 1), and also the smaller q makes DH computations
7704 much more efficient (160-bit exponentiation instead of 1024-bit
7705 exponentiation); so this provides a convenient way to support DHE
7706 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7707 utter importance to use
7708 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7710 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7711 when such DH parameters are used, because otherwise small subgroup
7712 attacks may become possible!
7715 *) Avoid memory leak in i2d_DHparams.
7718 *) Allow the -k option to be used more than once in the enc program:
7719 this allows the same encrypted message to be read by multiple recipients.
7722 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7723 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7724 it will always use the numerical form of the OID, even if it has a short
7728 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7729 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7730 otherwise bn_mod_exp was called. In the case of hardware keys for example
7731 no private key components need be present and it might store extra data
7732 in the RSA structure, which cannot be accessed from bn_mod_exp.
7733 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7734 private key operations.
7737 *) Added support for SPARC Linux.
7740 *) pem_password_cb function type incompatibly changed from
7741 typedef int pem_password_cb(char *buf, int size, int rwflag);
7743 ....(char *buf, int size, int rwflag, void *userdata);
7744 so that applications can pass data to their callbacks:
7745 The PEM[_ASN1]_{read,write}... functions and macros now take an
7746 additional void * argument, which is just handed through whenever
7747 the password callback is called.
7748 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7750 New function SSL_CTX_set_default_passwd_cb_userdata.
7752 Compatibility note: As many C implementations push function arguments
7753 onto the stack in reverse order, the new library version is likely to
7754 interoperate with programs that have been compiled with the old
7755 pem_password_cb definition (PEM_whatever takes some data that
7756 happens to be on the stack as its last argument, and the callback
7757 just ignores this garbage); but there is no guarantee whatsoever that
7760 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7761 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7762 problems not only on Windows, but also on some Unix platforms.
7763 To avoid problematic command lines, these definitions are now in an
7764 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7765 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7768 *) MIPS III/IV assembler module is reimplemented.
7771 *) More DES library cleanups: remove references to srand/rand and
7772 delete an unused file.
7775 *) Add support for the the free Netwide assembler (NASM) under Win32,
7776 since not many people have MASM (ml) and it can be hard to obtain.
7777 This is currently experimental but it seems to work OK and pass all
7778 the tests. Check out INSTALL.W32 for info.
7781 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7782 without temporary keys kept an extra copy of the server key,
7783 and connections with temporary keys did not free everything in case
7787 *) New function RSA_check_key and new openssl rsa option -check
7788 for verifying the consistency of RSA keys.
7789 [Ulf Moeller, Bodo Moeller]
7791 *) Various changes to make Win32 compile work:
7792 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7793 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7794 comparison" warnings.
7795 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7798 *) Add a debugging option to PKCS#5 v2 key generation function: when
7799 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7800 derived keys are printed to stderr.
7803 *) Copy the flags in ASN1_STRING_dup().
7804 [Roman E. Pavlov <pre@mo.msk.ru>]
7806 *) The x509 application mishandled signing requests containing DSA
7807 keys when the signing key was also DSA and the parameters didn't match.
7809 It was supposed to omit the parameters when they matched the signing key:
7810 the verifying software was then supposed to automatically use the CA's
7811 parameters if they were absent from the end user certificate.
7813 Omitting parameters is no longer recommended. The test was also
7814 the wrong way round! This was probably due to unusual behaviour in
7815 EVP_cmp_parameters() which returns 1 if the parameters match.
7816 This meant that parameters were omitted when they *didn't* match and
7817 the certificate was useless. Certificates signed with 'ca' didn't have
7819 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7821 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7822 The interface is as follows:
7823 Applications can use
7824 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7825 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7826 "off" is now the default.
7827 The library internally uses
7828 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7829 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7830 to disable memory-checking temporarily.
7832 Some inconsistent states that previously were possible (and were
7833 even the default) are now avoided.
7835 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7836 with each memory chunk allocated; this is occasionally more helpful
7837 than just having a counter.
7839 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7841 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7845 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7846 which largely parallels "options", but is for changing API behaviour,
7847 whereas "options" are about protocol behaviour.
7848 Initial "mode" flags are:
7850 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7851 a single record has been written.
7852 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7853 retries use the same buffer location.
7854 (But all of the contents must be
7858 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7861 *) Fix problems with no-hmac etc.
7862 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7864 *) New functions RSA_get_default_method(), RSA_set_method() and
7865 RSA_get_method(). These allows replacement of RSA_METHODs without having
7866 to mess around with the internals of an RSA structure.
7869 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7870 Also really enable memory leak checks in openssl.c and in some
7872 [Chad C. Mulligan, Bodo Moeller]
7874 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7875 up the length of negative integers. This has now been simplified to just
7876 store the length when it is first determined and use it later, rather
7877 than trying to keep track of where data is copied and updating it to
7879 [Steve Henson, reported by Brien Wheeler
7880 <bwheeler@authentica-security.com>]
7882 *) Add a new function PKCS7_signatureVerify. This allows the verification
7883 of a PKCS#7 signature but with the signing certificate passed to the
7884 function itself. This contrasts with PKCS7_dataVerify which assumes the
7885 certificate is present in the PKCS#7 structure. This isn't always the
7886 case: certificates can be omitted from a PKCS#7 structure and be
7887 distributed by "out of band" means (such as a certificate database).
7890 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7891 function prototypes in pem.h, also change util/mkdef.pl to add the
7892 necessary function names.
7895 *) mk1mf.pl (used by Windows builds) did not properly read the
7896 options set by Configure in the top level Makefile, and Configure
7897 was not even able to write more than one option correctly.
7898 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7901 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7902 file to be loaded from a BIO or FILE pointer. The BIO version will
7903 for example allow memory BIOs to contain config info.
7906 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7907 Whoever hopes to achieve shared-library compatibility across versions
7908 must use this, not the compile-time macro.
7909 (Exercise 0.9.4: Which is the minimum library version required by
7911 Note: All this applies only to multi-threaded programs, others don't
7915 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7916 through a BIO pair triggered the default case, i.e.
7917 SSLerr(...,SSL_R_UNKNOWN_STATE).
7920 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7921 can use the SSL library even if none of the specific BIOs is
7925 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7926 for the encoded length.
7927 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7929 *) Add initial documentation of the X509V3 functions.
7932 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7933 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7934 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7935 secure PKCS#8 private key format with a high iteration count.
7938 *) Fix determination of Perl interpreter: A perl or perl5
7939 _directory_ in $PATH was also accepted as the interpreter.
7940 [Ralf S. Engelschall]
7942 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7943 wrong with it but it was very old and did things like calling
7944 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7948 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7949 to use the new extension code.
7952 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7953 with macros. This should make it easier to change their form, add extra
7954 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7958 *) Add to configuration table a new entry that can specify an alternative
7959 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7960 according to Mark Crispin <MRC@Panda.COM>.
7964 *) DES CBC did not update the IV. Weird.
7967 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7968 Changing the behaviour of the former might break existing programs --
7969 where IV updating is needed, des_ncbc_encrypt can be used.
7972 *) When bntest is run from "make test" it drives bc to check its
7973 calculations, as well as internally checking them. If an internal check
7974 fails, it needs to cause bc to give a non-zero result or make test carries
7975 on without noticing the failure. Fixed.
7978 *) DES library cleanups.
7981 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7982 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7983 ciphers. NOTE: although the key derivation function has been verified
7984 against some published test vectors it has not been extensively tested
7985 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7989 *) Instead of "mkdir -p", which is not fully portable, use new
7990 Perl script "util/mkdir-p.pl".
7993 *) Rewrite the way password based encryption (PBE) is handled. It used to
7994 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7995 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7996 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7997 the 'parameter' field of the AlgorithmIdentifier is passed to the
7998 underlying key generation function so it must do its own ASN1 parsing.
7999 This has also changed the EVP_PBE_CipherInit() function which now has a
8000 'parameter' argument instead of literal salt and iteration count values
8001 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8004 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8005 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8006 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8007 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8008 value was just used as a "magic string" and not used directly its
8009 value doesn't matter.
8012 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8016 *) "linux-sparc64" configuration (ultrapenguin).
8017 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8018 "linux-sparc" configuration.
8019 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8021 *) config now generates no-xxx options for missing ciphers.
8024 *) Support the EBCDIC character set (work in progress).
8025 File ebcdic.c not yet included because it has a different license.
8026 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8028 *) Support BS2000/OSD-POSIX.
8029 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8031 *) Make callbacks for key generation use void * instead of char *.
8034 *) Make S/MIME samples compile (not yet tested).
8037 *) Additional typesafe stacks.
8040 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8044 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8046 *) New configuration variant "sco5-gcc".
8048 *) Updated some demos.
8049 [Sean O Riordain, Wade Scholine]
8051 *) Add missing BIO_free at exit of pkcs12 application.
8054 *) Fix memory leak in conf.c.
8057 *) Updates for Win32 to assembler version of MD5.
8060 *) Set #! path to perl in apps/der_chop to where we found it
8061 instead of using a fixed path.
8064 *) SHA library changes for irix64-mips4-cc.
8067 *) Improvements for VMS support.
8071 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8073 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8074 This also avoids the problems with SC4.2 and unpatched SC5.
8075 [Andy Polyakov <appro@fy.chalmers.se>]
8077 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8078 These are required because of the typesafe stack would otherwise break
8079 existing code. If old code used a structure member which used to be STACK
8080 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8081 sk_num or sk_value it would produce an error because the num, data members
8082 are not present in STACK_OF. Now it just produces a warning. sk_set
8083 replaces the old method of assigning a value to sk_value
8084 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8085 that does this will no longer work (and should use sk_set instead) but
8086 this could be regarded as a "questionable" behaviour anyway.
8089 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8090 correctly handle encrypted S/MIME data.
8093 *) Change type of various DES function arguments from des_cblock
8094 (which means, in function argument declarations, pointer to char)
8095 to des_cblock * (meaning pointer to array with 8 char elements),
8096 which allows the compiler to do more typechecking; it was like
8097 that back in SSLeay, but with lots of ugly casts.
8099 Introduce new type const_des_cblock.
8102 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8103 problems: find RecipientInfo structure that matches recipient certificate
8104 and initialise the ASN1 structures properly based on passed cipher.
8107 *) Belatedly make the BN tests actually check the results.
8110 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8111 to and from BNs: it was completely broken. New compilation option
8112 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8113 key elements as negative integers.
8116 *) Reorganize and speed up MD5.
8117 [Andy Polyakov <appro@fy.chalmers.se>]
8120 [Richard Levitte <richard@levitte.org>]
8122 *) New option -out to asn1parse to allow the parsed structure to be
8123 output to a file. This is most useful when combined with the -strparse
8124 option to examine the output of things like OCTET STRINGS.
8127 *) Make SSL library a little more fool-proof by not requiring any longer
8128 that SSL_set_{accept,connect}_state be called before
8129 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8130 in many applications because usually everything *appeared* to work as
8131 intended anyway -- now it really works as intended).
8134 *) Move openssl.cnf out of lib/.
8137 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8138 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8139 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8140 [Ralf S. Engelschall]
8142 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8143 handle PKCS#7 enveloped data properly.
8144 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8146 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8147 copying pointers. The cert_st handling is changed by this in
8148 various ways (and thus what used to be known as ctx->default_cert
8149 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8150 any longer when s->cert does not give us what we need).
8151 ssl_cert_instantiate becomes obsolete by this change.
8152 As soon as we've got the new code right (possibly it already is?),
8153 we have solved a couple of bugs of the earlier code where s->cert
8154 was used as if it could not have been shared with other SSL structures.
8156 Note that using the SSL API in certain dirty ways now will result
8157 in different behaviour than observed with earlier library versions:
8158 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8159 does not influence s as it used to.
8161 In order to clean up things more thoroughly, inside SSL_SESSION
8162 we don't use CERT any longer, but a new structure SESS_CERT
8163 that holds per-session data (if available); currently, this is
8164 the peer's certificate chain and, for clients, the server's certificate
8165 and temporary key. CERT holds only those values that can have
8166 meaningful defaults in an SSL_CTX.
8169 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8170 from the internal representation. Various PKCS#7 fixes: remove some
8171 evil casts and set the enc_dig_alg field properly based on the signing
8175 *) Allow PKCS#12 password to be set from the command line or the
8176 environment. Let 'ca' get its config file name from the environment
8177 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8181 *) Allow certificate policies extension to use an IA5STRING for the
8182 organization field. This is contrary to the PKIX definition but
8183 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8187 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8188 without disallowing inline assembler and the like for non-pedantic builds.
8191 *) Support Borland C++ builder.
8192 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8197 *) SHA-1 cleanups and performance enhancements.
8198 [Andy Polyakov <appro@fy.chalmers.se>]
8200 *) Sparc v8plus assembler for the bignum library.
8201 [Andy Polyakov <appro@fy.chalmers.se>]
8203 *) Accept any -xxx and +xxx compiler options in Configure.
8206 *) Update HPUX configuration.
8209 *) Add missing sk_<type>_unshift() function to safestack.h
8210 [Ralf S. Engelschall]
8212 *) New function SSL_CTX_use_certificate_chain_file that sets the
8213 "extra_cert"s in addition to the certificate. (This makes sense
8214 only for "PEM" format files, as chains as a whole are not
8218 *) Support verify_depth from the SSL API.
8219 x509_vfy.c had what can be considered an off-by-one-error:
8220 Its depth (which was not part of the external interface)
8221 was actually counting the number of certificates in a chain;
8222 now it really counts the depth.
8225 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8226 instead of X509err, which often resulted in confusing error
8227 messages since the error codes are not globally unique
8228 (e.g. an alleged error in ssl3_accept when a certificate
8229 didn't match the private key).
8231 *) New function SSL_CTX_set_session_id_context that allows to set a default
8232 value (so that you don't need SSL_set_session_id_context for each
8233 connection using the SSL_CTX).
8236 *) OAEP decoding bug fix.
8239 *) Support INSTALL_PREFIX for package builders, as proposed by
8243 *) New Configure options "threads" and "no-threads". For systems
8244 where the proper compiler options are known (currently Solaris
8245 and Linux), "threads" is the default.
8248 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8251 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8252 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8253 such as /usr/local/bin.
8256 *) "make linux-shared" to build shared libraries.
8257 [Niels Poppe <niels@netbox.org>]
8259 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8262 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8263 extension adding in x509 utility.
8266 *) Remove NOPROTO sections and error code comments.
8269 *) Partial rewrite of the DEF file generator to now parse the ANSI
8273 *) New Configure options --prefix=DIR and --openssldir=DIR.
8276 *) Complete rewrite of the error code script(s). It is all now handled
8277 by one script at the top level which handles error code gathering,
8278 header rewriting and C source file generation. It should be much better
8279 than the old method: it now uses a modified version of Ulf's parser to
8280 read the ANSI prototypes in all header files (thus the old K&R definitions
8281 aren't needed for error creation any more) and do a better job of
8282 translating function codes into names. The old 'ASN1 error code imbedded
8283 in a comment' is no longer necessary and it doesn't use .err files which
8284 have now been deleted. Also the error code call doesn't have to appear all
8285 on one line (which resulted in some large lines...).
8288 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8291 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8292 0 (which usually indicates a closed connection), but continue reading.
8295 *) Fix some race conditions.
8298 *) Add support for CRL distribution points extension. Add Certificate
8299 Policies and CRL distribution points documentation.
8302 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8305 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8306 8 of keying material. Merlin has also confirmed interop with this fix
8307 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8308 [Merlin Hughes <merlin@baltimore.ie>]
8310 *) Fix lots of warnings.
8311 [Richard Levitte <levitte@stacken.kth.se>]
8313 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8314 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8315 [Richard Levitte <levitte@stacken.kth.se>]
8317 *) Fix problems with sizeof(long) == 8.
8318 [Andy Polyakov <appro@fy.chalmers.se>]
8320 *) Change functions to ANSI C.
8323 *) Fix typos in error codes.
8324 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8326 *) Remove defunct assembler files from Configure.
8329 *) SPARC v8 assembler BIGNUM implementation.
8330 [Andy Polyakov <appro@fy.chalmers.se>]
8332 *) Support for Certificate Policies extension: both print and set.
8333 Various additions to support the r2i method this uses.
8336 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8337 return a const string when you are expecting an allocated buffer.
8340 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8341 types DirectoryString and DisplayText.
8344 *) Add code to allow r2i extensions to access the configuration database,
8345 add an LHASH database driver and add several ctx helper functions.
8348 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8349 fail when they extended the size of a BIGNUM.
8352 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8353 support typesafe stack.
8356 *) Fix typo in SSL_[gs]et_options().
8357 [Nils Frostberg <nils@medcom.se>]
8359 *) Delete various functions and files that belonged to the (now obsolete)
8360 old X509V3 handling code.
8363 *) New Configure option "rsaref".
8366 *) Don't auto-generate pem.h.
8369 *) Introduce type-safe ASN.1 SETs.
8372 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8373 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8375 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8376 that links with OpenSSL (well at least cause lots of warnings), but fear
8377 not: the conversion is trivial, and it eliminates loads of evil casts. A
8378 few STACKed things have been converted already. Feel free to convert more.
8379 In the fullness of time, I'll do away with the STACK type altogether.
8382 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8383 specified in <certfile> by updating the entry in the index.txt file.
8384 This way one no longer has to edit the index.txt file manually for
8385 revoking a certificate. The -revoke option does the gory details now.
8386 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8388 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8389 `-text' option at all and this way the `-noout -text' combination was
8390 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8391 [Ralf S. Engelschall]
8393 *) Make sure a corresponding plain text error message exists for the
8394 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8395 verify callback function determined that a certificate was revoked.
8396 [Ralf S. Engelschall]
8398 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8399 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8400 all available cipers including rc5, which was forgotten until now.
8401 In order to let the testing shell script know which algorithms
8402 are available, a new (up to now undocumented) command
8403 "openssl list-cipher-commands" is used.
8406 *) Bugfix: s_client occasionally would sleep in select() when
8407 it should have checked SSL_pending() first.
8410 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8411 the raw DSA values prior to ASN.1 encoding.
8414 *) Tweaks to Configure
8415 [Niels Poppe <niels@netbox.org>]
8417 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8421 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8424 *) New config option to avoid instructions that are illegal on the 80386.
8425 The default code is faster, but requires at least a 486.
8428 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8429 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8430 same as SSL2_VERSION anyway.
8433 *) New "-showcerts" option for s_client.
8436 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8437 application. Various cleanups and fixes.
8440 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8441 modify error routines to work internally. Add error codes and PBE init
8442 to library startup routines.
8445 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8446 packing functions to asn1 and evp. Changed function names and error
8447 codes along the way.
8450 *) PKCS12 integration: and so it begins... First of several patches to
8451 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8452 objects to objects.h
8455 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8456 and display support for Thawte strong extranet extension.
8459 *) Add LinuxPPC support.
8460 [Jeff Dubrule <igor@pobox.org>]
8462 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8463 bn_div_words in alpha.s.
8464 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8466 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8467 OAEP isn't supported when OpenSSL is built with RSAref.
8468 [Ulf Moeller <ulf@fitug.de>]
8470 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8471 so they no longer are missing under -DNOPROTO.
8472 [Soren S. Jorvang <soren@t.dk>]
8475 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8477 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8478 doesn't work when the session is reused. Coming soon!
8481 *) Fix a security hole, that allows sessions to be reused in the wrong
8482 context thus bypassing client cert protection! All software that uses
8483 client certs and session caches in multiple contexts NEEDS PATCHING to
8484 allow session reuse! A fuller solution is in the works.
8485 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8487 *) Some more source tree cleanups (removed obsolete files
8488 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8489 permission on "config" script to be executable) and a fix for the INSTALL
8491 [Ulf Moeller <ulf@fitug.de>]
8493 *) Remove some legacy and erroneous uses of malloc, free instead of
8495 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8497 *) Make rsa_oaep_test return non-zero on error.
8498 [Ulf Moeller <ulf@fitug.de>]
8500 *) Add support for native Solaris shared libraries. Configure
8501 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8502 if someone would make that last step automatic.
8503 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8505 *) ctx_size was not built with the right compiler during "make links". Fixed.
8508 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8509 except NULL ciphers". This means the default cipher list will no longer
8510 enable NULL ciphers. They need to be specifically enabled e.g. with
8511 the string "DEFAULT:eNULL".
8514 *) Fix to RSA private encryption routines: if p < q then it would
8515 occasionally produce an invalid result. This will only happen with
8516 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8519 *) Be less restrictive and allow also `perl util/perlpath.pl
8520 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8521 because this way one can also use an interpreter named `perl5' (which is
8522 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8523 installed as `perl').
8524 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8526 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8527 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8529 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8530 advapi32.lib to Win32 build and change the pem test comparision
8531 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8532 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8533 and crypto/des/ede_cbcm_enc.c.
8536 *) DES quad checksum was broken on big-endian architectures. Fixed.
8539 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8540 Win32 test batch file so it (might) work again. The Win32 test batch file
8541 is horrible: I feel ill....
8544 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8545 in e_os.h. Audit of header files to check ANSI and non ANSI
8546 sections: 10 functions were absent from non ANSI section and not exported
8547 from Windows DLLs. Fixed up libeay.num for new functions.
8550 *) Make `openssl version' output lines consistent.
8551 [Ralf S. Engelschall]
8553 *) Fix Win32 symbol export lists for BIO functions: Added
8554 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8555 to ms/libeay{16,32}.def.
8556 [Ralf S. Engelschall]
8558 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8559 fine under Unix and passes some trivial tests I've now added. But the
8560 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8561 added to make sure no one expects that this stuff really works in the
8562 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8563 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8565 [Ralf S. Engelschall]
8567 *) Fix the generation of two part addresses in perl.
8568 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8570 *) Add config entry for Linux on MIPS.
8571 [John Tobey <jtobey@channel1.com>]
8573 *) Make links whenever Configure is run, unless we are on Windoze.
8576 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8577 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8581 *) Add a useful kludge to allow package maintainers to specify compiler and
8582 other platforms details on the command line without having to patch the
8583 Configure script everytime: One now can use ``perl Configure
8584 <id>:<details>'', i.e. platform ids are allowed to have details appended
8585 to them (seperated by colons). This is treated as there would be a static
8586 pre-configured entry in Configure's %table under key <id> with value
8587 <details> and ``perl Configure <id>'' is called. So, when you want to
8588 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8589 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8590 now, which overrides the FreeBSD-elf entry on-the-fly.
8591 [Ralf S. Engelschall]
8593 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8596 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8597 on the `perl Configure ...' command line. This way one can compile
8598 OpenSSL libraries with Position Independent Code (PIC) which is needed
8599 for linking it into DSOs.
8600 [Ralf S. Engelschall]
8602 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8606 *) Cleaned up the LICENSE document: The official contact for any license
8607 questions now is the OpenSSL core team under openssl-core@openssl.org.
8608 And add a paragraph about the dual-license situation to make sure people
8609 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8610 to the OpenSSL toolkit.
8611 [Ralf S. Engelschall]
8613 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8614 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8615 Additonally cleaned up the `make links' target: Remove unnecessary
8616 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8617 to speed processing and no longer clutter the display with confusing
8618 stuff. Instead only the actually done links are displayed.
8619 [Ralf S. Engelschall]
8621 *) Permit null encryption ciphersuites, used for authentication only. It used
8622 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8623 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8627 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8628 signed attributes when verifying signatures (this would break them),
8629 the detached data encoding was wrong and public keys obtained using
8630 X509_get_pubkey() weren't freed.
8633 *) Add text documentation for the BUFFER functions. Also added a work around
8634 to a Win95 console bug. This was triggered by the password read stuff: the
8635 last character typed gets carried over to the next fread(). If you were
8636 generating a new cert request using 'req' for example then the last
8637 character of the passphrase would be CR which would then enter the first
8641 *) Added the new `Includes OpenSSL Cryptography Software' button as
8642 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8643 button and can be used by applications based on OpenSSL to show the
8644 relationship to the OpenSSL project.
8645 [Ralf S. Engelschall]
8647 *) Remove confusing variables in function signatures in files
8648 ssl/ssl_lib.c and ssl/ssl.h.
8649 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8651 *) Don't install bss_file.c under PREFIX/include/
8652 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8654 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8655 functions that return function pointers and has support for NT specific
8656 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8657 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8658 unsigned to signed types: this was killing the Win32 compile.
8661 *) Add new certificate file to stack functions,
8662 SSL_add_dir_cert_subjects_to_stack() and
8663 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8664 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8665 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8666 This means that Apache-SSL and similar packages don't have to mess around
8667 to add as many CAs as they want to the preferred list.
8670 *) Experiment with doxygen documentation. Currently only partially applied to
8672 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8673 openssl.doxy as the configuration file.
8676 *) Get rid of remaining C++-style comments which strict C compilers hate.
8677 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8679 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8680 compiled in by default: it has problems with large keys.
8683 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8684 DH private keys and/or callback functions which directly correspond to
8685 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8686 is needed for applications which have to configure certificates on a
8687 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8689 For the RSA certificate situation is makes no difference, but
8690 for the DSA certificate situation this fixes the "no shared cipher"
8691 problem where the OpenSSL cipher selection procedure failed because the
8692 temporary keys were not overtaken from the context and the API provided
8693 no way to reconfigure them.
8694 The new functions now let applications reconfigure the stuff and they
8695 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8696 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8697 non-public-API function ssl_cert_instantiate() is used as a helper
8698 function and also to reduce code redundancy inside ssl_rsa.c.
8699 [Ralf S. Engelschall]
8701 *) Move s_server -dcert and -dkey options out of the undocumented feature
8702 area because they are useful for the DSA situation and should be
8703 recognized by the users.
8704 [Ralf S. Engelschall]
8706 *) Fix the cipher decision scheme for export ciphers: the export bits are
8707 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8708 SSL_EXP_MASK. So, the original variable has to be used instead of the
8709 already masked variable.
8710 [Richard Levitte <levitte@stacken.kth.se>]
8712 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8713 [Richard Levitte <levitte@stacken.kth.se>]
8715 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8716 from `int' to `unsigned int' because it's a length and initialized by
8717 EVP_DigestFinal() which expects an `unsigned int *'.
8718 [Richard Levitte <levitte@stacken.kth.se>]
8720 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8721 script. Instead use the usual Shell->Perl transition trick.
8722 [Ralf S. Engelschall]
8724 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8725 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8726 -noout -modulus' as it's already the case for `openssl rsa -noout
8727 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8728 currently the public key is printed (a decision which was already done by
8729 `openssl dsa -modulus' in the past) which serves a similar purpose.
8730 Additionally the NO_RSA no longer completely removes the whole -modulus
8731 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8733 [Ralf S. Engelschall]
8735 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8736 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8737 [Arne Ansper <arne@ats.cyber.ee>]
8739 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8740 to be added. Now both 'req' and 'ca' can use new objects defined in the
8744 *) Add cool BIO that does syslog (or event log on NT).
8745 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8747 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8748 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8749 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8750 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8753 *) Add preliminary config info for new extension code.
8756 *) Make RSA_NO_PADDING really use no padding.
8757 [Ulf Moeller <ulf@fitug.de>]
8759 *) Generate errors when private/public key check is done.
8762 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8763 for some CRL extensions and new objects added.
8766 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8767 key usage extension and fuller support for authority key id.
8770 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8771 padding method for RSA, which is recommended for new applications in PKCS
8772 #1 v2.0 (RFC 2437, October 1998).
8773 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8774 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8775 against Bleichbacher's attack on RSA.
8776 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8779 *) Updates to the new SSL compression code
8780 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8782 *) Fix so that the version number in the master secret, when passed
8783 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8784 (because the server will not accept higher), that the version number
8785 is 0x03,0x01, not 0x03,0x00
8786 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8788 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8789 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8790 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8793 *) Support for RAW extensions where an arbitrary extension can be
8794 created by including its DER encoding. See apps/openssl.cnf for
8798 *) Make sure latest Perl versions don't interpret some generated C array
8799 code as Perl array code in the crypto/err/err_genc.pl script.
8800 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8802 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8803 not many people have the assembler. Various Win32 compilation fixes and
8804 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8808 *) Modify configure script 'Configure' to automatically create crypto/date.h
8809 file under Win32 and also build pem.h from pem.org. New script
8810 util/mkfiles.pl to create the MINFO file on environments that can't do a
8811 'make files': perl util/mkfiles.pl >MINFO should work.
8814 *) Major rework of DES function declarations, in the pursuit of correctness
8815 and purity. As a result, many evil casts evaporated, and some weirdness,
8816 too. You may find this causes warnings in your code. Zapping your evil
8817 casts will probably fix them. Mostly.
8820 *) Fix for a typo in asn1.h. Bug fix to object creation script
8821 obj_dat.pl. It considered a zero in an object definition to mean
8822 "end of object": none of the objects in objects.h have any zeros
8823 so it wasn't spotted.
8824 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8826 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8827 Masking (CBCM). In the absence of test vectors, the best I have been able
8828 to do is check that the decrypt undoes the encrypt, so far. Send me test
8829 vectors if you have them.
8832 *) Correct calculation of key length for export ciphers (too much space was
8833 allocated for null ciphers). This has not been tested!
8836 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8837 message is now correct (it understands "crypto" and "ssl" on its
8838 command line). There is also now an "update" option. This will update
8839 the util/ssleay.num and util/libeay.num files with any new functions.
8841 perl util/mkdef.pl crypto ssl update
8842 it will update them.
8845 *) Overhauled the Perl interface (perl/*):
8846 - ported BN stuff to OpenSSL's different BN library
8847 - made the perl/ source tree CVS-aware
8848 - renamed the package from SSLeay to OpenSSL (the files still contain
8849 their history because I've copied them in the repository)
8850 - removed obsolete files (the test scripts will be replaced
8851 by better Test::Harness variants in the future)
8852 [Ralf S. Engelschall]
8854 *) First cut for a very conservative source tree cleanup:
8855 1. merge various obsolete readme texts into doc/ssleay.txt
8856 where we collect the old documents and readme texts.
8857 2. remove the first part of files where I'm already sure that we no
8858 longer need them because of three reasons: either they are just temporary
8859 files which were left by Eric or they are preserved original files where
8860 I've verified that the diff is also available in the CVS via "cvs diff
8861 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8862 the crypto/md/ stuff).
8863 [Ralf S. Engelschall]
8865 *) More extension code. Incomplete support for subject and issuer alt
8866 name, issuer and authority key id. Change the i2v function parameters
8867 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8868 what that's for :-) Fix to ASN1 macro which messed up
8869 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8872 *) Preliminary support for ENUMERATED type. This is largely copied from the
8876 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8877 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8879 *) Make sure `make rehash' target really finds the `openssl' program.
8880 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8882 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8883 like to hear about it if this slows down other processors.
8886 *) Add CygWin32 platform information to Configure script.
8887 [Alan Batie <batie@aahz.jf.intel.com>]
8889 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8890 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8892 *) New program nseq to manipulate netscape certificate sequences
8895 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8899 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8900 but the BN code had some problems that would cause failures when
8901 doing certificate verification and some other functions.
8902 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8904 *) Add ASN1 and PEM code to support netscape certificate sequences.
8907 *) Add ASN1 and PEM code to support netscape certificate sequences.
8910 *) Add several PKIX and private extended key usage OIDs.
8913 *) Modify the 'ca' program to handle the new extension code. Modify
8914 openssl.cnf for new extension format, add comments.
8917 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8918 and add a sample to openssl.cnf so req -x509 now adds appropriate
8922 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8923 error code, add initial support to X509_print() and x509 application.
8926 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8927 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8928 stuff is currently isolated and isn't even compiled yet.
8931 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8932 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8933 Removed the versions check from X509 routines when loading extensions:
8934 this allows certain broken certificates that don't set the version
8935 properly to be processed.
8938 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8939 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8940 can still be regenerated with "make depend".
8943 *) Spelling mistake in C version of CAST-128.
8944 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8946 *) Changes to the error generation code. The perl script err-code.pl
8947 now reads in the old error codes and retains the old numbers, only
8948 adding new ones if necessary. It also only changes the .err files if new
8949 codes are added. The makefiles have been modified to only insert errors
8950 when needed (to avoid needlessly modifying header files). This is done
8951 by only inserting errors if the .err file is newer than the auto generated
8952 C file. To rebuild all the error codes from scratch (the old behaviour)
8953 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8954 or delete all the .err files.
8957 *) CAST-128 was incorrectly implemented for short keys. The C version has
8958 been fixed, but is untested. The assembler versions are also fixed, but
8959 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8960 to regenerate it if needed.
8961 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8962 Hagino <itojun@kame.net>]
8964 *) File was opened incorrectly in randfile.c.
8965 [Ulf Möller <ulf@fitug.de>]
8967 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8968 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8969 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8970 al: it's just almost always a UTCTime. Note this patch adds new error
8971 codes so do a "make errors" if there are problems.
8974 *) Correct Linux 1 recognition in config.
8975 [Ulf Möller <ulf@fitug.de>]
8977 *) Remove pointless MD5 hash when using DSA keys in ca.
8978 [Anonymous <nobody@replay.com>]
8980 *) Generate an error if given an empty string as a cert directory. Also
8981 generate an error if handed NULL (previously returned 0 to indicate an
8982 error, but didn't set one).
8983 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8985 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8988 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8989 parameters. This was causing a warning which killed off the Win32 compile.
8992 *) Remove C++ style comments from crypto/bn/bn_local.h.
8993 [Neil Costigan <neil.costigan@celocom.com>]
8995 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8996 based on a text string, looking up short and long names and finally
8997 "dot" format. The "dot" format stuff didn't work. Added new function
8998 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8999 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9000 OID is not part of the table.
9003 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9004 X509_LOOKUP_by_alias().
9007 *) Sort openssl functions by name.
9010 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9011 encryption from sample DSA keys (in case anyone is interested the password
9015 *) Make _all_ *_free functions accept a NULL pointer.
9016 [Frans Heymans <fheymans@isaserver.be>]
9018 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9020 [Anonymous <nobody@replay.com>]
9022 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9023 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9025 *) Don't blow it for numeric -newkey arguments to apps/req.
9026 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9028 *) Temp key "for export" tests were wrong in s3_srvr.c.
9029 [Anonymous <nobody@replay.com>]
9031 *) Add prototype for temp key callback functions
9032 SSL_CTX_set_tmp_{rsa,dh}_callback().
9035 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9036 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9039 *) X509_name_add_entry() freed the wrong thing after an error.
9040 [Arne Ansper <arne@ats.cyber.ee>]
9042 *) rsa_eay.c would attempt to free a NULL context.
9043 [Arne Ansper <arne@ats.cyber.ee>]
9045 *) BIO_s_socket() had a broken should_retry() on Windoze.
9046 [Arne Ansper <arne@ats.cyber.ee>]
9048 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9049 [Arne Ansper <arne@ats.cyber.ee>]
9051 *) Make sure the already existing X509_STORE->depth variable is initialized
9052 in X509_STORE_new(), but document the fact that this variable is still
9053 unused in the certificate verification process.
9054 [Ralf S. Engelschall]
9056 *) Fix the various library and apps files to free up pkeys obtained from
9057 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9060 *) Fix reference counting in X509_PUBKEY_get(). This makes
9061 demos/maurice/example2.c work, amongst others, probably.
9062 [Steve Henson and Ben Laurie]
9064 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9065 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9066 are no longer created. This way we have a single and consistent command
9067 line interface `openssl <command>', similar to `cvs <command>'.
9068 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9070 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9071 BIT STRING wrapper always have zero unused bits.
9074 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9077 *) Make the top-level INSTALL documentation easier to understand.
9080 *) Makefiles updated to exit if an error occurs in a sub-directory
9081 make (including if user presses ^C) [Paul Sutton]
9083 *) Make Montgomery context stuff explicit in RSA data structure.
9086 *) Fix build order of pem and err to allow for generated pem.h.
9089 *) Fix renumbering bug in X509_NAME_delete_entry().
9092 *) Enhanced the err-ins.pl script so it makes the error library number
9093 global and can add a library name. This is needed for external ASN1 and
9094 other error libraries.
9097 *) Fixed sk_insert which never worked properly.
9100 *) Fix ASN1 macros so they can handle indefinite length construted
9101 EXPLICIT tags. Some non standard certificates use these: they can now
9105 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9106 into a single doc/ssleay.txt bundle. This way the information is still
9107 preserved but no longer messes up this directory. Now it's new room for
9108 the new set of documenation files.
9109 [Ralf S. Engelschall]
9111 *) SETs were incorrectly DER encoded. This was a major pain, because they
9112 shared code with SEQUENCEs, which aren't coded the same. This means that
9113 almost everything to do with SETs or SEQUENCEs has either changed name or
9114 number of arguments.
9115 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9117 *) Fix test data to work with the above.
9120 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9121 was already fixed by Eric for 0.9.1 it seems.
9122 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9124 *) Autodetect FreeBSD3.
9127 *) Fix various bugs in Configure. This affects the following platforms:
9131 unixware-2.0-pentium
9135 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9136 before they are needed.
9139 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9143 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9145 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9146 changed SSLeay to OpenSSL in version strings.
9147 [Ralf S. Engelschall]
9149 *) Some fixups to the top-level documents.
9152 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9153 because the symlink to include/ was missing.
9154 [Ralf S. Engelschall]
9156 *) Incorporated the popular no-RSA/DSA-only patches
9157 which allow to compile a RSA-free SSLeay.
9158 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9160 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9161 when "ssleay" is still not found.
9162 [Ralf S. Engelschall]
9164 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9165 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9167 *) Updated the README file.
9168 [Ralf S. Engelschall]
9170 *) Added various .cvsignore files in the CVS repository subdirs
9171 to make a "cvs update" really silent.
9172 [Ralf S. Engelschall]
9174 *) Recompiled the error-definition header files and added
9175 missing symbols to the Win32 linker tables.
9176 [Ralf S. Engelschall]
9178 *) Cleaned up the top-level documents;
9179 o new files: CHANGES and LICENSE
9180 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9181 o merged COPYRIGHT into LICENSE
9182 o removed obsolete TODO file
9183 o renamed MICROSOFT to INSTALL.W32
9184 [Ralf S. Engelschall]
9186 *) Removed dummy files from the 0.9.1b source tree:
9187 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9188 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9189 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9190 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9191 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9192 [Ralf S. Engelschall]
9194 *) Added various platform portability fixes.
9197 *) The Genesis of the OpenSSL rpject:
9198 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9199 Young and Tim J. Hudson created while they were working for C2Net until
9201 [The OpenSSL Project]
9204 Changes between 0.9.0b and 0.9.1b [not released]
9206 *) Updated a few CA certificates under certs/
9209 *) Changed some BIGNUM api stuff.
9212 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9213 DGUX x86, Linux Alpha, etc.
9216 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9217 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9221 *) Add -strparse option to asn1pars program which parses nested
9223 [Dr Stephen Henson <shenson@bigfoot.com>]
9225 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9228 *) DSA fix for "ca" program.
9231 *) Added "-genkey" option to "dsaparam" program.
9234 *) Added RIPE MD160 (rmd160) message digest.
9237 *) Added -a (all) option to "ssleay version" command.
9240 *) Added PLATFORM define which is the id given to Configure.
9243 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9246 *) Extended the ASN.1 parser routines.
9249 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9252 *) Added a BN_CTX to the BN library.
9255 *) Fixed the weak key values in DES library
9258 *) Changed API in EVP library for cipher aliases.
9261 *) Added support for RC2/64bit cipher.
9264 *) Converted the lhash library to the crypto/mem.c functions.
9267 *) Added more recognized ASN.1 object ids.
9270 *) Added more RSA padding checks for SSL/TLS.
9273 *) Added BIO proxy/filter functionality.
9276 *) Added extra_certs to SSL_CTX which can be used
9277 send extra CA certificates to the client in the CA cert chain sending
9278 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9281 *) Now Fortezza is denied in the authentication phase because
9282 this is key exchange mechanism is not supported by SSLeay at all.
9285 *) Additional PKCS1 checks.
9288 *) Support the string "TLSv1" for all TLS v1 ciphers.
9291 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9292 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9295 *) Fixed a few memory leaks.
9298 *) Fixed various code and comment typos.
9301 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9302 bytes sent in the client random.
9303 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob