5 Changes between 1.0.0 and 1.1.0 [xx XXX xxxx]
7 *) Add signature printing for PSS. Add PSS OIDs.
8 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
10 *) Add algorithm specific signature printing. An individual ASN1 method
11 can now print out signatures instead of the standard hex dump.
13 More complex signatures (e.g. PSS) can print out more meaningful
14 information. Include DSA version that prints out the signature
18 *) Add -trusted_first option which attempts to find certificates in the
19 trusted store even if an untrusted chain is also supplied.
22 *) Initial experimental support for explicitly trusted non-root CAs.
23 OpenSSL still tries to build a complete chain to a root but if an
24 intermediate CA has a trust setting included that is used. The first
25 setting is used: whether to trust or reject.
28 *) New -verify_name option in command line utilities to set verification
32 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
33 Add CMAC pkey methods.
36 *) Experiemental regnegotiation in s_server -www mode. If the client
37 browses /reneg connection is renegotiated. If /renegcert it is
38 renegotiated requesting a certificate.
41 *) Add an "external" session cache for debugging purposes to s_server. This
42 should help trace issues which normally are only apparent in deployed
43 multi-process servers.
46 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
47 a few changes are required:
49 Add SSL_OP_NO_TLSv1_1 flag.
51 Update version checking logic to handle version 1.1.
52 Add explicit IV handling (ported from DTLS code).
53 Add command line options to s_client/s_server.
56 *) Experiemental password based recipient info support for CMS library:
60 *) Split password based encryption into PBES2 and PBKDF2 functions. This
61 neatly separates the code into cipher and PBE sections and is required
62 for some algorithms that split PBES2 into separate pieces (such as
66 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
67 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
68 BIO_set_cipher() and some obscure PEM functions were changed so they
69 can now return an error. The RAND changes required a change to the
70 RAND_METHOD structure.
73 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
74 a gcc attribute to warn if the result of a function is ignored. This
75 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
76 whose return value is often ignored.
79 Changes between 0.9.8m and 1.0.0 [25 Feb 2010]
81 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
82 context. The operation can be customised via the ctrl mechanism in
83 case ENGINEs want to include additional functionality.
86 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
89 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
90 output hashes compatible with older versions of OpenSSL.
91 [Willy Weisz <weisz@vcpc.univie.ac.at>]
93 *) Fix compression algorithm handling: if resuming a session use the
94 compression algorithm of the resumed session instead of determining
95 it from client hello again. Don't allow server to change algorithm.
98 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
99 change when encrypting or decrypting.
102 *) Add load_crls() function to apps tidying load_certs() too. Add option
103 to verify utility to allow additional CRLs to be included.
106 *) Update OCSP request code to permit adding custom headers to the request:
107 some responders need this.
110 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
112 [Julia Lawall <julia@diku.dk>]
114 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
115 needlessly dereferenced structures, used obsolete functions and
116 didn't handle all updated verify codes correctly.
119 *) Disable MD2 in the default configuration.
122 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
123 indicate the initial BIO being pushed or popped. This makes it possible
124 to determine whether the BIO is the one explicitly called or as a result
125 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
126 it handles reference counts correctly and doesn't zero out the I/O bio
127 when it is not being explicitly popped. WARNING: applications which
128 included workarounds for the old buggy behaviour will need to be modified
129 or they could free up already freed BIOs.
132 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
133 renaming to all platforms (within the 0.9.8 branch, this was
134 done conditionally on Netware platforms to avoid a name clash).
135 [Guenter <lists@gknw.net>]
137 *) Add ECDHE and PSK support to DTLS.
138 [Michael Tuexen <tuexen@fh-muenster.de>]
140 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
144 *) Add "missing" function EVP_MD_flags() (without this the only way to
145 retrieve a digest flags is by accessing the structure directly. Update
146 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
147 or cipher is registered as in the "from" argument. Print out all
148 registered digests in the dgst usage message instead of manually
149 attempting to work them out.
152 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
153 this allows the use of compression and extensions. Change default cipher
154 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
155 by default unless an application cipher string requests it.
158 *) Alter match criteria in PKCS12_parse(). It used to try to use local
159 key ids to find matching certificates and keys but some PKCS#12 files
160 don't follow the (somewhat unwritten) rules and this strategy fails.
161 Now just gather all certificates together and the first private key
162 then look for the first certificate that matches the key.
165 *) Support use of registered digest and cipher names for dgst and cipher
166 commands instead of having to add each one as a special case. So now
173 openssl dgst -sha256 foo
175 and this works for ENGINE based algorithms too.
179 *) Update Gost ENGINE to support parameter files.
180 [Victor B. Wagner <vitus@cryptocom.ru>]
182 *) Support GeneralizedTime in ca utility.
183 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
185 *) Enhance the hash format used for certificate directory links. The new
186 form uses the canonical encoding (meaning equivalent names will work
187 even if they aren't identical) and uses SHA1 instead of MD5. This form
188 is incompatible with the older format and as a result c_rehash should
189 be used to rebuild symbolic links.
192 *) Make PKCS#8 the default write format for private keys, replacing the
193 traditional format. This form is standardised, more secure and doesn't
194 include an implicit MD5 dependency.
197 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
198 committed to OpenSSL should pass this lot as a minimum.
201 *) Add session ticket override functionality for use by EAP-FAST.
202 [Jouni Malinen <j@w1.fi>]
204 *) Modify HMAC functions to return a value. Since these can be implemented
205 in an ENGINE errors can occur.
208 *) Type-checked OBJ_bsearch_ex.
211 *) Type-checked OBJ_bsearch. Also some constification necessitated
212 by type-checking. Still to come: TXT_DB, bsearch(?),
213 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
217 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
218 seconds to a tm structure directly, instead of going through OS
219 specific date routines. This avoids any issues with OS routines such
220 as the year 2038 bug. New *_adj() functions for ASN1 time structures
221 and X509_time_adj_ex() to cover the extended range. The existing
222 X509_time_adj() is still usable and will no longer have any date issues.
225 *) Delta CRL support. New use deltas option which will attempt to locate
226 and search any appropriate delta CRLs available.
228 This work was sponsored by Google.
231 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
232 code and add additional score elements. Validate alternate CRL paths
233 as part of the CRL checking and indicate a new error "CRL path validation
234 error" in this case. Applications wanting additional details can use
235 the verify callback and check the new "parent" field. If this is not
236 NULL CRL path validation is taking place. Existing applications wont
237 see this because it requires extended CRL support which is off by
240 This work was sponsored by Google.
243 *) Support for freshest CRL extension.
245 This work was sponsored by Google.
248 *) Initial indirect CRL support. Currently only supported in the CRLs
249 passed directly and not via lookup. Process certificate issuer
250 CRL entry extension and lookup CRL entries by bother issuer name
251 and serial number. Check and process CRL issuer entry in IDP extension.
253 This work was sponsored by Google.
256 *) Add support for distinct certificate and CRL paths. The CRL issuer
257 certificate is validated separately in this case. Only enabled if
258 an extended CRL support flag is set: this flag will enable additional
259 CRL functionality in future.
261 This work was sponsored by Google.
264 *) Add support for policy mappings extension.
266 This work was sponsored by Google.
269 *) Fixes to pathlength constraint, self issued certificate handling,
270 policy processing to align with RFC3280 and PKITS tests.
272 This work was sponsored by Google.
275 *) Support for name constraints certificate extension. DN, email, DNS
276 and URI types are currently supported.
278 This work was sponsored by Google.
281 *) To cater for systems that provide a pointer-based thread ID rather
282 than numeric, deprecate the current numeric thread ID mechanism and
283 replace it with a structure and associated callback type. This
284 mechanism allows a numeric "hash" to be extracted from a thread ID in
285 either case, and on platforms where pointers are larger than 'long',
286 mixing is done to help ensure the numeric 'hash' is usable even if it
287 can't be guaranteed unique. The default mechanism is to use "&errno"
288 as a pointer-based thread ID to distinguish between threads.
290 Applications that want to provide their own thread IDs should now use
291 CRYPTO_THREADID_set_callback() to register a callback that will call
292 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
294 Note that ERR_remove_state() is now deprecated, because it is tied
295 to the assumption that thread IDs are numeric. ERR_remove_state(0)
296 to free the current thread's error state should be replaced by
297 ERR_remove_thread_state(NULL).
299 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
300 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
301 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
302 application was previously providing a numeric thread callback that
303 was inappropriate for distinguishing threads, then uniqueness might
304 have been obtained with &errno that happened immediately in the
305 intermediate development versions of OpenSSL; this is no longer the
306 case, the numeric thread callback will now override the automatic use
308 [Geoff Thorpe, with help from Bodo Moeller]
310 *) Initial support for different CRL issuing certificates. This covers a
311 simple case where the self issued certificates in the chain exist and
312 the real CRL issuer is higher in the existing chain.
314 This work was sponsored by Google.
317 *) Removed effectively defunct crypto/store from the build.
320 *) Revamp of STACK to provide stronger type-checking. Still to come:
321 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
322 ASN1_STRING, CONF_VALUE.
325 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
326 RAM on SSL connections. This option can save about 34k per idle SSL.
329 *) Revamp of LHASH to provide stronger type-checking. Still to come:
330 STACK, TXT_DB, bsearch, qsort.
333 *) Initial support for Cryptographic Message Syntax (aka CMS) based
334 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
335 support for data, signedData, compressedData, digestedData and
336 encryptedData, envelopedData types included. Scripts to check against
337 RFC4134 examples draft and interop and consistency checks of many
338 content types and variants.
341 *) Add options to enc utility to support use of zlib compression BIO.
344 *) Extend mk1mf to support importing of options and assembly language
345 files from Configure script, currently only included in VC-WIN32.
346 The assembly language rules can now optionally generate the source
347 files from the associated perl scripts.
350 *) Implement remaining functionality needed to support GOST ciphersuites.
351 Interop testing has been performed using CryptoPro implementations.
352 [Victor B. Wagner <vitus@cryptocom.ru>]
354 *) s390x assembler pack.
357 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
361 *) Implement Opaque PRF Input TLS extension as specified in
362 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
363 official specification yet and no extension type assignment by
364 IANA exists, this extension (for now) will have to be explicitly
365 enabled when building OpenSSL by providing the extension number
366 to use. For example, specify an option
368 -DTLSEXT_TYPE_opaque_prf_input=0x9527
370 to the "config" or "Configure" script to enable the extension,
371 assuming extension number 0x9527 (which is a completely arbitrary
372 and unofficial assignment based on the MD5 hash of the Internet
373 Draft). Note that by doing so, you potentially lose
374 interoperability with other TLS implementations since these might
375 be using the same extension number for other purposes.
377 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
378 opaque PRF input value to use in the handshake. This will create
379 an interal copy of the length-'len' string at 'src', and will
380 return non-zero for success.
382 To get more control and flexibility, provide a callback function
385 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
386 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
390 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
393 Callback function 'cb' will be called in handshakes, and is
394 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
395 Argument 'arg' is for application purposes (the value as given to
396 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
397 be provided to the callback function). The callback function
398 has to return non-zero to report success: usually 1 to use opaque
399 PRF input just if possible, or 2 to enforce use of the opaque PRF
400 input. In the latter case, the library will abort the handshake
401 if opaque PRF input is not successfully negotiated.
403 Arguments 'peerinput' and 'len' given to the callback function
404 will always be NULL and 0 in the case of a client. A server will
405 see the client's opaque PRF input through these variables if
406 available (NULL and 0 otherwise). Note that if the server
407 provides an opaque PRF input, the length must be the same as the
408 length of the client's opaque PRF input.
410 Note that the callback function will only be called when creating
411 a new session (session resumption can resume whatever was
412 previously negotiated), and will not be called in SSL 2.0
413 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
414 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
415 for applications that need to enforce opaque PRF input.
419 *) Update ssl code to support digests other than SHA1+MD5 for handshake
422 [Victor B. Wagner <vitus@cryptocom.ru>]
424 *) Add RFC4507 support to OpenSSL. This includes the corrections in
425 RFC4507bis. The encrypted ticket format is an encrypted encoded
426 SSL_SESSION structure, that way new session features are automatically
429 If a client application caches session in an SSL_SESSION structure
430 support is transparent because tickets are now stored in the encoded
433 The SSL_CTX structure automatically generates keys for ticket
434 protection in servers so again support should be possible
435 with no application modification.
437 If a client or server wishes to disable RFC4507 support then the option
438 SSL_OP_NO_TICKET can be set.
440 Add a TLS extension debugging callback to allow the contents of any client
441 or server extensions to be examined.
443 This work was sponsored by Google.
446 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
447 OpenSSL should now compile cleanly on gcc 4.2
448 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
450 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
451 support including streaming MAC support: this is required for GOST
453 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
455 *) Add option -stream to use PKCS#7 streaming in smime utility. New
456 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
457 to output in BER and PEM format.
460 *) Experimental support for use of HMAC via EVP_PKEY interface. This
461 allows HMAC to be handled via the EVP_DigestSign*() interface. The
462 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
463 ENGINE support for HMAC keys which are unextractable. New -mac and
464 -macopt options to dgst utility.
467 *) New option -sigopt to dgst utility. Update dgst to use
468 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
469 alternative signing paramaters such as X9.31 or PSS in the dgst
473 *) Change ssl_cipher_apply_rule(), the internal function that does
474 the work each time a ciphersuite string requests enabling
475 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
476 removing ("!foo+bar") a class of ciphersuites: Now it maintains
477 the order of disabled ciphersuites such that those ciphersuites
478 that most recently went from enabled to disabled not only stay
479 in order with respect to each other, but also have higher priority
480 than other disabled ciphersuites the next time ciphersuites are
483 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
484 the same ciphersuites as with "HIGH" alone, but in a specific
485 order where the PSK ciphersuites come first (since they are the
486 most recently disabled ciphersuites when "HIGH" is parsed).
488 Also, change ssl_create_cipher_list() (using this new
489 funcionality) such that between otherwise identical
490 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
494 *) Change ssl_create_cipher_list() so that it automatically
495 arranges the ciphersuites in reasonable order before starting
496 to process the rule string. Thus, the definition for "DEFAULT"
497 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
498 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
499 This makes it much easier to arrive at a reasonable default order
500 in applications for which anonymous ciphers are OK (meaning
501 that you can't actually use DEFAULT).
502 [Bodo Moeller; suggested by Victor Duchovni]
504 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
505 processing) into multiple integers instead of setting
506 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
507 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
508 (These masks as well as the individual bit definitions are hidden
509 away into the non-exported interface ssl/ssl_locl.h, so this
510 change to the definition of the SSL_CIPHER structure shouldn't
511 affect applications.) This give us more bits for each of these
512 categories, so there is no longer a need to coagulate AES128 and
513 AES256 into a single algorithm bit, and to coagulate Camellia128
514 and Camellia256 into a single algorithm bit, which has led to all
517 Thus, among other things, the kludge introduced in 0.9.7m and
518 0.9.8e for masking out AES256 independently of AES128 or masking
519 out Camellia256 independently of AES256 is not needed here in 0.9.9.
521 With the change, we also introduce new ciphersuite aliases that
522 so far were missing: "AES128", "AES256", "CAMELLIA128", and
526 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
527 Use the leftmost N bytes of the signature input if the input is
528 larger than the prime q (with N being the size in bytes of q).
531 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
532 it yet and it is largely untested.
535 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
538 *) Initial incomplete changes to avoid need for function casts in OpenSSL
539 some compilers (gcc 4.2 and later) reject their use. Safestack is
540 reimplemented. Update ASN1 to avoid use of legacy functions.
543 *) Win32/64 targets are linked with Winsock2.
546 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
547 to external functions. This can be used to increase CRL handling
548 efficiency especially when CRLs are very large by (for example) storing
549 the CRL revoked certificates in a database.
552 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
553 new CRLs added to a directory can be used. New command line option
554 -verify_return_error to s_client and s_server. This causes real errors
555 to be returned by the verify callback instead of carrying on no matter
556 what. This reflects the way a "real world" verify callback would behave.
559 *) GOST engine, supporting several GOST algorithms and public key formats.
560 Kindly donated by Cryptocom.
563 *) Partial support for Issuing Distribution Point CRL extension. CRLs
564 partitioned by DP are handled but no indirect CRL or reason partitioning
565 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
566 selected via a scoring technique which handles IDP and AKID in CRLs.
569 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
570 will ultimately be used for all verify operations: this will remove the
571 X509_STORE dependency on certificate verification and allow alternative
572 lookup methods. X509_STORE based implementations of these two callbacks.
575 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
576 Modify get_crl() to find a valid (unexpired) CRL if possible.
579 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
580 this would be called X509_CRL_cmp() but that name is already used by
581 a function that just compares CRL issuer names. Cache several CRL
582 extensions in X509_CRL structure and cache CRLDP in X509.
585 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
586 this maps equivalent X509_NAME structures into a consistent structure.
587 Name comparison can then be performed rapidly using memcmp().
590 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
594 *) Allow digests to supply their own micalg string for S/MIME type using
595 the ctrl EVP_MD_CTRL_MICALG.
598 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
599 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
600 ctrl. It can then customise the structure before and/or after signing
604 *) New function OBJ_add_sigid() to allow application defined signature OIDs
605 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
606 to free up any added signature OIDs.
609 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
610 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
611 digest and cipher tables. New options added to openssl utility:
612 list-message-digest-algorithms and list-cipher-algorithms.
615 *) Change the array representation of binary polynomials: the list
616 of degrees of non-zero coefficients is now terminated with -1.
617 Previously it was terminated with 0, which was also part of the
618 value; thus, the array representation was not applicable to
619 polynomials where t^0 has coefficient zero. This change makes
620 the array representation useful in a more general context.
623 *) Various modifications and fixes to SSL/TLS cipher string
624 handling. For ECC, the code now distinguishes between fixed ECDH
625 with RSA certificates on the one hand and with ECDSA certificates
626 on the other hand, since these are separate ciphersuites. The
627 unused code for Fortezza ciphersuites has been removed.
629 For consistency with EDH, ephemeral ECDH is now called "EECDH"
630 (not "ECDHE"). For consistency with the code for DH
631 certificates, use of ECDH certificates is now considered ECDH
632 authentication, not RSA or ECDSA authentication (the latter is
633 merely the CA's signing algorithm and not actively used in the
636 The temporary ciphersuite alias "ECCdraft" is no longer
637 available, and ECC ciphersuites are no longer excluded from "ALL"
638 and "DEFAULT". The following aliases now exist for RFC 4492
639 ciphersuites, most of these by analogy with the DH case:
641 kECDHr - ECDH cert, signed with RSA
642 kECDHe - ECDH cert, signed with ECDSA
643 kECDH - ECDH cert (signed with either RSA or ECDSA)
644 kEECDH - ephemeral ECDH
645 ECDH - ECDH cert or ephemeral ECDH
651 AECDH - anonymous ECDH
652 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
656 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
657 Use correct micalg parameters depending on digest(s) in signed message.
660 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
661 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
664 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
665 an engine to register a method. Add ENGINE lookups for methods and
666 functional reference processing.
669 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
670 EVP_{Sign,Verify}* which allow an application to customise the signature
674 *) New -resign option to smime utility. This adds one or more signers
675 to an existing PKCS#7 signedData structure. Also -md option to use an
676 alternative message digest algorithm for signing.
679 *) Tidy up PKCS#7 routines and add new functions to make it easier to
680 create PKCS7 structures containing multiple signers. Update smime
681 application to support multiple signers.
684 *) New -macalg option to pkcs12 utility to allow setting of an alternative
688 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
689 Reorganize PBE internals to lookup from a static table using NIDs,
690 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
691 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
692 PRF which will be automatically used with PBES2.
695 *) Replace the algorithm specific calls to generate keys in "req" with the
699 *) Update PKCS#7 enveloped data routines to use new API. This is now
700 supported by any public key method supporting the encrypt operation. A
701 ctrl is added to allow the public key algorithm to examine or modify
702 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
706 *) Add a ctrl to asn1 method to allow a public key algorithm to express
707 a default digest type to use. In most cases this will be SHA1 but some
708 algorithms (such as GOST) need to specify an alternative digest. The
709 return value indicates how strong the prefernce is 1 means optional and
710 2 is mandatory (that is it is the only supported type). Modify
711 ASN1_item_sign() to accept a NULL digest argument to indicate it should
712 use the default md. Update openssl utilities to use the default digest
713 type for signing if it is not explicitly indicated.
716 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
717 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
718 signing method from the key type. This effectively removes the link
719 between digests and public key types.
722 *) Add an OID cross reference table and utility functions. Its purpose is to
723 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
724 rsaEncryption. This will allow some of the algorithm specific hackery
725 needed to use the correct OID to be removed.
728 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
729 structures for PKCS7_sign(). They are now set up by the relevant public
733 *) Add provisional EC pkey method with support for ECDSA and ECDH.
736 *) Add support for key derivation (agreement) in the API, DH method and
740 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
741 public and private key formats. As a side effect these add additional
742 command line functionality not previously available: DSA signatures can be
743 generated and verified using pkeyutl and DH key support and generation in
748 [Oliver Tappe <zooey@hirschkaefer.de>]
750 *) New make target "install_html_docs" installs HTML renditions of the
752 [Oliver Tappe <zooey@hirschkaefer.de>]
754 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
755 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
756 support key and parameter generation and add initial key generation
757 functionality for RSA.
760 *) Add functions for main EVP_PKEY_method operations. The undocumented
761 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
762 EVP_PKEY_{encrypt,decrypt}_old.
765 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
766 key API, doesn't do much yet.
769 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
770 public key algorithms. New option to openssl utility:
771 "list-public-key-algorithms" to print out info.
774 *) Implement the Supported Elliptic Curves Extension for
775 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
778 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
779 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
782 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
783 utilities such as rsa, dsa, dsaparam etc except they process any key
787 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
788 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
789 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
793 *) Initial support for pluggable public key ASN1.
794 De-spaghettify the public key ASN1 handling. Move public and private
795 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
796 algorithm specific handling to a single module within the relevant
797 algorithm directory. Add functions to allow (near) opaque processing
798 of public and private key structures.
801 *) Implement the Supported Point Formats Extension for
802 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
805 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
806 for the psk identity [hint] and the psk callback functions to the
807 SSL_SESSION, SSL and SSL_CTX structure.
810 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
814 SSL_CTX_use_psk_identity_hint
815 SSL_get_psk_identity_hint
817 SSL_use_psk_identity_hint
819 [Mika Kousa and Pasi Eronen of Nokia Corporation]
821 *) Add RFC 3161 compliant time stamp request creation, response generation
822 and response verification functionality.
823 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
825 *) Add initial support for TLS extensions, specifically for the server_name
826 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
827 have new members for a host name. The SSL data structure has an
828 additional member SSL_CTX *initial_ctx so that new sessions can be
829 stored in that context to allow for session resumption, even after the
830 SSL has been switched to a new SSL_CTX in reaction to a client's
831 server_name extension.
833 New functions (subject to change):
836 SSL_get_servername_type()
839 New CTRL codes and macros (subject to change):
841 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
842 - SSL_CTX_set_tlsext_servername_callback()
843 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
844 - SSL_CTX_set_tlsext_servername_arg()
845 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
847 openssl s_client has a new '-servername ...' option.
849 openssl s_server has new options '-servername_host ...', '-cert2 ...',
850 '-key2 ...', '-servername_fatal' (subject to change). This allows
851 testing the HostName extension for a specific single host name ('-cert'
852 and '-key' remain fallbacks for handshakes without HostName
853 negotiation). If the unrecogninzed_name alert has to be sent, this by
854 default is a warning; it becomes fatal with the '-servername_fatal'
857 [Peter Sylvester, Remy Allais, Christophe Renou]
859 *) Whirlpool hash implementation is added.
862 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
863 bn(64,32). Because of instruction set limitations it doesn't have
864 any negative impact on performance. This was done mostly in order
865 to make it possible to share assembler modules, such as bn_mul_mont
866 implementations, between 32- and 64-bit builds without hassle.
869 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
870 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
874 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
875 dedicated Montgomery multiplication procedure, is introduced.
876 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
877 "64-bit" performance on certain 32-bit targets.
880 *) New option SSL_OP_NO_COMP to disable use of compression selectively
881 in SSL structures. New SSL ctrl to set maximum send fragment size.
882 Save memory by seeting the I/O buffer sizes dynamically instead of
883 using the maximum available value.
886 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
887 in addition to the text details.
890 *) Very, very preliminary EXPERIMENTAL support for printing of general
891 ASN1 structures. This currently produces rather ugly output and doesn't
892 handle several customised structures at all.
895 *) Integrated support for PVK file format and some related formats such
896 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
897 these in the 'rsa' and 'dsa' utilities.
900 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
903 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
904 place for the (very old) "NETSCAPE" format certificates which are now
905 handled using new ASN1 code equivalents.
908 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
909 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
910 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
913 *) Modify CRL distribution points extension code to print out previously
914 unsupported fields. Enhance extension setting code to allow setting of
918 *) Add print and set support for Issuing Distribution Point CRL extension.
921 *) Change 'Configure' script to enable Camellia by default.
924 Changes between 0.9.8m and 0.9.8n [xx XXX xxxx]
926 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
927 could be crashed if the relevant tables were not present (e.g. chrooted).
928 [Tomas Hoger <thoger@redhat.com>]
930 Changes between 0.9.8l and 0.9.8m [xx XXX xxxx]
932 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
933 accommodate for stack sorting, always a write lock!).
936 *) On some versions of WIN32 Heap32Next is very slow. This can cause
937 excessive delays in the RAND_poll(): over a minute. As a workaround
938 include a time check in the inner Heap32Next loop too.
941 *) The code that handled flushing of data in SSL/TLS originally used the
942 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
943 the problem outlined in PR#1949. The fix suggested there however can
944 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
945 of Apache). So instead simplify the code to flush unconditionally.
946 This should be fine since flushing with no data to flush is a no op.
949 *) Handle TLS versions 2.0 and later properly and correctly use the
950 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
951 off ancient servers have a habit of sticking around for a while...
954 *) Modify compression code so it frees up structures without using the
955 ex_data callbacks. This works around a problem where some applications
956 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
957 restarting) then use compression (e.g. SSL with compression) later.
958 This results in significant per-connection memory leaks and
959 has caused some security issues including CVE-2008-1678 and
963 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
964 connect and renegotiate with servers which do not support RI.
965 Until RI is more widely deployed this option is enabled by default.
968 *) Add "missing" ssl ctrls to clear options and mode.
971 *) If client attempts to renegotiate and doesn't support RI respond with
972 a no_renegotiation alert as required by RFC5746. Some renegotiating
973 TLS clients will continue a connection gracefully when they receive
974 the alert. Unfortunately OpenSSL mishandled this alert and would hang
975 waiting for a server hello which it will never receive. Now we treat a
976 received no_renegotiation alert as a fatal error. This is because
977 applications requesting a renegotiation might well expect it to succeed
978 and would have no code in place to handle the server denying it so the
979 only safe thing to do is to terminate the connection.
982 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
983 peer supports secure renegotiation and 0 otherwise. Print out peer
984 renegotiation support in s_client/s_server.
987 *) Replace the highly broken and deprecated SPKAC certification method with
988 the updated NID creation version. This should correctly handle UTF8.
991 *) Implement RFC5746. Re-enable renegotiation but require the extension
992 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
993 turns out to be a bad idea. It has been replaced by
994 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
995 SSL_CTX_set_options(). This is really not recommended unless you
996 know what you are doing.
997 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
999 *) Fixes to stateless session resumption handling. Use initial_ctx when
1000 issuing and attempting to decrypt tickets in case it has changed during
1001 servername handling. Use a non-zero length session ID when attempting
1002 stateless session resumption: this makes it possible to determine if
1003 a resumption has occurred immediately after receiving server hello
1004 (several places in OpenSSL subtly assume this) instead of later in
1008 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1009 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1010 fixes for a few places where the return code is not checked
1012 [Julia Lawall <julia@diku.dk>]
1014 *) Add --strict-warnings option to Configure script to include devteam
1015 warnings in other configurations.
1018 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1019 makes it possible to install openssl libraries in locations which
1020 have names other than "lib", for example "/usr/lib64" which some
1022 [Steve Henson, based on patch from Jeremy Utley]
1024 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1025 X690 8.9.12 and can produce some misleading textual output of OIDs.
1026 [Steve Henson, reported by Dan Kaminsky]
1028 *) Delete MD2 from algorithm tables. This follows the recommendation in
1029 several standards that it is not used in new applications due to
1030 several cryptographic weaknesses. For binary compatibility reasons
1031 the MD2 API is still compiled in by default.
1034 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1038 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1039 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1041 [Guenter <lists@gknw.net>]
1043 *) Fix the server certificate chain building code to use X509_verify_cert(),
1044 it used to have an ad-hoc builder which was unable to cope with anything
1045 other than a simple chain.
1046 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1048 *) Don't check self signed certificate signatures in X509_verify_cert()
1049 by default (a flag can override this): it just wastes time without
1050 adding any security. As a useful side effect self signed root CAs
1051 with non-FIPS digests are now usable in FIPS mode.
1054 *) In dtls1_process_out_of_seq_message() the check if the current message
1055 is already buffered was missing. For every new message was memory
1056 allocated, allowing an attacker to perform an denial of service attack
1057 with sending out of seq handshake messages until there is no memory
1058 left. Additionally every future messege was buffered, even if the
1059 sequence number made no sense and would be part of another handshake.
1060 So only messages with sequence numbers less than 10 in advance will be
1061 buffered. (CVE-2009-1378)
1062 [Robin Seggelmann, discovered by Daniel Mentz]
1064 *) Records are buffered if they arrive with a future epoch to be
1065 processed after finishing the corresponding handshake. There is
1066 currently no limitation to this buffer allowing an attacker to perform
1067 a DOS attack with sending records with future epochs until there is no
1068 memory left. This patch adds the pqueue_size() function to detemine
1069 the size of a buffer and limits the record buffer to 100 entries.
1071 [Robin Seggelmann, discovered by Daniel Mentz]
1073 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1074 parent structure is freed. (CVE-2009-1379)
1077 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1078 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1081 [Ilya O. <vrghost@gmail.com>]
1083 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1085 *) Disable renegotiation completely - this fixes a severe security
1086 problem (CVE-2009-3555) at the cost of breaking all
1087 renegotiation. Renegotiation can be re-enabled by setting
1088 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1089 run-time. This is really not recommended unless you know what
1093 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1095 *) Don't set val to NULL when freeing up structures, it is freed up by
1096 underlying code. If sizeof(void *) > sizeof(long) this can result in
1097 zeroing past the valid field. (CVE-2009-0789)
1098 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1100 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1101 checked correctly. This would allow some invalid signed attributes to
1102 appear to verify correctly. (CVE-2009-0591)
1103 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1105 *) Reject UniversalString and BMPString types with invalid lengths. This
1106 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1107 a legal length. (CVE-2009-0590)
1110 *) Set S/MIME signing as the default purpose rather than setting it
1111 unconditionally. This allows applications to override it at the store
1115 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1116 to handle some structures.
1119 *) Improve efficiency of mem_gets: don't search whole buffer each time
1121 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1123 *) New -hex option for openssl rand.
1126 *) Print out UTF8String and NumericString when parsing ASN1.
1129 *) Support NumericString type for name components.
1132 *) Allow CC in the environment to override the automatically chosen
1133 compiler. Note that nothing is done to ensure flags work with the
1137 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1139 *) Properly check EVP_VerifyFinal() and similar return values
1141 [Ben Laurie, Bodo Moeller, Google Security Team]
1143 *) Enable TLS extensions by default.
1146 *) Allow the CHIL engine to be loaded, whether the application is
1147 multithreaded or not. (This does not release the developer from the
1148 obligation to set up the dynamic locking callbacks.)
1149 [Sander Temme <sander@temme.net>]
1151 *) Use correct exit code if there is an error in dgst command.
1152 [Steve Henson; problem pointed out by Roland Dirlewanger]
1154 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1155 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1158 *) Add experimental JPAKE support, including demo authentication in
1159 s_client and s_server.
1162 *) Set the comparison function in v3_addr_canonize().
1163 [Rob Austein <sra@hactrn.net>]
1165 *) Add support for XMPP STARTTLS in s_client.
1166 [Philip Paeps <philip@freebsd.org>]
1168 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1169 to ensure that even with this option, only ciphersuites in the
1170 server's preference list will be accepted. (Note that the option
1171 applies only when resuming a session, so the earlier behavior was
1172 just about the algorithm choice for symmetric cryptography.)
1175 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1177 *) Fix NULL pointer dereference if a DTLS server received
1178 ChangeCipherSpec as first record (CVE-2009-1386).
1181 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1182 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1185 *) The fix in 0.9.8c that supposedly got rid of unsafe
1186 double-checked locking was incomplete for RSA blinding,
1187 addressing just one layer of what turns out to have been
1188 doubly unsafe triple-checked locking.
1190 So now fix this for real by retiring the MONT_HELPER macro
1191 in crypto/rsa/rsa_eay.c.
1193 [Bodo Moeller; problem pointed out by Marius Schilder]
1195 *) Various precautionary measures:
1197 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1199 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1200 (NB: This would require knowledge of the secret session ticket key
1201 to exploit, in which case you'd be SOL either way.)
1203 - Change bn_nist.c so that it will properly handle input BIGNUMs
1204 outside the expected range.
1206 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1209 [Neel Mehta, Bodo Moeller]
1211 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1212 the load fails. Useful for distros.
1213 [Ben Laurie and the FreeBSD team]
1215 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1218 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1221 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1223 This work was sponsored by Logica.
1226 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1227 keystores. Support for SSL/TLS client authentication too.
1228 Not compiled unless enable-capieng specified to Configure.
1230 This work was sponsored by Logica.
1233 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1234 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1235 attribute creation routines such as certifcate requests and PKCS#12
1239 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1241 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1242 handshake which could lead to a cilent crash as found using the
1243 Codenomicon TLS test suite (CVE-2008-1672)
1244 [Steve Henson, Mark Cox]
1246 *) Fix double free in TLS server name extensions which could lead to
1247 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1250 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1252 Clear the error queue to ensure that error entries left from
1253 older function calls do not interfere with the correct operation.
1254 [Lutz Jaenicke, Erik de Castro Lopo]
1256 *) Remove root CA certificates of commercial CAs:
1258 The OpenSSL project does not recommend any specific CA and does not
1259 have any policy with respect to including or excluding any CA.
1260 Therefore it does not make any sense to ship an arbitrary selection
1261 of root CA certificates with the OpenSSL software.
1264 *) RSA OAEP patches to fix two separate invalid memory reads.
1265 The first one involves inputs when 'lzero' is greater than
1266 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1267 before the beginning of from). The second one involves inputs where
1268 the 'db' section contains nothing but zeroes (there is a one-byte
1269 invalid read after the end of 'db').
1270 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1272 *) Partial backport from 0.9.9-dev:
1274 Introduce bn_mul_mont (dedicated Montgomery multiplication
1275 procedure) as a candidate for BIGNUM assembler implementation.
1276 While 0.9.9-dev uses assembler for various architectures, only
1277 x86_64 is available by default here in the 0.9.8 branch, and
1278 32-bit x86 is available through a compile-time setting.
1280 To try the 32-bit x86 assembler implementation, use Configure
1281 option "enable-montasm" (which exists only for this backport).
1283 As "enable-montasm" for 32-bit x86 disclaims code stability
1284 anyway, in this constellation we activate additional code
1285 backported from 0.9.9-dev for further performance improvements,
1286 namely BN_from_montgomery_word. (To enable this otherwise,
1287 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1289 [Andy Polyakov (backport partially by Bodo Moeller)]
1291 *) Add TLS session ticket callback. This allows an application to set
1292 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1293 values. This is useful for key rollover for example where several key
1294 sets may exist with different names.
1297 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1298 This was broken until now in 0.9.8 releases, such that the only way
1299 a registered ENGINE could be used (assuming it initialises
1300 successfully on the host) was to explicitly set it as the default
1301 for the relevant algorithms. This is in contradiction with 0.9.7
1302 behaviour and the documentation. With this fix, when an ENGINE is
1303 registered into a given algorithm's table of implementations, the
1304 'uptodate' flag is reset so that auto-discovery will be used next
1305 time a new context for that algorithm attempts to select an
1307 [Ian Lister (tweaked by Geoff Thorpe)]
1309 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1310 implemention in the following ways:
1312 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1315 Lack of BER streaming support means one pass streaming processing is
1316 only supported if data is detached: setting the streaming flag is
1317 ignored for embedded content.
1319 CMS support is disabled by default and must be explicitly enabled
1320 with the enable-cms configuration option.
1323 *) Update the GMP engine glue to do direct copies between BIGNUM and
1324 mpz_t when openssl and GMP use the same limb size. Otherwise the
1325 existing "conversion via a text string export" trick is still used.
1326 [Paul Sheer <paulsheer@gmail.com>]
1328 *) Zlib compression BIO. This is a filter BIO which compressed and
1329 uncompresses any data passed through it.
1332 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1333 RFC3394 compatible AES key wrapping.
1336 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1337 sets string data without copying. X509_ALGOR_set0() and
1338 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1339 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1340 from an X509_ATTRIBUTE structure optionally checking it occurs only
1341 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1345 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1346 to get the expected BN_FLG_CONSTTIME behavior.
1347 [Bodo Moeller (Google)]
1351 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1352 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1353 - added some more tests to do_tests.pl
1354 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1355 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1356 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1357 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1358 - various changes to netware.pl to enable gcc-cross builds on Win32
1360 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1361 - various changes to fix missing prototype warnings
1362 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1363 - added AES, WHIRLPOOL and CPUID assembler code to build files
1364 - added missing AES assembler make rules to mk1mf.pl
1365 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1366 [Guenter Knauf <eflash@gmx.net>]
1368 *) Implement certificate status request TLS extension defined in RFC3546.
1369 A client can set the appropriate parameters and receive the encoded
1370 OCSP response via a callback. A server can query the supplied parameters
1371 and set the encoded OCSP response in the callback. Add simplified examples
1372 to s_client and s_server.
1375 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1377 *) Fix various bugs:
1378 + Binary incompatibility of ssl_ctx_st structure
1379 + DTLS interoperation with non-compliant servers
1380 + Don't call get_session_cb() without proposed session
1381 + Fix ia64 assembler code
1382 [Andy Polyakov, Steve Henson]
1384 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1386 *) DTLS Handshake overhaul. There were longstanding issues with
1387 OpenSSL DTLS implementation, which were making it impossible for
1388 RFC 4347 compliant client to communicate with OpenSSL server.
1389 Unfortunately just fixing these incompatibilities would "cut off"
1390 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1391 server keeps tolerating non RFC compliant syntax. The opposite is
1392 not true, 0.9.8f client can not communicate with earlier server.
1393 This update even addresses CVE-2007-4995.
1396 *) Changes to avoid need for function casts in OpenSSL: some compilers
1397 (gcc 4.2 and later) reject their use.
1398 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1401 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1402 RFC4507bis. The encrypted ticket format is an encrypted encoded
1403 SSL_SESSION structure, that way new session features are automatically
1406 If a client application caches session in an SSL_SESSION structure
1407 support is transparent because tickets are now stored in the encoded
1410 The SSL_CTX structure automatically generates keys for ticket
1411 protection in servers so again support should be possible
1412 with no application modification.
1414 If a client or server wishes to disable RFC4507 support then the option
1415 SSL_OP_NO_TICKET can be set.
1417 Add a TLS extension debugging callback to allow the contents of any client
1418 or server extensions to be examined.
1420 This work was sponsored by Google.
1423 *) Add initial support for TLS extensions, specifically for the server_name
1424 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1425 have new members for a host name. The SSL data structure has an
1426 additional member SSL_CTX *initial_ctx so that new sessions can be
1427 stored in that context to allow for session resumption, even after the
1428 SSL has been switched to a new SSL_CTX in reaction to a client's
1429 server_name extension.
1431 New functions (subject to change):
1433 SSL_get_servername()
1434 SSL_get_servername_type()
1437 New CTRL codes and macros (subject to change):
1439 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1440 - SSL_CTX_set_tlsext_servername_callback()
1441 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1442 - SSL_CTX_set_tlsext_servername_arg()
1443 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1445 openssl s_client has a new '-servername ...' option.
1447 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1448 '-key2 ...', '-servername_fatal' (subject to change). This allows
1449 testing the HostName extension for a specific single host name ('-cert'
1450 and '-key' remain fallbacks for handshakes without HostName
1451 negotiation). If the unrecogninzed_name alert has to be sent, this by
1452 default is a warning; it becomes fatal with the '-servername_fatal'
1455 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1457 *) Add AES and SSE2 assembly language support to VC++ build.
1460 *) Mitigate attack on final subtraction in Montgomery reduction.
1463 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1464 (which previously caused an internal error).
1467 *) Squeeze another 10% out of IGE mode when in != out.
1470 *) AES IGE mode speedup.
1471 [Dean Gaudet (Google)]
1473 *) Add the Korean symmetric 128-bit cipher SEED (see
1474 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1475 add SEED ciphersuites from RFC 4162:
1477 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1478 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1479 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1480 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1482 To minimize changes between patchlevels in the OpenSSL 0.9.8
1483 series, SEED remains excluded from compilation unless OpenSSL
1484 is configured with 'enable-seed'.
1485 [KISA, Bodo Moeller]
1487 *) Mitigate branch prediction attacks, which can be practical if a
1488 single processor is shared, allowing a spy process to extract
1489 information. For detailed background information, see
1490 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1491 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1492 and Necessary Software Countermeasures"). The core of the change
1493 are new versions BN_div_no_branch() and
1494 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1495 respectively, which are slower, but avoid the security-relevant
1496 conditional branches. These are automatically called by BN_div()
1497 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1498 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1499 remove a conditional branch.
1501 BN_FLG_CONSTTIME is the new name for the previous
1502 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1503 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1504 in the exponent causes BN_mod_exp_mont() to use the alternative
1505 implementation in BN_mod_exp_mont_consttime().) The old name
1506 remains as a deprecated alias.
1508 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1509 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1510 constant-time implementations for more than just exponentiation.
1511 Here too the old name is kept as a deprecated alias.
1513 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1514 the BN_BLINDING structure gets an independent copy of the
1515 modulus. This means that the previous "BIGNUM *m" argument to
1516 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1517 essentially becomes "const BIGNUM *m", although we can't actually
1518 change this in the header file before 0.9.9. It allows
1519 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1520 enable BN_FLG_CONSTTIME.
1522 [Matthew D Wood (Intel Corp)]
1524 *) In the SSL/TLS server implementation, be strict about session ID
1525 context matching (which matters if an application uses a single
1526 external cache for different purposes). Previously,
1527 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1528 set. This did ensure strict client verification, but meant that,
1529 with applications using a single external cache for quite
1530 different requirements, clients could circumvent ciphersuite
1531 restrictions for a given session ID context by starting a session
1532 in a different context.
1535 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1536 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1537 authentication-only ciphersuites.
1540 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1541 not complete and could lead to a possible single byte overflow
1542 (CVE-2007-5135) [Ben Laurie]
1544 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1546 *) Since AES128 and AES256 (and similarly Camellia128 and
1547 Camellia256) share a single mask bit in the logic of
1548 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1549 kludge to work properly if AES128 is available and AES256 isn't
1550 (or if Camellia128 is available and Camellia256 isn't).
1553 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1554 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1555 When a point or a seed is encoded in a BIT STRING, we need to
1556 prevent the removal of trailing zero bits to get the proper DER
1557 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1558 of a NamedBitList, for which trailing 0 bits need to be removed.)
1561 *) Have SSL/TLS server implementation tolerate "mismatched" record
1562 protocol version while receiving ClientHello even if the
1563 ClientHello is fragmented. (The server can't insist on the
1564 particular protocol version it has chosen before the ServerHello
1565 message has informed the client about his choice.)
1568 *) Add RFC 3779 support.
1569 [Rob Austein for ARIN, Ben Laurie]
1571 *) Load error codes if they are not already present instead of using a
1572 static variable. This allows them to be cleanly unloaded and reloaded.
1573 Improve header file function name parsing.
1576 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1577 or CAPABILITY handshake as required by RFCs.
1580 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1582 *) Introduce limits to prevent malicious keys being able to
1583 cause a denial of service. (CVE-2006-2940)
1584 [Steve Henson, Bodo Moeller]
1586 *) Fix ASN.1 parsing of certain invalid structures that can result
1587 in a denial of service. (CVE-2006-2937) [Steve Henson]
1589 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1590 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1592 *) Fix SSL client code which could crash if connecting to a
1593 malicious SSLv2 server. (CVE-2006-4343)
1594 [Tavis Ormandy and Will Drewry, Google Security Team]
1596 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1597 match only those. Before that, "AES256-SHA" would be interpreted
1598 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1599 the same strength classification in 0.9.7h) as we currently only
1600 have a single AES bit in the ciphersuite description bitmap.
1601 That change, however, also applied to ciphersuite strings such as
1602 "RC4-MD5" that intentionally matched multiple ciphersuites --
1603 namely, SSL 2.0 ciphersuites in addition to the more common ones
1604 from SSL 3.0/TLS 1.0.
1606 So we change the selection algorithm again: Naming an explicit
1607 ciphersuite selects this one ciphersuite, and any other similar
1608 ciphersuite (same bitmap) from *other* protocol versions.
1609 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1610 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1612 Since SSL 2.0 does not have any ciphersuites for which the
1613 128/256 bit distinction would be relevant, this works for now.
1614 The proper fix will be to use different bits for AES128 and
1615 AES256, which would have avoided the problems from the beginning;
1616 however, bits are scarce, so we can only do this in a new release
1617 (not just a patchlevel) when we can change the SSL_CIPHER
1618 definition to split the single 'unsigned long mask' bitmap into
1619 multiple values to extend the available space.
1623 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1625 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1626 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1628 *) Add AES IGE and biIGE modes.
1631 *) Change the Unix randomness entropy gathering to use poll() when
1632 possible instead of select(), since the latter has some
1633 undesirable limitations.
1634 [Darryl Miles via Richard Levitte and Bodo Moeller]
1636 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1637 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1638 cannot be implicitly activated as part of, e.g., the "AES" alias.
1639 However, please upgrade to OpenSSL 0.9.9[-dev] for
1640 non-experimental use of the ECC ciphersuites to get TLS extension
1641 support, which is required for curve and point format negotiation
1642 to avoid potential handshake problems.
1645 *) Disable rogue ciphersuites:
1647 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1648 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1649 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1651 The latter two were purportedly from
1652 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1655 Also deactivate the remaining ciphersuites from
1656 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1657 unofficial, and the ID has long expired.
1660 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1661 dual-core machines) and other potential thread-safety issues.
1664 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1665 versions), which is now available for royalty-free use
1666 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1667 Also, add Camellia TLS ciphersuites from RFC 4132.
1669 To minimize changes between patchlevels in the OpenSSL 0.9.8
1670 series, Camellia remains excluded from compilation unless OpenSSL
1671 is configured with 'enable-camellia'.
1674 *) Disable the padding bug check when compression is in use. The padding
1675 bug check assumes the first packet is of even length, this is not
1676 necessarily true if compresssion is enabled and can result in false
1677 positives causing handshake failure. The actual bug test is ancient
1678 code so it is hoped that implementations will either have fixed it by
1679 now or any which still have the bug do not support compression.
1682 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1684 *) When applying a cipher rule check to see if string match is an explicit
1685 cipher suite and only match that one cipher suite if it is.
1688 *) Link in manifests for VC++ if needed.
1689 [Austin Ziegler <halostatue@gmail.com>]
1691 *) Update support for ECC-based TLS ciphersuites according to
1692 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1693 TLS extensions, which are supported starting with the 0.9.9
1694 branch, not in the OpenSSL 0.9.8 branch).
1697 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1698 opaque EVP_CIPHER_CTX handling.
1701 *) Fixes and enhancements to zlib compression code. We now only use
1702 "zlib1.dll" and use the default __cdecl calling convention on Win32
1703 to conform with the standards mentioned here:
1704 http://www.zlib.net/DLL_FAQ.txt
1705 Static zlib linking now works on Windows and the new --with-zlib-include
1706 --with-zlib-lib options to Configure can be used to supply the location
1707 of the headers and library. Gracefully handle case where zlib library
1711 *) Several fixes and enhancements to the OID generation code. The old code
1712 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1713 handle numbers larger than ULONG_MAX, truncated printing and had a
1714 non standard OBJ_obj2txt() behaviour.
1717 *) Add support for building of engines under engine/ as shared libraries
1718 under VC++ build system.
1721 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1722 Hopefully, we will not see any false combination of paths any more.
1725 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1727 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1728 (part of SSL_OP_ALL). This option used to disable the
1729 countermeasure against man-in-the-middle protocol-version
1730 rollback in the SSL 2.0 server implementation, which is a bad
1731 idea. (CVE-2005-2969)
1733 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1734 for Information Security, National Institute of Advanced Industrial
1735 Science and Technology [AIST], Japan)]
1737 *) Add two function to clear and return the verify parameter flags.
1740 *) Keep cipherlists sorted in the source instead of sorting them at
1741 runtime, thus removing the need for a lock.
1744 *) Avoid some small subgroup attacks in Diffie-Hellman.
1745 [Nick Mathewson and Ben Laurie]
1747 *) Add functions for well-known primes.
1750 *) Extended Windows CE support.
1751 [Satoshi Nakamura and Andy Polyakov]
1753 *) Initialize SSL_METHOD structures at compile time instead of during
1754 runtime, thus removing the need for a lock.
1757 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1758 attempting to decrypt each encrypted key in turn. Add support to
1762 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1764 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1767 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1770 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1771 key into the same file any more.
1774 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1777 *) Add -utf8 command line and config file option to 'ca'.
1778 [Stefan <stf@udoma.org]
1780 *) Removed the macro des_crypt(), as it seems to conflict with some
1781 libraries. Use DES_crypt().
1784 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1785 involves renaming the source and generated shared-libs for
1786 both. The engines will accept the corrected or legacy ids
1787 ('ncipher' and '4758_cca' respectively) when binding. NB,
1788 this only applies when building 'shared'.
1789 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1791 *) Add attribute functions to EVP_PKEY structure. Modify
1792 PKCS12_create() to recognize a CSP name attribute and
1793 use it. Make -CSP option work again in pkcs12 utility.
1796 *) Add new functionality to the bn blinding code:
1797 - automatic re-creation of the BN_BLINDING parameters after
1798 a fixed number of uses (currently 32)
1799 - add new function for parameter creation
1800 - introduce flags to control the update behaviour of the
1801 BN_BLINDING parameters
1802 - hide BN_BLINDING structure
1803 Add a second BN_BLINDING slot to the RSA structure to improve
1804 performance when a single RSA object is shared among several
1808 *) Add support for DTLS.
1809 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1811 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1812 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1815 *) Remove buggy and incompletet DH cert support from
1816 ssl/ssl_rsa.c and ssl/s3_both.c
1819 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1820 the apps/openssl applications.
1823 *) Compile clean with "-Wall -Wmissing-prototypes
1824 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1825 DEBUG_SAFESTACK must also be set.
1828 *) Change ./Configure so that certain algorithms can be disabled by default.
1829 The new counterpiece to "no-xxx" is "enable-xxx".
1831 The patented RC5 and MDC2 algorithms will now be disabled unless
1832 "enable-rc5" and "enable-mdc2", respectively, are specified.
1834 (IDEA remains enabled despite being patented. This is because IDEA
1835 is frequently required for interoperability, and there is no license
1836 fee for non-commercial use. As before, "no-idea" can be used to
1837 avoid this algorithm.)
1841 *) Add processing of proxy certificates (see RFC 3820). This work was
1842 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1843 EGEE (Enabling Grids for E-science in Europe).
1846 *) RC4 performance overhaul on modern architectures/implementations, such
1847 as Intel P4, IA-64 and AMD64.
1850 *) New utility extract-section.pl. This can be used specify an alternative
1851 section number in a pod file instead of having to treat each file as
1852 a separate case in Makefile. This can be done by adding two lines to the
1855 =for comment openssl_section:XXX
1857 The blank line is mandatory.
1861 *) New arguments -certform, -keyform and -pass for s_client and s_server
1862 to allow alternative format key and certificate files and passphrase
1866 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1867 update associated structures and add various utility functions.
1869 Add new policy related verify parameters, include policy checking in
1870 standard verify code. Enhance 'smime' application with extra parameters
1871 to support policy checking and print out.
1874 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1875 Nehemiah processors. These extensions support AES encryption in hardware
1876 as well as RNG (though RNG support is currently disabled).
1877 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1879 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1882 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1883 [Andy Polyakov and a number of other people]
1885 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1886 implementation contributed by IBM.
1887 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1889 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1890 exponent rather than 'unsigned long'. There is a corresponding change to
1891 the new 'rsa_keygen' element of the RSA_METHOD structure.
1892 [Jelte Jansen, Geoff Thorpe]
1894 *) Functionality for creating the initial serial number file is now
1895 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1897 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1898 number file to 1, which is bound to cause problems. To avoid
1899 the problems while respecting compatibility between different 0.9.7
1900 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1901 CA.pl for serial number initialization. With the new release 0.9.8,
1902 we can fix the problem directly in the 'ca' utility.)
1905 *) Reduced header interdepencies by declaring more opaque objects in
1906 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1907 give fewer recursive includes, which could break lazy source code - so
1908 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1909 developers should define this symbol when building and using openssl to
1910 ensure they track the recommended behaviour, interfaces, [etc], but
1911 backwards-compatible behaviour prevails when this isn't defined.
1914 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1917 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1918 This will generate a random key of the appropriate length based on the
1919 cipher context. The EVP_CIPHER can provide its own random key generation
1920 routine to support keys of a specific form. This is used in the des and
1921 3des routines to generate a key of the correct parity. Update S/MIME
1922 code to use new functions and hence generate correct parity DES keys.
1923 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1924 valid (weak or incorrect parity).
1927 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1928 as looking them up. This is useful when the verified structure may contain
1929 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1930 present unless the new PKCS7_NO_CRL flag is asserted.
1933 *) Extend ASN1 oid configuration module. It now additionally accepts the
1936 shortName = some long name, 1.2.3.4
1939 *) Reimplemented the BN_CTX implementation. There is now no more static
1940 limitation on the number of variables it can handle nor the depth of the
1941 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1942 information can now expand as required, and rather than having a single
1943 static array of bignums, BN_CTX now uses a linked-list of such arrays
1944 allowing it to expand on demand whilst maintaining the usefulness of
1945 BN_CTX's "bundling".
1948 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1949 to allow all RSA operations to function using a single BN_CTX.
1952 *) Preliminary support for certificate policy evaluation and checking. This
1953 is initially intended to pass the tests outlined in "Conformance Testing
1954 of Relying Party Client Certificate Path Processing Logic" v1.07.
1957 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1958 remained unused and not that useful. A variety of other little bignum
1959 tweaks and fixes have also been made continuing on from the audit (see
1963 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1964 associated ASN1, EVP and SSL functions and old ASN1 macros.
1967 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1968 and this should never fail. So the return value from the use of
1969 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1970 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1973 *) BN_CTX_get() should return zero-valued bignums, providing the same
1974 initialised value as BN_new().
1975 [Geoff Thorpe, suggested by Ulf Möller]
1977 *) Support for inhibitAnyPolicy certificate extension.
1980 *) An audit of the BIGNUM code is underway, for which debugging code is
1981 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1982 is considered valid when processing BIGNUMs, and causes execution to
1983 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1984 further steps are taken to deliberately pollute unused data in BIGNUM
1985 structures to try and expose faulty code further on. For now, openssl will
1986 (in its default mode of operation) continue to tolerate the inconsistent
1987 forms that it has tolerated in the past, but authors and packagers should
1988 consider trying openssl and their own applications when compiled with
1989 these debugging symbols defined. It will help highlight potential bugs in
1990 their own code, and will improve the test coverage for OpenSSL itself. At
1991 some point, these tighter rules will become openssl's default to improve
1992 maintainability, though the assert()s and other overheads will remain only
1993 in debugging configurations. See bn.h for more details.
1994 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1996 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1997 that can only be obtained through BN_CTX_new() (which implicitly
1998 initialises it). The presence of this function only made it possible
1999 to overwrite an existing structure (and cause memory leaks).
2002 *) Because of the callback-based approach for implementing LHASH as a
2003 template type, lh_insert() adds opaque objects to hash-tables and
2004 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2005 to clean up those corresponding objects before destroying the hash table
2006 (and losing the object pointers). So some over-zealous constifications in
2007 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2008 objects as "const" and the lh_doall[_arg] callback wrappers are not
2009 prototyped to have "const" restrictions on the object pointers they are
2010 given (and so aren't required to cast them away any more).
2013 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2014 (speed) prefers to use its own implementation. The two implementations
2015 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2016 its object type properly exposed (MS_TM) instead of casting to/from "char
2017 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2018 aren't necessarily the greatest nomenclatures - but this is what was used
2019 internally to the implementation so I've used that for now.
2022 *) Ensure that deprecated functions do not get compiled when
2023 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2024 the self-tests were still using deprecated key-generation functions so
2025 these have been updated also.
2028 *) Reorganise PKCS#7 code to separate the digest location functionality
2029 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2030 New function PKCS7_set_digest() to set the digest type for PKCS#7
2031 digestedData type. Add additional code to correctly generate the
2032 digestedData type and add support for this type in PKCS7 initialization
2036 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2037 structure of type "other".
2040 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2041 sure the loop does correctly stop and breaking ("division by zero")
2042 modulus operations are not performed. The (pre-generated) prime
2043 table crypto/bn/bn_prime.h was already correct, but it could not be
2044 re-generated on some platforms because of the "division by zero"
2045 situation in the script.
2046 [Ralf S. Engelschall]
2048 *) Update support for ECC-based TLS ciphersuites according to
2049 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2050 SHA-1 now is only used for "small" curves (where the
2051 representation of a field element takes up to 24 bytes); for
2052 larger curves, the field element resulting from ECDH is directly
2053 used as premaster secret.
2054 [Douglas Stebila (Sun Microsystems Laboratories)]
2056 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2057 curve secp160r1 to the tests.
2058 [Douglas Stebila (Sun Microsystems Laboratories)]
2060 *) Add the possibility to load symbols globally with DSO.
2061 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2063 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2064 control of the error stack.
2067 *) Add support for STORE in ENGINE.
2070 *) Add the STORE type. The intention is to provide a common interface
2071 to certificate and key stores, be they simple file-based stores, or
2072 HSM-type store, or LDAP stores, or...
2073 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2076 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2077 pass a list of arguments to any function as well as provide a way
2078 for a function to pass data back to the caller.
2081 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2082 works like BUF_strdup() but can be used to duplicate a portion of
2083 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2087 *) Add the function sk_find_ex() which works like sk_find(), but will
2088 return an index to an element even if an exact match couldn't be
2089 found. The index is guaranteed to point at the element where the
2090 searched-for key would be inserted to preserve sorting order.
2093 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2094 takes an extra flags argument for optional functionality. Currently,
2095 the following flags are defined:
2097 OBJ_BSEARCH_VALUE_ON_NOMATCH
2098 This one gets OBJ_bsearch_ex() to return a pointer to the first
2099 element where the comparing function returns a negative or zero
2102 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2103 This one gets OBJ_bsearch_ex() to return a pointer to the first
2104 element where the comparing function returns zero. This is useful
2105 if there are more than one element where the comparing function
2109 *) Make it possible to create self-signed certificates with 'openssl ca'
2110 in such a way that the self-signed certificate becomes part of the
2111 CA database and uses the same mechanisms for serial number generation
2112 as all other certificate signing. The new flag '-selfsign' enables
2113 this functionality. Adapt CA.sh and CA.pl.in.
2116 *) Add functionality to check the public key of a certificate request
2117 against a given private. This is useful to check that a certificate
2118 request can be signed by that key (self-signing).
2121 *) Make it possible to have multiple active certificates with the same
2122 subject in the CA index file. This is done only if the keyword
2123 'unique_subject' is set to 'no' in the main CA section (default
2124 if 'CA_default') of the configuration file. The value is saved
2125 with the database itself in a separate index attribute file,
2126 named like the index file with '.attr' appended to the name.
2129 *) Generate muti valued AVAs using '+' notation in config files for
2133 *) Support for nameConstraints certificate extension.
2136 *) Support for policyConstraints certificate extension.
2139 *) Support for policyMappings certificate extension.
2142 *) Make sure the default DSA_METHOD implementation only uses its
2143 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2144 and change its own handlers to be NULL so as to remove unnecessary
2145 indirection. This lets alternative implementations fallback to the
2146 default implementation more easily.
2149 *) Support for directoryName in GeneralName related extensions
2153 *) Make it possible to link applications using Makefile.shared.
2154 Make that possible even when linking against static libraries!
2157 *) Support for single pass processing for S/MIME signing. This now
2158 means that S/MIME signing can be done from a pipe, in addition
2159 cleartext signing (multipart/signed type) is effectively streaming
2160 and the signed data does not need to be all held in memory.
2162 This is done with a new flag PKCS7_STREAM. When this flag is set
2163 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2164 is done after the data is output (and digests calculated) in
2165 SMIME_write_PKCS7().
2168 *) Add full support for -rpath/-R, both in shared libraries and
2169 applications, at least on the platforms where it's known how
2173 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2174 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2175 will now compute a table of multiples of the generator that
2176 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2177 faster (notably in the case of a single point multiplication,
2178 scalar * generator).
2179 [Nils Larsch, Bodo Moeller]
2181 *) IPv6 support for certificate extensions. The various extensions
2182 which use the IP:a.b.c.d can now take IPv6 addresses using the
2183 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2187 *) Added an ENGINE that implements RSA by performing private key
2188 exponentiations with the GMP library. The conversions to and from
2189 GMP's mpz_t format aren't optimised nor are any montgomery forms
2190 cached, and on x86 it appears OpenSSL's own performance has caught up.
2191 However there are likely to be other architectures where GMP could
2192 provide a boost. This ENGINE is not built in by default, but it can be
2193 specified at Configure time and should be accompanied by the necessary
2194 linker additions, eg;
2195 ./config -DOPENSSL_USE_GMP -lgmp
2198 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2199 testing availability of engines with "-t" - the old behaviour is
2200 produced by increasing the feature's verbosity with "-tt".
2203 *) ECDSA routines: under certain error conditions uninitialized BN objects
2204 could be freed. Solution: make sure initialization is performed early
2205 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2209 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2210 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2211 software implementations. For DSA and DH, parameter generation can
2212 also be overriden by providing the appropriate method callbacks.
2215 *) Change the "progress" mechanism used in key-generation and
2216 primality testing to functions that take a new BN_GENCB pointer in
2217 place of callback/argument pairs. The new API functions have "_ex"
2218 postfixes and the older functions are reimplemented as wrappers for
2219 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2220 declarations of the old functions to help (graceful) attempts to
2221 migrate to the new functions. Also, the new key-generation API
2222 functions operate on a caller-supplied key-structure and return
2223 success/failure rather than returning a key or NULL - this is to
2224 help make "keygen" another member function of RSA_METHOD etc.
2226 Example for using the new callback interface:
2228 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2232 BN_GENCB_set(&my_cb, my_callback, my_arg);
2234 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2235 /* For the meaning of a, b in calls to my_callback(), see the
2236 * documentation of the function that calls the callback.
2237 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2238 * my_callback should return 1 if it wants BN_is_prime_ex()
2239 * to continue, or 0 to stop.
2244 *) Change the ZLIB compression method to be stateful, and make it
2245 available to TLS with the number defined in
2246 draft-ietf-tls-compression-04.txt.
2249 *) Add the ASN.1 structures and functions for CertificatePair, which
2250 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2252 CertificatePair ::= SEQUENCE {
2253 forward [0] Certificate OPTIONAL,
2254 reverse [1] Certificate OPTIONAL,
2255 -- at least one of the pair shall be present -- }
2257 Also implement the PEM functions to read and write certificate
2258 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2260 This needed to be defined, mostly for the sake of the LDAP
2261 attribute crossCertificatePair, but may prove useful elsewhere as
2265 *) Make it possible to inhibit symlinking of shared libraries in
2266 Makefile.shared, for Cygwin's sake.
2269 *) Extend the BIGNUM API by creating a function
2270 void BN_set_negative(BIGNUM *a, int neg);
2271 and a macro that behave like
2272 int BN_is_negative(const BIGNUM *a);
2274 to avoid the need to access 'a->neg' directly in applications.
2277 *) Implement fast modular reduction for pseudo-Mersenne primes
2278 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2279 EC_GROUP_new_curve_GFp() will now automatically use this
2281 [Nils Larsch <nla@trustcenter.de>]
2283 *) Add new lock type (CRYPTO_LOCK_BN).
2286 *) Change the ENGINE framework to automatically load engines
2287 dynamically from specific directories unless they could be
2288 found to already be built in or loaded. Move all the
2289 current engines except for the cryptodev one to a new
2291 The engines in engines/ are built as shared libraries if
2292 the "shared" options was given to ./Configure or ./config.
2293 Otherwise, they are inserted in libcrypto.a.
2294 /usr/local/ssl/engines is the default directory for dynamic
2295 engines, but that can be overriden at configure time through
2296 the usual use of --prefix and/or --openssldir, and at run
2297 time with the environment variable OPENSSL_ENGINES.
2298 [Geoff Thorpe and Richard Levitte]
2300 *) Add Makefile.shared, a helper makefile to build shared
2301 libraries. Addapt Makefile.org.
2304 *) Add version info to Win32 DLLs.
2305 [Peter 'Luna' Runestig" <peter@runestig.com>]
2307 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2308 can be added using this API to created arbitrary PKCS#12
2309 files while avoiding the low level API.
2311 New options to PKCS12_create(), key or cert can be NULL and
2312 will then be omitted from the output file. The encryption
2313 algorithm NIDs can be set to -1 for no encryption, the mac
2314 iteration count can be set to 0 to omit the mac.
2316 Enhance pkcs12 utility by making the -nokeys and -nocerts
2317 options work when creating a PKCS#12 file. New option -nomac
2318 to omit the mac, NONE can be set for an encryption algorithm.
2319 New code is modified to use the enhanced PKCS12_create()
2320 instead of the low level API.
2323 *) Extend ASN1 encoder to support indefinite length constructed
2324 encoding. This can output sequences tags and octet strings in
2325 this form. Modify pk7_asn1.c to support indefinite length
2326 encoding. This is experimental and needs additional code to
2327 be useful, such as an ASN1 bio and some enhanced streaming
2330 Extend template encode functionality so that tagging is passed
2331 down to the template encoder.
2334 *) Let 'openssl req' fail if an argument to '-newkey' is not
2335 recognized instead of using RSA as a default.
2338 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2339 As these are not official, they are not included in "ALL";
2340 the "ECCdraft" ciphersuite group alias can be used to select them.
2341 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2343 *) Add ECDH engine support.
2344 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2346 *) Add ECDH in new directory crypto/ecdh/.
2347 [Douglas Stebila (Sun Microsystems Laboratories)]
2349 *) Let BN_rand_range() abort with an error after 100 iterations
2350 without success (which indicates a broken PRNG).
2353 *) Change BN_mod_sqrt() so that it verifies that the input value
2354 is really the square of the return value. (Previously,
2355 BN_mod_sqrt would show GIGO behaviour.)
2358 *) Add named elliptic curves over binary fields from X9.62, SECG,
2359 and WAP/WTLS; add OIDs that were still missing.
2361 [Sheueling Chang Shantz and Douglas Stebila
2362 (Sun Microsystems Laboratories)]
2364 *) Extend the EC library for elliptic curves over binary fields
2365 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2368 EC_GF2m_simple_method
2372 EC_GROUP_new_curve_GF2m
2373 EC_GROUP_set_curve_GF2m
2374 EC_GROUP_get_curve_GF2m
2375 EC_POINT_set_affine_coordinates_GF2m
2376 EC_POINT_get_affine_coordinates_GF2m
2377 EC_POINT_set_compressed_coordinates_GF2m
2379 Point compression for binary fields is disabled by default for
2380 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2383 As binary polynomials are represented as BIGNUMs, various members
2384 of the EC_GROUP and EC_POINT data structures can be shared
2385 between the implementations for prime fields and binary fields;
2386 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2387 are essentially identical to their ..._GFp counterparts.
2388 (For simplicity, the '..._GFp' prefix has been dropped from
2389 various internal method names.)
2391 An internal 'field_div' method (similar to 'field_mul' and
2392 'field_sqr') has been added; this is used only for binary fields.
2394 [Sheueling Chang Shantz and Douglas Stebila
2395 (Sun Microsystems Laboratories)]
2397 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2398 through methods ('mul', 'precompute_mult').
2400 The generic implementations (now internally called 'ec_wNAF_mul'
2401 and 'ec_wNAF_precomputed_mult') remain the default if these
2402 methods are undefined.
2404 [Sheueling Chang Shantz and Douglas Stebila
2405 (Sun Microsystems Laboratories)]
2407 *) New function EC_GROUP_get_degree, which is defined through
2408 EC_METHOD. For curves over prime fields, this returns the bit
2409 length of the modulus.
2411 [Sheueling Chang Shantz and Douglas Stebila
2412 (Sun Microsystems Laboratories)]
2414 *) New functions EC_GROUP_dup, EC_POINT_dup.
2415 (These simply call ..._new and ..._copy).
2417 [Sheueling Chang Shantz and Douglas Stebila
2418 (Sun Microsystems Laboratories)]
2420 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2421 Polynomials are represented as BIGNUMs (where the sign bit is not
2422 used) in the following functions [macros]:
2425 BN_GF2m_sub [= BN_GF2m_add]
2426 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2427 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2428 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2430 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2431 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2432 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2433 BN_GF2m_cmp [= BN_ucmp]
2435 (Note that only the 'mod' functions are actually for fields GF(2^m).
2436 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2438 For some functions, an the irreducible polynomial defining a
2439 field can be given as an 'unsigned int[]' with strictly
2440 decreasing elements giving the indices of those bits that are set;
2441 i.e., p[] represents the polynomial
2442 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2444 p[0] > p[1] > ... > p[k] = 0.
2445 This applies to the following functions:
2450 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2451 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2453 BN_GF2m_mod_sqrt_arr
2454 BN_GF2m_mod_solve_quad_arr
2458 Conversion can be performed by the following functions:
2463 bntest.c has additional tests for binary polynomial arithmetic.
2465 Two implementations for BN_GF2m_mod_div() are available.
2466 The default algorithm simply uses BN_GF2m_mod_inv() and
2467 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2468 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2469 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2471 [Sheueling Chang Shantz and Douglas Stebila
2472 (Sun Microsystems Laboratories)]
2474 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2475 functionality is disabled at compile-time.
2476 [Douglas Stebila <douglas.stebila@sun.com>]
2478 *) Change default behaviour of 'openssl asn1parse' so that more
2479 information is visible when viewing, e.g., a certificate:
2481 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2482 mode the content of non-printable OCTET STRINGs is output in a
2483 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2484 avoid the appearance of a printable string.
2485 [Nils Larsch <nla@trustcenter.de>]
2487 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2489 EC_GROUP_set_asn1_flag()
2490 EC_GROUP_get_asn1_flag()
2491 EC_GROUP_set_point_conversion_form()
2492 EC_GROUP_get_point_conversion_form()
2493 These control ASN1 encoding details:
2494 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2495 has been set to OPENSSL_EC_NAMED_CURVE.
2496 - Points are encoded in uncompressed form by default; options for
2497 asn1_for are as for point2oct, namely
2498 POINT_CONVERSION_COMPRESSED
2499 POINT_CONVERSION_UNCOMPRESSED
2500 POINT_CONVERSION_HYBRID
2502 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2505 EC_GROUP_get0_seed()
2506 EC_GROUP_get_seed_len()
2507 This is used only for ASN1 purposes (so far).
2508 [Nils Larsch <nla@trustcenter.de>]
2510 *) Add 'field_type' member to EC_METHOD, which holds the NID
2511 of the appropriate field type OID. The new function
2512 EC_METHOD_get_field_type() returns this value.
2513 [Nils Larsch <nla@trustcenter.de>]
2518 EC_POINT_point2hex()
2519 EC_POINT_hex2point()
2520 providing useful interfaces to EC_POINT_point2oct() and
2521 EC_POINT_oct2point().
2522 [Nils Larsch <nla@trustcenter.de>]
2524 *) Change internals of the EC library so that the functions
2525 EC_GROUP_set_generator()
2526 EC_GROUP_get_generator()
2527 EC_GROUP_get_order()
2528 EC_GROUP_get_cofactor()
2529 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2530 to methods, which would lead to unnecessary code duplication when
2531 adding different types of curves.
2532 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2534 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2535 arithmetic, and such that modified wNAFs are generated
2536 (which avoid length expansion in many cases).
2539 *) Add a function EC_GROUP_check_discriminant() (defined via
2540 EC_METHOD) that verifies that the curve discriminant is non-zero.
2542 Add a function EC_GROUP_check() that makes some sanity tests
2543 on a EC_GROUP, its generator and order. This includes
2544 EC_GROUP_check_discriminant().
2545 [Nils Larsch <nla@trustcenter.de>]
2547 *) Add ECDSA in new directory crypto/ecdsa/.
2549 Add applications 'openssl ecparam' and 'openssl ecdsa'
2550 (these are based on 'openssl dsaparam' and 'openssl dsa').
2552 ECDSA support is also included in various other files across the
2553 library. Most notably,
2554 - 'openssl req' now has a '-newkey ecdsa:file' option;
2555 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2556 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2557 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2558 them suitable for ECDSA where domain parameters must be
2559 extracted before the specific public key;
2560 - ECDSA engine support has been added.
2561 [Nils Larsch <nla@trustcenter.de>]
2563 *) Include some named elliptic curves, and add OIDs from X9.62,
2564 SECG, and WAP/WTLS. Each curve can be obtained from the new
2566 EC_GROUP_new_by_curve_name(),
2567 and the list of available named curves can be obtained with
2568 EC_get_builtin_curves().
2569 Also add a 'curve_name' member to EC_GROUP objects, which can be
2571 EC_GROUP_set_curve_name()
2572 EC_GROUP_get_curve_name()
2573 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2575 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2576 was actually never needed) and in BN_mul(). The removal in BN_mul()
2577 required a small change in bn_mul_part_recursive() and the addition
2578 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2579 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2580 bn_sub_words() and bn_add_words() except they take arrays with
2584 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2586 *) Cleanse PEM buffers before freeing them since they may contain
2588 [Benjamin Bennett <ben@psc.edu>]
2590 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2591 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2592 authentication-only ciphersuites.
2595 *) Since AES128 and AES256 share a single mask bit in the logic of
2596 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2597 kludge to work properly if AES128 is available and AES256 isn't.
2600 *) Expand security boundary to match 1.1.1 module.
2603 *) Remove redundant features: hash file source, editing of test vectors
2604 modify fipsld to use external fips_premain.c signature.
2607 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2608 run algorithm test programs.
2611 *) Make algorithm test programs more tolerant of whitespace.
2614 *) Have SSL/TLS server implementation tolerate "mismatched" record
2615 protocol version while receiving ClientHello even if the
2616 ClientHello is fragmented. (The server can't insist on the
2617 particular protocol version it has chosen before the ServerHello
2618 message has informed the client about his choice.)
2621 *) Load error codes if they are not already present instead of using a
2622 static variable. This allows them to be cleanly unloaded and reloaded.
2625 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2627 *) Introduce limits to prevent malicious keys being able to
2628 cause a denial of service. (CVE-2006-2940)
2629 [Steve Henson, Bodo Moeller]
2631 *) Fix ASN.1 parsing of certain invalid structures that can result
2632 in a denial of service. (CVE-2006-2937) [Steve Henson]
2634 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2635 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2637 *) Fix SSL client code which could crash if connecting to a
2638 malicious SSLv2 server. (CVE-2006-4343)
2639 [Tavis Ormandy and Will Drewry, Google Security Team]
2641 *) Change ciphersuite string processing so that an explicit
2642 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2643 will no longer include "AES128-SHA"), and any other similar
2644 ciphersuite (same bitmap) from *other* protocol versions (so that
2645 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2646 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2647 changes from 0.9.8b and 0.9.8d.
2650 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2652 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2653 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2655 *) Change the Unix randomness entropy gathering to use poll() when
2656 possible instead of select(), since the latter has some
2657 undesirable limitations.
2658 [Darryl Miles via Richard Levitte and Bodo Moeller]
2660 *) Disable rogue ciphersuites:
2662 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2663 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2664 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2666 The latter two were purportedly from
2667 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2670 Also deactive the remaining ciphersuites from
2671 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2672 unofficial, and the ID has long expired.
2675 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2676 dual-core machines) and other potential thread-safety issues.
2679 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2681 *) Adapt fipsld and the build system to link against the validated FIPS
2682 module in FIPS mode.
2685 *) Fixes for VC++ 2005 build under Windows.
2688 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2689 from a Windows bash shell such as MSYS. It is autodetected from the
2690 "config" script when run from a VC++ environment. Modify standard VC++
2691 build to use fipscanister.o from the GNU make build.
2694 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2696 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2697 The value now differs depending on if you build for FIPS or not.
2698 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2699 safely run with a non-FIPSed libcrypto, as it may crash because of
2700 the difference induced by this change.
2703 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2705 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2706 (part of SSL_OP_ALL). This option used to disable the
2707 countermeasure against man-in-the-middle protocol-version
2708 rollback in the SSL 2.0 server implementation, which is a bad
2709 idea. (CVE-2005-2969)
2711 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2712 for Information Security, National Institute of Advanced Industrial
2713 Science and Technology [AIST], Japan)]
2715 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2716 mainly for FIPS compliance and not fully integrated at this stage.
2719 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2720 the exponentiation using a fixed-length exponent. (Otherwise,
2721 the information leaked through timing could expose the secret key
2722 after many signatures; cf. Bleichenbacher's attack on DSA with
2726 *) Make a new fixed-window mod_exp implementation the default for
2727 RSA, DSA, and DH private-key operations so that the sequence of
2728 squares and multiplies and the memory access pattern are
2729 independent of the particular secret key. This will mitigate
2730 cache-timing and potential related attacks.
2732 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2733 and this is automatically used by BN_mod_exp_mont() if the new flag
2734 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2735 will use this BN flag for private exponents unless the flag
2736 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2737 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2739 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2741 *) Change the client implementation for SSLv23_method() and
2742 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2743 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2744 (Previously, the SSL 2.0 backwards compatible Client Hello
2745 message format would be used even with SSL_OP_NO_SSLv2.)
2748 *) Add support for smime-type MIME parameter in S/MIME messages which some
2752 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2753 a threadsafe manner. Modify rsa code to use new function and add calls
2754 to dsa and dh code (which had race conditions before).
2757 *) Include the fixed error library code in the C error file definitions
2758 instead of fixing them up at runtime. This keeps the error code
2759 structures constant.
2762 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2764 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2767 *) Fixes for newer kerberos headers. NB: the casts are needed because
2768 the 'length' field is signed on one version and unsigned on another
2769 with no (?) obvious way to tell the difference, without these VC++
2770 complains. Also the "definition" of FAR (blank) is no longer included
2771 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2772 some needed definitions.
2775 *) Undo Cygwin change.
2778 *) Added support for proxy certificates according to RFC 3820.
2779 Because they may be a security thread to unaware applications,
2780 they must be explicitely allowed in run-time. See
2781 docs/HOWTO/proxy_certificates.txt for further information.
2784 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2786 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2787 server and client random values. Previously
2788 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2789 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2791 This change has negligible security impact because:
2793 1. Server and client random values still have 24 bytes of pseudo random
2796 2. Server and client random values are sent in the clear in the initial
2799 3. The master secret is derived using the premaster secret (48 bytes in
2800 size for static RSA ciphersuites) as well as client server and random
2803 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2806 [Stephen Henson, reported by UK NISCC]
2808 *) Use Windows randomness collection on Cygwin.
2811 *) Fix hang in EGD/PRNGD query when communication socket is closed
2812 prematurely by EGD/PRNGD.
2813 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2815 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2818 *) Back-port of selected performance improvements from development
2819 branch, as well as improved support for PowerPC platforms.
2822 *) Add lots of checks for memory allocation failure, error codes to indicate
2823 failure and freeing up memory if a failure occurs.
2824 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2826 *) Add new -passin argument to dgst.
2829 *) Perform some character comparisons of different types in X509_NAME_cmp:
2830 this is needed for some certificates that reencode DNs into UTF8Strings
2831 (in violation of RFC3280) and can't or wont issue name rollover
2835 *) Make an explicit check during certificate validation to see that
2836 the CA setting in each certificate on the chain is correct. As a
2837 side effect always do the following basic checks on extensions,
2838 not just when there's an associated purpose to the check:
2840 - if there is an unhandled critical extension (unless the user
2841 has chosen to ignore this fault)
2842 - if the path length has been exceeded (if one is set at all)
2843 - that certain extensions fit the associated purpose (if one has
2847 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2849 *) Avoid a race condition when CRLs are checked in a multi threaded
2850 environment. This would happen due to the reordering of the revoked
2851 entries during signature checking and serial number lookup. Now the
2852 encoding is cached and the serial number sort performed under a lock.
2853 Add new STACK function sk_is_sorted().
2856 *) Add Delta CRL to the extension code.
2859 *) Various fixes to s3_pkt.c so alerts are sent properly.
2860 [David Holmes <d.holmes@f5.com>]
2862 *) Reduce the chances of duplicate issuer name and serial numbers (in
2863 violation of RFC3280) using the OpenSSL certificate creation utilities.
2864 This is done by creating a random 64 bit value for the initial serial
2865 number when a serial number file is created or when a self signed
2866 certificate is created using 'openssl req -x509'. The initial serial
2867 number file is created using 'openssl x509 -next_serial' in CA.pl
2868 rather than being initialized to 1.
2871 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2873 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2874 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2875 [Joe Orton, Steve Henson]
2877 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2879 [Joe Orton, Steve Henson]
2881 *) Make it possible to have multiple active certificates with the same
2882 subject in the CA index file. This is done only if the keyword
2883 'unique_subject' is set to 'no' in the main CA section (default
2884 if 'CA_default') of the configuration file. The value is saved
2885 with the database itself in a separate index attribute file,
2886 named like the index file with '.attr' appended to the name.
2889 *) X509 verify fixes. Disable broken certificate workarounds when
2890 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2891 keyUsage extension present. Don't accept CRLs with unhandled critical
2892 extensions: since verify currently doesn't process CRL extensions this
2893 rejects a CRL with *any* critical extensions. Add new verify error codes
2897 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2898 A clarification of RFC2560 will require the use of OCTET STRINGs and
2899 some implementations cannot handle the current raw format. Since OpenSSL
2900 copies and compares OCSP nonces as opaque blobs without any attempt at
2901 parsing them this should not create any compatibility issues.
2904 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2905 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2906 this HMAC (and other) operations are several times slower than OpenSSL
2910 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2911 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2913 *) Use the correct content when signing type "other".
2916 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2918 *) Fix various bugs revealed by running the NISCC test suite:
2920 Stop out of bounds reads in the ASN1 code when presented with
2921 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2923 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2925 If verify callback ignores invalid public key errors don't try to check
2926 certificate signature with the NULL public key.
2930 *) New -ignore_err option in ocsp application to stop the server
2931 exiting on the first error in a request.
2934 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2935 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2939 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2940 extra data after the compression methods not only for TLS 1.0
2941 but also for SSL 3.0 (as required by the specification).
2942 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2944 *) Change X509_certificate_type() to mark the key as exported/exportable
2945 when it's 512 *bits* long, not 512 bytes.
2948 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2949 blocks during encryption.
2952 *) Various fixes to base64 BIO and non blocking I/O. On write
2953 flushes were not handled properly if the BIO retried. On read
2954 data was not being buffered properly and had various logic bugs.
2955 This also affects blocking I/O when the data being decoded is a
2959 *) Various S/MIME bugfixes and compatibility changes:
2960 output correct application/pkcs7 MIME type if
2961 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2962 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2963 of files as .eml work). Correctly handle very long lines in MIME
2967 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2969 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2970 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2971 a protocol version number mismatch like a decryption error
2972 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2975 *) Turn on RSA blinding by default in the default implementation
2976 to avoid a timing attack. Applications that don't want it can call
2977 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2978 They would be ill-advised to do so in most cases.
2979 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2981 *) Change RSA blinding code so that it works when the PRNG is not
2982 seeded (in this case, the secret RSA exponent is abused as
2983 an unpredictable seed -- if it is not unpredictable, there
2984 is no point in blinding anyway). Make RSA blinding thread-safe
2985 by remembering the creator's thread ID in rsa->blinding and
2986 having all other threads use local one-time blinding factors
2987 (this requires more computation than sharing rsa->blinding, but
2988 avoids excessive locking; and if an RSA object is not shared
2989 between threads, blinding will still be very fast).
2992 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2993 ENGINE as defaults for all supported algorithms irrespective of
2994 the 'flags' parameter. 'flags' is now honoured, so applications
2995 should make sure they are passing it correctly.
2998 *) Target "mingw" now allows native Windows code to be generated in
2999 the Cygwin environment as well as with the MinGW compiler.
3002 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3004 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3005 via timing by performing a MAC computation even if incorrrect
3006 block cipher padding has been found. This is a countermeasure
3007 against active attacks where the attacker has to distinguish
3008 between bad padding and a MAC verification error. (CVE-2003-0078)
3010 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3011 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3012 Martin Vuagnoux (EPFL, Ilion)]
3014 *) Make the no-err option work as intended. The intention with no-err
3015 is not to have the whole error stack handling routines removed from
3016 libcrypto, it's only intended to remove all the function name and
3017 reason texts, thereby removing some of the footprint that may not
3018 be interesting if those errors aren't displayed anyway.
3020 NOTE: it's still possible for any application or module to have it's
3021 own set of error texts inserted. The routines are there, just not
3022 used by default when no-err is given.
3025 *) Add support for FreeBSD on IA64.
3026 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3028 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3029 Kerberos function mit_des_cbc_cksum(). Before this change,
3030 the value returned by DES_cbc_cksum() was like the one from
3031 mit_des_cbc_cksum(), except the bytes were swapped.
3032 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3034 *) Allow an application to disable the automatic SSL chain building.
3035 Before this a rather primitive chain build was always performed in
3036 ssl3_output_cert_chain(): an application had no way to send the
3037 correct chain if the automatic operation produced an incorrect result.
3039 Now the chain builder is disabled if either:
3041 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3043 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3045 The reasoning behind this is that an application would not want the
3046 auto chain building to take place if extra chain certificates are
3047 present and it might also want a means of sending no additional
3048 certificates (for example the chain has two certificates and the
3052 *) Add the possibility to build without the ENGINE framework.
3053 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3055 *) Under Win32 gmtime() can return NULL: check return value in
3056 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3059 *) DSA routines: under certain error conditions uninitialized BN objects
3060 could be freed. Solution: make sure initialization is performed early
3061 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3062 Nils Larsch <nla@trustcenter.de> via PR#459)
3065 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3066 checked on reconnect on the client side, therefore session resumption
3067 could still fail with a "ssl session id is different" error. This
3068 behaviour is masked when SSL_OP_ALL is used due to
3069 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3070 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3071 followup to PR #377.
3074 *) IA-32 assembler support enhancements: unified ELF targets, support
3075 for SCO/Caldera platforms, fix for Cygwin shared build.
3078 *) Add support for FreeBSD on sparc64. As a consequence, support for
3079 FreeBSD on non-x86 processors is separate from x86 processors on
3080 the config script, much like the NetBSD support.
3081 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3083 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3085 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3088 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3089 code (06) was taken as the first octet of the session ID and the last
3090 octet was ignored consequently. As a result SSLv2 client side session
3091 caching could not have worked due to the session ID mismatch between
3093 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3097 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3098 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3102 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3103 seems that in spite of existing for more than a year, many application
3104 author have done nothing to provide the necessary callbacks, which
3105 means that this particular engine will not work properly anywhere.
3106 This is a very unfortunate situation which forces us, in the name
3107 of usability, to give the hw_ncipher.c a static lock, which is part
3109 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3110 appear in 0.9.8 or later. We EXPECT application authors to have
3111 dealt properly with this when 0.9.8 is released (unless we actually
3112 make such changes in the libcrypto locking code that changes will
3113 have to be made anyway).
3116 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3117 octets have been read, EOF or an error occurs. Without this change
3118 some truncated ASN1 structures will not produce an error.
3121 *) Disable Heimdal support, since it hasn't been fully implemented.
3122 Still give the possibility to force the use of Heimdal, but with
3123 warnings and a request that patches get sent to openssl-dev.
3126 *) Add the VC-CE target, introduce the WINCE sysname, and add
3127 INSTALL.WCE and appropriate conditionals to make it build.
3128 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3130 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3131 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3132 edit numbers of the version.
3133 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3135 *) Introduce safe string copy and catenation functions
3136 (BUF_strlcpy() and BUF_strlcat()).
3137 [Ben Laurie (CHATS) and Richard Levitte]
3139 *) Avoid using fixed-size buffers for one-line DNs.
3140 [Ben Laurie (CHATS)]
3142 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3143 resizing buffers containing secrets, and use where appropriate.
3144 [Ben Laurie (CHATS)]
3146 *) Avoid using fixed size buffers for configuration file location.
3147 [Ben Laurie (CHATS)]
3149 *) Avoid filename truncation for various CA files.
3150 [Ben Laurie (CHATS)]
3152 *) Use sizeof in preference to magic numbers.
3153 [Ben Laurie (CHATS)]
3155 *) Avoid filename truncation in cert requests.
3156 [Ben Laurie (CHATS)]
3158 *) Add assertions to check for (supposedly impossible) buffer
3160 [Ben Laurie (CHATS)]
3162 *) Don't cache truncated DNS entries in the local cache (this could
3163 potentially lead to a spoofing attack).
3164 [Ben Laurie (CHATS)]
3166 *) Fix various buffers to be large enough for hex/decimal
3167 representations in a platform independent manner.
3168 [Ben Laurie (CHATS)]
3170 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3171 resizing buffers containing secrets, and use where appropriate.
3172 [Ben Laurie (CHATS)]
3174 *) Add BIO_indent() to avoid much slightly worrying code to do
3176 [Ben Laurie (CHATS)]
3178 *) Convert sprintf()/BIO_puts() to BIO_printf().
3179 [Ben Laurie (CHATS)]
3181 *) buffer_gets() could terminate with the buffer only half
3183 [Ben Laurie (CHATS)]
3185 *) Add assertions to prevent user-supplied crypto functions from
3186 overflowing internal buffers by having large block sizes, etc.
3187 [Ben Laurie (CHATS)]
3189 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3191 [Ben Laurie (CHATS)]
3193 *) Eliminate unused copy of key in RC4.
3194 [Ben Laurie (CHATS)]
3196 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3197 [Ben Laurie (CHATS)]
3199 *) Fix off-by-one error in EGD path.
3200 [Ben Laurie (CHATS)]
3202 *) If RANDFILE path is too long, ignore instead of truncating.
3203 [Ben Laurie (CHATS)]
3205 *) Eliminate unused and incorrectly sized X.509 structure
3207 [Ben Laurie (CHATS)]
3209 *) Eliminate unused and dangerous function knumber().
3210 [Ben Laurie (CHATS)]
3212 *) Eliminate unused and dangerous structure, KSSL_ERR.
3213 [Ben Laurie (CHATS)]
3215 *) Protect against overlong session ID context length in an encoded
3216 session object. Since these are local, this does not appear to be
3218 [Ben Laurie (CHATS)]
3220 *) Change from security patch (see 0.9.6e below) that did not affect
3221 the 0.9.6 release series:
3223 Remote buffer overflow in SSL3 protocol - an attacker could
3224 supply an oversized master key in Kerberos-enabled versions.
3226 [Ben Laurie (CHATS)]
3228 *) Change the SSL kerb5 codes to match RFC 2712.
3231 *) Make -nameopt work fully for req and add -reqopt switch.
3232 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3234 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3235 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3237 *) Make sure tests can be performed even if the corresponding algorithms
3238 have been removed entirely. This was also the last step to make
3239 OpenSSL compilable with DJGPP under all reasonable conditions.
3240 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3242 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3243 to allow version independent disabling of normally unselected ciphers,
3244 which may be activated as a side-effect of selecting a single cipher.
3246 (E.g., cipher list string "RSA" enables ciphersuites that are left
3247 out of "ALL" because they do not provide symmetric encryption.
3248 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3249 [Lutz Jaenicke, Bodo Moeller]
3251 *) Add appropriate support for separate platform-dependent build
3252 directories. The recommended way to make a platform-dependent
3253 build directory is the following (tested on Linux), maybe with
3256 # Place yourself outside of the OpenSSL source tree. In
3257 # this example, the environment variable OPENSSL_SOURCE
3258 # is assumed to contain the absolute OpenSSL source directory.
3259 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3260 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3261 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3262 mkdir -p `dirname $F`
3263 ln -s $OPENSSL_SOURCE/$F $F
3266 To be absolutely sure not to disturb the source tree, a "make clean"
3267 is a good thing. If it isn't successfull, don't worry about it,
3268 it probably means the source directory is very clean.
3271 *) Make sure any ENGINE control commands make local copies of string
3272 pointers passed to them whenever necessary. Otherwise it is possible
3273 the caller may have overwritten (or deallocated) the original string
3274 data when a later ENGINE operation tries to use the stored values.
3275 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3277 *) Improve diagnostics in file reading and command-line digests.
3278 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3280 *) Add AES modes CFB and OFB to the object database. Correct an
3281 error in AES-CFB decryption.
3284 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3285 allows existing EVP_CIPHER_CTX structures to be reused after
3286 calling EVP_*Final(). This behaviour is used by encryption
3287 BIOs and some applications. This has the side effect that
3288 applications must explicitly clean up cipher contexts with
3289 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3292 *) Check the values of dna and dnb in bn_mul_recursive before calling
3293 bn_mul_comba (a non zero value means the a or b arrays do not contain
3294 n2 elements) and fallback to bn_mul_normal if either is not zero.
3297 *) Fix escaping of non-ASCII characters when using the -subj option
3298 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3301 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3302 form for "surname", serialNumber has no short form.
3303 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3304 therefore remove "mail" short name for "internet 7".
3305 The OID for unique identifiers in X509 certificates is
3306 x500UniqueIdentifier, not uniqueIdentifier.
3307 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3310 *) Add an "init" command to the ENGINE config module and auto initialize
3311 ENGINEs. Without any "init" command the ENGINE will be initialized
3312 after all ctrl commands have been executed on it. If init=1 the
3313 ENGINE is initailized at that point (ctrls before that point are run
3314 on the uninitialized ENGINE and after on the initialized one). If
3315 init=0 then the ENGINE will not be iniatialized at all.
3318 *) Fix the 'app_verify_callback' interface so that the user-defined
3319 argument is actually passed to the callback: In the
3320 SSL_CTX_set_cert_verify_callback() prototype, the callback
3321 declaration has been changed from
3324 int (*cb)(X509_STORE_CTX *,void *);
3325 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3326 i=s->ctx->app_verify_callback(&ctx)
3327 has been changed into
3328 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3330 To update applications using SSL_CTX_set_cert_verify_callback(),
3331 a dummy argument can be added to their callback functions.
3332 [D. K. Smetters <smetters@parc.xerox.com>]
3334 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3335 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3337 *) Add and OPENSSL_LOAD_CONF define which will cause
3338 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3339 This allows older applications to transparently support certain
3340 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3341 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3342 load the config file and OPENSSL_add_all_algorithms_conf() which will
3343 always load it have also been added.
3346 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3347 Adjust NIDs and EVP layer.
3348 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3350 *) Config modules support in openssl utility.
3352 Most commands now load modules from the config file,
3353 though in a few (such as version) this isn't done
3354 because it couldn't be used for anything.
3356 In the case of ca and req the config file used is
3357 the same as the utility itself: that is the -config
3358 command line option can be used to specify an
3362 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3363 use "openssl_conf" if filename is NULL use default openssl config file.
3366 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3367 config section name. Add a new flag to tolerate a missing config file
3368 and move code to CONF_modules_load_file().
3371 *) Support for crypto accelerator cards from Accelerated Encryption
3372 Processing, www.aep.ie. (Use engine 'aep')
3373 The support was copied from 0.9.6c [engine] and adapted/corrected
3374 to work with the new engine framework.
3375 [AEP Inc. and Richard Levitte]
3377 *) Support for SureWare crypto accelerator cards from Baltimore
3378 Technologies. (Use engine 'sureware')
3379 The support was copied from 0.9.6c [engine] and adapted
3380 to work with the new engine framework.
3383 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3384 make the newer ENGINE framework commands for the CHIL engine work.
3385 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3387 *) Make it possible to produce shared libraries on ReliantUNIX.
3388 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3390 *) Add the configuration target debug-linux-ppro.
3391 Make 'openssl rsa' use the general key loading routines
3392 implemented in apps.c, and make those routines able to
3393 handle the key format FORMAT_NETSCAPE and the variant
3395 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3397 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3398 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3400 *) Add -keyform to rsautl, and document -engine.
3401 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3403 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3404 BIO_R_NO_SUCH_FILE error code rather than the generic
3405 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3408 *) Add new functions
3410 ERR_peek_last_error_line
3411 ERR_peek_last_error_line_data.
3412 These are similar to
3415 ERR_peek_error_line_data,
3416 but report on the latest error recorded rather than the first one
3417 still in the error queue.
3418 [Ben Laurie, Bodo Moeller]
3420 *) default_algorithms option in ENGINE config module. This allows things
3422 default_algorithms = ALL
3423 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3426 *) Prelminary ENGINE config module.
3429 *) New experimental application configuration code.
3432 *) Change the AES code to follow the same name structure as all other
3433 symmetric ciphers, and behave the same way. Move everything to
3434 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3435 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3437 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3438 [Ben Laurie and Theo de Raadt]
3440 *) Add option to output public keys in req command.
3441 [Massimiliano Pala madwolf@openca.org]
3443 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3444 (up to about 10% better than before for P-192 and P-224).
3447 *) New functions/macros
3449 SSL_CTX_set_msg_callback(ctx, cb)
3450 SSL_CTX_set_msg_callback_arg(ctx, arg)
3451 SSL_set_msg_callback(ssl, cb)
3452 SSL_set_msg_callback_arg(ssl, arg)
3454 to request calling a callback function
3456 void cb(int write_p, int version, int content_type,
3457 const void *buf, size_t len, SSL *ssl, void *arg)
3459 whenever a protocol message has been completely received
3460 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3461 protocol version according to which the SSL library interprets
3462 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3463 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3464 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3465 specification (change_cipher_spec(20), alert(21), handshake(22)).
3466 'buf' and 'len' point to the actual message, 'ssl' to the
3467 SSL object, and 'arg' is the application-defined value set by
3468 SSL[_CTX]_set_msg_callback_arg().
3470 'openssl s_client' and 'openssl s_server' have new '-msg' options
3471 to enable a callback that displays all protocol messages.
3474 *) Change the shared library support so shared libraries are built as
3475 soon as the corresponding static library is finished, and thereby get
3476 openssl and the test programs linked against the shared library.
3477 This still only happens when the keyword "shard" has been given to
3478 the configuration scripts.
3480 NOTE: shared library support is still an experimental thing, and
3481 backward binary compatibility is still not guaranteed.
3482 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3484 *) Add support for Subject Information Access extension.
3485 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3487 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3488 additional bytes when new memory had to be allocated, not just
3489 when reusing an existing buffer.
3492 *) New command line and configuration option 'utf8' for the req command.
3493 This allows field values to be specified as UTF8 strings.
3496 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3497 runs for the former and machine-readable output for the latter.
3500 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3501 of the e-mail address in the DN (i.e., it will go into a certificate
3502 extension only). The new configuration file option 'email_in_dn = no'
3503 has the same effect.
3504 [Massimiliano Pala madwolf@openca.org]
3506 *) Change all functions with names starting with des_ to be starting
3507 with DES_ instead. Add wrappers that are compatible with libdes,
3508 but are named _ossl_old_des_*. Finally, add macros that map the
3509 des_* symbols to the corresponding _ossl_old_des_* if libdes
3510 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3511 desired, the des_* symbols will be mapped to DES_*, with one
3514 Since we provide two compatibility mappings, the user needs to
3515 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3516 compatibility is desired. The default (i.e., when that macro
3517 isn't defined) is OpenSSL 0.9.6c compatibility.
3519 There are also macros that enable and disable the support of old
3520 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3521 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3522 are defined, the default will apply: to support the old des routines.
3524 In either case, one must include openssl/des.h to get the correct
3525 definitions. Do not try to just include openssl/des_old.h, that
3528 NOTE: This is a major break of an old API into a new one. Software
3529 authors are encouraged to switch to the DES_ style functions. Some
3530 time in the future, des_old.h and the libdes compatibility functions
3531 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3532 default), and then completely removed.
3535 *) Test for certificates which contain unsupported critical extensions.
3536 If such a certificate is found during a verify operation it is
3537 rejected by default: this behaviour can be overridden by either
3538 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3539 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3540 X509_supported_extension() has also been added which returns 1 if a
3541 particular extension is supported.
3544 *) Modify the behaviour of EVP cipher functions in similar way to digests
3545 to retain compatibility with existing code.
3548 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3549 compatibility with existing code. In particular the 'ctx' parameter does
3550 not have to be to be initialized before the call to EVP_DigestInit() and
3551 it is tidied up after a call to EVP_DigestFinal(). New function
3552 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3553 EVP_MD_CTX_copy() changed to not require the destination to be
3554 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3555 requires the destination to be valid.
3557 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3558 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3561 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3562 so that complete 'Handshake' protocol structures are kept in memory
3563 instead of overwriting 'msg_type' and 'length' with 'body' data.
3566 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3567 [Massimo Santin via Richard Levitte]
3569 *) Major restructuring to the underlying ENGINE code. This includes
3570 reduction of linker bloat, separation of pure "ENGINE" manipulation
3571 (initialisation, etc) from functionality dealing with implementations
3572 of specific crypto iterfaces. This change also introduces integrated
3573 support for symmetric ciphers and digest implementations - so ENGINEs
3574 can now accelerate these by providing EVP_CIPHER and EVP_MD
3575 implementations of their own. This is detailed in crypto/engine/README
3576 as it couldn't be adequately described here. However, there are a few
3577 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3578 were changed in the original introduction of ENGINE code have now
3579 reverted back - the hooking from this code to ENGINE is now a good
3580 deal more passive and at run-time, operations deal directly with
3581 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3582 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3583 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3584 they were not being used by the framework as there is no concept of a
3585 BIGNUM_METHOD and they could not be generalised to the new
3586 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3587 ENGINE_cpy() has been removed as it cannot be consistently defined in
3591 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3594 *) Change mkdef.pl to sort symbols that get the same entry number,
3595 and make sure the automatically generated functions ERR_load_*
3596 become part of libeay.num as well.
3599 *) New function SSL_renegotiate_pending(). This returns true once
3600 renegotiation has been requested (either SSL_renegotiate() call
3601 or HelloRequest/ClientHello receveived from the peer) and becomes
3602 false once a handshake has been completed.
3603 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3604 sends a HelloRequest, but does not ensure that a handshake takes
3605 place. SSL_renegotiate_pending() is useful for checking if the
3606 client has followed the request.)
3609 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3610 By default, clients may request session resumption even during
3611 renegotiation (if session ID contexts permit); with this option,
3612 session resumption is possible only in the first handshake.
3614 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3615 more bits available for options that should not be part of
3616 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3619 *) Add some demos for certificate and certificate request creation.
3622 *) Make maximum certificate chain size accepted from the peer application
3623 settable (SSL*_get/set_max_cert_list()), as proposed by
3624 "Douglas E. Engert" <deengert@anl.gov>.
3627 *) Add support for shared libraries for Unixware-7
3628 (Boyd Lynn Gerber <gerberb@zenez.com>).
3631 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3632 be done prior to destruction. Use this to unload error strings from
3633 ENGINEs that load their own error strings. NB: This adds two new API
3634 functions to "get" and "set" this destroy handler in an ENGINE.
3637 *) Alter all existing ENGINE implementations (except "openssl" and
3638 "openbsd") to dynamically instantiate their own error strings. This
3639 makes them more flexible to be built both as statically-linked ENGINEs
3640 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3641 Also, add stub code to each that makes building them as self-contained
3642 shared-libraries easier (see README.ENGINE).
3645 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3646 implementations into applications that are completely implemented in
3647 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3648 commands that can be used to configure what shared-library to load and
3649 to control aspects of the way it is handled. Also, made an update to
3650 the README.ENGINE file that brings its information up-to-date and
3651 provides some information and instructions on the "dynamic" ENGINE
3652 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3655 *) Make it possible to unload ranges of ERR strings with a new
3656 "ERR_unload_strings" function.
3659 *) Add a copy() function to EVP_MD.
3662 *) Make EVP_MD routines take a context pointer instead of just the
3663 md_data void pointer.
3666 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3667 that the digest can only process a single chunk of data
3668 (typically because it is provided by a piece of
3669 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3670 is only going to provide a single chunk of data, and hence the
3671 framework needn't accumulate the data for oneshot drivers.
3674 *) As with "ERR", make it possible to replace the underlying "ex_data"
3675 functions. This change also alters the storage and management of global
3676 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3677 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3678 index counters. The API functions that use this state have been changed
3679 to take a "class_index" rather than pointers to the class's local STACK
3680 and counter, and there is now an API function to dynamically create new
3681 classes. This centralisation allows us to (a) plug a lot of the
3682 thread-safety problems that existed, and (b) makes it possible to clean
3683 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3684 such data would previously have always leaked in application code and
3685 workarounds were in place to make the memory debugging turn a blind eye
3686 to it. Application code that doesn't use this new function will still
3687 leak as before, but their memory debugging output will announce it now
3688 rather than letting it slide.
3690 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3691 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3692 has a return value to indicate success or failure.
3695 *) Make it possible to replace the underlying "ERR" functions such that the
3696 global state (2 LHASH tables and 2 locks) is only used by the "default"
3697 implementation. This change also adds two functions to "get" and "set"
3698 the implementation prior to it being automatically set the first time
3699 any other ERR function takes place. Ie. an application can call "get",
3700 pass the return value to a module it has just loaded, and that module
3701 can call its own "set" function using that value. This means the
3702 module's "ERR" operations will use (and modify) the error state in the
3703 application and not in its own statically linked copy of OpenSSL code.
3706 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3707 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3708 the operation, and provides a more encapsulated way for external code
3709 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3710 to use these functions rather than manually incrementing the counts.
3712 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3715 *) Add EVP test program.
3718 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3721 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3722 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3723 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3724 These allow a CRL to be built without having to access X509_CRL fields
3725 directly. Modify 'ca' application to use new functions.
3728 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3729 bug workarounds. Rollback attack detection is a security feature.
3730 The problem will only arise on OpenSSL servers when TLSv1 is not
3731 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3732 Software authors not wanting to support TLSv1 will have special reasons
3733 for their choice and can explicitly enable this option.
3734 [Bodo Moeller, Lutz Jaenicke]
3736 *) Rationalise EVP so it can be extended: don't include a union of
3737 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3738 (similar to those existing for EVP_CIPHER_CTX).
3743 EVP_MD_CTX_init(&md); /* new function call */
3744 EVP_DigestInit(&md, EVP_sha1());
3745 EVP_DigestUpdate(&md, in, len);
3746 EVP_DigestFinal(&md, out, NULL);
3747 EVP_MD_CTX_cleanup(&md); /* new function call */
3751 *) Make DES key schedule conform to the usual scheme, as well as
3752 correcting its structure. This means that calls to DES functions
3753 now have to pass a pointer to a des_key_schedule instead of a
3754 plain des_key_schedule (which was actually always a pointer
3757 des_key_schedule ks;
3759 des_set_key_checked(..., &ks);
3760 des_ncbc_encrypt(..., &ks, ...);
3762 (Note that a later change renames 'des_...' into 'DES_...'.)
3765 *) Initial reduction of linker bloat: the use of some functions, such as
3766 PEM causes large amounts of unused functions to be linked in due to
3767 poor organisation. For example pem_all.c contains every PEM function
3768 which has a knock on effect of linking in large amounts of (unused)
3769 ASN1 code. Grouping together similar functions and splitting unrelated
3770 functions prevents this.
3773 *) Cleanup of EVP macros.
3776 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3777 correct _ecb suffix.
3780 *) Add initial OCSP responder support to ocsp application. The
3781 revocation information is handled using the text based index
3782 use by the ca application. The responder can either handle
3783 requests generated internally, supplied in files (for example
3784 via a CGI script) or using an internal minimal server.
3787 *) Add configuration choices to get zlib compression for TLS.
3790 *) Changes to Kerberos SSL for RFC 2712 compliance:
3791 1. Implemented real KerberosWrapper, instead of just using
3792 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3793 2. Implemented optional authenticator field of KerberosWrapper.
3795 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3796 and authenticator structs; see crypto/krb5/.
3798 Generalized Kerberos calls to support multiple Kerberos libraries.
3799 [Vern Staats <staatsvr@asc.hpc.mil>,
3800 Jeffrey Altman <jaltman@columbia.edu>
3801 via Richard Levitte]
3803 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3804 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3805 values for each of the key sizes rather than having just
3806 parameters (and 'speed' generating keys each time).
3809 *) Speed up EVP routines.
3812 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3813 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3814 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3815 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3817 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3818 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3819 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3822 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3824 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3827 *) Added the OS2-EMX target.
3828 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3830 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3831 to support NCONF routines in extension code. New function CONF_set_nconf()
3832 to allow functions which take an NCONF to also handle the old LHASH
3833 structure: this means that the old CONF compatible routines can be
3834 retained (in particular wrt extensions) without having to duplicate the
3835 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3838 *) Enhance the general user interface with mechanisms for inner control
3839 and with possibilities to have yes/no kind of prompts.
3842 *) Change all calls to low level digest routines in the library and
3843 applications to use EVP. Add missing calls to HMAC_cleanup() and
3844 don't assume HMAC_CTX can be copied using memcpy().
3845 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3847 *) Add the possibility to control engines through control names but with
3848 arbitrary arguments instead of just a string.
3849 Change the key loaders to take a UI_METHOD instead of a callback
3850 function pointer. NOTE: this breaks binary compatibility with earlier
3851 versions of OpenSSL [engine].
3852 Adapt the nCipher code for these new conditions and add a card insertion
3856 *) Enhance the general user interface with mechanisms to better support
3857 dialog box interfaces, application-defined prompts, the possibility
3858 to use defaults (for example default passwords from somewhere else)
3859 and interrupts/cancellations.
3862 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3863 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3866 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3867 tidy up some unnecessarily weird code in 'sk_new()').
3868 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3870 *) Change the key loading routines for ENGINEs to use the same kind
3871 callback (pem_password_cb) as all other routines that need this
3875 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3876 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3877 than this minimum value is recommended.
3880 *) New random seeder for OpenVMS, using the system process statistics
3881 that are easily reachable.
3884 *) Windows apparently can't transparently handle global
3885 variables defined in DLLs. Initialisations such as:
3887 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3889 wont compile. This is used by the any applications that need to
3890 declare their own ASN1 modules. This was fixed by adding the option
3891 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3892 needed for static libraries under Win32.
3895 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3896 setting of purpose and trust fields. New X509_STORE trust and
3897 purpose functions and tidy up setting in other SSL functions.
3900 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3901 structure. These are inherited by X509_STORE_CTX when it is
3902 initialised. This allows various defaults to be set in the
3903 X509_STORE structure (such as flags for CRL checking and custom
3904 purpose or trust settings) for functions which only use X509_STORE_CTX
3905 internally such as S/MIME.
3907 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3908 trust settings if they are not set in X509_STORE. This allows X509_STORE
3909 purposes and trust (in S/MIME for example) to override any set by default.
3911 Add command line options for CRL checking to smime, s_client and s_server
3915 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3916 are set then the CRL is looked up in the X509_STORE structure and
3917 its validity and signature checked, then if the certificate is found
3918 in the CRL the verify fails with a revoked error.
3920 Various new CRL related callbacks added to X509_STORE_CTX structure.
3922 Command line options added to 'verify' application to support this.
3924 This needs some additional work, such as being able to handle multiple
3925 CRLs with different times, extension based lookup (rather than just
3926 by subject name) and ultimately more complete V2 CRL extension
3930 *) Add a general user interface API (crypto/ui/). This is designed
3931 to replace things like des_read_password and friends (backward
3932 compatibility functions using this new API are provided).
3933 The purpose is to remove prompting functions from the DES code
3934 section as well as provide for prompting through dialog boxes in
3935 a window system and the like.
3938 *) Add "ex_data" support to ENGINE so implementations can add state at a
3939 per-structure level rather than having to store it globally.
3942 *) Make it possible for ENGINE structures to be copied when retrieved by
3943 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3944 This causes the "original" ENGINE structure to act like a template,
3945 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3946 operational state can be localised to each ENGINE structure, despite the
3947 fact they all share the same "methods". New ENGINE structures returned in
3948 this case have no functional references and the return value is the single
3949 structural reference. This matches the single structural reference returned
3950 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3954 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3955 needs to match any other type at all we need to manually clear the
3959 *) Changes to the "openssl engine" utility to include;
3960 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3961 about an ENGINE's available control commands.
3962 - executing control commands from command line arguments using the
3963 '-pre' and '-post' switches. '-post' is only used if '-t' is
3964 specified and the ENGINE is successfully initialised. The syntax for
3965 the individual commands are colon-separated, for example;
3966 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3969 *) New dynamic control command support for ENGINEs. ENGINEs can now
3970 declare their own commands (numbers), names (strings), descriptions,
3971 and input types for run-time discovery by calling applications. A
3972 subset of these commands are implicitly classed as "executable"
3973 depending on their input type, and only these can be invoked through
3974 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3975 can be based on user input, config files, etc). The distinction is
3976 that "executable" commands cannot return anything other than a boolean
3977 result and can only support numeric or string input, whereas some
3978 discoverable commands may only be for direct use through
3979 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3980 pointers, or other custom uses. The "executable" commands are to
3981 support parameterisations of ENGINE behaviour that can be
3982 unambiguously defined by ENGINEs and used consistently across any
3983 OpenSSL-based application. Commands have been added to all the
3984 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3985 control over shared-library paths without source code alterations.
3988 *) Changed all ENGINE implementations to dynamically allocate their
3989 ENGINEs rather than declaring them statically. Apart from this being
3990 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3991 this also allows the implementations to compile without using the
3992 internal engine_int.h header.
3995 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3996 'const' value. Any code that should be able to modify a RAND_METHOD
3997 should already have non-const pointers to it (ie. they should only
3998 modify their own ones).
4001 *) Made a variety of little tweaks to the ENGINE code.
4002 - "atalla" and "ubsec" string definitions were moved from header files
4003 to C code. "nuron" string definitions were placed in variables
4004 rather than hard-coded - allowing parameterisation of these values
4005 later on via ctrl() commands.
4006 - Removed unused "#if 0"'d code.
4007 - Fixed engine list iteration code so it uses ENGINE_free() to release
4008 structural references.
4009 - Constified the RAND_METHOD element of ENGINE structures.
4010 - Constified various get/set functions as appropriate and added
4011 missing functions (including a catch-all ENGINE_cpy that duplicates
4012 all ENGINE values onto a new ENGINE except reference counts/state).
4013 - Removed NULL parameter checks in get/set functions. Setting a method
4014 or function to NULL is a way of cancelling out a previously set
4015 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4016 and doesn't justify the extra error symbols and code.
4017 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4018 flags from engine_int.h to engine.h.
4019 - Changed prototypes for ENGINE handler functions (init(), finish(),
4020 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4023 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4024 to the algorithm using long division. The binary algorithm can be
4025 used only if the modulus is odd. On 32-bit systems, it is faster
4026 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4027 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4028 up to 450 bits. In 64-bit environments, the binary algorithm
4029 appears to be advantageous for much longer moduli; here we use it
4030 for moduli up to 2048 bits.
4033 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4034 could not support the combine flag in choice fields.
4037 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4038 extensions from a certificate request to the certificate.
4041 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4042 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4043 file: this allows the display of the certificate about to be
4044 signed to be customised, to allow certain fields to be included
4045 or excluded and extension details. The old system didn't display
4046 multicharacter strings properly, omitted fields not in the policy
4047 and couldn't display additional details such as extensions.
4050 *) Function EC_POINTs_mul for multiple scalar multiplication
4051 of an arbitrary number of elliptic curve points
4052 \sum scalars[i]*points[i],
4053 optionally including the generator defined for the EC_GROUP:
4054 scalar*generator + \sum scalars[i]*points[i].
4056 EC_POINT_mul is a simple wrapper function for the typical case
4057 that the point list has just one item (besides the optional
4061 *) First EC_METHODs for curves over GF(p):
4063 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4064 operations and provides various method functions that can also
4065 operate with faster implementations of modular arithmetic.
4067 EC_GFp_mont_method() reuses most functions that are part of
4068 EC_GFp_simple_method, but uses Montgomery arithmetic.
4070 [Bodo Moeller; point addition and point doubling
4071 implementation directly derived from source code provided by
4072 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4074 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4075 crypto/ec/ec_lib.c):
4077 Curves are EC_GROUP objects (with an optional group generator)
4078 based on EC_METHODs that are built into the library.
4080 Points are EC_POINT objects based on EC_GROUP objects.
4082 Most of the framework would be able to handle curves over arbitrary
4083 finite fields, but as there are no obvious types for fields other
4084 than GF(p), some functions are limited to that for now.
4087 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4088 that the file contains a complete HTTP response.
4091 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4092 change the def and num file printf format specifier from "%-40sXXX"
4093 to "%-39s XXX". The latter will always guarantee a space after the
4094 field while the former will cause them to run together if the field
4095 is 40 of more characters long.
4098 *) Constify the cipher and digest 'method' functions and structures
4099 and modify related functions to take constant EVP_MD and EVP_CIPHER
4103 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4104 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4107 *) Modify EVP_Digest*() routines so they now return values. Although the
4108 internal software routines can never fail additional hardware versions
4112 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4114 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4115 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4118 ERR_R_NESTED_ASN1_ERROR
4120 ERR_R_MISSING_ASN1_EOS
4121 were 4 .. 9, conflicting with
4122 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4124 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4125 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4127 Add new error code 'ERR_R_INTERNAL_ERROR'.
4130 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4134 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4135 sets the subject name for a new request or supersedes the
4136 subject name in a given request. Formats that can be parsed are
4137 'CN=Some Name, OU=myOU, C=IT'
4139 'CN=Some Name/OU=myOU/C=IT'.
4141 Add options '-batch' and '-verbose' to 'openssl req'.
4142 [Massimiliano Pala <madwolf@hackmasters.net>]
4144 *) Introduce the possibility to access global variables through
4145 functions on platform were that's the best way to handle exporting
4146 global variables in shared libraries. To enable this functionality,
4147 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4148 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4149 is normally done by Configure or something similar).
4151 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4152 in the source file (foo.c) like this:
4154 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4155 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4157 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4158 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4160 OPENSSL_DECLARE_GLOBAL(int,foo);
4161 #define foo OPENSSL_GLOBAL_REF(foo)
4162 OPENSSL_DECLARE_GLOBAL(double,bar);
4163 #define bar OPENSSL_GLOBAL_REF(bar)
4165 The #defines are very important, and therefore so is including the
4166 header file everywhere where the defined globals are used.
4168 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4169 of ASN.1 items, but that structure is a bit different.
4171 The largest change is in util/mkdef.pl which has been enhanced with
4172 better and easier to understand logic to choose which symbols should
4173 go into the Windows .def files as well as a number of fixes and code
4174 cleanup (among others, algorithm keywords are now sorted
4175 lexicographically to avoid constant rewrites).
4178 *) In BN_div() keep a copy of the sign of 'num' before writing the
4179 result to 'rm' because if rm==num the value will be overwritten
4180 and produce the wrong result if 'num' is negative: this caused
4181 problems with BN_mod() and BN_nnmod().
4184 *) Function OCSP_request_verify(). This checks the signature on an
4185 OCSP request and verifies the signer certificate. The signer
4186 certificate is just checked for a generic purpose and OCSP request
4190 *) Add OCSP_check_validity() function to check the validity of OCSP
4191 responses. OCSP responses are prepared in real time and may only
4192 be a few seconds old. Simply checking that the current time lies
4193 between thisUpdate and nextUpdate max reject otherwise valid responses
4194 caused by either OCSP responder or client clock inaccuracy. Instead
4195 we allow thisUpdate and nextUpdate to fall within a certain period of
4196 the current time. The age of the response can also optionally be
4197 checked. Two new options -validity_period and -status_age added to
4201 *) If signature or public key algorithm is unrecognized print out its
4202 OID rather that just UNKNOWN.
4205 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4206 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4207 ID to be generated from the issuer certificate alone which can then be
4208 passed to OCSP_id_issuer_cmp().
4211 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4212 ASN1 modules to export functions returning ASN1_ITEM pointers
4213 instead of the ASN1_ITEM structures themselves. This adds several
4214 new macros which allow the underlying ASN1 function/structure to
4215 be accessed transparently. As a result code should not use ASN1_ITEM
4216 references directly (such as &X509_it) but instead use the relevant
4217 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4218 use of the new ASN1 code on platforms where exporting structures
4219 is problematical (for example in shared libraries) but exporting
4220 functions returning pointers to structures is not.
4223 *) Add support for overriding the generation of SSL/TLS session IDs.
4224 These callbacks can be registered either in an SSL_CTX or per SSL.
4225 The purpose of this is to allow applications to control, if they wish,
4226 the arbitrary values chosen for use as session IDs, particularly as it
4227 can be useful for session caching in multiple-server environments. A
4228 command-line switch for testing this (and any client code that wishes
4229 to use such a feature) has been added to "s_server".
4230 [Geoff Thorpe, Lutz Jaenicke]
4232 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4233 of the form '#if defined(...) || defined(...) || ...' and
4234 '#if !defined(...) && !defined(...) && ...'. This also avoids
4235 the growing number of special cases it was previously handling.
4238 *) Make all configuration macros available for application by making
4239 sure they are available in opensslconf.h, by giving them names starting
4240 with "OPENSSL_" to avoid conflicts with other packages and by making
4241 sure e_os2.h will cover all platform-specific cases together with
4243 Additionally, it is now possible to define configuration/platform-
4244 specific names (called "system identities"). In the C code, these
4245 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4246 macro with the name beginning with "OPENSSL_SYS_", which is determined
4247 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4251 *) New option -set_serial to 'req' and 'x509' this allows the serial
4252 number to use to be specified on the command line. Previously self
4253 signed certificates were hard coded with serial number 0 and the
4254 CA options of 'x509' had to use a serial number in a file which was
4258 *) New options to 'ca' utility to support V2 CRL entry extensions.
4259 Currently CRL reason, invalidity date and hold instruction are
4260 supported. Add new CRL extensions to V3 code and some new objects.
4263 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4264 disable standard block padding (aka PKCS#5 padding) in the EVP
4265 API, which was previously mandatory. This means that the data is
4266 not padded in any way and so the total length much be a multiple
4267 of the block size, otherwise an error occurs.
4270 *) Initial (incomplete) OCSP SSL support.
4273 *) New function OCSP_parse_url(). This splits up a URL into its host,
4274 port and path components: primarily to parse OCSP URLs. New -url
4275 option to ocsp utility.
4278 *) New nonce behavior. The return value of OCSP_check_nonce() now
4279 reflects the various checks performed. Applications can decide
4280 whether to tolerate certain situations such as an absent nonce
4281 in a response when one was present in a request: the ocsp application
4282 just prints out a warning. New function OCSP_add1_basic_nonce()
4283 this is to allow responders to include a nonce in a response even if
4284 the request is nonce-less.
4287 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4288 skipped when using openssl x509 multiple times on a single input file,
4289 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4292 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4293 set string type: to handle setting ASN1_TIME structures. Fix ca
4294 utility to correctly initialize revocation date of CRLs.
4297 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4298 the clients preferred ciphersuites and rather use its own preferences.
4299 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4300 Internet Explorer by ensuring unchanged hash method during stepup.
4301 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4304 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4305 to aes and add a new 'exist' option to print out symbols that don't
4309 *) Additional options to ocsp utility to allow flags to be set and
4310 additional certificates supplied.
4313 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4314 OCSP client a number of certificate to only verify the response
4318 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4319 handle the new API. Currently only ECB, CBC modes supported. Add new
4322 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4323 Encryption Standard (AES) Ciphersuites for Transport Layer
4324 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4325 not enabled by default and were not part of the "ALL" ciphersuite
4326 alias because they were not yet official; they could be
4327 explicitly requested by specifying the "AESdraft" ciphersuite
4328 group alias. In the final release of OpenSSL 0.9.7, the group
4329 alias is called "AES" and is part of "ALL".)
4330 [Ben Laurie, Steve Henson, Bodo Moeller]
4332 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4333 request to response.
4336 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4337 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4338 extract information from a certificate request. OCSP_response_create()
4339 creates a response and optionally adds a basic response structure.
4340 OCSP_basic_add1_status() adds a complete single response to a basic
4341 response and returns the OCSP_SINGLERESP structure just added (to allow
4342 extensions to be included for example). OCSP_basic_add1_cert() adds a
4343 certificate to a basic response and OCSP_basic_sign() signs a basic
4344 response with various flags. New helper functions ASN1_TIME_check()
4345 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4346 (converts ASN1_TIME to GeneralizedTime).
4349 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4350 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4351 structure from a certificate. X509_pubkey_digest() digests the public_key
4352 contents: this is used in various key identifiers.
4355 *) Make sk_sort() tolerate a NULL argument.
4356 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4358 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4359 passed by the function are trusted implicitly. If any of them signed the
4360 response then it is assumed to be valid and is not verified.
4363 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4364 to data. This was previously part of the PKCS7 ASN1 code. This
4365 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4366 [Steve Henson, reported by Kenneth R. Robinette
4367 <support@securenetterm.com>]
4369 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4370 routines: without these tracing memory leaks is very painful.
4371 Fix leaks in PKCS12 and PKCS7 routines.
4374 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4375 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4376 effectively meant GeneralizedTime would never be used. Now it
4377 is initialised to -1 but X509_time_adj() now has to check the value
4378 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4379 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4380 [Steve Henson, reported by Kenneth R. Robinette
4381 <support@securenetterm.com>]
4383 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4384 result in a zero length in the ASN1_INTEGER structure which was
4385 not consistent with the structure when d2i_ASN1_INTEGER() was used
4386 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4387 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4388 where it did not print out a minus for negative ASN1_INTEGER.
4391 *) Add summary printout to ocsp utility. The various functions which
4392 convert status values to strings have been renamed to:
4393 OCSP_response_status_str(), OCSP_cert_status_str() and
4394 OCSP_crl_reason_str() and are no longer static. New options
4395 to verify nonce values and to disable verification. OCSP response
4396 printout format cleaned up.
4399 *) Add additional OCSP certificate checks. These are those specified
4400 in RFC2560. This consists of two separate checks: the CA of the
4401 certificate being checked must either be the OCSP signer certificate
4402 or the issuer of the OCSP signer certificate. In the latter case the
4403 OCSP signer certificate must contain the OCSP signing extended key
4404 usage. This check is performed by attempting to match the OCSP
4405 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4406 in the OCSP_CERTID structures of the response.
4409 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4410 and related routines. This uses the standard OpenSSL certificate
4411 verify routines to perform initial checks (just CA validity) and
4412 to obtain the certificate chain. Then additional checks will be
4413 performed on the chain. Currently the root CA is checked to see
4414 if it is explicitly trusted for OCSP signing. This is used to set
4415 a root CA as a global signing root: that is any certificate that
4416 chains to that CA is an acceptable OCSP signing certificate.
4419 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4420 extensions from a separate configuration file.
4421 As when reading extensions from the main configuration file,
4422 the '-extensions ...' option may be used for specifying the
4424 [Massimiliano Pala <madwolf@comune.modena.it>]
4426 *) New OCSP utility. Allows OCSP requests to be generated or
4427 read. The request can be sent to a responder and the output
4428 parsed, outputed or printed in text form. Not complete yet:
4429 still needs to check the OCSP response validity.
4432 *) New subcommands for 'openssl ca':
4433 'openssl ca -status <serial>' prints the status of the cert with
4434 the given serial number (according to the index file).
4435 'openssl ca -updatedb' updates the expiry status of certificates
4437 [Massimiliano Pala <madwolf@comune.modena.it>]
4439 *) New '-newreq-nodes' command option to CA.pl. This is like
4440 '-newreq', but calls 'openssl req' with the '-nodes' option
4441 so that the resulting key is not encrypted.
4442 [Damien Miller <djm@mindrot.org>]
4444 *) New configuration for the GNU Hurd.
4445 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4447 *) Initial code to implement OCSP basic response verify. This
4448 is currently incomplete. Currently just finds the signer's
4449 certificate and verifies the signature on the response.
4452 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4453 value of OPENSSLDIR. This is available via the new '-d' option
4454 to 'openssl version', and is also included in 'openssl version -a'.
4457 *) Allowing defining memory allocation callbacks that will be given
4458 file name and line number information in additional arguments
4459 (a const char* and an int). The basic functionality remains, as
4460 well as the original possibility to just replace malloc(),
4461 realloc() and free() by functions that do not know about these
4462 additional arguments. To register and find out the current
4463 settings for extended allocation functions, the following
4464 functions are provided:
4466 CRYPTO_set_mem_ex_functions
4467 CRYPTO_set_locked_mem_ex_functions
4468 CRYPTO_get_mem_ex_functions
4469 CRYPTO_get_locked_mem_ex_functions
4471 These work the same way as CRYPTO_set_mem_functions and friends.
4472 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4473 extended allocation function is enabled.
4474 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4475 a conventional allocation function is enabled.
4476 [Richard Levitte, Bodo Moeller]
4478 *) Finish off removing the remaining LHASH function pointer casts.
4479 There should no longer be any prototype-casting required when using
4480 the LHASH abstraction, and any casts that remain are "bugs". See
4481 the callback types and macros at the head of lhash.h for details
4482 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4485 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4486 If /dev/[u]random devices are not available or do not return enough
4487 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4489 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4490 /etc/entropy will be queried once each in this sequence, quering stops
4491 when enough entropy was collected without querying more sockets.
4494 *) Change the Unix RAND_poll() variant to be able to poll several
4495 random devices, as specified by DEVRANDOM, until a sufficient amount
4496 of data has been collected. We spend at most 10 ms on each file
4497 (select timeout) and read in non-blocking mode. DEVRANDOM now
4498 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4499 (previously it was just the string "/dev/urandom"), so on typical
4500 platforms the 10 ms delay will never occur.
4501 Also separate out the Unix variant to its own file, rand_unix.c.
4502 For VMS, there's a currently-empty rand_vms.c.
4505 *) Move OCSP client related routines to ocsp_cl.c. These
4506 provide utility functions which an application needing
4507 to issue a request to an OCSP responder and analyse the
4508 response will typically need: as opposed to those which an
4509 OCSP responder itself would need which will be added later.
4511 OCSP_request_sign() signs an OCSP request with an API similar
4512 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4513 response. OCSP_response_get1_basic() extracts basic response
4514 from response. OCSP_resp_find_status(): finds and extracts status
4515 information from an OCSP_CERTID structure (which will be created
4516 when the request structure is built). These are built from lower
4517 level functions which work on OCSP_SINGLERESP structures but
4518 wont normally be used unless the application wishes to examine
4519 extensions in the OCSP response for example.
4521 Replace nonce routines with a pair of functions.
4522 OCSP_request_add1_nonce() adds a nonce value and optionally
4523 generates a random value. OCSP_check_nonce() checks the
4524 validity of the nonce in an OCSP response.
4527 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4528 This doesn't copy the supplied OCSP_CERTID and avoids the
4529 need to free up the newly created id. Change return type
4530 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4531 This can then be used to add extensions to the request.
4532 Deleted OCSP_request_new(), since most of its functionality
4533 is now in OCSP_REQUEST_new() (and the case insensitive name
4534 clash) apart from the ability to set the request name which
4535 will be added elsewhere.
4538 *) Update OCSP API. Remove obsolete extensions argument from
4539 various functions. Extensions are now handled using the new
4540 OCSP extension code. New simple OCSP HTTP function which
4541 can be used to send requests and parse the response.
4544 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4545 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4546 uses the special reorder version of SET OF to sort the attributes
4547 and reorder them to match the encoded order. This resolves a long
4548 standing problem: a verify on a PKCS7 structure just after signing
4549 it used to fail because the attribute order did not match the
4550 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4551 it uses the received order. This is necessary to tolerate some broken
4552 software that does not order SET OF. This is handled by encoding
4553 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4554 to produce the required SET OF.
4557 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4558 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4559 files to get correct declarations of the ASN.1 item variables.
4562 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4563 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4564 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4565 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4566 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4567 ASN1_ITEM and no wrapper functions.
4570 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4571 replace the old function pointer based I/O routines. Change most of
4572 the *_d2i_bio() and *_d2i_fp() functions to use these.
4575 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4576 lines, recognice more "algorithms" that can be deselected, and make
4577 it complain about algorithm deselection that isn't recognised.
4580 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4581 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4582 to use new functions. Add NO_ASN1_OLD which can be set to remove
4583 some old style ASN1 functions: this can be used to determine if old
4584 code will still work when these eventually go away.
4587 *) New extension functions for OCSP structures, these follow the
4588 same conventions as certificates and CRLs.
4591 *) New function X509V3_add1_i2d(). This automatically encodes and
4592 adds an extension. Its behaviour can be customised with various
4593 flags to append, replace or delete. Various wrappers added for
4594 certifcates and CRLs.
4597 *) Fix to avoid calling the underlying ASN1 print routine when
4598 an extension cannot be parsed. Correct a typo in the
4599 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4602 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4603 entries for variables.
4606 *) Add functionality to apps/openssl.c for detecting locking
4607 problems: As the program is single-threaded, all we have
4608 to do is register a locking callback using an array for
4609 storing which locks are currently held by the program.
4612 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4613 SSL_get_ex_data_X509_STORE_idx(), which is used in
4614 ssl_verify_cert_chain() and thus can be called at any time
4615 during TLS/SSL handshakes so that thread-safety is essential.
4616 Unfortunately, the ex_data design is not at all suited
4617 for multi-threaded use, so it probably should be abolished.
4620 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4621 [Broadcom, tweaked and integrated by Geoff Thorpe]
4623 *) Move common extension printing code to new function
4624 X509V3_print_extensions(). Reorganise OCSP print routines and
4625 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4628 *) New function X509_signature_print() to remove duplication in some
4632 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4633 set (this was treated exactly the same as SET OF previously). This
4634 is used to reorder the STACK representing the structure to match the
4635 encoding. This will be used to get round a problem where a PKCS7
4636 structure which was signed could not be verified because the STACK
4637 order did not reflect the encoded order.
4640 *) Reimplement the OCSP ASN1 module using the new code.
4643 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4644 for its ASN1 operations. The old style function pointers still exist
4645 for now but they will eventually go away.
4648 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4649 completely replaces the old ASN1 functionality with a table driven
4650 encoder and decoder which interprets an ASN1_ITEM structure describing
4651 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4652 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4653 has also been converted to the new form.
4656 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4657 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4658 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4659 for negative moduli.
4662 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4663 of not touching the result's sign bit.
4666 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4670 *) Changed the LHASH code to use prototypes for callbacks, and created
4671 macros to declare and implement thin (optionally static) functions
4672 that provide type-safety and avoid function pointer casting for the
4673 type-specific callbacks.
4676 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4678 [Veers Staats <staatsvr@asc.hpc.mil>,
4679 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4681 *) Reformat the FAQ so the different questions and answers can be divided
4682 in sections depending on the subject.
4685 *) Have the zlib compression code load ZLIB.DLL dynamically under
4689 *) New function BN_mod_sqrt for computing square roots modulo a prime
4690 (using the probabilistic Tonelli-Shanks algorithm unless
4691 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4692 be handled deterministically).
4693 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4695 *) Make BN_mod_inverse faster by explicitly handling small quotients
4696 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4697 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4700 *) New function BN_kronecker.
4703 *) Fix BN_gcd so that it works on negative inputs; the result is
4704 positive unless both parameters are zero.
4705 Previously something reasonably close to an infinite loop was
4706 possible because numbers could be growing instead of shrinking
4707 in the implementation of Euclid's algorithm.
4710 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4711 sign of the number in question.
4713 Fix BN_is_word(a,w) to work correctly for w == 0.
4715 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4716 because its test if the absolute value of 'a' equals 'w'.
4717 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4718 it exists mostly for use in the implementations of BN_is_zero(),
4719 BN_is_one(), and BN_is_word().
4722 *) New function BN_swap.
4725 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4726 the exponentiation functions are more likely to produce reasonable
4727 results on negative inputs.
4730 *) Change BN_mod_mul so that the result is always non-negative.
4731 Previously, it could be negative if one of the factors was negative;
4732 I don't think anyone really wanted that behaviour.
4735 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4736 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4737 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4738 and add new functions:
4747 BN_mod_lshift1_quick
4751 These functions always generate non-negative results.
4753 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4754 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4756 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4757 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4758 be reduced modulo m.
4759 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4762 The following entry accidentily appeared in the CHANGES file
4763 distributed with OpenSSL 0.9.7. The modifications described in
4764 it do *not* apply to OpenSSL 0.9.7.
4766 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4767 was actually never needed) and in BN_mul(). The removal in BN_mul()
4768 required a small change in bn_mul_part_recursive() and the addition
4769 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4770 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4771 bn_sub_words() and bn_add_words() except they take arrays with
4776 *) In 'openssl passwd', verify passwords read from the terminal
4777 unless the '-salt' option is used (which usually means that
4778 verification would just waste user's time since the resulting
4779 hash is going to be compared with some given password hash)
4780 or the new '-noverify' option is used.
4782 This is an incompatible change, but it does not affect
4783 non-interactive use of 'openssl passwd' (passwords on the command
4784 line, '-stdin' option, '-in ...' option) and thus should not
4788 *) Remove all references to RSAref, since there's no more need for it.
4791 *) Make DSO load along a path given through an environment variable
4792 (SHLIB_PATH) with shl_load().
4795 *) Constify the ENGINE code as a result of BIGNUM constification.
4796 Also constify the RSA code and most things related to it. In a
4797 few places, most notable in the depth of the ASN.1 code, ugly
4798 casts back to non-const were required (to be solved at a later
4802 *) Make it so the openssl application has all engines loaded by default.
4805 *) Constify the BIGNUM routines a little more.
4808 *) Add the following functions:
4810 ENGINE_load_cswift()
4812 ENGINE_load_atalla()
4814 ENGINE_load_builtin_engines()
4816 That way, an application can itself choose if external engines that
4817 are built-in in OpenSSL shall ever be used or not. The benefit is
4818 that applications won't have to be linked with libdl or other dso
4819 libraries unless it's really needed.
4821 Changed 'openssl engine' to load all engines on demand.
4822 Changed the engine header files to avoid the duplication of some
4823 declarations (they differed!).
4826 *) 'openssl engine' can now list capabilities.
4829 *) Better error reporting in 'openssl engine'.
4832 *) Never call load_dh_param(NULL) in s_server.
4835 *) Add engine application. It can currently list engines by name and
4836 identity, and test if they are actually available.
4839 *) Improve RPM specification file by forcing symbolic linking and making
4840 sure the installed documentation is also owned by root.root.
4841 [Damien Miller <djm@mindrot.org>]
4843 *) Give the OpenSSL applications more possibilities to make use of
4844 keys (public as well as private) handled by engines.
4847 *) Add OCSP code that comes from CertCo.
4850 *) Add VMS support for the Rijndael code.
4853 *) Added untested support for Nuron crypto accelerator.
4856 *) Add support for external cryptographic devices. This code was
4857 previously distributed separately as the "engine" branch.
4858 [Geoff Thorpe, Richard Levitte]
4860 *) Rework the filename-translation in the DSO code. It is now possible to
4861 have far greater control over how a "name" is turned into a filename
4862 depending on the operating environment and any oddities about the
4863 different shared library filenames on each system.
4866 *) Support threads on FreeBSD-elf in Configure.
4869 *) Fix for SHA1 assembly problem with MASM: it produces
4870 warnings about corrupt line number information when assembling
4871 with debugging information. This is caused by the overlapping
4873 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4876 NCONF_get_number() has no error checking at all. As a replacement,
4877 NCONF_get_number_e() is defined (_e for "error checking") and is
4878 promoted strongly. The old NCONF_get_number is kept around for
4879 binary backward compatibility.
4880 Make it possible for methods to load from something other than a BIO,
4881 by providing a function pointer that is given a name instead of a BIO.
4882 For example, this could be used to load configuration data from an
4886 *) Fix for non blocking accept BIOs. Added new I/O special reason
4887 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4888 with non blocking I/O was not possible because no retry code was
4889 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4893 *) Added the beginnings of Rijndael support.
4896 *) Fix for bug in DirectoryString mask setting. Add support for
4897 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4898 to allow certificate printing to more controllable, additional
4899 'certopt' option to 'x509' to allow new printing options to be
4903 *) Clean old EAY MD5 hack from e_os.h.
4906 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4908 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4909 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4910 [Joe Orton, Steve Henson]
4912 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4914 *) Fix additional bug revealed by the NISCC test suite:
4916 Stop bug triggering large recursion when presented with
4917 certain ASN.1 tags (CVE-2003-0851)
4920 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4922 *) Fix various bugs revealed by running the NISCC test suite:
4924 Stop out of bounds reads in the ASN1 code when presented with
4925 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4927 If verify callback ignores invalid public key errors don't try to check
4928 certificate signature with the NULL public key.
4932 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4933 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4937 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4938 extra data after the compression methods not only for TLS 1.0
4939 but also for SSL 3.0 (as required by the specification).
4940 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4942 *) Change X509_certificate_type() to mark the key as exported/exportable
4943 when it's 512 *bits* long, not 512 bytes.
4946 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4948 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4949 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4950 a protocol version number mismatch like a decryption error
4951 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4954 *) Turn on RSA blinding by default in the default implementation
4955 to avoid a timing attack. Applications that don't want it can call
4956 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4957 They would be ill-advised to do so in most cases.
4958 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4960 *) Change RSA blinding code so that it works when the PRNG is not
4961 seeded (in this case, the secret RSA exponent is abused as
4962 an unpredictable seed -- if it is not unpredictable, there
4963 is no point in blinding anyway). Make RSA blinding thread-safe
4964 by remembering the creator's thread ID in rsa->blinding and
4965 having all other threads use local one-time blinding factors
4966 (this requires more computation than sharing rsa->blinding, but
4967 avoids excessive locking; and if an RSA object is not shared
4968 between threads, blinding will still be very fast).
4971 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4973 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4974 via timing by performing a MAC computation even if incorrrect
4975 block cipher padding has been found. This is a countermeasure
4976 against active attacks where the attacker has to distinguish
4977 between bad padding and a MAC verification error. (CVE-2003-0078)
4979 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4980 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4981 Martin Vuagnoux (EPFL, Ilion)]
4983 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4985 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4986 memory from it's contents. This is done with a counter that will
4987 place alternating values in each byte. This can be used to solve
4988 two issues: 1) the removal of calls to memset() by highly optimizing
4989 compilers, and 2) cleansing with other values than 0, since those can
4990 be read through on certain media, for example a swap space on disk.
4993 *) Bugfix: client side session caching did not work with external caching,
4994 because the session->cipher setting was not restored when reloading
4995 from the external cache. This problem was masked, when
4996 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4997 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5000 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5001 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5002 [Zeev Lieber <zeev-l@yahoo.com>]
5004 *) Undo an undocumented change introduced in 0.9.6e which caused
5005 repeated calls to OpenSSL_add_all_ciphers() and
5006 OpenSSL_add_all_digests() to be ignored, even after calling
5010 *) Change the default configuration reader to deal with last line not
5011 being properly terminated.
5014 *) Change X509_NAME_cmp() so it applies the special rules on handling
5015 DN values that are of type PrintableString, as well as RDNs of type
5016 emailAddress where the value has the type ia5String.
5017 [stefank@valicert.com via Richard Levitte]
5019 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5020 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5021 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5022 the bitwise-OR of the two for use by the majority of applications
5023 wanting this behaviour, and update the docs. The documented
5024 behaviour and actual behaviour were inconsistent and had been
5025 changing anyway, so this is more a bug-fix than a behavioural
5027 [Geoff Thorpe, diagnosed by Nadav Har'El]
5029 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5030 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5033 *) Fix initialization code race conditions in
5034 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5035 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5036 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5037 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5038 ssl2_get_cipher_by_char(),
5039 ssl3_get_cipher_by_char().
5040 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5042 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5043 the cached sessions are flushed, as the remove_cb() might use ex_data
5044 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5045 (see [openssl.org #212]).
5046 [Geoff Thorpe, Lutz Jaenicke]
5048 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5049 length, instead of the encoding length to d2i_ASN1_OBJECT.
5052 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5054 *) [In 0.9.6g-engine release:]
5055 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5056 [Lynn Gazis <lgazis@rainbow.com>]
5058 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5060 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5061 and get fix the header length calculation.
5062 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5063 Alon Kantor <alonk@checkpoint.com> (and others),
5066 *) Use proper error handling instead of 'assertions' in buffer
5067 overflow checks added in 0.9.6e. This prevents DoS (the
5068 assertions could call abort()).
5069 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5071 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5073 *) Add various sanity checks to asn1_get_length() to reject
5074 the ASN1 length bytes if they exceed sizeof(long), will appear
5075 negative or the content length exceeds the length of the
5077 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5079 *) Fix cipher selection routines: ciphers without encryption had no flags
5080 for the cipher strength set and where therefore not handled correctly
5081 by the selection routines (PR #130).
5084 *) Fix EVP_dsa_sha macro.
5088 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5089 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5090 that was added in OpenSSL 0.9.6d.
5092 As the countermeasure turned out to be incompatible with some
5093 broken SSL implementations, the new option is part of SSL_OP_ALL.
5094 SSL_OP_ALL is usually employed when compatibility with weird SSL
5095 implementations is desired (e.g. '-bugs' option to 's_client' and
5096 's_server'), so the new option is automatically set in many
5100 *) Changes in security patch:
5102 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5103 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5104 Air Force Materiel Command, USAF, under agreement number
5107 *) Add various sanity checks to asn1_get_length() to reject
5108 the ASN1 length bytes if they exceed sizeof(long), will appear
5109 negative or the content length exceeds the length of the
5110 supplied buffer. (CVE-2002-0659)
5111 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5113 *) Assertions for various potential buffer overflows, not known to
5115 [Ben Laurie (CHATS)]
5117 *) Various temporary buffers to hold ASCII versions of integers were
5118 too small for 64 bit platforms. (CVE-2002-0655)
5119 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5121 *) Remote buffer overflow in SSL3 protocol - an attacker could
5122 supply an oversized session ID to a client. (CVE-2002-0656)
5123 [Ben Laurie (CHATS)]
5125 *) Remote buffer overflow in SSL2 protocol - an attacker could
5126 supply an oversized client master key. (CVE-2002-0656)
5127 [Ben Laurie (CHATS)]
5129 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5131 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5132 encoded as NULL) with id-dsa-with-sha1.
5133 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5135 *) Check various X509_...() return values in apps/req.c.
5136 [Nils Larsch <nla@trustcenter.de>]
5138 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5139 an end-of-file condition would erronously be flagged, when the CRLF
5140 was just at the end of a processed block. The bug was discovered when
5141 processing data through a buffering memory BIO handing the data to a
5142 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5143 <ptsekov@syntrex.com> and Nedelcho Stanev.
5146 *) Implement a countermeasure against a vulnerability recently found
5147 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5148 before application data chunks to avoid the use of known IVs
5149 with data potentially chosen by the attacker.
5152 *) Fix length checks in ssl3_get_client_hello().
5155 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5156 to prevent ssl3_read_internal() from incorrectly assuming that
5157 ssl3_read_bytes() found application data while handshake
5158 processing was enabled when in fact s->s3->in_read_app_data was
5159 merely automatically cleared during the initial handshake.
5160 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5162 *) Fix object definitions for Private and Enterprise: they were not
5163 recognized in their shortname (=lowercase) representation. Extend
5164 obj_dat.pl to issue an error when using undefined keywords instead
5165 of silently ignoring the problem (Svenning Sorensen
5166 <sss@sss.dnsalias.net>).
5169 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5170 generators, i.e. generators other than 2 and 5. (Previously, the
5171 code did not properly initialise the 'add' and 'rem' values to
5172 BN_generate_prime().)
5174 In the new general case, we do not insist that 'generator' is
5175 actually a primitive root: This requirement is rather pointless;
5176 a generator of the order-q subgroup is just as good, if not
5180 *) Map new X509 verification errors to alerts. Discovered and submitted by
5181 Tom Wu <tom@arcot.com>.
5184 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5185 returning non-zero before the data has been completely received
5186 when using non-blocking I/O.
5187 [Bodo Moeller; problem pointed out by John Hughes]
5189 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5190 [Ben Laurie, Lutz Jaenicke]
5192 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5193 Yoram Zahavi <YoramZ@gilian.com>).
5196 *) Add information about CygWin 1.3 and on, and preserve proper
5197 configuration for the versions before that.
5198 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5200 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5201 check whether we deal with a copy of a session and do not delete from
5202 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5203 <izhar@checkpoint.com>.
5206 *) Do not store session data into the internal session cache, if it
5207 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5208 flag is set). Proposed by Aslam <aslam@funk.com>.
5211 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5215 *) [In 0.9.6d-engine release:]
5216 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5217 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5219 *) Add the configuration target linux-s390x.
5220 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5222 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5223 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5224 variable as an indication that a ClientHello message has been
5225 received. As the flag value will be lost between multiple
5226 invocations of ssl3_accept when using non-blocking I/O, the
5227 function may not be aware that a handshake has actually taken
5228 place, thus preventing a new session from being added to the
5231 To avoid this problem, we now set s->new_session to 2 instead of
5232 using a local variable.
5233 [Lutz Jaenicke, Bodo Moeller]
5235 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5236 if the SSL_R_LENGTH_MISMATCH error is detected.
5237 [Geoff Thorpe, Bodo Moeller]
5239 *) New 'shared_ldflag' column in Configure platform table.
5242 *) Fix EVP_CIPHER_mode macro.
5243 ["Dan S. Camper" <dan@bti.net>]
5245 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5246 type, we must throw them away by setting rr->length to 0.
5247 [D P Chang <dpc@qualys.com>]
5249 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5251 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5252 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5253 worked incorrectly for those cases where range = 10..._2 and
5254 3*range is two bits longer than range.)
5257 *) Only add signing time to PKCS7 structures if it is not already
5261 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5262 OBJ_ld_ce should be OBJ_id_ce.
5263 Also some ip-pda OIDs in crypto/objects/objects.txt were
5264 incorrect (cf. RFC 3039).
5265 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5267 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5268 returns early because it has nothing to do.
5269 [Andy Schneider <andy.schneider@bjss.co.uk>]
5271 *) [In 0.9.6c-engine release:]
5272 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5273 [Andy Schneider <andy.schneider@bjss.co.uk>]
5275 *) [In 0.9.6c-engine release:]
5276 Add support for Cryptographic Appliance's keyserver technology.
5277 (Use engine 'keyclient')
5278 [Cryptographic Appliances and Geoff Thorpe]
5280 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5281 is called via tools/c89.sh because arguments have to be
5282 rearranged (all '-L' options must appear before the first object
5284 [Richard Shapiro <rshapiro@abinitio.com>]
5286 *) [In 0.9.6c-engine release:]
5287 Add support for Broadcom crypto accelerator cards, backported
5289 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5291 *) [In 0.9.6c-engine release:]
5292 Add support for SureWare crypto accelerator cards from
5293 Baltimore Technologies. (Use engine 'sureware')
5294 [Baltimore Technologies and Mark Cox]
5296 *) [In 0.9.6c-engine release:]
5297 Add support for crypto accelerator cards from Accelerated
5298 Encryption Processing, www.aep.ie. (Use engine 'aep')
5299 [AEP Inc. and Mark Cox]
5301 *) Add a configuration entry for gcc on UnixWare.
5302 [Gary Benson <gbenson@redhat.com>]
5304 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5305 messages are stored in a single piece (fixed-length part and
5306 variable-length part combined) and fix various bugs found on the way.
5309 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5310 instead. BIO_gethostbyname() does not know what timeouts are
5311 appropriate, so entries would stay in cache even when they have
5313 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5315 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5316 faced with a pathologically small ClientHello fragment that does
5317 not contain client_version: Instead of aborting with an error,
5318 simply choose the highest available protocol version (i.e.,
5319 TLS 1.0 unless it is disabled). In practice, ClientHello
5320 messages are never sent like this, but this change gives us
5321 strictly correct behaviour at least for TLS.
5324 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5325 never resets s->method to s->ctx->method when called from within
5326 one of the SSL handshake functions.
5327 [Bodo Moeller; problem pointed out by Niko Baric]
5329 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5330 (sent using the client's version number) if client_version is
5331 smaller than the protocol version in use. Also change
5332 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5333 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5334 the client will at least see that alert.
5337 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5341 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5342 client receives HelloRequest while in a handshake.
5343 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5345 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5346 should end in 'break', not 'goto end' which circuments various
5347 cleanups done in state SSL_ST_OK. But session related stuff
5348 must be disabled for SSL_ST_OK in the case that we just sent a
5351 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5352 before just sending a HelloRequest.
5353 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5355 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5356 reveal whether illegal block cipher padding was found or a MAC
5357 verification error occured. (Neither SSLerr() codes nor alerts
5358 are directly visible to potential attackers, but the information
5359 may leak via logfiles.)
5361 Similar changes are not required for the SSL 2.0 implementation
5362 because the number of padding bytes is sent in clear for SSL 2.0,
5363 and the extra bytes are just ignored. However ssl/s2_pkt.c
5364 failed to verify that the purported number of padding bytes is in
5368 *) Add OpenUNIX-8 support including shared libraries
5369 (Boyd Lynn Gerber <gerberb@zenez.com>).
5372 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5373 'wristwatch attack' using huge encoding parameters (cf.
5374 James H. Manger's CRYPTO 2001 paper). Note that the
5375 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5376 encoding parameters and hence was not vulnerable.
5379 *) BN_sqr() bug fix.
5380 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5382 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5383 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5384 followed by modular reduction.
5385 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5387 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5388 equivalent based on BN_pseudo_rand() instead of BN_rand().
5391 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5392 This function was broken, as the check for a new client hello message
5393 to handle SGC did not allow these large messages.
5394 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5397 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5400 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5401 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5404 *) Rework the configuration and shared library support for Tru64 Unix.
5405 The configuration part makes use of modern compiler features and
5406 still retains old compiler behavior for those that run older versions
5407 of the OS. The shared library support part includes a variant that
5408 uses the RPATH feature, and is available through the special
5409 configuration target "alpha-cc-rpath", which will never be selected
5411 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5413 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5414 with the same message size as in ssl3_get_certificate_request().
5415 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5416 messages might inadvertently be reject as too long.
5417 [Petr Lampa <lampa@fee.vutbr.cz>]
5419 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5422 *) Modified SSL library such that the verify_callback that has been set
5423 specificly for an SSL object with SSL_set_verify() is actually being
5424 used. Before the change, a verify_callback set with this function was
5425 ignored and the verify_callback() set in the SSL_CTX at the time of
5426 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5427 to allow the necessary settings.
5430 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5431 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5432 done automatically (in contradiction to the requirements of the C
5433 standard). This made problems when used from OpenSSH.
5436 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5437 dh->length and always used
5439 BN_rand_range(priv_key, dh->p).
5441 BN_rand_range() is not necessary for Diffie-Hellman, and this
5442 specific range makes Diffie-Hellman unnecessarily inefficient if
5443 dh->length (recommended exponent length) is much smaller than the
5444 length of dh->p. We could use BN_rand_range() if the order of
5445 the subgroup was stored in the DH structure, but we only have
5450 BN_rand(priv_key, l, ...)
5452 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5458 RSA_eay_public_encrypt
5459 RSA_eay_private_decrypt
5460 RSA_eay_private_encrypt (signing)
5461 RSA_eay_public_decrypt (signature verification)
5463 (default implementations for RSA_public_encrypt,
5464 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5465 always reject numbers >= n.
5468 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5469 to synchronize access to 'locking_thread'. This is necessary on
5470 systems where access to 'locking_thread' (an 'unsigned long'
5471 variable) is not atomic.
5474 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5475 *before* setting the 'crypto_lock_rand' flag. The previous code had
5476 a race condition if 0 is a valid thread ID.
5477 [Travis Vitek <vitek@roguewave.com>]
5479 *) Add support for shared libraries under Irix.
5480 [Albert Chin-A-Young <china@thewrittenword.com>]
5482 *) Add configuration option to build on Linux on both big-endian and
5484 [Ralf Baechle <ralf@uni-koblenz.de>]
5486 *) Add the possibility to create shared libraries on HP-UX.
5489 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5491 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5492 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5493 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5494 PRNG state recovery was possible based on the output of
5495 one PRNG request appropriately sized to gain knowledge on
5496 'md' followed by enough consecutive 1-byte PRNG requests
5497 to traverse all of 'state'.
5499 1. When updating 'md_local' (the current thread's copy of 'md')
5500 during PRNG output generation, hash all of the previous
5501 'md_local' value, not just the half used for PRNG output.
5503 2. Make the number of bytes from 'state' included into the hash
5504 independent from the number of PRNG bytes requested.
5506 The first measure alone would be sufficient to avoid
5507 Markku-Juhani's attack. (Actually it had never occurred
5508 to me that the half of 'md_local' used for chaining was the
5509 half from which PRNG output bytes were taken -- I had always
5510 assumed that the secret half would be used.) The second
5511 measure makes sure that additional data from 'state' is never
5512 mixed into 'md_local' in small portions; this heuristically
5513 further strengthens the PRNG.
5516 *) Fix crypto/bn/asm/mips3.s.
5519 *) When only the key is given to "enc", the IV is undefined. Print out
5520 an error message in this case.
5523 *) Handle special case when X509_NAME is empty in X509 printing routines.
5526 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5527 positive and less than q.
5530 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5531 used: it isn't thread safe and the add_lock_callback should handle
5533 [Paul Rose <Paul.Rose@bridge.com>]
5535 *) Verify that incoming data obeys the block size in
5536 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5540 [Ulf Möller, Bodo Möller]
5542 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5543 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5544 when fixing the server behaviour for backwards-compatible 'client
5545 hello' messages. (Note that the attack is impractical against
5546 SSL 3.0 and TLS 1.0 anyway because length and version checking
5547 means that the probability of guessing a valid ciphertext is
5548 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5551 Before 0.9.5, the countermeasure (hide the error by generating a
5552 random 'decryption result') did not work properly because
5553 ERR_clear_error() was missing, meaning that SSL_get_error() would
5554 detect the supposedly ignored error.
5556 Both problems are now fixed.
5559 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5560 (previously it was 1024).
5563 *) Fix for compatibility mode trust settings: ignore trust settings
5564 unless some valid trust or reject settings are present.
5567 *) Fix for blowfish EVP: its a variable length cipher.
5570 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5571 parameters in DSA public key structures and return an error in the
5572 DSA routines if parameters are absent.
5575 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5576 in the current directory if neither $RANDFILE nor $HOME was set.
5577 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5578 caused some confusion to Windows users who haven't defined $HOME.
5579 Thus RAND_file_name() is changed again: e_os.h can define a
5580 DEFAULT_HOME, which will be used if $HOME is not set.
5581 For Windows, we use "C:"; on other platforms, we still require
5582 environment variables.
5584 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5585 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5586 having multiple threads call RAND_poll() concurrently.
5589 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5590 combination of a flag and a thread ID variable.
5591 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5592 flag), *other* threads can enter ssleay_add_bytes without obeying
5593 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5594 that they do not hold after the first thread unsets add_do_not_lock).
5597 *) Change bctest again: '-x' expressions are not available in all
5601 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5603 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5604 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5606 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5607 the default extension for executables, if any. Also, make the perl
5608 scripts that use symlink() to test if it really exists and use "cp"
5609 if it doesn't. All this made OpenSSL compilable and installable in
5613 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5614 If SEQUENCE is length is indefinite just set c->slen to the total
5615 amount of data available.
5616 [Steve Henson, reported by shige@FreeBSD.org]
5617 [This change does not apply to 0.9.7.]
5619 *) Change bctest to avoid here-documents inside command substitution
5620 (workaround for FreeBSD /bin/sh bug).
5621 For compatibility with Ultrix, avoid shell functions (introduced
5622 in the bctest version that searches along $PATH).
5625 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5626 with des_encrypt() defined on some operating systems, like Solaris
5630 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5631 On the Importance of Eliminating Errors in Cryptographic
5632 Computations, J. Cryptology 14 (2001) 2, 101-119,
5633 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5636 *) MIPS assembler BIGNUM division bug fix.
5639 *) Disabled incorrect Alpha assembler code.
5642 *) Fix PKCS#7 decode routines so they correctly update the length
5643 after reading an EOC for the EXPLICIT tag.
5645 [This change does not apply to 0.9.7.]
5647 *) Fix bug in PKCS#12 key generation routines. This was triggered
5648 if a 3DES key was generated with a 0 initial byte. Include
5649 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5650 (but broken) behaviour.
5653 *) Enhance bctest to search for a working bc along $PATH and print
5655 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5657 *) Fix memory leaks in err.c: free err_data string if necessary;
5658 don't write to the wrong index in ERR_set_error_data.
5661 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5665 *) Replace rdtsc with _emit statements for VC++ version 5.
5666 [Jeremy Cooper <jeremy@baymoo.org>]
5668 *) Make it possible to reuse SSLv2 sessions.
5671 *) In copy_email() check for >= 0 as a return value for
5672 X509_NAME_get_index_by_NID() since 0 is a valid index.
5673 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5675 *) Avoid coredump with unsupported or invalid public keys by checking if
5676 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5677 PKCS7_verify() fails with non detached data.
5680 *) Don't use getenv in library functions when run as setuid/setgid.
5681 New function OPENSSL_issetugid().
5684 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5685 due to incorrect handling of multi-threading:
5687 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5689 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5691 3. Count how many times MemCheck_off() has been called so that
5692 nested use can be treated correctly. This also avoids
5693 inband-signalling in the previous code (which relied on the
5694 assumption that thread ID 0 is impossible).
5697 *) Add "-rand" option also to s_client and s_server.
5700 *) Fix CPU detection on Irix 6.x.
5701 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5702 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5704 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5707 [This change does not apply to 0.9.7.]
5709 *) Use the cached encoding of an X509_NAME structure rather than
5710 copying it. This is apparently the reason for the libsafe "errors"
5711 but the code is actually correct.
5714 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5715 Bleichenbacher's DSA attack.
5716 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5717 to be set and top=0 forces the highest bit to be set; top=-1 is new
5718 and leaves the highest bit random.
5719 [Ulf Moeller, Bodo Moeller]
5721 *) In the NCONF_...-based implementations for CONF_... queries
5722 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5723 a temporary CONF structure with the data component set to NULL
5724 (which gives segmentation faults in lh_retrieve).
5725 Instead, use NULL for the CONF pointer in CONF_get_string and
5726 CONF_get_number (which may use environment variables) and directly
5727 return NULL from CONF_get_section.
5730 *) Fix potential buffer overrun for EBCDIC.
5733 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5734 keyUsage if basicConstraints absent for a CA.
5737 *) Make SMIME_write_PKCS7() write mail header values with a format that
5738 is more generally accepted (no spaces before the semicolon), since
5739 some programs can't parse those values properly otherwise. Also make
5740 sure BIO's that break lines after each write do not create invalid
5744 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5745 macros previously used would not encode an empty SEQUENCE OF
5746 and break the signature.
5748 [This change does not apply to 0.9.7.]
5750 *) Zero the premaster secret after deriving the master secret in
5754 *) Add some EVP_add_digest_alias registrations (as found in
5755 OpenSSL_add_all_digests()) to SSL_library_init()
5756 aka OpenSSL_add_ssl_algorithms(). This provides improved
5757 compatibility with peers using X.509 certificates
5758 with unconventional AlgorithmIdentifier OIDs.
5761 *) Fix for Irix with NO_ASM.
5762 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5764 *) ./config script fixes.
5765 [Ulf Moeller, Richard Levitte]
5767 *) Fix 'openssl passwd -1'.
5770 *) Change PKCS12_key_gen_asc() so it can cope with non null
5771 terminated strings whose length is passed in the passlen
5772 parameter, for example from PEM callbacks. This was done
5773 by adding an extra length parameter to asc2uni().
5774 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5776 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5777 call failed, free the DSA structure.
5780 *) Fix to uni2asc() to cope with zero length Unicode strings.
5781 These are present in some PKCS#12 files.
5784 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5785 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5786 when writing a 32767 byte record.
5787 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5789 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5790 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5792 (RSA objects have a reference count access to which is protected
5793 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5794 so they are meant to be shared between threads.)
5795 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5796 "Reddie, Steven" <Steven.Reddie@ca.com>]
5798 *) Fix a deadlock in CRYPTO_mem_leaks().
5801 *) Use better test patterns in bntest.
5804 *) rand_win.c fix for Borland C.
5807 *) BN_rshift bugfix for n == 0.
5810 *) Add a 'bctest' script that checks for some known 'bc' bugs
5811 so that 'make test' does not abort just because 'bc' is broken.
5814 *) Store verify_result within SSL_SESSION also for client side to
5815 avoid potential security hole. (Re-used sessions on the client side
5816 always resulted in verify_result==X509_V_OK, not using the original
5817 result of the server certificate verification.)
5820 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5821 SSL3_RT_APPLICATION_DATA, return 0.
5822 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5826 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5827 releases, have been re-implemented by renaming the previous
5828 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5829 and ssl3_read_internal, respectively, and adding 'peek' parameters
5830 to them. The new ssl[23]_{read,peek} functions are calls to
5831 ssl[23]_read_internal with the 'peek' flag set appropriately.
5832 A 'peek' parameter has also been added to ssl3_read_bytes, which
5833 does the actual work for ssl3_read_internal.
5836 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5837 the method-specific "init()" handler. Also clean up ex_data after
5838 calling the method-specific "finish()" handler. Previously, this was
5839 happening the other way round.
5842 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5843 The previous value, 12, was not always sufficient for BN_mod_exp().
5846 *) Make sure that shared libraries get the internal name engine with
5847 the full version number and not just 0. This should mark the
5848 shared libraries as not backward compatible. Of course, this should
5849 be changed again when we can guarantee backward binary compatibility.
5852 *) Fix typo in get_cert_by_subject() in by_dir.c
5853 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5855 *) Rework the system to generate shared libraries:
5857 - Make note of the expected extension for the shared libraries and
5858 if there is a need for symbolic links from for example libcrypto.so.0
5859 to libcrypto.so.0.9.7. There is extended info in Configure for
5862 - Make as few rebuilds of the shared libraries as possible.
5864 - Still avoid linking the OpenSSL programs with the shared libraries.
5866 - When installing, install the shared libraries separately from the
5870 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5872 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5873 and not in SSL_clear because the latter is also used by the
5874 accept/connect functions; previously, the settings made by
5875 SSL_set_read_ahead would be lost during the handshake.
5876 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5878 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5879 Previously, it would create entries for disableed algorithms no
5883 *) Added several new manual pages for SSL_* function.
5886 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5888 *) In ssl23_get_client_hello, generate an error message when faced
5889 with an initial SSL 3.0/TLS record that is too small to contain the
5890 first two bytes of the ClientHello message, i.e. client_version.
5891 (Note that this is a pathologic case that probably has never happened
5892 in real life.) The previous approach was to use the version number
5893 from the record header as a substitute; but our protocol choice
5894 should not depend on that one because it is not authenticated
5895 by the Finished messages.
5898 *) More robust randomness gathering functions for Windows.
5899 [Jeffrey Altman <jaltman@columbia.edu>]
5901 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5902 not set then we don't setup the error code for issuer check errors
5903 to avoid possibly overwriting other errors which the callback does
5904 handle. If an application does set the flag then we assume it knows
5905 what it is doing and can handle the new informational codes
5909 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5910 a general "ANY" type, as such it should be able to decode anything
5911 including tagged types. However it didn't check the class so it would
5912 wrongly interpret tagged types in the same way as their universal
5913 counterpart and unknown types were just rejected. Changed so that the
5914 tagged and unknown types are handled in the same way as a SEQUENCE:
5915 that is the encoding is stored intact. There is also a new type
5916 "V_ASN1_OTHER" which is used when the class is not universal, in this
5917 case we have no idea what the actual type is so we just lump them all
5921 *) On VMS, stdout may very well lead to a file that is written to
5922 in a record-oriented fashion. That means that every write() will
5923 write a separate record, which will be read separately by the
5924 programs trying to read from it. This can be very confusing.
5926 The solution is to put a BIO filter in the way that will buffer
5927 text until a linefeed is reached, and then write everything a
5928 line at a time, so every record written will be an actual line,
5929 not chunks of lines and not (usually doesn't happen, but I've
5930 seen it once) several lines in one record. BIO_f_linebuffer() is
5933 Currently, it's a VMS-only method, because that's where it has
5934 been tested well enough.
5937 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5938 it can return incorrect results.
5939 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5940 but it was in 0.9.6-beta[12].)
5943 *) Disable the check for content being present when verifying detached
5944 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5945 include zero length content when signing messages.
5948 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5949 BIO_ctrl (for BIO pairs).
5952 *) Add DSO method for VMS.
5955 *) Bug fix: Montgomery multiplication could produce results with the
5959 *) Add RPM specification openssl.spec and modify it to build three
5960 packages. The default package contains applications, application
5961 documentation and run-time libraries. The devel package contains
5962 include files, static libraries and function documentation. The
5963 doc package contains the contents of the doc directory. The original
5964 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5967 *) Add a large number of documentation files for many SSL routines.
5968 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5970 *) Add a configuration entry for Sony News 4.
5971 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5973 *) Don't set the two most significant bits to one when generating a
5974 random number < q in the DSA library.
5977 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5978 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5979 the underlying transport is blocking) if a handshake took place.
5980 (The default behaviour is needed by applications such as s_client
5981 and s_server that use select() to determine when to use SSL_read;
5982 but for applications that know in advance when to expect data, it
5983 just makes things more complicated.)
5986 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5990 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5991 work better on such systems.
5992 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5994 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5995 Update PKCS12_parse() so it copies the friendlyName and the
5996 keyid to the certificates aux info.
5999 *) Fix bug in PKCS7_verify() which caused an infinite loop
6000 if there was more than one signature.
6001 [Sven Uszpelkat <su@celocom.de>]
6003 *) Major change in util/mkdef.pl to include extra information
6004 about each symbol, as well as presentig variables as well
6005 as functions. This change means that there's n more need
6006 to rebuild the .num files when some algorithms are excluded.
6009 *) Allow the verify time to be set by an application,
6010 rather than always using the current time.
6013 *) Phase 2 verify code reorganisation. The certificate
6014 verify code now looks up an issuer certificate by a
6015 number of criteria: subject name, authority key id
6016 and key usage. It also verifies self signed certificates
6017 by the same criteria. The main comparison function is
6018 X509_check_issued() which performs these checks.
6020 Lot of changes were necessary in order to support this
6021 without completely rewriting the lookup code.
6023 Authority and subject key identifier are now cached.
6025 The LHASH 'certs' is X509_STORE has now been replaced
6026 by a STACK_OF(X509_OBJECT). This is mainly because an
6027 LHASH can't store or retrieve multiple objects with
6028 the same hash value.
6030 As a result various functions (which were all internal
6031 use only) have changed to handle the new X509_STORE
6032 structure. This will break anything that messed round
6033 with X509_STORE internally.
6035 The functions X509_STORE_add_cert() now checks for an
6036 exact match, rather than just subject name.
6038 The X509_STORE API doesn't directly support the retrieval
6039 of multiple certificates matching a given criteria, however
6040 this can be worked round by performing a lookup first
6041 (which will fill the cache with candidate certificates)
6042 and then examining the cache for matches. This is probably
6043 the best we can do without throwing out X509_LOOKUP
6044 entirely (maybe later...).
6046 The X509_VERIFY_CTX structure has been enhanced considerably.
6048 All certificate lookup operations now go via a get_issuer()
6049 callback. Although this currently uses an X509_STORE it
6050 can be replaced by custom lookups. This is a simple way
6051 to bypass the X509_STORE hackery necessary to make this
6052 work and makes it possible to use more efficient techniques
6053 in future. A very simple version which uses a simple
6054 STACK for its trusted certificate store is also provided
6055 using X509_STORE_CTX_trusted_stack().
6057 The verify_cb() and verify() callbacks now have equivalents
6058 in the X509_STORE_CTX structure.
6060 X509_STORE_CTX also has a 'flags' field which can be used
6061 to customise the verify behaviour.
6064 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6065 excludes S/MIME capabilities.
6068 *) When a certificate request is read in keep a copy of the
6069 original encoding of the signed data and use it when outputing
6070 again. Signatures then use the original encoding rather than
6071 a decoded, encoded version which may cause problems if the
6072 request is improperly encoded.
6075 *) For consistency with other BIO_puts implementations, call
6076 buffer_write(b, ...) directly in buffer_puts instead of calling
6079 In BIO_puts, increment b->num_write as in BIO_write.
6080 [Peter.Sylvester@EdelWeb.fr]
6082 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6083 BN_zero, we may not return a BIGNUM with an array consisting of
6087 *) Avoid calling abort() from within the library when problems are
6088 detected, except if preprocessor symbols have been defined
6089 (such as REF_CHECK, BN_DEBUG etc.).
6092 *) New openssl application 'rsautl'. This utility can be
6093 used for low level RSA operations. DER public key
6094 BIO/fp routines also added.
6097 *) New Configure entry and patches for compiling on QNX 4.
6098 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6100 *) A demo state-machine implementation was sponsored by
6101 Nuron (http://www.nuron.com/) and is now available in
6102 demos/state_machine.
6105 *) New options added to the 'dgst' utility for signature
6106 generation and verification.
6109 *) Unrecognized PKCS#7 content types are now handled via a
6110 catch all ASN1_TYPE structure. This allows unsupported
6111 types to be stored as a "blob" and an application can
6112 encode and decode it manually.
6115 *) Fix various signed/unsigned issues to make a_strex.c
6117 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6119 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6120 length if passed a buffer. ASN1_INTEGER_to_BN failed
6121 if passed a NULL BN and its argument was negative.
6122 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6124 *) Modification to PKCS#7 encoding routines to output definite
6125 length encoding. Since currently the whole structures are in
6126 memory there's not real point in using indefinite length
6127 constructed encoding. However if OpenSSL is compiled with
6128 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6131 *) Added BIO_vprintf() and BIO_vsnprintf().
6134 *) Added more prefixes to parse for in the the strings written
6135 through a logging bio, to cover all the levels that are available
6136 through syslog. The prefixes are now:
6138 PANIC, EMERG, EMR => LOG_EMERG
6139 ALERT, ALR => LOG_ALERT
6140 CRIT, CRI => LOG_CRIT
6141 ERROR, ERR => LOG_ERR
6142 WARNING, WARN, WAR => LOG_WARNING
6143 NOTICE, NOTE, NOT => LOG_NOTICE
6144 INFO, INF => LOG_INFO
6145 DEBUG, DBG => LOG_DEBUG
6147 and as before, if none of those prefixes are present at the
6148 beginning of the string, LOG_ERR is chosen.
6150 On Win32, the LOG_* levels are mapped according to this:
6152 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6153 LOG_WARNING => EVENTLOG_WARNING_TYPE
6154 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6158 *) Made it possible to reconfigure with just the configuration
6159 argument "reconf" or "reconfigure". The command line arguments
6160 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6161 and are retrieved from there when reconfiguring.
6165 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6167 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6170 *) The obj_dat.pl script was messing up the sorting of object
6171 names. The reason was that it compared the quoted version
6172 of strings as a result "OCSP" > "OCSP Signing" because
6173 " > SPACE. Changed script to store unquoted versions of
6174 names and add quotes on output. It was also omitting some
6175 names from the lookup table if they were given a default
6176 value (that is if SN is missing it is given the same
6177 value as LN and vice versa), these are now added on the
6178 grounds that if an object has a name we should be able to
6179 look it up. Finally added warning output when duplicate
6180 short or long names are found.
6183 *) Changes needed for Tandem NSK.
6184 [Scott Uroff <scott@xypro.com>]
6186 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6187 RSA_padding_check_SSLv23(), special padding was never detected
6188 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6189 version rollback attacks was not effective.
6191 In s23_clnt.c, don't use special rollback-attack detection padding
6192 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6193 client; similarly, in s23_srvr.c, don't do the rollback check if
6194 SSL 2.0 is the only protocol enabled in the server.
6197 *) Make it possible to get hexdumps of unprintable data with 'openssl
6198 asn1parse'. By implication, the functions ASN1_parse_dump() and
6199 BIO_dump_indent() are added.
6202 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6203 these print out strings and name structures based on various
6204 flags including RFC2253 support and proper handling of
6205 multibyte characters. Added options to the 'x509' utility
6206 to allow the various flags to be set.
6209 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6210 Also change the functions X509_cmp_current_time() and
6211 X509_gmtime_adj() work with an ASN1_TIME structure,
6212 this will enable certificates using GeneralizedTime in validity
6213 dates to be checked.
6216 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6217 negative public key encodings) on by default,
6218 NO_NEG_PUBKEY_BUG can be set to disable it.
6221 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6222 content octets. An i2c_ASN1_OBJECT is unnecessary because
6223 the encoding can be trivially obtained from the structure.
6226 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6227 not read locks (CRYPTO_r_[un]lock).
6230 *) A first attempt at creating official support for shared
6231 libraries through configuration. I've kept it so the
6232 default is static libraries only, and the OpenSSL programs
6233 are always statically linked for now, but there are
6234 preparations for dynamic linking in place.
6235 This has been tested on Linux and Tru64.
6238 *) Randomness polling function for Win9x, as described in:
6239 Peter Gutmann, Software Generation of Practically Strong
6243 *) Fix so PRNG is seeded in req if using an already existing
6247 *) New options to smime application. -inform and -outform
6248 allow alternative formats for the S/MIME message including
6249 PEM and DER. The -content option allows the content to be
6250 specified separately. This should allow things like Netscape
6251 form signing output easier to verify.
6254 *) Fix the ASN1 encoding of tags using the 'long form'.
6257 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6258 STRING types. These convert content octets to and from the
6259 underlying type. The actual tag and length octets are
6260 already assumed to have been read in and checked. These
6261 are needed because all other string types have virtually
6262 identical handling apart from the tag. By having versions
6263 of the ASN1 functions that just operate on content octets
6264 IMPLICIT tagging can be handled properly. It also allows
6265 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6266 and ASN1_INTEGER are identical apart from the tag.
6269 *) Change the handling of OID objects as follows:
6271 - New object identifiers are inserted in objects.txt, following
6272 the syntax given in objects.README.
6273 - objects.pl is used to process obj_mac.num and create a new
6275 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6278 This is currently kind of a hack, and the perl code in objects.pl
6279 isn't very elegant, but it works as I intended. The simplest way
6280 to check that it worked correctly is to look in obj_dat.h and
6281 check the array nid_objs and make sure the objects haven't moved
6282 around (this is important!). Additions are OK, as well as
6283 consistent name changes.
6286 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6289 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6290 The given file adds to whatever has already been seeded into the
6291 random pool through the RANDFILE configuration file option or
6292 environment variable, or the default random state file.
6295 *) mkstack.pl now sorts each macro group into lexical order.
6296 Previously the output order depended on the order the files
6297 appeared in the directory, resulting in needless rewriting
6301 *) Patches to make OpenSSL compile under Win32 again. Mostly
6302 work arounds for the VC++ problem that it treats func() as
6303 func(void). Also stripped out the parts of mkdef.pl that
6304 added extra typesafe functions: these no longer exist.
6307 *) Reorganisation of the stack code. The macros are now all
6308 collected in safestack.h . Each macro is defined in terms of
6309 a "stack macro" of the form SKM_<name>(type, a, b). The
6310 DEBUG_SAFESTACK is now handled in terms of function casts,
6311 this has the advantage of retaining type safety without the
6312 use of additional functions. If DEBUG_SAFESTACK is not defined
6313 then the non typesafe macros are used instead. Also modified the
6314 mkstack.pl script to handle the new form. Needs testing to see
6315 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6316 the default if no major problems. Similar behaviour for ASN1_SET_OF
6317 and PKCS12_STACK_OF.
6320 *) When some versions of IIS use the 'NET' form of private key the
6321 key derivation algorithm is different. Normally MD5(password) is
6322 used as a 128 bit RC4 key. In the modified case
6323 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6324 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6325 as the old Netscape_RSA functions except they have an additional
6326 'sgckey' parameter which uses the modified algorithm. Also added
6327 an -sgckey command line option to the rsa utility. Thanks to
6328 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6329 algorithm to openssl-dev.
6332 *) The evp_local.h macros were using 'c.##kname' which resulted in
6333 invalid expansion on some systems (SCO 5.0.5 for example).
6334 Corrected to 'c.kname'.
6335 [Phillip Porch <root@theporch.com>]
6337 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6338 a STACK of email addresses from a certificate or request, these look
6339 in the subject name and the subject alternative name extensions and
6340 omit any duplicate addresses.
6343 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6344 This makes DSA verification about 2 % faster.
6347 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6348 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6349 plus overhead for 1024 bit moduli).
6350 This makes exponentiations about 0.5 % faster for 1024 bit
6351 exponents (as measured by "openssl speed rsa2048").
6354 *) Rename memory handling macros to avoid conflicts with other
6356 Malloc => OPENSSL_malloc
6357 Malloc_locked => OPENSSL_malloc_locked
6358 Realloc => OPENSSL_realloc
6359 Free => OPENSSL_free
6362 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6363 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6366 *) CygWin32 support.
6367 [John Jarvie <jjarvie@newsguy.com>]
6369 *) The type-safe stack code has been rejigged. It is now only compiled
6370 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6371 by default all type-specific stack functions are "#define"d back to
6372 standard stack functions. This results in more streamlined output
6373 but retains the type-safety checking possibilities of the original
6377 *) The STACK code has been cleaned up, and certain type declarations
6378 that didn't make a lot of sense have been brought in line. This has
6379 also involved a cleanup of sorts in safestack.h to more correctly
6380 map type-safe stack functions onto their plain stack counterparts.
6381 This work has also resulted in a variety of "const"ifications of
6382 lots of the code, especially "_cmp" operations which should normally
6383 be prototyped with "const" parameters anyway.
6386 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6387 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6388 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6389 where all of 'md' is used each time the PRNG is used, but 'state'
6390 is used only indexed by a cyclic counter. As entropy may not be
6391 well distributed from the beginning, 'md' is important as a
6392 chaining variable. However, the output function chains only half
6393 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6394 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6395 in all of 'state' being rewritten, with the new values depending
6396 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6399 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6400 the handshake is continued after ssl_verify_cert_chain();
6401 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6402 can lead to 'unexplainable' connection aborts later.
6403 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6405 *) Major EVP API cipher revision.
6406 Add hooks for extra EVP features. This allows various cipher
6407 parameters to be set in the EVP interface. Support added for variable
6408 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6409 setting of RC2 and RC5 parameters.
6411 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6414 Remove lots of duplicated code from the EVP library. For example *every*
6415 cipher init() function handles the 'iv' in the same way according to the
6416 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6417 for CFB and OFB modes they zero ctx->num.
6419 New functionality allows removal of S/MIME code RC2 hack.
6421 Most of the routines have the same form and so can be declared in terms
6424 By shifting this to the top level EVP_CipherInit() it can be removed from
6425 all individual ciphers. If the cipher wants to handle IVs or keys
6426 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6429 Change lots of functions like EVP_EncryptUpdate() to now return a
6430 value: although software versions of the algorithms cannot fail
6431 any installed hardware versions can.
6434 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6435 this option is set, tolerate broken clients that send the negotiated
6436 protocol version number instead of the requested protocol version
6440 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6441 i.e. non-zero for export ciphersuites, zero otherwise.
6442 Previous versions had this flag inverted, inconsistent with
6443 rsa_tmp_cb (..._TMP_RSA_CB).
6444 [Bodo Moeller; problem reported by Amit Chopra]
6446 *) Add missing DSA library text string. Work around for some IIS
6447 key files with invalid SEQUENCE encoding.
6450 *) Add a document (doc/standards.txt) that list all kinds of standards
6451 and so on that are implemented in OpenSSL.
6454 *) Enhance c_rehash script. Old version would mishandle certificates
6455 with the same subject name hash and wouldn't handle CRLs at all.
6456 Added -fingerprint option to crl utility, to support new c_rehash
6460 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6463 *) Fix for SSL server purpose checking. Server checking was
6464 rejecting certificates which had extended key usage present
6465 but no ssl client purpose.
6466 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6468 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6469 is a little unclear about how a blank password is handled.
6470 Since the password in encoded as a BMPString with terminating
6471 double NULL a zero length password would end up as just the
6472 double NULL. However no password at all is different and is
6473 handled differently in the PKCS#12 key generation code. NS
6474 treats a blank password as zero length. MSIE treats it as no
6475 password on export: but it will try both on import. We now do
6476 the same: PKCS12_parse() tries zero length and no password if
6477 the password is set to "" or NULL (NULL is now a valid password:
6478 it wasn't before) as does the pkcs12 application.
6481 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6482 perror when PEM_read_bio_X509_REQ fails, the error message must
6483 be obtained from the error queue.
6486 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6487 it in ERR_remove_state if appropriate, and change ERR_get_state
6488 accordingly to avoid race conditions (this is necessary because
6489 thread_hash is no longer constant once set).
6492 *) Bugfix for linux-elf makefile.one.
6495 *) RSA_get_default_method() will now cause a default
6496 RSA_METHOD to be chosen if one doesn't exist already.
6497 Previously this was only set during a call to RSA_new()
6498 or RSA_new_method(NULL) meaning it was possible for
6499 RSA_get_default_method() to return NULL.
6502 *) Added native name translation to the existing DSO code
6503 that will convert (if the flag to do so is set) filenames
6504 that are sufficiently small and have no path information
6505 into a canonical native form. Eg. "blah" converted to
6506 "libblah.so" or "blah.dll" etc.
6509 *) New function ERR_error_string_n(e, buf, len) which is like
6510 ERR_error_string(e, buf), but writes at most 'len' bytes
6511 including the 0 terminator. For ERR_error_string_n, 'buf'
6513 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6515 *) CONF library reworked to become more general. A new CONF
6516 configuration file reader "class" is implemented as well as a
6517 new functions (NCONF_*, for "New CONF") to handle it. The now
6518 old CONF_* functions are still there, but are reimplemented to
6519 work in terms of the new functions. Also, a set of functions
6520 to handle the internal storage of the configuration data is
6521 provided to make it easier to write new configuration file
6522 reader "classes" (I can definitely see something reading a
6523 configuration file in XML format, for example), called _CONF_*,
6524 or "the configuration storage API"...
6526 The new configuration file reading functions are:
6528 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6529 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6531 NCONF_default, NCONF_WIN32
6533 NCONF_dump_fp, NCONF_dump_bio
6535 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6536 NCONF_new creates a new CONF object. This works in the same way
6537 as other interfaces in OpenSSL, like the BIO interface.
6538 NCONF_dump_* dump the internal storage of the configuration file,
6539 which is useful for debugging. All other functions take the same
6540 arguments as the old CONF_* functions wth the exception of the
6541 first that must be a `CONF *' instead of a `LHASH *'.
6543 To make it easer to use the new classes with the old CONF_* functions,
6544 the function CONF_set_default_method is provided.
6547 *) Add '-tls1' option to 'openssl ciphers', which was already
6548 mentioned in the documentation but had not been implemented.
6549 (This option is not yet really useful because even the additional
6550 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6553 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6554 OpenSSL-based applications) load shared libraries and bind to
6555 them in a portable way.
6556 [Geoff Thorpe, with contributions from Richard Levitte]
6558 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6560 *) Make sure _lrotl and _lrotr are only used with MSVC.
6562 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6563 (the default implementation of RAND_status).
6565 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6566 to '-clrext' (= clear extensions), as intended and documented.
6567 [Bodo Moeller; inconsistency pointed out by Michael Attili
6570 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6571 was larger than the MD block size.
6572 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6574 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6575 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6576 using the passed key: if the passed key was a private key the result
6577 of X509_print(), for example, would be to print out all the private key
6581 *) des_quad_cksum() byte order bug fix.
6582 [Ulf Möller, using the problem description in krb4-0.9.7, where
6583 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6585 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6587 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6589 *) For easily testing in shell scripts whether some command
6590 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6591 returns with exit code 0 iff no command of the given name is available.
6592 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6593 the output goes to stdout and nothing is printed to stderr.
6594 Additional arguments are always ignored.
6596 Since for each cipher there is a command of the same name,
6597 the 'no-cipher' compilation switches can be tested this way.
6599 ('openssl no-XXX' is not able to detect pseudo-commands such
6600 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6603 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6606 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6607 is set; it will be thrown away anyway because each handshake creates
6609 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6610 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6611 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6612 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6615 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6616 'Q' and 'R' lose their special meanings (quit/renegotiate).
6617 This is part of what -quiet does; unlike -quiet, -ign_eof
6618 does not suppress any output.
6621 *) Add compatibility options to the purpose and trust code. The
6622 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6623 accepts a certificate or CA, this was the previous behaviour,
6624 with all the associated security issues.
6626 X509_TRUST_COMPAT is the old trust behaviour: only and
6627 automatically trust self signed roots in certificate store. A
6628 new trust setting X509_TRUST_DEFAULT is used to specify that
6629 a purpose has no associated trust setting and it should instead
6630 use the value in the default purpose.
6633 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6634 and fix a memory leak.
6637 *) In util/mkerr.pl (which implements 'make errors'), preserve
6638 reason strings from the previous version of the .c file, as
6639 the default to have only downcase letters (and digits) in
6640 automatically generated reasons codes is not always appropriate.
6643 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6644 using strerror. Previously, ERR_reason_error_string() returned
6645 library names as reason strings for SYSerr; but SYSerr is a special
6646 case where small numbers are errno values, not library numbers.
6649 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6650 converts DSA parameters into DH parameters. (When creating parameters,
6651 DSA_generate_parameters is used.)
6654 *) Include 'length' (recommended exponent length) in C code generated
6655 by 'openssl dhparam -C'.
6658 *) The second argument to set_label in perlasm was already being used
6659 so couldn't be used as a "file scope" flag. Moved to third argument
6663 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6664 instead of RAND_bytes for encryption IVs and salts.
6667 *) Include RAND_status() into RAND_METHOD instead of implementing
6668 it only for md_rand.c Otherwise replacing the PRNG by calling
6669 RAND_set_rand_method would be impossible.
6672 *) Don't let DSA_generate_key() enter an infinite loop if the random
6673 number generation fails.
6676 *) New 'rand' application for creating pseudo-random output.
6679 *) Added configuration support for Linux/IA64
6680 [Rolf Haberrecker <rolf@suse.de>]
6682 *) Assembler module support for Mingw32.
6685 *) Shared library support for HPUX (in shlib/).
6686 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6688 *) Shared library support for Solaris gcc.
6689 [Lutz Behnke <behnke@trustcenter.de>]
6691 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6693 *) PKCS7_encrypt() was adding text MIME headers twice because they
6694 were added manually and by SMIME_crlf_copy().
6697 *) In bntest.c don't call BN_rand with zero bits argument.
6698 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6700 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6701 case was implemented. This caused BN_div_recp() to fail occasionally.
6704 *) Add an optional second argument to the set_label() in the perl
6705 assembly language builder. If this argument exists and is set
6706 to 1 it signals that the assembler should use a symbol whose
6707 scope is the entire file, not just the current function. This
6708 is needed with MASM which uses the format label:: for this scope.
6709 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6711 *) Change the ASN1 types so they are typedefs by default. Before
6712 almost all types were #define'd to ASN1_STRING which was causing
6713 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6717 *) Change names of new functions to the new get1/get0 naming
6718 convention: After 'get1', the caller owns a reference count
6719 and has to call ..._free; 'get0' returns a pointer to some
6720 data structure without incrementing reference counters.
6721 (Some of the existing 'get' functions increment a reference
6722 counter, some don't.)
6723 Similarly, 'set1' and 'add1' functions increase reference
6724 counters or duplicate objects.
6727 *) Allow for the possibility of temp RSA key generation failure:
6728 the code used to assume it always worked and crashed on failure.
6731 *) Fix potential buffer overrun problem in BIO_printf().
6732 [Ulf Möller, using public domain code by Patrick Powell; problem
6733 pointed out by David Sacerdote <das33@cornell.edu>]
6735 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6736 RAND_egd() and RAND_status(). In the command line application,
6737 the EGD socket can be specified like a seed file using RANDFILE
6741 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6742 Some CAs (e.g. Verisign) distribute certificates in this form.
6745 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6746 list to exclude them. This means that no special compilation option
6747 is needed to use anonymous DH: it just needs to be included in the
6751 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6752 EVP_MD_type. The old functionality is available in a new macro called
6753 EVP_MD_md(). Change code that uses it and update docs.
6756 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6757 where the 'void *' argument is replaced by a function pointer argument.
6758 Previously 'void *' was abused to point to functions, which works on
6759 many platforms, but is not correct. As these functions are usually
6760 called by macros defined in OpenSSL header files, most source code
6761 should work without changes.
6764 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6765 sections with information on -D... compiler switches used for
6766 compiling the library so that applications can see them. To enable
6767 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6768 must be defined. E.g.,
6769 #define OPENSSL_ALGORITHM_DEFINES
6770 #include <openssl/opensslconf.h>
6771 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6772 [Richard Levitte, Ulf and Bodo Möller]
6774 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6778 *) Change the 'other' type in certificate aux info to a STACK_OF
6779 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6780 the required ASN1 format: arbitrary types determined by an OID.
6783 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6784 argument to 'req'. This is not because the function is newer or
6785 better than others it just uses the work 'NEW' in the certificate
6786 request header lines. Some software needs this.
6789 *) Reorganise password command line arguments: now passwords can be
6790 obtained from various sources. Delete the PEM_cb function and make
6791 it the default behaviour: i.e. if the callback is NULL and the
6792 usrdata argument is not NULL interpret it as a null terminated pass
6793 phrase. If usrdata and the callback are NULL then the pass phrase
6794 is prompted for as usual.
6797 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6798 the support is automatically enabled. The resulting binaries will
6799 autodetect the card and use it if present.
6800 [Ben Laurie and Compaq Inc.]
6802 *) Work around for Netscape hang bug. This sends certificate request
6803 and server done in one record. Since this is perfectly legal in the
6804 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6805 the bugs/SSLv3 entry for more info.
6808 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6811 *) Add -rand argument to smime and pkcs12 applications and read/write
6815 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6818 *) Add command line password options to the remaining applications.
6821 *) Bug fix for BN_div_recp() for numerators with an even number of
6825 *) More tests in bntest.c, and changed test_bn output.
6828 *) ./config recognizes MacOS X now.
6831 *) Bug fix for BN_div() when the first words of num and divsor are
6832 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6835 *) Add support for various broken PKCS#8 formats, and command line
6836 options to produce them.
6839 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6840 get temporary BIGNUMs from a BN_CTX.
6843 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6847 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6848 include a #define from the old name to the new. The original intent
6849 was that statically linked binaries could for example just call
6850 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6851 link with digests. This never worked becayse SSLeay_add_all_digests()
6852 and SSLeay_add_all_ciphers() were in the same source file so calling
6853 one would link with the other. They are now in separate source files.
6856 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6859 *) Use a less unusual form of the Miller-Rabin primality test (it used
6860 a binary algorithm for exponentiation integrated into the Miller-Rabin
6861 loop, our standard modexp algorithms are faster).
6864 *) Support for the EBCDIC character set completed.
6865 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6867 *) Source code cleanups: use const where appropriate, eliminate casts,
6868 use void * instead of char * in lhash.
6871 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6872 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6873 this the server could overwrite ephemeral keys that the client
6877 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6878 using 50 iterations of the Rabin-Miller test.
6880 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6881 iterations of the Rabin-Miller test as required by the appendix
6882 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6883 As BN_is_prime_fasttest includes trial division, DSA parameter
6884 generation becomes much faster.
6886 This implies a change for the callback functions in DSA_is_prime
6887 and DSA_generate_parameters: The callback function is called once
6888 for each positive witness in the Rabin-Miller test, not just
6889 occasionally in the inner loop; and the parameters to the
6890 callback function now provide an iteration count for the outer
6891 loop rather than for the current invocation of the inner loop.
6892 DSA_generate_parameters additionally can call the callback
6893 function with an 'iteration count' of -1, meaning that a
6894 candidate has passed the trial division test (when q is generated
6895 from an application-provided seed, trial division is skipped).
6898 *) New function BN_is_prime_fasttest that optionally does trial
6899 division before starting the Rabin-Miller test and has
6900 an additional BN_CTX * argument (whereas BN_is_prime always
6901 has to allocate at least one BN_CTX).
6902 'callback(1, -1, cb_arg)' is called when a number has passed the
6903 trial division stage.
6906 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6910 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6913 *) New function BN_pseudo_rand().
6916 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6917 bignum version of BN_from_montgomery() with the working code from
6918 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6922 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6923 made it impossible to use the same SSL_SESSION data structure in
6924 SSL2 clients in multiple threads.
6927 *) The return value of RAND_load_file() no longer counts bytes obtained
6928 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6929 to seed the PRNG (previously an explicit byte count was required).
6930 [Ulf Möller, Bodo Möller]
6932 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6933 used (char *) instead of (void *) and had casts all over the place.
6936 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6939 *) Retain source code compatibility for BN_prime_checks macro:
6940 BN_is_prime(..., BN_prime_checks, ...) now uses
6941 BN_prime_checks_for_size to determine the appropriate number of
6942 Rabin-Miller iterations.
6945 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6946 DH_CHECK_P_NOT_SAFE_PRIME.
6947 (Check if this is true? OpenPGP calls them "strong".)
6950 *) Merge the functionality of "dh" and "gendh" programs into a new program
6951 "dhparam". The old programs are retained for now but will handle DH keys
6952 (instead of parameters) in future.
6955 *) Make the ciphers, s_server and s_client programs check the return values
6956 when a new cipher list is set.
6959 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6960 ciphers. Before when the 56bit ciphers were enabled the sorting was
6963 The syntax for the cipher sorting has been extended to support sorting by
6964 cipher-strength (using the strength_bits hard coded in the tables).
6965 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6967 Fix a bug in the cipher-command parser: when supplying a cipher command
6968 string with an "undefined" symbol (neither command nor alphanumeric
6969 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6970 an error is flagged.
6972 Due to the strength-sorting extension, the code of the
6973 ssl_create_cipher_list() function was completely rearranged. I hope that
6974 the readability was also increased :-)
6975 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6977 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6978 for the first serial number and places 2 in the serial number file. This
6979 avoids problems when the root CA is created with serial number zero and
6980 the first user certificate has the same issuer name and serial number
6984 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6985 the new code. Add documentation for this stuff.
6988 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6989 X509_*() to X509at_*() on the grounds that they don't handle X509
6990 structures and behave in an analagous way to the X509v3 functions:
6991 they shouldn't be called directly but wrapper functions should be used
6994 So we also now have some wrapper functions that call the X509at functions
6995 when passed certificate requests. (TO DO: similar things can be done with
6996 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6997 things. Some of these need some d2i or i2d and print functionality
6998 because they handle more complex structures.)
7001 *) Add missing #ifndefs that caused missing symbols when building libssl
7002 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7003 NO_RSA in ssl/s2*.c.
7004 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7006 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7007 has a return value which indicates the quality of the random data
7008 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7009 error queue. New function RAND_pseudo_bytes() generates output that is
7010 guaranteed to be unique but not unpredictable. RAND_add is like
7011 RAND_seed, but takes an extra argument for an entropy estimate
7012 (RAND_seed always assumes full entropy).
7015 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7016 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7017 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7018 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7019 false-positive rate of at most 2^-80 for random input.
7022 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7025 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7026 in the 0.9.5 release), this returns the chain
7027 from an X509_CTX structure with a dup of the stack and all
7028 the X509 reference counts upped: so the stack will exist
7029 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7032 Also make SSL_SESSION_print() print out the verify return
7036 *) Add manpage for the pkcs12 command. Also change the default
7037 behaviour so MAC iteration counts are used unless the new
7038 -nomaciter option is used. This improves file security and
7039 only older versions of MSIE (4.0 for example) need it.
7042 *) Honor the no-xxx Configure options when creating .DEF files.
7045 *) Add PKCS#10 attributes to field table: challengePassword,
7046 unstructuredName and unstructuredAddress. These are taken from
7047 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7048 international characters are used.
7050 More changes to X509_ATTRIBUTE code: allow the setting of types
7051 based on strings. Remove the 'loc' parameter when adding
7052 attributes because these will be a SET OF encoding which is sorted
7056 *) Initial changes to the 'req' utility to allow request generation
7057 automation. This will allow an application to just generate a template
7058 file containing all the field values and have req construct the
7061 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7062 used all over the place including certificate requests and PKCS#7
7063 structures. They are currently handled manually where necessary with
7064 some primitive wrappers for PKCS#7. The new functions behave in a
7065 manner analogous to the X509 extension functions: they allow
7066 attributes to be looked up by NID and added.
7068 Later something similar to the X509V3 code would be desirable to
7069 automatically handle the encoding, decoding and printing of the
7070 more complex types. The string types like challengePassword can
7071 be handled by the string table functions.
7073 Also modified the multi byte string table handling. Now there is
7074 a 'global mask' which masks out certain types. The table itself
7075 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7076 is useful when for example there is only one permissible type
7077 (as in countryName) and using the mask might result in no valid
7081 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7082 SSL_get_peer_finished to allow applications to obtain the latest
7083 Finished messages sent to the peer or expected from the peer,
7084 respectively. (SSL_get_peer_finished is usually the Finished message
7085 actually received from the peer, otherwise the protocol will be aborted.)
7087 As the Finished message are message digests of the complete handshake
7088 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7089 be used for external authentication procedures when the authentication
7090 provided by SSL/TLS is not desired or is not enough.
7093 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7094 the host supports BWX extension and if Compaq C is present on the
7095 $PATH. Just exploiting of the BWX extension results in 20-30%
7096 performance kick for some algorithms, e.g. DES and RC4 to mention
7097 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7101 *) Add support for MS "fast SGC". This is arguably a violation of the
7102 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7103 weak crypto and after checking the certificate is SGC a second one
7104 with strong crypto. MS SGC stops the first handshake after receiving
7105 the server certificate message and sends a second client hello. Since
7106 a server will typically do all the time consuming operations before
7107 expecting any further messages from the client (server key exchange
7108 is the most expensive) there is little difference between the two.
7110 To get OpenSSL to support MS SGC we have to permit a second client
7111 hello message after we have sent server done. In addition we have to
7112 reset the MAC if we do get this second client hello.
7115 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7116 if a DER encoded private key is RSA or DSA traditional format. Changed
7117 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7118 format DER encoded private key. Newer code should use PKCS#8 format which
7119 has the key type encoded in the ASN1 structure. Added DER private key
7120 support to pkcs8 application.
7123 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7124 ciphersuites has been selected (as required by the SSL 3/TLS 1
7125 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7126 is set, we interpret this as a request to violate the specification
7127 (the worst that can happen is a handshake failure, and 'correct'
7128 behaviour would result in a handshake failure anyway).
7131 *) In SSL_CTX_add_session, take into account that there might be multiple
7132 SSL_SESSION structures with the same session ID (e.g. when two threads
7133 concurrently obtain them from an external cache).
7134 The internal cache can handle only one SSL_SESSION with a given ID,
7135 so if there's a conflict, we now throw out the old one to achieve
7139 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7140 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7141 some routines that use cipher OIDs: some ciphers do not have OIDs
7142 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7146 *) Simplify the trust setting structure and code. Now we just have
7147 two sequences of OIDs for trusted and rejected settings. These will
7148 typically have values the same as the extended key usage extension
7149 and any application specific purposes.
7151 The trust checking code now has a default behaviour: it will just
7152 check for an object with the same NID as the passed id. Functions can
7153 be provided to override either the default behaviour or the behaviour
7154 for a given id. SSL client, server and email already have functions
7155 in place for compatibility: they check the NID and also return "trusted"
7156 if the certificate is self signed.
7159 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7160 traditional format into an EVP_PKEY structure.
7163 *) Add a password callback function PEM_cb() which either prompts for
7164 a password if usr_data is NULL or otherwise assumes it is a null
7165 terminated password. Allow passwords to be passed on command line
7166 environment or config files in a few more utilities.
7169 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7170 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7171 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7172 Update documentation.
7175 *) Support for ASN1 "NULL" type. This could be handled before by using
7176 ASN1_TYPE but there wasn't any function that would try to read a NULL
7177 and produce an error if it couldn't. For compatibility we also have
7178 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7179 don't allocate anything because they don't need to.
7182 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7184 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7186 *) Rebuild of the memory allocation routines used by OpenSSL code and
7187 possibly others as well. The purpose is to make an interface that
7188 provide hooks so anyone can build a separate set of allocation and
7189 deallocation routines to be used by OpenSSL, for example memory
7190 pool implementations, or something else, which was previously hard
7191 since Malloc(), Realloc() and Free() were defined as macros having
7192 the values malloc, realloc and free, respectively (except for Win32
7193 compilations). The same is provided for memory debugging code.
7194 OpenSSL already comes with functionality to find memory leaks, but
7195 this gives people a chance to debug other memory problems.
7197 With these changes, a new set of functions and macros have appeared:
7199 CRYPTO_set_mem_debug_functions() [F]
7200 CRYPTO_get_mem_debug_functions() [F]
7201 CRYPTO_dbg_set_options() [F]
7202 CRYPTO_dbg_get_options() [F]
7203 CRYPTO_malloc_debug_init() [M]
7205 The memory debug functions are NULL by default, unless the library
7206 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7207 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7208 gives the standard debugging functions that come with OpenSSL) or
7209 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7210 provided by the library user) must be used. When the standard
7211 debugging functions are used, CRYPTO_dbg_set_options can be used to
7212 request additional information:
7213 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7214 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7216 Also, things like CRYPTO_set_mem_functions will always give the
7217 expected result (the new set of functions is used for allocation
7218 and deallocation) at all times, regardless of platform and compiler
7221 To finish it up, some functions that were never use in any other
7222 way than through macros have a new API and new semantic:
7225 CRYPTO_dbg_realloc()
7228 All macros of value have retained their old syntax.
7229 [Richard Levitte and Bodo Moeller]
7231 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7232 ordering of SMIMECapabilities wasn't in "strength order" and there
7233 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7237 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7238 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7239 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7241 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7242 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7243 functionality to handle multipart/signed properly) and a utility
7244 called 'smime' to call all this stuff. This is based on code I
7245 originally wrote for Celo who have kindly allowed it to be
7246 included in OpenSSL.
7249 *) Add variants des_set_key_checked and des_set_key_unchecked of
7250 des_set_key (aka des_key_sched). Global variable des_check_key
7251 decides which of these is called by des_set_key; this way
7252 des_check_key behaves as it always did, but applications and
7253 the library itself, which was buggy for des_check_key == 1,
7254 have a cleaner way to pick the version they need.
7257 *) New function PKCS12_newpass() which changes the password of a
7261 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7262 dynamic mix. In both cases the ids can be used as an index into the
7263 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7264 functions so they accept a list of the field values and the
7265 application doesn't need to directly manipulate the X509_TRUST
7269 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7273 *) Modify the way the V3 extension code looks up extensions. This now
7274 works in a similar way to the object code: we have some "standard"
7275 extensions in a static table which is searched with OBJ_bsearch()
7276 and the application can add dynamic ones if needed. The file
7277 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7278 updated whenever a new extension is added to the core code and kept
7279 in ext_nid order. There is a simple program 'tabtest.c' which checks
7280 this. New extensions are not added too often so this file can readily
7281 be maintained manually.
7283 There are two big advantages in doing things this way. The extensions
7284 can be looked up immediately and no longer need to be "added" using
7285 X509V3_add_standard_extensions(): this function now does nothing.
7286 [Side note: I get *lots* of email saying the extension code doesn't
7287 work because people forget to call this function]
7288 Also no dynamic allocation is done unless new extensions are added:
7289 so if we don't add custom extensions there is no need to call
7290 X509V3_EXT_cleanup().
7293 *) Modify enc utility's salting as follows: make salting the default. Add a
7294 magic header, so unsalted files fail gracefully instead of just decrypting
7295 to garbage. This is because not salting is a big security hole, so people
7296 should be discouraged from doing it.
7299 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7300 digest to be passed on the command line but it only used this
7301 parameter when signing a certificate. Modified so all relevant
7302 operations are affected by the digest parameter including the
7303 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7304 DSA key was used because it didn't fix the digest.
7307 *) Initial certificate chain verify code. Currently tests the untrusted
7308 certificates for consistency with the verify purpose (which is set
7309 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7311 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7312 this is because it will reject chains with invalid extensions whereas
7313 every previous version of OpenSSL and SSLeay made no checks at all.
7315 Trust code: checks the root CA for the relevant trust settings. Trust
7316 settings have an initial value consistent with the verify purpose: e.g.
7317 if the verify purpose is for SSL client use it expects the CA to be
7318 trusted for SSL client use. However the default value can be changed to
7319 permit custom trust settings: one example of this would be to only trust
7320 certificates from a specific "secure" set of CAs.
7322 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7323 which should be used for version portability: especially since the
7324 verify structure is likely to change more often now.
7326 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7327 to set them. If not set then assume SSL clients will verify SSL servers
7330 Two new options to the verify program: -untrusted allows a set of
7331 untrusted certificates to be passed in and -purpose which sets the
7332 intended purpose of the certificate. If a purpose is set then the
7333 new chain verify code is used to check extension consistency.
7336 *) Support for the authority information access extension.
7339 *) Modify RSA and DSA PEM read routines to transparently handle
7340 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7341 public keys in a format compatible with certificate
7342 SubjectPublicKeyInfo structures. Unfortunately there were already
7343 functions called *_PublicKey_* which used various odd formats so
7344 these are retained for compatibility: however the DSA variants were
7345 never in a public release so they have been deleted. Changed dsa/rsa
7346 utilities to handle the new format: note no releases ever handled public
7347 keys so we should be OK.
7349 The primary motivation for this change is to avoid the same fiasco
7350 that dogs private keys: there are several incompatible private key
7351 formats some of which are standard and some OpenSSL specific and
7352 require various evil hacks to allow partial transparent handling and
7353 even then it doesn't work with DER formats. Given the option anything
7354 other than PKCS#8 should be dumped: but the other formats have to
7355 stay in the name of compatibility.
7357 With public keys and the benefit of hindsight one standard format
7358 is used which works with EVP_PKEY, RSA or DSA structures: though
7359 it clearly returns an error if you try to read the wrong kind of key.
7361 Added a -pubkey option to the 'x509' utility to output the public key.
7362 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7363 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7364 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7365 that do the same as the EVP_PKEY_assign_*() except they up the
7366 reference count of the added key (they don't "swallow" the
7370 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7371 CRLs would fail if the file contained no certificates or no CRLs:
7372 added a new function to read in both types and return the number
7373 read: this means that if none are read it will be an error. The
7374 DER versions of the certificate and CRL reader would always fail
7375 because it isn't possible to mix certificates and CRLs in DER format
7376 without choking one or the other routine. Changed this to just read
7377 a certificate: this is the best we can do. Also modified the code
7378 in apps/verify.c to take notice of return codes: it was previously
7379 attempting to read in certificates from NULL pointers and ignoring
7380 any errors: this is one reason why the cert and CRL reader seemed
7381 to work. It doesn't check return codes from the default certificate
7382 routines: these may well fail if the certificates aren't installed.
7385 *) Code to support otherName option in GeneralName.
7388 *) First update to verify code. Change the verify utility
7389 so it warns if it is passed a self signed certificate:
7390 for consistency with the normal behaviour. X509_verify
7391 has been modified to it will now verify a self signed
7392 certificate if *exactly* the same certificate appears
7393 in the store: it was previously impossible to trust a
7394 single self signed certificate. This means that:
7395 openssl verify ss.pem
7396 now gives a warning about a self signed certificate but
7397 openssl verify -CAfile ss.pem ss.pem
7401 *) For servers, store verify_result in SSL_SESSION data structure
7402 (and add it to external session representation).
7403 This is needed when client certificate verifications fails,
7404 but an application-provided verification callback (set by
7405 SSL_CTX_set_cert_verify_callback) allows accepting the session
7406 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7407 but returns 1): When the session is reused, we have to set
7408 ssl->verify_result to the appropriate error code to avoid
7410 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7412 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7413 case in PKCS7_dataInit() where the signed PKCS7 structure
7414 didn't contain any existing data because it was being created.
7415 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7417 *) Add a salt to the key derivation routines in enc.c. This
7418 forms the first 8 bytes of the encrypted file. Also add a
7419 -S option to allow a salt to be input on the command line.
7422 *) New function X509_cmp(). Oddly enough there wasn't a function
7423 to compare two certificates. We do this by working out the SHA1
7424 hash and comparing that. X509_cmp() will be needed by the trust
7428 *) SSL_get1_session() is like SSL_get_session(), but increments
7429 the reference count in the SSL_SESSION returned.
7430 [Geoff Thorpe <geoff@eu.c2.net>]
7432 *) Fix for 'req': it was adding a null to request attributes.
7433 Also change the X509_LOOKUP and X509_INFO code to handle
7434 certificate auxiliary information.
7437 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7441 *) Add the possibility to add extra information to the memory leak
7442 detecting output, to form tracebacks, showing from where each
7443 allocation was originated: CRYPTO_push_info("constant string") adds
7444 the string plus current file name and line number to a per-thread
7445 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7446 is like calling CYRPTO_pop_info() until the stack is empty.
7447 Also updated memory leak detection code to be multi-thread-safe.
7450 *) Add options -text and -noout to pkcs7 utility and delete the
7451 encryption options which never did anything. Update docs.
7454 *) Add options to some of the utilities to allow the pass phrase
7455 to be included on either the command line (not recommended on
7456 OSes like Unix) or read from the environment. Update the
7457 manpages and fix a few bugs.
7460 *) Add a few manpages for some of the openssl commands.
7463 *) Fix the -revoke option in ca. It was freeing up memory twice,
7464 leaking and not finding already revoked certificates.
7467 *) Extensive changes to support certificate auxiliary information.
7468 This involves the use of X509_CERT_AUX structure and X509_AUX
7469 functions. An X509_AUX function such as PEM_read_X509_AUX()
7470 can still read in a certificate file in the usual way but it
7471 will also read in any additional "auxiliary information". By
7472 doing things this way a fair degree of compatibility can be
7473 retained: existing certificates can have this information added
7474 using the new 'x509' options.
7476 Current auxiliary information includes an "alias" and some trust
7477 settings. The trust settings will ultimately be used in enhanced
7478 certificate chain verification routines: currently a certificate
7479 can only be trusted if it is self signed and then it is trusted
7483 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7484 The problem was that one of the replacement routines had not been working
7485 since SSLeay releases. For now the offending routine has been replaced
7486 with non-optimised assembler. Even so, this now gives around 95%
7487 performance improvement for 1024 bit RSA signs.
7490 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7491 handling. Most clients have the effective key size in bits equal to
7492 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7493 A few however don't do this and instead use the size of the decrypted key
7494 to determine the RC2 key length and the AlgorithmIdentifier to determine
7495 the effective key length. In this case the effective key length can still
7496 be 40 bits but the key length can be 168 bits for example. This is fixed
7497 by manually forcing an RC2 key into the EVP_PKEY structure because the
7498 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7499 the key length and effective key length are equal.
7502 *) Add a bunch of functions that should simplify the creation of
7503 X509_NAME structures. Now you should be able to do:
7504 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7505 and have it automatically work out the correct field type and fill in
7506 the structures. The more adventurous can try:
7507 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7508 and it will (hopefully) work out the correct multibyte encoding.
7511 *) Change the 'req' utility to use the new field handling and multibyte
7512 copy routines. Before the DN field creation was handled in an ad hoc
7513 way in req, ca, and x509 which was rather broken and didn't support
7514 BMPStrings or UTF8Strings. Since some software doesn't implement
7515 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7516 using the dirstring_type option. See the new comment in the default
7517 openssl.cnf for more info.
7520 *) Make crypto/rand/md_rand.c more robust:
7521 - Assure unique random numbers after fork().
7522 - Make sure that concurrent threads access the global counter and
7523 md serializably so that we never lose entropy in them
7524 or use exactly the same state in multiple threads.
7525 Access to the large state is not always serializable because
7526 the additional locking could be a performance killer, and
7527 md should be large enough anyway.
7530 *) New file apps/app_rand.c with commonly needed functionality
7531 for handling the random seed file.
7533 Use the random seed file in some applications that previously did not:
7535 dsaparam -genkey (which also ignored its '-rand' option),
7538 x509 (when signing).
7539 Except on systems with /dev/urandom, it is crucial to have a random
7540 seed file at least for key creation, DSA signing, and for DH exchanges;
7541 for RSA signatures we could do without one.
7543 gendh and gendsa (unlike genrsa) used to read only the first byte
7544 of each file listed in the '-rand' option. The function as previously
7545 found in genrsa is now in app_rand.c and is used by all programs
7546 that support '-rand'.
7549 *) In RAND_write_file, use mode 0600 for creating files;
7550 don't just chmod when it may be too late.
7553 *) Report an error from X509_STORE_load_locations
7554 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7557 *) New function ASN1_mbstring_copy() this copies a string in either
7558 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7559 into an ASN1_STRING type. A mask of permissible types is passed
7560 and it chooses the "minimal" type to use or an error if not type
7564 *) Add function equivalents to the various macros in asn1.h. The old
7565 macros are retained with an M_ prefix. Code inside the library can
7566 use the M_ macros. External code (including the openssl utility)
7567 should *NOT* in order to be "shared library friendly".
7570 *) Add various functions that can check a certificate's extensions
7571 to see if it usable for various purposes such as SSL client,
7572 server or S/MIME and CAs of these types. This is currently
7573 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7574 verification. Also added a -purpose flag to x509 utility to
7575 print out all the purposes.
7578 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7582 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7583 for, obtain and decode and extension and obtain its critical flag.
7584 This allows all the necessary extension code to be handled in a
7585 single function call.
7588 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7589 platforms. See crypto/rc4/rc4_enc.c for further details.
7592 *) New -noout option to asn1parse. This causes no output to be produced
7593 its main use is when combined with -strparse and -out to extract data
7594 from a file (which may not be in ASN.1 format).
7597 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7598 when producing the local key id.
7599 [Richard Levitte <levitte@stacken.kth.se>]
7601 *) New option -dhparam in s_server. This allows a DH parameter file to be
7602 stated explicitly. If it is not stated then it tries the first server
7603 certificate file. The previous behaviour hard coded the filename
7607 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7608 a public key to be input or output. For example:
7609 openssl rsa -in key.pem -pubout -out pubkey.pem
7610 Also added necessary DSA public key functions to handle this.
7613 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7614 in the message. This was handled by allowing
7615 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7616 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7618 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7619 to the end of the strings whereas this didn't. This would cause problems
7620 if strings read with d2i_ASN1_bytes() were later modified.
7621 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7623 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7624 data and it contains EOF it will end up returning an error. This is
7625 caused by input 46 bytes long. The cause is due to the way base64
7626 BIOs find the start of base64 encoded data. They do this by trying a
7627 trial decode on each line until they find one that works. When they
7628 do a flag is set and it starts again knowing it can pass all the
7629 data directly through the decoder. Unfortunately it doesn't reset
7630 the context it uses. This means that if EOF is reached an attempt
7631 is made to pass two EOFs through the context and this causes the
7632 resulting error. This can also cause other problems as well. As is
7633 usual with these problems it takes *ages* to find and the fix is
7634 trivial: move one line.
7635 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7637 *) Ugly workaround to get s_client and s_server working under Windows. The
7638 old code wouldn't work because it needed to select() on sockets and the
7639 tty (for keypresses and to see if data could be written). Win32 only
7640 supports select() on sockets so we select() with a 1s timeout on the
7641 sockets and then see if any characters are waiting to be read, if none
7642 are present then we retry, we also assume we can always write data to
7643 the tty. This isn't nice because the code then blocks until we've
7644 received a complete line of data and it is effectively polling the
7645 keyboard at 1s intervals: however it's quite a bit better than not
7646 working at all :-) A dedicated Windows application might handle this
7647 with an event loop for example.
7650 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7651 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7652 will be called when RSA_sign() and RSA_verify() are used. This is useful
7653 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7654 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7655 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7656 This necessitated the support of an extra signature type NID_md5_sha1
7657 for SSL signatures and modifications to the SSL library to use it instead
7658 of calling RSA_public_decrypt() and RSA_private_encrypt().
7661 *) Add new -verify -CAfile and -CApath options to the crl program, these
7662 will lookup a CRL issuers certificate and verify the signature in a
7663 similar way to the verify program. Tidy up the crl program so it
7664 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7665 less strict. It will now permit CRL extensions even if it is not
7666 a V2 CRL: this will allow it to tolerate some broken CRLs.
7669 *) Initialize all non-automatic variables each time one of the openssl
7670 sub-programs is started (this is necessary as they may be started
7671 multiple times from the "OpenSSL>" prompt).
7672 [Lennart Bang, Bodo Moeller]
7674 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7675 removing all other RSA functionality (this is what NO_RSA does). This
7676 is so (for example) those in the US can disable those operations covered
7677 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7681 *) Non-copying interface to BIO pairs.
7682 (still largely untested)
7685 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7686 ASCII string. This was handled independently in various places before.
7689 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7690 UTF8 strings a character at a time.
7693 *) Use client_version from client hello to select the protocol
7694 (s23_srvr.c) and for RSA client key exchange verification
7695 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7698 *) Add various utility functions to handle SPKACs, these were previously
7699 handled by poking round in the structure internals. Added new function
7700 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7701 print, verify and generate SPKACs. Based on an original idea from
7702 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7705 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7708 *) Allow the config file extension section to be overwritten on the
7709 command line. Based on an original idea from Massimiliano Pala
7710 <madwolf@comune.modena.it>. The new option is called -extensions
7711 and can be applied to ca, req and x509. Also -reqexts to override
7712 the request extensions in req and -crlexts to override the crl extensions
7716 *) Add new feature to the SPKAC handling in ca. Now you can include
7717 the same field multiple times by preceding it by "XXXX." for example:
7720 this is the same syntax as used in the req config file.
7723 *) Allow certificate extensions to be added to certificate requests. These
7724 are specified in a 'req_extensions' option of the req section of the
7725 config file. They can be printed out with the -text option to req but
7726 are otherwise ignored at present.
7729 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7730 data read consists of only the final block it would not decrypted because
7731 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7732 A misplaced 'break' also meant the decrypted final block might not be
7733 copied until the next read.
7736 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7737 a few extra parameters to the DH structure: these will be useful if
7738 for example we want the value of 'q' or implement X9.42 DH.
7741 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7742 provides hooks that allow the default DSA functions or functions on a
7743 "per key" basis to be replaced. This allows hardware acceleration and
7744 hardware key storage to be handled without major modification to the
7745 library. Also added low level modexp hooks and CRYPTO_EX structure and
7746 associated functions.
7749 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7750 as "read only": it can't be written to and the buffer it points to will
7751 not be freed. Reading from a read only BIO is much more efficient than
7752 a normal memory BIO. This was added because there are several times when
7753 an area of memory needs to be read from a BIO. The previous method was
7754 to create a memory BIO and write the data to it, this results in two
7755 copies of the data and an O(n^2) reading algorithm. There is a new
7756 function BIO_new_mem_buf() which creates a read only memory BIO from
7757 an area of memory. Also modified the PKCS#7 routines to use read only
7761 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7762 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7763 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7764 but a retry condition occured while trying to read the rest.
7767 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7768 NID_pkcs7_encrypted by default: this was wrong since this should almost
7769 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7770 the encrypted data type: this is a more sensible place to put it and it
7771 allows the PKCS#12 code to be tidied up that duplicated this
7775 *) Changed obj_dat.pl script so it takes its input and output files on
7776 the command line. This should avoid shell escape redirection problems
7780 *) Initial support for certificate extension requests, these are included
7781 in things like Xenroll certificate requests. Included functions to allow
7782 extensions to be obtained and added.
7785 *) -crlf option to s_client and s_server for sending newlines as
7786 CRLF (as required by many protocols).
7789 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7791 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7792 [Ralf S. Engelschall]
7794 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7795 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7797 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7801 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7802 DH parameters/keys (q is lost during that conversion, but the resulting
7803 DH parameters contain its length).
7805 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7806 much faster than DH_generate_parameters (which creates parameters
7807 where p = 2*q + 1), and also the smaller q makes DH computations
7808 much more efficient (160-bit exponentiation instead of 1024-bit
7809 exponentiation); so this provides a convenient way to support DHE
7810 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7811 utter importance to use
7812 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7814 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7815 when such DH parameters are used, because otherwise small subgroup
7816 attacks may become possible!
7819 *) Avoid memory leak in i2d_DHparams.
7822 *) Allow the -k option to be used more than once in the enc program:
7823 this allows the same encrypted message to be read by multiple recipients.
7826 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7827 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7828 it will always use the numerical form of the OID, even if it has a short
7832 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7833 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7834 otherwise bn_mod_exp was called. In the case of hardware keys for example
7835 no private key components need be present and it might store extra data
7836 in the RSA structure, which cannot be accessed from bn_mod_exp.
7837 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7838 private key operations.
7841 *) Added support for SPARC Linux.
7844 *) pem_password_cb function type incompatibly changed from
7845 typedef int pem_password_cb(char *buf, int size, int rwflag);
7847 ....(char *buf, int size, int rwflag, void *userdata);
7848 so that applications can pass data to their callbacks:
7849 The PEM[_ASN1]_{read,write}... functions and macros now take an
7850 additional void * argument, which is just handed through whenever
7851 the password callback is called.
7852 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7854 New function SSL_CTX_set_default_passwd_cb_userdata.
7856 Compatibility note: As many C implementations push function arguments
7857 onto the stack in reverse order, the new library version is likely to
7858 interoperate with programs that have been compiled with the old
7859 pem_password_cb definition (PEM_whatever takes some data that
7860 happens to be on the stack as its last argument, and the callback
7861 just ignores this garbage); but there is no guarantee whatsoever that
7864 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7865 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7866 problems not only on Windows, but also on some Unix platforms.
7867 To avoid problematic command lines, these definitions are now in an
7868 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7869 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7872 *) MIPS III/IV assembler module is reimplemented.
7875 *) More DES library cleanups: remove references to srand/rand and
7876 delete an unused file.
7879 *) Add support for the the free Netwide assembler (NASM) under Win32,
7880 since not many people have MASM (ml) and it can be hard to obtain.
7881 This is currently experimental but it seems to work OK and pass all
7882 the tests. Check out INSTALL.W32 for info.
7885 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7886 without temporary keys kept an extra copy of the server key,
7887 and connections with temporary keys did not free everything in case
7891 *) New function RSA_check_key and new openssl rsa option -check
7892 for verifying the consistency of RSA keys.
7893 [Ulf Moeller, Bodo Moeller]
7895 *) Various changes to make Win32 compile work:
7896 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7897 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7898 comparison" warnings.
7899 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7902 *) Add a debugging option to PKCS#5 v2 key generation function: when
7903 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7904 derived keys are printed to stderr.
7907 *) Copy the flags in ASN1_STRING_dup().
7908 [Roman E. Pavlov <pre@mo.msk.ru>]
7910 *) The x509 application mishandled signing requests containing DSA
7911 keys when the signing key was also DSA and the parameters didn't match.
7913 It was supposed to omit the parameters when they matched the signing key:
7914 the verifying software was then supposed to automatically use the CA's
7915 parameters if they were absent from the end user certificate.
7917 Omitting parameters is no longer recommended. The test was also
7918 the wrong way round! This was probably due to unusual behaviour in
7919 EVP_cmp_parameters() which returns 1 if the parameters match.
7920 This meant that parameters were omitted when they *didn't* match and
7921 the certificate was useless. Certificates signed with 'ca' didn't have
7923 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7925 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7926 The interface is as follows:
7927 Applications can use
7928 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7929 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7930 "off" is now the default.
7931 The library internally uses
7932 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7933 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7934 to disable memory-checking temporarily.
7936 Some inconsistent states that previously were possible (and were
7937 even the default) are now avoided.
7939 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7940 with each memory chunk allocated; this is occasionally more helpful
7941 than just having a counter.
7943 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7945 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7949 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7950 which largely parallels "options", but is for changing API behaviour,
7951 whereas "options" are about protocol behaviour.
7952 Initial "mode" flags are:
7954 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7955 a single record has been written.
7956 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7957 retries use the same buffer location.
7958 (But all of the contents must be
7962 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7965 *) Fix problems with no-hmac etc.
7966 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7968 *) New functions RSA_get_default_method(), RSA_set_method() and
7969 RSA_get_method(). These allows replacement of RSA_METHODs without having
7970 to mess around with the internals of an RSA structure.
7973 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7974 Also really enable memory leak checks in openssl.c and in some
7976 [Chad C. Mulligan, Bodo Moeller]
7978 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7979 up the length of negative integers. This has now been simplified to just
7980 store the length when it is first determined and use it later, rather
7981 than trying to keep track of where data is copied and updating it to
7983 [Steve Henson, reported by Brien Wheeler
7984 <bwheeler@authentica-security.com>]
7986 *) Add a new function PKCS7_signatureVerify. This allows the verification
7987 of a PKCS#7 signature but with the signing certificate passed to the
7988 function itself. This contrasts with PKCS7_dataVerify which assumes the
7989 certificate is present in the PKCS#7 structure. This isn't always the
7990 case: certificates can be omitted from a PKCS#7 structure and be
7991 distributed by "out of band" means (such as a certificate database).
7994 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7995 function prototypes in pem.h, also change util/mkdef.pl to add the
7996 necessary function names.
7999 *) mk1mf.pl (used by Windows builds) did not properly read the
8000 options set by Configure in the top level Makefile, and Configure
8001 was not even able to write more than one option correctly.
8002 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8005 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8006 file to be loaded from a BIO or FILE pointer. The BIO version will
8007 for example allow memory BIOs to contain config info.
8010 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8011 Whoever hopes to achieve shared-library compatibility across versions
8012 must use this, not the compile-time macro.
8013 (Exercise 0.9.4: Which is the minimum library version required by
8015 Note: All this applies only to multi-threaded programs, others don't
8019 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8020 through a BIO pair triggered the default case, i.e.
8021 SSLerr(...,SSL_R_UNKNOWN_STATE).
8024 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8025 can use the SSL library even if none of the specific BIOs is
8029 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8030 for the encoded length.
8031 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8033 *) Add initial documentation of the X509V3 functions.
8036 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8037 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8038 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8039 secure PKCS#8 private key format with a high iteration count.
8042 *) Fix determination of Perl interpreter: A perl or perl5
8043 _directory_ in $PATH was also accepted as the interpreter.
8044 [Ralf S. Engelschall]
8046 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8047 wrong with it but it was very old and did things like calling
8048 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8052 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8053 to use the new extension code.
8056 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8057 with macros. This should make it easier to change their form, add extra
8058 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8062 *) Add to configuration table a new entry that can specify an alternative
8063 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8064 according to Mark Crispin <MRC@Panda.COM>.
8068 *) DES CBC did not update the IV. Weird.
8071 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8072 Changing the behaviour of the former might break existing programs --
8073 where IV updating is needed, des_ncbc_encrypt can be used.
8076 *) When bntest is run from "make test" it drives bc to check its
8077 calculations, as well as internally checking them. If an internal check
8078 fails, it needs to cause bc to give a non-zero result or make test carries
8079 on without noticing the failure. Fixed.
8082 *) DES library cleanups.
8085 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8086 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8087 ciphers. NOTE: although the key derivation function has been verified
8088 against some published test vectors it has not been extensively tested
8089 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8093 *) Instead of "mkdir -p", which is not fully portable, use new
8094 Perl script "util/mkdir-p.pl".
8097 *) Rewrite the way password based encryption (PBE) is handled. It used to
8098 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8099 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8100 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8101 the 'parameter' field of the AlgorithmIdentifier is passed to the
8102 underlying key generation function so it must do its own ASN1 parsing.
8103 This has also changed the EVP_PBE_CipherInit() function which now has a
8104 'parameter' argument instead of literal salt and iteration count values
8105 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8108 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8109 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8110 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8111 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8112 value was just used as a "magic string" and not used directly its
8113 value doesn't matter.
8116 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8120 *) "linux-sparc64" configuration (ultrapenguin).
8121 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8122 "linux-sparc" configuration.
8123 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8125 *) config now generates no-xxx options for missing ciphers.
8128 *) Support the EBCDIC character set (work in progress).
8129 File ebcdic.c not yet included because it has a different license.
8130 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8132 *) Support BS2000/OSD-POSIX.
8133 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8135 *) Make callbacks for key generation use void * instead of char *.
8138 *) Make S/MIME samples compile (not yet tested).
8141 *) Additional typesafe stacks.
8144 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8148 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8150 *) New configuration variant "sco5-gcc".
8152 *) Updated some demos.
8153 [Sean O Riordain, Wade Scholine]
8155 *) Add missing BIO_free at exit of pkcs12 application.
8158 *) Fix memory leak in conf.c.
8161 *) Updates for Win32 to assembler version of MD5.
8164 *) Set #! path to perl in apps/der_chop to where we found it
8165 instead of using a fixed path.
8168 *) SHA library changes for irix64-mips4-cc.
8171 *) Improvements for VMS support.
8175 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8177 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8178 This also avoids the problems with SC4.2 and unpatched SC5.
8179 [Andy Polyakov <appro@fy.chalmers.se>]
8181 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8182 These are required because of the typesafe stack would otherwise break
8183 existing code. If old code used a structure member which used to be STACK
8184 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8185 sk_num or sk_value it would produce an error because the num, data members
8186 are not present in STACK_OF. Now it just produces a warning. sk_set
8187 replaces the old method of assigning a value to sk_value
8188 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8189 that does this will no longer work (and should use sk_set instead) but
8190 this could be regarded as a "questionable" behaviour anyway.
8193 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8194 correctly handle encrypted S/MIME data.
8197 *) Change type of various DES function arguments from des_cblock
8198 (which means, in function argument declarations, pointer to char)
8199 to des_cblock * (meaning pointer to array with 8 char elements),
8200 which allows the compiler to do more typechecking; it was like
8201 that back in SSLeay, but with lots of ugly casts.
8203 Introduce new type const_des_cblock.
8206 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8207 problems: find RecipientInfo structure that matches recipient certificate
8208 and initialise the ASN1 structures properly based on passed cipher.
8211 *) Belatedly make the BN tests actually check the results.
8214 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8215 to and from BNs: it was completely broken. New compilation option
8216 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8217 key elements as negative integers.
8220 *) Reorganize and speed up MD5.
8221 [Andy Polyakov <appro@fy.chalmers.se>]
8224 [Richard Levitte <richard@levitte.org>]
8226 *) New option -out to asn1parse to allow the parsed structure to be
8227 output to a file. This is most useful when combined with the -strparse
8228 option to examine the output of things like OCTET STRINGS.
8231 *) Make SSL library a little more fool-proof by not requiring any longer
8232 that SSL_set_{accept,connect}_state be called before
8233 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8234 in many applications because usually everything *appeared* to work as
8235 intended anyway -- now it really works as intended).
8238 *) Move openssl.cnf out of lib/.
8241 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8242 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8243 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8244 [Ralf S. Engelschall]
8246 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8247 handle PKCS#7 enveloped data properly.
8248 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8250 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8251 copying pointers. The cert_st handling is changed by this in
8252 various ways (and thus what used to be known as ctx->default_cert
8253 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8254 any longer when s->cert does not give us what we need).
8255 ssl_cert_instantiate becomes obsolete by this change.
8256 As soon as we've got the new code right (possibly it already is?),
8257 we have solved a couple of bugs of the earlier code where s->cert
8258 was used as if it could not have been shared with other SSL structures.
8260 Note that using the SSL API in certain dirty ways now will result
8261 in different behaviour than observed with earlier library versions:
8262 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8263 does not influence s as it used to.
8265 In order to clean up things more thoroughly, inside SSL_SESSION
8266 we don't use CERT any longer, but a new structure SESS_CERT
8267 that holds per-session data (if available); currently, this is
8268 the peer's certificate chain and, for clients, the server's certificate
8269 and temporary key. CERT holds only those values that can have
8270 meaningful defaults in an SSL_CTX.
8273 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8274 from the internal representation. Various PKCS#7 fixes: remove some
8275 evil casts and set the enc_dig_alg field properly based on the signing
8279 *) Allow PKCS#12 password to be set from the command line or the
8280 environment. Let 'ca' get its config file name from the environment
8281 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8285 *) Allow certificate policies extension to use an IA5STRING for the
8286 organization field. This is contrary to the PKIX definition but
8287 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8291 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8292 without disallowing inline assembler and the like for non-pedantic builds.
8295 *) Support Borland C++ builder.
8296 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8301 *) SHA-1 cleanups and performance enhancements.
8302 [Andy Polyakov <appro@fy.chalmers.se>]
8304 *) Sparc v8plus assembler for the bignum library.
8305 [Andy Polyakov <appro@fy.chalmers.se>]
8307 *) Accept any -xxx and +xxx compiler options in Configure.
8310 *) Update HPUX configuration.
8313 *) Add missing sk_<type>_unshift() function to safestack.h
8314 [Ralf S. Engelschall]
8316 *) New function SSL_CTX_use_certificate_chain_file that sets the
8317 "extra_cert"s in addition to the certificate. (This makes sense
8318 only for "PEM" format files, as chains as a whole are not
8322 *) Support verify_depth from the SSL API.
8323 x509_vfy.c had what can be considered an off-by-one-error:
8324 Its depth (which was not part of the external interface)
8325 was actually counting the number of certificates in a chain;
8326 now it really counts the depth.
8329 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8330 instead of X509err, which often resulted in confusing error
8331 messages since the error codes are not globally unique
8332 (e.g. an alleged error in ssl3_accept when a certificate
8333 didn't match the private key).
8335 *) New function SSL_CTX_set_session_id_context that allows to set a default
8336 value (so that you don't need SSL_set_session_id_context for each
8337 connection using the SSL_CTX).
8340 *) OAEP decoding bug fix.
8343 *) Support INSTALL_PREFIX for package builders, as proposed by
8347 *) New Configure options "threads" and "no-threads". For systems
8348 where the proper compiler options are known (currently Solaris
8349 and Linux), "threads" is the default.
8352 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8355 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8356 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8357 such as /usr/local/bin.
8360 *) "make linux-shared" to build shared libraries.
8361 [Niels Poppe <niels@netbox.org>]
8363 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8366 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8367 extension adding in x509 utility.
8370 *) Remove NOPROTO sections and error code comments.
8373 *) Partial rewrite of the DEF file generator to now parse the ANSI
8377 *) New Configure options --prefix=DIR and --openssldir=DIR.
8380 *) Complete rewrite of the error code script(s). It is all now handled
8381 by one script at the top level which handles error code gathering,
8382 header rewriting and C source file generation. It should be much better
8383 than the old method: it now uses a modified version of Ulf's parser to
8384 read the ANSI prototypes in all header files (thus the old K&R definitions
8385 aren't needed for error creation any more) and do a better job of
8386 translating function codes into names. The old 'ASN1 error code imbedded
8387 in a comment' is no longer necessary and it doesn't use .err files which
8388 have now been deleted. Also the error code call doesn't have to appear all
8389 on one line (which resulted in some large lines...).
8392 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8395 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8396 0 (which usually indicates a closed connection), but continue reading.
8399 *) Fix some race conditions.
8402 *) Add support for CRL distribution points extension. Add Certificate
8403 Policies and CRL distribution points documentation.
8406 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8409 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8410 8 of keying material. Merlin has also confirmed interop with this fix
8411 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8412 [Merlin Hughes <merlin@baltimore.ie>]
8414 *) Fix lots of warnings.
8415 [Richard Levitte <levitte@stacken.kth.se>]
8417 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8418 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8419 [Richard Levitte <levitte@stacken.kth.se>]
8421 *) Fix problems with sizeof(long) == 8.
8422 [Andy Polyakov <appro@fy.chalmers.se>]
8424 *) Change functions to ANSI C.
8427 *) Fix typos in error codes.
8428 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8430 *) Remove defunct assembler files from Configure.
8433 *) SPARC v8 assembler BIGNUM implementation.
8434 [Andy Polyakov <appro@fy.chalmers.se>]
8436 *) Support for Certificate Policies extension: both print and set.
8437 Various additions to support the r2i method this uses.
8440 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8441 return a const string when you are expecting an allocated buffer.
8444 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8445 types DirectoryString and DisplayText.
8448 *) Add code to allow r2i extensions to access the configuration database,
8449 add an LHASH database driver and add several ctx helper functions.
8452 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8453 fail when they extended the size of a BIGNUM.
8456 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8457 support typesafe stack.
8460 *) Fix typo in SSL_[gs]et_options().
8461 [Nils Frostberg <nils@medcom.se>]
8463 *) Delete various functions and files that belonged to the (now obsolete)
8464 old X509V3 handling code.
8467 *) New Configure option "rsaref".
8470 *) Don't auto-generate pem.h.
8473 *) Introduce type-safe ASN.1 SETs.
8476 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8477 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8479 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8480 that links with OpenSSL (well at least cause lots of warnings), but fear
8481 not: the conversion is trivial, and it eliminates loads of evil casts. A
8482 few STACKed things have been converted already. Feel free to convert more.
8483 In the fullness of time, I'll do away with the STACK type altogether.
8486 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8487 specified in <certfile> by updating the entry in the index.txt file.
8488 This way one no longer has to edit the index.txt file manually for
8489 revoking a certificate. The -revoke option does the gory details now.
8490 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8492 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8493 `-text' option at all and this way the `-noout -text' combination was
8494 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8495 [Ralf S. Engelschall]
8497 *) Make sure a corresponding plain text error message exists for the
8498 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8499 verify callback function determined that a certificate was revoked.
8500 [Ralf S. Engelschall]
8502 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8503 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8504 all available cipers including rc5, which was forgotten until now.
8505 In order to let the testing shell script know which algorithms
8506 are available, a new (up to now undocumented) command
8507 "openssl list-cipher-commands" is used.
8510 *) Bugfix: s_client occasionally would sleep in select() when
8511 it should have checked SSL_pending() first.
8514 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8515 the raw DSA values prior to ASN.1 encoding.
8518 *) Tweaks to Configure
8519 [Niels Poppe <niels@netbox.org>]
8521 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8525 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8528 *) New config option to avoid instructions that are illegal on the 80386.
8529 The default code is faster, but requires at least a 486.
8532 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8533 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8534 same as SSL2_VERSION anyway.
8537 *) New "-showcerts" option for s_client.
8540 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8541 application. Various cleanups and fixes.
8544 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8545 modify error routines to work internally. Add error codes and PBE init
8546 to library startup routines.
8549 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8550 packing functions to asn1 and evp. Changed function names and error
8551 codes along the way.
8554 *) PKCS12 integration: and so it begins... First of several patches to
8555 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8556 objects to objects.h
8559 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8560 and display support for Thawte strong extranet extension.
8563 *) Add LinuxPPC support.
8564 [Jeff Dubrule <igor@pobox.org>]
8566 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8567 bn_div_words in alpha.s.
8568 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8570 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8571 OAEP isn't supported when OpenSSL is built with RSAref.
8572 [Ulf Moeller <ulf@fitug.de>]
8574 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8575 so they no longer are missing under -DNOPROTO.
8576 [Soren S. Jorvang <soren@t.dk>]
8579 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8581 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8582 doesn't work when the session is reused. Coming soon!
8585 *) Fix a security hole, that allows sessions to be reused in the wrong
8586 context thus bypassing client cert protection! All software that uses
8587 client certs and session caches in multiple contexts NEEDS PATCHING to
8588 allow session reuse! A fuller solution is in the works.
8589 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8591 *) Some more source tree cleanups (removed obsolete files
8592 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8593 permission on "config" script to be executable) and a fix for the INSTALL
8595 [Ulf Moeller <ulf@fitug.de>]
8597 *) Remove some legacy and erroneous uses of malloc, free instead of
8599 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8601 *) Make rsa_oaep_test return non-zero on error.
8602 [Ulf Moeller <ulf@fitug.de>]
8604 *) Add support for native Solaris shared libraries. Configure
8605 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8606 if someone would make that last step automatic.
8607 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8609 *) ctx_size was not built with the right compiler during "make links". Fixed.
8612 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8613 except NULL ciphers". This means the default cipher list will no longer
8614 enable NULL ciphers. They need to be specifically enabled e.g. with
8615 the string "DEFAULT:eNULL".
8618 *) Fix to RSA private encryption routines: if p < q then it would
8619 occasionally produce an invalid result. This will only happen with
8620 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8623 *) Be less restrictive and allow also `perl util/perlpath.pl
8624 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8625 because this way one can also use an interpreter named `perl5' (which is
8626 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8627 installed as `perl').
8628 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8630 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8631 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8633 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8634 advapi32.lib to Win32 build and change the pem test comparision
8635 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8636 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8637 and crypto/des/ede_cbcm_enc.c.
8640 *) DES quad checksum was broken on big-endian architectures. Fixed.
8643 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8644 Win32 test batch file so it (might) work again. The Win32 test batch file
8645 is horrible: I feel ill....
8648 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8649 in e_os.h. Audit of header files to check ANSI and non ANSI
8650 sections: 10 functions were absent from non ANSI section and not exported
8651 from Windows DLLs. Fixed up libeay.num for new functions.
8654 *) Make `openssl version' output lines consistent.
8655 [Ralf S. Engelschall]
8657 *) Fix Win32 symbol export lists for BIO functions: Added
8658 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8659 to ms/libeay{16,32}.def.
8660 [Ralf S. Engelschall]
8662 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8663 fine under Unix and passes some trivial tests I've now added. But the
8664 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8665 added to make sure no one expects that this stuff really works in the
8666 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8667 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8669 [Ralf S. Engelschall]
8671 *) Fix the generation of two part addresses in perl.
8672 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8674 *) Add config entry for Linux on MIPS.
8675 [John Tobey <jtobey@channel1.com>]
8677 *) Make links whenever Configure is run, unless we are on Windoze.
8680 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8681 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8685 *) Add a useful kludge to allow package maintainers to specify compiler and
8686 other platforms details on the command line without having to patch the
8687 Configure script everytime: One now can use ``perl Configure
8688 <id>:<details>'', i.e. platform ids are allowed to have details appended
8689 to them (seperated by colons). This is treated as there would be a static
8690 pre-configured entry in Configure's %table under key <id> with value
8691 <details> and ``perl Configure <id>'' is called. So, when you want to
8692 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8693 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8694 now, which overrides the FreeBSD-elf entry on-the-fly.
8695 [Ralf S. Engelschall]
8697 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8700 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8701 on the `perl Configure ...' command line. This way one can compile
8702 OpenSSL libraries with Position Independent Code (PIC) which is needed
8703 for linking it into DSOs.
8704 [Ralf S. Engelschall]
8706 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8710 *) Cleaned up the LICENSE document: The official contact for any license
8711 questions now is the OpenSSL core team under openssl-core@openssl.org.
8712 And add a paragraph about the dual-license situation to make sure people
8713 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8714 to the OpenSSL toolkit.
8715 [Ralf S. Engelschall]
8717 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8718 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8719 Additonally cleaned up the `make links' target: Remove unnecessary
8720 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8721 to speed processing and no longer clutter the display with confusing
8722 stuff. Instead only the actually done links are displayed.
8723 [Ralf S. Engelschall]
8725 *) Permit null encryption ciphersuites, used for authentication only. It used
8726 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8727 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8731 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8732 signed attributes when verifying signatures (this would break them),
8733 the detached data encoding was wrong and public keys obtained using
8734 X509_get_pubkey() weren't freed.
8737 *) Add text documentation for the BUFFER functions. Also added a work around
8738 to a Win95 console bug. This was triggered by the password read stuff: the
8739 last character typed gets carried over to the next fread(). If you were
8740 generating a new cert request using 'req' for example then the last
8741 character of the passphrase would be CR which would then enter the first
8745 *) Added the new `Includes OpenSSL Cryptography Software' button as
8746 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8747 button and can be used by applications based on OpenSSL to show the
8748 relationship to the OpenSSL project.
8749 [Ralf S. Engelschall]
8751 *) Remove confusing variables in function signatures in files
8752 ssl/ssl_lib.c and ssl/ssl.h.
8753 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8755 *) Don't install bss_file.c under PREFIX/include/
8756 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8758 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8759 functions that return function pointers and has support for NT specific
8760 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8761 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8762 unsigned to signed types: this was killing the Win32 compile.
8765 *) Add new certificate file to stack functions,
8766 SSL_add_dir_cert_subjects_to_stack() and
8767 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8768 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8769 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8770 This means that Apache-SSL and similar packages don't have to mess around
8771 to add as many CAs as they want to the preferred list.
8774 *) Experiment with doxygen documentation. Currently only partially applied to
8776 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8777 openssl.doxy as the configuration file.
8780 *) Get rid of remaining C++-style comments which strict C compilers hate.
8781 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8783 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8784 compiled in by default: it has problems with large keys.
8787 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8788 DH private keys and/or callback functions which directly correspond to
8789 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8790 is needed for applications which have to configure certificates on a
8791 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8793 For the RSA certificate situation is makes no difference, but
8794 for the DSA certificate situation this fixes the "no shared cipher"
8795 problem where the OpenSSL cipher selection procedure failed because the
8796 temporary keys were not overtaken from the context and the API provided
8797 no way to reconfigure them.
8798 The new functions now let applications reconfigure the stuff and they
8799 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8800 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8801 non-public-API function ssl_cert_instantiate() is used as a helper
8802 function and also to reduce code redundancy inside ssl_rsa.c.
8803 [Ralf S. Engelschall]
8805 *) Move s_server -dcert and -dkey options out of the undocumented feature
8806 area because they are useful for the DSA situation and should be
8807 recognized by the users.
8808 [Ralf S. Engelschall]
8810 *) Fix the cipher decision scheme for export ciphers: the export bits are
8811 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8812 SSL_EXP_MASK. So, the original variable has to be used instead of the
8813 already masked variable.
8814 [Richard Levitte <levitte@stacken.kth.se>]
8816 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8817 [Richard Levitte <levitte@stacken.kth.se>]
8819 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8820 from `int' to `unsigned int' because it's a length and initialized by
8821 EVP_DigestFinal() which expects an `unsigned int *'.
8822 [Richard Levitte <levitte@stacken.kth.se>]
8824 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8825 script. Instead use the usual Shell->Perl transition trick.
8826 [Ralf S. Engelschall]
8828 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8829 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8830 -noout -modulus' as it's already the case for `openssl rsa -noout
8831 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8832 currently the public key is printed (a decision which was already done by
8833 `openssl dsa -modulus' in the past) which serves a similar purpose.
8834 Additionally the NO_RSA no longer completely removes the whole -modulus
8835 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8837 [Ralf S. Engelschall]
8839 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8840 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8841 [Arne Ansper <arne@ats.cyber.ee>]
8843 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8844 to be added. Now both 'req' and 'ca' can use new objects defined in the
8848 *) Add cool BIO that does syslog (or event log on NT).
8849 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8851 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8852 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8853 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8854 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8857 *) Add preliminary config info for new extension code.
8860 *) Make RSA_NO_PADDING really use no padding.
8861 [Ulf Moeller <ulf@fitug.de>]
8863 *) Generate errors when private/public key check is done.
8866 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8867 for some CRL extensions and new objects added.
8870 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8871 key usage extension and fuller support for authority key id.
8874 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8875 padding method for RSA, which is recommended for new applications in PKCS
8876 #1 v2.0 (RFC 2437, October 1998).
8877 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8878 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8879 against Bleichbacher's attack on RSA.
8880 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8883 *) Updates to the new SSL compression code
8884 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8886 *) Fix so that the version number in the master secret, when passed
8887 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8888 (because the server will not accept higher), that the version number
8889 is 0x03,0x01, not 0x03,0x00
8890 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8892 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8893 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8894 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8897 *) Support for RAW extensions where an arbitrary extension can be
8898 created by including its DER encoding. See apps/openssl.cnf for
8902 *) Make sure latest Perl versions don't interpret some generated C array
8903 code as Perl array code in the crypto/err/err_genc.pl script.
8904 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8906 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8907 not many people have the assembler. Various Win32 compilation fixes and
8908 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8912 *) Modify configure script 'Configure' to automatically create crypto/date.h
8913 file under Win32 and also build pem.h from pem.org. New script
8914 util/mkfiles.pl to create the MINFO file on environments that can't do a
8915 'make files': perl util/mkfiles.pl >MINFO should work.
8918 *) Major rework of DES function declarations, in the pursuit of correctness
8919 and purity. As a result, many evil casts evaporated, and some weirdness,
8920 too. You may find this causes warnings in your code. Zapping your evil
8921 casts will probably fix them. Mostly.
8924 *) Fix for a typo in asn1.h. Bug fix to object creation script
8925 obj_dat.pl. It considered a zero in an object definition to mean
8926 "end of object": none of the objects in objects.h have any zeros
8927 so it wasn't spotted.
8928 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8930 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8931 Masking (CBCM). In the absence of test vectors, the best I have been able
8932 to do is check that the decrypt undoes the encrypt, so far. Send me test
8933 vectors if you have them.
8936 *) Correct calculation of key length for export ciphers (too much space was
8937 allocated for null ciphers). This has not been tested!
8940 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8941 message is now correct (it understands "crypto" and "ssl" on its
8942 command line). There is also now an "update" option. This will update
8943 the util/ssleay.num and util/libeay.num files with any new functions.
8945 perl util/mkdef.pl crypto ssl update
8946 it will update them.
8949 *) Overhauled the Perl interface (perl/*):
8950 - ported BN stuff to OpenSSL's different BN library
8951 - made the perl/ source tree CVS-aware
8952 - renamed the package from SSLeay to OpenSSL (the files still contain
8953 their history because I've copied them in the repository)
8954 - removed obsolete files (the test scripts will be replaced
8955 by better Test::Harness variants in the future)
8956 [Ralf S. Engelschall]
8958 *) First cut for a very conservative source tree cleanup:
8959 1. merge various obsolete readme texts into doc/ssleay.txt
8960 where we collect the old documents and readme texts.
8961 2. remove the first part of files where I'm already sure that we no
8962 longer need them because of three reasons: either they are just temporary
8963 files which were left by Eric or they are preserved original files where
8964 I've verified that the diff is also available in the CVS via "cvs diff
8965 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8966 the crypto/md/ stuff).
8967 [Ralf S. Engelschall]
8969 *) More extension code. Incomplete support for subject and issuer alt
8970 name, issuer and authority key id. Change the i2v function parameters
8971 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8972 what that's for :-) Fix to ASN1 macro which messed up
8973 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8976 *) Preliminary support for ENUMERATED type. This is largely copied from the
8980 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8981 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8983 *) Make sure `make rehash' target really finds the `openssl' program.
8984 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8986 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8987 like to hear about it if this slows down other processors.
8990 *) Add CygWin32 platform information to Configure script.
8991 [Alan Batie <batie@aahz.jf.intel.com>]
8993 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8994 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8996 *) New program nseq to manipulate netscape certificate sequences
8999 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9003 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9004 but the BN code had some problems that would cause failures when
9005 doing certificate verification and some other functions.
9006 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9008 *) Add ASN1 and PEM code to support netscape certificate sequences.
9011 *) Add ASN1 and PEM code to support netscape certificate sequences.
9014 *) Add several PKIX and private extended key usage OIDs.
9017 *) Modify the 'ca' program to handle the new extension code. Modify
9018 openssl.cnf for new extension format, add comments.
9021 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9022 and add a sample to openssl.cnf so req -x509 now adds appropriate
9026 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9027 error code, add initial support to X509_print() and x509 application.
9030 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9031 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9032 stuff is currently isolated and isn't even compiled yet.
9035 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9036 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9037 Removed the versions check from X509 routines when loading extensions:
9038 this allows certain broken certificates that don't set the version
9039 properly to be processed.
9042 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9043 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9044 can still be regenerated with "make depend".
9047 *) Spelling mistake in C version of CAST-128.
9048 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9050 *) Changes to the error generation code. The perl script err-code.pl
9051 now reads in the old error codes and retains the old numbers, only
9052 adding new ones if necessary. It also only changes the .err files if new
9053 codes are added. The makefiles have been modified to only insert errors
9054 when needed (to avoid needlessly modifying header files). This is done
9055 by only inserting errors if the .err file is newer than the auto generated
9056 C file. To rebuild all the error codes from scratch (the old behaviour)
9057 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9058 or delete all the .err files.
9061 *) CAST-128 was incorrectly implemented for short keys. The C version has
9062 been fixed, but is untested. The assembler versions are also fixed, but
9063 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9064 to regenerate it if needed.
9065 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9066 Hagino <itojun@kame.net>]
9068 *) File was opened incorrectly in randfile.c.
9069 [Ulf Möller <ulf@fitug.de>]
9071 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9072 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9073 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9074 al: it's just almost always a UTCTime. Note this patch adds new error
9075 codes so do a "make errors" if there are problems.
9078 *) Correct Linux 1 recognition in config.
9079 [Ulf Möller <ulf@fitug.de>]
9081 *) Remove pointless MD5 hash when using DSA keys in ca.
9082 [Anonymous <nobody@replay.com>]
9084 *) Generate an error if given an empty string as a cert directory. Also
9085 generate an error if handed NULL (previously returned 0 to indicate an
9086 error, but didn't set one).
9087 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9089 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9092 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9093 parameters. This was causing a warning which killed off the Win32 compile.
9096 *) Remove C++ style comments from crypto/bn/bn_local.h.
9097 [Neil Costigan <neil.costigan@celocom.com>]
9099 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9100 based on a text string, looking up short and long names and finally
9101 "dot" format. The "dot" format stuff didn't work. Added new function
9102 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9103 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9104 OID is not part of the table.
9107 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9108 X509_LOOKUP_by_alias().
9111 *) Sort openssl functions by name.
9114 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9115 encryption from sample DSA keys (in case anyone is interested the password
9119 *) Make _all_ *_free functions accept a NULL pointer.
9120 [Frans Heymans <fheymans@isaserver.be>]
9122 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9124 [Anonymous <nobody@replay.com>]
9126 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9127 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9129 *) Don't blow it for numeric -newkey arguments to apps/req.
9130 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9132 *) Temp key "for export" tests were wrong in s3_srvr.c.
9133 [Anonymous <nobody@replay.com>]
9135 *) Add prototype for temp key callback functions
9136 SSL_CTX_set_tmp_{rsa,dh}_callback().
9139 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9140 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9143 *) X509_name_add_entry() freed the wrong thing after an error.
9144 [Arne Ansper <arne@ats.cyber.ee>]
9146 *) rsa_eay.c would attempt to free a NULL context.
9147 [Arne Ansper <arne@ats.cyber.ee>]
9149 *) BIO_s_socket() had a broken should_retry() on Windoze.
9150 [Arne Ansper <arne@ats.cyber.ee>]
9152 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9153 [Arne Ansper <arne@ats.cyber.ee>]
9155 *) Make sure the already existing X509_STORE->depth variable is initialized
9156 in X509_STORE_new(), but document the fact that this variable is still
9157 unused in the certificate verification process.
9158 [Ralf S. Engelschall]
9160 *) Fix the various library and apps files to free up pkeys obtained from
9161 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9164 *) Fix reference counting in X509_PUBKEY_get(). This makes
9165 demos/maurice/example2.c work, amongst others, probably.
9166 [Steve Henson and Ben Laurie]
9168 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9169 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9170 are no longer created. This way we have a single and consistent command
9171 line interface `openssl <command>', similar to `cvs <command>'.
9172 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9174 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9175 BIT STRING wrapper always have zero unused bits.
9178 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9181 *) Make the top-level INSTALL documentation easier to understand.
9184 *) Makefiles updated to exit if an error occurs in a sub-directory
9185 make (including if user presses ^C) [Paul Sutton]
9187 *) Make Montgomery context stuff explicit in RSA data structure.
9190 *) Fix build order of pem and err to allow for generated pem.h.
9193 *) Fix renumbering bug in X509_NAME_delete_entry().
9196 *) Enhanced the err-ins.pl script so it makes the error library number
9197 global and can add a library name. This is needed for external ASN1 and
9198 other error libraries.
9201 *) Fixed sk_insert which never worked properly.
9204 *) Fix ASN1 macros so they can handle indefinite length construted
9205 EXPLICIT tags. Some non standard certificates use these: they can now
9209 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9210 into a single doc/ssleay.txt bundle. This way the information is still
9211 preserved but no longer messes up this directory. Now it's new room for
9212 the new set of documenation files.
9213 [Ralf S. Engelschall]
9215 *) SETs were incorrectly DER encoded. This was a major pain, because they
9216 shared code with SEQUENCEs, which aren't coded the same. This means that
9217 almost everything to do with SETs or SEQUENCEs has either changed name or
9218 number of arguments.
9219 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9221 *) Fix test data to work with the above.
9224 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9225 was already fixed by Eric for 0.9.1 it seems.
9226 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9228 *) Autodetect FreeBSD3.
9231 *) Fix various bugs in Configure. This affects the following platforms:
9235 unixware-2.0-pentium
9239 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9240 before they are needed.
9243 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9247 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9249 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9250 changed SSLeay to OpenSSL in version strings.
9251 [Ralf S. Engelschall]
9253 *) Some fixups to the top-level documents.
9256 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9257 because the symlink to include/ was missing.
9258 [Ralf S. Engelschall]
9260 *) Incorporated the popular no-RSA/DSA-only patches
9261 which allow to compile a RSA-free SSLeay.
9262 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9264 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9265 when "ssleay" is still not found.
9266 [Ralf S. Engelschall]
9268 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9269 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9271 *) Updated the README file.
9272 [Ralf S. Engelschall]
9274 *) Added various .cvsignore files in the CVS repository subdirs
9275 to make a "cvs update" really silent.
9276 [Ralf S. Engelschall]
9278 *) Recompiled the error-definition header files and added
9279 missing symbols to the Win32 linker tables.
9280 [Ralf S. Engelschall]
9282 *) Cleaned up the top-level documents;
9283 o new files: CHANGES and LICENSE
9284 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9285 o merged COPYRIGHT into LICENSE
9286 o removed obsolete TODO file
9287 o renamed MICROSOFT to INSTALL.W32
9288 [Ralf S. Engelschall]
9290 *) Removed dummy files from the 0.9.1b source tree:
9291 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9292 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9293 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9294 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9295 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9296 [Ralf S. Engelschall]
9298 *) Added various platform portability fixes.
9301 *) The Genesis of the OpenSSL rpject:
9302 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9303 Young and Tim J. Hudson created while they were working for C2Net until
9305 [The OpenSSL Project]
9308 Changes between 0.9.0b and 0.9.1b [not released]
9310 *) Updated a few CA certificates under certs/
9313 *) Changed some BIGNUM api stuff.
9316 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9317 DGUX x86, Linux Alpha, etc.
9320 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9321 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9325 *) Add -strparse option to asn1pars program which parses nested
9327 [Dr Stephen Henson <shenson@bigfoot.com>]
9329 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9332 *) DSA fix for "ca" program.
9335 *) Added "-genkey" option to "dsaparam" program.
9338 *) Added RIPE MD160 (rmd160) message digest.
9341 *) Added -a (all) option to "ssleay version" command.
9344 *) Added PLATFORM define which is the id given to Configure.
9347 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9350 *) Extended the ASN.1 parser routines.
9353 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9356 *) Added a BN_CTX to the BN library.
9359 *) Fixed the weak key values in DES library
9362 *) Changed API in EVP library for cipher aliases.
9365 *) Added support for RC2/64bit cipher.
9368 *) Converted the lhash library to the crypto/mem.c functions.
9371 *) Added more recognized ASN.1 object ids.
9374 *) Added more RSA padding checks for SSL/TLS.
9377 *) Added BIO proxy/filter functionality.
9380 *) Added extra_certs to SSL_CTX which can be used
9381 send extra CA certificates to the client in the CA cert chain sending
9382 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9385 *) Now Fortezza is denied in the authentication phase because
9386 this is key exchange mechanism is not supported by SSLeay at all.
9389 *) Additional PKCS1 checks.
9392 *) Support the string "TLSv1" for all TLS v1 ciphers.
9395 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9396 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9399 *) Fixed a few memory leaks.
9402 *) Fixed various code and comment typos.
9405 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9406 bytes sent in the client random.
9407 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob